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
|
|---|---|---|---|---|---|
Mycobacterium tuberculosis ( Mtb ) primarily resides in the lung but can also persist in extrapulmonary sites . Macrophages are considered the prime cellular habitat in all tissues . Here we demonstrate that Mtb resides inside adipocytes of fat tissue where it expresses stress-related genes . Moreover , perigonadal fat of Mtb-infected mice disseminated the infection when transferred to uninfected animals . Adipose tissue harbors leukocytes in addition to adipocytes and other cell types and we observed that Mtb infection induces changes in adipose tissue biology depending on stage of infection . Mice infected via aerosol showed infiltration of inducible nitric oxide synthase ( iNOS ) or arginase 1 ( Arg1 ) -negative F4/80+ cells , despite recruitment of CD3+ , CD4+ and CD8+ T cells . Gene expression analysis of adipose tissue of aerosol Mtb-infected mice provided evidence for upregulated expression of genes associated with T cells and NK cells at 28 days post-infection . Strikingly , IFN-γ-producing NK cells and Mtb-specific CD8+ T cells were identified in perigonadal fat , specifically CD8+CD44-CD69+ and CD8+CD44-CD103+ subpopulations . Gene expression analysis of these cells revealed that they expressed IFN-γ and the lectin-like receptor Klrg1 and down-regulated CD27 and CD62L , consistent with an effector phenotype of Mtb-specific CD8+ T cells . Sorted NK cells expressed higher abundance of Klrg1 upon infection , as well . Our results reveal the ability of Mtb to persist in adipose tissue in a stressed state , and that NK cells and Mtb-specific CD8+ T cells infiltrate infected adipose tissue where they produce IFN-γ and assume an effector phenotype . We conclude that adipose tissue is a potential niche for Mtb and that due to infection CD8+ T cells and NK cells are attracted to this tissue .
In 2015 , tuberculosis ( TB ) affected 10 . 4 million individuals leading to 1 . 8 million deaths globally [1 , 2] . TB is primarily a disease of the lung , which serves as port of entry and site of disease manifestation . In the lung , Mycobacterium tuberculosis ( Mtb ) is preferentially entrapped in granulomas [3] . Approximately one-third of the global population suffers from latent TB infection , which persists without apparent clinical signs of disease , and which can be reactivated to active TB at later time points [4] . The niches where Mtb persists remain incompletely understood and both pulmonary and extrapulmonary sites have been proposed [5] . Adipose tissue , which harbors pathogens such as Trypanosoma cruzi [6 , 7] constitutes 15–25% of the total body mass [8] and is a rich source of hormones and inflammatory cytokines that participate in host defence against infectious agents [6 , 9] . In vitro studies revealed that murine adipocytes release TNF , IL-6 , IL-12p40 and IL-10 upon Mtb infection [10] . These cells are also susceptible to infection with Chlamydia pneumoniae , influenza A , respiratory syncytial virus [11] and human immunodeficiency virus ( HIV ) [12 , 13] . While some epidemiological studies suggest that obesity is inversely associated with TB [14–16] , others have found a positive genetic association between the two diseases [17] . Obesity and changing patterns of diet are associated with type 2 diabetes , and complex interrelations between nutrition , obesity , diabetes , and TB are increasingly appreciated [18] . Therefore , we embarked on a systematic in vivo study towards better understanding of the role of adipose tissue in Mtb infection . Resting macrophages are the preferred habitat of Mtb , which turn into effector cells after appropriate activation . Adipose tissue includes diverse cell types such as monocytes , F4/80+ macrophages [19] , CD4+ and CD8+ T cells [20 , 21] , endothelial cells , and vascular smooth muscle cells [6] and proportions of these cell populations vary under different pathophysiologic conditions [19 , 20] . Here , we identified Mtb in adipocytes of fat tissue after aerosol infection of mice and expression of stress-related genes in Mtb within human and mouse adipocytes . We also demonstrated the capacity of adipose tissue to carry Mtb when transferred to uninfected animals . Finally , we identified Mtb-specific effector CD8+ T cells and NK cells expressing IFN-γ in adipose tissue after aerogenic Mtb infection . We conclude that adipose tissue provides a potential sanctuary for Mtb in vivo and that Mtb persistence markedly affects adipose tissue biology .
When human and murine adipocytes were cultured with Mtb in vitro , approximately 80% of the bacterial inoculum was engulfed within 24 h ( Fig 1A , 1B and 1C ) , consistent with published data on Mtb uptake by mouse adipocytes [5] . In a control experiment , professional phagocytes , such as the human macrophage cell line THP-1 , took up nearly 98% bacteria under comparable conditions . To determine whether Mtb can replicate inside human and murine adipocytes we counted CFUs between 4 h and 6 days post-infection in presence or absence of the cell-impermeable antibiotic amikacin . Over time , numbers of CFUs remained constant indicating that Mtb did not replicate inside adipocytes ( Fig 1D , 1E and 1F ) , again corroborating previous findings [5] . During persistence in the host under stress conditions [22] , Mtb becomes dormant and expresses a specific gene program under the control of the dormancy survival regulon ( dosR , Rv3133c ) [23 , 24] . These genes include hspX ( Rv2031 ) , which encodes alpha-crystallin , and lat ( Rv3290c ) encoding a lysine aminotransferase [25] , which are considered as mycobacterial stress markers [26] . In Mtb-infected human adipocytes after 48 h , both dosR and lat were upregulated ( Fig 1G ) while in the murine cell line 3T3-L1 , dosR and hspX were induced ( Fig 1H ) . We note that relative expression levels varied between experiments , likely reflecting biological variations in commitment to dormancy and conclude that Mtb is internalized by human and mouse adipocytes where it ceases from replication and becomes stressed . We next evaluated whether Mtb resides in adipose tissue after aerosol infection of mice . For these experiments we evaluated the Mtb-CFUs present in the whole perigonadal fat pad , which was extensively washed to exclude possible blood contamination . Perigonadal fat pad was evaluated because obesity studies have reported its association with insulin resistance [27 , 28] . At 14 and 28 days post-infection , CFUs of Mtb were not only detected in lung and spleen but also in perigonadal fat of infected mice ( Fig 2A ) . These organs also harbored Mtb at 56 and 90 days post-infection ( S1A Fig ) . Mtb was detectable in perigonadal fat of 17% to 83% of mice at 14 and 28 days post-infection throughout the experiments , while dissemination from lung to spleen was observed in all animals at these time points . The abundance of Mtb in perigonadal fat tended to increase with inoculum dose via aerosol ( 50 CFUs: no colonies; 200 CFUs: logCFUs from 0 . 1 to 0 . 9; 4 out of 6 animals showed colonies ) ( Fig 2B ) . Mtb-CFUs were also observed in subcoutaneous fat after aerosol infection ( S1B Fig ) . In addition to adipocytes , fat tissue contains numerous cell types including macrophages [19] and T cells [20 , 21] , which together form the stromal vascular fraction ( SVF ) . When adipose fraction and SVF were mechanically and enzymatically separated , Mtb was identified both in the adipose fraction and in the SVF between 14 and 28 days post-infection ( Fig 2C ) indicating that adipose cells as well as leukocytes present in the fat tissue harbor Mtb . To investigate whether Mtb residing in perigonadal fat had entered a stressed state we evaluated the expression of the latency genes dosR , hspX and lat . Importantly , Mtb residing in the adipose fraction and in SFV expressed the dosR , hspX and lat genes ( Fig 2D ) , indicating activation of the stress-related program . The variability in the extent of the gene responses between adipose fraction and SVF could indicate different metabolic adaptations of Mtb present in those tissues . PCR reaction directed at the Mtb-specific IS6110 insertion sequence confirmed the presence of Mtb in adipose tissue from aerosol-infected mice and revealed again considerable variations between individual animals ( Fig 2E ) . Similar to experiments in which Mtb from perigonadal fat was enumerated , in these experiments on average approximately 50% of samples were positive for Mtb . Confocal microscopy of perigonadal fat from i . v . infected mice consistently confirmed Mtb residence inside adipose tissue in vivo ( Fig 2F ) . To further validate the presence of Mtb in adipose tissue of mice we infected prospective donor animals i . v . Fourteen days after infection we collected the perigonadal fat , washed it extensively to avoid blood contamination , homogenized it and injected it subcutaneously to naïve recipient mice . Mtb was detected in lung , spleen , perigonadal and subcutaneous fat from previously naïve recipient mice confirming that Mtb had been present in adipose tissue of donor animals ( Fig 2G ) . Mtb was not only found in the spleen and lung of donor mice ( S1C Fig ) but also in perigonadal fat of control mice infected at the same time as donors ( S1D Fig ) . Adipose tissue composition varies depending on metabolic changes . However , little is known about alterations in adipose tissue during infectious diseases including TB . Immunohistologic examination at day 28 post-aerosol infection revealed that F4/80+ macrophages , CD3+ , CD4+ and CD8+ cells localized in perigonadal fat ( Fig 3A and 3C ) , whereas no positive staining for myeloperoxidase ( MPO ) abundantly expressed in neutrophils , or for the pan B cell marker B220 was observed ( Fig 3C ) . Surprisingly , neither arginase 1 ( Arg1 ) ( M2 type ) nor inducible nitric oxide synthase ( iNOS ) ( M1 type ) could be detected in F4/80+ macrophages infiltrating perigonadal fat ( Fig 3B and 3C ) indicating that other cell types expressed these enzymes in adipose tissue of infected animals . The higher numbers of F4/80+ macrophages present in perigonadal fat after aerosol infection point to these cells as likely hosts of Mtb in the SFV as depicted in Fig 2C . Additionally , perigonadal fat tissue from Mtb-infected mice exhibited a marked staining for CD3 indicating the presence of T cells and both CD4+ and CD8+ T cell subsets outnumbered those in uninfected tissue ( Fig 3A and 3C ) . At day 14 post-infection , no signs of leukocyte infiltration were noted . Immunohistology of infected tissue also revealed that adipocyte size did not change after infection ( S2A Fig ) , which is in agreement with the absence of higher abundance of free fatty acids in serum of infected animals ( S2B Fig ) . In summary , different leukocyte populations infiltrate perigonadal fat after aerogenic Mtb infection indicating phenotypic alterations in this tissue . We determined the global gene expression profile of perigonadal adipose tissue after Mtb aerosol infection . The global transcription profile did not reveal statistically significant changes at day 14 post-infection , ( Fig 4A and 4B ) whereas at day 28 , statistically significant differences in gene expression were observed ( Fig 4A and 4B ) . Several differentially regulated genes support the presence of CD8+ T cells in perigonadal fat of infected mice ( Fig 4A ) , including Cd3d , Cd8a , Ifng , Gzmb , Fasl , Il2rg , Lat , Tbx21 and Tcrb-j . Increased transcription of Il12rb1 , Itgal , Cd274 , Clec9a , Clec7 , Il18bp and Tnf suggested a myeloid signature and the upregulated genes Nkg7 , Nrc1 , Klrg1 and Klrk1 are characteristic of the NK cell lineage ( Fig 4A ) . In addition , MHC-I and MHC-II ( Tap1 , H2-q2 , H2-m3 , Uba7 , Ube2l6 , H2-dmb1 , Cd74 ) , Toll-like receptor ( TLR ) pathways and chemokines/chemokine receptors were also differentially regulated ( Fig 4B ) including Ccl5 , Ccl8 , Ccl4 , Cxcl9 , Ccr5 , Ccr7 , Cxcr3 and Cxcr6 . The most highly upregulated genes comprised IFN-γ-regulated genes such as Stat1 , Irf9 , Igtp , Irgm1 and Gbp6 ( Fig 4B ) . Since differential metabolism of adipose tissue between males and females has been reported [29] , female and male mice were analysed separately . Even though responses were comparable between genders , males and females expressed comparable transcription profiles in qualitative terms , with females showing slightly stronger responses ( Fig 4A and 4B ) . Taken together the gene expression profile of perigonadal fat in response to Mtb aerosol infection revealed marked prevalence of T cell activation-associated genes ( Fig 4A ) . To validate these results , qPCR analyses of selected genes were performed in samples from female mice . No significant changes were observed in perigonadal fat and in the lung at day 14 post-infection with the exception of IFN-γ in the latter organ ( S3A Fig ) . In contrast , at day 28 , genes associated with T cells , including Cd8a , Ccr7 and Tbx21 were highly upregulated in perigonadal fat of Mtb-infected mice together with cytokines and chemokines , including genes encoding IFN-γ , CCL5 and CXCL9 and the chemokine receptor CXCR3 ( Fig 5A ) . Gene expression pattern in perigonadal fat at day 28 post-infection was distinct from subcutaneous fat where only Cd3d was upregulated ( Fig 5A ) . This profile also diverged from that observed in lungs of infected mice , where Cd4 , rather than Cd8a , was highly expressed while Ccr7 remained unaltered ( Fig 5A ) . We extended our analysis to later time points post-infection with Mtb . Expression of Cd3d , Cd8a , Ifn , Ccl5 and Tbx21 remained highly upregulated 56 days post-infection whereas Ccr7 did not ( S3B Fig ) . A similar pattern was observed in the lung where in contrast to perigonadal fat Cd4 was also upregulated ( S3B Fig ) . Finally , we characterized the gene expression of adipocytes and SVF separately . As expected , the SVF was enriched in leukocyte-associated genes ( Fig 5B ) while only Ccl5 was enriched in the adipocyte fraction of Mtb-infected mice ( Fig 5B ) . Thus , changes in gene expression in perigonadal fat induced by Mtb were primarily a consequence of leukocyte infiltration and this infiltration was tissue-specific since subcutaneous fat or lung showed a different gene expression pattern . Even though Mtb was not detected in adipose tissue of all infected mice ( Fig 2 ) , gene expression patterns in perigonadal fat after infection was similar in all animals ( Fig 5 ) . For an in-depth characterization of leukocytes infiltrating perigonadal fat , we performed FACS analysis of the SVF at different time points after Mtb aerogenic infection . At day 14 post-infection , numbers of CD4+ and CD8+ T cells in perigonadal fat remained unaltered in contrast to lung tissue where higher numbers of both populations were identified ( S4 Fig ) . Rare CD8+ CD44+ TB10 . 4+ T cells specific for a representative Mtb antigen were not detected in any of the tissues ( S4 Fig ) consistent with the notion that antigen-specific T cells at this time point do not accumulate in the lung in sufficient numbers to limit Mtb replication [30] . Similarly , at day 14 post-infection , numbers of CD4+ , CD8+ or TCRβ– NK1 . 1+ ( NK ) cells producing IFN-γ in the perigonadal fat or lung remained unaltered ( S5 and S6 Figs ) . Numbers of IL-4-producing cells did not change in perigonadal fat during infection whereas in the lung the number of IL-4-producing NK cells was reduced ( S5 and S6 Figs ) . In agreement with the gene expression analysis , at day 28 post-infection , the number of total CD8+ T cells as well as Mtb-specific CD8+ T cells ( CD8+ CD44+ TB10 . 4+ ) in perigonadal fat increased while the number of CD4+ T cells remained unaltered in contrast to the lung ( Fig 6 ) . Mtb-specific CD8+ T cells represented an average of 25 . 5% of the total CD8+ T cell population ( 25 . 5 ± 6 . 9% , n = 12 ) indicating that perigonadal fat became highly enriched in Mtb-specific CD8+ T cells . During infection , CD8+ T cells in perigonadal fat expressed an effector phenotype ( CD8+ CD44– CD69+ ) that was also seen in the lung where , in addition , CD4+ T cells with similar characteristics were identified ( S7 Fig ) . It is known that the αE integrin CD103 is expressed in pathogen-specific CD8+ T cells in peripheral tissues [31]; here a cell population of CD8+ CD44– CD103+ phenotype was identified in perigonadal fat but not in lung of infected mice ( S8 Fig ) . FACS analysis identified NK cells as major source of IFN-γ ( Fig 7 ) . In contrast , in the lungs , CD4+ and CD8+ T cells were major IFN-γ producers in addition to NK cells ( S9 Fig ) , and these populations also produced IL-4 . In summary , at day 28 post-infection , perigonadal fat from Mtb-infected mice harbored higher numbers of total CD8+ T cells displaying an activated phenotype as well as Mtb-specific CD8+ T cells and IFN-γ-producing NK cells . At day 28 post-aerosol infection , Mtb-specific CD8+ T cells and NK cells from perigonadal fat were sorted and selected genes were evaluated simultaneously ( S10A and S10B Fig ) . Sorted total CD8+ T cells and NK cells were sorted from uninfected mice served as controls . The TB10 . 4 tetramer was selected as a representative antigen for sorting infiltrating Mtb-specific CD8+ T cells from infected mice [32] . We are aware that other Mtb specificities were still to be expected in the CD8+ TB10 . 4- population . Therefore , we opted to compare CD8+ TB10 . 4+ cells from infected mice to total CD8+ cells from uninfected ones . TB10 . 4-specific CD8+ T cells from Mtb-infected mice expressed higher levels of Ifng than total CD8+ T cells from uninfected mice while the levels of Tnf , ( Fig 8A ) and Tgfb1 ( S10C Fig ) were unaltered . In contrast , expression of the cytokines Il10 and Il17a in TB10 . 4-specific CD8+ T cells , total CD8+ T cells of uninfected mice and NK cells from perigonadal fat were low ( S10A Fig ) . Among transcription factors evaluated , TB10 . 4-specific CD8+ T cells from perigonadal fat expressed lower abundance of Rorc and Eomes than CD8+ T cells from uninfected mice ( Fig 8A ) whereas in the lungs these cells showed lower Gata3 levels instead ( Fig 8B ) . No significant differences were observed for Tbx21 ( T-bet ) in TB10 . 4-specific CD8+ T cells and NK cells from infected mice in perigonadal fat or lung possibly due to high variations between samples ( Fig 8A and 8B ) . Microarray analyses revealed that several chemokine and chemokine receptor genes were differentially regulated in perigonadal fat upon infection ( Fig 4B ) . Expression levels of the Ccl5 transcript in total CD8+ and TB10 . 4-specific CD8+ T cells of perigonadal fat were comparable ( Fig 8A ) . In contrast , in the lung TB10 . 4-specific CD8+ T cells expressed a higher abundance of Ccl5 than total CD8+ T cells from uninfected animals ( Fig 8B ) . Ccr5 and Cxcr3 expression was also lower in TB10 . 4-specific CD8+ T cells from perigonadal fat and Ccr7 remained unaffected after infection ( Fig 8A ) . In the lung , the levels of Cxcr6 were marginally higher in TB10 . 4-specific CD8+ T cells ( S10D Fig ) . Surface expression of CD62L is down-regulated on T cells after antigen-specific activation [33] . Accordingly , Mtb-specific CD8+ T cells from perigonadal fat and lung exhibited lower levels of the CD62L transcript ( Sell ) than total CD8+ T cells from uninfected mice . Consistent with an effector rather than a memory phenotype , TB10 . 4-specific CD8+ T cells in perigonadal fat and lung showed lower levels of transcription of Sell ( CD62L ) than cells from uninfected mice ( Fig 8A and 8B ) . On the other hand , Cd69 , Fasl ( Fig 8A and 8B ) , Cd44 , and Icos ( S10C and S10D Fig ) were not affected while the MHC-I family member H2-q9 expressed lower levels of transcription in cells from perigonadal fat ( S10C Fig ) . Transcripts of the IFN-γ-induced molecules Irgm1 , Irf9 and Tap1 , as well as Zap70 , Ube4b and Nkg7 were not affected by Mtb infection in TB10 . 4-specific CD8+ T cells ( S10C Fig ) . Killer cell lectin-like receptor subfamily G member 1 ( KLRG-1 ) is an inhibitory C-type lectin expressed on NK cells and activated CD8+ T cells and a marker for terminally differentiated effector lymphocytes [34] . Both TB10 . 4-specific CD8+ T cells and NK cells in perigonadal fat showed higher abundance of Klrg1 after Mtb infection ( Fig 8A ) . In addition , NK cells from perigonadal fat upon Mtb infection exhibited higher levels of the integrin Itgam ( CD11b ) ( Fig 8A ) . The appearance of CD11b in murine NK cells corresponds to progressive acquisition of effector functions [35] . Thus , at day 28 post-aerosol infection perigonadal fat tissue was enriched in TB10 . 4-specific CD8+ effector T cells and activated NK cells .
Our study reveals that Mtb enters and persists in adipose tissue where it transitions into a stress mode , and induces influx of NK cells and Mtb-specific effector CD8+ T cells to the site of its residence . We take our findings as evidence that adipose tissue serves as a potential sanctuary for persistent Mtb . Several pathogens exploit adipose tissue for persistence . For example , P . berghei accumulates in mouse adipose tissue in form of schizont-infected erythrocytes [36] and in malaria patients the presence of P . falciparum in adipose tissue is apparently related to parasite survival [37] . T . cruzi , the causative agent of Chagas disease was also detected in biopsies of some infected individuals [7] . Similarly , in our experiments Mtb was identified in adipose tissue of approximately half of aerosol-infected mice and the pathogen was localized in both adipocytes and SVF . It is tempting to speculate that Mtb finds suitable conditions for persistence in adipocytes due to the higher abundance of triacylglycerol in adipocyte lipid droplets , which is critical for onset of dormancy in Mtb [38] . Human and murine adipocytes differed in the stressed-related genes upregulated by Mtb upon infection probably reflecting differences in host-pathogen interactions due to differences between species [39] . Not mutually exclusive , macrophages present in SVF are also a potential host for Mtb . Our experiments show that adipose-tissue resident Mtb is able to colonize lung , spleen and perigonadal fat when transferred to naïve animals pointing out that Mtb is viable despite its stress-related status . Further studies precisely characterizing the gene expression profile of Mtb in adipose tissue are needed for deeper understanding of metabolic changes occurring in the mycobacteria . Our experiments reveal marked impact of Mtb on immune surveillance including leukocyte infiltration in adipose tissue . Nippostrongylus brasiliensis infection affects adipose tissue metabolism by increasing the number of eosinophils in perigonadal fat [40] . Simian immunodeficiency virus ( SIV ) in adipose tissue of macaques causes influx of leukocytes and activated CD4+ T cells have been identified in adipose tissue from HIV+ individuals [13] . In our study F4/80+ macrophages and T cells infiltrated adipose tissue by day 28 of aerogenic Mtb infection while no infiltration was detected at day 14 reflecting the onset of the adaptive immune response by day 21 [41 , 42] . At day 28 almost 25% of the CD8+ T cells infiltrating perigonadal fat were Mtb-specific indicating that these cells are specifically attracted to the tissue after infection . Leukocyte infiltration into adipose tissue has been mostly studied in relation to obesity rather than infection [43 , 44] . Here we characterized the infiltration of adipose tissue of lean mice after infection . Whether the same pattern is observed in obese mice remains to be clarified . Such studies will help provide deeper insights into the relationship between obesity , type 2 diabetes and TB . Global gene expression and fluorescence-activated cell sorting analyses provided compelling evidence for NK and Mtb-specific CD8+ T cells infiltration as well as IFN-γ induction at day 28 post-infection . T cells enter adipose tissue after high fat diet [20] coincident with onset of insulin resistance [44] . In a mouse model of experimental TB , no changes in serum insulin levels were observed upon infection ( uninfected: 1 . 70 ± 0 . 46 ng/ml; Mtb: 4 . 14 ± 2 . 94 ng/ml; data representative of 3 independent experiments; 4 to 5 animals per group ) indicating that CD8+ T cells do not induce insulin resistance . We assume that Mtb-specific CD8+ T cells likely control intracellular Mtb in adipose tissue by virtue of lysis of Mtb-infected cells and IFN-γ secretion [45] . Enrichment of CXCR6 Mtb-specific CD8+ T cells in lung , which correlates with protective immunity [46] is consistent with this assumption . A similar role for IFN-γ-producing NK cells can be inferred since these cells numerically increase in the lung after infection [47] and are also able to lyse infected cells [48 , 49] . Finally , IFN-γ has been associated with phenotypic polarization towards pro-inflammatory M1 cells in adipose tissue in obesity [19] and IFN-γ-producing NK cells have been linked to insulin resistance [50] . Local production of IFN-γ activates macrophages and T cells which , in turn , upregulate IFN-γ-regulated molecules . These include STAT1 and Irgm1 together with interferon-induced Gbp1 and Gbp5 and distinct chemokines including CXCL9 , which amplify the inflammatory response through a feed-forward loop , resulting in chronic immune activation in TB . Gbp1 , Stat1 and Tap1 , which are part of a biosignature for subclinical TB in humans [51] were enriched in perigonadal fat during Mtb infection . CD8+ adipose tissue infiltrating cells expressed CD69 and downregulated CD62L transcripts suggesting that these activated cells contribute to control of Mtb in fat tissue . Interestingly , after Mtb infection we identified a unique population of CD8+ T cells present in perigonadal fat that was absent in the lung: CD8+ CD44– CD103+ . This population plays a balancing role in an inflammatory model of chronic murine ileitis [52] and the αE integrin CD103 is expressed by pathogen-specific CD8+ T cells in peripheral tissues [31] . We therefore feel confident that CD8+ T cells in adipose tissue contribute to regional control of Mtb and regulation of inflammation . In sum , this is the first report to describe persistence of Mtb expressing stress-related genes in adipocytes and the recruitment of activated immune cells to adipose tissue . We also demonstrate that Mtb present in adipose tissue can be transferred to naïve animals . Together , our findings point to adipose tissue as a potential reservoir for persistent Mtb . Better understanding of the role of adipose tissue in Mtb infection will provide the basis for rational intervention measures directed at comorbidity of TB and type 2 diabetes for which strong epidemiological evidence exists [14–18] .
Human adipocytes were obtained from plastic surgery waste tissue from patients of the Department of General , Visceral , Vascular and Thoracic Surgery ( Sub-area: Plastic and Reconstructive Surgery ) , Charité University Medical Center in Berlin according to the regulations of and approval by the Ethics Committee of Charité , University Medical Center , Berlin , Germany ( reference number EA1/249/11 ) . All human samples were anonymized . Animal procedures were performed in accordance with the German “Tierschutzgesetz in der Fassung vom 18 . Mai 2006 ( BGBI . IS . 1207 ) ” and the guideline 2010/63/EU from the European Union and the European Convention for the protection of vertebrate animals used for experimental and other scientific purposes . Animal protocols were approved by the ethics committee and the Berlin state authorities ( LAGeSo , reference number G179/12 ) . The preadipocyte human cell strain Simpson-Golabi-Behmel syndrome ( SGBS ) , kindly provided by Prof . M . Wabitsch ( Department of Pediatrics and Adolescent Medicine , University of Ulm , Germany ) , was cultured and differentiated as described [53] . Human adipose tissue ( see Ethic statements ) was separated from skin , minced and digested with type I collagenase ( Worthington Chemicals ) at 200 U/ml and 3 ml/g in HBSS for 1 h at 37°C under shaking . Samples were centrifuged at 400 g for 10 min , erythrocytes in the pellet were lysed and cells were filtered through 70 μm and 40 μm cell strainers . After a density gradient ( Biochrome ) separation , the interface was collected and cell depletion performed with anti-CD45 ( 5B1; Miltenyi ) , anti-CD31 antibodies ( WM59; Serotec ) and anti-fluorescein isothiocyanate ( FITC ) beads ( Miltenyi ) . Cells were cultured in Dulbecco’s modified Eagle medium ( DMEM ) :F12 ( Gibco ) 50% heat-inactivated fetal calf serum ( FCS; Gibco ) and then in 20% serum for differentiation as described [53] . Murine 3T3-L1 cell line was grown in DMEM 10% FCS ( Gibco ) , 4 mM L-glutamine ( Biochrome ) , 1 mM sodium pyruvate ( Biochrome ) , and 1 . 5 g/L sodium bicarbonate ( Gibco ) until confluency . Differentiation was initiated by addition of 1 μg/ml bovine insulin , 0 . 5 mM 3-isobutyl-1-methylxanthine ( IBMX ) , and 1 μM dexamethasone ( all from Sigma ) . At day 4 , medium was replaced with medium containing 1 μg/ml of bovine insulin only and differentiation was complete at day 8 . Cells were infected with Mtb H37Rv at a multiplicity of infection ( MOI ) of 5 . Bacteria were grown in Middlebrook 7H9 broth ( BD Biosciences ) supplemented with 0 . 05% glycerol , Tween 80 and 10% albumin dextrose catalase ( ADC ) growth supplement ( BD Biosciences ) . Single bacteria resuspended in culture medium were obtained from early log-phase cultures . After 2 h of infection , 200 μg/ml of amikacin ( Sigma ) was added . After 4 h , cells were washed . At different time points , cells were lysed with 0 . 1% Triton X-100 ( ICN Biomedicals ) and plated on Middlebrook 7H10 agar plates . Mtb colonies were enumerated after 3 to 6 weeks of incubation at 37°C . Human or murine adipocytes were infected with Mtb-GFP at MOI 20 for 24 h . Mtb was detected with an anti-Mtb antibody ( Abcam ) and secondary antibody conjugated with Alexa 546 ( Invitrogen ) . Nuclei were stained with 4’ , 6-diamidino-2-phenylindole ( DAPI ) . Small pieces of paraformaldehyde ( PFA ) -fixed Mtb-infected adipose tissue were embedded in low melting point agarose ( 3 . 3% in PBS ) . The agarose was gelled on ice and the block containing the adipose tissue was sectioned at 300 μm thickness using a Leica VT 1000 S vibratome . Sections were stained with the DNA-intercalating dye Draq5 , mounted and analyzed using a Leica TCS5 confocal microscope . Female and male 8- to 10-week-old C57BL/6 mice were kept under specific pathogen-free conditions at the Max Planck Institute for Infection Biology in Berlin , Germany ( see Ethic statements ) . Mtb strain H37Rv was grown in Middlebrook 7H9 broth ( BD Biosciences ) supplemented with 0 . 2% glycerol , 0 . 05% Tween 80 , and 10% ADC enrichment ( BD Biosciences ) until logarithmic growth phase before storage at –80°C . Animals were aerosol-infected with 50–200 colony-forming units ( CFUs ) Mtb , using a Glas-Col inhalation exposure system . At given time points , serial dilutions of lung , spleen or undiluted perigonadal fat homogenates were plated onto Middlebrook 7H11 plates . CFUs were counted after 3 ( lungs and spleen ) or 6 ( perigonadal fat ) weeks of incubation at 37°C . PCR was performed by incubation of the DNA samples with the IS6110 Mtb insertion sequence primers 5’-CGTGAGGGCATCGAGGTGGC-3’ and 5’-GCGTAGGCGTCG GTGACAAA-3’ and the Hot Star Mix ( Qiagen ) . Gene expression microarray studies were carried out with the SurePrint G3 Mouse GE 8×60K Microarray Kit ( Agilent Technologies , product number G4852A ) . Microarray data were deposited in the NCBI’s Gene Expression Omnibus ( GEO accession number GSE83554 ) . Microarrays were background corrected , normalized and statistically analysed with limma [54] with moderated t-test for significance of the factors ( sex and experimental group ) and the interaction between sex and treatments for both day 14 and day 28 samples . P-values were corrected for multiple comparisons with the Benjamini-Hochberg method . Genes were tested for enrichment in functional associations using the R package tmod [55] with the CERNO test with Benjamini-Hochberg correction . Detailed R scripts used in the analysis are available upon request . Perigonadal adipose tissue was placed in DMEM ( Gibco ) with 20 mM HEPES ( Gibco ) and 10 mg/ml free fatty acid–bovine serum albumin ( FFA-BSA; Sigma ) and minced to fine pieces . Samples were extensively washed to eliminate all traces of blood and incubated with 280 U/ml collagenase type I ( Worthington ) , and 50 U/ml DNAse ( Roche ) for 45 min under shaking and passed through a 250 μm mesh ( Pierce ) . Cells were centrifuged at 1 , 000 g for 10 min , and resuspended in PBS , 5% EDTA , 0 . 2% FFA-BSA . Single-cell suspensions from lungs of mice were prepared as previously described [56] . Immune cells were stained with antibodies against T cell receptor-beta chain ( TCRβ ) ( H57-597; BD Biosciences ) , CD4 ( RM4-5; BD Biosciences ) , CD8 ( 56–6 . 7; BD Biosciences ) , CD44 ( IM7 ) , CD69 ( H1 . 2F3; BD Biosciences ) , CD103 ( 2E7; eBioscience ) , NK1 . 1 ( PK136; eBioscience ) , CD62L ( MEL-14; eBioscience ) , IFNγ ( XMG1 . 2; BD ) , and IL-4 ( 11B11; eBioscience ) . H-2Kb:TB10 . 4 ( 4–11: IMYNYPAM ) tetramers were prepared in-house . To stain for intracellular cytokines , cells were incubated with brefeldin A 10 μg/ml , ionomycin 1 μg/ml and phorbol 12 , 13-dibutyrate 50 ng/ml ( all from Sigma ) for 4 h at 37°C , 5% CO2 and permeabilized with Cytofix/Cytoperm kit ( BD Biosciences ) according to manufacturer’s instructions . Cells were acquired on a Canto II flow cytometer ( BD Biosciences ) and analyzed with FACSDiva ( BD Biosciences ) software . For sorting experiments cells from three animals were pooled and stained with H-2Kb:TB10 . 4 ( 4–11: IMYNYPAM ) tetramers , antibodies against TCRβ ( H57-597; BD Biosciences ) , CD4 ( RM4-5; BD Biosciences ) , CD8 ( 56–6 . 7; BD Biosciences ) , CD44 ( IM7 ) and NK1 . 1 ( PK136; eBioscience ) and sorted on a FACSAria II ( BD Biosciences ) . Paraffin sections were dewaxed and stained histochemically with hematoxylin and eosin ( HE ) for overview . For immunohistochemistry , sections were subjected to a heat-induced epitope retrieval step except for the detection of B cells prior to incubation with antibodies against CD3 ( code A0452; Dako ) , B220 ( RA3-6B2; BD Bioscience ) , myeloperoxidase ( MPO; code 9661; Cell Signaling ) , CD4 ( 4SM95; eBioscience ) , or CD8 ( 4SM15; eBioscience ) . This step was followed by incubation with biotinylated secondary antibodies ( Dianova ) . For detection , alkaline phosphatase ( AP ) -labelled streptavidin and chromogen RED ( both Dako ) were employed . For detection of macrophages , sections were subjected to protein-induced epitope retrieval employing protease ( Sigma ) prior to incubation with anti-F4/80 ( BM8; eBioscience ) followed by incubation with biotinylated rabbit anti-rat ( Dako ) secondary antibody . Biotin was detected using AP-labelled streptavidin ( Dako ) and AP was visualized with chromogen RED ( Dako ) . For detection of classically activated ( M1 ) macrophages the sections were subjected to a heat-induced epitope retrieval step prior to incubation with anti-inducible nitric oxide synthase ( iNOS ) ( code ab15323; Abcam , ) . The EnVision+ System , HRP Labelled Polymer Anti-Rabbit ( Dako ) was used for detection . Nuclei were counterstained with hematoxylin ( Merck ) . For the detection of alternatively-activated ( M2 ) macrophages , dewaxed sections were incubated with anti-arginase 1 ( N20; Santa Cruz ) followed by incubation with biotinylated donkey anti-goat ( Dianova ) . For detection AP-labelled streptavidin and chromogen RED ( both Dako ) were employed . After color development , sections were subjected to protein-induced epitope retrieval as described above prior to incubation with anti-F4/80 ( BM8; eBioscience ) . Alexa488-labelled secondary antibody ( Invitrogen ) was used for detection . Nuclei were counterstained with DAPI ( Sigma ) . Negative controls were performed by omitting the primary antibody . Images were acquired using the AxioImager Z1 microscope ( Carl Zeiss MicroImaging ) . All evaluations were performed in a blinded manner . Perigonadal fat samples was collected in TRIzol total RNA isolation reagent ( Invitrogen ) and RNA was isolated as previously described [57] . For mRNA quantification , RNA was reverse-transcribed to cDNA , and qRT-PCR was performed according to manufacturer’s instructions ( BioRad ) TaqMan qRT-PCR assays with specific probes for mouse Actinb ( NM_007393 . 4 ) , Cd8a ( NM_001081110 . 2 ) , Cd3d ( NM_013487 . 3 ) , Cd4 ( NM_013488 . 2 ) , Ifng ( NM_008337 . 3 ) , Il4 ( NM_021283 . 2 ) , Ccl5 ( NM_013653 . 3 ) , Ccr7 ( NM_007719 . 2 ) , Cxcl9 ( NM_008599 . 4 ) , Cxcr3 ( NM_009910 . 3 ) , Tbx21 ( NM_019507 . 2 ) , Rorc ( NM_011281 . 2 ) , Gata3 ( NM_008091 . 3 ) ( Applied Biosystems ) were used . TaqMan probes for the mycobacterial genes sigA ( Rv2703 ) , dosR ( Rv3133c ) , lat ( Rv3290c ) , and hspX ( Rv2031 ) were designed by the manufacturer ( Applied Biosystems ) . Samples from perigonadal fat were pre-amplified with TaqMan PreAmp Master Mix according to manufacturer’s protocol ( Applied Biosystems ) . All probes were normalized to β-actin as internal control ( Applied Biosystems ) , except for the quantification of mycobacterial genes where sigA was the internal control . All fold changes were calculated using the ΔΔCt method [58] and normalized to the lowest value in each group . Amplifications were performed with Step One Plus ( Applied Biosystems ) . Gene expression of sorted CD8+ T cells from uninfected mice , Mtb-specific CD8+ T cells from infected mice and NK cells from uninfected and infected mice were analyzed simultaneously using the 48 . 48 Dynamic Array Integrated Fluidic Circuits ( IFCs; Fluidigm ) . Triplicates of 100 sorted cells were collected in a 96-well PCR plate ( Eppendorf ) containing CellDirect Reaction mix ( Life Technologies ) with Ambion SUPERase-In ( Ambion ) and stored at –80°C . Pre-amplification of genes by reverse transcription and cDNA synthesis ( 18 cycles ) was performed using Cells Direct One-Step qPCR Kit ( Life Technologies ) and TaqMan gene expression assay mix ( Applied Biosystems ) . The cDNA and the single TaqMan assays were then loaded in a microfluidic chip ( Fluidigm ) using Fluidigm 48 . 48 IFC Controller MX according to manufacturer's protocol and quantitative PCR was run using the Data Collection Software ( 36 cycles; Fluidigm ) . Data were exported with the Real-time PCR Analysis Software ( Fluidigm ) and analyzed with Microsoft Office Excel . mRNA amounts were normalized to β-actin ( NM_007393 . 4 ) expression . To compare data from different animals , tissues and chips fold change ( 2–[ ( ∆Ct ) reference−∆Ct ( value ) ] ) in transcripts was calculated relative to splenic CD8+ T cells , which were sorted in each plate as internal reference [59] . The following transcripts were evaluated: Cd8a ( NM_001081110 . 2 ) , Cd3d ( NM_013487 . 3 ) , Cd4 ( NM_013488 . 2 ) , Ifng ( NM_008337 . 3 ) , Il4 ( NM_021283 . 2 ) , Ccl5 ( NM_013653 . 3 ) , Ccr7 ( NM_007719 . 2 ) , Cxcl9 ( NM_008599 . 4 ) , Cxcr3 ( NM_009910 . 3 ) , Tbx21 ( NM_019507 . 2 ) , Rorc ( NM_011281 . 2 ) , Gata3 ( NM_008091 . 3 ) , Eomes ( NM_001164789 . 1 ) , Serpine1 ( NM_008871 . 2 ) , Ube4b ( NM_022022 . 3 ) , Fasl ( NM_001205243 . 1 ) , Cxcr6 ( NM_030712 . 4 ) , Tnf ( NM_001278601 . 1 ) , Il10 ( NM_010548 . 2 ) , Il2 ( NM_008366 . 3 ) , Tlr2 ( NM_011905 . 3 ) , Irgm1 ( NM_008326 . 1 ) , Irf9 ( NM_008394 . 3 ) , Itgax ( NM_021334 . 2 ) , Tgfb1 ( NM_011577 . 1 ) , Il17a ( NM_010552 . 3 ) , Itgae ( NM_008399 . 2 ) , Pdcd1 ( NM_008798 . 2 ) , Tap1 ( NM_001161730 . 1 ) , Nkg7 ( NM_024253 . 4 ) , Foxp3 ( NM_001199347 . 1 ) , Il15 ( NM_008357 . 2 ) , Itgam ( NM_008401 . 2 ) , Cd27 ( NM_001033126 . 2 ) , Sell ( NM_001164059 . 1 ) , Cd44 ( NM_009851 . 2 ) , Ccl2 ( NM_011333 . 3 ) , Ccr5 ( NM_009917 . 5 ) , Zap70 ( NM_009539 . 2 ) , Fcgr1 ( NM_010186 . 5 ) , Icos ( NM_017480 . 2 ) , Cd69 ( NM_001033122 . 3 ) , Klrg1 ( NM_016970 . 1 ) , Ncr1 ( NM_010746 . 3 ) , Cd5 ( NM_007650 . 3 ) , H2-q7/h2-q9 ( NM_001201460 . 1 ) , and Nampt ( NM_021524 . 2 ) ( Applied Biosystems ) . Differences were analyzed using Student’s t-test ( parametric groups ) or Mann–Whitney U test ( nonparametric groups ) . P values <0 . 05 were considered statistically significant .
|
In 2015 , tuberculosis ( TB ) affected 10 . 4 million individuals causing 1 . 8 million deaths per year . Yet , a much larger group– 2 billion people–harbors latent TB infection ( LTBI ) without clinical symptoms , but at lifelong risk of reactivation . The physiological niches of Mycobacterium tuberculosis ( Mtb ) persistence remain incompletely defined and both pulmonary and extrapulmonary sites have been proposed . Adipose tissue constitutes 15–25% of total body mass and is an active production site for hormones and inflammatory mediators . The increasing prevalence of obesity , has led to greater incidence of type 2 diabetes . These patients suffer from three times higher risk of developing TB , pointing to a potential link between adipose tissue and TB pathogenesis . In individuals with LTBI , Mtb survives in a stressed , non-replicating state with low metabolic activity and resting macrophages serve as preferred habitat and become effectors after appropriate stimulation . Here we demonstrate that Mtb can infect and persist within adipocytes where it upregulates stress-related genes . In vivo , relative proportions of leukocyte subsets infiltrating adipose tissue varied under different conditions of infection . During natural aerosol Mtb infection , distinct leukocyte subsets , including mononuclear phagocytes , Mtb-specific CD8+ T cells and NK cells infiltrated adipose tissue and became activated . Thus , our study shows that adipose tissue is not only a potential reservoir for this pathogen but also undergoes significant alteration during TB infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immunology",
"adipocytes",
"connective",
"tissue",
"cells",
"cytotoxic",
"t",
"cells",
"bacteria",
"lipids",
"white",
"blood",
"cells",
"animal",
"cells",
"fats",
"gene",
"expression",
"t",
"cells",
"connective",
"tissue",
"biological",
"tissue",
"actinobacteria",
"biochemistry",
"cell",
"biology",
"anatomy",
"mycobacterium",
"tuberculosis",
"adipose",
"tissue",
"nk",
"cells",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms"
] |
2017
|
Mycobacterium tuberculosis infection modulates adipose tissue biology
|
Relatively little is known about the filarial proteins that interact with the human host . Although the filarial genome has recently been completed , protein profiles have been limited to only a few recombinants or purified proteins of interest . Here , we describe a large-scale proteomic analysis using microcapillary reverse-phase liquid chromatography-tandem-mass spectrometry to identify the excretory-secretory ( ES ) products of the L3 , L3 to L4 molting ES , adult male , adult female , and microfilarial stages of the filarial parasite Brugia malayi . The analysis of the ES products from adult male , adult female , microfilariae ( Mf ) , L3 , and molting L3 larvae identified 852 proteins . Annotation suggests that the functional and component distribution was very similar across each of the stages studied; however , the Mf contributed a higher proportion to the total number of identified proteins than the other stages . Of the 852 proteins identified in the ES , only 229 had previous confirmatory expressed sequence tags ( ESTs ) in the available databases . Moreover , this analysis was able to confirm the presence of 274 “hypothetical” proteins inferred from gene prediction algorithms applied to the B . malayi ( Bm ) genome . Not surprisingly , the majority ( 160/274 ) of these “hypothetical” proteins were predicted to be secreted by Signal IP and/or SecretomeP 2 . 0 analysis . Of major interest is the abundance of previously characterized immunomodulatory proteins such as ES-62 ( leucyl aminopeptidase ) , MIF-1 , SERPIN , glutathione peroxidase , and galectin in the ES of microfilariae ( and Mf-containing adult females ) compared to the adult males . In addition , searching the ES protein spectra against the Wolbachia database resulted in the identification of 90 Wolbachia-specific proteins , most of which were metabolic enzymes that have not been shown to be immunogenic . This proteomic analysis extends our knowledge of the ES and provides insight into the host–parasite interaction .
The parasitic nematodes , Brugia malayi and Wuchereria bancrofti are lymphatic dwelling filariae that induce a spectrum of clinical manifestations ( ranging from clinically asymptomatic [or subclinical] microfilaremia to the deforming elephantiasis ) felt to reflect the nature of the filarial specific immune response . The filarial parasite differentiates into a series of morphologically distinct forms in the vertebrate and mosquito hosts during its life cycle , and is comprised of five major stages , separated by 4 molts . The microfilariae undergo two molts in the mosquito before they develop into the infective L3 stage . Once the L3's gain access to the human host they develop over months into adults after 2 molts . After mating , the females release large numbers of microfilariae that move into the blood stream from which they are taken up by the mosquito to complete the life cycle . Each of the filarial life cycle stages has both unique and common biologic characteristics . Because each stage of parasite development may be antigenically distinct , filarial infections are often characterized by a series of discrete immune responses that evolve at different times during the course of infection . Moreover , each stage of parasite development may also entail a change in tissue tropism introducing a compartmental feature to the responses induced by the filarial parasite . Although the composition of the excretory-secretory ( ES ) products of the filarial parasites is limited and largely uncharacterized , they have been used as a source in the identification of potential diagnostics [1]–[3] and/or vaccine candidates [4] , [5] . Given the potential role of the ES in regulating the host immune system and subsequent pathology associated with the immune response , identifying the individual components of the ES has been of considerable interest . Early studies on the characterization of the ES were limited by both technical/practical constraints along with the low abundance of these proteins often precluding molecular identification [6] , [7] . The availability of high throughput and sensitive techniques along with genomic data has facilitated the identification of these proteins . A recent study using comparative analysis of the soluble extracts of mixed adult Brugia malayi ( BmA ) antigen and the excretory-secretory ( BES ) products of these mixed adults has helped to identify a portion of the secretome contributed by the two adult stages of the parasites [8] . Another study focusing on the stage-specific identification of the ES was reported while this manuscript was under review [9] . The ES data available from these recent studies [8] , [9] were based on separation of the ES by gel-electrophoresis techniques followed by LC-MS/MS . The present study utilized a high-throughput , shot-gun proteomic approach to identify comprehensively the secretome of multiple stages of Brugia malayi starting with the infective L3 larvae , and moving to those molting to L4 , through the adult parasite stages ( male and female individually ) and to the microfilariae ( L1 ) .
Adult Brugia malayi male , female parasites , microfilariae and the L3 larvae were obtained from the Filariasis Research Reagent Repository Center ( Athens , Georgia , USA ) The adult parasites were washed three times in RPMI media supplemented with antibiotics ( 100 U/ml penicillin , 100 µg/ml streptomycin , and 0 . 25 µg/ml of Amphotericin B ) . The microfilariae were centrifuged at 800×g for 10 mins . The pellet was re-suspended in 10 ml of fresh RPMI-1640 and layered on Ficoll-Hypaque and centrifuged at 1500 rpm for 30 mins . The supernatant was discarded and the pellet containing microfilariae was washed with RPMI-1640 . Contaminating RBC were lysed by treating with ACLK solution and washed again . The animal procedures were conducted in accordance with the ACUC guidelines at the National Institutes of Health and at the University of Georgia . The adult male ( 1–2 worms/ml ) , adult female worms ( 1–2 worms/ml ) and microfilariae ( 0 . 25×106 mf/ml ) were cultured separately in serum-free RPMI 1640 ( GIBCO ) supplemented with 5 g/L glucose and antibiotic-antimycotic ( Invitrogen , 100 U/ml penicillin , 100 µg/ml streptomycin , and 0 . 25 µg/ml of Amphotericin B ) . The spent media was collected and replaced with fresh media every 24 to 48 hours to a maximum time of 7 days . The medium collected was filtered through 0 . 2 µM filters ( Millipore ) and stored , pooled and concentrated using Amicon Ultrafilters with 3 kDa cut-off membranes . The ES was stored at −80°C until use . Protein concentrations were estimated based on OD280 using a Nanodrop ND-1000 Spectrophotometer ( Thermo Fisher Scientific , San Jose , CA ) . The L3 larvae were cultured as described previously [10] with slight modification . Briefly , 5 to 10 larvae were cultured in 96-well cluster-well plates in 200 µl of serum-free α-MEM supplemented with ( ribonucleosides and deoxyribonucleosides ) , at 37°C with 5% carbon dioxide . The medium was supplemented with penicillin ( 100 U/ml ) , streptomycin ( 100 µg/ml ) , 0 . 25 µg/ml amphotericin , 2 µg/ml of ceftazidime and 2 µg/ml of ciprofloxacin . Ascorbic acid ( Sigma , St . Louis ) was added to a final concentration of 75 µM on day 5 of culture . The media prior to the addition of ascorbic acid was collected every 24–48 hours for analysis of pre-molting ES ( L3-ES ) and the spent media during and after molting collected as “molting ES ( L3-MES ) ” . The spent media were pooled from various batches , concentrated and subjected to microcapillary reversed-phase liquid chromatography-tandem mass spectrometry ( LC-MS/MS ) . The ES samples ( a minimum of 50 µg ) were digested with trypsin at 37°C overnight . The digests were resolved by strong cation exchange LC and analyzed using LC-MS/MS . Microcapillary RPLC was performed using an Agilent 1100 nanoflow LC system coupled online with a linear ion trap-Fourier Transform ( LIT-FT ) mass spectrometer . Reversed-phase columns were slurry-packed in-house with 5 µm , 300 Å pore size C-18 stationary phase in 75 µm i . d . ×10 cm fused silica capillaries with a flame pulled tip . After sample injection , the column was washed for 30 min with 98% mobile phase A ( 0 . 1% formic acid in water ) at 0 . 5 µl/min and peptides were eluted using a linear gradient of 2% mobile phase B ( 0 . 1% formic acid in acetonitrile ) to 40% solvent B in 110 minutes at 0 . 25 µl/min , then to 98% B for an additional 30 minutes . The LIT-FT mass spectrometer was operated in a data-dependent mode in which each full MS scan was followed by seven MS/MS scans wherein the seven most abundant molecular ions were dynamically selected for collision-induced dissociation ( CID ) using a normalized collision energy of 35% . Proteins were identified by searching the LC-MS/MS data using SEQUEST against the Brugia malayi database downloaded from The Institute for Genomic Research ( TIGR ) and the Wolbachia database from New England Biolabs . Methionine oxidation and phosphorylations on serine , threonine and tyrosine were included as dynamic modifications in the database search . Only tryptic peptides with up to two missed cleavage sites that met the criteria [delta correlation ( ΔCn ) > = 0 . 08 and charge state dependent cross correlation scores ( Xcorr> = 1 . 9 for [M+H]1+ , > = 2 . 2 for [M+2H]2+ and > = 3 . 1 for [M+3H]3+] were considered legitimately identified . For functional annotation of the transcripts , the program Blast× [11] was used to compare nucleotide sequences to the non-redundant ( NR ) protein database of the National Center for Biotechnology Information ( NCBI ) and to the gene ontology database [12] . The tool rpsBlast [13] was used to search for conserved protein domains in the Pfam [14] , SMART [15] , KOG [16] and conserved domain databases [17] . We also compared the transcripts with other nucleotide sequences of C . elegans , and Wolbachia . All BLAST comparisons were done with the complexity filter off , but segments of the same nucleotide of 20 bases or larger were masked . To identify possible transcripts coding for secreted proteins , the BMA-pep sequences ( obtained from TIGR Brugia malayi database ) , were submitted to the SignalP server [18] to identify translation products that could be secreted . In addition , non-classical secretion analysis was carried out using the SecretomeP2 . 0 [19] . Transmembrane helices were predicted using the TMHMM program [20] . Functional annotation of the transcripts was based on all the comparisons above and manually curated . Following inspection of all these results , transcripts were classified into various functional classes , with further subdivisions based on function and/or protein families . The entire annotated database ( with embedded links ) is accessible at http://exon . niaid . nih . gov/transcriptome/Bm-secretome/Brugia-Secretome-Web . zip . A standalone version can also be accessed and downloaded at http://exon . niaid . nih . gov/transcriptome/Bm-secretome/Brugia-Secretome-SA . tar . gz . DNA was isolated from the culture media of parasites and used to amplify with primers ( wsp forward primer: GATGAGGAAACTAGTTACTA; wsp reverse primer: CCAAATAGCGAGCTCCAGC ) targeting the Wolbachia wsp gene . As positive controls , genomic DNA isolated from Brugia malayi and Wuchereria bancrofti were used .
Analysis of data from the filarial genomic project indicated that 1 , 022 , 933 peptides could theoretically be generated from tryptic digestion of the 11 , 610 predicted proteins of Brugia malayi [21] . Our proteomic analysis revealed the Brugia malayi secretome to be surprisingly complex . The identified ES proteins can be found in concise form in Table S1 and in a complete annotated table with embedded links at Final Secretome-Web ( http://exon . niaid . nih . gov/transcriptome/Bm-secretome/Brugia-Secretome-Web . zip ) . From a total of 1040 unique and 65 non-unique ( defined as peptides that matched 100% with more than one filarial protein ) peptides identified from the adult male , adult female and microfilarial stages of the parasite , 170 , 239 and 540 proteins , respectively were identified ( Figure 1 ) . Interestingly , the ES products of the L3 larvae was limited to the identification of unique peptides that matched only 5 proteins with another 25 proteins being identified in the L3/L4 molting stage ( L3-MES ) . In addition , non-unique peptides ( to which a definitive protein match could not be made ) identifying 9 proteins from adult males and 12 from the adult females , 36 proteins from microfilariae and 3 proteins each from the L3-ES and L3/L4 ES ( L3-MES ) were detected ( Table S2 , Table S3 , Table S4 , Table S5 , and Table S6 , highlighted in yellow ) . Of the 852 distinct ( or different ) proteins identified in the secretome , 215 ( 25 . 2% ) would have been predicted based on clustered expressed sequence tagged ( EST ) sequences . Among the 852 proteins identified in the ES products , 274 are annotated as ‘hypothetical’ . The identification of these “predicted proteins” as clearly being measurable proteins provides validation that these are bona fide gene products . Moreover , each of these 274 proteins have orthologues in C . elegans ( Final Secretome-Web ) ( http://exon . niaid . nih . gov/transcriptome/Bm-secretome/Brugia-Secretome-Web . zip ) . Comparative analysis of these 852 proteins between the secretome and the somatic proteome ( unpublished ) of Brugia malayi suggest that 118 were exclusive to the ES products , of which 89 ( 72% ) were hypothetical proteins or uncharacterized proteins with no known function . On a weight per volume basis , male adult parasites ( about 20% of the female body weight ) produced 8–9% of the ES produced by an adult female . The female parasites were found to excrete/secrete not only more protein per worm but also a greater number of proteins , many of which have previously been shown to be gender specific [7] . The overwhelming majority of secreted molecules were stage-specific ( Figure 1 ) . As can be seen , 119/170 ( 70 . 0% ) of the ES products of the adult males , 157/239 ( 65 . 6% ) of the ES products of the adult females , 457/540 ( 84 . 6% ) of the ES products of the microfilariae , and 12/27 ( 44% ) and 5/8 ( 62 . 5% ) molting L3-ES and L3-ES respectively , were not shared between or among the stages studied . Fifteen percent of the 852 proteins identified within the ES were predicted to contain either 5′ signal peptide sequences or signal anchors . However , ∼50% of the total number of ES proteins of Brugia malayi predicted to be non-secretory by SignalP , would be predicted to be secretory by Secretome P . ( Final Secretome-Web ) ( http://exon . niaid . nih . gov/transcriptome/Bm-secretome/Brugia-Secretome-Web . zip ) . Combining the results obtained using both methods , results in 54 . 5% ( 437/802 ) of the proteins being predicted as having the necessary motifs to be secreted . Using conventional biochemically-based methods approximately 20 secreted or excreted proteins of Brugia malayi and other related parasitic nematodes have been identified [22]–[35] . Our analyses were able to corroborate the overwhelming majority of these except Bm20 , BmTGH-1 , GST , cathepsin , and SXP-1 . More recently , higher throughput methods have been applied to characterize the proteomes of mixed adult males and females [8] or of adults and microfilariae [9] , Our study identified 40 of the 80 proteins identified in Hewitson et . al . [8] and 90 of the 228 identified by Moreno et . al [9] , but failed to detect the remaining proteins that include protein disulphide isomerase ( PDI ) , BmR1 , and tropomyosin . In addition we were able to identify some of the proteins such as phosphatidylethanolamine binding proteins , phosphofructokinase and cyclophilin-5 that were detected by Hewitson et . al . , [8] and could not be found in the study by Moreno and Geary ( 9 ) . Moreover , in these two stages ( adult females and males cultured separately ) , we were able to identify an additional 320 proteins , listed in Table S1 and in Final Secretome-Web ( http://exon . niaid . nih . gov/transcriptome/Bm-secretome/Brugia-Secretome-Web . zip ) . Proteins were classified into main functional classes based widely on the KOG classification of the C . elegans orthologues from Wormbase release WS184 , with some adaptations for uncharacterized protein group comprising of hypothetical or uncharacterized conserved or unknown proteins . Metabolic processes involved in carbohydrates , amino acids , nuclear and energy were collectively classified into the category of metabolism . From the entire annotated secretome ( Figure 2 ) , 46% of the proteins do not have any known function and/or are predicted hypothetical genes . As mentioned above , these data emphasize that not only are the predicted hypothetical proteins actually being synthesized , but they might also play an important role at the host parasite interface . When examined by parasite stage , it is clear that none of the stages appeared to be biased toward a particular set of functional processes ( Figure 3 ) . A complete list of the functionally classified proteins is given in Final Secretome-Web ( http://exon . niaid . nih . gov/transcriptome/Bm-secretome/Brugia-Secretome-Web . zip ) . Spectral counts have been adopted to reflect the relative abundance of individual proteins [36] . The number of times a particular peptide is seen in the spectra ( Spectral Counts ) can be correlated with abundance of that particular protein . The abundance does not represent a true molar concentration of the proteins in the ES , but rather serves to differentiate between highly abundant proteins from those of lesser abundance . The number of peptides and the peptide ( s ) identified for each protein in the various stages are listed ( Table S1 , Table S2 , Table S3 , Table S4 , Table S5 , and Table S6 ) . The transition of L3s to the L4 stage reveals the identification of several immunologically important parasite proteins . Analysis of the proteins secreted by the L3 larvae indicates an abundance of the abundant larval transcripts ( ALT-1 and ALT-2; Bm1_26880 , Bm1_28735 ) that are not detected after the L3's molt to L4s . Although cysteine proteases ( i . e . , cathepsins ) have been implicated in molting of the L3 to L4 stage [37] , [38] , we were not able to identify these proteins in the secretome irrespective of the stage examined . The Brugia malayi homologue of LL20 ( Bm1_50995 ) , which has been identified as a nematode polyprotein allergen ( NPA ) , is an immunologically important antigen that is comprised of subunits that form a ladder 15 kDa increments on SDS-PAGE gels . Peptides matching to LL20 were found to be abundantly secreted not only by the molting L3s but also by the adult females and the microfilariae ( Table S1 , Table S3 , and Table S6 ) . Another molecule that was found to be abundantly secreted only by the molting L3s is the DJ-1 family protein ( Bm1_07685 ) . Among several of the parasite proteins identified in the ES that were “non-unique” was the larval allergen ( Bm1_17875 ) known to be among the most abundantly released by the L3 larvae and , by sequence analysis , is related to the ALT family of proteins . The larval allergen ( see Table S5 ) forms the majority of the L3-ES . Parasite encoded stress-inducible secretory proteins such as thioredoxin peroxidases , superoxide dismutases , glutathione peroxidases have the capacity to neutralize the reactive oxygen species and form a potential immune evasive strategy for parasite survival . During the molting stage of the L3 larvae , thioredoxins ( TRX; Bm1_46700 , Bm1_46705 ) and glutathione peroxidases ( GPX; Bm1_40465 ) were found to be secreted ( Figure 4 and Table S6 ) . Analysis of the ESTs for GPX shows that the L4 stage showed comparatively higher gene expression than in L3 ( L3 = 0; L4 = 7 ) . Another protein found to be abundant in the ES at the time of the L3/L4 transition and not in any of the other stages examined is the intermediate filament protein ( Bm1_45215 ) , that has orthologues in O . volvulus ( Ov1CF or OV-IF ) , A . lumbricoides IF-A and C . elegans ( IFA-1 ) . IFA-1 is required for survival past the L1 stage and is predicted to function as a structural component of the cytoskeleton . In addition , several proteins known to modulate the immune system of the host , leucyl aminopeptidase ( ES-62 ) [39] , gamma-glutamyl transpeptidase ( γ-GT ) [40] , macrophage migration inhibitory factors ( Bm-MIF ) [29] , [41] , and the galectins [42] , were detected in the molting ES during the L3/L4 molt . Compared to the L3 and L4 larvae that interface with the host over a relatively short period , the adult and microfilarial stages provide at both a quantitative and temporal level the overwhelmingly abundant levels of ES proteins . Several extracellular proteins including collagens , ankyrins , cuticulins and other associated proteins; and cytoskeletal proteins such as myosin and kinesin-like molecules could be detected . Galectins ( Bm1_46740 , Bm1_24940 ) or galactoside binding lectins , have been found to be abundant in the ES ( Figure 4 ) , and by inference could influence the host immune response and modulate ( or mediate ) pathologic responses . A number of enzymes involved in the general metabolism of parasites have been proved to be immunologically important . Gamma glutamyl transpeptidase ( γ-GT – Bm1_09950 ) is a multifunctional enzyme involved in the metabolism of glutathione and has been found to be released by the adult females and microfilariae but not adult males . γ-GT specific IgE antibodies appear to be associated not only with tropical pulmonary eosinophilia but also possible resistance to filarial infections [40] . This enzyme could also be involved in the catabolism of immunomodulatory and pro-inflammatory cysteinyl leukotrienes [43] . Triose-phosphate isomerase ( TPI – Bm1_29130 ) an enzyme involved in the conversion of dihydroxy-acetone to glyceraldehyde-3-phosphate was specifically identified abundantly in the ES of the adult female , and not in the adult male or microfilariae . TPI has been previously reported in the secretions of Schistosomes and Haemonchus contortus , suggesting it has an important , but yet unidentified role , in host-parasite interactions . Previous studies on the Brugia secreted products also found TPI to be the most abundantly expressed protein by the adults and microfilariae [8] , [9] . Incidentally , TPI is the most abundant protein in the ES based on relative abundance of unique peptides identified not only in the adult females but also in the molting L3 larvae ( Figure 4 ) . The other glycolytic enzymes detected at lower abundance levels are enolase ( Bm1_24115 ) and transaldolase ( Bm1_04195 ) . While transaldolase was maximally observed in the secretions of microfilariae and sparingly in adults , enolase was detectable only in the secretions of the adult male parasites . Previously identified immunomodulatory proteins such as BmMIF-1 ( Bm1_28435 ) and ES-62 ( Leucyl aminopeptidase , Bm1_56305 ) were found to be abundant in the ES of adult females and microfilariae . Among the cytokine homologues of the parasite ( macrophage migration inhibitory factor ( MIF ) and TGF-β ) , only Bm-MIF was identified in high abundance in the microfilariae and the adult females . Based on the abundance levels , it is not clear whether these proteins are the secretions of the uterine microfilariae or the adult female per se . At the transcriptional level , the females have a 2 . 4 fold increased expression of BmMIF -1 than in the adult males ( data not shown ) . In addition , MIF–2 ( Bm1_16680 ) was detected in the ES of the molting L3 ( L3-MES ) . ES-62 or leucyl aminopeptidase is a phosphorylcholine containing glycoprotein that is secreted only by the post infective larvae in A . viteae . Similar to A . viteae , the ES-62 homologue from Brugia malayi was found to be secreted by the L3s molting to L4s ( Table S1 and Table S6 ) . Of the surface associated proteins , the anti-oxidant products glutathione peroxidase ( GPX , Bm1_40465 ) ) and superoxide dismutase ( SOD , Bm1_48045 ) are important proteins suggested to be involved in immune evasion that have been detected in the microfilariae and the adults parasites . Abundance analysis indicates that SOD is more specific to the ES of microfilariae compared to adults , while GPX could be detected in each of the parasite stages with molting L3/L4 larvae and microfilariae containing greater amounts than the adults ( Figure 4 ) . In a similar fashion , the cystatins ( CPI-2 , Bm1_56600 ) and serpins ( Bm1_25935 , Bm1_25945 , and Bm1_28525 ) have been found to be released in greater amounts by microfilariae compared to adults . The most abundant microfilarial transcript from Brugia malayi encodes serine protease inhibitors ( serpins ) and was found only in the microfilarial stage . Another protease inhibitor-like molecule that was prominently identified is the phosphatidyl-ethanolamine binding protein ( Bm1_31500 , Bm1_41005 ) . It is not yet clear if the parasite encoded or phosphatidyl-ethanolamine binding proteins or the orthologues from other helminths ( Ov-16 ) have protease-inhibitory activity . One of the proteins of interest that was identified was the vespid venom allergen homologue-like ( VAH; Bm1_14040 ) protein that was detected in the ES of microfilariae . Peptides matching the microfilarial sheath protein ( Bm-SHP3; Bm1_50600 , Bm1_50585 ) were also found to be secreted by the microfilariae alone and not by the adult parasites . Endochitinase ( Bm1_06755 , Bm1_17035 ) and nematode cuticle collagens ( Bm1_27595 ) were found to be released in greater amounts by the microfilariae compared to the adults . Results from the secretome analysis indicate the presence of chitinases ( CHI-1 and CHI-2 ) only in the microfilarial stages including the uterine/new born microfilariae ( unpublished ) . Nematode cuticle collagens and its associated enzymes that are involved in cuticle synthesis ( peptidyl-prolyl cis-trans isomerase ( PPI , cyclophilins ( Bm1_24035 ) / FKBP ( Bm1_35815 , Bm1_56870 ) ) , prolyl-4-hydroxylase ( PHY ) , cyclophilins and FKBPs ) were easily detected in the secretome . In the present study 90 proteins ( listed in Table S7 ) of Wolbachia origin were found in the ES of the adults and the microfilariae , with the microfilariae contributing maximally . Single peptide hits matching Wolbachia proteins formed a majority ( 68 proteins ) of the proteins identified in the adult ES , while the Wolbachia proteins represented by more than one peptide seemed to be involved in nuclear metabolism ( based on functional annotation ) . In contrast , the Wolbachia proteins were not detectable in the infective L3 larvae or the molting larvae . Classification of the Wolbachia proteins did not result in the identification of any particular class of proteins that could be mapped to specific known biochemical pathways . Functional annotation of the Wolbachia proteins in the ES suggests a wide range from uncharacterized proteins , as well as those involved in nuclear regulation , metabolic enzymes , protein synthesis and modification ( Figure 5 ) . Further , to evaluate if Wolbachia themselves were released into the culture media [these intracellular bacteria being a surrogate for potential death of the in vitro cultured parasites] , DNA isolated from the ES of microfilariae , adult males and females was assayed for Wolbachia surface protein ( wsp ) gene by PCR . First , there was only a very minimal amount of DNA present in the culture media based on 260/280 nm ODs; secondly what little DNA there was did not contain Wolbachia ( Figure S1 ) .
Since the filarial parasites develop through multiple stages , how the different stages orchestrate the production of ES proteins becomes important for understanding successful parasite development and survival . Interestingly our data suggest that the filarial parasites utilize many secretory proteins in an as yet unknown but stage specific manner . The identification of ALT proteins only in the infective L3 larval stage corroborates previous EST analyses [53] . Furthermore , the importance of ALT proteins in invasion [54] , [55] suggests a primary parasitic adaptation by secreting molecules that help in the establishment of the infection . An interesting and intriguing aspect of the present study is the observation that the microfilariae appear to secrete/excrete more ( at a quantitative level and weight/volume basis ) proteins than do the adult stages . Some of these proteins identified in this stage , such as the serpins , had previously been shown to be microfilarial-specific [56] . Interestingly , microarray analysis indicated a 4-fold upregulation of the serpins in the adult females ( containing uterine microfilariae ) compared to the adult males ( data not shown ) . Among the protease inhibitors , CPI-2 is constitutively expressed transcriptionally by all stages of the parasite life cycle [27] , [57] , [58] , though , in the present study , the CPI-2 protein was found to be secreted only by the microfilariae . Interestingly , expression profiles of CPI-2 [57] have shown CPI-2 mRNA expression to be higher in L1 ( microfilaria ) and L3s compared to the adults stages [27] , [58] . Like the serpins , endochitinases are microfilarial specific proteins that are absent in the adult worms , with synthesis starting in newborn microfilariae 2 days after birth [59] . The possible functional role played by these chitinases have been reviewed previously [60] but appear to be restricted to the microfilarial stage . Recent advances in the understanding of the molting process bring into light the roles for several peptidases and specifically the metallo-peptidases that may be potential candidates for chemotherapeutic interventions [61]–[63] . In this study , apart from leucyl aminopeptidase ( ES-62 ) we were unable to identify other aminopeptidases in the molting L3 ES . Though cathepsins have been implicated in the molting process , they were not detected in the ES of molting L3 larvae . Transcriptional analysis of the molting process , currently underway , should address this more conclusively . Although each stage of the parasite has a cuticle , it is apparent that the microfilariae expend a large amount of energy in the maintenance and generation of their cuticular components ( e . g . peptidyl-prolyl cis-trans isomerase ( PPI , Cyclophilins ( Bm1_24035 ) / FKBP ( Bm1_35815 , Bm1_56870 ) ) , prolyl-4-hydroxylase ( PHY ) . Cyclophilins ( CYPs ) and FKBPs are large multi-member families in nematodes [64] . Representative members of the ‘cyp’ genes have been identified in B . malayi and other parasitic nematodes [64]–[69] as well . Whether these parasite-derived molecules function as human homologues is unknown . The success of a parasite to establish infection depends on its abilities to subvert the host immune defenses and survive in the host for a long time without eliciting an inflammatory reaction . In the current study we have identified a number of proteins that have been implicated such as the abundant larval transcripts , thioredoxins , and glutathione peroxidases in evading the host immune responses . Analyses of the anti-free radical defenses of Brugia malayi indicate the parasite can secrete GPX , TRX and SOD . Glutathione peroxidases ( GPX ) are surface associated proteins that protect the parasite by neutralizing the reactive-oxygen intermediate attack , or modifying and removing host immunomodulatory lipids . Thioredoxins ( TRXs ) are a large family of anti-oxidant proteins produced by a wide range of organisms in defense against toxic hydroxyl radicals that damage proteins , lipids and DNA . Despite the lack of a signal sequences , thioredoxins are secreted by the parasites and were only found in the ES of the molting L3's . Together with GPX , TRX provides the parasite with essential defense against reactive oxygen species . In the current study , superoxide dismutase ( SOD ) was been found to be present most abundantly in the ES of the microfilariae , and finding that differs to that which has been described previously [70] . Though DJ-1 was found to be secreted by the mixed adult parasites [8] and in the adults and microfilariae [9] , our study identified this protein to be abundant in the molting L3 larval secretions . Very little is known about the functional role of DJ-1 except its association with Parkinson's disease [71] . The multitude of functional groups within the DJ1/ThiJ/PfpI family makes it difficult to infer any specific role of the protein in the parasite . Analysis of homologous sequences to DJ-1 suggests that DJ-1 may have a role in thiamine metabolism . However , the association of DJ-1 with chaperones [72] , [73] and its KOG classification ( KOG2764 ) suggests a particular role for it in evading the host immune system . Although the LL20 ladder protein ( NPA ) was found to be abundantly present in the ES of the molting L3's and the adult female , interest in NPAs is not limited to their abundance but extends to their intrinsic role as potential allergens [74] and their potential role in the sequestration of immunologically important signaling lipids that can , in turn , modulate local host inflammatory reactions [74] , [75] . Several known immunomodulatory molecules of filarial parasites were identified in the ES in the present study . Among them were ES-62 , serpins , galectins and MIF-1 . ES-62 has been shown to modulate key signal transduction pathways associated with immune cell activation and polarization [76] . Leucine aminopeptidases have also been implicated in the metabolism of cysteinyl leukotrienes as has gamma-glutamyl transpeptidase [43] . Given the abundance of ES-62 and gamma glutamyl transpeptidase in the secretome of the parasite their role in modulating lymphatic ( and vascular ) endothelial cells exposed to these proteins in situ is of obvious importance . Another major component of the Brugia malayi secretome were the galectins . Galectins are atypical secreted proteins that occur both extra- and intracellular and have been identified from a number of filarial parasites [77]–[79] . Accumulation of galectins occurs at endothelial interfaces by binding to extracellular proteins and promoting endothelial proliferation [80] . The ability of galectins to bind specifically to IgE [78] , to regulate alternative macrophage activation [81] , and inhibit lymphocyte trafficking [82] suggests additional potential roles for these proteins . Another interesting but unexplained feature of the ES products is the identification of cytochrome C ( Bm1_28235 ) and histone-like molecules . These proteins are amongst the most abundant proteins of the Brugia malayi secretome ( Table S1 ) . Studies on the ES proteins of other parasites [83] , [84] have also identified these molecules , though no specific functional implications have been attributed to them . Similar findings of nuclear , cytoskeletal and mitochondrial proteins in the two recently published studies [8] , [9] in the ES proteins of the filarial parasites suggest that this is due to the active release of the proteins and not due to chance identification of these proteins contaminated by dying worms . Interactome analysis of the B . malayi ES products , based on the C . elegans interactome was unable to demonstrate networks of interacting proteins ( data not shown ) . The Wolbachia endosymbiont is very essential for Brugia worms , as antibiotic treatment that targets the Wolbachia results in the loss of adult worms viability [85] . Molecules of Wolbachia origin released by the parasite has been postulated to influence the inflammatory responses associated with infection [86] . Microfilarial ES contained a majority of the Wolbachia proteins identified . Most of these proteins were present in extremely low amounts in the ES . It is possible that dying parasites or disintegrating embryo's shed out by the adult worms could release Wolbachia proteins . In our system at least , though there were no dead adult or larval worms in culture , dying microfilariae could have resulted in the identification of some of the low abundant Wolbachia proteins . Wolbachia DNA , however , could not be detected in highly concentrated microfilarial ES ( Figure S1 ) suggesting that the gene products were indeed exported from their intracellular location . It is possible that LC-MS/MS data acquired of spots excised from the 2D-gels or 1D-gels [8] , [9] followed by LC-MS/MS was unable to identify these very low abundance proteins . This study provides a comprehensive examination of the secretome from most stages of the Brugia malayi parasite . The secretome is comprised of a large array of proteins many of which have been shown previously to modulate ( either directly or indirectly ) the host immune system , potentially promoting their survival . In addition , many “hypothetical proteins” have been detected whose presence provides independent corroboration of the Brugia Genome Project's gene prediction algorithms . The application of high-throughput proteomic techniques in the secretome analyses resulted in the identification of low abundance proteins that might not be detected using gel electrophoresis-based methods . Typical SignalP analyses could not account for the most of the ES products and suggests that new methods for these predictions need to be tailored to helminths . It is expected to have variations in the secretion profiles when the parasites are exposed to the host immune attack . Thus , this Brugia malayi secretome , in addition to the previously published secretomes of the filarial parasite Brugia malayi [8] , [9] , should help enable the study of the host-parasite interface in much greater detail than has been possible heretofore .
|
Human lymphatic filariasis caused by the nematode parasites Brugia malayi and Wuchereria bancrofti are a major cause of concern in tropical countries . Studies over several decades have identified various proteins of these parasites that have highlighted their role in host–parasite interactions and possible chemotherapeutic and prophylactic interventions . The availability of the parasite genome facilitates the identification of all of the proteins of the parasite that could interact with the host . In this study , we have attempted to identify the excretory-secretory proteins of the various stages of the parasite that could be maintained in vitro for a limited period utilizing a high-throughput proteomics approach . We observe and report that the parasites expend resources to secrete out various molecules that they utilize to evade the host immune system and modulate its responses . Further , this study also provides information on the predicted hypothetical proteins to be bonafide proteins and thus a catalogue of the excretory-secretory proteins towards a better understanding of the host–parasite interactions .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/helminth",
"infections",
"biotechnology/protein",
"chemistry",
"and",
"proteomics"
] |
2009
|
Brugia malayi Excreted/Secreted Proteins at the Host/Parasite Interface: Stage- and Gender-Specific Proteomic Profiling
|
RNA-protein interactions are critical in many biological processes , yet how such interactions affect the evolution of both partners is still unknown . RNA and protein structures are impacted very differently by mechanisms of genomic change . While most protein families are identifiable at the nucleotide level across large phylogenetic distances , RNA families display far less nucleotide similarity and are often only shared by closely related bacterial species . Ribosomal protein S15 has two RNA binding functions . First , it is a ribosomal protein responsible for organizing the rRNA during ribosome assembly . Second , in many bacterial species S15 also interacts with a structured portion of its own transcript to negatively regulate gene expression . While the first interaction is conserved in most bacteria , the second is not . Four distinct mRNA structures interact with S15 to enable regulation , each of which appears to be independently derived in different groups of bacteria . With the goal of understanding how protein-binding specificity may influence the evolution of such RNA regulatory structures , we examine whether examples of these mRNA structures are able to interact with , and regulate in response to , S15 homologs from organisms containing distinct mRNA structures . We find that despite their shared RNA binding function in the rRNA , S15 homologs have distinct RNA recognition profiles . We present a model to explain the specificity patterns observed , and support this model by with further mutagenesis . After analyzing the patterns of conservation for the S15 protein coding sequences , we also identified amino acid changes that alter the binding specificity of an S15 homolog . In this work we demonstrate that homologous RNA-binding proteins have different specificity profiles , and minor changes to amino acid sequences , or to RNA structural motifs , can have large impacts on RNA-protein recognition .
RNA-protein interactions and ribonucleoprotein complexes play key roles in many cellular processes including transcriptional regulation , translation , epigenetic regulation , and post-transcriptional silencing [1] . Furthermore , aberrant RNA-protein recognition and binding has been linked to multiple human disease states [1–4] , as well as promoting cancer metastasis [5] , and tumorigenesis [6–8] . Efforts to characterize RNA-protein interactions have focused on identification of RNA binding sites using experimental data coupled with motif-finders . These approaches typically identify short conserved sequences ( k-mers ) that may also have specific positioning within a predicted RNA secondary structure [9 , 10] . However , many RNA-protein interactions involve complex three-dimensional interactions that are not easily captured by k-mer descriptions[11] . Several studies have tried to identify general rules governing RNA-protein interactions using the growing collection of structural data [12 , 13] . Although limited by the availability of non-redundant data , these studies show that both electrostatic interactions and shape complementarity , on the part of both the RNA and the protein , are important for recognition . Due to the complexity of RNA-protein interactions , and the challenges associated with in-depth characterization of RNA binding sites , relatively few studies have assessed how the specificities of RNA-binding proteins may be conserved , or altered over evolutionary time . Many eukaryotic RNA-binding proteins appear to have conserved recognition motifs [14] . However , there may be multiple modes of binding for a single protein ( e . g . PUF ( Pumilio and FBF ) RNA-binding proteins ) , and minor changes to a protein sequence can have specific effects on RNA recognition [15] . Due to the nature of the genetic code , the direct impacts of genomic change on the structure of proteins and RNA are very different . RNA secondary structure is more conserved than sequence within RNA families [16] . Amino acid sequences of proteins tend to be much more highly conserved than nucleotide sequences of structured RNAs , and it is often difficult or impossible to follow the vertical inheritance of any but the most conserved structured RNAs ( e . g . the ribosome ) across large evolutionary distances [17] . For many RNA regulatory functions it appears that there are different RNA structures that accomplish very similar or identical biological functions in different bacterial phyla , which adds to the complexity of describing these RNA-protein interactions over evolutionary time . Bacteria commonly use portions of their mRNA transcripts as cis-acting regulatory elements ( riboregulators ) . Such regulators typically alter RNA structure in response to cellular cues including small molecules ( riboswitches ) , tRNAs ( t-boxes ) , or proteins [18 , 19] . A classic example of the structural diversity that has arisen across different bacterial phyla are the many distinct riboswitch classes that bind the small molecule S-adenosyl methionine ( SAM ) [20–24] . From structural data it is clear that at least three of these RNAs interact with their ligand ( SAM ) in fundamentally different ways , suggesting completely independent derivation [25–28] . Furthermore , this example is far from unique . Two distinct riboswitch classes interact with the second messenger c-di-GMP [29 , 30] , and three such classes with the nucleoside prequeosine-1 [31–33] . The existence of multiple unique RNA architectures responsible for analogous biological functions is not limited to RNA-small molecule interactions . This phenomenon is also apparent for riboregulators interacting with protein partners . Multiple mRNA regulatory structures have been identified that perform autogenous regulation in response to ribosomal proteins bL20 , uS4 , and uS15 [34–37] . From even this small set of RNA-protein interactions , we see that distinct RNA architectures in different bacterial phyla can successfully perform analogous biological functions by interacting with homologous protein binding partners . In some cases there is obvious similarity between the mRNA and rRNA binding-sites , suggesting that the protein recognizes the same tertiary structure features [38 , 39] . However , there are several examples where this similarity is not obvious [40–43] . In such cases , it remains unclear how much of the mRNA structural diversity observed is due to independent derivation of the similar tertiary structure , or if differences between homologous protein partners lead to distinct RNA-binding profiles . To assess these questions , we have focused on ribosomal protein S15 which has two RNA-binding roles . S15 is a primary rRNA binding protein that is responsible for organizing the 16S rRNA during assembly of the small ribosomal subunit . The primary rRNA-S15 recognition site is formed where helices H20 , H21 , and H22 come together to yield a three-way junction ( 3WJ ) that contains the base-triple GGC [44] ( Fig 1A and 1B ) . The secondary S15-recognition site is a GU/GC motif that is ~10 nucleotide pairs ( 1 helical turn ) distal to the 3WJ in H22 ( Fig 1A and 1B ) . Together the 3WJ and the GU/GC motif form a bipartite S15-recognition surface in the rRNA . Not surprisingly , this region of the rRNA is highly conserved across all bacterial species to ensure proper ribosome assembly and function [45] ( Fig 1C ) . S15 also interacts with an RNA structure found in the 5’-UTR of its own transcript enabling negative regulation of its own expression in many bacterial species . However , unlike the rRNA , this regulatory structure is not conserved [35] . In different groups of bacteria , four experimentally validated [37 , 44 , 46 , 47] , as well as two predicted [37] , mRNA structures with distinct architectures interact with their respective S15 homologs to enable gene regulation ( Fig 2A–2D ) . In this study we assess the interactions between examples of each of the S15-interacting mRNA families and their respective S15 homologs . These natural mRNA-protein interactions provide us with an opportunity to explore whether the diverse RNA architectures present similar tertiary structure surfaces to the protein , or if the different S15 protein homologs have distinct RNA recognition profiles . We assess RNA-S15 recognition using a translationally fused β-galactosidase reporter to characterize biologically relevant regulatory interactions , and use in vitro binding assays to directly quantify the RNA-protein interactions . We find that the results of the regulatory assays and in vitro assays largely agree and together show that there are differences between S15 homologs that result in specific recognition of the diverse mRNA structures . Furthermore , we analyze the conservation of S15 amino acid sequences from species showing different recognition patterns and identify amino acid mutations responsible for these specificity changes . Together our results suggest that even highly conserved RNA-binding proteins have distinct RNA recognition profiles , and that co-evolution has occurred between bacterial S15 homologs and their respective mRNA regulators .
To explore whether S15 homologs can specifically recognize different mRNA architectures to allow regulation within the cell , we utilized a β-galactosidase reporter assay . This functional assay directly tests the regulatory interaction between an mRNA and ribosomal protein S15 and enables the mRNA to fold into a biologically relevant structure . One plasmid contains an mRNA-LacZ fusion ( pRNA ) which was constructed by cloning the 5’-UTR through the first 5–9 codons of rpsO in-frame with lacZ and downstream of an IPTG-inducible promoter . A second plasmid ( pS15 ) includes a full-length rpsO open reading frame ( encoding S15 ) under the control of the pBAD33 L-arabinose inducible promoter . The plasmids have compatible replication origins and different antibiotic markers allowing them to be stably maintained in the same bacterium . For regulatory assays the plasmids are co-transformed into an E . coli K12:ΔrpsO strain that lacks endogenous S15 [48] ( strain confirmed through PCR screening , S1 Fig ) . Each of the S15 homologs complemented this strain , enabling much faster growth when protein expression was induced ( Fig 1E ) . Cells containing a pRNA and a pS15 are grown with and without L-arabinose , and at stationary phase the reporter is induced for 30 minutes with the addition of IPTG . Subsequently , the β-galactosidase activity of + and–L-arabinose cultures started from a single colony are compared to indicate whether a given mRNA structure enables S15-dependent regulation of β-galactosidase expression . Given the short induction time during stationary phase , we did not observe any noticeable growth changes upon induction of individual mRNA reporter constructs . The four experimentally validated riboregulators and their respective S15 homologs from Escherichia coli ( Ec-mRNA , Ec-S15 ) , Geobacillus kaustophilus ( Gk-mRNA , Gk-S15 ) , Thermus thermophilus ( Tt-mRNA , Tt-S15 ) , and Rhizobium radiobacter ( Rr-mRNA , Rr-S15 ) were each examined using the β-galactosidase reporter assay . We confirmed all native mRNA-S15 regulatory interactions ( Fig 2F ) by directly comparing fold repression of pS15 to pBAD33 with no insert ( pEMPTY ) . In each case we find that the native regulatory interaction can be detected using our assay in the surrogate organism . However , the unregulated levels of β-galactosidase expression using each mRNA riboregulator affects the resulting fold-repression ( S2 Fig ) . The Ec-mRNA showed the highest β-galactosidase activity ( ~5 , 000–10 , 000 Miller Units ) whereas the remaining mRNAs tested were all within a similar range ( ~1000–2000 Miller Units ) . To further ensure the significance of our observed interactions , a mutation abolishing the native binding interaction was introduced into each mRNA . In each case repression was reduced , typically to levels comparable to that observed for pEMPTY ( ~2-fold ) , although the Tt-mRNA-M1 does retain some regulatory activity ( Fig 2F ) . To determine whether the distinct mRNA architectures contain a shared tertiary structure or binding motifs , we examined all inter-species interactions using our regulatory assay . These results show that each mRNA structure has a specific set of S15 homologs to which it responds . For the mRNA regulator from E . coli , Ec-mRNA , both Rr-S15 and Tt-S15 successfully regulated β-galactosidase expression , yet do so more modestly than its native binding partner , Ec-S15 ( Fig 3A ) . The mutation abolishing the native RNA-protein interaction ( Ec-mRNA-M1 , derived from [49] ) deregulated reporter expression in response to both Rr-S15 and Tt-S15 . Gk-S15 did not regulate the Ec-mRNA or its mutant . These results suggest that these three S15 homologs , Ec-S15 , Tt-S15 , and Rr-S15 , interact with this mRNA in a similar fashion to regulate gene expression . The inability of this mRNA to respond to Gk-S15 suggests that Gk-S15 requires a regulatory motif or structure not found in Ec-mRNA . In contrast , the mRNA from R . radiobacter , Rr-mRNA , regulates gene expression in response to all the S15 homologs ( Fig 3B ) . A mutation to Rr-mRNA in the main stem was sufficient to deregulate expression in response to both Ec-S15 and Rr-S15 . However , this mutation did not impact the convincing regulation observed in response to Tt-S15 and Gk-S15 ( >10-fold repression observed ) . This suggests that the Ec-S15 and Rr-S15 homologs utilize similar determinants to recognize the mRNA , but that the Gk-S15 and Tt-15 homologs may be recognizing alternative motifs that are not impacted by the mutation . The mRNA from T . thermophilus , Tt-mRNA , displayed regulatory activity in response to all the S15 homologs . A mutation to the 3WJ ( derived from [50] ) diminishes Tt-mRNA’s response to Tt-S15 , Ec-S15 , and Gk-S15 homologs ( Fig 3C ) . However , this mutation does not completely abolish regulation in response to any of the proteins , and had no effect on regulation in response to Rr-S15 ( Fig 3C ) . These results have two potential interpretations . First , Tt-S15 , Ec-S15 , and Gk-S15 proteins may recognize Tt-mRNA in a different manner than Rr-S15 , and therefore a mutation to the binding site for Tt-S15 may not impact binding and regulation in response to Rr-S15 . A second explanation is that the relatively modest 6-fold regulation observed for Rr-S15 is an artifact of our regulatory assay . The mRNA from G . kaustophilus , Gk-mRNA , is also responsive to all S15 homologs tested ( Fig 3D ) . Like the Rr-RNA , the convincing regulatory responses to Ec-S15 , Rr-S15 , and Tt-S15 were not diminished by the mutation to Gk-mRNA ( a truncation used during in vitro studies in [40] expected to disrupt the 3WJ ) , while regulation in response to Gk-S15 was abolished by this mutation . Like the Rr-mRNA , this data suggests that that the binding determinants for Tt-S15 , Ec-S15 , and Rr-S15 on Gk-mRNA are different from those of Gk-S15 , and that different S15 homologs utilize distinct features to recognize the same mRNA . Together , the regulatory assays show that there is extensive , but not universal cross-reactivity in the inter-species mRNA-S15 regulatory interactions . However , results obtained with mRNA mutants suggest that even mRNAs recognized by multiple S15 homologs are recognized using different determinants . In particular , for both the Gk-mRNA and the Rr-mRNA , mutations that abolish native interactions have little or no impact on interactions with other S15 homologs . Given that many of our mutations that abolish native interactions still allowed regulation in response to other protein homologs , we used in vitro nitrocellulose filter-binding assays to directly measure the strength of RNA-protein binding interactions to corroborate our findings . All four S15 homologs were purified and nitrocellulose filter binding assays were performed for all cross-species interactions . We find that the dissociation constants for native interactions are in the 2–20 nM range . However , the native interactions were not always the strongest interactions . For example , Gk-S15 bound Tt-mRNA with an affinity that was almost an order of magnitude smaller than Tt-S15 ( 0 . 35 nM vs . 2 . 11 nM ) . We were unsuccessful in demonstrating Ec-mRNA interactions with any S15 homolog including its native binding partner; therefore it was omitted from further study . The native Ec-mRNA interaction with Ec-S15 has been characterized in vitro in the past ( KD = 231 nM ) [51] . Notably , this value is significantly higher than those that we measured for the other native interactions . Although a 3’-terminal [32P]pCp has been previously shown to decrease the KD four-fold in truncated versions of this RNA [51] , we found that labeling the full-length mRNA with [32P]pCp did not change our result . We did not explicitly test the truncated RNA since we are primarily interested in the wild-type interaction . Aside from our inability to measure interactions with Ec-mRNA , we find that our in vitro findings closely follow the results of the regulatory assays . The Rr-mRNA was able to interact with all S15 homologs in vitro , and all are relatively strong interactions with dissociation constants ranging from 1 to ~30 nM ( Table 1 ) . The inactivating mutation ( Rr-mRNA-M1 ) abolished interaction with Ec-S15 , but had little impact on interactions with the Gk-S15 or Tt-S15 homologs . These data corroborate our results from the regulatory assay indicating that Gk-S15 and Tt-S15 interact with the Rr-mRNA-M1 , and further indicates that Ec-S15 , Gk-S15 , and Tt-S15 homologs use distinct features to recognize this mRNA . Tt-mRNA binds strongly to both Tt-S15 and Gk-S15 , which corroborates our in vivo regulation findings . Conversely , Ec-S15 and Rr-S15 both do not bind Tt-mRNA in vitro , which makes interpreting the regulatory assay results less clear . They both displayed modest regulatory activity in vivo . Mutating Tt-mRNA decreased the regulatory response to Ec-S15 , yet did not significantly impact the response to Rr-S15 ( Fig 3C ) . However , neither Ec-S15 or Rr-S15 were able to bind this mutant in vitro . In comparison to Gk-S15 and Tt-S15 , the dissociation constants measured for Ec-S15 and Rr-S15 tend to be significantly higher for all measured S15-mRNA interactions , indicating that perhaps these proteins behave less well in vitro . Alternatively , relatively high levels of noise in our regulatory assay ( even empty vector controls typically display 2–3 fold repression ) may bias our findings . Therefore , Ec-S15 may be able to regulate gene expression using Tt-mRNA structure because the regulatory activity decreased with the mutated mRNA . However , whether Rr-S15 interacts with the Tt-mRNA to allow regulation remains unclear . In addition , although regulation of the Tt-mRNA by Gk-S15 is significantly reduced by the Tt-mRNA-M1 mutation , Tt-mRNA-M1 is not sufficient to completely abolish in vitro binding of Gk-S15 . However , the measured KD is over two-orders of weaker ( 0 . 35 nM vs 76 . 3 nM ) , and the maximum fraction of RNA bound by the protein is <20% ( S3 Fig ) , indicating that the in vitro interaction may be non-specific . Gk-mRNA interacted with all four S15 homologs in vitro ( Table 1 ) . The strongest interaction was with Tt-S15 , roughly an order of magnitude stronger than the native Gk-S15 interaction , and roughly three orders of magnitude stronger than with Ec-S15 and Rr-S15 . In addition , the Ec-S15 and Tt-S15 homologs retain strong interactions with the Gk-mRNA-M1 . This suggests that the retained regulation for this mutant in response to Ec-S15 and Tt-S15 is because these homologs still bind the mutant mRNA . In addition , while we do not measure any interaction between Rr-S15 and the Gk-mRNA-M1 ( up to 250 nM Rr-S15 ) , the interaction between Rr-S15 and Gk-mRNA is relatively weak in comparison to the other S15 homologs ( KD ~200 nM ) . Therefore , Rr-S15 may bind Gk-mRNA-M1 weakly , yet this interaction is sufficient to regulate reporter expression within cells . In conclusion , our in vitro results with the Gk-mRNA suggest that the regulatory interactions we observed between Ec-S15 , Rr-S15 , Tt-S15 and the Gk-mRNA and its mutant ( Gk-mRNA-M1 ) are indeed due to differences in the way that the proteins interact with the mRNA . In summary , we find that measuring cross-species interactions between S15 homologs and diverse mRNA structures using both regulatory assays and in vitro binding assays shows that the two approaches largely agree . While in isolation each type of assay is prone to various artifacts ranging from poor in vitro binding properties , to likely differences in protein expression levels in the surrogate organism , the large extent of agreement between our two assays significantly strengthens our conclusions . Overall , we find that Tt-S15 and Gk-S15 bind very tightly in vitro . This may be due to many factors including that the Tt-S15 and Gk-S15 homologs are both from thermophiles and may be more stable resulting in better in vitro binding characteristics . We also assessed Tt- and Gk-S15 in vitro binding at 55°C and found that no significant differences were detected at the higher temperature . To combine our in vitro and regulatory results into a single determination of whether or not an interaction occurs , we consider all measureable dissociation constants as viable interactions . For regulatory interactions , we consider all interactions that are significantly reduced by a mutation to the RNA , or corroborated by in vitro data as viable interactions ( Fig 4F ) . Using this criterion there is only a single ambiguous interaction , which is that between Rr-S15 and the Tt-mRNA . In addition , we did not detect an in vitro interaction between Ec-S15 and the Tt-mRNA , although regulation was observed for this pairing ( ~ 10 fold repression ) , and it is reduced by the Tt-mRNA-M1 , suggesting that it is not an artifact . From our collected data it is clear that there is extensive cross-reactivity , but that S15 homologs often recognize mRNAs using different characteristics , as demonstrated by the very divergent responses of different S15 homologs to the mutated mRNA structures . Using this data we can start to assess what RNA structural motifs result in these differences . The rRNA binding site for S15 is bipartite , consisting of a three-way junction ( 3WJ ) and a GU/GC motif approximately one helical turn away from the 3WJ ( Fig 4A ) . Previous studies have established that the E . coli mRNA mimics of the GU/GC motif [51] , and that the Tt-mRNA mimics the G-G-C base-triple found in the 3WJ of the rRNA [50] . However , in both of these cases it is clear that while the mRNA is contacted at a second position consistent with bi-partite binding , the second position bears limited resemblance to the rRNA . In the case of Ec-mRNA , the second binding site occurs within the co-axially stacked pseudoknot [51] ( Fig 4B ) , and in the case of the Tt-mRNA , the long H2 stem is necessary for binding , but the GU/GC motif is replaced by a G•G mismatch [50] ( Fig 4D ) . In contrast , the Gk-mRNA appears to contain mimics of both binding determinants . The 3WJ is mimicked in the multi-stem junction and a GU/GC motif is apparent approximately one helical turn away from this junction [46] ( Fig 4E ) . In the case of the Rr-mRNA , far less data exists concerning which bases are necessary for binding . However , a GU/GC motif is apparent in the most conserved portion of the Rr-mRNA [37] , and like the Tt-mRNA , the junction of the stems is important for retaining interaction with its native binding partner [37] ( Fig 4C ) . Taking our results in conjunction with previously published results , the data suggests that each of the mRNA structures mimics a portion of the rRNA . Both Ec-mRNA and Tt-mRNA contain a direct mimic for a portion of the binding site , while Gk- and Rr- mRNAs likely contain both portions ( Fig 4 ) . The inactivating mutations for each of the mRNAs target different portions of these rRNA binding sites . Rr-mRNA-M1 and Ec-mRNA-M1 both target putative GU/GC motifs , the Gk-mRNA-M1 is a truncation that presumably disrupts the 3-way junction , and the Tt-mRNA-M1 also targets the 3-way junction . The partial mimicry of the S15 rRNA binding site potentially explains the regulatory differences we observe for the S15 homologs . Our observations suggest that Ec-S15 and Rr-S15 preferentially recognize the GU/GC motif . Regulatory interactions between both Ec-S15 and Rr-S15 and several mRNAs are significantly impacted when this region is mutated ( Ec-mRNA-M1 , and Rr-mRNA-M1 ) . The regulatory interaction between Ec-S15 and Rr-S15 does not appear to be impacted by Gk-mRNA-M1 , a mutant targeting the putative 3WJ . Tt-mRNA lacks the GU/GC motif , and while Ec-S15 appears to regulate gene expression using Tt-mRNA , this interaction could not be reproduced in vitro . In contrast , the Gk-S15 appears to preferentially interact with a mimic of the three-dimensional motif formed at the helical junction . Gk-S15 does not interact with the Ec-mRNA ( which lacks a mimic of the junction ) , it is not impacted by the Rr-mRNA-M1 mutation that targets the GU/GC motif , and mutations that impact the junction result in lack of regulatory activity ( Tt-mRNA-M1 , and Gk-mRNA-M1 ) . Finally , the Tt-S15 appears to regulate gene expression with any mRNA structure that contains either portion of the rRNA binding site . The Ec-mRNA and Tt-mRNA each contain an obvious mimic for a single portion of the rRNA binding site , and mutations to these regions prevent gene regulation in response to Tt-S15 . The Gk-mRNA and Rr-mRNA are presumed to contain mimics of the entire rRNA binding site , and mutations that impact only one of these regions do not affect the regulatory interaction with Tt-S15 . In summary , we propose that the four S15 homologs preferentially recognize different sections of the naturally occurring mRNA regulators . To test our model for S15 interaction we constructed a second mutation of Gk-mRNA targeting the putative GU/GC motif ( Fig 5A ) . We hypothesized that this mutant should abolish regulation and binding of Ec-S15 and Rr-S15 , and have less of an impact on the Tt-S15 and Gk-S15 interactions . The interaction between this mutant mRNA and all four S15 homologs was assessed using both our regulatory assay and in vitro binding assay . We find that this mutation indeed abolishes regulation of ß-galactosidase expression in response to Ec-S15 and Rr-S15 , and reduces regulation in response to Tt-S15 ( Fig 5B ) . In addition , this mutation abolishes in vitro interactions with each of these proteins ( Fig 5C ) . Gk-S15 weakly binds this mutant ( the dissociation constant is nearly two orders of magnitude higher than that for the native interaction ) , but displays significant regulatory activity . These results are consistent with our proposal that while the GU/GC motif alone is not sufficient to enable interaction between Gk-mRNA and its native binding partner , it is sufficient to allow interactions between Gk-mRNA and the other three S15 homologs . In summary , our results indicate that homologous proteins , even those that recognize the same RNA structures , do so using different structural determinants . Our specificity data as well as existing studies indicate that the determinants for mRNA and rRNA binding are distinct [40 , 52] . We hypothesize that , depending on the RNA regulator present in the organism , the positions in S15 under strong selection are different . Such positions may be responsible for mRNA as opposed to rRNA recognition . To explore this hypothesis , we analyzed the rpsO coding sequences from sequenced microbial genomes containing each class of mRNA regulator . For the E . coli , G . kaustophilus , and R . radiobacter RNAs there are high-quality RNA alignments that provide a list of genomes containing each mRNA regulator [35–37] . For each class of RNA regulator we constructed alignments of the corresponding S15 protein coding sequences , which we will refer to by their species type ( e . g . alignment of S15 sequences from organisms containing homologs of the Ec-mRNA will be referred to as the Ec-alignment ) . S15 is typically well-conserved and the alignments contain few if any gapped regions . The Gk-alignment was the largest at 202 sequences; the Ec-alignment had 165 sequences , and the Rr-alignment 65 sequences . In the case of the T . thermophilus mRNA regulator , no RNA alignment exists , and a cursory BLAST search did not return hits to the mRNA outside the Thermus genus . Both the Rr-S15 and Tt-S15 were omitted from further analysis due to the limited sequence alignments that could be constructed for them . In addition , this choice allows us to focus on the differences between Gk- and Ec-S15 , which display very different RNA interaction behaviors based on our data . Previous mutagenesis studies for both Ec-S15 and Gk-S15 suggest they use similar , but not exactly the same , residues in recognition of their mRNA and rRNA [46 , 51 , 52] ( S4 Fig ) . To systematically assess which positions might be under selective pressure , we used the tool Rate4site to evaluate each of the alignments [53] . Rate4Site returns a Z-score for each position indicating the extent of conservation . Statistical significance of the Z-score depends on the overall extent of conservation over the entire protein sequence . Therefore , due to the small size and the high degree of conservation in our alignments , no site had statistically significant Z-scores ( even those that are completely conserved ) . However , the Z-score may be used as a rough indicator of conservation [53] ( Fig 6A ) . There are many positions that are strongly conserved ( Z-score < -0 . 1 ) , however most of these have the same amino acid conserved in both alignments ( e . g . position 28 , which is a strongly conserved glutamine ) ( Fig 6B ) . Positions 2 , 40 , 58 , and 61 show evidence of strong conservation of different amino acids in the two alignments ( e . g . at position 2 an alanine is conserved in the Gk-alignment , but a serine in the Ec alignment ) . Positions 9 , 18 , 71 , 72 , 73 , and 79 are strongly conserved in one alignment , but highly variable ( Z-score > 0 . 5 ) in the other ( e . g . position 18 is a conserved histidine in the Gk-alignment , but quite variable in the Ec-alignment ) . Additionally , both the N- and C-termini of the proteins show high degrees of variability in both alignments compared with the central portion that is expected to make direct contacts with the RNA . To determine whether the positions identified above contribute to our observed interaction specificity ( Fig 4 ) , we focused on the Ec- versus Gk-alignment differences , with the goal of identifing amino acid changes that would enable Gk-S15 to recognize and regulate gene expression of its 3WJ-mutant ( Gk-mRNA-M1 ) , or Ec-mRNA , both of which had no regulatory activity with Gk-S15 . Several of the positions identified are not expected to contact the RNA based on structural data ( positions 2 , 4 , 9 , and 79 ) [54] ( Fig 6C ) , or are the same in the Ec-S15 and Gk-S15 sequences ( position 73 ) ( Fig 6C ) . Therefore we assessed whether Gk-S15 carrying the sextuple mutation to positions H18D , N40Q , K58R , G61S , R71K , and K72R ( Gk-S15-6MUT ) would regulate gene regulation with Ec-mRNA , or Gk-mRNA-M1 ( Fig 6A ) . We find that Gk-S15-6MUT is capable of regulating gene expression with both Gk-mRNA-M1 and Ec-mRNA ( Fig 6D and 6E ) . Furthermore , this interaction appears to be specific as it is abolished in the Ec-mRNA-M1 . This result suggests that one or more of the altered positions are responsible for recognition of these mRNA structures ( Fig 6D and 6E ) . We speculate these residues contribute to higher affinity recognition of the GU/GC motif or possibly play a role in stabilizing a secondary binding site on the mRNA , independent of the GU/GC . When tested with Gk-mRNA-M2 , Gk-S15-6MUT retains significant regulatory activity , evocative of that displayed by the Tt-S15 ( Fig 5B ) . Our results suggest that Gk-S15-6MUT still recognizes the 3WJ , and the presence of either motif is sufficient to allow gene regulation ( Fig 6D and 6E ) . To further assess whether the diversity present in the N- and C-termini of the protein play a significant role in recognition , we also created a series of chimeric proteins for Gk-S15 and Gk-S15-6MUT where the N- and C-terminal sections were swapped from Gk-S15 to Ec-S15 ( S5A–S5C Fig ) . From these studies we found that the N-terminal residues from Ec-S15 typically decreased the extent of regulation across the board ( S5D Fig ) . This could be due to several factors including potential deleterious interactions between the N-terminus and other portions of the protein structure ( the N-terminus represents 12 changes between Ec-S15 and Gk-S15 ) , as well as differences in protein expression levels . The N-termini of protein coding sequences have been implicated in the past in determining expression levels[55–58] . We also found that chimeras with swapped C-terminal portions behaved very similarly , likely due to the small number of amino acid changes ( three ) between the two sequences . In summary , the N- and C-terminal regions of the protein are unlikely to play a large role in mRNA recognition and gene regulation , but do impact the extent of regulation observed in our regulatory assay due to alterations in effective protein concentration . The goal of this study was to assess how the differences between S15 homologs may contribute to the diversity of mRNA regulators that arise across different bacterial phyla to allow gene regulation . This work shows how the rRNA binding site for S15 may be partially mimicked in the four different mRNA regulators . We demonstrate that S15 homologs have distinct RNA binding profiles , and that even when recognizing the same RNA , different homologs are using distinct sequence features . These results suggest that either S15 has co-evolved with its mRNA regulators , or that differences between the ancestral S15 proteins lead to the development of a diverse array of RNA regulators that we observe in nature today .
The pRNA plasmid was constructed by modifying the reporter plasmid ptrc-Ec-mRNA-GFP from [37] . First , the ptrc promoter was replaced with the plac promoter . Complementary oligonucleotides of the lac promoter sequence flanked by the cohesive ends corresponding to a XhoI site ( 5’ ) and a EcoRI site ( 3’ ) were phosphorylated using T4 Polynucleotide Kinase , annealed , then ligated into ptrc-EcmRNA-GFP [37] digested with XhoI and EcoRI using Quick Ligase . Second , the lacZ gene was amplified from E . coli genomic DNA using Phusion DNA polymerase and primers containing restriction sites SalI and XbaI . The PCR product was digested and ligated into ptrc-RNA-GFP digested using the same enzymes ( GFP was excised in this process ) . This new plasmid , pBS2-Ec-RNA , was sequence verified . Finally , the lac repressor coding sequence ( lacIQ ) was cloned into pBS2-Ec-RNA at the XhoI site . The lacIQ gene flanked by XhoI sites was amplified from E . coli genomic DNA ( Strain NCM534 , K12 derivative , Yale E . coli Genetic Stock Center #8256 ) using Taq DNA polymerase to generate pBS3-RNA . The plasmid sequence was verified by Sanger sequencing . All mRNA sequences were cloned into the pBS3-RNA plasmid as a translational fusion with lacZ using primers containing EcoRI and SalI restriction sites ( See S6 Fig for overview of plasmid , and S3 Table for list of primers ) . Translational fusions were constructed such that the first 9 amino acids originating from E . coli or R . radiobacter rpsO , 5 amino acids from T . thermophilus rpsO , or 4 amino acids from G . kaustophilus rpsO , were appended to the N-terminus the lacZ sequence . The lacZ sequence requires a start codon from the fused rpsO sequence . All enzymes for molecular biology were purchased from New England Biolabs unless otherwise noted . Mutations to the mRNAs were constructed by site-directed mutagenesis ( S3 Table ) . pS15 protein expression plasmids were constructed by amplifying the rpsO open reading frame from genomic DNA with a forward primer containing a SacI site plus a strong ribosome binding site that matched the E . coli ribosome binding site preceding rpsO and an 8 nucleotide linker ( S3 Table ) preceding the rpsO start site . The native ribosome-binding sites preceding rpsO from both G . kaustophilus and T . thermophilus were tested . However , these did not allow sufficient protein production to complement the ΔrpsO strain and were consequently abandoned . The reverse primer contained an XbaI site . After digestion , the PCR product was cloned into the pBAD33 vector ( ATCC 87402 ) digested with the same enzymes . All pS15 were sequence verified . The Gk-S15-6MUT sextuple mutant was created using site-directed mutagenesis with primers listed on S3 Table and chimeras created by PCR assembly using pEc-S15 , pGk-S15 , or pGk-S15-6MUT as template DNA . K12: ΔrpsO E . coli cells were transformed with a pS15 and a single colony picked to grow cultures +/- 15 mM L-arabinose for ~16 hours in LB + 34 ug/mL chloramphenicol . Cultures were diluted to OD600 = 0 . 01 in 0 . 5 mL of fresh medium 24-well plates , and OD600 was measured for 27 . 5 hours . Each pS15 was performed 3+ replicates . Doubling times were calculated by taking the inverse of the slope of ln ( OD600 ) in exponential phase readings . K12: ΔrpsO E . coli cells ( kind gift from Gloria Culver , [48] ) were co-transformed with pRNA and pS15 plasmid ( made competent using the Z-competent buffer system , Zymo Research ) . Although this strain does contain a chromosomal copy of lacZ , we find that it is significantly repressed by the lacIQ allele present on our reporter plasmid such that the background levels of β-galactosidase expression from the native lacZ are < 10–20% of those that we observe from our reporter carried on a multi-copy plasmid ( S2 and S7 Figs ) . However , no doubt some of the experimental variation and background that we observe is due to this additional copy . For our assays , a single colony was used to start overnight cultures , grown +/- L-arabinose ( 15 mM ) at 37°C , then diluted the next day to OD600 = 0 . 15 in fresh media ( LB + 100 ug/mL ampicillin + 34 ug/mL chloramphenicol +/- 15 mM L-arabinose ) . At stationary phase ( 5 hours after dilution ) 1 mM IPTG was added to induce β-galactosidase expression . After 30 minutes , 100 ug/mL spectinomycin was used to stop initiation of protein translation , and the cultures assayed immediately according to Miller [59] to determine the levels of reporter expression . Fold repression = ( Miller units of–L-arabinose ) / ( Miller units of + L-arabinose ) . All RNA/S15 combinations were examined with 3+ independent replicates ( typically 4 ) . To determine the significance , all fold repression values were compared as indicated in S1 Table ( data on Figs 2 and 3 ) and S2 Table ( data on Figs 5 and 6 ) using a Welch’s single-tailed T-test in Microsoft Excel . Regulation was considered biologically significant if greater than 3-fold repression was observed , and the fold-repression was significantly different ( p<0 . 05 ) than that observed with an empty pBAD33 vector . DNA corresponding to the 5’-UTR of the rpsO gene was PCR amplified using species-specific primers with the T7-promoter sequence added within the forward primer sequence . Genomic DNA extracted from the each species was used as template . Indicated mutations were inserted to a DNA sequence using PCR primers containing the mutation . T7 RNA polymerase [60] was used to transcribe RNA and transcription reactions were purified by 6% denaturing PAGE . Bands were visualized using UV shadow , excised , and the RNA eluted ( in 200 mM NaCl , 1 mM EDTA ph 8 , 10 mM Tris-HCl pH 7 . 5 ) and ethanol precipitated . Purified RNA ( 10 pmol ) was 5’-labeled with 32P-ATP and purified as previously described [61] . pCp labeling was performed using T4 RNA ligase with 50 pmol RNA and 50 pmol of [32P]-pCp . 3’-labeled RNA was isolated using Ambion MEGAclear kit . The rpsO open reading frame was PCR amplified using whole genomic DNA and species-specific primers . It was cloned into pET-HT overexpression vector similarly to previously described [62] . Sequence verified plasmid was transformed into chemically competent BL-21 cells ( DE3 ) . Protein expression and purification for all four S15 homologs was conducted as described previously [37] . A fixed amount of 5’32P-labeled RNA ( 1000 cpm , <1 nM ) was renatured for 15 minutes at 42°C , then incubated with serial dilution of S15 in Buffer A ( 50 mM-Tris/Acetate , pH 7 . 5 , 20 mM Mg-acetate , 270 mM KCl , 5 mM dithiothreitol , 0 . 02% bovine serum albumin[63] ) for 30 minutes at 25°C . For RNAs originating from thermophillic organisms , assays were also conducted at 55°C , but these either did not yield a productive interaction , or the results were not significantly different from those observed at 25°C . Nitrocellulose membrane ( GE Healthcare ) was used to collect RNA-S15 complexes and positively charged nylon membrane ( GE Healthcare ) was used to collect unbound RNA under suction in a filter binding apparatus . Membranes were air-dried 5 minutes and the fraction bound quantified by imaging membranes on a phosphorimager screen . Radioactivity counts per sample on each membrane were measured using GE Healthcare STORM 820 phosphorimager and ImageQuant . For each sample the fraction bound ( Fb ) corresponds to the ( counts nitrocellulose ) / ( counts nitrocellulose + counts nylon ) . To determine the KD and the maximum fraction bound ( Max% ) , the resulting values were fit to the equation: Fb = ( Max%*[S15] ) / ( [S15]+KD ) where [S15] corresponds to the concentration of S15 in the reaction . The residuals were minimized using the Solver function in Microsoft Excel to find both the Max% and the KD . KD values given in Table 1 represent the mean of 3 or more independent binding assays ± the standard deviation . Amino acid sequences corresponding to the rpsO open reading frame from all bacterial species carrying each mRNA regulator were gathered based on existing RNA alignments [35–37] . These sequences were aligned using ClustalW [64] , and the alignments analyzed using Rate4site [53] .
|
RNA-protein interactions are important for many biological processes , and aberrant interactions have been implicated in many human diseases . However , how the specificity of such interactions may change over time and impact the evolution of both partners is still largely unexplored . In this work we examine four homologs of ribosomal protein S15 from diverse bacterial species . While S15 interacts with a conserved rRNA site , each S15 homolog regulates its own expression using a different mRNA structure . We measured the in vitro and regulatory interactions for all pairwise combinations between four S15 homologs and regulatory mRNA structures from diverse species . We find that each S15 homolog interacts with a different set of mRNA structures and their mutants . This work demonstrates that even highly conserved proteins such as S15 can have distinct RNA-binding repertoires that are likely the result of selection on both the mRNA and the protein sequence over time .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Co-evolution of Bacterial Ribosomal Protein S15 with Diverse mRNA Regulatory Structures
|
The two-component signalling system ( TCS ) comprising a histidine kinase ( HK ) and a response regulator ( RR ) is the predominant bacterial sense-and-response machinery . Because bacterial cells usually encode a number of TCSs to adapt to various ecological niches , the specificity of a TCS is in the centre of regulation . Specificity of TCS is defined by the capability and velocity of phosphoryl transfer between a cognate HK and a RR . Here , we provide genetic , enzymology and structural data demonstrating that the second messenger cyclic-di-GMP physically and specifically binds to RavS , a HK of the phytopathogenic , gram-negative bacterium Xanthomonas campestris pv . campestris . The [c-di-GMP]-RavS interaction substantially promotes specificity between RavS and RavR , a GGDEF–EAL domain-containing RR , by reinforcing the kinetic preference of RavS to phosphorylate RavR . [c-di-GMP]-RavS binding effectively decreases the phosphorylation level of RavS and negatively regulates bacterial swimming . Intriguingly , the EAL domain of RavR counteracts the above regulation by degrading c-di-GMP and then increasing the level of phosphorylated RavS . Therefore , RavR acts as a bifunctional phosphate sink that finely controls the level of phosphorylated RavS . These biochemical processes interactively modulate the phosphoryl flux between RavS-RavR and bacterial lifestyle transition . Our results revealed that c-di-GMP acts as an allosteric effector to dynamically modulate specificity between HK and RR .
The two-component signalling system ( TCS ) is one of the predominant molecular machineries used by almost all bacteria to monitor and adaptively respond to environmental cues [1 , 2] . The prototypical TCS is composed of a membrane-bound histidine kinase ( HK ) and a cytosolic response regulator ( RR ) . Upon detecting a stimulus , HK autophosphorylates an invariant histidine residue within its dimerization and histidine phosphotransfer ( DHp ) domain and then catalyses the transfer of the phosphoryl group onto a conserved aspartic acid within the receiver ( REC ) domain of the cognate RR [3] . The activated RR then modulates bacterial adaptation by controlling gene transcription or cellular behaviour [4 , 5] . There is a high level of specificity between a HK and its cognate RR , which is quantified by the kinetic preference during phosphotransfer [6 , 7] . The complexity of TCS regulation was revealed after three decades of extensive investigations . Bacterial cells dynamically and elegantly regulate time , rhythm , space and flux of the phosphotransfer between the HK and RR to adapt to diverse ecological niches [8] . For example , hybrid-type HK-mediated phosphorelay , phosphatases , auxiliary proteins and small RNAs are involved in TCS regulation [9–13] . Cyclic di-GMP ( c-di-GMP ) is a ubiquitous second messenger involved in regulating bacterial physiology [14–17] . This signalling chemical interacts with riboswitches or protein as effectors to control a diverse set of processes , such as virulence , biofilm formation , motility , cell division and quorum-sensing [18–20] . The c-di-GMP effectors include the PilZ family of proteins , transcription factors , ATPases , transporters and various metabolic enzymes [15 , 21 , 22] . Intrinsically , the TCS is extensively involved in c-di-GMP signalling because approximately 6% of bacterial RRs encode three types of protein modules that control c-di-GMP turnover [18] . Among them , GGDEF domains have diguanylate cyclase activity to biosynthesize c-di-GMP from two GTP molecules [23] , whereas EAL or HD-GYP domains contains phosphodiesterase activities to degrade c-di-GMP into pGpG or GMP , respectively [23–25] . Some RRs encoding degenerate GGDEF or EAL domains can act as c-di-GMP receptors [14 , 26] . Recently , c-di-GMP was found to interact directly with CckA , an essential HK regulating cell division of Caulobacter crescentus , to inhibit kinase activity and activate phosphatase activity of CckA , which triggers DNA replication and initiation of the cell cycle [27–29] . This evidence indicates that c-di-GMP and TCS signalling are tightly associated . However , how c-di-GMP modulates the specificity between HKs and RRs and the biological significance of this signalling crosstalk remain unresolved . The gram-negative bacterium Xanthomonas campestris pv . campestris is a model organism for studying plant pathology . This bacterium is the causative agent of black rot disease in various cruciferous plants and causes substantial loss of vegetable production yields worldwide [30] . The X . campestris pv . campestris genome encodes a large number of TCS proteins ( approximately 52 HKs and 54 RRs ) [31] , and a number of these proteins , including RpfC-RpfG , RavS-RavR and PcrK-PcrR , are regulators of bacterial virulence that are associated with c-di-GMP turnover [32–36] . Among these regulators , RavS-RavR was reported to constitute a putative TCS . RavS is a membrane-bound HK with two tandem PAS domains as potential intracellular sensors , and RavR is a RR that contains N-terminal degenerate GGDEF and EAL domains [37] . In the vicinity of the ravR-ravS locus , a HK gene named ravA was identified and encodes a protein that may phosphorylate RavR [38] . Inactivation of ravA and ravR causes a significant decrease in virulence [37 , 38] . Therefore , RavS-RavA-RavR is likely to form a “three-component signalling system” that most likely has complex associations and outputs . However , the regulatory relationship and protein phosphorylation processes among RavS-RavA-RavR , and the potential role of c-di-GMP in the dynamics of the signalling system , remain unknown . In this study , we reveal that c-di-GMP regulates the HK-RR phosphotransfer flux . c-di-GMP physically binds to the catalytic and ATP-binding domain ( CA ) of RavS to significantly enhance the phosphotransferase activity of RavS towards RavR . This process efficiently decreases the phosphorylation level of RavS . In controlling bacterial swimming and flagellar development , the REC domain of RavR acts as a highly efficient phosphate sink to decrease the phosphorylated RavS level , which is subject to the control of c-di-GMP . Epistasis analysis also revealed that unlike typical TCSs , the RR RavR is upstream of the HK RavS in regulation . In addition , the EAL domain of RavR degrades c-di-GMP to reduce the RavS-RavR phosphotransfer , thereby maintaining a high level of phosphorylated RavS . Therefore , RavR subtly modulates the phosphorylation level of RavS , which is critical in controlling bacterial swimming and flagellar biogenesis . In this process , c-di-GMP controls the phosphotransfer flux from the HK to the RR and plays a critical role during the lifestyle transition between swimming and bacterial virulence .
The genomic organization of ravS , ravR and ravA and the putative structures of their products are shown in Fig 1A . Western blotting revealed that RavS is located in the membrane and cytoplasm , and that RavR is a cytoplasmic protein ( Fig 1B ) . Although RavA does not contain a recognisable transmembrane helix , repeated experiments showed that RavA is also located in both the membrane and cytoplasm ( Fig 1B ) , which suggests that the HK has a membrane-anchored peptide or that the protein contains an unrecognised transmembrane region . Semi-quantitative western blotting demonstrated that the estimated ratio of RavA:RavR:RavS is approximately 10:29:17 , amounting to approximately 1890 , 5510 and 3230 molecules , respectively , in a bacterial cell grown under the tested conditions ( Fig 1C and S1 Fig ) . The estimated concentrations of RavA , RavR and RavS in a cell were 2 . 05 ± 0 . 53 , 5 . 99 ± 0 . 46 , and 3 . 51 ± 0 . 27 μM , respectively ( average bacterial cell volume was defined as 1 . 53 ± 0 . 23 μM3 ) . This HK:RR ratio is remarkably higher than the general trend observed that RR molecules are often several dozen times more than the HK [39] . To further investigate the roles of ravA , ravR and ravS in controlling physiological processes , we performed phenotypic profiling of the in-frame deletion mutants of ΔravA , ΔravR and ΔravS . These mutants have phenotypes similar to those of the wild-type ( WT ) strain in growth in rich NYG medium , secretion of extracellular enzymes and resistance to oxidative stress ( S2 Fig ) . However , these mutants exhibited remarkable but differing phenotypic changes in bacterial virulence and motility . Deletion of ravA and ravR substantially decreased bacterial virulence against the host plant Brassica oleracea cv . Zhonggan 11 ( Fig 1D and 1E ) . Genetic complementation of the two genes , whose expression levels were controlled by their native promoters , fully or partially restored virulence to levels of the WT strain ( Fig 1D and 1E ) . Although deletion of ravS has no impact on bacterial virulence , a genetic complementary strain ( ΔravS-ravS ) showed significant attenuation in virulence ( Fig 1D and 1E ) , which might be caused by overexpression of ravS in a medium-copy , broad-host vector ( pHM2 ) . Accordingly , the production of extracellular polysaccharides ( EPS ) , a major virulence determinate of X . campestris pv . campestris , decreased significantly to 43% and 42% of the levels of the WT strain in the ΔravA and ΔravR mutants , respectively , whereas the EPS levels were similar to the WT strain in their genetic complementary strains ( Fig 1F ) . EPS production remained nearly unchanged in the ravS null mutant but decreased significantly in the ravS complementary strain ( 31% compared with that of the WT strain , ( Fig 1F ) ) . These results indicate that ravA and ravR are positive regulators of bacterial virulence but overexpression of ravS decreased bacterial virulence . As shown in Fig 1G and 1H , compared with the WT , deletions of ravA and ravR increased swimming motility by approximately two-fold , but deletion of ravS almost completely eradicated the swimming capability of the mutant . Genetic complementation successfully restored the observed deficiencies to levels similar to that of WT strain ( Fig 1G and 1H ) . Since bacterial swimming motility is controlled mainly by flagella , bacterial flagella were observed by transmission electron microscopy ( TEM ) . As measured by the flagellar cell ratio ( Fig 2A and 2B ) , ravA ( flagellar cell ratio = 71 . 6% ) and ravR ( 77 . 1% ) mutant strains had a higher frequency of flagella in cells and significantly longer flagella than observed in the WT strain ( 32 . 5% ) . The average flagellar lengths of ΔravA and ΔravR mutant strains and the WT strain were 4 . 60 , 4 . 84 and 2 . 82 μm ( Fig 2C ) , respectively . However , inactivation of ravS resulted in complete loss of flagella ( Fig 2A–2C ) . Genetic complementation of ravA , ravR or ravS in the corresponding mutant significantly suppressed these deficiencies in flagellar morphology . ravA-ravR-ravS is located in the vicinity of gene clusters that encode flagella-associated proteins and may regulate the expression of these genes ( S3 Fig ) . We selected eight genes to examine their transcription levels in various strains and found that fliC ( encoding flagellin ) is a representative gene that is controlled by ravA-ravR-ravS ( S3A Fig ) . Deletion of ravA and ravR significantly increased the mRNA levels of fliC to 363% and 402% of the levels of the WT strain , respectively ( Fig 2D ) . In contrast , mutation of ravS caused a significant decrease in the fliC transcription level to 5 . 6% of the level of the WT strain , while genetic complementation of ravS increased fliC mRNA to 253% of the level of the WT strain ( Fig 2D ) . Based on these results , in controlling the development of bacterial flagella and fliC transcription , both ravA and ravR act as negative regulators but ravS is a positive regulator . In addition , because type III secretion ( T3SS ) genes ( hrp ) are critical in virulence of X . campestris , we selected three hrp cluster genes ( hrpD , hrpF and hrpG ) to measure their expression levels . As shown in S3C Fig , the mRNA levels of hrpF did not have remarkable change in the ravR and ravA mutants . However , the amounts of hrpG and hrpD transcripts were significantly decreased comparing to the WT strain . Therefore , we further analysed the expression level of hrpG because HrpG is an OmpR-family response regulator to control the transcription of an AraC-family transcription factor HrpX , and the latter binds to the plant inducible promoters ( PIP ) of some hrp gene clusters to modulate the expressions of these T3SS genes [30 , 40] . In the ravR and ravA mutants , the hrpG transcripts significantly decreased to the 5 . 2% and 2 . 2% levels of the WT strain , while genetic complementation of the two genes restored the hrpG expression to 111 . 1% and 75 . 6% levels of the WT strain ( S3D Fig ) . This result reveals that ravR and ravA positively modulate the transcription of hrpG . Epistatic analysis was performed to dissect the regulatory relationships among ravR and ravA or ravS . Although deletion of ravA or ravR significantly decreased the bacterial virulence and production of EPS ( Fig 1 ) , the double mutants ( ΔravA-ΔravS or ΔravR-ΔravS ) exhibited virulence levels similar to those of the WT strain ( Fig 3A and 3B and S4A Fig ) . Similarly , inactivation of ravS effectively suppressed the increased swimming motility , flagellar cell ratio and flagellar length that were caused by ravA or ravR deletion ( Fig 3C–3E , and S4B and S4C Fig ) . The fliC expression level of the ravA or ravR mutant was increased significantly ( Fig 2D ) . However , in the double mutants , fliC mRNA levels decreased significantly to the level of ΔravS ( Fig 3F ) . Furthermore , two double mutants on ΔravA and ΔravR backgrounds were constructed with the conserved phosphorylation residue ( His500 ) of RavS substituted by Ala ( ΔravA-ravSH500A and ΔravR-ravSH500A ) . The phenotypic changes in the single mutant ravSH500A were similar to those observed for the ΔravS mutant , and in the double mutants , analyses revealed that the ravSH500A point mutation effectively suppressed all the tested phenotypic deficiencies that were caused by the ravA or ravR deletion ( Fig 3A–3F and S4A–S4C Fig ) . Collectively , these genetic results suggest that ravS is downstream of ravR and ravA in regulation . The point mutant ravSH500A , in which RavS is constitutively dephosphorylated , exhibits reduced swimming capacity , which suggests that a high level of phosphorylated RavS ( RavS~P ) positively controls bacterial swimming but negatively modulates virulence . Furthermore , the suppressive function of ravS depends completely on the function of the conserved His500 residue , which suggests that RavA and RavR may negatively regulate the RavS~P level in the epistasis relationship . Intriguingly , the RR RavR is upstream of RavS ( HK ) in regulation . The protein has an EAL domain and a GGDEF domain that associated with c-di-GMP turnover . Using 32P-labelled c-di-GMP produced by a tDGC cyclase ( S5A Fig ) [41] , we confirmed that the EAL domain of RavR can form a homodimer in vitro and has phosphodiesterase activity to degrade c-di-GMP into pGpG , which depends on the E-A-L motif of the domain ( Fig 4A , S5B Fig and S5C Fig ) . The GGDEF signature motif of RavR is degenerate as “GSDEM” and did not have diguanylate cyclase activity to synthesize c-di-GMP from GTP ( S5B Fig ) , suggest that RavR does not encode cyclase activity . Phosphorylation of RavR by RavA did not affect its phosphodiesterase activity ( S5D Fig and S5E Fig ) , consistent with the result that inactivation of two HK genes ( ravS or ravA ) did not remarkably affect the intracellular concentration of c-di-GMP ( S5F Fig ) . Additionally , deletion of the coding sequences of the EAL domain or ravR gene increased the intracellular c-di-GMP to 185 . 9% and 150 . 0% of the level of the WT strain , respectively ( Fig 4B ) . These results suggest RavR mainly acts as a phosphodiesterase . We further genetically analysed the interaction between the phosphor-receiving domain ( REC ) and the EAL domain of RavR and their regulatory relationship in controlling ravS . Similar to the null mutant of ravR ( ΔravR ) , point mutation of the conserved Asp496 residue within the REC domain of RavR decreased bacterial virulence and EPS production ( Fig 4C–4E ) but increased bacterial swimming ( Fig 4F and 4G ) , flagellar cell ratio ( Fig 4H and 4I ) , flagellar length ( Fig 4J ) and the transcriptional level of fliC ( Fig 4K ) , which suggests that phosphorylation residue Asp496 of RavR is indispensable for regulation . Surprisingly , inactivation of RavR phosphodiesterase activity by an in-frame deletion of the coding sequence of its EAL domain ( ravRΔEAL ) caused phenotypic changes that were completely different from those of the ravR null mutant ( ΔravR ) or ravRD496A point mutant . The virulence level and EPS production of ravRΔEAL were unaffected ( Fig 4C–4E ) rather than reduced . The swimming motility of the ravRΔEAL mutant was reduced significantly ( Fig 4F and 4G ) rather than showing a dramatic increase that was observed for ΔravR or the ravRD496A mutant . TEM analysis revealed the absence of flagellum in the bacterial cells of the ravRΔEAL mutant; thus , its flagellar cell ratio was zero ( Fig 4H–4J ) . Additionally , the fliC mRNA level of the ravRΔEAL mutant was reduced significantly ( Fig 4K ) . We then constructed a double mutant of ravRΔEAL ( D496A ) to investigate the epistatic relationship between the EAL domain and Asp496 . ravRΔEAL ( D496A ) phenocopied the ravRD496A mutant displaying a decrease in virulence and EPS production but increases in swimming capability , flagellar length , flagellar cell ratio and fliC transcription level ( Fig 4A–4K ) . This genetic evidence suggests that RavR has two opposite functions in regulating bacterial swimming . Full-length RavR negatively controls swimming and flagellar development in an Asp496-dependent manner , whereas its EAL domain positively controls swimming and flagellar development . Additionally , epistatic analysis revealed that the phosphorylation of Asp496 potentially downstream of the PDE activity of EAL domain in regulating bacterial swimming . The biochemical relationship will be dissected in the following experiments . Based on the above results , the genetic code of the conserved His500 residue of the downstream HK gene ravS was mutated on the backgrounds of ravRD496A and ravRΔEAL ( D496A ) to construct triple mutants . As predicted , the pattern of phenotypic changes in these three triple mutants was suppressed by the mutation and similar to that of ΔravS or ravSH500A ( Fig 4A–4K ) . The phosphorylation process and regulation of RavA-RavR-RavS have not been investigated previously . In vitro phosphorylation assays using autoradiography and Phos-tag gel analysis revealed that full-length RavA autophosphorylates its conserved RavAHis164 residue because the RavAH564A substitution completely inactivated RavA autophosphorylation ( S6A Fig , S6B Fig and S6D Fig ) . The addition of RavR or RavRΔEAL into the reaction mixture remarkably decreased the RavA phosphorylation level ( RavA~P ) , although no recognisable signal of phosphorylated RavR ( RavR~P or RavRΔEAL~P ) was detected in autoradiography ( S6A Fig , S6B Fig and S6D Fig ) . Recombinant RavR with a conserved Asp496 substitution did not cause a decrease in the RavA~P level ( S6A Fig , S6B Fig and S6D Fig ) , which suggests that RavA transfers the phosphoryl group onto Asp496 of RavR and that RavR~P is highly unstable . Unlike isotope-labelling method , Phos-tag gel analysis successfully detected a RavR~P band in the RavA-RavR phosphotransfer reaction under an extremely high ATP concentration ( 2 mM , S6D Fig , lane 5 ) , providing an evidence to demonstrate that RavA phosphorylates RavR . For RavS , repeated efforts failed to detect any signal representing the autophosphorylated RavS-containing cytosolic region ( RavSΔTrM , S6C Fig ) . However , a truncated RavS that lacks the N-terminal transmembrane helix and the two PAS domains ( RavSΔN ) exhibited strong autokinase activity in the presence of ATP ( Fig 5A , lane 1 and S6C Fig ) . The addition of RavRΔEAL to the reaction mixture slightly decreased the level of RavS~P after 30 min of co-incubation ( Fig 5A , lane 2 , S6E Fig , lane 3 ) . No signals representing RavR~P were observed by the isotope-labelling and Phos-tag gel analysis ( Fig 5A , lane 3 , S6E Fig and S6F Fig ) and substitution of the conserved Asp496 of RavR ( RavRΔEAL ( D496A ) ) eliminated its activity to decrease the RavS~P level . Together , these biochemical analyses demonstrated that RavA and RavS are bona fide HKs . RavA can phosphorylate RavR at residue Asp496 , and RavR~P is highly unstable . Since ravS is downstream of ravR during the regulation of bacterial swimming and RavR is a functional phosphodiesterase that degrades c-di-GMP , we reasoned that c-di-GMP modulates the activity of RavS in controlling bacterial swimming . We analysed the impact of c-di-GMP on the enzymatic activities of RavS to test this hypothesis . In the absence or presence of c-di-GMP ( Fig 5A , lane 1 and lane 4 ) , the autokinase activity of RavSΔN was unchanged , which suggests that c-di-GMP does not affect RavSΔN autophosphorylation . However , when RavRΔEAL was added to the RavS-RavR phosphotransfer reaction , the presence of c-di-GMP in the mixture remarkably increased RavS-RavR phosphotransfer because the RavSΔN~P level immediately decreased when compared with the level in the absence of c-di-GMP ( Fig 5A , lane 5 vs . lane 2 , and Fig 5B ) . The time ( T50 ) that signal intensity of the RavS~P decreased to the half of the initiation stage ( before adding RavRΔEAL ) was estimated: without c-di-GMP , the T50 value of RavS~P is 38 . 2 min . When 15 μM c-di-GMP is in presence , the T50 value of RavS~P remarkably decreased to 16 . 9 min . Since some HKs or RRs have phosphatase activity to dephosphorylate HKs , it is possible that c-di-GMP positively regulates RavS or RavR phosphatase activity to decrease the RavS~P level . However , addition of c-di-GMP alone did not affect RavSΔN~P levels ( Fig 5A , lane 4 vs . lane 1 , and Fig 5B ) and the addition of an inactive RavRΔEAL ( D496A ) did not have any impact on RavS phosphorylation levels , regardless of the presence or absence of c-di-GMP ( Fig 5A , lanes 3 and 6 ) . These results indicate that c-di-GMP enhanced significantly RavS phosphotransferase activity towards RavR . This c-di-GMP-triggered kinetic preference effectively decreases the RavS~P level . Along with the elevation of c-di-GMP concentrations , the RavS-RavR phosphotransfer , which was measured by the intensity of the RavS~P signal , gradually increased . This suggests that the impact of c-di-GMP on RavS autokinase activity is dose-dependent ( Fig 5C ) . The estimated value of the median effect concentration ( EC50 ) of c-di-GMP was approximately 15 . 58 μM ( Fig 5D ) . Furthermore , the regulation of c-di-GMP on the RavS-RavR phosphotransfer to decrease the RavS~P level is highly specific because that other nucleotides and derivatives , including GTP , GMP , cGMP and c-di-AMP , did not have any impact on RavS phosphotransferase activity ( Fig 5E ) . Next , microscale thermophoresis ( MST ) was used to determine whether c-di-GMP binds directly to RavS . 2′-Fluo-AHC-c-di-GMP ( fl-c-di-GMP ) was used in the assay since the analogue contains a carboxyfluorescein that has excitation and emission wavelengths of 497 nm and 520 nm , respectively , which can be detected directly by a MST instrument [42] . As shown in Fig 5F , fl-c-di-GMP directly bound a truncated , soluble RavS protein lacking the transmembrane region ( RavSΔ ) . The dissociation constant was 6 . 06 ± 0 . 92 μM , which is in the range of physiological concentrations of c-di-GMP in bacterial cells [22] . As controls , 2′-Fluo-AHC-cGMP ( fl-cGMP ) or 2′-Fluo-AHC-c-di-AMP ( fl-c-di-AMP ) bound RavSΔTrM with weaker affinity ( fl-cGMP , 64 . 4 ± 9 . 43 μM ) or completely failed to bind the protein ( fl-c-di-AMP , Fig 5F ) . Furthermore , to determine the c-di-GMP binding region on RavS , we expressed and purified various RavS fragments ( S7A Fig ) . MST assays revealed that fl-c-di-GMP bound RavS with the DHp-CA or CA region , whereas this second messenger did not bind the PAS-A , PAS-B and DHp regions of RavS ( Fig 6A and 6B ) . Of note , a lower binding affinity of c-di-GMP with the CA domain than the DHp-CA domain was observed , which suggests that the DHp domain indirectly affects the interaction ( Fig 6B ) . MST analysis also revealed that fl-c-di-GMP did not bind to RavRΔEAL , RavA and an unrelated HK VgrS since the Kd values were greater than 80 μM ( S7B Fig ) , which did not have physiological significance . Therefore , these results revealed that c-di-GMP specifically binds to RavS and the CA domain of RavS is the possible interacting region of c-di-GMP . Similar to most HKs , the three-dimensional ( 3D ) structure of RavS has not been elucidated . As an alternative , molecular docking was used to predict the structural interaction of the RavS–c-di-GMP complex . We predicted the DHp-CA ( residues E464–R697 ) structure of RavS by homology modelling with the crystal structures of other HKs , such as CpxA-HDC ( PDB ID: 4BIU ) , DivL ( 4EW8 ) and VraS ( 4GT8 ) , as templates . As shown in Fig 6C , the putative structure of DHp-CA is composed of seven α-helices , four β-strands and random crimps . The active centre includes the helices of α3 , α4 , α5 and α6 and strands β3 and β4 , and is highly similar to other DHp-CA domains of HKs . The 3D structure of c-di-GMP was extracted from the crystal structure of the c-di-GMP-VCA0042 complex ( PDB ID: 2RDE ) [43] . Autodock 4 . 0 software was used for molecular docking . The results showed that helices α4 and α7 and strands β3 and β4 of the RavS DHp-CA domain formed a c-di-GMP binding pocket and 10 residues in the vicinity of the pocket potentially interact with c-di-GMP ( Fig 6C and 6D ) . Among these residues , Ser636 , Asp638 , Asp640 , Ile653 , Glu657 and Leu666 putatively interact with c-di-GMP via hydrophobic interactions and van der Waals forces , whereas Asp637 , Gly639 , Leu652 and Arg656 form hydrogen bonds with c-di-GMP ( Fig 6D ) . In contrast to the other residues , the amino hydrogens of Arg656 and the phosphate group oxygens of c-di-GMP formed two hydrogen bonds with bond lengths of 2 . 3 Å and 2 . 5 Å , respectively ( Fig 6D ) . These results suggest that c-di-GMP potentially interacts with the CA region of RavS and that Arg656 is a key residue in the interaction . To experimentally verify the above prediction , we expressed and purified 10 recombinant DHp-CA proteins in which each of the predicted c-di-GMP-interacting amino acids were separately substituted to Ala . MST assays revealed that the recombinant proteins DHpCAD637A , DHp-CAG639A , DHp-CAD640A , DHp-CAI653A , DHp-CAE657A and DHp-CAL666A exhibited no significant change in c-di-GMP binding affinity . In contrast , DHp-CAS636A , DHp-CAD638A and DHp-CAL652A exhibited 1–2-fold increases in Kd values when compared with that of the positive control ( DHp-CA protein ) , which suggests that these substitutions disrupt the protein-nucleotide interaction . Notably , the Kd value of the c-di-GMP–DHp-CAR656A interaction was 62 . 4 ± 3 . 19 μM , which showed a 10-fold decrease in binding affinity ( Fig 6E ) . Consistent with this result , an in vitro phosphotransfer assay revealed that c-di-GMP failed to stimulate the phosphotransferase activity of RavSΔN ( R656A ) ~P towards RavRΔEAL and the RavSΔN ( R656A ) ~P level did not change noticeably or decreased slightly in the presence of c-di-GMP ( 15 or 100 μM , Fig 6F ) . Thermal shift assay revealed that the DHp-CAR656A is as stable as DHp-CA protein ( S8A Fig ) , and RavSΔN and RavSΔN ( R656A ) have similar autokinase activity ( S8B Fig ) , suggesting that the Arg656 substitution did not remarkably affect the structure of the recombinant protein . Collectively , these results demonstrate that Arg656 of RavS is a key residue in c-di-GMP binding . Substitution of Arg656 leads to dissociation between c-di-GMP and RavS such that the RavS-RavR phosphotransfer is not accelerated . Under these circumstances , the majority of RavS remains phosphorylated . Based on the above results , we constructed a ravSR656A point mutant to determine whether the dissociation of c-di-GMP from RavS affects bacterial phenotypes . Western blotting revealed that RavS is stable in this mutant ( S8C Fig ) . As shown in Fig 7A and 7B , the virulence of the ravSR656A mutant was attenuated substantially and the production of EPS by this mutant had also decreased significantly ( Fig 7C ) . The swimming zone diameter of the ravSR656A mutant was 2 . 38 ± 0 . 10 cm , which was much larger than that of the WT strain ( 1 . 16 ± 0 . 08 cm ) ( Fig 7D and 7E ) . Therefore , we measured the parameters of bacterial flagella and found that the ravSR656A mutant had longer flagella ( 4 . 79 ± 1 . 31 μm , Fig 7F and 7H ) . The ravSR656A mutant flagellar cell ratio was two-times the level of the WT strain ( Fig 7G ) and the fliC mRNA level had increased by 292% when compared with that of the WT strain ( Fig 7I ) . Therefore , the pattern of phenotypic changes of the ravSR656A point mutant is similar to that of the ravS overexpression strain ( ΔravS-ravS; S9A–S6F Figs ) but is quite different from that of the ravS deletion or ravSH500A mutant whose virulence and EPS production were unaffected , whereas the tested parameters of swimming motility were reduced significantly ( Fig 3 ) . Thus , these results suggest that the ravSR656A substitution , which eliminated the binding capability of c-di-GMP , mimics the constitutive phosphorylated state of RavS . Subsequently , the conserved His500 residue was mutated on the genetic background of ravSR656A . Compared with the ravSR656A mutant , the double mutant ravSR656A-H500A exhibited restored virulence and EPS production levels ( Fig 7A–7C ) . The swimming motility ( Fig 7D and 7E ) , flagellar cell ratio ( Fig 7F and 7G ) , flagellar length ( Fig 7H ) and fliC mRNA levels ( Fig 7I ) of the double mutant were all significantly reduced or undetectable . Therefore , point mutation of His500 completely suppressed the mutational effects caused by the Arg656 point mutation . Together with the biochemical results demonstrating that c-di-GMP did not bind ravSR656A and that the phosphorylation level of RavSR656A is insensitive to c-di-GMP inhibition , these findings suggest that c-di-GMP associates with the Arg656 residue to negatively regulate the RavS phosphotransferase activity towards RavR . The aforementioned genetic and biochemical results prompted the following hypothesis: a high level of RavS~P positively regulates bacterial swimming but is detrimental to virulence; however , dephosphorylation of RavS reduces the expression of fliC and is negative in modulating bacterial swimming . Since c-di-GMP binds to RavS to inhibit the phosphotransferase activity of RavS towards RavR , RavR is likely to be a special phosphate sink that controls the RavS~P level through opposite mechanisms: 1 ) the REC domain of RavR receives the phosphoryl group from RavS~P , which decreases the RavS~P level and then negatively regulates bacterial swimming . Additionally , c-di-GMP inhibits bacterial swimming by further enhancing the RavS-RavR phosphotransfer; and 2 ) the EAL domain of RavR degrades c-di-GMP and then inhibits the c-di-GMP-triggered RavS-RavR phosphotransfer . This process prefers to maintain the RavS~P level and positively regulates bacterial swimming . This hypothesis is true when constitutive dissociation by genetic manipulation of the binding event between c-di-GMP and RavS results in a high RavS~P level . As a consequence , this process will positively modulate bacterial swimming independent of the c-di-GMP concentration in cells . To test this hypothesis , we point mutated Arg656 of ravS to Ala to construct a double mutant on the ravRΔEAL genetic background ( ravRΔEALravSR656A ) . The RavR phosphodiesterase activity of this mutant was genetically inactivated and recombinant RavS did not bind c-di-GMP because Arg656 is essential for the protein-[c-di-GMP] interaction . As shown in Fig 7 , the point mutation of ravSR656A effectively suppressed the phenotypic effects ( i . e . , virulence and EPS production ) caused by the EAL domain deletion ( Fig 7A–7C ) and bacterial swimming was restored completely ( 1 . 48 ± 0 . 08 cm; Fig 7D and 7E ) . The number of flagella and flagellar cell ratio were increased significantly when compared with the level of the WT strain ( Fig 7F–7H ) . The fliC mRNA level of the ravRΔEALravSR656A mutant was also significantly increased from 9 . 4% ( ravRΔEAL ) to 199% of the level of the WT strain ( Fig 7I ) . Similarly , overexpression of ravS on the ravRΔEAL ΔravS background ( ravRΔEALΔravS-OE-ravS; S9A Fig ) , which mimics constitutive phosphorylation of RavS , also effectively suppressed the mutational effects of the EAL domain deletion on bacterial swimming ( S6 Fig ) . Furthermore , the suppression of the ravSR656A mutation and ravRΔEAL deletion completely depends on the conserved His500 site of RavS . The triple mutant of ravRΔEALravSR656A ( H500A ) exhibited performance similar to that of ravRΔEAL in all of the tested parameters associated with swimming or flagella development ( Fig 7A–7I ) . Collectively , these genetic analyses support the hypothesis regarding the regulation of RavR in controlling the RavS~P level . Asp469 within the REC domain of RavR receives the phosphoryl group from RavS to reduce the RavS~P level , whereas the EAL domain of RavR degrades c-di-GMP , which binds to RavS and promotes the specificity between RavS-RavR by broadening the phosphotransfer flux . Dissociation of RavS–c-di-GMP leads to enhanced bacterial swimming , which depends on the high phosphorylation level of RavS .
Kinetic preferences between HKs and RRs define the specificity of TCS regulation [7 , 44] , which is quantified by the velocity of a HK to phosphorylate a RR [6] . In the present study , genetic , biophysical and biochemical data revealed that c-di-GMP substantially enhanced the specificity between RavS and RavR of X . campestris pv . campestris . RavS physically interacts with c-di-GMP via its CA region . This interaction remarkably accelerated phosphoryl transfer from RavS~P towards RavR , resulting in efficient dephosphorylation of RavS~P . Intriguingly , the EAL domain-containing RavR acts as a phosphate sink of RavS~P . Stimulated by c-di-GMP , the REC domain of RavR receives a phosphoryl group from RavS to decrease the RavS~P level , which negatively modulates bacterial swimming . However , the EAL domain of RavR degrades c-di-GMP . Degradation of c-di-GMP revokes the accelerating effect of c-di-GMP on RavS-RavR phosphotransfer , leading to a high level of RavS~P , which positively regulates bacterial swimming ( Fig 8 ) . Therefore , these results suggest that RavS is a pivotal node for integrating the regulatory pathways of c-di-GMP and TCS , and RavR is a bifunctional regulator that finely controls the phosphorylation state of RavS ( Fig 8 ) . The crosstalk between c-di-GMP and RavS–RavR phosphoryl transfer regulates a bacterial lifestyle transition from virulence to freely swimming . c-di-GMP plays a pleiotropic role in controlling the physiological processes of bacterial cells . This cyclic dinucleotide is a versatile , flexible molecule and the regulatory function of c-di-GMP is usually exerted by binding to a diverse array of effectors to elicit allosteric changes [14 , 18 , 45] . However , the intracellular concentration of c-di-GMP is complex and changed dynamically [46] . The local , subcellular c-di-GMP concentrations are critical and have profound effects on physiological processes [14] . Currently , the reported Kd values between c-di-GMP and proteins range from 0 . 1 μM ( PizD ) to 14 . 5 μM ( STING ) [22] . Our MST assay revealed that c-di-GMP physically binds to RavS in vitro with a Kd of 6 . 06 μM ( Fig 5F ) , which is similar to the previously identified effector CLP of Xanthomonas campestris pv . campestris [47] , suggesting that RavS is a bona fide c-di-GMP effector . The binding of c-di-GMP to RavS substantially increased its phosphotransferase activity towards RavR but did not affect RavS autokinase and phosphatase activities ( Fig 5A and 5B ) . In TCS regulation , phosphotransfer from HK to RR is generally controlled by their specificity , which is defined by key ‘specificity residues’ in these proteins [6 , 44] . However , our results showed that c-di-GMP did not bind the RR RavR ( S7B Fig ) , thus excluding the possibility that c-di-GMP increases RavR phosphatase activity towards RavS~P . Since the DHp and CA domains of HK are associated , we reasoned that the binding of c-di-GMP to the CA domain confers a conformational switch in the DHp domain , which leads to more efficient phosphoryl transfer from the DHp domain to the REC domain of RavR . Details on c-di-GMP-mediated enhancement of phosphoryl transfer require further structural analysis . According to previous studies , c-di-GMP binds to diverse regions of proteins , such as the degenerate GGDEF and EAL domains , PilZ domain and the interdomain linker [22] . In the absence of the RavS structure , we employed a molecular docking method together with mutational analysis to define residues essential to interaction with c-di-GMP ( Fig 6 ) . The analysis suggests that Arg656 within the helix α7 region of the CA domain potentially forms two hydrogen bonds with c-di-GMP ( Fig 6C and 6D ) . Substitution of this residue eliminated RavS–c-di-GMP binding in vitro ( Fig 6E ) . Additionally , point mutant ravSR656A exhibited enhanced bacterial swimming , elongation of flagella , a higher ratio of flagellar cells and increased fliC transcription levels ( Fig 7 ) . Similar to the overexpression of ravS ( S9 Fig ) , the ravSR656A mutation suppressed completely the mutational effects of the ravRΔEAL deletion that decreased swimming motility ( Fig 7 ) . Therefore , Arg656 within the CA region is an essential residue for binding c-di-GMP . This finding is in accordance with the preference of Arg or Asp/Glu residues to form hydrogen bonds with the phosphate group of c-di-GMP [22] . Compared with CckA of C . crescentus , which also binds to c-di-GMP , both HKs contain two PAS domains and a DHp-CA core; however , RavS lacks the C-terminal REC domain and is not a hybrid-type HK . When interacting with c-di-GMP , the autokinase activity of CckA was inhibited but its phosphatase activity was increased [27–29] . However , c-di-GMP binding did not affect the RavS autophosphorylation . It increases the phosphoryl flux from RavS towards RavR . The structural basis of this difference is unclear . Intriguingly , RavR plays a special role in regulation . Previous studies on c-di-GMP turnover found a general trend in which an elevated cellular concentration of c-di-GMP leads to a reduction in motility behaviours [14 , 18 , 48] . Therefore , inactivation of EAL domain-containing proteins usually increases c-di-GMP concentrations and inhibits bacterial swimming . In this study , deletion of the ravR gene increased the cellular concentration of c-di-GMP ( Fig 4B ) but this increase promoted bacterial swimming substantially rather than reducing this activity ( Fig 1G ) . Further investigation revealed that deletion of the EAL domain of RavR caused completely opposite phenotypic changes as the swimming zone was significantly reduced ( Fig 4F ) . This phenotype led us to examine the function of the EAL and REC domains of RavR . We found that RavR acts as a phosphate sink of RavS . Since the point mutation of ravRD496A restored the reduced swimming motility caused by EAL deletion ( Fig 4F–4K ) , the Asp496-dependent phosphoryl-receiving activity of RavR functions to regulate RavS~P levels by c-di-GMP-triggered RavS-RavR phosphotransfer , rather by directly modulating the PDE activity of the EAL domain , as revealed by S5D–S5F Fig . Based on these findings , the EAL domain degrades cellular c-di-GMP to release the negative regulation of c-di-GMP at the RavS~P level by promoting RavS-RavR phosphotransfer . Therefore , RavR is a bidirectional regulator that controls the RavS~P level through degrading c-di-GMP and accepting phosphoryl group from RavS . As genetic and biochemical analyses revealed , both HKs RavA and RavS phosphorylate RavR , and all three proteins control the virulence and swimming of X . campestris pv . campestris ( Figs 1 and 5 ) . Among these proteins , RavA and RavR compose a typical TCS that regulates bacterial virulence . As a previous study revealed , ravR is downstream of ravA in regulation [37 , 38] . In X . oryzae pv . oryzae , a close relative of X . campestris pv . campestris , ravA-ravR orthologues ( pdeK-pdeR ) have the same relationship in modulating bacterial blight disease against rice ( Orzya sativa ) [49] . As mentioned above , ravR is upstream of ravS in regulating bacterial swimming because RavR acts as a phosphate sink to dephosphorylate RavS ( Fig 3 ) . Our results further suggest that the regulation of bacterial virulence by RavA–RavR also depends on RavS because the ravS mutation suppressed the deficiencies of ravA and ravR mutants in virulence as well as in swimming ( Figs 3 and 7 ) . Since RavS and RavA are HKs that detect environmental and intracellular stimuli , it is important to know what signals were monitored by the two HKs . Because RavA lacks a recognizable signal input domain , it is challenging to predict the stimulus detected by RavA . However , RavS contains two intracellular PAS domains . PAS domain was found to sense low oxygen , redox , light or binds to small chemicals [50 , 51] . A previous study also demonstrated that hypoxia sensing by RavS is involved in regulating virulence of X . campestris [37] . In our study , we found that the truncated RavS without the N-terminal transmembrane and PAS domains had a higher level of autokinase activity , indicating that the PAS domains autoinhibit the phosphorylation of RavS . If this hypothesis is true , the signals detected by PAS domains , such as low oxygen and still to-be-identified stimuli , will regulate RavS phosphorylation and play important roles in controlling virulence and motility of X . campestris . In vitro phosphotransfer assays revealed that upon phosphorylation by RavA or RavS , there is no signal representing the observed RavR~P , which suggests that RavR~P is highly unstable ( Fig 5A and S6 Fig ) . This biochemical feature of RavR may have an essential role in regulation because regardless of whether the phosphoryl groups are obtained from RavS or RavA or even if the transfer speed is enhanced by c-di-GMP , RavR can be promptly uncharged and avoid rapid saturation in phosphorylation . The unstable nature of RavR~P make the protein to be uncharged by dephosphorylation very quickly [52 , 53] . Furthermore , considering that dissociation of the interaction between RavS and c-di-GMP either by substitution of Arg656 or by overexpression of ravS led to virulence attenuation , we speculate that RavR~P dephosphorylation has an important role in regulating the lifestyle transition between virulence and swimming motility . Future investigations are needed to dissect the molecular mechanism controlling the half-life of RavR~P in the context of RavA-RavR-RavS regulation .
All bacterial strains and recombinant vectors used in this work are listed in S1 Table . The WT strain of X . campestris pv . campestris ATCC 33913 and recombinant strains were grown at 28°C in rich NYG medium ( 5 g L−1 tryptone , 3 g L−1 yeast extract and 20 g L−1 glycerol , pH 7 . 0 ) . We used 210 medium ( 5 g L−1 sucrose , 8 g L–1 casein enzymatic hydrolysate , 4 g L−1 yeast extract , 3 g L−1 K2HPO4 and 0 . 3 g L−1 MgSO4·7H2O , pH 7 . 0 ) to prepare for electro-component cell production and the TGM medium ( 10 g L−1 tryptone , 5 g L−1 yeast extract , 20 g L−1 glycerol , 10 g L−1 glucose , 0 . 7 g L–1 K2HPO4 , and 0 . 25 g L−1 MgSO4 . 7H2O ) for the EPS production assay . Escherichia coli ( E . coli ) DH5α cells were used as a host for the construction of all recombinant vectors . The E . coli BL21 ( DE3 ) strain was used as a host for expressing recombinant proteins with the pET30a , pET22b or pET28b vectors ( Novagen , USA ) , and the E . coli TB1 strain was used as a host for expressing recombinant proteins with the pMAL-p2X vector ( Novagen ) . Appropriate antibiotics were added to the media when needed at the following concentrations: kanamycin ( 50 μg ml–1 ) , spectinomycin ( 150 μg ml–1 ) and ampicillin ( 100 μg ml–1 ) . Electro-competent cells were prepared by washing bacterial cells thoroughly three times with ice-cold glycerol ( 10% ) . The transformation conditions of bacterial cells were set at 1 . 8 kV cm–1 , 25 μF and 200 Ω and conducted in a Bio-Rad Pulser XCell ( Bio-Rad , USA ) . General molecular biology experiments , including PCR , DNA ligation , enzyme restriction and western blotting , followed the protocols from Molecular Cloning unless otherwise mentioned . All in-frame deletion ( markerless ) mutants were constructed using the suicide vector pK18mobsacB by a homologous , double crossover method . Briefly , the 5′ and 3′ genomic sequences of a targeted region were amplified using the primers listed in S2 Table , and correct PCR products were ligated into pK18mobsacB . Point mutants were constructed from the corresponding deletion mutants using pK18mobsacB with the point mutated genes . The recombinant pK18mobsacB vector was electroporated into X . campestris pv . campestris competent cells to generate single-crossover mutants by selection on NYG plates containing kanamycin . Next , single-crossover mutants were cultured in NYG medium for 1–2 h and then grown on NYG plates containing 10% sucrose to select second-round homologous crossovers . Candidate bacterial mutants were verified by PCR and sequencing . To genetically complement mutants , a full-length gene with a native promoter was amplified using primers listed in S2 Table and ligated into the broad-host vector pHM2 . Then , the pHM2 recombinant vector was used to complement the mutated genes . The recombinant medium-copy , broad-host-range vector pBBR1MCS2 that carries full-length sequences of the genes of interest ( under the control of the Plac promoter ) was used to construct overexpression strains . Briefly , bacterial culture grown to OD600 nm = 0 . 5 and 0 . 1 ml or 1 . 0 ml cell cultures were transferred to into two new tubes , A and B , respectively . The tube B were centrifuged for 10 min at 4°C at 5 , 000 g . Cell pellets in tube B were resuspended in 80 μl ddH2O and 20 μl 5×sample loading buffer and the mixtures were boiled for 15 min before being subjected to12% ( w/v ) SDS-PAGE gels for electrophoresis analysis . Proteins were then electrotransferred onto a Fluoro Trans polyvinylidene fluoride membrane ( PALL , USA ) following the manufacturer's protocol and were detected with immunoblotting by the antiserum against RavS , RavA or RavR and HRP-conjugate goat anti-rabbit IgG secondary antibody ( Abmart , China ) at a dilution of 1:5 , 000 . The protein bands were visualized by Enhanced ECL Chemiluminescent Substrate Kit ( Yeasen , China ) . The densities of protein bands were determined by Image J . The amounts of individual protein bands were calculated from the standard curves derived from a set of purified and quantified RavS , RavA or RavR protein standards run on the same blot . Protein levels were normalised to bacterial cells , which were counted in a serially diluted bacterial culture from tube A that was grown on NYG plates for 48 hours . To estimate the concentration of a protein in a cell , morphology of 30 bacterial cells were selected and their volumes were calculated as cylinders . Each experiment was performed with three biological replicates . Plant inoculation and virulence assays were conducted using six-week-old cabbage cultivar Brassica oleracea cv . Zhonggan 11 as host plants . The WT strain of X . campestris pv . campestris ATCC 33913 and sterile 10 mM MgCl2 were used as positive and negative controls , respectively . All bacterial strains were cultured overnight in NYG medium containing appropriate antibiotics . Cells were collected and washed with 10 mM MgCl2 and the concentrations were adjusted to OD600 = 0 . 1 before inoculation using sterile scissors . After inoculation , the plants were kept in a greenhouse at 25–30°C and relative humidity >80% . The lesion length was measured 10–12 d after inoculation . For each strain , at least 30 cutting sites were made to evaluate the virulence level . An assay of EPS production was conducted according to a previous study with small modifications [37 , 54] . Bacterial strains were cultured at 28°C in NYG medium until the OD600 = 0 . 8 . Then , 200 μL subculture was inoculated into 20 mL TGM medium and cultured at 28°C for 72 h before measurement . To determine the production of EPS , the supernatants of 12 mL bacterial cultures were separated by centrifugation at 25 , 000 g for 30 min . A total of 24 mL ethanol and 1 . 2 mL saturated KCl were added into the supernatants and the mixtures were kept at 4°C for 4 h . The precipitated EPS was pelleted by centrifugation at 25 , 000 g for 30 min and washed twice with 12 mL 95% ethanol . The bacterial cells and EPS were dried at 60°C overnight before determination of their dry weights . EPS production was quantified by the ratio of the weight of EPS vs . the dry weight of bacterial cells . NYG semi-soft agar ( 0 . 15% agar ) motility plates were used to determine the swimming capability of bacterial strains . Semi-soft plates were prepared and allowed to cool and dry at room temperature for 3 h prior to inoculation . Strains were cultured at 28°C overnight and adjusted to OD600 = 0 . 8 . Then , 2 . 5 μL of the mixture was transferred to the motility plates and incubated at 28°C for 28 h . The plates were photographed and the diameter of the migration zone of bacteria was measured . Assays were performed with at least three biological replicates , each containing 10 repeats . Statistical analysis was performed using the two-tailed Student’s t-test . Bacterial strains were cultured at 28°C for 36 h . Negative staining and TEM observations were performed . Briefly , bacterial cells were acquired from NYG plates and gently suspended in 40 μL ddH2O for 15 min . The bacterial turbid liquid was then spotted onto 400 mesh carbon-coated copper grids , which were glow discharged for 15 s immediately prior to use . Samples on the grids were negatively stained twice with 0 . 15% uranyl acetate before blotting with filter paper . Pictures were taken with a JEM-1400 electron microscope ( JEOL , Japan ) at an operating voltage of 80 kV . The number of flagellated cells ( Nf ) and non-flagellar cells ( Nnf ) in the TEM images were counted and the flagellated cell ratio = Nf/ ( Nf + Nnf ) determined . Assays were performed with three biological replicates and the number of counted cells in each assay was larger than 100 . Statistical analysis was performed using the two-tailed Student’s t-test . To measure flagellar length , the tagged image file ( TIF ) format pictures of cells were transformed to the AutoCAD Drawing Database ( DWG ) format and then opened by auto Desk auto CAD 2014 software . Flagella were traced by the PLINE model and the relative lengths were calculated by the PROPERTIES model . The scale segment of each picture was measured in the same way to calculate the flagellar length . Assays were repeated with three biological replicates and the number of measured flagella in each assay was larger than 30 . qRT-PCR was used to measure the level of mRNA . Total RNA was extracted by TRIzol ( Invitrogen , USA ) . The DNA contamination in total RNA samples was eliminated by RNAase-free DNase I ( Ambion , USA ) . The first strand of cDNA was generated using random primers ( Promega , USA ) and Superscript III reverse transcriptase ( Invitrogen ) . qRT-PCR was conducted using Maxima SYBR Green ( Fermentas , USA ) in a DNA Engine Option 2 System ( Bio-Rad ) , according to the manufacturer’s instructions . Amplification of tmRNA was used as an internal control . Generally , a qRT-PCR experiment was repeated independently three times with three technical repeats of each sample . A representative of all the biological repeats was selected and reported . C-terminal , His6-tagged recombinant proteins were produced by corresponding recombinant pET30a , pET22b or pET28b ( Novagen ) vectors that were transformed into the E . coli BL21 ( DE3 ) strain . His6-tagged proteins were extracted and purified using affinity chromatography with Ni-NTA agarose beads ( Novagen ) , according to the manufacturer′s instructions . N-terminal MBP-tagged recombinant proteins were produced by the corresponding recombinant pMAL-P2X ( Novagen ) vector that was transformed into the E . coli TB1 strain . If necessary , the TEV protease cleavage sequence was appended at the N-terminus of the target protein . The MBP fusion protein was purified and the elute buffer was exchanged to the reaction buffer ( 50 mM Tris-HCl , pH 8 . 0 , 50 mM NaCl , 0 . 5 mM EDTA and 1 mM DTT ) , then incubated with TEV protease at a ratio of 1:100 for 8 h at 4°C . Cleaved MBP or uncleaved fusion protein was removed by MBP resin and the TEV protease was removed by Ni-NTA agarose beads . Purified target protein was obtained by size exclusion chromatography with a Superdex 75 10/300 GL column ( GE Healthcare , Piscataway , NJ , USA ) . Purified proteins were concentrated using Centricon YM-10 columns ( Millipore ) and the elute buffer was changed to the storage buffer ( 50 mM Tris-HCl , pH 8 . 0 , 0 . 5 mM EDTA , 50 mM NaCl and 5% glycerol ) for further use or storage at −80°C . Protein preparations were examined for purity by SDS-PAGE and quantified by a Bradford assay ( Bio-Rad ) . Polyclonal antiserums of RavR , RavS and RavA were prepared by immunising rabbits with approximately 3 mg purified , soluble proteins . The polyclonal antibodies of RpfC and HPPK were reported by our previous studies [32 , 55] . The purified EAL protein was concentrated by VivaSpin Turbo , 5K MWCO ( Sartorius , German ) . A total of 15 μM protein and 300 μM c-di-GMP are incubated in phosphodiesterase reaction buffer ( 50 mM Tris-HCl pH 7 . 5 , 250 mM NaCl , 25 mM KCl , 10 mM MgCl2 and 2 mM DTT ) at 4°C for 30 min and subsequently subjected to gel filtration with Fast Protein Liquid Chromatography AKTA Purifier 10 with Frac-900 ( GE Healthcare , USA ) . The ATKA system was pre-equilibrated with phosphodiesterase reaction buffer at a flow rate of 0 . 5 ml/min and then applied to a Superdex 75 10/300 GL column to separate dimer and monomer . The elution profiles were collected at A280 and confirmed by SDS-PAGE and Coomassie brilliant blue staining . An in vitro autophosphorylation assay was conducted as described in our previous study [32 , 55] . Purified protein was incubated with 10 μM ATP containing 10 μCi [γ-32P]ATP ( PerkinElmer , USA ) in the reaction buffer ( 50 mM Tris-HCl , pH 7 . 8 , 25 mM NaCl , 25 mM KCl , 5 mM MgCl2 , 2 mM DTT ) for the indicated time at 28°C . If necessary , c-di-GMP was added to the mixture at the same time as ATP . For the phosphoryl transfer assay , purified RR protein or its derivative was added into the reaction mixture containing the phosphorylated HK . The autophosphorylation or phosphoryl transfer reaction was terminated by adding 5× SDS-PAGE loading buffer . Phosphorylated proteins were separated by 12% SDS-PAGE . After electrophoresis , gels were exposed to a phosphor screen ( GE Healthcare ) and the autoradiographic signals were detected by a Typhoon FLA7000 ( GE Healthcare ) . Phos-tag acrylamide gels were prepared according to the instructions described by the manufacturer with minor modifications . Phos-tag acrylamide running gels contained 8% or 12% ( w/v ) 29:1 acrylamide:N , N”-methylene-bis-acrylamide , 375 mM Tris pH 8 . 8 , 0 . 1% ( w/v ) SDS . All the gels were copolymerized with 50 μM Phos-tag acrylamide and 100 μM MnCl2 . All stacking gels contained 5% ( w/v ) 29:1 acrylamide:N , N”-methylene-bis-acrylamide , 125 mM Tris pH 6 . 8 , 0 . 1% ( w/v ) SDS . All Phostag acrylamide containing gels were run at 4°C for 3 . 5 hours under constant voltage ( 150 V ) . Coomassie brilliant blue staining was used to detect the phosphorylated proteins . The interaction between c-di-GMP nucleotides and various proteins was measured by using MST with a fluorescein-labelled c-di-GMP ( 2′-Fluo-AHC-c-di-GMP , abbreviated as fl-c-di-GMP; 2′-Fluo-AHC-c-di-AMP , fl-c-di-AMP; 2′-Fluo-AHC-cGMP , fl-cGMP; Biolog , Germany ) . These chemicals contain carboxyfluorescein that has excitation and emission wavelengths of 497 nm and 520 nm , respectively , which can be detected directly by a MST instrument . The experiments were performed on a Monolith NT . 115 device using standard treated capillaries ( NanoTemper Technologies , Germany ) . The concentration of the particular protein varied from 0 . 036 to 150 μM with a 2-fold gradient and the concentration of fl-c-di-GMP was constant at 20 nM . Fluorescence intensity due to thermophoresis was recorded using the blue channel optics of the instrument ( λex = 470 ± 15 nm , λem = 520 ± 10 nm ) during a 30 s period of infrared laser heating at 80% of the maximum laser power , followed by a 5 s cooling period . Measurements were performed in buffer containing 20 mM HEPES , pH 7 . 5 , 150 mM NaCl , 2 mM MgCl2 , 2 mM DTT and 0 . 05% Tween 20 . The KD Fit of NanoTemper Analysis software ( ver . 1 . 5 . 41 ) was used for fitting the curve and calculation of the dissociation constant ( Kd ) . Assays were repeated with at least three biological replicates in triplicate . Thermal shift assays were performed as previously described with minor modification[32] . Briefly , purified protein was added to the reaction to a final concentration of 5 μM in the MST buffer [20 mM HEPES , pH 7 . 5 , 150 mM NaCl , 2 mM MgCl2 , 2 mM DTT , 0 . 05% Tween 20 and 1:500 dilution of SYPRO Orange Dye ( Invitrogen , USA ) ] . A melt curve protocol was run on a Bio-Rad qPCR instrument ( Bio-Rad , USA ) . The fluorescence was measured using the ROX reporter with a temperature gradient of 20–95°C in 1 . 0°C increments at 30 second intervals . Melt curve data were trimmed to three data points after maximum and the data were plotted with Boltzmann model to obtain the temperature midpoint of unfolding ( Tmid ) of the protein using Prism 5 . 0 software ( GraphPad ) . Three biological replicates were assayed in triplicate and statistical significance was determined with two-tailed Student’s t-test . [32P]-labelled c-di-GMP was chemically synthesised using a labelled [32P]GTP ( 3000 Ci/mmol , PerkinElmer , USA ) and the purified His6-tagged enzyme tDGC; a protein construct with a key residue mutation ( R158A ) of diguanylate cyclase from Thermotoga maritime [41] . Five micromolar tDGC and 20 μCi [32P]GTP ( mixed with 50 μM cold GTP ) were added to a mixture of 20 μL reaction buffer ( 300 mM NaCl , 50 mM Tris-HCl , pH 7 . 5 , 20 mM MgCl2 , and 2 mM DTT ) . After 3 h at 45°C , the reaction was terminated by heating at 98°C for 10 min . The precipitated protein was removed by centrifugation at 20 , 000 g for 5 min . Radioactive [32P]c-di-GMP was tested by separation on a polyethyleneimine-cellulose plate ( 1:1 . 5 ( v/v ) saturated ( NH4 ) 2SO4 and 1 . 5 M KH2PO4 , pH 3 . 6 ) . The mixture contained more than 95% [32P]c-di-GMP without further purification . c-di-GMP extraction was performed as described previously with modifications [34] . Briefly , 8 mL bacterial culture grown to OD600 = 0 . 8 was centrifuged at 5000 g and 4°C for 10 min . Cell pellets were resuspended in 2× 800 μL NYG medium and 0 . 2 and 1 . 0 mL of the cell resuspension was transferred to two tubes , A and B , respectively . The tubes were centrifuged at 5 , 000 g and 4°C for 10 min . The cell pellet in tube B was resuspended in 300 μL extraction solution ( 40% acetonitrile , 40% methanol and 20% water ) , incubated on ice for 15 min and lysed by a non-contact ultrasonication system ( Bioruptor UCD-200 , Diagenode , Belgium ) for 10 min with an alternating 30 s power on and 30 s power off procedure . Samples were then centrifuged at 20 , 000 g and 4°C for 10 min and the supernatant was transferred into a 2 . 0 mL tube . Extraction was repeated twice with 200 μL extraction solution but heating at 95°C was omitted . The combined supernatant fluids of three extractions were completely evaporated . c-di-GMP powders were resuspended in 100 μL HPLC-grade ddH2O and analysed by liquid chromatography-tandem mass spectrometry ( LC-MS/MS ) on an AB SCIEX QTRAP 4500 system ( AB SCIEX , USA ) . A Synergi Hydro-RP 80A LC column ( 4 μM , 150 × 2 mm , Phenomenex , Torrance , CA , USA ) was used for reversed-phase liquid chromatography . Solvent A was 0 . 1% acetic acid in 10 mM ammonium acetate and solvent B was 0 . 1% formic acid in methanol . The gradient used was as follows: time ( t ) = 0–4 min , 98% solvent A , 2% solvent B; t = 10–15 min , 5% solvent A , 95% solvent B . The injection volume was 5 μL and the flow rate for chromatography was 200 μL/min . The amount of c-di-GMP in samples was calculated with a standard curve generated from pure c-di-GMP ( Sigma-Aldrich , St . Louis , MO , USA ) suspended in ddH2O ( Biolog , Germany ) . c-di-GMP levels were normalised to bacterial cells , which were counted in a serially diluted bacterial culture in tube A that was grown on NYG plates for 48 h . Each c-di-GMP quantification experiment was performed with three biological replicates . The levels of c-di-GMP were compared with those of WT with the two-tailed Student’s t-test . For the DGC activity assay , purified proteins were added to the reaction buffer ( 300 mM NaCl , 50 mM Tris-HCl , pH 7 . 5 , 20 mM MgCl2 , 2 mM DTT ) and 20 μCi [32P]GTP ( mixed with 50 μM cold GTP ) was added . After 1 h at 28°C , the reaction was terminated by the addition of an equal volume of 0 . 5 M EDTA , pH 8 . 0 . For the PDE activity assay , the purified RavR protein was added to the reaction buffer consisting of 50 mM Tris-HCl , pH 7 . 5 , 250 mM NaCl , 25 mM KCl , 10 mM MgCl2 and 2 mM DTT . Reactions were initiated by the addition of ~1 μM of [32P]c-di-GMP substrate . Reactions were incubated at 28°C for 30 min before termination by adding an equal volume of 0 . 5 M EDTA , pH 8 . 0 . For the DGC and PDE assays , reaction products were mixed with an equal volume of running buffer consisting of 1:1 . 5 ( v/v ) saturated ( NH4 ) 2SO4 and 1 . 5 M KH2PO4 , pH 3 . 6 . Two microlitres of the reaction mixture was spotted and dried onto Cellulose PEI TLC plates ( Selecto Scientific , USA ) . Plates were developed in running buffer , air-dried and exposed to a storage phosphor screen ( GE Healthcare ) and then the autoradiographic signals were recorded on a Typhoon FLA7000 ( GE Healthcare ) . The three-dimensional structure of the DHp-CA domain of RavS was generated by homology modelling methods . Homology modelling was carried out using Modeller 9 . 10 software . The known structures of CpxAHDC ( 4BIU ) , DivL ( 4EW8 ) and VraS ( 4GT8 ) were used for multiple template homology modelling . The structure of c-di-GMP was extracted from the c-di-GMP-VCA0042 structure complex ( PDB ID: 2RDE ) . The structure of DHp-CA was treated by adding hydrogen atoms , calculating the charge and combining nonpolar hydrogens . Docking calculations were carried out using Autodock 4 . 0 software . Affinity ( grid ) maps of 40 × 40 × 40 Å grid points were generated using the Autogrid program . After 100 calculations , the lowest energy conformation was chosen for c-di-GMP and protein interaction analysis . There were no animal experiments in this study .
|
c-di-GMP is a multifunctional bacterial second messenger that controls various physiological processes . The nucleotide derivative binds to riboswitches or proteins as effectors during regulation . Here , we found that c-di-GMP physically binds to a histidine kinase , RavS , of a plant pathogenic bacterium . The binding event significantly enhanced the phosphotransferase activity of RavS to phosphorylate a response regulator , RavR . This process tightly modulates the phosphorylation level of RavS , which is important to the lifestyle transition of the bacterium between virulence and swimming motility . Therefore , our results reveal that c-di-GMP controls the bacterial two-component signalling , one of the dominant mechanisms of bacterial cells in adaptation to various environmental stimuli .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"phosphorylation",
"cell",
"motility",
"swimming",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"xanthomonas",
"campestris",
"pathogens",
"biological",
"locomotion",
"mutation",
"plant",
"science",
"plant",
"pathology",
"cellular",
"structures",
"and",
"organelles",
"bacteria",
"cell",
"swimming",
"proteins",
"recombinant",
"proteins",
"plant",
"bacterial",
"pathogens",
"pathogen",
"motility",
"biochemistry",
"point",
"mutation",
"cell",
"biology",
"post-translational",
"modification",
"virulence",
"factors",
"physiology",
"plant",
"pathogens",
"genetics",
"biology",
"and",
"life",
"sciences",
"xanthomonas",
"organisms",
"flagella"
] |
2019
|
Cyclic-di-GMP binds to histidine kinase RavS to control RavS-RavR phosphotransfer and regulates the bacterial lifestyle transition between virulence and swimming
|
Gastric cancer development is strongly correlated with infection by Helicobacter pylori possessing the effector protein CagA . Using a transgenic Drosophila melanogaster model , we show that CagA expression in the simple model epithelium of the larval wing imaginal disc causes dramatic tissue perturbations and apoptosis when CagA-expressing and non-expressing cells are juxtaposed . This cell death phenotype occurs through activation of JNK signaling and is enhanced by loss of the neoplastic tumor suppressors in CagA-expressing cells or loss of the TNF homolog Eiger in wild type neighboring cells . We further explored the effects of CagA-mediated JNK pathway activation on an epithelium in the context of oncogenic Ras activation , using a Drosophila model of metastasis . In this model , CagA expression in epithelial cells enhances the growth and invasion of tumors in a JNK-dependent manner . These data suggest a potential role for CagA-mediated JNK pathway activation in promoting gastric cancer progression .
Infection with Helicobacter pylori is the strongest risk factor for the development of gastric carcinoma , which is the second most common cause of cancer-related death worldwide [1] . Although approximately half the world's population is infected with H . pylori , most of those individuals will develop simple gastritis and remain asymptomatic . However , 10–15% of infected subjects will develop duodenal ulcers and 1% will develop gastric adenocarcinoma [2] . This dramatic variability in clinical outcome of H . pylori infection is not well-understood , but likely results from the consequences of long-term interactions between the bacterium and its human host . Specific bacterial and host genetic factors have been shown to affect H . pylori pathogenesis . Strains that possess the cag pathogenicity island ( cag PAI ) , which encodes a type IV secretion system used to inject the CagA effector protein directly into gastric epithelial cells , are much more virulent [3] . Once inside host cells , CagA is tyrosine phosphorylated on conserved carboxyl terminal EPIYA motifs by Src family kinases . Variability in the number and composition of these phosphorylation motifs also correlates with differences in the carcinogenic potential of H . pylori strains [4] . Host genetic factors that can influence the progression and ultimate disease outcome of H . pylori pathogenesis include polymorphisms that enhance expression of certain cytokines [2] , and genetic changes that occur during progression from normal mucosa to gastric carcinoma such as loss of tumor suppressors and activation of oncogenes [5] . Although development of a complex disease like gastric cancer requires the cooperation of many bacterial and host genetic factors , it is clear that the CagA effector protein is an important driver of disease progression . CagA has been shown to interact with a multitude of host cell proteins belonging to several conserved signaling pathways [6] , and these interactions are thought to promote carcinogenesis upon H . pylori infection . The majority of these interactions were discovered using cell culture models in which CagA expression can disrupt processes such as tight junction formation , motility and cytoskeleton dynamics . However , which interactions between CagA and host cell signaling pathways trigger the processes that lead to gastric cancer remains unclear [4] . Obtaining more specific information about the relative importance of CagA's interactions with host cell proteins will require investigation of their downstream effects on intact epithelial tissue . In order to examine the effects of both bacterial and host genetic factors , our group has developed a system in which Drosophila melanogaster is used to model pathogenesis of the H . pylori virulence factor CagA [7] . There are several properties that make this model organism well-suited for studying the pathogenic effects of CagA expression . First , many canonical cell signaling pathways have been extensively characterized in Drosophila and show high conservation with the homologous pathways in humans . Also , genetic tools like the GAL4/UAS system allow expression of CagA in specific cells within an epithelium and examination of how CagA-expressing cells interact with neighboring wild type cells . Finally , we can easily manipulate host genes using resources generated by the rich Drosophila research community to assess potential effects on CagA-induced phenotypes . In addition , our model allows us to test whether CagA's interactions are phosphorylation-dependent through expression of a mutant form of CagA known as CagAEPISA , in which the EPIYA phosphorylation motifs have been deleted or mutated [8] . Use of this model has already provided insight into CagA's role in manipulating receptor tyrosine kinases , the Rho signaling pathway and epithelial junctions [7]–[10] . Epithelial polarity is one important feature of host cells known to be perturbed by CagA . Strains of H . pylori that encode CagA are exclusively able to cause localized disruption of apicobasal polarity in order to colonize a polarized monolayer of tissue culture cells [11] . CagA-positive strains of H . pylori have also been shown to cause apoptosis in both cultured gastric cancer cells and human gastric biopsies [12] , [13] , although the role of CagA-dependent apoptosis in H . pylori pathogenesis remains controversial . Loss of epithelial cell polarity has been shown to induce apoptotic cell death or promote tumorigenesis in different cellular and genetic contexts [14] . Cell death resulting from polarity disruption can trigger compensatory proliferation in order to replace lost cells , but this process can become tumorigenic in the presence of genetic alterations that block apoptosis [15] . This mechanism has been proposed to explain how the ability of CagA to disrupt cell polarity and induce apoptosis may be linked to its tumorigenic potential , but the host cell signaling pathways that could mediate these downstream effects have not been identified [16] . An important host signaling pathway that triggers apoptosis downstream of cell polarity disruption is the c-Jun NH2-terminal kinase ( JNK ) pathway . JNK is a stress-activated protein kinase with numerous upstream activators including cytokines , mitogens , osmotic stress , ultraviolet radiation and loss of cell polarity [17] . JNK-mediated apoptosis plays a role in several physiological processes including morphogenetic apoptosis and classical cell competition in which slow-growing cells are eliminated by their wild type neighbors . The JNK pathway also triggers apoptosis in response to a unique form of cell competition known as intrinsic tumor suppression where JNK activation performs a cell editing function by removing aberrant cells that arise within an epithelium , thus enhancing the resilience of epithelia to insult . Both expression of the tumor necrosis factor ( TNF ) homolog Eiger ( Egr ) and the presence of wild type cells within an epithelium are required for JNK pathway activation downstream of cell polarity disruption , and their absence can lead to tumor formation [18] . Furthermore , JNK signaling has been shown to switch from a proapoptotic to a progrowth role in the presence of oncogenic Ras [19] . These functions of the JNK pathway are well-established in Drosophila , and likely also relevant in mammals given the high conservation of this pathway throughout evolution [20] . Bacterial activation of JNK signaling has also demonstrated importance in enhancing epithelial robustness . During oral infection of Drosophila with the human pathogen Pseudomonas aeruginosa , the bacterium activates JNK signaling in the intestinal epithelium to trigger apoptosis and subsequent compensatory proliferation , thereby stimulating epithelial renewal . The same effect was not seen during infection with an avirulent strain of P . aeruginosa that does not secrete the virulence factor pyocyanin , suggesting a role for this effector protein in activating JNK signaling in response to damage induced by the bacterium [21] . Similar to the adult Drosophila intestine , the larval imaginal disc epithelia are particularly resistant to the effects of stress-induced apoptosis and can recover after losing over 50% of their cells during development to produce normal adult structures [22] . This inherent epithelial resilience makes the imaginal discs a relevant tissue in which to examine potential effects of JNK-dependent apoptosis mediated by a bacterial virulence factor . In this study , we discovered a role for the CagA virulence factor in activating JNK signaling . We used transgenic Drosophila to express CagA in the developing wing imaginal disc , a simple polarized epithelial structure formed during larval stages of development . We found that CagA expression caused a distinct pattern of cell death in which apoptotic cells are basally extruded from the epithelium . In addition we showed that this apoptosis phenotype is enhanced by coexpression with Basket ( Bsk ) , the Drosophila homolog of JNK , and suppressed by coexpression with a dominant-negative form of Bsk . From these results , we conclude that expression of CagA triggers JNK pathway activation which causes apoptosis in an intact epithelium . Furthermore , we used a Drosophila model of metastasis to show that CagA expression can enhance the growth and invasion of tumors generated by expression of activated Ras . This increase in tumorigenic capacity is suppressed by coexpression with dominant-negative Bsk , leading us to conclude that CagA promotes tumor growth and invasion through JNK pathway activation .
In order to examine the effects of expressing the H . pylori effector protein CagA on an intact epithelium , we used the GAL4/UAS system to drive its expression in the wing imaginal disc . The Drosophila wing begins to form during early larval life when it exists as a primordial sac which contains both a simple columnar epithelium and the squamous epithelium of the peripodial membrane [23] . Cells within the wing imaginal disc proliferate extensively in larval stages followed by disc evagination during pupation , resulting in the adult wing structure . This developmental process is distinct from that of the eye imaginal disc used to model CagA pathogenesis previously [7]–[10] , which undergoes systematic differentiation during larval stages . In addition , the fate of imaginal disc cells is specified early in development [24] which allowed us to express CagA in distinct regions of the wing disc ( Figure 1A ) . We expressed CagA with various GAL4 drivers specific to the wing ( Figure S1A ) , and determined that both the level of CagA protein and the region in which it is expressed affect the resulting larval and adult wing phenotypes ( Figure S1B–S1L ) . We focused our subsequent analysis on two different GAL4 drivers which express CagA either in a subset of wing cells or throughout the wing imaginal disc: beadex-GAL4 ( bx-GAL4 ) is expressed specifically in cells of the columnar epithelium that give rise to the dorsal surface of the wing blade ( Figure 1B ) , and 765-GAL4 is expressed ubiquitously throughout the wing . A membrane-localized GFP construct ( mGFP ) was used to visualize the expression domain . Expressing CagA with the 765-GAL4 ubiquitous wing driver did not cause any observable phenotype ( Figure 1C ) . However , expressing CagA with the bx-GAL4 dorsal wing driver caused clusters of apoptotic cells to form near the center of the expression domain in wing imaginal discs from third instar larvae ( Figure 1D ) . This phenotype was dose-dependent , since expressing two copies of CagA increased both the size and number of apoptotic clusters formed ( Figure 1E ) . A similar phenotype has been shown to result from localized JNK pathway activation in the wing imaginal disc epithelium but does not occur upon more ubiquitous activation [25] . Interestingly , although expressing one copy of CagAEPISA with the bx-GAL4 driver did not cause a phenotype ( Figure 1F ) , expressing two copies induced formation of small apoptotic clusters within the expression domain ( Figure 1G ) . This reduction in apoptosis induction suggests that the phenomenon does not require phosphorylated CagA , but that CagAEPISA is a less potent activator of cell death . This observation is consistent with data obtained from transgenic expression of CagAEPISA in the eye imaginal disc epithelium , where less severe phenotypes were shown to result from differential cellular localization of the phosphorylation-resistant form of CagA . Whereas wild type CagA was highly enriched at the apical membrane in eye imaginal disc epithelial cells , CagAEPISA was expressed diffusely throughout the cytoplasm . We propose that the inability of phosphorylation-resistant CagA to localize apically within an epithelium influences its interactions with host cell proteins and their resulting effects on the epithelial tissue [9] . Cells within the apoptotic clusters generated by CagA expression were extruded from the basal surface of the wing imaginal disc epithelium . Further examination of this tissue revealed an enrichment of matrix metalloproteinases , which break down basement membrane , specifically in cells located directly apical to the apoptotic clusters ( Figure 1H ) . This observation indicates that apoptotic cells generated by CagA expression are actively removed from the wing epithelium and not passively lost during development of the imaginal disc . Many complex cellular interactions are required during wing disc development to ensure proper formation of the adult wing structure ( Figure 1I ) . While this process did not appear to be disrupted by ubiquitous expression of CagA in the wing ( Figure 1J ) , CagA expression specifically in the dorsal wing caused a dose-dependent disruption of the imaginal disc epithelium ( Figure S1M and S1N ) which affected the overall appearance of the adult wing ( Figure 1K and 1L ) . This phenomenon also did not require phosphorylated CagA since expression of CagAEPISA caused a less severe dose-dependent disruption of the adult wing ( Figure 1M and 1N ) . The observation that ubiquitous expression of CagA in the wing does not cause apoptosis or epithelial disruption suggests that wild type cells surrounding those which express CagA are required to produce both phenotypes . This is consistent with the previous observation that JNK-dependent apoptosis is only triggered when aberrant cells within an epithelium are surrounded by wild type cells [26] . Taken together , these data prompted us to examine a potential role for JNK signaling in the apoptosis and epithelial disruption phenotypes resulting from localized expression of CagA in the wing imaginal disc . Several aspects of the apoptosis phenotype caused by CagA expression in the wing imaginal disc suggested an interaction between CagA and the JNK pathway . In order to determine the nature of this potential interaction , we examined the effects of expressing several forms of Bsk , the Drosophila homolog of JNK , on the CagA-induced wing phenotype . Ectopic overexpression of wild type Bsk with the bx-GAL4 dorsal wing driver generated small apoptotic clusters ( Figure 2A ) , indicating that the presence of excess JNK in the wing can phenocopy CagA expression . Furthermore , the cell death phenotype caused by CagA expression in the wing was dramatically enhanced by coexpression with wild type Bsk ( Figure 2B ) . Coexpression of Bsk with CagAEPISA also caused a substantial amount of apoptosis in the wing imaginal disc , suggesting that this interaction is not dependent on phosphorylated CagA ( Figure 2C ) . As expected , expression of a dominant-negative form of Bsk ( BskDN ) alone did not cause apoptosis in the wing imaginal disc ( Figure 2D ) . Significantly , coexpression of BskDN with CagA almost completely suppressed the apoptosis phenotype caused by CagA expression ( Figure 2E ) , indicating that blocking JNK signaling suppresses CagA-dependent cell death in the wing . These data suggest that CagA expression triggers wing imaginal disc apoptosis through JNK pathway activation . We also examined the effects of JNK pathway modulation on the epithelial disruption phenotype caused by CagA expression . Although ectopic overexpression of wild type Bsk with bx-GAL4 caused only a minor adult wing phenotype in the form of extra vein material ( Figure 2F ) , coexpression of Bsk with CagA dramatically enhanced the epithelial disruption phenotype ( Figure 2G ) . Ectopic overexpression of Bsk with CagAEPISA was not sufficient to induce epithelial disruption ( Figure 2H ) . Expression of BskDN also gave rise to only subtle vein defects in an otherwise normal adult wing ( Figure 2I ) . Interestingly , BskDN expression was not able to rescue but instead enhanced the epithelial disruption caused by CagA expression ( Figure 2J ) . One explanation for this apparent contradiction is that blocking JNK signaling prevents the induction of apoptosis that is required to remove aberrant CagA-expressing cells from within the epithelium , which are then allowed to accumulate and lead to a more severe disruption of the adult structure . We tested this hypothesis using the apoptosis inhibitor p35 , a baculovirus-derived suicide substrate for effector caspases . Overexpressing p35 alone with bx-GAL4 did not produce a phenotype ( Figure S2A and S2C ) , while coexpressing p35 with CagA effectively blocked apoptosis but enhanced disruption of the adult wing epithelium ( Figure S2B and S2D ) . This observation is consistent with the inhibition of apoptosis resulting in more severe CagA-dependent adult phenotypes . Enhancement and suppression of CagA-induced apoptosis in the wing imaginal disc was quantified using a method we developed to measure the percentage of the expression domain that is caspase-positive . These quantitative data showed that both the enhancement of CagA-induced apoptosis seen with coexpression of ectopic Bsk , and its suppression upon expression of BskDN were statistically significant ( Figure 2K ) . In order to further examine the genetic interaction between CagA and JNK signaling , we used a lacZ reporter allele of puckered ( puc ) , the main component of a negative feedback loop in the JNK pathway . This construct has been used extensively as a readout for JNK pathway activation in Drosophila tissue using antibody staining for β-galactosidase ( β-gal ) . Expressing CagA in combination with puc-lacZ in the dorsal wing imaginal disc demonstrated that cells adjacent to those undergoing apoptosis are activating JNK signaling ( Figure 2L ) . Upregulation of puc-lacZ correlated with phosphorylation of JNK , verifying that specific activation of JNK signaling results from CagA expression ( Figure S2E ) . These data provide additional evidence that CagA expression activates JNK signaling in the wing imaginal disc epithelium . JNK signaling is activated by a complex set of signals including TNF and loss of epithelial polarity ( Figure 2M ) . To examine the mechanism through which CagA activates JNK signaling , we used the bx-GAL4 driver to express CagA in combination with RNAi-mediated knockdown of known epithelial polarity determinants and examined wing imaginal discs for enhancement of the apoptosis phenotype ( Figure 3A ) . We tested a panel of polarity proteins , many of which caused apoptosis when knocked down in the absence of CagA expression ( Table S1 ) . We chose to target a protein from each of the previously described complexes whose localization and function establish epithelial cell polarity [27] , and to simplify our analysis we selected polarity proteins that did not cause an apoptosis phenotype when knocked down on their own ( Figure S3A–S3C ) . When tested in combination with CagA expression , we found that RNAi-mediated knockdown of neither the junctional protein Bazooka ( Baz ) , nor the apical protein Crumbs ( Crb ) enhanced apoptosis ( Figure S3D and S3E ) . In addition , knockdown of Par1 , which has been shown to interact with CagA in tissue culture cells [28] , did not enhance the apoptosis phenotype caused by CagA expression in this context ( Figure S3F ) . Interestingly , RNAi-mediated knockdown of the basolateral protein Discs Large ( Dlg ) did not cause a significant phenotype ( Figure 3B ) but markedly enhanced the apoptosis caused by CagA expression ( Figure 3C ) . The same effect was seen with knockdown of Lethal Giant Larvae ( Lgl ) , another basolateral protein ( Figure S3G and S3H ) . The genes encoding these polarity proteins are known as neoplastic tumor suppressor genes ( nTSGs ) because their loss causes tumor formation in Drosophila [29] , and generating clones of cells which lack this specific class of polarity determinants has been shown to trigger JNK-dependent apoptosis in imaginal discs [30] . Our data suggest that nTSGs normally suppress CagA-mediated JNK pathway activation and subsequent apoptosis in the wing imaginal disc . Disruption of the nTSGs activates JNK signaling through endocytosis of the TNF homolog Egr [31] . Homozygous egr mutant animals are viable and , as expected , no apoptosis was observed in their wing imaginal discs ( Figure S3I ) . Conversely , ectopic overexpression of wild type Egr in the dorsal wing imaginal disc caused a severe apoptosis phenotype ( Figure S3J ) , consistent with previous data showing Egr to be a potent activator of cell death in Drosophila epithelia [32] . We made the unexpected observation that expression of CagA in the dorsal wing disc of an egr mutant animal enhanced the apoptosis phenotype ( Figure 3D ) . Interestingly , RNAi-mediated knockdown of Egr alone in the dorsal wing with bx-GAL4 did not cause a phenotype ( Figure S3K ) or enhance apoptosis when coexpressed with CagA ( Figure S3L ) . This observation suggests that loss of Egr in wild type cells surrounding the CagA expression domain is responsible for the enhanced apoptosis phenotype seen in the wing imaginal discs of egr mutant animals expressing CagA . Recent data has demonstrated that loss of nTSGs in clones of imaginal disc cells causes Egr-dependent activation of nonapoptotic JNK signaling in their wild type neighbors . JNK activation in surrounding wild type cells leads to induction of a phagocytic pathway which triggers engulfment of polarity-deficient cells within the clone [30] . A similar mechanism can be invoked to explain the enhancement of CagA-induced apoptosis seen in egr mutant wing imaginal discs . Loss of Egr in the wild type cells surrounding the expression domain may prevent engulfment of CagA-expressing cells . This would increase the number of aberrant cells available to undergo apoptosis upon CagA-mediated activation of JNK signaling via another parallel upstream pathway . We hypothesize that multiple cellular consequences of CagA expression can activate JNK signaling combinatorially . Supporting this view , we demonstrated that CagA-induced apoptosis was enhanced by ectopic overexpression with a wild type form of the small GTPase Rho1 ( Figure 3E ) , another upstream activator of the JNK pathway that did not cause a phenotype when overexpressed alone ( Figure S3M ) , and which our group has shown is activated by CagA [9] . Enhancement of CagA-induced apoptosis in the wing imaginal disc was quantified using the previously described method . These data showed significant enhancement of apoptosis with coexpression of CagA and knockdown of nTSGs , ubiquitous loss of Egr or overexpression of Rho1 . Knockdown of several other polarity proteins or Egr in CagA-expressing cells did not enhance the apoptosis phenotype ( Figure 3F ) . Overexpression of Rho1 , ubiquitous or localized loss of Egr and knockdown of the other polarity proteins alone did not induce significant apoptosis in the wing imaginal disc ( Figure S3N ) . These observations suggest that specific polarity protein complexes within the cell , as well as other upstream activators are responsible for transducing the signals that lead to JNK pathway activation upon CagA expression in the wing imaginal disc ( Figure 3G ) . The finding that CagA activates the JNK pathway is intriguing in light of recent evidence indicating that activation of JNK signaling can switch from proapoptotic to progrowth in the presence of oncogenic Ras [19] . In order to examine a potential role for CagA-mediated JNK pathway activation in promoting tumorigenesis , we used a slight variation of a previously established Drosophila metastasis model to create whole eye clones expressing an activated form of the Ras oncogene ( RasV12 ) in epithelial cells of the eye imaginal disc using the eyeless ( ey ) driver with the FLP/FRT system to generate primary tumors [33] . We then evaluated the size of GFP-marked tumors in whole larvae ( Figure 4A ) and dissected cephalic complexes ( Figure 4B ) in order to determine whether coexpression of CagA could enhance the growth and invasive potential of these tumor cells through activation of the JNK signaling pathway . Expression of RasV12 alone in whole eye clones caused overgrowth of eye imaginal disc cells which resulted in tumor formation ( Figure 4C ) , as previously described [34] . Although generating whole eye clones expressing either GFP alone ( Figure S4A ) or with CagA ( Figure S4B ) was not tumorigenic , coexpression of CagA enhanced the growth of tumors generated by RasV12 expression ( Figure 4D ) . Whole eye clones expressing CagAEPISA were also not tumorigenic ( Figure S4C ) , and when combined with RasV12 expression caused only a minor enhancement of tumor growth ( Figure 4E ) . As expected , coexpression of BskDN did not affect the growth of tumors generated by RasV12 expression alone ( Figure 4F ) . However , BskDN expression caused a severe reduction in the growth of tumors expressing both RasV12 and CagA ( Figure 4G ) . Quantification of these data was accomplished by determining the size of dissected cephalic complexes of each genotype and showed a significant growth enhancement with combined expression of RasV12 and CagA , which was suppressed by coexpression of BskDN ( Figure 4H ) . These data demonstrate that expression of CagA can enhance the growth of tumors generated by expression of RasV12 in a JNK-dependent manner . Generating whole eye clones that express RasV12 alone most commonly caused either a mildly invasive phenotype characterized by the migration of a small number of GFP-positive cells along one edge of the ventral nerve cord ( VNC ) , or a noninvasive phenotype in which cells within the optic lobe approached but did not migrate into the VNC ( Figure 5A ) . Whole eye clones expressing either GFP alone ( Figure S5A ) or with CagA ( Figure S5B ) were not invasive , but coexpression of CagA with RasV12 resulted in a much larger number of GFP-positive tumor cells migrating from both optic lobes into the VNC ( Figure 5B ) . These cells were not terminally differentiated , as indicated by a lack of staining with the neuron-specific ElaV antibody , and phalloidin staining showed a morphology distinct from other cells in the VNC ( Figure S5D ) . Expressing CagAEPISA in whole eye clones also did not produce an invasive phenotype ( Figure S5C ) , and coexpression of CagAEPISA with RasV12 caused a less pronounced enhancement of the mild invasion caused by expression of RasV12 alone ( Figure 5C ) , suggesting that the phosphorylation-resistant form of CagA is less effective at promoting tumor progression . Coexpression of BskDN did not affect the invasive phenotype generated by RasV12 expression alone ( Figure 5D ) , but BskDN expression caused a dramatic reduction in the invasive capacity of tumors expressing both RasV12 and CagA ( Figure 5E ) . These data show that CagA expression can enhance the invasion of RasV12-expressing tumor cells through JNK activation . In order to determine the significance of CagA's enhancement of invasion , we used a previously described method [35] to categorize invasive phenotypes into four distinct classes which represent a progression from non-invasive to severe invasion of the VNC ( Figure 5F ) . Quantitation of the percentage of cephalic complexes exhibiting each class of VNC invasion showed a significant difference between expression of RasV12 alone and in combination with CagA , which was suppressed by coexpression of BskDN ( Figure 5G ) .
Infection of tissue culture cells with H . pylori has been shown to activate JNK signaling , but a role for CagA in this process remains controversial [36]–[38] . Additionally , these experiments were performed in nonpolar AGS cells , so if polarity disruption plays a role in JNK pathway activation downstream of CagA , as our data suggest , these cell culture models may not reveal this interaction . JNK pathway activation has also been shown to result from infection with several pathogenic bacteria in epithelial cell culture models of infection [39] . Interestingly , the enteroinvasive bacterium Shigella flexneri was shown to activate JNK and upregulate TNFα expression in both infected and adjacent uninfected epithelial cells in culture [40] , similar to our data showing that JNK-mediated tissue responses to CagA expression involve a cell-nonautonomous requirement for TNF/Egr . The distribution of H . pylori during infection of the gastric epithelium is known to be heterogeneous [2] . We therefore hypothesize that interactions between cells containing CagA protein and uninfected neighboring cells could also be important for pathogenesis of H . pylori . Our data suggest that CagA is an important mediator of JNK pathway activation during H . pylori infection , and identify several host proteins involved in this process . We observe genetic interaction between CagA and nTSGs , but not junctional proteins involved in polarity . This is consistent with recent data from tissue culture cells which demonstrated that CagA-positive strains of H . pylori specifically disrupt apicobasal polarity in a polarized monolayer prior to affecting the integrity of cellular junctions [11] . Disruption of nTSGs has been shown to cause JNK-dependent apoptosis , and more recent data indicates that elimination of polarity-deficient cells is dependent on their location within the wing imaginal disc due to varying levels of dMyc throughout the tissue [41] . The extent of aberrant cell removal differs significantly with respect to established gradients of Wnt/Wingless , dMyc and Hippo-Salvador-Warts pathway activation that ensure proper development of the wing [42] . We propose that the extent of variation observed upon CagA expression in the wing with different GAL4 drivers is due to spatial variation in these host cell signaling pathways . Our data also suggest that CagA can activate JNK-dependent apoptosis through multiple upstream pathways . The observation that overexpression of Rho1 enhances CagA-dependent apoptosis in the wing imaginal disc epithelium is consistent with previous data from our group demonstrating a role for CagA in activating the Rho pathway to disrupt epithelial patterning [9] . Use of the unique genetic tools available in Drosophila has provided important insight into potential interactions between CagA-expressing cells and neighboring wild type cells . Our observation that loss of TNF/Egr in wild type cells surrounding those expressing CagA can enhance apoptosis , presumably by reducing engulfment of CagA-expressing cells , indicates that the genetic state of uninfected cells may also play a role in H . pylori pathogenesis . This finding is important with respect to the established function of TNF/Egr-dependent JNK activation in cell competition induced by intrinsic tumor suppression . Our data suggest that the presence of CagA protein induces changes in signaling and morphology which cause an epithelial cell to be outcompeted by its wild type neighbors through a local mechanism that requires TNF/Egr in the neighboring epithelial cells . Interestingly , Drosophila immune cells known as hemocytes have also demonstrated the ability to remove polarity-deficient cells from an epithelium through a more global extrinsic tumor suppression mechanism that is TNF/Egr-dependent [43] . Although we have not explored a role for hemocytes in removal of CagA-expressing wing epithelial cells , it is possible that a related mechanism may occur during H . pylori infection of the human stomach through immune surveillance mediated by TNF . Although this specific cytokine is an important component of the initial immune response to infection with a pathogen , TNF is also known to promote tumor progression specifically in the context of chronic inflammation or in the presence of activated Ras [43] , [44] . We hypothesize that TNF functions to suppress tumor initiation resulting from the presence of CagA protein in gastric epithelial cells through several mechanisms , but that the inflammatory environment created by prolonged infection with H . pylori and the emergence of oncogenic mutations over time cause TNF to promote progression of gastric cancer . Since it was first discovered , JNK has been demonstrated to have both pro-tumorigenic and tumor suppressor functions in different cell types and organs . Studies in Drosophila have helped shed light on the genetic contexts in which JNK activation functions to promote tumor progression , namely in the presence of oncogenic Ras [45] . Recently , JNK was shown to be required for activated KRas-induced lung tumor formation in mice [46] , suggesting a conserved function of JNK activation in cooperating with activated Ras to promote tumorigenesis in mammals . A potential role for JNK pathway activation has also been explored in mammalian gastric cancer . Activation of JNK signaling has been detected in human gastric cancer samples , and mice lacking JNK1 exhibit a decrease in gastric apoptosis and an attenuation of gastric tumor development induced by the chemical carcinogen N-methyl-N-nitrosourea [47] . A role for H . pylori in the context of mammalian gastric cancers induced by cooperation between JNK and Ras signaling has not been explored . Our finding that CagA expression can induce JNK-dependent apoptosis in a polarized epithelium is interesting with respect to data suggesting that JNK signaling has evolved as a cell editing mechanism to remove aberrant cells from within an epithelium [18] . Activation of JNK signaling could represent a host response aimed at removing cells containing CagA protein from the gastric epithelium . Similarly , P . aeruginosa-mediated activation of JNK signaling in the intestinal epithelium of Drosophila can trigger epithelial renewal as a host defense mechanism . However , this process can become pathogenic and lead to dramatic overproliferation of intestinal cells in animals harboring oncogenic Ras mutations [21] . In H . pylori infection , which can persist for many years before the development of gastric cancer , JNK-mediated apoptosis could be an effective mechanism to limit pathogenic effects on the gastric epithelium . However , this process of tissue editing can also increase cell turnover , contributing to accumulation of genetic mutations in host cells . Our data show that acquisition of an oncogenic mutation in host epithelial cells experiencing CagA-mediated JNK pathway activation can promote tumor progression , suggesting that this potential host defense strategy can become tumorigenic in certain genetic contexts ( Figure 6B ) . Transgenic expression of CagA was recently found to cause neoplastic transformation in a mouse model , providing evidence for CagA's role as a bacterial oncoprotein in mammals [48] . The low incidence and delayed development of gastrointestinal tumors in these mice was attributed to lower expression of CagA in the surviving animals , as higher expression was assumed to be lethal during embryogenesis . Additionally , secondary mutations were identified in the tumors , but their potential cooperation with host cell signaling pathways activated by CagA expression was not addressed [48] . Infection with CagA-positive H . pylori is also known to induce an invasive phenotype in tissue culture cells [49] , but potential effects of the oncogenic mutations present in these immortalized cell lines is unknown . Although we did not demonstrate the sufficiency of CagA to induce tumor phenotypes in our Drosophila model , our data support a crucial role for CagA in promoting tumor progression in combination with oncogene activation . We believe that using an inducible expression system in Drosophila allowed us to bypass the toxicity observed upon CagA expression in mice and cell culture models , thus revealing novel interactions between CagA and host cell proteins with downstream effects on apoptosis and tumorigenesis . Although half the world's population is thought to be infected with H . pylori , a small percentage of those individuals will develop gastric cancer [2] . This observation indicates that , in addition to the presence of the cag PAI in more virulent strains , host genetics must also play a crucial role in determining the outcome of H . pylori infection . Our results suggest that a change in host genetics during long-term association with H . pylori could cause JNK activation to switch from conferring a protective function against CagA-induced cellular changes to enabling tumor progression . Data collected from tissue biopsies indicate that Ras mutation may play a role in the development of gastric cancer in human patients [50] , and our data put forward the idea that enhanced tumorigenic potential created by cooperation between JNK pathway activation via the bacterial genetic factor CagA and sporadic activation of Ras in host cells could drive gastric cancer formation in a subset of H . pylori infections .
The following fly stocks were used: UAS-CagA , UAS-CagAEPISA [7]; bx-GAL4 , sd-GAL4 , ap-GAL4 , en-GAL4 , ptc-GAL4 , hs-FLP , Act>y+>GAL4 , UAS-GFP . S65T , UAS-mCD8::GFP ( mGFP ) , UAS-bsk . B ( Bsk1 ) , UAS-bsk . A-Y ( Bsk2 ) , UAS-bsk . K53R ( BskDN ) , egrMB06803 ( egr−/− ) , UAS-Dlg-RNAi , UASp-FLAG . Rho1 ( Rho1 ) , UAS-Ras85D . V12 ( RasV12 ) , UAS-Crb-RNAi , UAS-Patj-RNAi , UAS-Cora-RNAi , UAS-Cdc42-RNAi ( from Bloomington Stock Center ) ; 765-GAL4 ( provided by Ross Cagan , Mount Sinai School of Medicine ) ; pucE69 ( puc-lacZ ) , Regg1GS9830 ( UAS-Egr ) ( provided by Michael Galko , MD Anderson Cancer Center ) ; UAS-Lgl-RNAi , UAS-Baz-RNAi , UAS-Par1-RNAi , UAS-Scrib-RNAi , UAS-Par6-RNAi , UAS-aPKC-RNAi , UAS-Mir-RNAi ( provided by Chris Doe , University of Oregon ) ; UAS-Egr-RNAi ( from Vienna Drosophila Resource Center ) ; ey-FLP; Act>y+>GAL4 , UAS-GFP ( provided by Tory Herman , University of Oregon ) . Flies were raised at 25°C using standard methods . Whole eye clones were generated as previously described [33] without the GAL80 repressor to express transgenes in all cells that give rise to the eye-antennal disc . FLP-out clones were generated by subjecting each 4–6 hour collection of embryos to one hour of heat-shock at 37°C , then dissecting wing discs approximately 96–120 hours later . Larval tissues were fixed and stained using standard protocols . The following primary antibodies were used: rabbit anti-active caspase-3 ( 1∶200; BD Pharmingen ) , mouse anti-Mmp1 ( 1∶50; Developmental Studies Hybridoma Bank ) , mouse anti-β-galatosidase ( 1∶500; Sigma ) rat anti-ElaV ( 1∶10; Developmental Studies Hybridoma Bank ) , rabbit anti-β-galatosidase ( 1∶200; MP Biomedicals ) and mouse anti-phospho-SAPK/JNK ( 1∶100; Cell Signaling Technology ) . Both Cy3 and Cy5-conjugated secondary antibodies were used ( 1∶200; Jackson ImmunoResearch ) , as well as Alexa Fluor 546 and Alexa Fluor 633 phalloidin ( 1∶40; Molecular Probes ) . Intact adult wings were mounted in a 1∶1 mixture of lactic acid and ethanol . Adult wings , intact larvae and whole cephalic complexes were visualized using light microscopy or GFP fluorescence on a Zeiss dissecting microscope . Wing imaginal discs , ventral nerve cords and cephalic complexes were visualized on a Nikon confocal microscope . Images were processed using Adobe Photoshop , where levels were adjusted to optimize the signal-to-noise ratio in each color channel while maintaining similar levels of background noise and desired signal between channels and images . Adult wing images were removed from their background using the Extract filter in Adobe Photoshop . XZ confocal planes were created using the Reslice function in Image J . Projections of confocal cross sections were created using the Merge to HDR command in Adobe Photoshop . Apoptosis was quantified by selecting the single confocal cross section of each wing imaginal disc exhibiting the highest level of active caspase-3 staining and manually tracing the expression domain , then determining the percentage of this domain showing active caspase-3 staining using the Threshold function in Image J . Cephalic complex size was quantified using the Threshold function in Image J to determine the area of the tissue in µm2 . Graphs were created with GraphPad Prism software , which was also used to calculate two-tailed p values using the unpaired t test with Welch's correction for apoptosis quantitation . The statistical significance of differences in metastatic potential for each genotype was calculated using Excel to determine two-tailed p values using the unpaired t test .
|
The gastric pathogen Helicobacter pylori infects an estimated 50% of the world's population and is a major risk factor for the development of gastric cancer . Strains of H . pylori that can inject the CagA effector protein into host cells are known to be more virulent , but the potential contributions of host genetics to pathogenesis are not well-understood . Using transgenic Drosophila melanogaster , we show that the genetic context of both the host cells in which CagA is expressed and their neighboring cells changes CagA's effects on epithelial tissue . When CagA is expressed in a subset of cells within an epithelium , it disrupts tissue integrity and induces apoptosis through activation of JNK signaling , a pathway that functions to remove aberrant cells from an epithelium . CagA's proapoptotic effects are inhibited by neoplastic tumor suppressor genes in CagA-expressing cells , and by the tumor necrosis factor homolog Eiger in neighboring cells . In contrast , when CagA is coexpressed with oncogenic Ras in a Drosophila model of metastasis , it enhances the growth and invasion of tumors in a JNK-dependent manner . Our study demonstrates how changes in host genetics can cooperate with activation of JNK signaling by the bacterial virulence factor CagA to promote tumorigenesis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"death",
"cellular",
"stress",
"responses",
"cancer",
"genetics",
"microbiology",
"host-pathogen",
"interaction",
"signaling",
"in",
"selected",
"disciplines",
"epithelial",
"cells",
"animal",
"models",
"c-jun",
"n-terminal",
"kinase",
"signaling",
"cascade",
"drosophila",
"melanogaster",
"model",
"organisms",
"apoptotic",
"signaling",
"bacterial",
"pathogens",
"signaling",
"in",
"cellular",
"processes",
"oncogenic",
"signaling",
"microbial",
"pathogens",
"ras",
"signaling",
"biology",
"pathogenesis",
"signal",
"transduction",
"gram",
"negative",
"cell",
"biology",
"genetics",
"cellular",
"types",
"molecular",
"cell",
"biology",
"genetics",
"of",
"disease",
"genetics",
"and",
"genomics",
"signaling",
"cascades"
] |
2012
|
Transgenic Expression of the Helicobacter pylori Virulence Factor CagA Promotes Apoptosis or Tumorigenesis through JNK Activation in Drosophila
|
Vibrio cholerae infections cluster in households . This study's objective was to quantify the relative contribution of direct , within-household exposure ( for example , via contamination of household food , water , or surfaces ) to endemic cholera transmission . Quantifying the relative contribution of direct exposure is important for planning effective prevention and control measures . Symptom histories and multiple blood and fecal specimens were prospectively collected from household members of hospital-ascertained cholera cases in Bangladesh from 2001–2006 . We estimated the probabilities of cholera transmission through 1 ) direct exposure within the household and 2 ) contact with community-based sources of infection . The natural history of cholera infection and covariate effects on transmission were considered . Significant direct transmission ( p-value<0 . 0001 ) occurred among 1414 members of 364 households . Fecal shedding of O1 El Tor Ogawa was associated with a 4 . 9% ( 95% confidence interval: 0 . 9%–22 . 8% ) risk of infection among household contacts through direct exposure during an 11-day infectious period ( mean length ) . The estimated 11-day risk of O1 El Tor Ogawa infection through exposure to community-based sources was 2 . 5% ( 0 . 8%–8 . 0% ) . The corresponding estimated risks for O1 El Tor Inaba and O139 infection were 3 . 7% ( 0 . 7%–16 . 6% ) and 8 . 2% ( 2 . 1%–27 . 1% ) through direct exposure , and 3 . 4% ( 1 . 7%–6 . 7% ) and 2 . 0% ( 0 . 5%–7 . 3% ) through community-based exposure . Children under 5 years-old were at elevated risk of infection . Limitations of the study may have led to an underestimation of the true risk of cholera infection . For instance , available covariate data may have incompletely characterized levels of pre-existing immunity to cholera infection . Transmission via direct exposure occurring outside of the household was not considered . Direct exposure contributes substantially to endemic transmission of symptomatic cholera in an urban setting . We provide the first estimate of the transmissibility of endemic cholera within prospectively-followed members of households . The role of direct transmission must be considered when planning cholera control activities .
Cholera disproportionately affects less-developed areas of Asia , Africa , and Latin America , leading to an estimated 3–5 million cases and 100–130 thousand deaths per year [1] . Vibrio cholerae O1/O139 transmission is associated with two general modes of exposure to infection . Community-to-person transmission results from ingestion of contaminated water from environmental sources [2] . Direct transmission results from exposure to food , water , and surfaces shared by a cluster of individuals , such as a household , and contaminated by an infectious member of the cluster [1] . The relative contributions of community-to-person and direct exposure to endemic cholera transmission are subject to ongoing debate [2]–[7] . Community-to-person exposure is a well-established mode of transmission for cholera infection [8]–[11] , whereas the relative contribution of direct exposure is poorly quantified . Evidence of herd immunity from cholera vaccine studies [12]–[14] , reports from cholera outbreak investigations [15]–[17] , the clustering of V . cholerae clones by household [18] , and mathematical modeling of epidemics [3]–[5] suggest that direct transmission is an important component of overall transmission . John Snow himself believed that cholera transmission had a direct as well as a waterborne component . After describing many case studies of potential cholera transmission in England during the mid 1800's , Snow stated: “Besides the facts above mentioned , which prove that cholera is communicated from person to person , …” [19] . The effectiveness of interventions for the control of endemic cholera depends , in part , upon the prevailing mode of transmission . Dominant transmission through direct exposure is expected to generate clusters of cases within close contact groups , such as households . Transmission primarily via community-to-person exposure is expected to generate a more diffuse spatial pattern of cases within a population . Knowledge of the dominant mode of transmission is important for informing the selection of optimal intervention strategies , such as vaccination strategies for preventing ongoing cholera transmission in Haiti [20] . Here , we estimate the probabilities of endemic V . cholerae O1/O139 transmission through 1 ) exposure to community sources of infection and 2 ) direct exposure within the household . The modifying role of covariates and aspects of the natural history of V . cholerae infection are also described . To our knowledge , this study provides the first estimates of the transmission potential of endemic cholera via direct exposure among the members of prospectively-followed households .
The Ethical and Research Review Committees of the International Centre for Diarrhoeal Disease Research in Dhaka , Bangladesh ( icddr , b ) and the Institutional Review Board of the Massachusetts General Hospital reviewed and approved this study . Written informed consent was collected from all participants , with consent provided by a parent or legal guardian for participants under 18 years-old . This analysis was conducted using de-identified data . A case-ascertained [21] study was conducted from January 2001 to May 2006 among households of Dhaka , Bangladesh , as previously described [18] , [22]–[24] . Individuals older than six months of age presenting to the hospital of the icddr , b with severe acute watery diarrhea , stool positive for V . cholerae infection by culture , and without a history of significant co-morbidities were selected for inclusion as index cases . Antimicrobial therapy was provided to all index cases , as per the standard clinical practice for the management of acute watery diarrhea at the icddr , b . The study timeline is expressed relative to the enrollment day of the index case ( day 1 ) . Written informed consent was requested from the members of the index case's household on day 2 of the study . Households are defined as individuals who ate from the same cooking pot during the preceding three days . For children under 18 years of age , informed consent was requested from a parent or legal guardian . Consenting household members were enrolled if they were not participating in other icddr , b studies . Figure 1 describes the data collection schedule for each household . Upon enrollment , stool samples were obtained from the index cases . Study staff visited each household on days 2 through 7 , 14 , and 21 . On day 2 , information about age and sex was collected for all enrolled household members , along with an eight-day clinical history for preceding symptoms of diarrheal disease , i . e . , for days −7 to 1 ( there was no day 0 ) . Similar seven-day clinical histories were collected on days 7 , 14 , and 21 . Rectal swabs were collected from all study participants on days 2 through 7 , 14 , and 21 . Blood specimens were collected from all study participants on days 2 , 4 , 7 , and 21 . Using previously-described methods [22] , stool samples and rectal swabs were tested for V . cholerae , and blood specimens were assayed for vibriocidal antibody titers and ABO blood type . The serogroup ( O1 or O139 ) and serotype of serogroup O1 El Tor biotype ( Ogawa or Inaba ) were determined for positive stool specimens .
The study enrolled 1491 individuals living in 399 households ( one index case each ) . We excluded 24 households with only one enrolled member ( the index case ) from the analysis , because they impart no information about transmission . An additional 11 households ( 53 members ) were excluded from the analysis for lack of serum vibriocidal antibody titer measurements . The resulting study population of 1414 people lived in 364 households . Table S1 in Text S1 provides the frequency of the households included in the analysis by the number of enrolled members , infections among the enrolled , and the serogroup-serotype of infection . The 53 enrolled study members who were excluded from the analysis because of missing vibriocidal antibody titer data did not substantially differ from the study population in the distributions of age ( p<0 . 580 ) , sex ( p<0 . 981 ) , the prevalence of O blood group ( p<0 . 320 ) , the attack rate for watery diarrhea ( p<0 . 508 ) , or the proportion of members with at least one stool/rectal swab specimen positive for V . cholerae ( p<0 . 098 ) ( Table 1 ) . Non-primary cholera infections were significantly younger than index cholera infections ( p<0 . 001 ) , but not uninfected household contacts ( p<0 . 200 ) ( Table 1 ) . The sex ratio and the prevalence of O blood group were similar for infected and uninfected household contacts ( p<0 . 896 and p<0 . 145 , respectively ) . When compared to household contacts , index cases were more commonly female ( p<0 . 027 ) and of the O blood group ( p<0 . 004 ) . Among the 1050 household contacts , 22 . 5% ( 318 ) developed V . cholerae infection ( Table 1 ) . The proportion of non-index V . cholerae infections with watery diarrhea was 56 . 7% ( 180 of 318 ) . The number of days between the symptom onset dates of the primary ( onset of symptoms on or before onset in the household index infection ) and non-primary symptomatic cholera cases in a household was not uniformly distributed ( Figure 3 ) ; the mean ( standard deviation ) duration was 9 ( 4 ) days . For each serogroup-serotype , we rejected the null hypothesis of no direct transmission ( p-value<0 . 001 ) . The SAR estimates were 4 . 9% ( 95% CI: 0 . 9%–22 . 8% ) for , 3 . 7% ( 95% CI: 0 . 7%–16 . 6% ) for , and 8 . 2% ( 95% CI: 2 . 1%–27 . 1% ) for . The CPI estimates for a comparable 11-day period were 2 . 5% ( 0 . 8%–8 . 0% ) for , 3 . 4% ( 1 . 7%–6 . 7% ) for , and 2 . 0% ( 0 . 5%–7 . 3% ) for . Serogroup-serotype specific estimates for the risk of infection resulting from a month ( 30 days ) of community-to-person exposure were 6 . 7% ( 2 . 1%–20 . 4% ) for , 9 . 0% ( 4 . 5%–17 . 3% ) for , and 5 . 3% ( 1 . 4%–18 . 7% ) for . For all serogroup-serotypes combined , the estimated risk of cholera infection associated with community-to-person exposure varied by calendar month of the year ( Figure S2 ) . In the univariate and multivariate transmission models , children under five years of age were significantly more susceptible than adults 18 years and older to cholera infection ( Table 2 ) . The susceptibility of children 5–17 years-old to cholera infection did not significantly differ from that of adults 18 years and older . There was no evidence of significant differences in the susceptibility of contacts based upon either sex or ABO blood group . Serum vibriocidal antibody titers demonstrated some protection against infection by O1 El Tor Ogawa and O139 . This protection was only statistically significant for O1 El Tor Ogawa in the multivariate model , providing an estimated 9% ( 95% CI: 1%–17% ) reduction in susceptibility to infection per two-fold higher initial titer for serum vibriocidal antibody . The magnitude of protection against O139 infection was larger than the effect for O1 El Tor Ogawa , but lacked statistical significance in every model . The unadjusted transmission model , both including ( -and- ) and excluding ( -only ) transmission through direct exposure , demonstrated adequate fit to the study data ( Figure S3 in Text S1 ) . For both unadjusted transmission models , chi-squared tests rejected the null hypothesis for a difference between the observed and expected final size distributions ( p-value< = 0 . 001 ) . There was little evidence to suggest that the quality of the fit of the unadjusted transmission model including direct transmission differed from that of the model excluding this mode of exposure .
This study and analysis have several limitations that are expected to result in underestimation of the true serogroup-serotype specific household secondary attack rates for V . cholerae infection . Initial serum vibriocidal antibody titers may incompletely account for levels of pre-existing immunity to V . cholerae infection [30] . In addition , we only considered direct transmission within the households of study participants . Some of the transmission attributed to community-to-person exposure may have actually resulted from direct exposure outside of the participant's own household . Theoretically , the observed pattern of cholera infections could have been solely attributable to community-to-person transmission . The inability to discern between transmission of V . cholerae through exposure to direct versus short-duration community-based sources of infection has been demonstrated through mathematical modeling investigations using mass action transmission models [3] , [4] . If the only source of exposure to cholera infections in one of our study households was a common community-based source of infection that occurred on a single day , a pattern of illness onset dates consistent with acute direct transmission could occur . If this scenario occurred in every household enrolled in this study , our statistical model would errantly attribute some or all transmission to direct exposure . If a community-based source of infection exposed the members of a household for a period longer than one day ( a plausible scenario ) , our statistical model would be able to differentiate between transmission due to that source and that resulting from direct exposure within the household . The fact that the mean time between the onset of symptoms in primary and associated non-primary symptomatic cholera cases was longer than the maximum length of the incubation period supports the assertion that direct and/or multi-day community-based sources of exposure to infection were operating in this population during this study . The current study categorizes the cholera infections by bacterial phenotype , i . e . , biotype , serogroup , and serotype . The primary limitation of using this categorization schema relates to the substantial genetic variation evident among vibrios of the same bacterial phenotype . Since the probability of direct transmission of cholera within households is likely to be strongly associated with genetic similarity of the infecting vibrios , this analysis would certainly have been enhanced by incorporation of measures of the genetic similarity between the cholera bacteria isolated from the participants in the same household . If this type of information had been available for the current study , our analysis could have either 1 ) estimated genotype specific SAR's ( analogous to the approach used here to estimate SAR's by bacterial phenotype ) or 2 ) directly incorporated measures of genetic distance into the likelihood for the transmission model . Whole genome sequencing has been used characterize genetic variability among clinical V . cholerae isolates , for example , to characterize the origins of the recent outbreak of cholera in Haiti [34] . Ongoing work by the authors [35] seeks to address this limitation through a combination of ongoing field studies and statistical methodologic research . Cholera remains an important public health issue for low-resource settings with limited public health facilities . Recent experiences in Zimbabwe and Haiti underscore the urgent need for effective intervention strategies [1] , [36] . Our results demonstrate that exposure through direct contact significantly contributes to the endemic transmission of cholera infection . Greater emphasis needs to be placed on implementing interventions targeting transmission through direct exposure . Consideration should be given to evaluating the utility of pre-emptive administration of antimicrobial agents to the household contacts of patients with cholera in areas lacking adequate sanitation . Vaccination of entire households prior to the onset of seasonal transmission may allow for additional control of transmission by bolstering existing immunity among members , thereby reducing the level of direct transmission and protecting against community-to-person transmission from other contaminated sources .
|
Since John Snow's ground-breaking investigations of the devastating outbreaks in 19th-century London , cholera has been considered the quintessential waterborne human infection , transmitting via fecal contamination of environmental water sources . Recently , renewed interest has been paid to the potential importance of transmission through direct exposure within close-contact groups , such as , via fecal contamination of surfaces , food , or drinking water within households . Significant direct transmission of cholera within close contact groups would represent a new target for innovative prevention and control strategies . We estimated the probability of transmission 1 ) via direct contact within 364 urban households located in an endemic cholera setting ( Dhaka , Bangladesh ) and 2 ) via exposure to sources located outside of these households . In this setting we estimated a 4 to 8 percent probability of becoming infected with cholera via direct exposure within households in this setting versus a 2 to 3 percent likelihood of infection due to exposure to external sources over a comparable time period . Our results demonstrate that direct ( within-household ) transmission is a significant component of endemic cholera transmission , suggesting that biomedical and behavioral-modification interventions specifically targeting this mode of transmission could substantially reduce the cholera burden in this type of setting .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"tropical",
"diseases",
"microbiology",
"health",
"care",
"multivariate",
"analysis",
"bacterial",
"diseases",
"mathematics",
"statistics",
"(mathematics)",
"global",
"health",
"neglected",
"tropical",
"diseases",
"population",
"biology",
"statistical",
"inference",
"research",
"and",
"analysis",
"methods",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"cholera",
"disease",
"ecology",
"epidemiology",
"mathematical",
"and",
"statistical",
"techniques",
"hygiene",
"ecology",
"sanitization",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"statistical",
"methods",
"organisms",
"maximum",
"likelihood",
"estimation"
] |
2014
|
Household Transmission of Vibrio cholerae in Bangladesh
|
Mutations in the BREVIPEDICELLUS ( BP ) gene of Arabidopsis thaliana condition a pleiotropic phenotype featuring defects in internode elongation , the homeotic conversion of internode to node tissue , and downward pointing flowers and pedicels . We have characterized five mutant alleles of BP , generated by EMS , fast neutrons , x-rays , and aberrant T–DNA insertion events . Curiously , all of these mutagens resulted in large deletions that range from 140 kbp to over 900 kbp just south of the centromere of chromosome 4 . The breakpoints of these mutants were identified by employing inverse PCR and DNA sequencing . The south breakpoints of all alleles cluster in BAC T12G13 , while the north breakpoint locations are scattered . With the exception of a microhomology at the bp-5 breakpoint , there is no homology in the junction regions , suggesting that double-stranded breaks are repaired via non-homologous end joining . Southwestern blotting demonstrated the presence of nuclear matrix binding sites in the south breakpoint cluster ( SBC ) , which is A/T rich and possesses a variety of repeat sequences . In situ hybridization on pachytene chromosome spreads complemented the molecular analyses and revealed heretofore unrecognized structural variation between the Columbia and Landsberg erecta genomes . Data mining was employed to localize other large deletions around the HY4 locus to the SBC region and to show that chromatin modifications in the region shift from a heterochromatic to euchromatic profile . Comparisons between the BP/HY4 regions of A . lyrata and A . thaliana revealed that several chromosome rearrangement events have occurred during the evolution of these two genomes . Collectively , the features of the region are strikingly similar to the features of characterized metazoan chromosome fragile sites , some of which are associated with karyotype evolution .
Genome integrity depends upon the coordination of replicon and centriole duplication , chromatin condensation , and the assembly and action of the spindle apparatus . Several checkpoints regulate the progression of the chromosomal and cytoskeletal events [1] , and repair systems are recruited as needed to correct replication errors and lesions caused by intrinsic and extrinsic mutagens . Intrinsic mutations may result from the interaction of DNA with reactive metabolites ( e . g . hydroxyl radicals ) and through the activation of mobile genetic elements . Forward genetics proceeds by employing mutagens , which can range from simple chemical mutagens such as ethyl methanesulfonate ( EMS ) , that typically induce base substitutions , to insertional mutagens such as viral and T-DNA integration , to ionizing radiation that is often associated with rearrangements and/or deletions . Characterization of naturally occurring and induced mutants has dramatically accelerated our understanding of all cellular processes and has led to the discovery of small regulatory RNA molecules and epigenetic modifications of chromatin , two areas of intense investigation in contemporary biology . The eukaryotic nucleus is organized hierarchically . The basic level of chromatin organization centers on the nucleosome and through various levels of organization and compaction , nucleosome strings are organized into large loop domains that are anchored to the nuclear matrix [2] . Superimposed on this , specific chromosome boundaries or territories exist within the nucleus , further defining the association of specific inter- and intrachromosomal domains [3] . Double strand breaks ( DSB ) , created in the context of recombination activities , or due to mutagen exposure , must be repaired to ensure chromosome integrity . The juxtapositioning of specific chromosomes likely underpins the recurrent nature of specific rearrangements , for example the reciprocal translocation between human chromosomes 9 and 22 associated with chronic myelogenous leukemia . The evolution of chromosomes has been intensively studied in many systems and is being revolutionized by high throughput sequencing technologies . In plants , genome duplication followed by translocations , inversions , centromere shifts , and the activity of endogenous mobile elements are deemed to be responsible for dispersed blocks of synteny that are observed between distantly related species [4] , [5] . The breakpoints for some of these rearrangements have now been mapped onto reference genomes , and this has facilitated a high-resolution comparison between Arabidopsis thaliana and A . lyrata [6] , which are believed to have evolved from a common ancestor . A series of chromosome fusion and other rearrangement events , coupled with DNA loss , has reduced the chromosome number from eight to five and the amount of DNA by approximately 40% . Within the Brassicaceae , it is clear that 24 conserved chromosomal blocks have been rearranged during evolution to constitute the genomes of mustard family members [7] . At present , it is unknown whether structures at the boundaries of the blocks promote recombination/rearrangement events or if repressive structures exist elsewhere to maintain the syntenic blocks . We previously reported that the brevipedicellus phenotype of Arabidopsis is due to loss-of function of the KNAT1 homeodomain protein-encoding gene and that unusually large deletions occur with a high frequency [8] , [9] . Here we report the characterization of the breakpoint junctions of five bp deletion alleles and the conservation of their features with metazoan chromosome fragile sites , some of which are associated with chromosome breakage events that occurred during the evolution of eukaryotic karyotypes .
Our previous work documented that five different brevipedicellus alleles were not simple alterations in the gene sequence , but rather were due to large deletions of at least 150 kbp [8] . Information on all reported bp alleles can be found in Table S1 . The original bp-1 mutant , isolated by Koornneef and coworkers [10] , was generated by EMS mutagenesis , an alkylating agent that typically induces G to A transition mutations . We identified the bp-2 and bp-3 mutants from a fast neutron mutagenized population; bp-5 is the result of an aberrant T-DNA insertion , and bp-11 is an x-ray induced mutant . All five of these alleles exhibit large deletions . The other characterized bp mutants appear to be simple base changes induced by EMS ( bp-4 , 6–8 , and 10 ) , or are insertional mutants ( bp-9 ) . Lastly , Venglat et al . [11] reported the isolation of a bp-2 mutant from a promoter tagged population; this mutant was later characterized as a point mutation . Delineation of the boundaries of these deletion alleles would permit an analysis of breakpoint regions and perhaps provide clues as to why the region appears to be prone to such extreme segmental deletion events . We therefore employed a six-phase strategy to determine the breakpoints of the five deletions , followed by in-silico analyses to search for motifs that might be important determinants of either generating the deletion or limiting the extent of the lesion . First , for each allele , at least one breakpoint was generally localized by employing PCR , using sets of primers that span a region of approximately 1 . 2 Mbp . The absence of a PCR amplification product was interpreted to mean that the region was part of the deletion ( Figure 1 ) . Next , DNA gel blotting was employed to find a restriction fragment length polymorphism ( RFLP ) between the mutant and parental ( either Columbia or Ler ) DNA . In phase 3 , mutant DNA was cleaved with the enzyme identified by RFLP analysis and the DNA was ligated under conditions designed to produce intramolecular events , generating circular products . This DNA was used as a template for inverse PCR ( iPCR ) to amplify sequences adjacent to known DNA , including the repaired breakpoints . DNA sequencing was then employed to determine the breakpoint junction sequences . Based on this information , new primers were designed to amplify mutant DNA across the breakpoints to give rise to products of predicted size and sequence and thus validate the deletion ( phase 5 analysis ) . Lastly , computer algorithms were employed to analyze the breakpoint junctions , to search for common features/motifs . Table 1 summarizes the five bp deletion alleles . Sequencing of iPCR products revealed that both bp-1 and bp-2 north junction sequences are composed of highly repeated sequences that are found predominantly in centromeric regions on all five chromosomes . For bp-1 , the breakpoint occurs within a tandem ( AG ) n motif that is broadly distributed and may be a member of the LIMPET transposon family [12] . The north junction sequence could not be unequivocally localized for the following reasons . First , we found that a substantial number of primer sets , spanning BACs F5K24 , F10A2 , T3E15 , and F28D6 , did not give rise to PCR products with Landsberg erecta DNA templates ( and therefore were PCR negative for all bp alleles in an Ler background; see Tables S2 , S3 , S4 for primer sequences , locations and amplicon status ) . Other primer sets generated AFLPs wherein the PCR product size differed between Ler and Col , but which can be useful tools for map-based cloning ( e . g . primer sets 76/77 and T3E15-2 in Figure 1 ) . Second , sequencing of iPCR products revealed that the junction sequence adjacent to the south bp-1 breakpoint is most similar to sequences found on other chromosomes , particularly chromosomes 1 and 3 . Intriguingly , the percent match for all of these sequences is only 80–91% . The best matches on chromosome 4 include a sequence near 1 Mb , which could be interpreted as an inversion involving the centromere . Additionally there is homology to a region centered at about 4 . 3 Mb that has a polarity opposite that expected from sequencing the iPCR product . Subsequent pachytene chromosome in situ hybridization discounted the first possibility as cytological landmarks and the distances between fluorescent probe signals are normally distributed in bp-1 ( see below ) . Although a complex rearrangement may be involved , it is more likely that sequence divergence between the Columbia and Landsberg erecta genomes is responsible for this disparity . Based on the locations of the primer sets , bp-1 has suffered a deletion of at least 400 . 6 kbp ( south breakpoint to primer set F1K3 ) and possibly as much as 900 kbp ( south breakpoint to homologous region at 4 . 3 Mb ) . The bp-2 south breakpoint is located 13 . 7 kbp south of the BP gene , within the third intron of At4g01870 , encoding a putative IP3 kinase . The sequence north of this breakpoint exhibits significant homology to a centromeric satellite sequence that is highly repeated on all chromosomes . The best match on chromosome 4 is to BAC F13J5 ( 94% ) , followed by the adjacent BAC , F15N16 ( 93% ) . It is noteworthy that the bp-2 allele , also in the Ler background , is PCR positive for three primer sets within BACs T3E15 , F28D6 and F14G16 , all of which are south of F15N16 according to the AGI reference sequence . The simplest explanation of these data is that either Columbia or Ler has suffered an inversion that positions the F13J5 and F15N16 sequences closer to the BP locus . Additionally or alternatively , sequence divergence between the centromeric satellites in Col vs . Ler might account for the less than perfect sequence homology we found . As is the case for bp-1 , the repetitive nature of the bp-2 flanking sequence prohibits a phase 5 PCR analysis to amplify across the breakpoints and thus the extent of the bp-2 deletion cannot be unequivocally known . However , with the Columbia BAC tiling path as a reference , the deletion in bp-2 is at least 851 kbp based on the localized south breakpoint at AGI coordinate 5164664 , and the first PCR positive fragment on the north end ( in BAC T3E15 centering at 4 . 3 Mbp ) . The locations of both junctions for bp-3 , bp-5 and bp-11 were unequivocally determined . bp-3 , another fast neutron generated allele , suffered a precise deletion of 386 , 634 bp , with a single guanosine residue inserted between the two breakpoints . bp-5 , which presumably arose due to an aberrant T-DNA integration event , suffered a 140 , 996 bp deletion , but possesses a 19 bp ‘filler’ sequence ( 5′ TCCATGTAGTAAGGTAATT3′ ) at the junction that is 90% identical to a sequence on chromosome 3 . The phase 5 PCR product sequence validates the deletion boundaries and the foreign insertion sequence , but its origin is unknown . Lastly , bp-11 is an x-ray induced allele in which the precise excision of 925 , 158 bp occurred . To provide complementary data on the breakpoint locations determined by our molecular analyses , and to investigate the extent of the deletions in bp-1 and bp-2 , we coupled cytological analyses with pachytene chromosome fluorescence in situ hybridization ( FISH ) , using five probesets that span chromosome 4 from 0 . 7 Mbp to 6 . 4 Mbp ( Figure 2 ) . In Columbia , an inversion event involving pericentric heterochromatin has resulted in a heterochromatic knob in this ecotype , which has no counterpart in Ler [13] , [14] . Cytological examination of DAPI stained chromosomes enabled us to measure distances between the NOR4 ribosomal gene cluster at the north end and other cytological landmarks: the heterochromatic knob ( hk4S , in Columbia backgrounds ) , and CEN4 ( Table 2 ) . In addition , the size of hk4S and CEN4 could be evaluated . These analyses revealed that the distance from NOR4 to CEN4 and the size of CEN4 are approximately equal in both parental lines and in all bp alleles . In the Columbia based alleles , the size of hk4S is also similar to the wildtype parent line , as was expected . The probesets for FISH were chosen to evaluate distances between distinct cytological landmarks and the probe target sequences . The two rhodamine-associated probesets RED1 and RED2 bracket hk4S in Columbia ( Figure 2a ) and measurements of NOR4 to RED1 revealed no significant differences in either of the two parental ecotypes , or in the bp alleles . As expected , due to the pericentric inversion in Columbia that generated hk4S , the RED1 to RED2 interval is smaller in Columbia than in Ler , and the distances for the bp alleles are similar to their parent ecotypes . We conclude that the bp-associated deletions do not involve chromatin north of the centromere . In this regard , the possibility that the bp-1 north junction sequence , which exhibits homology to the 1 Mbp region suggestive of an inversion involving the centromere , can be ruled out in favor of an event occurring south of the centromere in a region that has diverged between Columbia and Ler . To correlate the size of the deletions with cytological measurements , three additional probesets were employed . GREEN 1 sequences are located very near CEN4 at 4 . 3 Mbp , while two additional GREEN probes bind in the 6–6 . 4 Mbp region , south of the BP locus at 5 . 15 Mbp ( Figure 2 ) . We expected that the extent of the deletions , as gauged by molecular analyses and sequence comparisons , could be roughly correlated with changes in the distance between GREEN1 and GREEN2 ( G1/G2 ) . Indeed , this is the general trend we observed as bp-5 ( 141 Kbp ) and bp-11 ( 925 Kbp ) deletions gave intersignal distances of 2 . 1 µm and 1 . 7 µm , respectively , compared to the wildtype Columbia distance of 2 . 7 µm . For the Ler-based alleles bp-1 and bp-2 , for which the north breakpoints could not be established , we observed similar reductions in signal distances compared to the Ler parental background with bp-2 exhibiting a shorter G1/G2 length , and thus a larger deletion , than bp-1 . To our surprise , we found that the G1/G2 distance in Columbia is markedly different from Ler , being on average about 2 . 7 µm in Columbia , but over 5 µm for Ler . These measurements were reproducible and statistically significant , implying that a major polymorphism exists between these two ecotypes . Because the RED2 to GREEN1 distances do not vary significantly between Col and Ler and because GREEN1 localizes very close to the DAPI stained centromere in both ecotypes , it seems likely that the G1/G2 polymorphism is due to an indel occurring between them , or possibly an inversion that moved the G2/G3 probeset sequences in a manner similar to the event that created hk4S . The latter possibility might be resolved by using different colored G2/G3 probes in future experiments . Simple BLASTn searches coupled with gene annotations available through the TAIR database permit the identification of genetic elements at breakpoint regions . As the bp deletions are found in the pericentromeric region , there are many occurrences of transposons , pseudogenes , and repetitive sequences . Some of these may be expressed , based on annotated cDNAs and/or expressed sequence tags that map to these sequences , but due to their repetitive nature , it is not clear if they are the actively expressed copies . Discounting transposon/repeat-associated sequences , in the region of 4 . 2 Mbp to 5 . 25 Mbp , there are 31 annotation units for which cDNA/EST annotations exist , and another 16 predicted genes for which there are no cDNA/EST sequences . The largest deletion , bp-11 , which spans over 925 kbp , has lost 27 genes , while bp-5 , the shortest deletion allele spanning 141 kbp , is missing 10 genes ( see Table S5 ) . Within the set of 31 genes , there are two pseudogenes , nine that encode unknown proteins , and three that encode proteins with conserved domains of unknown function . Genes with more complete annotations are mostly members of gene families , though a few are single copy genes . Under normal growth conditions , the Ler based alleles are indistinguishable from one another , and the Columbia based alleles are also similar to one another . It must be appreciated that the bp phenotype is enhanced in the Ler background due to the absence of the ERECTA protein kinase [9] . Our initial work with bp-2 demonstrated that the bp mutant phenotype can be rescued by transformation with a wildtype gene [8]; thus under normal growth conditions , the deleted genes seem to be dispensable . A suite of bioinformatics algorithms ( BLAST , MEME , RepeatMasker ) was used to interrogate the breakpoint sequences to discover commonalities that might inform our understanding of how the lesions were generated/repaired . Our strategy was to analyze north and south donor sequences . A north donor consists of 1 kbp of DNA north of the north breakpoint linked to 1 kbp of adjacent DNA that was deleted . Similarly , a south donor consists of 1 kbp of DNA that was deleted adjacent to 1 kbp of DNA south of the south breakpoints . As the breakpoint regions lie in the pericentromeric chromatin , most of the breakpoint donor sequences were found to possess one or more known repeats and/or transposon remnants ( Figure 3 ) . There is no common sequence motif shared by all alleles , but several short repeats of 20–50 nucleotides ( motifs 1–5 , see also Figure S1 for alignments and their p-values ) are conserved in three to four alleles . The north flanking sequences are predominated by satellite ( bp-2 ) , Athila/gypsy transposons ( bp-3 , bp-5 , bp-11 ) or LIMPET elements ( bp-1 ) , most of which are abundant and dispersed throughout the pericentromeric region . The south flanking regions , which cluster within 80 kbp of each other in BAC T12G13 , tend to be sparse in repetitive sequences , but are A/T rich and 80% possess motif 5 , a T-rich element that is also found in two of the north flanking regions . Statistically , the p-values for motifs 1–5 range from 2 . 8×10−22 to 5×10−6 , lending credibility to their association with lesion formation and/or repair . Sequences at the breakpoints offer little evidence of homologous recombination . Although both the north and south flanking regions of bp-1 contain sequences with homology to LIMPET elements , the junctions are not a continuous sequence in the repaired product , which argues against homologous recombination . A simplistic explanation for this lesion is that it results from illegitimate recombination between ( AG ) n rich repeat sequences located at approximately 4 . 2 and 5 . 2 Mbp , but the significance of the flanking LIMPET sequences cannot be evaluated . Additionally , the close proximity of a ( CT ) n rich sequence juxtaposes two complementary sequences that could generate a long hairpin structure with the breakpoint occurring in a loop of four nucleotides at its apex . Complex secondary structures are predicted at or near all of the donor sequences , with the exception of bp-1 ( see Figure S2 ) . The apparent lack of homology between the paired north and south flanking sequences indicates that the DNA damage is likely repaired by non-homologous end joining ( NHEJ ) . In only one instance ( bp-5 ) is there microhomology between the 5′ and 3′ donor sequence junctions , which might portend the involvement of microhomology-mediated recombination . Matrix attachment region ( MAR ) prediction algorithms suggested the possibility that some of the sequences near the breakpoint junctions contain binding sites for the nuclear matrix . There are several known components of the nuclear matrix , including AHL1 , an AT hook domain containing protein that has been shown by biochemical and cytological assays to be associated with the nuclear matrix [15] . We cloned an AHL1 cDNA behind an inducible promoter and expressed it as a His-tagged protein in E . coli . Southwestern blot analysis was then undertaken with end-labeled probes , including a positive control ( a plastocyanin gene , PC , [16] ) , a negative control ( a histone H1 gene fragment , At1g06760 ) and several fragments near the breakpoint junctions ( 2S , 3N , 3S , 5S , 5N , and 11S ) . Figure 4 shows that the histone H1 probe does not bind to the AHL1 protein . The positive control PC1 probe as well as the 2S , 3N and 5N probes bound weakly , but above background levels , while the 3S , 5S and 11S probes exhibited strong binding to AHL1 . We conclude that these probes , representing regions predicted to contain MAR binding sites , do indeed bind to a known matrix associated protein . Intriguingly , the 3S , 5S and 11S probes are located within 26 kbp of each other at the south end of BAC T12G13 , which harbors all of the south end deletion breakpoints . It is conceivable that BAC T12G13 contains sequences that organize chromatin loops and that in the bp deletion mutants , resection of the initial lesions is limited by either a complex chromatin structure ( e . g . boundary element possessing extensive secondary structure ) and/or an attachment point on the nuclear matrix . The lack of an ordered array of large clones of Ler DNA precludes both direct Col/Ler synteny comparisons as well as the construction of genome array chips for chromatin immunoprecipitation analysis . Nevertheless , as two of our bp alleles are derived from Columbia , we suspected that data mining might prove fruitful for correlating the breakpoint regions with established chromatin features and genetic data . We reasoned that segments of the genome that are active in recombination might have features that could predispose them for rearrangement events . High-resolution recombination mapping along chromosome 4 revealed very little recombination in the hk4S and CEN4 regions , as expected , but also identified regions which exhibit high levels of recombination ( hotspots , [17] ) . Figure 5 shows that the 5–6 Mbp region is very active in recombination , with one of the hotspots in the same region where the south breakpoints of all bp alleles cluster . Interestingly , this region also includes the HY4 locus , where multiple large deletion alleles have been reported [18] . A . thaliana and A . lyrata likely evolved from a common ancestor and several chromosomal fusion and rearrangement events have occurred to reduce both the size of the genome and the number of linkage groups in A . thaliana [6] , [7] . Comparisons of the genomic sequences of the two species reveal that several regions of the ancestral chromosomes six and seven were fused to generate A . thaliana chromosome 4 [6] , [7 and references therein] . Importantly , some of the major rearrangement events also map to the region where the south breakpoints are clustered and in other areas , for example in the 16–17 Mbp region where other rearrangement events have occurred , recombination hotspots also exist ( Figure 5B ) . Comparative sequence analysis of the BP/HY4 region of the A . thaliana and A . lyrata genomes indicate that one or more segments of these genomes underwent transposition/inversion events , providing additional evidence that the region is recombinogenic and prone to chromosome rearrangement events that are possibly associated with speciation ( Figure 5C ) . Epigenomic data mining revealed that the defined north end breakpoints ( bp-3 , bp-5 and bp-11 ) possess common histone modifications , specifically H3K27me1 , H3K9me2 and H4K20me1 , and in addition exhibit 5–15% 5-methylcytosine ( Figure 6 , see also Figure S3 ) . This combination of chromatin modifications is associated with transposable element rich regions found in pericentric heterochromatin [19] . The south end breakpoints , with the exception of bp-1 , differ markedly from the north ends , sharing no chromatin marks and exhibiting very low levels of 5-methylcytosine . The bp-2 south breakpoint , located within an expressed gene , contains ubiquitinated H2B , H3K27me3 and H3K4me2 , all typically associated with euchromatin [19] . The other clustered south breakpoints contain either of the two latter modifications ( bp-5 , bp-11 ) , but are generally devoid of chromatin marks . However , the local regions possess a variety of modifications and the four basic chromatin states [19] are interspersed ( see Figure S3 ) . It is conceivable that a particular combination of chromatin modifications may promote genome instability , or , as we observed for the clustered south breakpoints , a dearth of chromatin modifications and lack of 5-methylcytosine may be indicative of a chromatin state that underpins genomic instability/recombinogenic potential . In any event , features inherent to BAC T12G13 represent a boundary element or transition zone in which the chromatin state switches from one bearing heterochromatic marks to one indicative of a more euchromatic state ( Figure 6B ) . The large deletions that we document for the BP locus , as well as the close proximity of the HY4 locus where numerous large deletions have also been reported [18] , encourages speculation that the region is generally unstable and prone to deletions and other gross chromosomal rearrangement events . We therefore examined literature reports and employed mutant germplasm search engines to catalog the locations of deletions greater than 25 bp to ascertain if the BP/HY4 region is overrepresented in this data set . Figure S4 shows that the 5 bp alleles and the 15 hy4 alleles reported by Bruggemann and coworkers [18] , constitute 20 of the 120 deletion mutations associated with structural genes . We conclude that the BP/HY4 region is on average more prone to deletion/rearrangement events than most other regions of the genome .
Chromosome fragile sites have long been recognized in metazoans as regions which are prone to breakage when replication stress occurs , and activation ( breakage ) of some are implicated in various disease states and cancer [26] , [27] . Some of the common fragile sites that have been characterized at the molecular level are associated with breakpoints that delineate homologous syntenic blocks that have persisted during the evolution of eukaryote karyotypes [28] . Although a wealth of data correlates recurrent deletions and rearrangements with these fragile sites , very little is known about the initial breakage events , but rather only the sequences of junction regions provide clues as to the repair pathway . Some fragile site breakpoints possess microhomologies indicative of replication slippage or recombination based repair , whereas others possess no homology at the junctions . In the latter case , NHEJ must occur and likely involves structural features ( e . g . attachment points on the nuclear matrix ) that juxtapose free ends . Common fragile sites are A/T rich , possess various repeats and classes of repetitive DNAs ( particularly LINE/LTR type elements ) , are gene-poor regions , and occupy large tracts of eukaryotic genomes , sometimes extending over hundreds of thousands of nucleotides [29] , [30] . Our analyses of the SBC revealed that four of the five alleles harbor Athila/LTR retrotransposon elements at or near the breakpoints , which are A/T rich . The A/T rich sequences may contribute to secondary structure formation ( see below ) and could serve to anchor the chromatin fiber to MAR regions , which are known to be enriched for enzymes of DNA metabolism [31] . Indeed , the inhibition of topoisomerase I by camptothecin almost completely eliminates CFS breakage in cultured mammalian cells [32] . Fragile sites are found at the interface of chromosome R and G bands , classically defined as being early and late replicating , respectively [33] . Recent studies suggest that CFS activation ( breakage ) may be due to differential utilization of replication origins such that the CFS zone may experience replication stalling or fork collapse [34] , [35] , and secondary structures likely contribute to this process [36] . The paucity of replicon initiation in these regions necessitates the use of more distant origins and delays replicon completion , in part explaining their location at R/G boundaries . In Arabidopsis , high resolution profiling of replication in suspension cultured cells has revealed the locations of origins and their replication timing [37] , [38] . The south breakpoint cluster in T12G13 is situated between two distantly separated origins that span 271 Kbp; for comparison , the interorigin median and mean distances along chromosome 4 are 51 . 1 and 77 . 2 kbp , respectively . We propose that the south breakpoint cluster possesses the hallmark features of metazoan chromosome fragile sites , harboring sequences that promote replication fork collapse , and that loss of DNA may be limited by association of local sequences with the nuclear matrix , where topoisomerase/ligase activities can coordinate repair by a NHEJ ( or other ) mechanism . In addition to the deletions associated with BP/HY4 , this stochastic process may occasionally generate rearrangements and chromosomal fusion events that underpin chromosome evolution . Arabidopsis has been extensively used for comparative evolutionary genomics studies in plants . The 1001 Genomes Project has generated a wealth of sequence information , and numerous SNPs and indels are known to exist in many ecotypes [39]–[41] . The vast majority of this work has employed high throughput sequencing technologies , generating short sequence reads that are mapped onto scaffolds of reference genomes . While this strategy is useful and saturation can be achieved , it has two major disadvantages . First , some syntenic relationships may be masked , as many inversions and translocations cannot be accurately mapped . This is also the case for sequences associated with repetitive DNAs . Second , some copy number variants and their locations may go undetected and unrecognized , depending on the extent of divergence amongst them and their flanking sequences . While paired end mapping can be used to discover and verify structural variation [42] , only through the sequencing of contiguous long clones ( e . g . BACs ) can such sequence anomalies be accurately located and quantified . In Arabidopsis , several such studies have revealed striking departures from the Columbia reference genome . For example , comparative sequence analysis of a region around 100 map units on chromosome 1 in Columbia vs . Ler revealed that the region in Columbia is approximately 135 kbp , while the comparable region in Ler is only 71 kbp [43] . This appears to be due to several gross rearrangement events including two different ecotype specific duplications , a deletion and several ecotype specific transposition events that ultimately led to the shorter Ler region having fewer R genes than its Columbia counterpart [43] . In a similar vein , Lai et al . [44] resequenced a 371 kbp region of chromosome 3 from Columbia and Ler , mapping these reads onto both the Columbia reference genome as well as the more recent Wellcome Trust generated Ler draft genome , and discovered 61 misassemblies and large structural variants not represented by the draft genome . Lastly , the quartet mutation has afforded the possibility of tetrad analysis in Arabidopsis , and Lu and coworkers [45] sequenced the F1 genomes of a Col/Ler cross , reporting numerous known and heretofore unknown SNPs and indels . Analysis of the indels revealed that a deletion of 19 . 2 kbp encompasses the region of the bp-11 north breakpoint . These high resolution sequencing projects reveal that the genomes of Arabidopsis accessions are highly polymorphic and subject to rapid change as a result of normal meiotic recombination events and other rearrangements that may be due to replication errors and mobile element activity . While the power and economics of high throughput sequencing technologies cannot be disputed , discovering and mapping large structural variation in comparative genomics studies ultimately will depend upon more painstaking mapping and the generation and analysis of long clones . Fluorescent in situ hybridization or chromosome painting offers a middle ground for analyzing gross chromosomal architecture . Our experiments reveal that the two most commonly used Arabidopsis ecotypes , Columbia and Landsberg erecta , possess at least two major structural differences on chromosome 4 . The heterochromatic knob , which can be readily detected by DAPI staining , arose in Columbia via a pericentric inversion , moving a heterochromatic region towards the north end of the chromosome [13] , [14] . South of the centromere , there appears to be an indel that generates a significant length difference between the two GREEN probeset signals located at 4 . 3 Mbp and 6 Mbp of the Columbia reference genome . The precise nature of this polymorphism is unknown , but a comparison of homologous regions in A . lyrata and A . thaliana ( Columbia ) indicates that on two contigs that span the region , A . lyrata possesses over 600 kbp more DNA than Columbia . It is possible that this region has persisted in the Ler genome , accounting for the longer GREEN1/GREEN2 intervals that we observed . Given that similar polymorphisms are likely to exist on the other chromosomes , which could involve the mobilization of large blocks of sequence to perhaps distant sites ( e . g . an inversion involving over 1 Mbp ) , map-based cloning of some genes could be complicated and require protracted efforts . Advances in chromatin immunoprecipitation have permitted high throughput approaches for elucidation of the histone codes associated with genome elements . Roudier et al . , [19] conducted integrative epigenomic mapping in Arabidopsis , tracking 12 chromatin modifications to define four primary chromatin states associated with different coding and noncoding sequences . As in other organisms , the Arabidopsis pericentromeric DNA is enriched for several histone modifications , in particular methylation at H3K27 , H3K9 , and at H4K20 residues . In the context of maintaining genome integrity , H4K20 modifications are intriguing . In mammals , the H4K20 methylation state plays important roles in DNA damage repair as well as in class switch recombination ( CSR ) during the maturation of antibody producing cells [46] . CSR employs a NHEJ mechanism to exchange constant regions of immunoglobulin genes during B cell differentiation [47] . In mouse cells in which H4K20 methyl transferases are silenced , H4K20me1 accumulates , and this is associated with translocations and deletions of the IgH locus [46] . To our knowledge , there are no reports in Arabidopsis on the distribution and role of either di- or trimethylated H4K20 , but these modifications might prove to associate with the instability that we observe .
The alleles bp-1 , bp-2 , bp-3 , and bp-5 were described by Douglas et al . [8] . bp-11 was obtained through ABRC ( CS3161 ) . Table S1 contains information on all characterized bp mutants . Plants were grown in environmental growth chambers with a 16 hr day/8 hr night cycle at 22°C under fluorescent lighting of approximately 100 µE/m2 . Bacterial artificial chromosome clones were obtained from ABRC and DNA was prepared by employing Qiagen midi-prep columns . General molecular techniques were carried out as described by Sambrook et al . [48] . Genomic DNA for PCR was prepared using Sigma GeneElute columns . A list of primers used for determining the presence or absence of a locus , as well as for use in inverse PCR and phase 5 PCR is given in Tables S2 , S3 , S4 . Breakpoint junctions were cloned by employing inverse PCR . Genomic DNA was digested with a restriction enzyme that was shown by DNA gel blotting to give rise to an RFLP . The digested DNA was purified , diluted to approximately 1 µg/ml and subjected to overnight ligation to promote recircularization . Inverse PCR primers were then employed in PCR reactions to generate products containing the breakpoint junctions . These molecules were cloned into pJet1 . 2 ( Fermentas ) and sequenced ( The Centre for Applied Genomics , Toronto ) . For Southwestern analysis , a cDNA encoding the MAR binding protein AHL1 [15] was cloned by conducting RT-PCR on silique RNA , using oligo dT to prime first strand synthesis , and two primers: MAR Forward Nhe: 5′ AGGCTAGCGTCTTAAATATGGAGTCTACC 3′ and MAR Backward Bgl II: 5′ AAAGATCTGATTTCAAGTTACATTGACATTAATATCGG 3′ . The underlined sequences represent engineered Nhe I and Bgl II sites that were used to clone the cDNA into the expression vector pRSET B ( Invitrogen ) . Authenticated clones were mobilized into BL21 cells and expression of AHL1 induced by the addition of IPTG to a final concentration of 1 mM . After several hours of growth , induced and uninduced cultures were harvested by centrifugation and resuspended in Laemmli buffer ( 100 µl per 1 ml of culture ) , boiled , and stored at −20°C until needed . For Southwestern blotting , aliquots of protein extracts from induced and uninduced cultures were subjected to SDS-PAGE and transferred to nitrocellulose . The membrane was placed in 3% gelatin/TBS and kept at 4°C overnight . Protein refolding and blocking was carried out for 2 hours in a solution of 5% non-fat dry milk ( Carnation ) , 20 mM Tris , pH7 . 6 , 150 mM NaCl , 10 mM MgCl2 , 0 . 25 mM DTT , 0 . 05% Tween-20 , and 10 µg/ml salmon sperm DNA . DNA fragments to be used for probes were generated by PCR ( see Table S6 for primer set information ) and cloned into pJet1 . 2 ( Fermentas ) . Probe DNA was prepared by fill-in reactions using alpha 32P-dCTP and the Klenow fragment of DNA polymerase I ( Invitogen ) on restriction fragments . Binding was performed in the refolding buffer except that the non-fat milk concentration was reduced to 0 . 5% . The membranes were gently agitated at room temperature for 2 hours in approximately 3 ml of liquid containing the radiolabeled probes , then washed four times in 20 mMTris pH7 . 5 , 200 mM NaCl , 10 mM MgCl2 , 0 . 05% Tween-20 , 0 . 1% Triton ×100 , 0 . 25 mM DTT and 10 µg/ml salmon sperm DNA . Autoradiography was then employed to detect binding . In no case was binding detected using uninduced extracts . Probe labeling was assessed by gel electrophoresis and autoradiography and all probes were deemed to be of comparable specific activity . The BLASTn search tool , the MEME algorithm ( National Biomedical Computation Resource website; version 4 . 9 . 0; http://meme . sdsc . edu/meme/cgi-bin/meme . cgi; Accessed 2012 Nov 7 [49] ) and the RepeatMasker algorithm ( Institute for Systems Biology website , open-3 . 3 . 0 . Available: http://www . repeatmasker . org/cgi-bin/WEBRepeatMasker . Accessed 2012 Nov 7 . [50] ) were employed to identify common motifs and repetitive elements . Potential secondary structures were determined by employing the Mold algorithm [51] . The EPIGARA database ( Arabidopsis epigenetics and epigenomics group website; version 1 . 69; http://epigara . biologie . ens . fr/index . html . Accessed 2012 Nov 7 . ) of chromatin modifications was interrogated for locations of modifications and their proximity to the bp breakpoint junctions . Data was extracted from ChIP/chip studies conducted by Roudier et al . [19] and array based profiling of methylated DNA [52] . The pie charts in Figure 6 were generated by taking the IP/input log ratio for each region and dividing this by the sum of all IP/inputs for the region . Staged floral buds of wildtype Columbia and the Columbia based bp-5 and bp-11 alleles , along with Landsberg erecta ( Ler ) and the Ler derived bp-1 and bp-2 alleles were used as the starting materials . Pachytene chromosome spreads were prepared and identified according to the method of Stronghill and Hasenkampf [53] . Spreads were then subjected to fluorescence in situ hybridization as described by Lysak et al . [54] . Five chromosome 4 probesets , comprised of bacterial artificial chromosomes ( BACs ) , were used to determine the contour distances between these five sequences , which span chromosome 4 from 0 . 8 Mbp to 6 . 4 Mbp ( AGI coordinates ) . Probes were generated by employing a Nick Translation Kit ( Roche ) . Two of the five probesets ( north of the centromere ) were biotin-labeled BAC clones bracketing the heterochromatic knob region in Columbia: Red 1: BACs T7B11 , T2H3 and Red 2: BACs T4B21 , T1J1 , T32N4 . Detection of these signals was facilitated by goat anti-biotin antibodies ( Vector Laboratories ) and a secondary donkey anti-goat Cy3 conjugated antibody ( Jackson Immunoresearch ) . Three additional probesets were DIG-labeled BAC clones ( south of centromere ) , bracketing the region of the BP locus: Green 1: BACs F28D6 , T3E15; Green 2: BACs T15G18 , T25P22 and Green 3: T9A4 , F24G24 . These hybridization signals were detected by employing a mouse anti-DIG primary antibody and a donkey anti-mouse FITC conjugated secondary antibody ( Jackson Immunoresearch ) . Spreads were examined using a Zeiss axiophot epifluorescent microscope and a Plan-Neofluar 100×/1 . 3NA oil immersion objective lens . Northern Eclipse 5 . 0 software was used to capture images and measure the distance between FISH signals . Merged images were created using Photoshop CS5 software .
|
Chromosome evolution involves both small-scale ( e . g . single nucleotide ) changes , as well as large-scale rearrangements such as inversions , translocations , and fusion events . We investigated mutations of the BREVIPEDICELLUS gene of Arabidopsis , which is a master regulator of inflorescence architecture . These mutations are not due to single nucleotide changes , but rather to large deletions , some spanning nearly one million base pairs . Molecular and biochemical analyses reveal that the chromosome breakpoints cluster in an area that is rich in repetitive elements and harbor multiple binding sites for nuclear matrix proteins . Data mining revealed intriguing correlations between the breakpoint cluster and hotspots of genetic recombination , regions of the chromosome that have undergone several rearrangement events during evolution , and changes in histone protein modifications . We propose that these unstable regions are chromosome fragile sites that assist in marking a boundary between gene-poor , transcriptionally repressed centromeric chromatin and a more relaxed chromatin domain that is gene-rich .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"developmental",
"biology",
"plant",
"science",
"plant",
"growth",
"and",
"development",
"molecular",
"development",
"plant",
"genomics",
"plant",
"biology",
"biology"
] |
2012
|
Chromosome Fragile Sites in Arabidopsis Harbor Matrix Attachment Regions That May Be Associated with Ancestral Chromosome Rearrangement Events
|
Type 1 Epstein-Barr virus ( EBV ) strains immortalize B lymphocytes in vitro much more efficiently than type 2 EBV , a difference previously mapped to the EBNA-2 locus . Here we demonstrate that the greater transforming activity of type 1 EBV correlates with a stronger and more rapid induction of the viral oncogene LMP-1 and the cell gene CXCR7 ( which are both required for proliferation of EBV-LCLs ) during infection of primary B cells with recombinant viruses . Surprisingly , although the major sequence differences between type 1 and type 2 EBNA-2 lie in N-terminal parts of the protein , the superior ability of type 1 EBNA-2 to induce proliferation of EBV-infected lymphoblasts is mostly determined by the C-terminus of EBNA-2 . Substitution of the C-terminus of type 1 EBNA-2 into the type 2 protein is sufficient to confer a type 1 growth phenotype and type 1 expression levels of LMP-1 and CXCR7 in an EREB2 . 5 cell growth assay . Within this region , the RG , CR7 and TAD domains are the minimum type 1 sequences required . Sequencing the C-terminus of EBNA-2 from additional EBV isolates showed high sequence identity within type 1 isolates or within type 2 isolates , indicating that the functional differences mapped are typical of EBV type sequences . The results indicate that the C-terminus of EBNA-2 accounts for the greater ability of type 1 EBV to promote B cell proliferation , through mechanisms that include higher induction of genes ( LMP-1 and CXCR7 ) required for proliferation and survival of EBV-LCLs .
Epstein-Barr Virus ( EBV ) is a B-lymphotropic gamma herpesvirus which persistently infects over 90% of the adult population world-wide . EBV infection is usually asymptomatic , although in some cases the virus can be the causative agent of infectious mononucleosis [1] . EBV is also involved in some B cell cancers , such as Burkitt's Lymphoma ( BL ) , Hodgkin's Lymphoma and lymphoproliferative disease in immunocompromised hosts , in addition to various epithelial tumors , for example nasopharyngeal carcinoma ( NPC ) and gastric cancer [2] . In vitro , EBV can efficiently immortalize primary B cells , converting them into permanently growing lymphoblastoid cell lines ( LCLs ) , in which the viral genome is maintained in an episomal form . A specific viral gene expression program , known as latency III [3] , is activated in LCLs and this involves expression of six nuclear antigens ( EBNA-1 , -2 , -3A , -3B , -3C and -LP ) , three membrane proteins ( LMP-1 , -2A and -2B ) , and various non-coding small RNAs ( EBERs and miRNAs ) . Together these viral gene products activate the resting B cells and sustain their continuous proliferation . [1] . Work with recombinant EBV viruses has demonstrated that at least five of the latent proteins ( EBNA-1 , -2 , -3A , -3C and LMP-1 ) are essential for EBV-mediated B cell transformation , while EBNA-LP and LMP-2 can contribute to optimal efficiency of this process [4]-[11] . EBV isolates can be classified as type 1 or type 2 ( also known as type A and B ) based on linked sequence variation in the latent genes EBNA-2 , EBNA-3A , -3B , -3C and EBNA-LP . The most divergent locus between the two viral types is EBNA-2 , which defines the viral types . There is about 50% difference in the predicted primary amino acid sequence , which is strikingly high if compared to the extremely low degree of variation in the rest of the EBV genome [12]–[17] . Type 1 strains of EBV are ubiquitous in the world , whereas type 2 strains are frequent in some areas of Central Africa and some other parts of the world where malaria is endemic . In these regions , type 1 and type 2 EBV have approximately equal prevalence [18] . The major biological difference between the two viral types is that type 1 EBV immortalizes B cells in vitro much more efficiently than type 2 EBV [19] . Experiments with a recombinant type 2 EBV virus carrying a type 1 EBNA-2 sequence showed that this virus gained a type 1 immortalization phenotype , demonstrating that the difference in transformation efficiency is determined by the EBNA-2 locus [5] . The in vitro transforming activities of type 1 and type 2 EBV also correlate with the frequency of tumor formation in SCID mice inoculated with type 1 or type 2 EBV in vitro-transformed LCLs [20] , [21] . Despite the clear difference in growth phenotype in vitro , there is so far no obvious disease association or in vivo phenotype known for type 1 and type 2 EBV strains , although one study reported that type 1 EBV strains are significantly more likely to cause infectious mononucleosis , compared to type 2 strains [22] . Upon EBV infection of naïve B cells in vitro , EBNA-2 and EBNA-LP are the first viral latent proteins expressed . EBNA-2 is a transcription factor that can activate expression of the viral Cp and LMP promoters and many cell genes crucial for B cell survival and proliferation [1] , [23]–[26] . Activation of some of these promoters by EBNA-2 is enhanced by cooperation with EBNA-LP [27]-[31] . Sequence comparison analysis of the EBNA-2 allele from type 1 and type 2 EBV and different primate lymphocryptoviruses , led to identification of nine evolutionary conserved regions ( CR1 to CR9 ) which represent much of the total sequence homology between type 1 and type 2 EBNA-2 proteins and define some functional domains and important structures of the protein [13] , [32] ( see diagram below ) . CRs 1 to 4 at the N-terminus of the protein correspond to two self-association domains [33] , [34] and a poly-proline region , which consists of a variable number of consecutive proline residues , depending on the virus strain . CRs 5 to 9 are located in the C-terminal half of the EBNA-2 protein , separated from the N-terminal cluster of CRs by the diversity region , where the sequence similarity between type 1 and type 2 EBNA-2 is very low . CR5 is involved in association with the host DNA-binding protein RBP-Jk and CR6 mediates the interaction with the cell protein SKIP , which facilitates EBNA-2/RBP-Jk complex formation [35] , [36] . CR8 is part of an acidic transactivation domain ( TAD ) , which mediates gene transcriptional activation , whereas CR9 coincides with a nuclear localization signal ( NLS ) . An additional karyophilic signal is represented by the RG sequence , which consists of an 18-amino acid stretch rich in arginines and glycines [32] . Four domains of EBNA-2 have been shown to be involved in EBNA-2/EBNA-LP cooperation: the TAD , amino acids 1 to 58 , the RG motif and the CR7 [31] , [37] . Detailed mutational analysis of the type 1 EBNA-2 protein led to mapping the regions essential for transformation . These regions are mostly also essential for transactivation of the LMP-1 promoter , suggesting that transformation and transactivation functions of EBNA-2 are closely related [38] . In the N-terminal half of EBNA-2 , residues 3 to 30 have been shown to be required for induction of LMP-1 expression and , consequently , for immortalization maintenance , using the EREB2 . 5 trans-complementation system [39] . CR4 has been shown to contribute to EBNA-2 mediated immortalization of B cells , as mutant viruses with a deleted CR4 are drastically impaired in B cell transformation [40] . B cell infection experiments with recombinant viruses suggested that a minimum of seven prolines in the poly-proline region is required for transformation , whereas in the EREB2 . 5 trans-complementation assay , the whole poly-proline region was found to be dispensable for EBNA-2-mediated immortalization maintenance [41] , [42] . In the C-terminal half of the protein , the transactivation domain and the RBP-Jk-binding region are absolutely required for B cell immortalization and LMP-1 induction [38] , [43] . The RG motif has been demonstrated to be important for optimal B lymphocyte transformation efficiency [44] . However , a deletion mutant EBNA-2 protein lacking the RG sequence displayed a four-fold increase in the activation of the LMP-1 promoter in reporter assays , compared to wild-type EBNA-2 , suggesting that the RG domain is a negative regulator of EBNA-2 activity on the LMP-1 promoter [44] . EBNA-2 mechanisms of transcriptional activation have been identified by many studies , which have mainly involved type 1 EBNA-2 , whereas less is known about the type 2 EBNA-2 . EBNA-2 does not bind directly to DNA but is tethered to EBNA-2 responsive promoters by interacting with various cell DNA-bound transcription factors . For example , it interacts through its conserved WWPP motif with the transcriptional repressor RBP-Jk ( CBF1 ) , thereby converting RBP-Jk to the transcriptionally active form [35] , [45] . Additional cell sequence-specific transcription factors are involved in EBNA-2 recruitment at some target promoters , such as Spi-1/PU . 1 , AP-2 and AUF1 [46]–[50] . EBNA-2-mediated gene transcription is activated by the transactivation domain which interacts with several cell basal transcription factors ( TFIIH , TAF40 , TFIIB and p100/TFIIE ) and histone acetyltransferases ( p300/CBP and PCAF ) [51]–[54] . Other EBNA-2-interacting partners are involved in chromatin remodelling , such as proteins of the cell SWI-SNF complex [55] , [56] . EBNA-LP contains multiple copies of an N-terminal 66-aa domain encoded by two exons ( W1 and W2 ) , located within each of the BamHI W repeats of the EBV genome , and a unique C-terminal 45-aa domain encoded by the Y1 and Y2 exons within the downstream BamHI Y fragment [57]–[59] . Due to variations in the number of BamHI W repeats in different viruses and alternative splicing between the repeated W1W2 exons and the unique Y1Y2 exons , EBNA-LP proteins of different sizes can be expressed in EBV-infected cells . Genetic studies showed that recombinant viruses lacking the C-terminal 45 amino acids display a markedly lower immortalization efficiency compared to wild-type viruses and require feeder cells , suggesting that EBNA-LP is important but not essential for EBV-induced immortalization [6] , [9] . Importantly , EBNA-LP has been shown to cooperate with EBNA-2 and enhance EBNA-2 transcriptional activation of both viral ( LMP-1/LMP-2B , Cp ) and cell ( cyclin D2 ) target promoters [27]–[31] . Most of the cooperative function of EBNA-LP maps to the W1W2 repeats , whereas the unique carboxyl-terminal Y1Y2 domains modulate EBNA-LP function [28] , [29] . Only proteins with 2 or more copies of the W1W2 repeat are capable of cooperating with EBNA-2 and viruses with two BamHI W repeats are immortalization-competent [17] , [29] , [60] , [61] . LMP-1 is required for establishment of B cell transformation in vitro [8] and is also required for continuous proliferation of EBV-infected LCLs [62] . Regulation of LMP-1 by EBNA-2 is complex and involves many cell proteins , including RBP-Jk , PU . 1 , AP-2 , SWI-SNF , CBP/p300 , ATF/CREB [46]-[49] , [54] , [63] . Unlike other EBNA-2 target promoters ( e . g . LMP-2A ) , the EBNA-2/RBP-Jk interaction plays only a minor role in EBNA-2-induced activation of the LMP-1 promoter [64] . Since the EBNA-2 domains that are essential for B cell transformation and LMP-1 induction are similar , transactivation of LMP-1 by EBNA-2 is considered to play a key role in EBNA-2-induced B lymphocyte transformation [65] . Several studies have used microarray analysis to identify human genes that are targets of type 1 EBNA-2 [23]–[25] but until recently little was known about the ability of type 2 EBNA-2 to regulate gene expression . In earlier reports the abilities of type 1 and type 2 EBNA-2 to up-regulate gene expression were compared only on two individual promoters , LMP-1 and CD23 [66] , [67] . Recently we compared the host genes induced by type 1 EBNA-2 to those induced by the type 2 . Only a few genes were found to be differentially regulated ( CXCR7 , MARCKS , IL1β and ADAMDEC ) , with a stronger induction by type 1 EBNA-2 [26] . Among these , CXCR7 was the most differentially regulated gene and was also shown to be required for proliferation of EBV-infected LCLs . Expression of MARCKS , IL1β and ADAMDEC was very low in LCLs and may not be significant [26] . CXCR7 ( RDC1 , CMKOR1 ) is a G-protein coupled chemokine receptor with seven trans-membrane domains [68] , [69] . CXCR7 binds with high affinity to the inflammatory and homing chemokines CXCL12/SDF-1 and CXCL11/ITAC , which are also ligands for CXCR4 and CXCR3 respectively [70] , [71] . CXCR7 is believed to be an atypical chemokine receptor since neither coupling of the receptor to G-proteins nor CXCR7-mediated triggering of typical chemokine responses , such as chemotaxis or receptor-induced calcium mobilization , could be demonstrated [70]-[73] . Therefore , CXCR7 has been proposed to be mainly a modulator of CXCR4 and CXCR3 signalling , possibly acting as a “decoy” receptor to scavenge and sequester CXCL12/SDF-1 and CXCL11/ITAC [74] , [75] . This is suggested by the observation that CXCR7 expression on mature B cells inversely correlates with the activity of CXCR4 [76] . Alternatively , a function for CXCR7 as co-receptor for CXCR4 has also been proposed , as the two receptors can heterodimerize [77] . Analysis of CXCR7 expression in normal human leukocytes revealed that the mRNA is broadly expressed in various fractions of resting PBMCs , including B cells [76] , [78] , whereas the protein seems to be strictly expressed at the plasma membrane of monocytes and B cells [76] . However , surface CXCR7 was not consistently detected on B cells in another study [78] . CXCR7 has been implicated in tumorigenesis based on the observation that ectopic expression of the gene induces tumor formation in nude mice [79] . Moreover , CXCR7 protein has been shown to be expressed on many human and mouse tumor cell lines and to confer a strong growth and survival advantage to the cells [71] , [80]–[82] . In this study we show that LMP-1 and CXCR7 mRNAs are differentially induced by type 1 and type 2 EBV during the early stages of infection of primary B cells with recombinant EBV viruses , consistent with differential induction of these genes being the basis for the weaker ability of type 2 EBV strains to transform primary B cells . EBNA-2 type , rather than the type of EBNA-LP , is shown to be the major determinant of the differential induction of LMP-1 oncoprotein and the C-terminal region of EBNA-2 is responsible for the superior ability of the type 1 protein to maintain proliferation of B cells . Within this region , the RG , CR7 and TAD are the minimum type 1 domains required for type 1 growth phenotype and sufficient LMP-1/CXCR7 expression levels . The results demonstrate a mechanism for the enhanced capability of type 1 strains of EBV to transform B cells into LCLs .
We previously identified CXCR7 as the most differentially regulated type-specific EBNA-2-target gene in the human genome , being more strongly induced by type 1 EBNA-2 in a BL cell background [26] . RNAi experiments demonstrated that CXCR7 is required for proliferation of EBV-LCLs [26] . We also showed , in an EBV-positive BL cell line with oestrogen-regulated EBNA-2 proteins , that type 1 EBNA-2 induced LMP-1 protein more rapidly and to a higher extent compared to the type 2 [26] . This has been now verified at the RNA level using the same system . Briefly , this consists of Daudi cells ( which lack the EBNA-2 locus [83] and therefore expression of LMP-1 ) engineered to stably express type 1 or type 2 chimaeric EBNA-2 proteins fused to the hormone-binding domain of the oestrogen receptor ( ER-EBNA-2 ) [26] . In the absence of oestrogen , the ER-tagged EBNA-2 proteins are localized to the cytoplasm and therefore inactive [84] . Addition of oestrogen results in normal localization of the EBNA-2 fusion proteins to the nucleus and therefore in activation of the LMP-1 gene from the resident EBV genome . In a time-course experiment of oestrogen stimulation over 48 hours , total cell RNA from the cell lines containing type 1 or type 2 ER-EBNA-2 was analyzed by ribonuclease protection assay ( RPA ) using an LMP-1 specific probe ( Figure 1 A ) . When type 1 EBNA-2 function was activated , LMP-1 mRNA levels were strongly induced as soon as 8 hours after the stimulation and were maintained thereafter throughout the time-course . In contrast , after induction of type 2 EBNA-2 function , markedly lower levels of LMP-1 transcripts were detected . Analysis by qRT-PCR for LMP-1 mRNA gave a similar result ( Figure 1 B ) and qRT-PCR analysis was also performed for CXCR7 transcripts levels in the same experiment ( Figure 1 C ) . These were found to be significantly higher in type 1 EBNA-2 cell lines stimulated with oestrogen than in type 2 cells that had undergone the same treatment . Agarose-gel electrophoresis analysis of total cell RNA used for the RPA analysis confirmed equal loading and no degradation of the samples ( data not shown ) . The somewhat higher expression of type 2 ER-EBNA-2 in the Daudi cell line ( Figure S1 and [26] ) further emphasizes the superior induction of target genes by type 1 ER-EBNA-2 . The more rapid induction of LMP-1 by type 1 EBNA-2 was also confirmed in the type 2 EBV P3HR1 cl . 16 BL cell line . The resident EBV genome in P3HR1 cl . 16 line also lacks the EBNA-2 locus and the Y1Y2-coding region of EBNA-LP , having a deletion similar but not identical to that present in Daudi cells [83] , [85] . Plasmids expressing type 1 or type 2 ER-EBNA-2 ( p554-4 and its type 2 derivative , described in Materials and Methods and in [26] , [86] ) were transfected into P3HR1 cl . 16 cells and stable cell lines were selected with G418 . In P3HR1 cl . 16 cells , despite the absence of EBNA-2 , basal expression of LMP-1 can still be detected , albeit at very low levels ( Figure 1 D ) . Nevertheless , when a time-course experiment of oestrogen stimulation was performed , a clear increase of LMP-1 expression levels above background was detected by western blotting when type 1 EBNA-2 function was activated ( Figure 1 D ) . LMP-1 induction was detected as soon as 4 hours after oestrogen stimulation; after 8 hours LMP-1 expression level reached its maximum and was thereafter maintained throughout the time-course . Conversely , when type 2 EBNA-2 function was activated by oestrogen , LMP-1 levels were induced only moderately above the background , with a modest peak at 20 hours followed by a slight decrease ( Figure 1 D ) . As EBNA-LP is important in cooperation with EBNA-2 for induction of LMP-1 [28] , [29] , we determined which isoforms of EBNA-LP are expressed in normal Daudi and P3HR1 cl . 16 cells and in the derivative stable cell lines bearing the oestrogen-inducible EBNA-2 proteins ( Figures S1 and S2 ) . In several previous reports , depending on the antibodies used , one or two species of EBNA-LP ( around 31 and 37 kDa ) have been detected in Daudi cells , both lacking the Y1Y2 domain [87]–[90] . Here we used the 4D3 antibody and observed the 37 kDa isoform ( Figure S1 , lane 4 ) , which corresponds to a 4-repeat species , as determined by comparison with the EBNA-LPs expressed in B95-8 and in 293 cells transiently transfected with plasmids coding for 3- or 7- repeat EBNA-LP as standards ( Figure S1 , lanes 1 , 2 and 3 ) . P3HR1 cl . 16 cells have been reported to express Y1Y2-deleted EBNA-LP species migrating as a major band of around 28–30 kDa and a minor band of about 48–50 kDa on a SDS-PAGE [91] and this was confirmed by our western blot analysis with the 4D3 antibody ( Figure S2 , lane 4 ) . We identified these species as containing 3 and 6 W repeats , respectively ( by comparison with B95-8 ) . In both cell lines these truncated EBNA-LP proteins are type 2 , since they were detected with the 4D3 antibody , which detects both type 1 and type 2 EBNA-LP [89] , but not with the JF186 antibody , which is type 1-specific [92] . In the stable cell lines expressing ER-EBNA-2 proteins several EBNA-LP isoforms were detected , with 2 to 4 repeats in Daudi:ER-EBNA-2 T1/T2 ( Figure S1 , lanes 5 to 8 ) and with 2 to 7 repeats in P3HR1 cl . 16:ER-EBNA-2 T1/T2 ( Figure S2 , lanes 5 to 8 ) . These species are type 1 , since they can be detected with the JF186 antibody , and are expressed from the plasmids that code for the ER-EBNA-2 proteins ( p554-4 and its type 2 derivative [26] , [86] ) . In fact , these vectors contain not only the ER-EBNA-2 open reading frame , but also 2 W repeats and the Y fragment , both of type 1 sequence . The expression of EBNA-LP species that are larger in size than that predicted simply from the number of W fragments has been already reported by others [92]-[94] and may indicate splicing of RNA transcribed more than once around the plasmid . In both systems ( Daudi and P3HR1 cl . 16:ER-EBNA-2 T1/T2 ) several isoforms of full-length type 1 EBNA-LP are present but no full-length type 2 EBNA-LP is available . To exclude the possibility that type 1 EBNA-LP proteins are able to enhance the ability of type 1 ER-EBNA-2 to induce LMP-1 but not that of the type 2 ER-EBNA-2 , we performed an assay for EBNA-LP function , similar to that used in earlier studies [17] , [29] , whereby the effect of EBNA-LP on EBNA-2-induced expression of LMP-1 from the resident EBV genome can be examined . This assay was adapted to compare the cooperative effect of type 1 and type 2 EBNA-LP on type 1 and type 2 EBNA-2-induced expression of LMP-1 in Daudi cells . Transfection of vectors expressing type 1 or type 2 EBNA-2 alone did not produce any significant induction of LMP-1 , despite the strong expression of both types of EBNA-2 proteins , as assessed by immunoblotting ( Figure 2 ) . Likewise , no LMP-1 was detected when plasmids coding for either type 1 or type 2 EBNA-LP bearing 3 repeats were transfected , even though both EBNA-LP species were clearly detected on a western blot as an approximately 30 kDa band . Co-transfection of type 1 EBNA-2 and type 1 EBNA-LP plasmids produced a clear induction of LMP-1 protein . A similar , but slightly lower level of induction was observed when the type 2 EBNA-LP was co-expressed with type 1 EBNA-2 . In contrast , up-regulation of LMP-1 induced by the type 2 EBNA-2 , in the presence of either type of EBNA-LP , was very modest and equal with either type 1 or type 2 EBNA-LP ( Figure 2 ) . These results indicate that the weaker LMP-1 up-regulation induced by type 2 EBNA-2 can be ascribed to the EBNA-2 type and is not affected by the type of EBNA-LP present . Type 1 EBNA-2 up-regulates LMP-1 better than type 2 EBNA-2 , in the presence of either type 1 or type 2 EBNA-LP . These observations support the conclusion that , in the Daudi and P3HR1 cl . 16:ER-EBNA-2 T1/T2 cell lines , the impaired ability of type 2 EBNA-2 to induce LMP-1 is not due to sub-optimal cooperation with the full-length type 1 EBNA-LP isoforms present in the systems , but is rather due to the type 2 EBNA-2 protein . Having showed , in our previous report [26] and in the present study in assays in cell lines , that type 1 and type 2 EBNA-2 differentially regulate LMP-1 and CXCR7 , we examined whether this phenomenon occurs also during infection of primary B cells with BAC-derived EBV recombinant viruses , expressing either type 1 or type 2 EBNA-2 . Early stages of infection were studied since the differences in kinetics of gene regulation induced by the two types of EBNA-2 appeared rapidly in our cell line-based systems . Normalized amounts of the EB viruses expressing type 1 or type 2 EBNA-2 ( described in Materials and Methods ) were used to infect aliquots of 2×106 human primary B cells , purified from peripheral blood . Samples were taken at 1 , 2 , 4 and 8 days post-infection and examined by western blotting and quantitative RT-PCR ( qRT-PCR ) for EBNA-2-target gene expression . Similar amounts of type 1 and type 2 EBNA-2 proteins were expressed , as determined by the EBNA-2 immunoblot ( Figure 3 A ) . In type 1 EBV-infected cells , LMP-1 transcripts were detected as soon as 2 days post-infection and gradually increased thereafter reaching typical LCL levels of expression by day 8 , whereas in type 2-infected cells LMP-1 induction was delayed , with overall lower levels than those observed in type 1 infection ( Figure 3 B ) . These different kinetics of LMP-1 mRNA induction were consistently observed in 4 independent infection experiments , although the overall level of expression and precise timing varied slightly in different experiments , this variability perhaps being due to different B cell donors being used each time . As we were able to perform these experiments only on a small scale , LMP-1 protein could not be detected by western blotting at these early times after infection . qRT-PCR analysis for CXCR7 mRNA revealed that uninfected resting B cells ( harvested at day 0 ) already expressed substantial basal amounts of mRNA ( Figure 3 C ) , an observation consistent with other reports [76] , [78] . Following viral infection , CXCR7 mRNA increased up to 2 days after infection and then settled back to the LCL level ( Figure 3 C ) . On the other hand , in type 2 infection the mRNA for CXCR7 was consistently at about a 4-fold lower level than in type 1 virus-infected cells ( Figure 3 C ) . This trend in CXCR7 mRNA levels was observed in 3 independent experiments , with minor differences in the overall levels of expression . Expression of the chemokine receptor CXCR4 , which shares the ligand CXCL12/SDF-1 with CXCR7 , has been shown to be down-regulated in primary B cells upon infection by EBV [95]-[97] . Uninfected primary B cells were found to express high levels of CXCR4 mRNA ( Figure 3 D ) , in agreement with other reports [98] . Upon viral infection , CXCR4 transcript levels were effectively repressed in cells infected with both type 1 and type 2 EBNA-2 recombinant EB viruses ( Figure 3 D ) . In order to establish LCLs , serial dilutions of preparations of EBV-BACs expressing type 1 or type 2 EBNA-2 were used to infect primary B lymphocytes ( 106 cells ) and their outgrowth was monitored over time . As expected , type 1 EBV-BAC-infected cells grew very rapidly , leading to LCLs within approximately 1 month post-infection , whereas type 2 transformants were much more difficult to expand and yielded cell lines only after approximately 4 months of culture . The growth phenotype observed with our recombinant type 1 and type 2 EBNA-2-expressing viruses is consistent with the growth behaviour originally reported for the natural strains of EBV B95-8 ( type 1 prototype ) and AG876 ( type 2 prototype ) [19] . These differences in growth kinetics , which were revealed by microscopic observation of the cells in culture , were mirrored by the counts of proliferating cells assessed at a single time-point after infection , for each of the cell lines originally infected with serial dilutions of virus ( Figure S3 ) . For example , the amount of live cells counted in type 1 LCLs originally infected with either 1600 or 320 GRUs of viral preparation was increased compared to the corresponding type 2 transformants by 53 and 22-fold , respectively ( Figure S3 ) . Once LCLs were established , hygromycin selection was applied to ensure the cells contained the BAC-EBV constructs , which bear the hygromycin-resistance gene , and not an EBV genome that might potentially have been endogenous at a low level in the donor B cells . The fully established type 1 and type 2 EBNA-2 LCLs were then further checked by examining the pattern of expression of all the EBV latent antigens , using western blotting ( Figure S4 ) . The EBV latent antigens examined are expressed at similar levels and display the expected size . EBNA-LP isoforms with slightly variable number of W repeats ( from 3 to 5 ) were detected in type 1 and type 2 LCLs , but this is well above the minimum required for cooperation with EBNA-2 [17] , [60] . Some differences were observed in EBNA-3A and -3C expression levels among the several clones analyzed , but do not seem to be consistently associated with one specific type of LCL ( Figure S4 ) . We previously developed a functional assay that distinguishes the ability of type 1 and type 2 EBNA-2 to maintain B cell proliferation using the EREB2 . 5 cell line [26] , [86] . EREB2 . 5 cells contain a P3HR1 EBV genome lacking the EBNA-2 locus and a separate plasmid ( p554-4 ) expressing the ER-EBNA-2 fusion protein . Hence , EBNA-2 activity and therefore survival and proliferation of the cells are dependent on oestrogen . In the absence of oestrogen , EBNA-2 function is inhibited by retention of the conditional ER-EBNA-2 protein in the cytoplasm but cell proliferation can be rescued if an OriP plasmid expressing constitutive type 1 EBNA-2 protein is transfected into the cells . In this situation the wild-type type 1 EBNA-2 functionally replaces the ER-EBNA-2 fusion protein . In contrast , if a wild-type type 2 EBNA-2-expressing OriP vector is transfected into oestrogen-starved cells , the cells undergo growth arrest . In this assay only type 1 but not type 2 constitutive EBNA-2 protein can provide essential EBNA-2 functions in trans . The trans-complementation assay was used in the present study to identify the regions of the type 1 EBNA-2 protein that can complement the deficiency of the type 2 protein to maintain proliferation of LCLs . For this , several type 1/type 2 EBNA-2 chimaeras were generated ( by swapping sequences between the two types of protein ) and these were tested for their ability to sustain cell growth of oestrogen-starved EREB2 . 5 cells . The complete set of chimaeras tested and the respective growth phenotype observed in the trans-complementation assay are shown in Figure 4 B . The chimaeras were precise swaps of EBNA-2 sequence without extraneous amino acids at the point of fusion . OriP-p294 plasmids expressing the chimaeric or wild-type type 1 or type 2 EBNA-2 proteins were Amaxa-transfected into EREB2 . 5 cells that had been growing normally in the presence of oestrogen . Immediately after transfection , oestrogen was withdrawn from the culture medium and cell proliferation was monitored over time by counting numbers of live cells ( Figure 4 C ) . When the C-terminal region of type 1 EBNA-2 protein spanning from the WWPP motif ( codons 323–326 ) , within the RBP-Jk binding domain , to the end of the protein was swapped at homologous amino acid positions into the type 2 EBNA-2 protein ( chimaera 2 ) , live cell counts were similar to those obtained with wild-type type 1 EBNA-2 in the EREB2 . 5 growth assay ( Figure 4 C ) . However , a null growth phenotype , similar to wild-type type 2 or empty vector control , was observed with the converse chimaera ( chimaera 1 ) , generated by joining together the type 1 sequence from codon 1 to the WWPP motif with type 2 sequences spanning from the WWPP motif to the last codon of the protein ( Figure 4 C ) . Therefore , the C-terminus of type 1 EBNA-2 is able to confer a type 1 growth phenotype to type 2 EBNA-2 in the EREB2 . 5 trans-complementation assay . The C-terminus of the EBNA-2 protein includes the RG motif , the conserved region 7 ( CR7 ) , the transactivation domain ( TAD ) and the nuclear localization signal ( NLS ) ( Figure 4 A ) . To further map the minimum type 1 sequences within the C-terminus of EBNA-2 responsible for the higher B cell immortalization efficiency , additional chimaeras were constructed by swapping different combinations of the RG , CR7 , TAD and NLS regions from the type 1 into the type 2 protein ( chimaeras 3 to 7 ) . Only when the type 2 RG , CR7 and TAD regions were together replaced with homologous type 1 sequences ( chimaera 7 ) , accumulation of proliferating cells was restored to almost wild-type type 1 levels ( Figure 4 C ) . In contrast , chimaeras from 3 to 6 behaved like wild-type type 2 EBNA-2 , as they were unable to sustain proliferation ( Figure 4 C ) . Long-term culture of oestrogen-independent EREB2 . 5 cells expressing the different EBNA-2 chimaeras and wild-type proteins resulted in establishment of continuously proliferating LCLs only in the case of chimaeras 2 and 7 and wild-type type 1 EBNA-2 . These results demonstrate that , within the C-terminus of the type 1 EBNA-2 , the RG , CR7 and TAD are the minimum type 1 sequences required to complement the lack of ability of type 2 EBNA-2 to rescue growth of oestrogen-depleted EREB2 . 5 cells . To ensure that the lack of growth of oestrogen-depleted EREB2 . 5 cells was not due to either lack of expression or inappropriate protein localization of the chimaeras , western blot and immunofluorescence experiments were performed . Western blot analysis for EBNA-2 on extracts from transfected EREB2 . 5 cells grown without oestrogen revealed that all the chimaeric proteins display the expected size and are expressed at equal levels , similar to wild-type type 1 and type 2 EBNA-2 proteins ( Figure 4 D ) . Levels of EBNA-2 expressed from the OriP plasmids was similar to those normally found in LCLs ( Figure 4 D ) . For the analysis , the PE2 antibody was used , which was previously shown to detect the type 1 and the type 2 EBNA-2 with equal efficiency on a western blot [26] . Immunofluorescence analysis of HeLa cells transiently transfected with OriP-p294 plasmids , expressing the chimaeric and wild-type EBNA-2 proteins , confirmed that all the chimaeras are appropriately localized to the nucleus of the cells with exclusion from nucleoli ( Figure S5A ) , with a pattern similar to that of wild-type type 1 and type 2 EBNA-2 ( Figure S5B and [99] ) . Expression of the EBNA-2-direct viral target and major EBV oncogene LMP-1 was assessed in the EREB2 . 5 growth assay . Western blot analysis revealed that at 3 and 5 days after transfection , in oestrogen-starved cells expressing type 1 EBNA-2 and chimaeras 2 and 7 , LMP-1 protein expression was maintained at levels similar to those detected in EREB2 . 5 cells normally grown in the presence of oestrogen ( Figure 5 A ) . In contrast , LMP-1 expression was lost in type 2 EBNA-2 or empty vector-transfected cells ( Figure 5 A ) . qRT-PCR analysis for the mRNA of the cell gene CXCR7 showed increased levels following expression of either type 1 EBNA-2 or chimaeras 2 and 7 , with a 14-fold increase ( type 1 EBNA-2 ) or approximately 7-fold increase ( chimaeras 2 and 7 ) , compared to non-transfected cells , at 1 week after transfection ( Figure 5 B ) . In contrast , when type 2 EBNA-2-expressing vector was transfected , CXCR7 mRNA was maintained at the same levels normally observed in non-transfected cells ( fold-increase of 1 ) up to 5 days after transfection but after this time-point it was not further induced and the cells died ( Figure 5 B ) . These results demonstrate that LMP-1 protein and CXCR7 mRNA expression levels , induced by type 1 and type 2 EBNA-2 and chimaeras 2 and 7 , correlate with the growth phenotype in the EREB2 . 5 trans-complementation system . Loss of LMP-1 and CXCR7 expression in type 2 EBNA-2-expressing cells is rapidly followed by growth arrest of the cells , whereas in cells expressing type 1 EBNA-2 , chimaera 2 or 7 , maintenance of LMP-1 and CXCR7 expression is accompanied by sustained proliferation of the cells . Therefore the C-terminus of type 1 EBNA-2 in chimaeras 2 and 7 is sufficient to confer induction of type 1 growth phenotype and approximately type 1 expression levels of LMP-1 and CXCR7 . In our mapping analysis performed to identify the type 1 regions of EBNA-2 important for B cell growth , EBNA-2 sequences from the EBV strains B95-8 ( type 1 prototype ) and AG876 ( type 2 prototype ) were used . To ensure that the functional differences being mapped are typical of EBV type sequences , the C-terminus of EBNA-2 from additional type 1 and type 2 isolates was sequenced and compared to the prototypes B95-8 and AG876 . The EBV strains chosen for the analysis were derived from laboratory LCL , BL and NPC cell lines ( Table 1 ) . Comparison of type 1 isolates to the type 1 strain of reference B95-8 showed that , in the region examined , there is a very high percentage of identity ( > 98% ) with only a few nucleotide sequence variations detected . Most of the changes are silent but a few produce a change in the amino acid coded . Changes in the amino acid sequence were almost always located outside the known EBNA-2 functional domains; only in 2 strains ( Mak 1 and MABA , Table 1 ) were there changes within a functional domain ( TAD and NLS respectively ) . Similarly , alignment of the sequence coding for the C-terminus of EBNA-2 from several type 2 isolates against the prototype strain AG876 , indicated that the sequence identity is almost 100% , with only 2 strains ( Alouek and Wewak , Table 1 ) displaying nucleotide variations ( 2 silent and 1 non-silent , but localized outside functional domains ) . These results indicate that the functional differences between type 1 and type 2 EBNA-2 , which we have mapped to this region , are representative of viral type sequence . EBNA-2-mediated activation of both cell ( cyclin D2 ) and viral promoters ( LMP-1/LMP-2B and Cp ) can be greatly augmented by cooperation with the EBV latent protein EBNA-LP , which has been shown to be required for optimal virus-induced B cell transformation [9] , [27]–[29] . Several EBNA-LP proteins , of different type and size , are believed to be present in the EREB2 . 5 trans-complementation system . To address whether these EBNA-LP species may influence the growth phenotype induced by the wild-type ( type 1 and type 2 ) and chimaeric EBNA-2 proteins in oestrogen-starved EREB2 . 5 cells we first determined which specific EBNA-LP isoforms are expressed in the EREB2 . 5 cell line and in the derivative oestrogen-independent LCLs established with type 1 EBNA-2 and chimaeras 2 and 7 ( Figure 6 ) . For this , western blot analysis was performed comparing two anti-EBNA-LP antibodies: JF186 ( type 1-specific ) and 4D3 , which recognizes both types of EBNA-LP [89] , [92] . Using the 4D3 antibody , a type 2 EBNA-LP species of around 52 kDa was detected in non-transfected EREB2 . 5 cells , normally grown in oestrogen-supplemented medium ( Figure 6 A , lane EREB2 . 5 + est ) , and in all the oestrogen-autonomous LCLs ( Figure 6 A and B ) . This isoform is encoded by the resident P3HR1 genome and is known to lack the Y1 and Y2 domains [100] . Both antibodies detected a type 1 EBNA-LP isoform containing 3 W repeats , as determined by comparison with lysates from 293 cells transiently transfected with a plasmid expressing 3-repeat EBNA-LP ( Figure 6 A , lane T1ELP3R ) . This EBNA-LP is encoded by the endogenous plasmid p554-4 and it was detected not only in non-transfected oestrogen-dependent EREB2 . 5 cells ( Figure 6 A , lane EREB2 . 5 + est ) , but also in all the oestrogen-independent LCLs established by transfecting chimaeras 2 and 7 or type 1 EBNA-2 , analyzed over a period of 6 and 3 months ( Figure 6 A and B , respectively ) . This observation suggests that in these experiments the p554-4 plasmid is not completely lost from the oestrogen-free LCLs . This was corroborated by the EBNA-2 immunoblot that showed continuous expression of ER-EBNA-2 over time , although at variable levels . It is likely that some of the p554-4 plasmid has become integrated in these cells since , after extraction of low molecular weight DNA from the oestrogen-independent LCLs , only the OriP-p294 plasmid ( expressing constitutive EBNA-2 proteins ) but not the p554-4 vector was rescued ( data not shown ) . Since only a type 1 full-length EBNA-LP is present in the EREB2 . 5 assay , it might be that a type-specific cooperation between EBNA-2 and EBNA-LP would be important for optimal proliferation in the EREB2 . 5 growth system and that the lack of a full-length type 2 EBNA-LP species could hamper the growth-promoting abilities of type 2 EBNA-2 . We tested this issue directly by transfecting plasmids expressing either type 1 or type 2 EBNA-2 into EREB2 . 5 cells , in the presence or absence of a plasmid expressing a 3-repeat type 2 EBNA-LP , to check whether the presence of a full-length type 2 EBNA-LP ( homologous to the endogenous type 1 3-repeat EBNA-LP ) would allow the type 2 EBNA-2 to sustain cell proliferation in an oestrogen-independent fashion ( Figure 7 ) . Accumulation of proliferating cells was assessed each week after transfection and oestrogen withdrawal over a period of 3 weeks ( Figure 7 A ) . Cells expressing type 1 or type 2 EBNA-2 alone displayed proliferation levels similar to those observed in Figure 4 C . Addition of the type 2 EBNA-LP did not produce any significant effect on growth of either type 1 or type 2 EBNA-2-expressing cells or control cells . There were no significant differences in the levels of EBNA-2 proteins comparing the type 1 and type 2 forms , as assessed by western blotting with the PE2 antibody ( Figure 7 B ) . Immunoblotting with the 4D3 antibody confirmed good expression of the type 2 3-repeat EBNA-LP from the transfected pSNOC vector . This EBNA-LP isoform displays a slightly higher molecular weight on SDS-PAGE compared to the endogenous type 1 3-repeat EBNA-LP ( Figure 7 B ) because of the substantial differences ( 12 amino acids ) in the primary amino acid sequences between type 1 and type 2 EBNA-LPs . This experiment demonstrates that the inability of type 2 EBNA-2 to rescue growth of oestrogen-depleted EREB2 . 5 cells is not due to the absence of a full-length type 2 EBNA-LP in the system .
The most striking biological difference between type 1 and type 2 EBV is that type 1 strains are far better than type 2 at inducing primary B lymphocyte transformation in vitro and the EBNA-2 gene is the key determinant of this difference [5] , [19] . Following EBV infection of naïve B cells in vitro , EBNA-2 drives B cell transformation into LCLs by acting as a potent transcriptional activator of both viral and cell genes which in turn will then cause survival and proliferation of the infected cells [1] , [23]–[26] . Our results demonstrate a mechanism that can account for the superior ability of type 1 EBV to produce LCLs . Primary B cell infection experiments using BAC-derived recombinant EB viruses showed that higher and more rapid induction of some EBNA-2-target genes ( which are required for continuous proliferation of EBV-infected LCLs ) by type 1 EBNA-2-expressing virus correlates with the enhanced B cell transforming capability of this virus ( Figure 3 and Figure S3 ) . In this study we have focussed our analysis on two genes , among all the direct target genes of EBNA-2 , based on their crucial importance in B cell proliferation in the context of EBV infection and these are the viral oncogene LMP-1 and the cell gene CXCR7 . Among the viral EBNA-2-targets , LMP-1 is the major EBV oncogene and is required for initiation of primary B cell transformation in vitro [8] and also for maintenance of continuous proliferation of EBV-LCLs [62] . Our time-course analysis of induction of LMP-1 and CXCR7 during the initial stages of infection with type 1 or type 2 EBNA-2 BAC-EBV , revealed that with type 2 EBNA-2 LMP-1 induction is delayed and weaker compared to type 1 EBV-infected cells , and CXCR7 expression levels are not maintained consistently throughout the time-points analyzed with type 2 EBNA-2 ( Figure 3 B and C ) . However , once a type 2 LCL is eventually established ( up to 4 months after infection , with a delay of approximately 3 months compared to type 1 LCLs ) , LMP-1 and CXCR7 expression levels are similar to those detected in type 1 LCLs ( Figure 3 B and C ) , confirming that those genes are essential for long-term proliferation of EBV-LCLs . Our results indicate that the key difference in transformation efficiency between type 1 and type 2 EBV is determined by differences in induction of the pro-survival genes LMP-1 and CXCR7 occurring during the early stages of infection . Differential regulation of LMP-1 and CXCR7 by the two types of EBNA-2-expressing viruses during infection is consistent with the results obtained in cell line based assays . We have shown that in Daudi and P3HR1 cl . 16 cell line backgrounds , using oestrogen-regulated EBNA-2 proteins , LMP-1 protein and CXCR7 mRNA are more potently and rapidly induced by the type 1 EBNA-2 ( Figure 1 and [26] ) . Our finding that the CXCR7 gene , together with LMP-1 , accounts for the greater transforming potential of type 1 EBV , compared to type 2 , is particularly relevant in light of the fact that two other oncogenic viruses , Kaposi's Sacroma-associated herpesvirus ( KSHV ) and human T-lymphotropic virus type 1 ( HTLV-1 ) , have been shown to up-regulate this gene during infection . CXCR7 induction by HTLV-1 Tax protein has been reported to have a pro-survival effect on HTLV-1-infected T cells and in KSHV-infected hDMVECs induction of CXCR7 is essential for spindle cell transformation , much as occurs with EBV-mediated B cell transformation [78] , [79] , [101]-[103] . Targeting of CXCR7 may therefore represent a common mechanism exploited by oncogenic viruses to induce cell transformation . CXCR7 binds to the same ligand as CXCR4 ( CXCL12/SDF-1 ) and has been proposed to function as a modulator of CXCR4-mediated signalling [77] , [104] , [105] . Several reports indicate that CXCR4 is down-regulated in primary B cells upon infection by EBV [95]–[97] and this was confirmed by our qRT-PCR time-course analysis during infection with type 1 and type 2 EBNA-2 recombinant viruses ( Figure 3 D ) . At least five EBV latent proteins have been found potentially to be involved in modulation of CXCR4 expression in EBV-LCLs . Stable expression of EBNA-2 and LMP-1 in the EBV-negative BJAB cell line was shown to induce down-regulation of CXCR4 [96] . Moreover , whole genome microarray studies indicated that CXCR4 expression is down-regulated by EBNA-3A and EBNA-3B in EBNA-3A and -3B knock-out LCLs , but up-regulated by EBNA-3C in LCLs with conditional EBNA-3C protein [106]–[109] . In our infection experiments CXCR4 was effectively repressed in both type 1 and type 2 EBNA-2-virus infected cells by the day 8 time-point ( Figure 3 D ) . In both type 1 and type 2 infections , the kinetics of CXCR4 regulation were almost complementary to that of LMP-1 ( Figure 3 B and D ) , consistent with the notion that both EBNA-2 and LMP-1 contribute to but are not the only determinants of CXCR4 regulation . From the EBNA-2 chimaeras tested , chimaeras 2 and 7 indicated that the RG , CR7 and TAD are the minimum type 1 sequences required to confer a type 1 growth phenotype to a type 2 EBNA-2 protein in the EREB2 . 5 trans-complementation assay ( Figure 4 B and C ) . This result was quite surprising and unexpected as most of the sequence variation between type 1 and type 2 EBNA-2 lies in the N-terminal half of the protein , mostly in the diversity region ( Figure 4 A ) and not in the C-terminus . In fact , in the C-terminus the sequence homology between the two types of EBNA-2 proteins is 62% , whereas in the N-terminus this is 48% and only 30% in the diversity region . Importantly , in the EREB2 . 5 assay the growth phenotype correlated with the ability of the EBNA-2 proteins to induce LMP-1 and CXCR7 genes . Type 1 EBNA-2 and chimaeras 2 and 7 maintained high levels of expression of those genes , which in turn drive continuous proliferation of the cells leading to LCLs establishment ( Figures 5 and 6 ) , whereas in cells expressing type 2 EBNA-2 , LMP-1 and CXCR7 were not maintained at normal physiological levels ( Figure 5 ) . These results show that the C-terminus of type 1 EBNA-2 protein confers the ability to give sufficient expression levels of those genes that are essential for EBV-LCL proliferation ( LMP-1 , CXCR7 ) . In summary , the data indicate that the mechanism that accounts for the superior ability of type 1 EBNA-2 to promote B cell growth , compared to the type 2 , is higher induction of LMP-1 and CXCR7 mediated by the C-terminal part of the protein . Although the differential effects on LMP-1 and CXCR7 would be sufficient to cause the reduced transformation by type 2 EBV , we have not shown that these are the only genes involved . Attempts to complement the deficiency in transformation by co-transfecting LMP-1 and CXCR7 expression plasmids in the EREB2 . 5 growth assay were not successful ( data not shown ) . However , the technical difficulty of maintaining the correct level of all the plasmids in the cells ( EBV , p554-4 , LMP-1 and CXCR7 expression vectors ) may well account for that result . The TAD of EBNA-2 is a strong transcriptional activator and is essential for EBNA-2-mediated activation of viral and cell genes and B cell transformation , since viruses lacking this domain are immortalization-incompetent [38] , [65] . The TAD mediates transcriptional activation at EBNA-2-target promoters by recruiting histone acetyltransferases [54] and several host transcription factors [51]–[53] . The RG sequence is involved in protein-protein and protein-nucleic acid interactions and is important for efficient B cell growth transformation [44] . The RG element down-regulates EBNA-2 activation of the LMP-1 promoter but not of the Cp [44] . The CR7 has been shown to be dispensable for B cell growth transformation , although is important for transcriptional cooperation with EBNA-LP [44] . In addition to the CR7 , the RG and TAD domains have also been shown to be involved in EBNA-2/EBNA-LP transcriptional cooperation . In fact , the TAD of EBNA-2 has been demonstrated to be required for EBNA-LP co-activation with EBNA-2 in the context of LMP-1-promoter reporter assays and represents a specific binding site for EBNA-LP , as demonstrated by in vitro binding assays [31] . Moreover , the RG and CR7 elements are able to down-regulate the high intrinsic transcriptional activity of EBNA-2 TAD in the context of Gal4 DNA-binding fusions , suggesting a modulatory activity of these domains on EBNA-LP function [37] . EBNA-LP has been shown to enhance EBNA-2-mediated transcriptional activation of both viral and cell promoters ( LMP-1/LMP-2B , Cp and cyclin D2 ) [27]–[31]; the mechanism of cooperative function is not entirely clear but has been proposed to involve interaction with histone deacetylase 4 [110] . EBNA-2 and EBNA-LP proteins do not significantly associate in lymphoblasts , as demonstrated by co-immunoprecipitation and immunofluorescence assays in previous studies [30] , [31] and it has therefore been suggested that EBNA-LP co-activation with EBNA-2 takes place only when transient unstable interactions are established , which would allow recruitment of positive transcriptional regulators to the EBNA-2 TAD at EBNA-2-specific promoters [31] . Our mapping studies suggest that the difference in LMP-1/CXCR7 activation and therefore in growth phenotype observed between type 1 and type 2 EBNA-2 is determined by different mechanisms of transcriptional activation , mediated by the TAD of EBNA-2 and modulated by cooperation with EBNA-LP , through the RG , CR7 and TAD EBNA-2 elements . Our results are consistent with the crucial roles played by the EBNA-2 TAD-mediated transcriptional activation and the EBNA-LP cooperative function during the EBV-driven B cell transformation process [28] , [29] , [38] . Ongoing additional detailed mutational analysis of the RG , CR7 and TAD sequences may identify the amino acids that are the key determinants of the differing abilities of type 1 and type 2 EBNA-2 to transactivate target genes , and therefore sustain B cell growth . At the biochemical level the differences between type 1 and type 2 transcriptional activation may be determined by differences in cell transcription factors or histone acetyltransferases recruited by type 1 EBNA-2 TAD , compared to the type 2 , resulting in higher induction of LMP-1 and CXCR7 . It is noteworthy in this context that EBNA-2 regulation of the LMP-1 promoter occurs mainly through other cell transcription factors such as PU . 1 , rather than RBP-Jk [46]–[49] , [54] , [63] . It will be interesting to determine in the future whether a similar situation exists at the CXCR7 promoter . We envisage the most likely mechanism for the differential regulation of LMP-1 and CXCR7 compared to other EBNA-2 target genes would be that their promoters may both require specific transcription factors that cooperate with EBNA-2 , transcription factors not required for most EBNA-2-responsive genes . We are currently investigating the CXCR7 promoter that is active in EBV-LCLs to test this hypothesis . Bearing in mind the importance of EBNA-LP for enhancement of the ability of EBNA-2 to activate genes essential for proliferation , such as LMP-1 [28] , [29] , we determined which EBNA-LP species are expressed in the EREB2 . 5 assay and detected a type 1 full-length EBNA-LP and a type 2 Y1Y2-truncated isoform ( Figure 6 ) . Even though the Y1Y2 region has been shown to have only a modulatory effect on EBNA-LP cooperative function [28] , [29] , one could speculate that the type 1 full-length EBNA-LP might cooperate more efficiently with the type 1 EBNA-2 than with the type 2 EBNA-2 , thereby determining the null growth phenotype of the latter EBNA-2 type in the assay . However , the experiment illustrated in Figure 7 clearly demonstrates that this is not the case , since even in the presence of a full-length type 2 EBNA-LP , the inability of type 2 EBNA-2 at maintaining cell proliferation was not complemented . Therefore the EREB2 . 5 growth assay functionally distinguishes between the ability of type 1 and type 2 EBNA-2 to sustain cell proliferation , which is likely to be enhanced by cooperation with EBNA-LP but is not influenced by the type of EBNA-LP present in the system . A role in protection from caspase-induced apoptosis has been recently proposed for the Y1Y2-truncated type 2 EBNA-LP species in BL cell lines [88] , however the possible contributions of this additional function of the EBNA-LP mutant to our EREB2 . 5 growth assay have not yet been explored . Likewise , in the Daudi and P3HR1 cl . 16 systems with oestrogen-inducible EBNA-2 proteins used to analyse the time-course induction of LMP-1 and CXCR7 by type 1 and type 2 EBNA-2 ( Figure 1 ) , several full-length type 1 EBNA-LP isoforms are expressed from the ER-EBNA-2-coding vectors , in addition to the type 2 Y1Y2-truncated species , encoded by the endogenous EBV genomes ( Figures S1 and S2 ) . Similarly to the EREB2 . 5 assay , it might be speculated that the absence of a full-length type 2 EBNA-LP could account for the weaker induction of LMP-1/CXCR7 by type 2 EBNA-2 . By using a functional assay for EBNA-LP cooperative function [17] , [29] based on the Daudi cell line , we demonstrated that this is not the case ( Figure 2 ) . Co-expression of type 2 EBNA-2 with either type 1 or type 2 full-length EBNA-LP in the Daudi cell line induced only a weak up-regulation of the endogenous LMP-1 gene , which was markedly lower than that observed when type 1 EBNA-2 was co-expressed with either type 1 or type 2 EBNA-LP . Therefore the weaker induction of LMP-1 by type 2 EBNA-2 is not affected by the EBNA-LP type . In all our assays for EBNA-LP cooperative function an EBNA-LP construct with 3 repeat W1W2 domains has been used , which is above the minimum ( 2 repeats ) required for optimal EBNA-LP function [17] , [29] , [60] . Our infection assays are similar to the EREB2 . 5 system and to the Daudi/P3HR1 cl . 16:ER-EBNA-2 systems in that both the type 1 and the type 2 EBNA-2 BAC viruses express full-length type 1 EBNA-LP species ( Figure S4 ) , containing at least 3 W1W2 repeats . Based on the results discussed before and illustrated in Figures 2 and 7 , showing that the weaker induction of LMP-1 in Daudi and the inability to maintain long-term growth in the EREB2 . 5 assay by type 2 EBNA-2 is not affected by the type of EBNA-LP present , we can infer that the gene-expression pattern ( Figure 3 ) and the growth phenotype ( Figure S3 ) induced by type 2 EBNA-2-expressing BAC virus in the infection experiments is not affected by the absence of a type 2 EBNA-LP . Moreover , the low-efficiency transformation phenotype produced by our type 2 EBNA-2 BAC recombinant EBV is similar to that reported for the wild-type type 2 AG876 EBV strain virus [19] . Our finding that type 2 EBNA-2 cannot rescue proliferation of oestrogen-starved EREB2 . 5 cells when expressed in trans in the EREB2 . 5 assay ( Figure 4 B and C ) is somehow surprising , if one considers that a type 2 EBV virus is able to eventually establish LCLs upon infection of primary B cells , albeit with very low efficiency ( Figures 3 and S3 and [19] ) . This difference might be either ascribed to the truncated EBNA-LP protein present in the EREB2 . 5 assay , or , more likely , to additional factors linked to the virus/B cell interaction occurring during the process of infection . These events occurring during the early stages of infection might act as compensatory mechanisms for the poor ability of type 2 EBNA-2 to induce B cell proliferation . LCL outgrowth following initial EBV infection of naïve resting B cells is a stochastic event whereby a few cells with the correct levels of expression of those genes required for B cell proliferation ( e . g . CXCR7 , LMP-1 ) are selected and clonally expanded . It seems likely that slower and lower expression of LMP-1/CXCR7 induced by type 2 EBNA-2-expressing virus during the initial stages of infection accounts for the low transforming efficiency displayed by this viral type . However , once a few cells infected with type 2 EBV eventually express the correct levels of LMP-1/CXR7 , these cells are selected and can be amplified to give rise to an LCL . In fact , in both type 1 and type 2 established LCLs , LMP-1 and CXCR7 expression levels were ultimately approximately similar ( Figure 3 B and C and [26] ) . In this study we have identified a mechanism that accounts for a very remarkable biological difference between type 1 and type 2 EBV , namely the poorer in vitro transforming ability of type 2 strains of EBV , compared to type 1 . This mechanism consists of lower and delayed induction of both viral ( LMP-1 ) and host ( CXCR7 ) genes which are essential for continuous proliferation of EBV-infected LCLs and is mediated by the C-terminal region of the EBNA-2 protein .
P3HR1 clone 16 , Daudi and Raji are EBV-positive BL cell lines [111]-[113] . B cell lines were maintained in RPMI 1640 medium ( Gibco-BRL ) supplemented with 10 to 20% ( v/v ) heat-inactivated foetal bovine serum ( FBS , BioWhittaker ) and antibiotics ( 100 units/ml penicillin and 100 units/ml streptomycin , Gibco-BRL ) . EREB2 . 5 cells [86] contain a conditional EBNA-2 protein regulated by oestrogen and were maintained in RPMI 1640 medium supplemented with 10% FBS , antibiotics and 1 µM β-oestradiol ( Sigma ) . Daudi:ER-EBNA-2 T1/T2 [26] and P3HR1 cl . 16:ER-EBNA-2 T1/T2 stable cell lines were grown in RPMI 1640 medium supplemented with 10% FBS , antibiotics and 400 or 500 µg/mL G418 ( Calbiochem ) respectively . 293 and HeLa cell lines were grown in Dulbecco Modified Eagle Medium ( DMEM , Gibco-BRL ) supplemented with 10% FBS and antibiotics . 293 cell lines stably transfected with EBV-BAC constructs were maintained in DMEM with 10% FBS , antibiotics and 100 µg/ml Hygromycin B ( Roche ) . The OriP-p294 plasmids used for expression of EBNA-2 , type 1 and type 2 , and the EBNA-2 chimaeras , were derived from the CMVpEBNA-1 plasmid described in [114] . The EBNA-1 coding region of the CMVpEBNA-1 plasmid was extracted and replaced with a multiple-cloning site ( MCS ) comprising HindIII , BamHI and NotI restriction sites , generating the OriP-p294 MCS plasmid . To generate the wild-type type 1 EBNA-2-expressing plasmid , the region comprising one BamHI W repeat and the type 1 EBNA-2 coding region from the p554 plasmid ( kind gift from Bettina Kempkes , described in [86] , [115] ) was cloned as a BglII – NotI fragment into OriP-p294 MCS , between BamHI and NotI sites . To generate the homologous vector expressing a type 2 EBNA-2 , the fragment encompassing one BamHI W repeat and the type 2 EBNA-2 sequence from the pAG1 vector ( described in [26]; this is essentially homologous to p554 , but carries a type 2 EBNA-2 ) was extracted by BglII – NotI digestion and cloned between BamHI and NotI sites in the OriP-p294 MCS vector . This procedure produced the two OriP-p294 plasmids with either type 1 or type 2 EBNA-2 sequences ( OriP-p294 E2T1/T2 ) . The expression of the EBNA-2 proteins from these plasmids is driven by the EBV Wp promoter . Additional features of these vectors include the EBV origin of replication , OriP , and a hygromycin-resistance gene . To generate the EBNA-2 chimaeric sequences , two pBluescript plasmids carrying type 1 or type 2 EBNA-2 coding sequences between EcoRI and NotI sites , extracted from the p554 and pAG1 plasmids respectively [26] , [86] , were used . For chimaeras 1 and 2 , a fragment spanning between two BstXI sites ( one in the EBNA-2 ORF , mapping to the WWPP amino acid motif , codons 323-326 , within the RBP-Jk domain , and one in the vector backbone , downstream NotI site ) was swapped between the two constructs . To generate chimaera 3 , the region comprising the CR7 , TAD and NLS elements from type 1 EBNA-2 was Pfu-PCR amplified using a primer with overhanging SacII ( 5′ ) restriction site and the universal T3 primer . A SacII restriction site was artificially inserted by site-directed mutagenesis , using the QuickChange Site-directed mutagenesis kit ( Stratagene ) , in type 2 EBNA-2 sequence , without altering the amino acid sequence ( codons 337-338 , between the RG sequence and the CR7 ) . The type 1 PCR amplicon was then cloned as a SacII – NotI fragment at homologous positions in the type 2 EBNA-2 vector . A similar strategy was used to generate chimaera 4 . Briefly , a PCR amplicon , consisting of the TAD and the NLS sequences of type 1 EBNA-2 , was generated by Pfu-DNA polymerase PCR using an oligonucleotide carrying an overhanging EcoRV ( 5′ ) restriction site and the universal primer T3 . This PCR product was then cloned as an EcoRV – NotI fragment into the pBluescript plasmid carrying type 2 EBNA-2 ORF , in which a silent EcoRV site had been previously introduced by in vitro mutagenesis at codons 388-389 , between the CR7 and the TAD . To produce chimaera 5 , the RG sequence from type 1 EBNA-2 was amplified by PCR with Pfu DNA polymerase using primers with overhanging BstXI ( 5′ ) and SacII ( 3′ ) sites and cloned between the same restriction sites into the type 2 EBNA-2 pBluescript plasmid , carrying the artificial SacII site . Likewise , to make the chimaera 6 construct , the type 1 sequences encompassing the RG and the CR7 domains were Pfu-PCR amplified with primers bearing overhanging BstXI ( 5′ ) and EcoRV ( 3′ ) restriction sites , the PCR product was BstXI – EcoRV digested and then cloned between the same positions into the type 2 EBNA-2 pBluescript vector with the engineered EcoRV sequence . For construction of chimaera 7 , an artificial BstBI site was inserted in the type 2 EBNA-2 pBluescript plasmid , at the end of the TAD sequence ( codons 429-430 ) . The RG , CR7 and TAD sequences were amplified by Pfu-PCR from the type 1 EBNA-2 vector , using primers with overhanging BstXI ( 5′ ) and BstBI ( 3′ ) sites . The PCR product was then cloned into the type 2 EBNA-2 plasmid , between the same restriction sites . The identity of all the chimaeric EBNA-2 sequences was screened by restriction mapping and by sequencing ( data not shown ) . These were subcloned as EcoRI – NotI fragments from the pBluescript background into an intermediate vector ( pSuper backbone ) bearing the BglII – NotI region from the pAG1 vector ( described above ) , which consists of one BamHI W repeat and the EcoRI – NotI EBNA-2 coding sequence . The BglII – NotI sequences , comprising the BamHI W repeat and the chimaeric EBNA-2 ORFs , were then cloned into the OriP-p294 MCS expression vector , between BamHI and NotI sites . Transcription of the chimaeric EBNA-2 ORFs in these plasmids is driven by the EBV Wp promoter , similarly to the type 1 and type 2 EBNA-2 constructs . The pSNOC vector expressing a type 1 3-repeat EBNA-LP coding sequence is described elsewhere [116] . Briefly , this OriP expression plasmid contains the G418-resistance locus and the CMV immediate early promoter and the SV-40 poly A sequences , expressing the EBNA-LP coding sequence . To generate a homologous type 2 3-repeat EBNA-LP-expressing plasmid , a cDNA clone from the type 2 cell line C2+BL16 was obtained by Pfu-PCR using oligonucleotides 5′- ATG CGG CCA TGT AGG CCC ACT T -3′ and 5′- TGC CCA ACC ACA GGT TCA GGC A-3′ , complementary to sequences within the C1 and YH exons , respectively , of AG876 EBV strain ( coordinates 11403-11425 and 36141-36119 ) . The amplified product was subcloned into pCR2 . 1-TOPO vector ( Invitrogen ) via TA cloning strategy and sequenced to verify its identity to the published AG876 EBNA-LP sequence . The type 2 3-repeat EBNA-LP cDNA was then cloned as an EcoRI fragment into the pSNOC empty vector . The p554-4 plasmid , bearing the coding sequences for a type 1 ER-EBNA-2 and 2 BamHI W repeats was a kind gift from Bettina Kempkes [86] , [115] . The p554-4 derivative expressing type 2 ER-EBNA-2 is described elsewhere [26] . These two vectors were used to generate the stable cell lines P3HR1 cl . 16:ER-EBNA-2 T1/T2 ( described below ) . EBNA-2 in the type 1 B95-8 EBV-BAC [117] was substituted with the type 2 EBNA-2 gene by RecA-mediated homologous recombination . The final targeting construct used for this was cloned between the BamHI and Psp0MI sites of pKov-Kan-ΔCm p4487 . 1 [118] and was called p121 ( Figure S6 ) . Starting from the type 2 EBNA-2 pBluescript plasmid ( described above ) , type 1 upstream flanking sequences were added by cloning a HindIII – EcoRI fragment from the p554 vector [86] , [115] using a HindIII site in the Bluescript polylinker . Type 1 downstream flanking sequence was then added by substituting in a PmeI to NotI fragment from p554 . An XhoI site in the pBluescript polylinker upstream of the HindIII site was converted into a BglII site using an oligonucleotide adaptor and then the whole assembly was cloned as a BglII – NotI fragment between the BamHI and Psp0MI sites of p4487 . 1 . The resulting type 2 EBNA-2 with type 1 flanking sequences contained type 2 EBV from 36201 – 38008 of the AG876 sequence [119] . B95-8 EBV-BAC ( p2089 , chloramphenicol resistant ) in E . Coli strain DH10B was kindly provided by W . Hammerschmidt [117] . For clarity , the name T1 E2 EBV-BAC is used for this plasmid in this paper . RecA recombination to produce the type 2 EBNA-2 BAC mutant used the same methods as described previously for construction of EBER mutants [120] . The integrity of the type 2 BAC was screened at each stage of recombination by miniprep DNA isolation , restriction digest and pulsed field gel electrophoresis ( PFGE ) in order to control for any second-site mutations that may have occurred . Substitution of the type 1 EBNA-2 ORF with the type 2 sequence introduced an additional EcoRI restriction site into the EBV-BAC , in front of the starting ATG of EBNA-2 open reading frame . Digestion with this restriction enzyme followed by PFGE analysis therefore enabled diagnosis of correct insertion of the type 2 EBNA-2 sequence ( data not shown ) . EcoRI digestion revealed also that the type 2 EBNA-2 construct had lost 2 copies ( out of 6 ) of the BamHI W repeat compared to the parental type 1 construct . This was confirmed by digestion with another restriction enzyme , EcoRV ( data not shown ) and was the only gross rearrangement that had occurred , as the pattern of the remaining bands was identical in the type 1 and the type 2 EBNA-2 BAC in both EcoRI and EcoRV digestions ( data not shown ) . Because the number of BamHI W repeats varies considerably in different EBV strains , from 3 to 12 [94] and because the repeat number found in the type 2 BAC is within this range , this deletion was not considered to be undesirable . EBV-BAC DNA was purified using EndoFree plasmid Maxi kit ( Qiagen ) and 1 µg was transfected into 293 cells with Lipofectamine2000 ( Invitrogen ) according to the manufacturer's instructions . Stable 293 cell clones were isolated under Hygromycin B selection ( 100 µg/ml ) , clonal cell lines were expanded and these were then screened for integrity of the EBV-BAC DNA by episome rescue and restriction digestion or PCR analysis ( data not shown ) . A 293 clone containing the T1 E2 EBV-BAC , used to generate the type 2 BAC , was also produced . Infectious virus was produced from the 293 stable cell lines containing either T1 E2 EBV-BAC ( with type 1 EBNA-2 ) or T2 E2 EBV-BAC ( with type 2 EBNA-2 ) by transient transfection of BZLF1 and BALF4-expressing plasmids using Lipofectamine2000 ( Invitrogen ) [117] , [121] , [122] . After 4 days , the virus-containing supernatants were harvested , filtered ( 0 . 45 µm pore size ) and titred by infecting Raji cells and counting GFP-positive cells ( Green Raji Units , GRU ) on an inverted fluorescent microscope , following treatment with 5 nM TPA ( Sigma ) and 5 mM sodium butyrate ( Sigma ) , to enhance GFP expression . Primary B cells were isolated from mixed-donors buffy coat residues by negative selection using RosetteSEP Human B cell Enrichment cocktail ( Stemcell Technologies ) as described by the manufacturer's instructions . The isolated cells were analyzed by flow cytometry for purity using fluorescein-isothiocyanate-conjugated anti-CD20 antibody ( Dako ) . 2×106 cells were infected with 5000 GRUs of recombinant viruses and cultured in RPMI 1640 medium supplemented with 20% FBS and antibiotics . At the desired time-points cells were harvested and processed for RNA or protein extraction . For LCL establishment , 106 cells were infected with 5-fold serial dilutions of recombinant EBVs starting with 8000 GRUs . Cells were grown initially in RPMI 1640 medium with 20% FBS and antibiotics and every 2-3 days half the medium was replaced . Once LCLs had grown out in large culture volumes , the FBS level in the medium was reduced to 10% and cells were selected in Hygromycin B ( Roche ) at a concentration of 150 µg/ml for 3 weeks . After this time , hygromycin selection was discontinued and cells were not used for experimental investigation during hygromycin selection . To generate P3HR1 cl . 16:ER-EBNA-2 T1/T2 stable cell lines , 2×106 cells were transfected with 4 µg of p554-4 [86] , [115] or pERT2 [26] plasmids expressing ER-EBNA-2 type 1 and type 2 respectively using Neon transfection system ( Invitrogen ) set at 1300 V , 30 pulse width , 1 pulse . 24 hours after transfection G418 ( Calbiochem ) selection was added at a concentration of 500 µg/ml . Selected cell lines were grown out and tested for expression of ER-EBNA-2 T1/T2 by western blotting . For EBNA-2/EBNA-LP cooperation assays in Daudi cells , the Neon system ( Invitrogen ) was used to transiently transfect 2×106 cells with 4 µg of the OriP-p294 plasmids expressing type 1 or type 2 EBNA-2 and 2 . 4 µg DNA of the pSNOC plasmids expressing either type 1 or type 2 EBNA-LP . Empty vector ( OriP-p294 or pSNOC ) was added to normalize the total amount of DNA transfected ( 6 . 2 µg ) . Electroporation conditions were 1400 V , 30 pulse width and 1 pulse , which routinely gave approximately 40% transfection efficiency ( assessed as GFP-positive cells when transfecting a GFP-expressing plasmid ) and 80% of cell viability . For the EREB2 . 5 growth assay , 6×106 EREB2 . 5 cells were transfected with 5 µg of OriP-p294 plasmids bearing type 1/type 2 EBNA-2 sequences , either wild-type or chimaeric , using the Amaxa system ( Lonza ) , program A-23 . β-oestradiol was withdrawn immediately after transfection in order to select for β-oestradiol – independent lymphoblastoid cell lines . Cells were incubated overnight in 2 ml of complete medium supplemented with 20% FBS and antibiotics , but without β-oestradiol , in 12-well plates . The following day each transfected sample was diluted up to 10 ml with culture medium and plated into 5 wells of a 24-well plate . At the desired time-points cell samples were harvested for cell count analysis and RNA or protein extraction . For immunofluorescence experiments , 5×106 HeLa cells , grown on coverslips ( VWR ) in 6-well dishes , were transfected with 4 µg of OriP-p294 plasmids encoding either type 1 , type 2 EBNA-2 or the chimaeric EBNA-2 proteins using Lipofectamine2000 ( Invitrogen ) as recommended by the manufacturer's instructions . Total cell RNA was extracted from growing cells using Trizol reagent ( Invitrogen ) and then quantified by measuring its absorbance at 260 nm . RNA samples were digested with RQ1 RNase-free DNase ( Promega ) . cDNA was prepared using Protoscript M-Mulv First strand cDNA Synthesis kit ( New England Biolabs ) using random primer as recommended by the manufacturer's instructions . To detect the expression of the EBNA-2-regulated genes , quantitative RT-PCR ( qRT-PCR ) was used with the primer pairs listed in Table S2 . qRT-PCR was performed on an ABI 7900HT real-time PCR machine using Absolute QPCR SYBR Green ROX Mix ( Thermo Scientific ) . The cycling conditions were: 95°C for 15 min , followed by 40 cycles of 15 sec at 95°C , 30 sec at 60°C and 40 sec at 72°C on a fast block . Dissociation curve analysis was performed at the end of each run to ensure absence of non-specific products . Quantification of mRNA levels was carried out using the Delta-Delta Ct method . Results were analyzed with SDS2 . 3 software ( Applied Biosystems ) . mRNA levels for each target gene analyzed were normalized on the housekeeping gene GAPDH used as a loading control . The LMP-1-specific probe for use in RPA was generated by PCR of genomic DNA derived from Daudi cells using the following primers , designed based on the EBV genome sequence: forward 5′-CAC TCA TAA CGA TGT ACA GC-3′ ( coordinates 168881 to 169000 in EBV wild-type , accession number NC007605 ) and reverse 5′- CAA GAA ACA CGC GTT ACT C-3′ ( coordinates 169103 to 169121 in EBV wild-type , accession number NC007605 ) . The PCR product ( 241 bp ) was cloned into pCR2 . 1-TOPO vector and sequenced and the plasmid was then linearized by BamHI digestion at the 5′ end of the PCR product . 1 µg of the linearized plasmid was in vitro-transcribed with T7 enzyme using the MAXIscript SP6/T7 in vitro transcription kit ( Ambion ) , according to the manufacturer's instructions , for production of 32P-labelled antisense RNA probe . RPAs were carried out using RPAIII kit ( Ambion ) as recommended by the manufacturer's instructions . Briefly , 20 µg of total cell RNA were hybridized overnight at 42°C with 50000 cpm of the probe . As negative control , an equivalent amount of yeast RNA was included in a hybridization reaction . Single-stranded RNA was digested with an RNase A/T1 mixture for 30 min at 37°C . Protected fragments were precipitated and separated on a polyacrylamide gel and the gel was analyzed on a phosphoimager . Cell samples were lysed for 15 min on ice in 2 volumes of RIPA lysis buffer ( 150 mM NaCl , 1% v/v Nonidet-P40 , 0 . 5% v/v deoxycholic acid , 0 . 1% w/v SDS , 50 mM Tris/HCl , pH 8 . 0 ) supplemented with 1 mM PMSF ( Fluka ) and Complete protease inhibitors ( Roche ) . Samples were centrifuged at 16000 g for 15 min , to pellet cell debris , and the supernatant was harvested . Proteins were fractionated by SDS-PAGE and transferred onto a nitrocellulose membrane . After blocking for 1 h at room-temperature with 5% milk powder in PBS/0 . 1% Tween-20 ( Sigma ) , membranes were incubated overnight at 4°C with the primary antibody diluted in blocking buffer . The primary antibodies used are listed in Table S1 . The secondary antibodies were horseradish peroxidise-conjugated anti-mouse ( GE Healthcare ) or anti-rabbit or anti-human IgG ( Dako ) and were used at a dilution of 1/2000 for 1 h at room-temperature . Bound immunocomplexes were detected by using ECL western blotting detection reagents ( Amersham ) . 24 hours after transfection , HeLa cells were fixed using 4% paraformaldehyde and permeabilized with 0 . 25% Triton-X 100 ( Sigma ) in PBS . Primary anti-EBNA-2 antibody ( PE2 clone ) was diluted 1/100 in blocking buffer ( 1% bovine serum albumin in PBS ) and cells were incubated for 45 min at room-temperature . The coverslips were washed 3 times with PBS and TRITC-conjugated anti-mouse secondary antibody ( Sigma ) was applied at a dilution of 1/5000 in blocking buffer for 45 min at room-temperature . After 3 washes in PBS the coverslips were mounted in Mowiol and visualized using a Zeiss 510 Meta laser confocal microscope . Database entries for genes mentioned in this manuscript include: CXCR7 ( RefSeq NM 020311 . 2 ) , EBV type 1 ( accession number AJ507799 ) and EBV type 2 ( accession number DQ279927 ) . Annotation in the EBV genome shows EBNA-2 and LMP-1 genes .
|
Epstein-Barr virus ( EBV ) is a common human virus that is involved in several types of cancer and directly causes human B lymphocytes to proliferate when they become infected . EBV occurs naturally as two different viral types ( type 1 and type 2 ) . The genomes of these viruses are mostly very similar but they differ in a few genes , particularly the EBNA-2 gene . For many years it has been known that type 1 EBV is much more effective than type 2 EBV at causing B lymphocyte proliferation and this difference is mediated by the EBNA-2 gene . Here we have shown that the greater ability of type 1 EBNA-2 to cause B cell proliferation is due to superior induction of the EBV LMP-1 and the cell CXCR7 genes , both of which are required for growth of EBV-infected lymphocytes . We mapped the section of type 1 EBNA-2 responsible for this to the C-terminus of the protein , including the transactivation and EBNA-LP interaction domains . The results provide a mechanism for the long-standing question of the functional difference between these two major types of EBV and will be important in understanding the significance of the EBV types in human infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology"
] |
2011
|
C-Terminal Region of EBNA-2 Determines the Superior Transforming Ability of Type 1 Epstein-Barr Virus by Enhanced Gene Regulation of LMP-1 and CXCR7
|
We hypothesized that variants in genes expressed as a consequence of interactions between ovarian cancer cells and the host micro-environment could contribute to cancer susceptibility . We therefore used a two-stage approach to evaluate common single nucleotide polymorphisms ( SNPs ) in 173 genes involved in stromal epithelial interactions in the Ovarian Cancer Association Consortium ( OCAC ) . In the discovery stage , cases with epithelial ovarian cancer ( n = 675 ) and controls ( n = 1 , 162 ) were genotyped at 1 , 536 SNPs using an Illumina GoldenGate assay . Based on Positive Predictive Value estimates , three SNPs—PODXL rs1013368 , ITGA6 rs13027811 , and MMP3 rs522616—were selected for replication using TaqMan genotyping in up to 3 , 059 serous invasive cases and 8 , 905 controls from 16 OCAC case-control studies . An additional 18 SNPs with Pper-allele<0 . 05 in the discovery stage were selected for replication in a subset of five OCAC studies ( n = 1 , 233 serous invasive cases; n = 3 , 364 controls ) . The discovery stage associations in PODXL , ITGA6 , and MMP3 were attenuated in the larger replication set ( adj . Pper-allele≥0 . 5 ) . However genotypes at TERT rs7726159 were associated with ovarian cancer risk in the smaller , five-study replication study ( Pper-allele = 0 . 03 ) . Combined analysis of the discovery and replication sets for this TERT SNP showed an increased risk of serous ovarian cancer among non-Hispanic whites [adj . ORper-allele 1 . 14 ( 1 . 04–1 . 24 ) p = 0 . 003] . Our study adds to the growing evidence that , like the 8q24 locus , the telomerase reverse transcriptase locus at 5p15 . 33 , is a general cancer susceptibility locus .
Ovarian cancer is the seventh leading cause of cancer mortality among women globally , accounting for 4 . 2% of cancer deaths [1] , due in part to the lack of practical screening methods and detectable symptoms in the early stages of tumor progression [2] . Although the aetiology of ovarian cancer has not been fully elucidated , it is generally agreed that family history of ovarian or breast cancer is the most important risk factor for epithelial ovarian cancer [3] . Hereditary ovarian cancer occurring in breast/ovarian cancer families has been linked to mutations in the BRCA1 and BRCA2 genes , while cases occurring in association with Lynch syndrome have been linked to mutations in MSH2 and MLH1 [4] , [5] . Given that only 3% to 5% of ovarian cancer cases present from high-risk families and residual family history associations [2] , it is likely that several low-penetrance genes with relatively common alleles that confer slightly increased risk may account for a portion of the risk of non-familial ovarian cancer . The Ovarian Cancer Association Consortium ( OCAC ) was established in 2005 to provide a forum for the identification and validation of common low-penetrance ovarian cancer susceptibility polymorphisms with increased power [6] . OCAC recently conducted a genome-wide association study ( GWAS ) and identified the first susceptibility locus associated with invasive ovarian cancer risk [7] . A number of hypotheses have been put forward to explain the pathogenesis of ovarian cancer [8] , [9] , including that of incessant ovulation which causes repeated minor trauma to the surface of the ovary , leading to proliferation of ovarian epithelium and repair of the ovulatory wound [10] . However , it has also been hypothesized that fallopian tube epithelial cells migrating to the ovulatory wound could serve as precursors to ovarian cancer [11] . Research in the past two decades compellingly suggests that the neighbors of cancer cells , collectively termed stroma , are not uninvolved bystanders [12] and studies involving three-dimensional cell culture models underscore the involvement of the extracellular matrix surrounding cancer cells in the signalling pathways that promote cell survival [13] . Fibroblasts with a carcinoma-promoting phenotype [carcinoma-associated fibroblasts ( CAFs ) ] residing in the breast cancer microenvironment lack the ability of normal fibroblasts to attenuate the growth of neighbouring transformed epithelial cells [14] . In addition , xenograft models have shown that CAFs accelerate cancer progression through their ability to secrete stromal cell-derived factor 1 [15] . Furthermore , expression profiling of ovarian tumor samples has identified a group of high-grade invasive cancers characterized by a reactive stromal gene expression signature and extensive desmoplasia , which confer an inherently poor prognosis [16] . If this CAF-dependent model of tumorigenesis is correct , it assigns a key role to the neighboring stroma in cancer initiation . We therefore hypothesized that subtle variation in the expression or function of genes expressed as a consequence of interactions between ovarian cancer cells and the host micro-environment could contribute to ovarian cancer susceptibility . We used a two-stage approach to comprehensively evaluate common variation in 173 genes selected for their putative role in stromal-epithelial interactions using a tagging-SNP approach and data from sixteen case-control studies participating in the Ovarian Cancer Association Consortium ( OCAC ) .
Candidate gene selection and justification are provided in Text S1 and Table S1 . Characteristics of all case-control studies that contributed data to discovery and replication analyses are provided in Table S2 . Comparison of the mean age at diagnosis for cases and age at interview for controls showed that cases were significantly older compared to controls ( p<0 . 05 ) . Figure S1 provides an overview of SNP and cases-controls numbers analysed in the discovery and replication stages of this study . Discovery samples consisted of serous invasive cases from the AUS ( 550 cases and 1 , 101 controls ) and MAY ( 125 cases and 61 controls; all non-Hispanic Whites ) studies . AUS participants were not selected for ethnicity , but comprised of predominantly non-Hispanic White women . Of the 1 , 837 women with genotype data , three were excluded by PLINK default thresholds because >10% of SNPs failed genotyping for these individuals . Of the 1 , 536 single nucleotide polymorphisms ( SNPs ) genotyped , 1 , 309 SNPs passed our initial quality control ( QC ) criteria , and of these , seven were excluded by PLINK default thresholds . The remaining 1 , 302 SNPs were subject to further pruning as follows: 37 SNPs with significantly different frequencies of missing genotype data between cases and controls ( PMiss<0 . 05 ) ; 296 SNPs with duplicate discordance and/or failure to meet Hardy-Weinberg equilibrium ( HWE ) criteria ( 0 . 001<PHWE<0 . 05 ) . Of the remaining 969 SNPs analysed in the discovery stage , 59 SNPs with PTrend<0 . 05 were considered for the replication study ( see Table S3 ) . Based on positive predictive value ( PPV ) estimates , the three SNPs selected for replication using TaqMan genotyping by the 16 OCAC studies were PODXL ( podocalyxin-like ) rs1013368 ( PPV 33 . 1% ) , ITGA6 ( integrin , alpha 6 ) rs13027811 ( PPV 4 . 5% ) and MMP3 ( matrix metallopeptidase 3 ) rs522616 ( PPV 4 . 4% ) ( Table 1 ) . These 16 OCAC studies included all histologic subtypes , and ethnicities . An additional 18 SNPs with PTrend<0 . 05 which fitted into the iPLEX design were selected for replication by a subset of five of the 16 OCAC studies [AUS ( additional samples not in the discovery set ) , MAL , SEA , UKO , and USC] . FGF2 rs17473132 included among the 18 selected SNPs ( PTrend = 0 . 008 ) has been previously reported elsewhere [17] and is therefore excluded from this report . Replication sample sizes varied by SNP depending on which participating OCAC study met QC criteria; MAY , NCO , NEC and NHS failed QC for PODXL rs1013368 , and GER and STA failed QC for ITGA6 rs13027811 . Table 2 provides the risk estimates adjusted for age and study site for SNPs included in the replication analysis . There was no evidence of between-study heterogeneity for any replication SNP with the exception of TERT rs7726159 ( p = 0 . 005 ) ( Table S4 ) . Further examination of the site-specific Odds Ratios ( ORs ) showed that this was driven in part by the smaller USC study , the exclusion of which resulted in a p-value for between-study heterogeneity of 0 . 09 . The associations observed in the discovery set for the three SNPs selected based on PPV values ( PODXL rs1013368 , ITGA6 rs13027811 , and MMP3 rs522616 ) , were completely attenuated in the larger replication analysis of 16 case control studies ( adj . Pper-allele≥0 . 5 ) ( Table 2 ) . However , adjusted log additive estimates for TERT ( telomerase reverse transcriptase ) rs7726159 retained a statistically significant p-value in the replication study of non-Hispanic White serous invasive cases and controls ( Pper-allele = 0 . 03 ) , and showed evidence of log additive effects across genotypes . We re-analysed this SNP combining discovery and replication data and observed some evidence of between-study heterogeneity ( p = 0 . 027 ) which again improved with the exclusion of the smaller studies ( USC and MAY; p = 0 . 16 ) . Risk estimates for serous invasive ovarian cancer adjusted for age and study site remained statistically significant in the combined dataset [adj . ORper-allele 1 . 14 ( 1 . 04–1 . 24 ) p = 0 . 003; Table 3] . Likewise , in exploratory analyses of genotype data on all ethnicities stratified by histological subtype , a increased risk associated with this SNP was observed for serous invasive cases in models adjusted for age , site and ethnicity [adj . ORper-allele 1 . 17 ( 1 . 08–1 . 27 ) p = 7 . 21×10−5] . TERT rs7726159 was also associated with serous borderline tumors , but not with any other invasive or borderline subtypes ( Table 4 , and Figure 1 ) . For MMP7 rs17098236 , the combined age- and site-adjusted estimate from the log additive model suggested an association with serous ovarian cancer but the point estimates were not in the same direction as those obtained in discovery analysis ( 0 . 84 vs . 1 . 19; see Table S3 and Table 2 ) . All other SNPs in the smaller replication study failed to replicate the significant associations observed in the discovery sample .
Herein we report a large-scale analysis of 1 , 309 SNPs in 173 genes selected for their putative role in stromal epithelial cross talk , using a two-stage design for assessment of ovarian cancer risk . In the discovery stage we used data from two OCAC case-control studies ( AUS and MAY ) of predominantly non-Hispanic White women , and observed that SNPs in several genes were associated with risk of serous tumours in unadjusted log-additive models ( Table S3 ) . The most significant associations observed ( PODXL rs1013368 , ITGA6 rs13027811 , and MMP3 rs522616; PTrend≤0 . 001; Table 1 ) were then genotyped in a total of sixteen OCAC studies including additional samples from discovery studies ( AUS and MAY ) , and also from non-serous histologies and all ethnicities . None of these three SNPs were significantly associated with ovarian cancer risk ( Pper-allele≥0 . 5 ) . The power of the replication sample to detect the odds ratios observed in the discovery set at a type 1 error rate of 0 . 05 assuming log additive effects was >99 . 9% for all three SNPs . Combining discovery and replication data would have provided greater power to detect a significant effect [18] , but this was not considered for these SNPs because estimates were unequivocally null in replication analysis and/or in the opposite direction compared to the smaller discovery dataset . We analysed an additional 18 SNPs , including one in FGF2 reported elsewhere [17] in a second smaller replication study using five case-control studies from OCAC , and found evidence of an allelic association between TERT rs7726159 and serous tumors ( Table 2 ) . Although the PPV for TERT rs7726159 was 1 . 4% , it was not selected for the larger replication stage in all sixteen OCAC case-control studies because of limited resources . Our estimate from the replication study , adjusted for age and study site , showed an overall 12% increased risk of serous ovarian cancer associated with each minor allele among non-Hispanic Whites . Site-specific estimates were also statistically significant in case-control studies with the largest samples sizes ( SEA , AUS and MAL ) ( Table 3 ) . We detected significant study heterogeneity in this combined sample of all studies ( p = 0 . 027 ) , and this effect was attenuated when the smallest sample sizes ( USC and MAY ) were removed from the dataset; p = 0 . 16 ) . Inclusion of data on all ethnicities additionally adjusted for race resulted in a significance level ( adj . Pper-allele = 7 . 21×10−5 ) that met the conservative Bonferroni adjustment for multiple testing ( 0 . 05/21 = adj . Pper-allele≤0 . 0024 ) . In addition , the estimates from log-additive models for TERT rs7726159 in the combined discovery and replication non-Hispanic White samples would almost meet Bonferroni adjustment ( adj . Pper-allele = 0 . 003 ) . TERT encodes the catalytic subunit of telomerase and activation of telomerase has been implicated in human cell immortalization and cancer cell pathogenesis . TERT was selected as a candidate gene because it serves as an epithelial stem cell marker [19] and we hypothesized that cross-talk modifies critical aspects of epithelial transformation . TERT is a ribonucleoprotein enzyme that maintains telomere ends , and is essential for the replication of chromosomes and suppression of cell senescence . Telomere dysfunction is associated with genomic instability and consequently increased risk of tumor formation [20] . The rs7726159 variant resides in intron 3 of TERT and has no obvious functional significance , but it could be in linkage disequilibrium with another functional or causal SNP within the gene . An alternative explanation for the observed association is population stratification , which occurs when allele frequencies differ with population subgroups , or when cases and controls are drawn from different subgroups . We suggest that this is not a likely explanation because cases and controls were drawn from the same source populations within each study , and replication analyses were restricted to non-Hispanic White women or adjusted for ethnicity where applicable . However , it is possible that the association with serous ovarian cancer may vary across populations because of interaction with other genes or environmental factors , and additional studies would be required to confirm these findings . Although TERT variants have not been previously reported to be associated with ovarian cancer , a recent meta-analysis of two GWAS identified another SNP in TERT , rs2736100 , as significantly associated with gliomas ( OR = 1 . 27; P = 1 . 50×10−17 ) [21] . GWAS have found that rs2736100 is also associated with lung cancer ( OR = 1 . 14; P = 4×10−6 ) [22] and more specifically , with the adenocarcinoma subtype ( OR = 1 . 23; P = 3 . 02×10−7 ) [23] ( Figure 2A ) . Associations have also been reported between the TERT- CLPTM1L ( cleft lip and palate transmembrane 1-like gene - cisplatin resistance-related protein 9- ) locus and lung cancer ( rs402710; OR = 1 . 17; P = 2×10−7 ) [22] , basal cell carcinoma ( rs401681; OR = 1 . 20; P = 4 . 8×10−9 ) [24] , pancreatic cancer ( rs401681; OR 1 . 19; ( P = 3 . 66×10−7 ) [25] , and multiple cancer types that are known to originate in the epithelium , including bladder , prostate and cervical cancer [26] . We genotyped rs2736100 in the discovery samples and found a borderline , but inverse , association with serous ovarian cancer [OR = 0 . 88 ( 0 . 77–1 . 01 ) PTrend = 0 . 06] . We also found a borderline association with rs11133719 and serous ovarian cancer risk [OR = 0 . 81 ( 0 . 67–0 . 98 ) PTrend = 0 . 025] in discovery samples . Linkage disequilibrium ( LD ) estimation between the 11 TERT SNPs that we genotyped in stage 1 in 1 , 047 non-Hispanic White controls showed a moderate pairwise correlation between rs2736100 and rs7726159 ( r2 = 0 . 43; Figure 2B ) but rs7726159 , which we selected from NIEHS , is not in HapMap and so has not been genotyped in GWAS of ovarian or other cancers . Further analysis of this locus is necessary in order to definitively identify the causal SNP ( s ) at this locus . To our knowledge , this is the first comprehensive evaluation of genes involved in stromal epithelial cross-talk and serous ovarian cancer . Candidate gene and SNP selection for discovery stage analysis was aimed at optimizing the likelihood of detecting a signal by including tagging and putatively functional SNPs with minor allele frequency ( MAF ) >5% . Although a tagSNP approach has been shown to improve the power of the study for common variants [27] , modest effects from SNPs with low MAFs may remain undetected . This was illustrated in a recent re-analysis of two SNPs in the DCN gene that failed to achieve the minimal PTrend≤0 . 05 in stage 1 analysis , but conferred a small but significantly decreased risk of serous ovarian cancer in a combined analysis of data from two additional studies [28] . We therefore suggest caution in interpreting null findings , and the need for large discovery and replication studies . Our discovery study was reasonably well powered , so the failure to find any associations with SNPs in genes involved in stromal epithelial cross-talk , except in DCN and TERT , suggests that genetic variation in this pathway is not a major determinant of serous ovarian cancer risk . In summary , we have identified an association between TERT rs7726159 and serous ovarian cancer in a large sample of non-Hispanic White women participating in five OCAC case-control studies . We plan to further our investigation of this SNP and others in linkage disequilibrium with it , to determine whether TERT , CLPTM1L or another gene in the region is the functional target of this association . Our study adds to the growing evidence that , as well as the 8q24 locus [21] , [29] , [30]–[32] , the TERT-CLPTM1L locus at 5p15 . 33 , is a general cancer susceptibility locus . This is particularly interesting given the key roles of c-MYC ( the nearest gene to the 8q24 locus ) and TERT in tumorigenesis . TERT and MYC are both expressed in normal and transformed proliferating cells , and can induce immortalization when constitutively expressed [33] . The TERT promoter contains numerous MYC binding sites that mediate TERT transcriptional activation [34] , suggesting that TERT is a target of MYC activity . Although TERT variants have not been previously reported to be associated with ovarian cancer , multiple genome-wide association studies have reported associations with this locus and risk of other cancers . Further analyses of this locus , including fine mapping , resequencing and functional assays , will be necessary to definitively identify the causal SNP ( s ) .
Approval from respective human research ethics committees was obtained , and all participants provided written informed consent . Sixteen OCAC case-control studies ( summarized in Table S2 ) contributed data to this two-stage risk analysis . Samples in the discovery stage were derived from two case-control studies , AUS ( 550 cases and 1 , 101 controls ) and MAY ( 125 cases and 61 controls ) . Cases in the discovery set were all diagnosed with serous carcinoma of the ovary , fallopian tube or peritoneum , and most of the participants were non-Hispanic white women . Cases and controls from an additional 14 OCAC studies , as well as an additional 284 AUS and 477 MAY samples , including cases with other histologies , were included in a stage 2 analysis designed to replicate the most promising SNPs from the discovery stage . Fifteen studies used population-based case and control ascertainment , and one ( MAY ) was clinic-based . All studies have been previously described [7] , [35] , [36] . The final combined dataset of all discovery and replication samples consisted of a total of 10 , 067 controls ( 9 , 953 were self-classified as non-Hispanic White ) and 5 , 976 ovarian cancer cases of all histologies and morphologies , including 3 , 734 serous invasive cases ( 3 , 710 were self-classified as non-Hispanic Whites ) ( Table S2 ) . Our approach and our choice of candidate genes was based on extensive preliminary data we have accumulated from gene expression profiles of co-cultured of theca fibroblast and epithelial ovarian cells ( I . Haviv , personal communication ) , and expression profiles of murine ovarian epithelial cells identifying candidates that are regulated through the estrus cycle [37] , [38] ( see Text S1 ) . A compiled list of candidates was uploaded on the Ingenuity Pathway Analysis web interface and GeneSpring GX in order to obtain further candidates inferred from the literature . Prioritisation based on literature evidence for a plausible role in oncogenesis resulted in a list of 255 candidate genes of interest including CXCL9 , CTGF , LCN2 , DCN , and VIL2 . CXCL9 is associated with ovarian cancer survival and acts by recruiting T-cells and inducing immune surveillance [39] , and is expressed in epithelial cells co-cultured with fibroblasts . CTGF is likely to be the driver of the CAF phenotype . CTGF ( TGFβ-stimulated ) expression is associated with desmoplastic stroma [40] and elevated angiogenesis [41] . LCN2 , DCN and VIL2 were regulated through the murine estrus cycle , and appear to be hormone responsive ( either directly or indirectly ) [37] . Furthermore , comparison with expression profiles of human ovarian carcinomas [42] , [43] showed that all three are differentially expressed in tumors compared with normal epithelial cells . Further details for candidate gene selection and justification are provided in Text S1 and Table S1 . We identified SNPs within 5 kb of these 255 genes ( 58 , 114 SNPs in total from dbSNP , Ensembl , the International HapMap Consortium [44] , Perlegen Sciences [45] , SeattleSNPs [pga . mbt . washington . edu/] , NIEHS SNPs [http://egp . gs . washington . edu] , and the Innate Immunity PGA [http://www . nhlbi . nih . gov/resources/pga/] . We used the binning algorithm of ldSelect [46] to identify 4 , 567 tagSNPs among these ( r2>0 . 8 ) and minor allele frequencies ( MAFs ) >0 . 05 based on the most informative available source ( 84% of genes used HapMap , 10% used SeattleSNPs , 3% used Perlegen Sciences , 2% used NIEHS SNPs , and 1% used Innate Immunity PGA ) . We prioritized the list to 166 genes based on known function and the number of bins in each gene ( excluding genes with a large number of bins ) , in an attempt to identify ∼1 , 500 key SNPs . Based on Illumina design scores , we picked the best tagSNP in each bin ( or two tagSNPs , if there were >10 tagSNPs in a bin but none of them had an optimal design score ) . We also used PATROCLES ( www . patrocles . org , ) to identify supplemental SNPs with MAFs>0 . 05 in microRNA binding sites or non-synonymous SNPs from public databases to the potential SNP list . This identified an additional 170 miRNA binding site SNPs and nsSNPs with Illumina design scores>0 . 6 . In total this gave 1 , 410 tagSNPs , miRNA binding site SNPs and nsSNPs . In order to reach the final total of 1 , 536 SNPs for the Illumina GoldenGate assay , we added tagSNPs in another 12 candidate genes with MAF≥0 . 01 . The final list of 1 , 536 SNPs included 106 supplemental SNPs and 1 , 430 tagSNPs in 173 genes ( see Table S1 ) . The discovery samples were predominantly non-Hispanic White women with serous ovarian cancer and controls derived from two studies , the AUS and MAY studies , and were genotyped using the Illumina GoldenGate assay and Illumina BeadStudio software [47] , [48] . Plates were prepared containing randomly mixed cases and controls , with two duplicated samples and one blank per plate ( n = 20 ) . The Illumina GoldenGate assay was performed according to the manufacturer's instructions . Following completion of the assay , all plates were analysed using Illumina BeadStudio software version 3 . 1 . 0 . 0 . The original raw genotype dataset contained genotype information for 1 , 920 samples ( including blanks and duplicates ) and 1 , 536 SNPs . Following automatic clustering , SNPs were ranked using their GenTrain score ( ranging from 0 to 1 ) and those with GenTrain scores<0 . 5 were manually checked and adjusted according to Illumina guidelines . Samples with call rates below 95% and SNPs with call rates below 98% were excluded . A total of 1 , 292 SNPs passed this initial quality control ( QC ) . Genotyping quality was also assessed using tests for Hardy-Weinberg equilibrium ( HWE ) . Plots were examined for SNPs with significant deviations from HWE in controls ( 0 . 001<P<0 . 05 ) and the genotype data was excluded if the clustering was found to be suboptimal . SNPs with PHWE<0 . 001 were excluded from analysis . In addition , we genotyped 17 SNPs in CXCL9 , CTGF , LCN2 , DCN , and VIL2 , that had not been amenable to the Illumina GoldenGate assay or failed QC criteria , at the Queensland Institute of Medical Research using MALDI-TOF mass spectrophotometric mass determination of allele-specific primer extension products with Sequenom's MassARRAY platform and iPLEX Gold technology . The final discovery dataset for analysis consisted of 675 cases and 1 , 162 controls with genotype data on 1 , 309 SNPs . The three SNPs in PODXL , ITGA6 and MMP3 selected for replication by all participating OCAC sites ( with the exception of MMP3 at the MAY site ) were genotyped with the TaqMan allele discrimination assay ( Taqman Applied Biosystems , Foster City , CA ) , using primers designed by Assays-by-Design ( Applied Biosystems ) . MAY genotyping of MMP3 rs522616 was performed as part of a 1 , 536 Illumina Golden Gate Assay at the Mayo Clinic with cases and controls randomly mixed within each plate . Additional genotyping details are provided elsewhere [49] . Samples from five OCAC case-control studies ( MAL , SEA , UKO , USC and additional samples from AUS ) were genotyped for these and other replication SNPs , at the Queensland Institute of Medical Research using Sequenom iPLEX Gold technology . Primer design was carried out according Sequenom's guidelines using MassARRAY Assay Design software ( version 1 . 0 ) . Multiplex PCR amplification of fragments containing target SNPs was performed using Qiagen HotStart Taq Polymerase and a Perkin Elmer GeneAmp 2400 thermal cycler with 10 ng genomic DNA in 384 well plates . Shrimp Alkaline Phosphatase and allele-specific primer extension reactions were carried out according to manufacturer's instructions for iPLEX GOLD chemistry . Assay data were analysed using Sequenom TYPER software ( Version 3 . 4 ) . Only replication SNPs that met OCAC's QC criteria ( including >95% call rate , and >98% concordance between duplicates ) were included in the analysis [50] . The primary test for association in stage 1 was univariate analyses of the relationship between SNP genotypes and risk of serous ovarian cancer using the PLINK v0 . 99 Whole Genome Association Analysis toolset ( http://pngu . mgh . harvard . edu/purcell/plink/ ) [51] . Single-marker basic allelic association ( χ2 1df ) tests ( –assoc option ) analyses were performed on each of the 1 , 309 post-QC SNPs in a total of 1 , 837 women . PLINK default thresholds were utilized , resulting in further exclusions: maximum missing genotypes per person≤0 . 10 ( –mind option ) , maximum failed genotypes per SNP≤0 . 10 ( –geno option ) , MAF≥0 . 01 ( –maf option ) . Summary statistics were obtained for each SNP on the frequency of missing genotype data among cases and controls as well as a comparison of ‘missingness’ between cases and controls using the Fisher's exact test ( –test-missing option ) . Deviations from expected HWE proportions were analysed using the Fisher's exact test and the MAFs were also estimated for all SNPs . The Cochran Armitage Trend test ( χ2 1df ) assuming the log additive model ( –model option ) was performed to test the association between the minor allele of each SNP and serous ovarian tumors . Selection of stage 1 SNPs for replication analyses in stage 2 was prioritized as follows: first , SNPs with at least one failed duplicate , SNPs with a significantly different proportion of missing genotype data between cases and controls ( PMiss<0 . 05 ) , SNPs not conforming to HWE criteria ( see Genotyping and quality control ) for either cases , controls or both , and SNPs with no significant trend in allelic dose response ( PTrend>0 . 05 ) were excluded; secondly , we estimated from the remaining SNPs which were likely to be the best predictors of serous ovarian cancer risk by calculating the positive predictive value ( PPV ) using the PTrend values , the power of the study to detect this association , and the prior probability of 0 . 0001 [52] . Cases and controls from up to 14 additional studies participating in OCAC were included in replication analyses . We selected the three SNPs with the highest PPV for the larger replication analysis by all studies . Some additional individuals from AUS and MAY ( not in the discovery set ) were included in the replication analysis . Replication samples were examined to determine the distribution of race/ethnicity across studies , and analyses were restricted to White non-Hispanic women with serous invasive ovarian tumors . Significant differences by study site between age at interview for controls and age and diagnosis for cases were assessed using the Student's t-test for comparison of means . The MAF for each SNP was estimated from the control population for each study . The combined odds ratios ( OR ) and their 95% confidence intervals ( 95% CIs ) were obtained from unconditional logistic regression models for each SNP genotype . Assuming a log additive model of inheritance , the per-allele ORs and their 95% CIs associated with serous invasive ovarian cancer in non-Hispanic Whites for each SNP selected for replication were estimated by fitting the number of rare alleles carried as a continuous covariate . Separate comparisons for women with one copy ( heterozygotes ) and women with two copies ( rare homozygotes ) of the minor allele vs . those with no copies ( reference homozygotes ) were conducted for all replication SNPs . Between-study heterogeneity was assessed using the likelihood ratio test to compare logistic regression models with and without a genotype-by-study interaction term . Risk estimates from all replication analyses were adjusted for age at diagnosis for cases or age at interview for controls and study site . Exploratory analyses combining all ethnicities were additionally adjusted for ethnicity . Forest plots generated in exploratory analyses according to histological subtype were obtained using the rmeta library ( v2 . 15 ) implemented in the R project for Statistical Computing ( http://www . r-project . org/ ) , and LD plots were generated using Haploview v4 . 1 [53] . All tests for association were two-tailed , and unless otherwise specified , statistical significance was assessed at p<0 . 05 and tests for association in stage 2 were performed in STATA v . 9 . 0 ( StataCorp , USA ) .
|
In this article , we report the findings from a large-scale analysis of common variation in genes that are expressed as a consequence of interactions between ovarian cancer cells and their host micro-environment that could influence serous ovarian cancer risk . We evaluated 1 , 302 common variants within or near 173 genes in two large case-control studies from the Ovarian Cancer Association Consortium ( OCAC ) and selected three variants for further evaluation in sixteen OCAC studies and an additional 18 for evaluation in five OCAC studies . We observed a significantly increased risk of serous ovarian cancer associated with a variant in the telomerase reverse transcriptase ( TERT ) gene . Although TERT variants have not been previously shown to contribute to ovarian cancer risk , several studies have recently reported associations between TERT variants and other forms of cancer , including gliomas , lung cancer , adenocarcinoma , basal cell carcinoma , prostate cancer , and multiple other cancers . TERT encodes a protein that is essential for the replication and maintenance of chromosomal integrity during cell division . In cancer cells , TERT has been linked to genomic instability and tumour cell proliferation . Further studies are necessary to confirm our findings and to investigate the mechanisms for the observed association .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"public",
"health",
"and",
"epidemiology/epidemiology",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"genetics",
"and",
"genomics/complex",
"traits",
"genetics",
"and",
"genomics/cancer",
"genetics"
] |
2010
|
Evaluation of Candidate Stromal Epithelial Cross-Talk Genes Identifies Association between Risk of Serous Ovarian Cancer and TERT, a Cancer Susceptibility “Hot-Spot”
|
Aedes aegypti poses a serious risk to human health due to its wide global distribution , high vector competence for several arboviruses , frequent human biting , and ability to thrive in urban environments . Pyrethroid insecticides remain the primary means of controlling adult A . aegypti populations during disease outbreaks . As a result of decades of use , pyrethroid resistance is a global problem . Cytochrome P450 monooxygenase ( CYP ) -mediated detoxification is one of the primary mechanisms of pyrethroid resistance . However , the specific CYP ( s ) responsible for resistance have not been unequivocally determined . We introgressed the resistance alleles from the resistant A . aegypti strain , Singapore ( SP ) , into the genetic background of the susceptible ROCK strain . The resulting strain ( CKR ) was congenic to ROCK . Our primary goal was to determine which CYPs in SP are linked to resistance . To do this , we first determined which CYPs overexpressed in SP are also overexpressed in CKR , with the assumption that only the CYPs linked to resistance will be overexpressed in CKR relative to ROCK . Next , we determined whether any of the overexpressed CYPs were genetically linked to resistance ( cis-regulated ) or not ( trans-regulated ) . We found that CYP6BB2 , CYP6Z8 , CYP9M5 and CYP9M6 were overexpressed in SP as well as in CKR . Based on the genomic sequences and polymorphisms of five single copy CYPs ( CYP4C50 , 6BB2 , 6F2 , 6F3 and 6Z8 ) in each strain , none of these genes were linked to resistance , except for CYP6BB2 , which was partially linked to the resistance locus . Hence , overexpression of these four CYPs is due to a trans-regulatory factor ( s ) . Knowledge on the specific CYPs and their regulators involved in resistance is critical for resistance management strategies because it aids in the development of new control chemicals , provides information on potential environmental modulators of resistance , and allows for the detection of resistance markers before resistance becomes fixed in the population .
Aedes aegypti is an important pest capable of transmitting four important human disease viruses: dengue , yellow fever , chikungunya , and Zika . Dengue , for example , causes morbidity and mortality in 141 countries across the tropical and subtropical regions of the world and is estimated to be a risk to over 50% of the world’s population [1] . Yellow fever is an endemic disease in the tropical regions of Africa and South America with a recently rising number of cases in Brazil [2 , 3] . Chikungunya is a disease new to the Americas as of 2013 [4] that often causes debilitating joint pains in addition to flu-like symptoms . Zika was introduced to the Americas in 2015 [5] and has generated great concerns due to its association with birth defects and Guillain-Barré syndrome [6] . Given that A . aegypti has a wide global distribution , high vector competence for several arboviruses , frequently bites humans and thrives in urban environments , it poses a serious risk to human health . Insecticides are still the primary means to control A . aegypti in endemic areas . More specifically , pyrethroids are the most widely used class of insecticides for control of adult A . aegypti [7] in the past three decades . As a result of this continued use , pyrethroid resistance in A . aegypti is a global problem [8] . Cytochrome P450 monooxygenase ( CYP ) -mediated detoxification is one of the primary mechanisms of pyrethroid resistance in mosquitoes . CYPs are a large family of enzymes that metabolize both endogenous substrates and xenobiotics , such as insecticides . A . aegypti have approximately 160 CYP genes [9] . Several studies have directly ( e . g . in vivo and/or in vitro metabolism ) or indirectly ( reduction in resistance with the CYP inhibitor piperonyl butoxide ( PBO ) ) implicated CYPs as a mechanism of pyrethroid resistance in A . aegypti [8] . Elucidating the specific CYP ( s ) responsible for resistance is challenging because of the large number of CYPs and because CYP-mediated resistance can be due to overexpression of a CYP or to a mutation in the open reading frame of a CYP [10] . However , identifying these CYPs is extremely important to manage resistance because it allows us to detect resistance markers and stop insecticide use before resistance becomes fixed in the population [11] . Knowledge of the specific CYPs may also aid in the development of new insecticides and resistance inhibitors as well as allow us to better understand the influence of environmental xenobiotics in the development of insecticide resistance [12] . Most studies done to identify the CYPs responsible for resistance in A . aegypti have looked at changes in expression levels using unrelated strains [9 , 13–16] . However , when strains of different origins are used , it is not possible to determine the exact relationship between the overexpressed genes and insecticide resistance , because CYP expression can vary for reasons unrelated to insecticide resistance . For example , CYP9M9 was overexpressed in the SBE strain , relative to the BORA strain [13] , even though both were susceptible strains . Increased transcription of a CYP resulting in resistance could be due to a change in the regulatory region of the CYP [17] , to a change in a CYP regulatory protein , or an increase in the copy numbers of the CYP through gene amplification [18 , 19] . These processes of increasing CYP expression would give different outcomes . First , a mutation in a specific CYP that leads to increased expression ( cis-regulation ) would be expected to show a specific increase in only that CYP , and the resistance would map to that CYP . In contrast , if resistance is due to a mutation in a gene regulatory protein ( trans-regulation ) , there could potentially be multiple CYPs whose expression are elevated in the resistant strain , even if only one of them is responsible for the resistance . In addition , the resistance locus would not map to the CYP that is overexpressed . There are now several cases where CYP overexpression is found in insecticide resistant strains , and the overexpression is due to trans-regulation of the CYP . Examples include CYP6D1 in Musca domestica [20] , 6A2 and 6A8 in Drosophila melanogaster [21] , 6BJa/b , 6BJ1 , 9Z25 , and 9Z29 in Leptinotarsa decemlineata , and 4G7 , 4G14 and 6BQ in Tribolium castaneum [22–24] . Gene amplification could also lead to increased expression of a single CYP or multiple CYP genes occurring in tandem depending on the length of the duplicated region . In this case , the resistance locus could map either to one or more of the duplicate CYPs . CYPs as a group are rapidly evolving genes [25] and are frequently polymorphic within and between strains [26 , 27] . When a mutation causing resistance occurs and is under high selection pressure such that the resistance allele becomes fixed in the strain or population , the region near this mutation will have decreased amounts of polymorphisms relative to the rest of the genome [28–35] . Thus , reduced abundance of single nucleotide polymorphisms ( SNPs ) are useful to detect resistance loci [31–33] . The footprint of the region around the resistance locus will decrease in time , as recombination introduces back variation , but this will be a slow process . Furthermore , if a mutation in a CYP causes resistance , we would expect the CYP to have a single unique allele in the resistant strain , but be polymorphic in susceptible strains . One of the best-characterized pyrethroid resistant strains of A . aegypti is Singapore ( SP ) . SP developed a 1650-fold resistance to permethrin ( relative to the susceptible SMK strain ) after 10 generations of selection [19] . Pyrethroid resistance in SP is due to CYP-mediated detoxification and target site insensitivity ( V1016G+S989P mutations in the voltage sensitive sodium channel [Vssc] ) . CYP-mediated resistance was unambiguously demonstrated in SP through in vitro and in vivo metabolism experiments and by PBO suppression of the resistance . Nine CYP genes ( CYP9M6 , 9M5 , 9M4 , 6Z8 , 6Z7 , 6F3 , 6F2 , 6BB2 and 4C50 ) in SP were overexpressed >3-fold relative to the susceptible SMK strain [19] . Overexpression of four of these ( CYP6Z7 , 9M4 , 9M5 and 9M6 ) was due , in part , to gene amplification . The genetic linkage of these CYPs , or the genetic linage of their overexpression , relative to resistance has not been investigated . In order to understand which CYPs in SP map to the resistance locus , we introgressed the resistance from SP into the genetic background of the susceptible ROCK strain resulting in CYP+KDR:ROCK ( CKR ) , a resistant strain congenic to ROCK . We then asked two questions . First , which CYPs overexpressed in SP are also overexpressed in CKR relative to ROCK ? Second , are any of the overexpressed CYP genes cis- ( map to a resistance locus ) or trans-regulated ( do not map to a resistance locus ) ?
Two parental strains of A . aegypti were used: Rockefeller ( ROCK ) , an insecticide-susceptible strain which originated from the Caribbean [36] and has been reared without exposure to insecticides for several decades , and Singapore ( SP ) , a pyrethroid resistant strain in which the mechanisms of resistance have been well studied [19] . SP is resistant to permethrin due to two mutations in Vssc , V1016G+S989P ( referred to as kdr ) , and CYP-mediated detoxification , but not by hydrolases or decreased cuticular penetration [19] . A third strain , CYP+KDR:ROCK ( CKR ) , was isolated from crossing ROCK with SP followed by four backcrosses and permethrin selections . CKR is congenic to ROCK , but resistant to pyrethroids due to CYP-mediated resistance and to Vssc mutations S989P+V1016G . The procedure for isolating CKR is illustrated in Fig 1 . In short , unmated ROCK females were crossed en masse with SP males . Unmated F1 females were backcrossed with ROCK males and unmated BC1 females were selected with a permethrin dose that killed at least 60% . BC1 females that survived were backcrossed to ROCK males . This process was repeated for the BC2 and BC3 generations , again using doses of permethrin that gave approximately 60% mortality . To ensure that we retained all the resistance alleles , both male and unmated female BC3 were selected with permethrin ( ~60% kill ) and crossed with each other prior to backcrossing to ROCK again . At BC4 both males and unmated females were selected with permethrin ( ~60% kill ) and reared en masse . Males and unmated females from the following three generations were selected with permethrin ( ~60% kill ) and reared en masse . At BC4F4 , kdr homozygosity was confirmed by allele-specific polymerase chain reaction ( ASPCR ) , ( n = 190 ) following our established protocol [37] . The resultant strain was named CKR ( Fig 1 ) . Mosquitoes were reared at 27˚C ( ± 1˚C ) with 70–80% relative humidity , and a photoperiod of 14L:10D . Females were blood fed using membrane-covered water-jacketed glass feeders with cow blood ( Owasco Meat Co . , Moravia , NY ) . Adults were maintained on 10% sugar water in cages approximately 35 x 25 x 25 cm holding ≤ 1000 mosquitoes . Larvae ( ~400–600 ) were reared in 27 . 5 x 21 . 5 x 7 . 5 cm containers with 1 L distilled water and fed Cichlid Gold fish food pellets ( Hikari , Hayward , CA ) ( ground pellets for 1st instar and medium size pellets for 2nd to 4th instars ) . Food pellets were given daily as needed . Adult bioassays were done by topical application using 3- to 7-day-old mated females . Permethrin ( 99 . 5% pure , 24 . 1% cis , 75 . 8% trans , Chem Service ) and piperonyl butoxide ( PBO ) ( 90% , Sigma-Aldrich ) were diluted in acetone ( VWR , Radnor , PA , USA ) for the bioassays . Mosquitoes were briefly anesthetized with CO2 and held on ice . A 0 . 22 μL drop of permethrin in acetone was applied to the thorax of each mosquito using a Hamilton PB-600 repeating dispenser equipped with a 10-μL syringe . Controls were treated with acetone only . At least five doses were used per bioassay with at least three giving mortality values between 0 and 100% and each containing 20 mosquitoes . Mosquitoes were given a cotton ball saturated with distilled water and held at 25˚C . A minimum of four replicates over at least two days and two cages were done per strain . Mortality was defined as mosquitoes that were ataxic after 24 h . Probit analysis [38] , as adapted to personal computer use [39] using Abbott’s [40] correction for control mortality , was used to calculate the LD50 and the 95% confidence intervals ( CI ) . All of the bioassay data fit a line ( chi-square test ) . Resistance ratios ( RR ) were calculated by dividing the LD50 of the resistant strain ( SP or CKR ) by the LD50 of ROCK . Significant differences were determined by calculating the RRs for the minimum and maximum LD50 values based on the 95% CI . If the minimum and maximum RR values did not overlap , they were deemed significantly different . Bioassays using the synergist PBO was performed as described above , except that 2 . 5 μg PBO ( maximum sublethal dose for the ROCK strain ) was applied to each mosquito 2 h prior to permethrin application . For this , the mosquitoes were anesthetized on ice twice , once for PBO and once for permethrin application . Two controls were run: double acetone and an acetone plus PBO application . Ten 5–7 days old mated female mosquitos were pooled into 2 mL micro tubes ( Starstedt AG & Co . , Nümbrecht , Germany ) containing 500 μL of TRIzol reagent ( Invitrogen , Carlsbad , CA , USA ) per replicate and four replicate tubes were prepared . The mosquitos were pulverized at 4 . 5 m/s for 20 s with an MP FastPrep 24 bead beater ( MP-Biomedicals , Santa Ana , CA , USA ) . The RNA content was extracted following Invitrogen’s TRIzol reagent protocol . The concentration of RNA was measured using a NanoDrop 2000 spectrophotometer ( Thermo Fisher Scientific Inc . , Waltham , MA , USA ) and then diluted to 10 μg/50μL with nuclease-free water to standardize the concentration between the tubes . DNA was removed by DNase treatment ( TURBO DNA-free kit , Invitrogen ) following the manufacturer’s instructions . Complete digestion of DNA was confirmed by lack of PCR amplification of the 5’ UTR of CYP6Z7 ( S1 Table ) determined by visual inspection on an ethidium bromide-stained 1% agarose gel . Complementary DNA ( cDNA ) was synthesized with 1 μg of total RNA per reaction using the Promega GoScript Reverse Transcription System kit ( Promega , Madison , WI , USA ) and random primers per the manufacturer’s instructions . The cDNA pools were then diluted 1:5 using nuclease-free water before use in real time quantitative polymerase chain reaction ( RT-qPCR ) . RT-qPCR plates were set up with three cDNA biological replicates and two technical replicates of each biological replicate . Two strains were compared at a time; first ROCK and SP , then ROCK and CKR . For each strain comparison , the nine CYPs were run along with two internal control genes , ribosomal protein S3 ( RPS3 ) and eukaryotic translation elongation factor 1-alpha ( EF1α ) . Plates were spun down at 2100 RPM for 1 minute to ensure the liquid had reached the bottom of the wells . The reaction volume ( 20 μL ) contained 10 μL of 2 × iQTM SYBR Green SuperMix , 7 . 4 μL of nuclease-free water , 0 . 8 μL of 10 μM of each specific primer ( S1 Table ) , and 1 μL of first-strand cDNA template . The qPCR was performed in a CFX ConnectReal-Time PCR Detection System ( Bio-Rad , Hercules , CA , USA ) thermocycler with the following program: an initial denaturation and enzyme activation at 95°C for 10 min followed by 40 cycles of denaturation at 95°C for 10 s , annealing at 60°C for 10 s with a plate read , and extension at 72°C for 10 s . An automatic dissociation step cycle was added for melting curve analysis . Relative quantification analysis was performed using the amplification efficiency-corrected ΔΔCt method [41] . The change in Ct value of each strain between the target CYP gene and the reference gene ( RPS3 or EF1α ) represents a ΔCt value , while the change in ΔCt value of a CYP between the susceptible strain ( ROCK ) and a resistant strain ( SP or CKR ) represents a ΔΔCt , or fold-expression difference value . This is synonymous to , and will be referred to as , an R/S value in this paper . Amplification efficiency for each gene and strain was determined using LinRegPCR with a 20% exclusion of outliers from the median value , along with a manual correction of the window of linearity to fit the straight continuous set of data points in the log-linear phase of the amplification plots [42] . Data were normalized to the two endogenous controls for both strain comparisons . Multiple t-tests were conducted to determine the significance of the R/S ratios . Genomic DNA was extracted using two methods; 1 ) an isopropanol precipitation method from whole bodies of pooled mosquitoes , and 2 ) an alkali extraction from the hind legs of individual mosquitoes . The isopropanol extraction was conducted as follows: eight whole mosquitoes were placed in 2 mL tubes ( Starstedt Inc . , Nümbrecht , Germany ) containing ten 2 . 3-mm diameter zirconia/silica beads ( BioSpec Products , Bartlesville , OK , USA ) and 400 μl Buffer A ( 100 mM Tris-HCl , pH 8 . 0 , 100 mM EDTA , 100 mM NaCl , 0 . 5% SDS , ddH2O ) . Samples were homogenized and 800 μl of 4 . 3 M LiCl and 1 . 4 M KOAc solution was added followed by centrifugation ( 14 , 100 x g for 10 min ) and collection of supernatant . Next , 570 μl isopropanol was added , mixed , centrifuged ( 14 , 100 x g for 10 min ) , and supernatant removed to isolate DNA pellet . Tubes containing the pellet were centrifuged ( 14 , 100 x g for 30 sec ) once more with 500 μl of 70% EtOH . The supernatant was removed , then the DNA pellet was dried and resuspended in ddH2O . The alkali extraction method was conducted as follows: legs of individual mosquitoes were placed in individual wells of a 96-well PCR plate ( BioRad , Hercules , CA , USA ) containing three 2 . 3-mm diameter zirconia/silica beads and 10 μl 0 . 2 M NaOH per well . The leg samples were beaten for 1–2 min on a vortex mixer at maximum speed and then incubated for 10 min at 70˚C . Ten μL of neutralization buffer ( 360 mM Tris-HCl , pH 8 . 0 and 10 mM EDTA ) and 80 μL ddH2O were then added to each well . PCR was carried out using 2 μl template gDNA , 10 μl PrimeSTAR GXL Buffer ( Takara Bio Inc . , Shiga , Japan ) , 4 μl dNTP Mixture , 1 μl PrimeSTAR GXL DNA Polymerase , 2 μl forward and reverse primer mix ( S1 Table , S1 Fig ) , 31 μl ddH20 and the following thermocycler conditions: 95˚C for 3 min , 37 x ( 98˚C for 10 sec , 60˚C for 15 sec , 68˚C for 3 min ) and 68˚C for 10 min . Our goal was to examine the polymorphisms in the CYPs that are overexpressed in the SP strain [19] to look for markers or mutations that can link the CYPs to resistance . To do this we sequenced CYP gDNA from pools of mosquitoes from the SP and ROCK strains . If polymorphisms were found in the pools , eight additional mosquitos were sequenced individually . If no polymorphism were found in the pools , this process was repeated until we had high confidence in the polymorphisms in each strain . For genes where SNPs were found , the SP sequences were compared to ROCK sequences to check for any reliable and unique SNPs in SP . We then sequenced from CKR any CYP with a reliable marker to examine if it was inherited from the ROCK or SP parent strain ( i . e . linked to resistance or not ) . ROCK and SP gDNA were sequenced and aligned for each of the single copy CYPs ( 4C50 , 6BB2 , 6F2 , 6F3 , and 6Z8 ) that are overexpressed in the SP strain . Four of the nine overexpressed CYPs found by Kasai et al . [19] could not be used to search for SNPs due to multiple gene copies; these were CYP6Z7 , 9M4 , 9M5 and 9M6 . CYP sequences were determined by Sanger sequencing using PCR products treated with ExoSAP ( Thermo Fisher Scientific , Waltham , MA , USA ) and sequenced at the Cornell University Biotechnology Resource Center ( BRC ) . Sequence alignments were carried out on DNASTAR’s Lasergene software , EditSeq and SeqMan Pro ( Madison , WI ) . SNPs were searched for using the SNP Report feature in SeqMan Pro and confirmed by visually inspecting the alignments and chromatograms .
CKR and SP were both resistant to permethrin ( 110- and 360-fold respectively ) relative to ROCK ( Table 1 ) . The resistance was PBO suppressible in both CKR and SP , lowering the RR to 77- and 70-fold , respectively , confirming CYP-mediated resistance in both strains . The 3-fold difference in the resistance ratio ( RR ) between the CKR and SP strains suggests some minor mechanism of resistance was lost during the isolation of the CKR strain . To determine if overexpression was genetically linked to permethrin resistance , CYP expression was quantified in both the SP and CKR strains relative to ROCK . CYP6BB2 , 6Z8 , 9M5 and 9M6 were overexpressed in both SP and CKR indicating that their overexpression is linked to resistance . Seven CYPs are significantly overexpressed in the SP strain relative to ROCK: CYP6BB2 , 6F2 , 6F3 , 6Z7 , 6Z8 , 9M5 and 9M6 ( Fig 2 ) . The level of increased expression in SP was 169 for CYP9M6 ( p = 7 . 0 x 10−6 ) , 54 for CYP9M5 ( p = 6 . 6 x 10−6 ) , 7 . 5 for CYP6BB2 ( p = 0 . 01 ) , 5 . 2 for CYP6Z7 ( p = 0 . 01 ) , 3 . 8 for CYP6Z8 ( p = 2 . 2 x 10−4 ) , 1 . 7 for CYP6F3 ( p = 0 . 03 ) , and 1 . 4 for CYP6Z2 ( p = 0 . 03 ) . For CKR relative to ROCK , only CYP6BB2 , CYP6Z8 , CYP9M5 and CYP9M6 were significantly overexpressed ( Fig 2 ) . The fold- change in expression ( R/S ) in CKR were 76 for CYP9M6 ( p = 0 . 05 ) , 21 for CYP9M5 ( p = 0 . 02 ) , 6 . 9 for CYP6BB2 ( p = 2 . 8 x 10−3 ) , and 1 . 5 for CYP6Z8 ( p = 0 . 03 ) . When comparing the two resistant strains , SP had a higher level of expression of CYP6F2 ( p = 4 . 0 x 10−3 ) , 6F3 ( p = 3 . 0 x 10−3 ) , 6Z7 ( p = 0 . 01 ) , 6Z8 ( p = 4 . 8 x 10−4 ) , CYP9M5 ( p = 1 . 5 x 10−3 ) , and CYP9M6 ( p = 0 . 02 ) compared to CKR . On average , the expression in SP was about 2 . 4-fold greater than seen for CKR . CYP4C50 , 6BB2 and 9M4 , and were not significantly different between SP and CKR ( Fig 2 ) . The basal CYP transcription levels ( ΔCt values ) were also investigated for each strain relative to RPS3 and EF1α . Expression levels of all nine CYPs were readily detectable ( S2 Fig , S3 Fig ) and similar between both endogenous controls . Genomic DNA of five single copy CYPs ( 4C50 , 6BB2 , 6F2 , 6F3 , 6Z8 ) were sequenced from both the ROCK and SP strains ( see S1 Fig for diagram of genes and approximate primer locations ) . GenBank accession numbers for consensus CYP sequences are listed in S2 Table . These CYPs were selected because they are overexpressed in SP , but not duplicated [19] . SP had more SNPs than ROCK in CYP4C50 , 6F2 , 6F3 and 6Z8 ( Table 2 ) . The frequency of polymorphisms per kilobase ( kb ) ranged from 23–53 in SP and from 18–36 in ROCK . However , there were no strain-specific polymorphisms ( i . e . neither a unique nucleotide between strains at a non-polymorphic site , nor a SNP in which both nucleotides differed between strains ) . This is not what would be expected for a gene at a resistance locus and leads us to conclude that the CYP4C50 , 6F2 , 6F3 and 6Z8 genes are not linked to resistance , even though their increased expression was ( see above ) . In contrast to these four CYPs , there were many less SNPs detected in CYP6BB2 , which had 0 . 08 and zero SNPs per kb in ROCK and SP , respectively ( Table 2 ) . CYP6BB2 had a unique , strain-specific synonymous polymorphism ( thymine in ROCK and cytosine in SP at position 1595 ) that was homozygous in both strains . This allowed us to test if CYP6BB2 was linked to resistance by sequencing this CYP from the CKR strain . We found seven individuals homozygous for the SP allele , 10 heterozygotes , and four homozygous for the ROCK allele in the CKR strain . This indicates a partial genetic linkage of CYP6BB2 and the resistance locus . Based on the methods used to isolate the CKR strain , a measurement of the linkage was not possible .
Our bioassay results generated compelling data that we had isolated a strain ( CKR ) that had CYP-mediated resistance and kdr from the SP strain . Most of the resistance in SP was recaptured in the isolation of the CKR strain , although the 3-fold lower permethrin resistance in CKR ( 110-fold ) compared to SP ( 360-fold ) reveals that some resistance alleles may have been lost in the selection process . This can happen if the resistance factor is recessive and/or has a high fitness cost [43] . Interestingly , the RR value to SP is nearly 5-fold lower than that reported in Kasai et al . 2014 . There are at least three possible explanations for this . First , different batches of permethrin were used and this is known to alter the expression of resistance [44] . Second , different susceptible strains were used and this can cause differences in levels of resistance reported [45] . Third , the SP strain may have lost some resistance while being maintained in the lab since it was received in 2014 . Bioassays with PBO , reduced the RR to 77- and 70-fold in CKR and SP respectively , confirming the involvement of CYP-mediated resistance in both strains . The suppression of resistance with PBO was incomplete ( kdr alone confers 40-fold resistance [37] ) as is commonly seen in strains with CYP-mediated resistance [19 , 46] . Both CKR and SP are homozygous for the S989P+V1016G Vssc mutations . Our study both validates previous work [19] and provides new information about the basis of CYP-mediated resistance in SP . Consistent with what was previously reported [19] , we find elevated expression of CYP6BB2 , 6Z7 , 6Z8 , 9M5 and 9M6 in SP . Given that we used a different susceptible strain in this study , and still found increased expression of these CYPs in SP , strengthens the hypothesis that these CYPs are involved in resistance . Further , our data provides evidence that the overexpression of four CYPs in the SP strain are genetically linked to resistance: CYP6BB2 , 6Z8 , 9M5 and 9M6 . For CYP9M5 and 9M6 ( but not 6BB2 or 6Z8 ) , this is due in part to gene duplication [19] . How many different transcriptional regulation genes might be involved in insecticide resistance is an important , but unanswered question . Thus far , several different transcription factors have been implicated in insecticide resistance , including Gfi-1 in M . domestica [17 , 47] cap n collar C ( CncC ) and muscle aponeurosis fibromatosis ( Maf ) family transcription factors in Tribolium castaneum [24] . Identification of the mutation responsible for the increased expression of CYPs in the SP strain would expand our knowledge about this important evolutionary process and would provide a means by which the population genetics of this resistance could be studied . There is clearly some evolutionary plasticity in CYP-mediated resistance [48] , but identification of the transcriptional regulatory factors , the mutations that cause CYP overexpression and the geographic frequency of these mutations are needed before we will start to have a satisfactory understanding of this important mechanism of resistance . CYP genes are generally highly polymorphic . For example , in Anopheles gambiae CYPs have an average SNP frequency of 1 every 26 bp compared to the 1 every 34 bp genome average [26] . Determining the genetic diversity of A . aegypti has proven to be a challenging task due to the large genome size and high percentage of repetitive transposable elements [49] . One study found the average SNP frequency in the A . aegypti genome to be 12 per kb , however estimates of average nucleotide diversity ( π ) have varied greatly , ranging from about 0 . 001 to 0 . 015 [50] [51 , 52] . We found that the CYPs we studied ( with the exception of 6BB2 ) had a frequency of polymorphisms similar to the average reported for An . gambiae [26] with an average SNP frequency of 1 per 36 bp in ROCK and 1 per 26 bp in SP . However , CYP6BB2 also had little variation ( only two SNPs in the 2430 bp sequenced ) in ROCK . The low level of polymorphism in CYP6BB2 appears more to do with the stains we used , rather than the gene per se , as wild A . aegypti populations from Uganda and Senegal had 164 CYP6BB2 SNPs , from Mexico there were 45 SNPs , and in populations from Sri Lanka there were no SNPs [50] . Overall , our results suggest that CYP-mediated resistance in SP is due to a trans-regulatory factor ( s ) that is capable of increasing the expression of multiple CYPs . The overexpression of four CYPs ( CYP6BB2 , 6Z8 , 9M5 and 9M6 ) were linked to resistance . However , sequencing of the five single-copy CYPs that were found to be overexpressed in the SP strain , revealed that none of them showed expected signs of being at the resistance locus , except for CYP6BB2 which showed partial linkage to a resistance locus . Given that three of the CYPs have multiple copies in SP precluded us from being able to evaluate their linkage to resistance . Hopefully more sequence information will become available for these amplicons in the future which would allow for testing of linkage . Based on these results and other studies [20–24 , 53–55] , it appears that trans-regulation of CYP expression may be a common mechanism of insecticide resistance .
|
Cytochrome P450 monooxygenases ( CYPs ) are one of the most important mechanism of insecticide resistance in mosquitoes . These CYP enzymes break down insecticides into non-toxic forms that can be readily excreted . An increase of CYP-mediated detoxification is commonly found in pyrethroid resistant Aedes aegypti , however the link between specific CYPs and the resistance loci have not been clearly established for this species . In this study , we measured the expression levels of nine candidate CYPs in two strains highly resistant to permethrin: SP , a field collected strain that was selected with permethrin for high levels of resistance and CKR , a strain that contains the resistance mechanisms from SP , but that is congenic ( i . e . has the genetic background ) to the insecticide susceptible strain , ROCK . We found seven overexpressed CYPs in SP and four in CKR , confirming their involvement in resistance . Next , we sequenced the CYP genes ( with the exception of the duplicated ones ) to determine if the genes themselves are located in the resistance locus ( meaning their expression is cis-regulated ) or not ( meaning their expression is trans-regulated ) . We found no reduced polymorphisms in any of the resistant strain ( SP ) CYPs , suggesting that the overexpression of these CYPs ( and thus CYP-mediated resistance ) is trans-regulated .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"animals",
"molecular",
"biology",
"techniques",
"molecular",
"genetics",
"insect",
"vectors",
"research",
"and",
"analysis",
"methods",
"agrochemicals",
"infectious",
"diseases",
"artificial",
"gene",
"amplification",
"and",
"extension",
"aedes",
"aegypti",
"hyperexpression",
"techniques",
"molecular",
"biology",
"disease",
"vectors",
"insects",
"agriculture",
"genetic",
"loci",
"arthropoda",
"insecticides",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"gene",
"expression",
"and",
"vector",
"techniques",
"mosquitoes",
"eukaryota",
"polymerase",
"chain",
"reaction",
"heredity",
"genetics",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"genetic",
"linkage",
"organisms"
] |
2018
|
CYP-mediated permethrin resistance in Aedes aegypti and evidence for trans-regulation
|
The vigorous cytokine response of immune cells to Gram-negative bacteria is primarily mediated by a recognition molecule , Toll-like receptor 4 ( TLR4 ) , which recognizes lipopolysaccharide ( LPS ) and initiates a series of intracellular NF-κB–associated signaling events . Recently , bladder epithelial cells ( BECs ) were reported to express TLR4 and to evoke a vigorous cytokine response upon exposure to LPS . We examined intracellular signaling events in human BECs leading to the production of IL-6 , a major urinary cytokine , following activation by Escherichia coli and isolated LPS . We observed that in addition to the classical NF-κB–associated pathway , TLR4 triggers a distinct and more rapid signaling response involving , sequentially , Ca2+ , adenylyl cyclase 3–generated cAMP , and a transcriptional factor , cAMP response element–binding protein . This capacity of BECs to mobilize secondary messengers and evoke a more rapid IL-6 response might be critical in their role as first responders to microbial challenge in the urinary tract .
The innate immune system is the first line of defense against infection and is thought to primarily be mediated by phagocytic immune cells such as macrophages and dendritic cells . These cells recognize microorganisms via a limited number of germline-encoded pattern recognition receptors ( PRRs ) that recognize microbial components known as pathogen-associated molecular patterns , which are essential for the survival of the microorganism and , therefore , difficult for the microorganism to alter [1] . Several classes of PRRs , including Toll-like receptors ( TLRs ) and cytoplasmic receptors , recognize distinct microbial components and directly activate immune cells , triggering intracellular signaling cascades that rapidly induce the expression of a variety of inflammatory cytokines that initiate a variety of overlapping immune responses . One of the best known PRRs is TLR4 , which recognizes the major Gram-negative bacterial surface component lipopolysaccharide ( LPS ) [1] . Studies on TLR4 signaling in monocytes , macrophages , and dendritic cells have revealed that engagement of TLR4 by LPS triggers a signaling cascade involving several intracytoplasmic and nuclear transcriptional factors . TLR4 activation first engages a set of adaptor family members that link TLR4 to the serine/threonine kinases . These kinases mediate phosphorylation and ubiquitination of various substrates , eventually resulting in the activation of the transcriptional factor NF-κB , which regulates the expression of several immunomodulatory cytokines [2] . The urinary tract is extremely intractable to infection by most pathogens . This is attributable to a large extent on the multifaceted innate immune defenses of the bladder and , in particular , bladder epithelial cells ( BECs ) . These cells selectively exfoliate upon bacterial colonization and undergo re-epithelialization as a mechanism to reduce bacterial load in the bladder . They are also a major source of proinflammatory cytokines and chemokines in the urinary tract following bacterial infection [3 , 4] . These BEC-derived mediators are responsible for the vigorous neutrophil response , which is responsible for early clearance of infecting bacteria [5] . A prominent mediator released by BECs is IL-6 and it is , by far , the single most prominent cytokine detected in the urine of infected patients [6] . IL-6 is known to mobilize and amplify both local as well as systemic innate immune defenses against infection [7] . The production of some of the earliest indicators of inflammation in the body such as the acute phase proteins has been directly related to production of this cytokine [7] . Considering the large number of pathogens capable of infecting the urinary tract , it is remarkable that uropathogenic Escherichia coli ( UPEC ) account for over 85% of urinary tract infections in patients without underlying predisposing factors [8] . The singular success of UPEC in the urinary tract has been attributed to bacterial surface expression of filamentous fimbrial appendages , called type 1 fimbriae [9] . These structures promote avid bacterial binding to uroplakin 1a molecules on the surface of BECs , triggering bacterial invasion of these cells [10 , 11] . In their intracellular location , UPEC avoid elimination by the flushing actions of urine [12] . Recent studies have suggested additional traits on UPEC that account for their success as uropathogens . These include their capacity to block apoptosis and exfoliation of infected BECs [13] as well as inhibit the ability of BECs to mount a cytokine responses [14] . Although several genes on UPEC have been implicated in inhibiting cytokine production , the underlying mechanism remains elusive , a problem exacerbated , at least partly , by the fact that most of our current understanding of TLR4 signaling is based almost exclusively on cells of hematopoietic origin [1] . Here , we sought to better define LPS/TLR4 signaling pathway in BECs . We were especially interested in defining the role , if any , of two second messengers , Ca2+ and cyclic adenosine monophosphate ( cAMP ) , since these low molecular weight diffusible molecules have been globally implicated in cellular signaling events , including cytokine responses . We investigated the IL-6 responses of human BECs to E . coli and to isolated LPS . Our studies demonstrated that the IL-6 response triggered by TLR4 in BECs involves not only the classical NF-κB–associated pathway , but also a distinct pathway involving Ca2+ , cAMP , and cAMP response element–binding protein ( CREB ) . Interestingly , the latter pathway resulted in a significant IL-6 response , which is evident at least 3 h before the NF-κB–associated pathway .
A K-12 laboratory E . coli strain ORN103 ( pSH2 ) and a UPEC type 1 fimbriated and non-hemolytic strain CI5 were utilized in this study [15–17] . Bacteria and the human BEC line 5637 ( ATCC HTB-9 ) and primary BECs were cultured as described previously [11 , 18] . Human airway epithelial cells ( 16-HBE ) were cultured in DMEM plus 4 mM glutamine and 10% FBS , and the human monocytic cell line ( Mono Mac 6 ) was cultured as described previously [19] . IL-6 secretion was tested by using the human IL-6 ELISA kit ( R&D Systems , http://www . rndsystems . com ) according to the manufacturer's protocol . Cell viability was not affected by any of the pharmacological agents employed , as assessed by trypan blue exclusion assays . Ratiometric Ca2+ imaging was performed as described previously [20] . The Fura-2 calcium imaging calibration kit ( Molecular Probes , http://probes . invitrogen . com ) was used to calibrate fluorimetric analyses to quantify intracellular calcium concentrations . Intracellular concentrations of cAMP were determined using a cAMP enzyme immunoassay kit ( Sigma , http://www . sigmaaldrich . com ) according to the manufacturer's instructions . Total cellular RNA was isolated using an RNeasy purification system ( Qiagen , http://www . qiagen . com ) . Two micrograms of total RNA was reverse transcribed and amplified with gene-specific primers using the RT-PCR System kit ( Bio-Rad , http://www . bio-rad . com ) . The primer sequences for the genes and expected product sizes are summarized in Table S1 . We confirmed that the adenylyl cyclase ( AC ) isotype-specific primers were functional by undertaking reverse transcription ( RT ) –PCR on total RNA from HEK cells ( a positive control cell , where all ACs except AC-4 and AC-8 were expressed ) [21] . Detailed methods and target sequences , including GenBank accession numbers for the genes mentioned in this study , are described in Text S1 . Detailed methods and materials are described in Text S1 . Nuclear extraction kit ( Chemicon , http://www . chemicon . com ) was used for performing a nuclear extraction , and the active form of NF-κB contained in the nuclear extract was detected using an NF-κB p65 Transcription Factor assay system ( Chemicon ) . We employed the Noshift transcription factor assay system ( Novagen , http://www . emdbiosciences . com/html/NVG/home . html ) to assay binding of CREB to CRE oligonucleotides . BECs were cultured and exposed for 1 h to E . coli ORN103 ( pSH2 ) , and nuclear extracts were collected following the vendor's recommendation ( Novagen ) . To detect the binding of CREB to the CRE site of the IL-6 promoter , CRE oligonucleotides from IL-6 promoter region were synthesized and end-labeled by biotinylation . The CRE oligonucleotide sequences utilized were the same as those used previously [22] .
Although there are several data implicating type 1 fimbriae and its adhesive subunit , FimH , as the determinant largely responsible on UPEC for triggering endocytic responses from BECs [10 , 11] , a recent study has reported LPS as the primary determinant on UPEC responsible for evoking the cytokine response from BECs [23] . We initiated our studies by examining the role of LPS in mediating the IL-6 response of BECs following exposure to E . coli . This was undertaken by comparing the IL-6 response of the human BEC line 5637 to E . coli in the presence and absence of polymyxin B ( PMB ) , which binds to the lipid A portion of LPS and blocks its recognition by host cells [23] . The E . coli strain we selected for our studies was a well-characterized laboratory strain of E . coli ORN103 ( pSH2 ) expressing recombinant type 1 fimbriae , including the adhesive subunit , FimH . We employed this laboratory strain rather than a UPEC strain because UPEC strains express multiple genes capable of suppressing cytokine responses in BECs [14] . We observed a strong IL-6 response from BECs following exposure to the laboratory E . coli that was significantly reduced following pretreatment of the bacteria with PMB ( Figure 1A ) . For comparative purposes , shown in Figure 1A is the PMB-mediated inhibition of the IL-6 responses of BECs to soluble E . coli LPS . Due to the possibility of lipoprotein contamination of LPS prepared by trichloroacetic acid ( TCA ) or phenol-chloroform-petroleum ether ( PCP ) extraction , LPS ultra purified by ion exchange chromatography and verified to contain <1% protein was used in this study ( Sigma; E . coli 055:B5 LPS ) . To confirm that the LPS on E . coli was the primary determinant responsible for activating BECs , we sought to show that the activation of BECs involved TLR4 , the signaling receptor for LPS . Using RNA interference ( RNAi ) techniques , we generated BECs where expression of TLR4 was appreciably knocked down . Densitometric quantification of message levels in the knockdown ( KD ) BECs revealed that the expression of TLR4 was reduced by 49% ( Figure 1B ) . Shown in Figure 1C is the IL-6 response of control ( transfected with control vector ) BECs and of the KD BECs to E . coli and LPS . Compared to control BECs , significant reduction in the IL-6 response to both E . coli and LPS was observed with the KD cells ( Figure 1C ) . For the most part , the reduction in the IL-6 response paralleled the degree of KD of TLR4 in the BECs ( Figure 1B–1C ) . Taken together , these data confirm that LPS is the primary determinant on E . coli responsible for triggering the IL-6 response , and the intracellular signaling triggered by LPS involves TLR4 on BECs . Since intracellular Ca2+ ( [Ca2+]i ) has been implicated in important cellular processes , including IL-6 secretion [20 , 24] , we examined the involvement , if any , of this second messenger in the IL-6 response of BECs to E . coli . We investigated whether exposure of BECs to E . coli induced an increase in [Ca2+]i by performing ratiometric imaging on Fura-2/AM-loaded 5637 BECs . A unique pattern of Ca2+ influx into exposed BECs to E . coli was observed ( Figure 2A ) . BEC [Ca2+]i was constant before bacterial exposure and increased rapidly , within 1 min , after E . coli exposure , returning to baseline levels within 5 min ( Figure 2A ) . To determine whether the E . coli–induced increase of [Ca2+]i was essential for the BEC IL-6 response to bacterial exposure , we examined IL-6 secretion by BECs following bacterial exposure with or without pretreatment with NiCl2 , a general Ca2+ channel inhibitor [25] , or BAPTA-AM , a [Ca2+]i chelator [26] ( Figure 2B ) . Whereas BEC IL-6 secretion was readily induced after exposure to E . coli , pretreatment of BECs with NiCl2 or BAPTA-AM before bacterial exposure completely abolished IL-6 secretion by exposed BECs to E . coli . In addition , a general inducer of calcium influx , ionophore A23187 , was able to directly induce BEC IL-6 production in the absence of E . coli , demonstrating the importance of [Ca2+]i increases in initiating this response . We also investigated whether purified E . coli LPS was capable of inducing an increase in [Ca2+]i in BECs . LPS was seen to induce a similar , but delayed , increase in [Ca2+]i compared to that caused by E . coli ( Figure 2A and 2C ) . The LPS-induced [Ca2+]i peak occurred ~5 min after the addition of 100 μg/ml LPS . Disrupting the LPS-induced [Ca2+]i increase with NiCl2 or BAPTA-AM pretreatment before LPS exposure greatly reduced IL-6 production by BECs ( Figure 2D ) . Pretreatment of LPS with PMB almost completely abrogated the [Ca2+]i response of BEC ( Figure 2E ) . Taken together , these observations provide strong evidence indicating that the IL-6 response of BECs to E . coli involves a sharp increase in [Ca2+]i levels . Although LPS appears to be the primary bacterial component responsible for elevation of [Ca2+]i , bacteria-associated LPS evoked a faster [Ca2+]i response in BECs compared to that of soluble LPS . Intracellular cAMP is an important second messenger in several signaling pathways , including IL-6 response [27–29] . Exposing BECs to E . coli for 1 h demonstrated a 2 . 7-fold increase in intracellular cAMP , which was blocked by inhibiting AC activity with the compound MDL-12 , 330A ( MDL ) ( Figure 3A ) . The increase in intracellular cAMP following bacterial exposure was dependent on both bacteria-associated LPS and an increase in [Ca2+]i as shown , respectively , by pretreating the bacteria with PMB or pretreating the BECs with NiCl2 ( Figure 3B ) . This E . coli–induced [Ca2+]i-dependent cAMP production was found to be an important step in the cytokine response of BECs to bacterial exposure , since inhibition of ACs with MDL reduced BEC IL-6 expression by ~75% ( Figure 3C ) . In addition , a membrane-permeable cAMP analog , dibutyryl cAMP , induced a greater than 3-fold increase in BEC IL-6 production in the absence of bacterial exposure , demonstrating the importance of intracellular cAMP in inducing BEC IL-6 production . However , a membrane-permeable cAMP analog ( 8-CPT-cAMP ) that does not activate the classical cAMP-target protein , protein kinase A ( PKA ) , but only activates the recently discovered cAMP-target protein Epac ( exchange protein activated by cAMP ) [30] , did not induce the production of BEC IL-6 , indicating that PKA is involved in the downstream induction of IL-6 production by exposed BECs to E . coli ( Figure 3C ) . Forskolin activates ACs , the enzymes that produce intracellular cAMP , by a direct mechanism [31] , which should bypass the need for an increase in [Ca2+]i that is observed with E . coli–induced intracellular cAMP production . As shown in Figure 3D , direct activation of AC by forskolin led to a dramatic production of IL-6 that was not inhibited by NiCl2 , indicating that the increase in [Ca2+]i that occurred after E . coli exposure preceded the production of intracellular cAMP . Neither NiCl2 nor PMB treatment affected forskolin-induced IL-6 production , demonstrating that these agents did not have a detrimental effect on protein synthesis in general ( Figure 3D ) . Thus , the IL-6 response to E . coli evoked by BECs involves another secondary messenger , cAMP , which acts downstream of the Ca2+ response . Because there are currently ten known isoforms of mammalian ACs [32] it was of interest to determine which AC was responsible for the E . coli–induced increase of intracellular cAMP in BECs . First , we determined which AC isoforms were actually expressed in BECs . RT-PCR was performed on total cellular RNA , using primers specific for each known AC isoform and only mRNA for AC isoforms AC-3 , AC-4 , AC-6 , and AC-7 was detectable in BECs ( Figure 4A ) . We confirmed that the other AC isotype-specific primers used were functional by undertaking RT-PCR on total RNA from human embryonic kidney ( HEK ) cells , positive control cells , where all ACs except AC-4 and AC-8 were expressed [21] ( unpublished data ) . RNAi was utilized to minimize the expression of each AC , which was verified by AC isotype-specific RT-PCR ( Figure 4B ) . Following E . coli exposure , intracellular cAMP , as well as IL-6 secretion , rose significantly in all of the KDs , except for the KD of AC-3 , indicating that AC-3 is the BEC AC isoform linked to the IL-6 response following E . coli exposure ( Figure 4C and 4D ) . It is noteworthy that of the four AC isoforms expressed by BECs , only AC-3 is known to be activated by increases in [Ca2+]i [33 , 34] . The KD of AC-3 also abrogated the production of intracellular cAMP ( Figure 4E ) and expression of IL-6 ( Figure 4F ) following exposure of BECs to purified LPS . Interestingly , in the AC-3 KD BECs , forskolin-induced IL-6 expression was largely unaffected ( Figure 4F ) . The appreciable IL-6 response to forskolin suggested that a general increase in intracellular cAMP was sufficient to signal IL-6 secretion in BECs . The absence of any reduction in the IL-6 response to forskolin in AC-3 KD BECs is attributable to the presence of other isoforms of ACs in these cells , which were directly activated by forskolin . Remarkably , when AC-3–specific Western blotting was performed on BECs before and after bacterial exposure or exposure to purified LPS , no discernible increase in the expression of AC-3 protein was observed , indicating that an increase in activity , rather than expression , of AC-3 occurred following infection ( Figure 4G and 4H ) . Finally , when we examined for increase in [Ca2+]i in the AC-3 KD BECs following exposure to E . coli , we found that it was comparable to that seen in wild-type ( WT ) BECs ( Figure 4I ) , which is consistent with the idea that the rise in [Ca2+]i preceded any rise in intracellular cAMP . Next , we sought to connect the Ca2+- and cAMP-dependent signaling events described in this study to the classical NF-κB–associated signaling pathway mediated by TLR4 . To link cAMP to the classical pathway , we examined its effects on the translocation of the transcriptional factor NF-κB from the cytoplasm to the nucleus following bacterial exposure . Remarkably , when we examined control-transfected BECs and AC-3 KD BECs 1 h following exposure to E . coli ( which corresponds to the time we observed significant secondary messenger responses ) for nuclear translocation , we found little or no translocation of NF-κB in either cell type ( Figure 5A ) . However , when we increased the incubation time to 2 h following exposure to E . coli , we detected a marked increase in translocation of NF-κB in control-transfected BECs , and an identical increase was also seen in AC-3 KD BECs ( Figure 5B ) . This finding revealed that ( i ) the cAMP responses in BECs preceded the nuclear translocation of NF-κB by significant amounts of time , and ( ii ) these cAMP responses did not appear to impact the NF-κB–associated signaling pathway . These observations raised the intriguing possibility that secondary messengers such as cAMP may be acting via an independent pathway . To see whether the regulatory effect of cAMP on the IL-6 response was at the transcriptional level , we compared IL-6 mRNA levels in WT BECs , control-transfected BECs , and AC-3 KD BECs before and 1 h after exposure to E . coli . We observed a marked increase in IL-6 mRNA in WT BECs and control-transfected BECs but not in AC-3 KD BECs ( Figure 5C ) , indicating that the AC-3–mediated elevation in intracellular cAMP was regulating the IL-6 response at the transcriptional level . It is pertinent to also note the time frame of when these assays were undertaken . Here , mRNA for IL-6 was detected in control-transfected BECs as early as 1 h after exposure to E . coli ( Figure 5C ) . Considering that nuclear translocation of NF-κB was detectable only after 2 h ( Figure 5B ) , this cAMP-regulated pathway appears to be activated sooner than the classical pathway . Interestingly , when we examined AC-3 KD BECs for IL-6 mRNA 6 h after exposure to E . coli , we detected similar amounts of message in AC-3 KD BECs and control-transfected BECs ( Figure 5D ) , indicating that the classical NF-κB–mediated pathway was still functional in AC-3 KD BECs . Thus , the IL-6 response in BECs appears to originate from two distinct pathways: the NF-κB–associated pathway and a separate but speedier pathway involving Ca2+ and cAMP . One mechanism through which cAMP may directly affect transcription of IL-6 is by promoting phosphorylation of CREB , which binds to CRE in the IL-6 promoter region [35] . An increase in intracellular cAMP levels activates PKA , whose catalytic subunits enter the nucleus and phosphorylates CREB [36] . Upon phosphorylation , CREB promotes the recruitment of various transcriptional co-activators that promote transcription of target genes with consensus sites for CREB , such as IL-6 [37 , 38] . When we examined for CREB phosphorylation in BECs exposed to forskolin and calcium ionophore A23187 , two potent elevators of intracellular cAMP , we observed a marked increase in CREB phosphorylation on the Western blots ( Figure 5E ) consistent with the idea that CREB phosphorylation occurred following elevation of intracellular cAMP levels . To see if bacterial exposure also triggered phosphorylation of CREB , CREB protein from extracts of BECs before and after exposure to E . coli was probed for phosphorylation . Following E . coli exposure , we found an appreciable increase in phosphorylation of CREB in non-transfected BECs and control-transfected BECs , but not in AC-3 KD BECs ( Figure 5F ) . To extend these findings , we investigated the binding of CREB to the CRE sites on the IL-6 promoter region . Nuclear extracts of BECs were obtained after 1 h incubation with E . coli ORN103 ( pSH2 ) and then incubated with the biotinylated oligonucleotides corresponding to CRE on the IL-6 promoter . Binding of CREB to CRE was assessed by a colorimetric assay . We found that CREB bound to CRE oligonucleotides but not to a scrambled oligonucleotide sequence of identical length ( Figure 5G ) . Thus , cAMP appears to be modulating IL-6 responses through the binding of the phosphorylated transcriptional factor CREB to the CRE site on the IL-6 promoter region . Since there are other inflammatory mediators such as IL-1α , IL-1β , and IL-8 with consensus CRE sites in their promoter region that have known to be activated during urinary tract infection [14 , 39 , 40] , we examined whether production of any of these mediators was modulated by the cAMP/CREB pathway following exposure to E . coli . We compared message levels for IL-1α , IL-1β , and IL-8 in WT and AC-3 KD BECs after 1 h exposure to E . coli ORN103 ( pSH2 ) . We found that mRNA levels for IL-8 but not IL-1α or IL-1β appeared to be regulated by the TLR4/cAMP/CREB pathway . Thus , in addition to IL-6 , production of IL-8 by BECs appears to be under the regulation of the novel signaling pathway . Based on the evaluation of transcriptional messages , the IL-6 response of BECs is mediated by two separate signaling pathways with different expression kinetics . To verify this observation , we compared the kinetics of IL-6 secretion in WT and AC-3 KD BECs following exposure to E . coli . We found that whereas appreciable IL-6 secretion ( arbitrarily defined as 4-fold over unstimulated controls ) was observed as early as 6 h in WT BECs , a comparable amount of IL-6 was only produced in AC-3 KD BECs after about 9 h ( Figure 6A ) . By 12 h , however , the amounts of IL-6 secretion were not significantly different between both cell types , suggesting that the IL-6 responses of AC-3 KD BECs eventually caught up to that of the WT BECs ( Figure 6A ) . A similar profile was obtained when we substituted the laboratory E . coli strain with a UPEC strain , CI5 ( Figure 6B ) . To assess the relative contribution of the cAMP-mediated pathway to the IL-6 response of BECs , we examined E . coli–elicited IL-6 responses of WT BECs after selective inhibition of the NF-κB pathway with pyrrolidine dithiocarbamate [41] . This agent has been reported not to inhibit CREB activity [42] . We found that although the early kinetics of the IL-6 responses were not significantly different from untreated WT BECs , the amounts of IL-6 secreted , especially by the 12 h period of incubation , were significantly reduced . Thus , while the cAMP/CREB–mediated IL-6 response was an early one , the amounts of IL-6 generated by this pathway were significantly less than that produced by the classical NF-κB–associated pathway . Additionally , we examined whether the IL-6 responses to E . coli mediated by other human cells involved the two second messengers , Ca2+ and cAMP . Monolayers of the human monocytic cell line Mono Mac 6 , and the human bronchial epithelial cell line 16-HBE , were exposed to E . coli ORN103 ( pSH2 ) as before in the presence of inhibitors of either Ca2+ response or cAMP response , and IL-6 secretion was measured . We found that whereas both cell lines evoked appreciable IL-6 responses to E . coli , neither one of these responses were reduced by inhibitors of calcium ( NiCl2 ) or cAMP ( PKA inhibitor ) signaling ( Figure 6C and 6D ) . Thus , the two secondary messengers , Ca2+ and cAMP , appear to be important mediators of the IL-6 responses in BECs but not in other cell types . Since BECs express other TLRs such as TLR2 and TLR3 [23 , 43] , it was of interest to investigate whether known ligands for TLR2 and TLR3 also triggered the cAMP/CREB pathway . We assessed the phosphorylation levels of CREB before and 6 h after exposure to lipoteichoic acid ( TLR2 ligand ) or polyinosine-polycytidylic acid ( TLR3 ligand ) and found that both TLR ligands induced significant phosphorylation of CREB ( Figure 6E ) . Thus , the CREB pathway appears to be activated by TLRs other than TLR4 . Since the secondary messenger/CREB pathway was detectable only in immortalized human BECs , it was important to validate our observation of the existence of a cAMP/CREB pathway in primary human BECs . Therefore , we investigated whether freshly isolated and cultured human BECs would secrete IL-6 through a Ca2+- and cAMP-dependent mechanism when exposed to UPEC strain CI5 [16] . We cultured primary bladder cells obtained from fresh bladder biopsies as described previously [44] . These cells exhibited characteristics of primary BECs , including expression of uroplakin 1a , a marker of the asymmetrical unit membrane , the junctional complex protein ZO1 , as well as cytokeratin , all of which are hallmarks of terminal differentiation in bladder umbrella cells ( unpublished data ) . We observed that , after the stimulation with UPEC strain CI5 , primary BECs exhibited elevation in [Ca2+]i ( Figure 7A ) . The intensity of the response was lower than that observed with the laboratory strain of E . coli ( unpublished data ) . This is consistent with the fact that UPEC express multiple virulence factors , some of which exhibit disparate effects on [Ca2+]i levels [20] . As demonstrated previously with 5637 BECs , this response was significantly abrogated following pretreatment with the general Ca2+ channel blocker , NiCl2 ( Figure 7B ) . Significant elevation in intracellular cAMP levels and IL-6 secretion was observed in primary BECs following exposure to UPEC ( Figure 7C and 7D ) . Both the increase in intracellular cAMP levels and IL-6 secretion were inhibitable by NiCl2 , once again confirming the importance of [Ca2+]i to the BEC IL-6 response ( Figure 7C and 7D ) . Also shown in Figure 7D is the IL-6 response of these primary cells to the laboratory E . coli strain ORN103 . Notice that it is markedly higher than the response to the UPEC strain . Taken together , these findings support the notion that although the IL-6 response of primary BECs to UPEC is dampened , it involves a Ca2+- and cAMP-dependent mechanism .
The bladder and the upper urinary tract are typically sterile , which is attributable , at least in part , to the highly efficient immune system monitoring these sites . One of the principal effectors of immune surveillance is the epithelial cell lining the urinary tract [3 , 45] . In addition to serving as a barrier against urine , BECs function as first responders , mobilizing multiple innate immune responses against microorganisms . Mediating microbial recognition on the surfaces of epithelial cells are PRRs , which recognize specific microbial products and activate intracellular signaling events leading to secretion of various inflammatory and immunoregulatory cytokines [43] . That epithelial cells possess PRRs such as TLR4 and contribute to immune surveillance has only recently been recognized [23 , 46] . For a long time it was assumed that PRRs were exclusively found on immune cells of hematopoietic origin and , therefore , most of our current information regarding PRR-mediated signal transduction is largely based on these cells [1 , 47] . Although there is no conclusive data suggesting cell-specific TLR4 signaling , there have been suggestions that LPSs evoke intracellular signaling reactions in Kupffer cells [48] and tracheal epithelial cells [49] that are absent in polymorphonuclear leukocytes [50] . Here , we report the existence of a distinct TLR4-mediated signaling pathway leading to IL-6 secretion that is present in BECs but absent in other human cell types . This novel signaling pathway detected in BECs is independent of the classical pathway involving the transcriptional element NF-κB and contains two well-known secondary messengers , Ca2+ and cAMP , which mobilize a different transcriptional element , CREB . The existence of this pathway only became evident to us because of our focus on second messengers in TLR signaling and because our assays for IL-6 production were undertaken at earlier incubation periods than the more traditional 24–48 h incubation time points [7] . At the later incubation periods , the contribution of this novel pathway to the IL-6 response is superseded by the traditional NF-κB–mediated pathway . Evidence for the involvement of [Ca2+]i in the early BEC IL-6 response comes from the finding that a flux in [Ca2+]i was observed within a minute of exposure to E . coli , and inhibiting this flux with Ca2+ channel inhibitors or [Ca2+]i chelators inhibited the IL-6 response ( Figure 2B ) . Since a general inducer of calcium influx such as ionophore A23187 was able to induce IL-6 production from BECs even in the absence of E . coli , increase in [Ca2+]i appears sufficient to trigger the IL-6 response from BECs ( Figure 2B ) . Evidence for the role of cAMP in the IL-6 response comes from the finding that the IL-6 response to E . coli was closely associated with a 3-fold increase in intracellular levels of cAMP ( Figure 3A ) . In addition , inhibition of cAMP-generating ACs significantly reduced the IL-6 response of BECs to E . coli ( Figure 3C ) . As with the [Ca2+]i flux , merely enhancing intracellular levels of cAMP with a membrane-permeable cAMP analog induced significant IL-6 release from BECs even in the absence of E . coli ( Figure 3C ) . Thus , the two secondary messengers are sufficient , as well as necessary , for the early IL-6 secretion in BECs . That Ca2+ response preceded the cAMP production following bacterial stimulation was deduced from the findings that ( i ) the earliest detectable increase in intracellular cAMP levels was observed 15 min following exposure to E . coli ( unpublished data ) , whereas Ca2+ responses could be seen within a minute of bacterial exposure ( Figure 2A ) , and ( ii ) the Ca2+ flux following exposure to E . coli remained largely unaffected in AC-3 KD BECs , while the cAMP response was abrogated ( Figure 4I ) . Since enhancement of intracellular cAMP specifically required an increase in [Ca2+]i , we suspected that a Ca2+-inducible form of AC was responsible . BECs were found to express mRNA for AC-3 , AC-4 , AC-6 , and AC-7 , but only RNAi KD of AC-3 , a Ca2+-inducible AC isoform , inhibited E . coli–induced intracellular cAMP production and subsequent IL-6 expression . To identify where in the signaling cascade cAMP was exerting its effects , we compared IL-6 message levels in control and AC-3 KD BECs . The absence of IL-6 message in AC-3 KD BECs after 1 h of exposure to bacteria suggested that cAMP was regulating IL-6 production at the transcriptional level rather than at the levels of translation or cytokine secretion . Since translocation of NF-κB into the nucleus of BECs following exposure to E . coli was largely unaffected in the AC-3 KD BECs ( Figure 5B ) , cAMP may not be exerting its effect through altering NF-κB . Interestingly , one way that cAMP can directly promote expression of certain genes is to activate PKA , which translocates to the nucleus , where it phophorylates the transcriptional factor CREB [36] . Upon phosphorylation , CREB is believed to promote transcription of a number of genes , including IL-6 , IL-8 , IL-1α , and IL-1β , which possess consensus CRE sites on their promoter region [36 , 37 , 51–53] . Interestingly , in BECs , only IL-6 and IL-8 appear to be regulated by the cAMP/CREB pathway ( Figure 5 ) . That cAMP was modulating phosphorylation of CREB was evident from CREB phosphorylation in control BECs following exposure to E . coli but not in AC-3 KD BECs . Thus , taken together , our cumulative data reveal the existence of a distinct TLR4-activated signaling pathway in BECs involving Ca2+ , cAMP , and phosphorylated CREB . A diagrammatic representation of the proposed TLR4-initiated Ca2+- , cAMP- and CREB-dependent pathway , as well as the NF-κB pathway in BECs , is shown in Figure 8 . Although in the figure we have indicated that the cAMP/CREB pathway in BECs is activated by TLR4 , our data also suggest that TLR2 and TLR3 activation may also trigger this pathway ( Figure 6E ) . Our analysis of the kinetics of IL-6 secretion in WT and AC-3 KD BECs has revealed that this novel second messenger/CREB-mediated pathway is mediating a faster IL-6 response than the classical NF-κB–mediated pathway . Following exposure of BECs to E . coli ORN103 ( pSH2 ) , marked phosphorylation of CREB was observed at least 1 h before nuclear translocation of NF-κB was evident . Indeed , the earliest evidence of nuclear translocation of NF-κB in BECs following exposure to E . coli was at 2 h ( Figure 5B ) . Another piece of evidence implicating the second messenger/CREB pathway in a rapid and distinct IL-6 response was the observation that a message for IL-6 was detectable in control BECs 1 h following exposure to E . coli , whereas no message for IL-6 was seen in AC-3 KD BECs . However , by 6 h after the classical NF-κB signaling pathways had been activated , there was very little difference in the amounts of mRNA in both cell types ( Figure 5D ) . Consistent with this finding , the kinetics of IL-6 secretion by WT BECs and AC-3 KD BECs following exposure to E . coli ORN103 ( pSH2 ) revealed a 3-h lag in the latter's response , but by 12 h the amounts of IL-6 secreted were comparable ( Figure 6A ) . Thus , the rapid and vigorous inflammatory responses to infection typically observed in the urinary tract may be attributable , at least in part , to this distinct cAMP-dependent signaling pathway . The relevance of the early IL-6 response by BECs may be linked to their role as first responders . One of the primary cell types in the urinary tract reacting to BEC-generated IL-6 are also BECs . These cells possess IL-6 receptors [54] and this cytokine , acting in autocrine fashion , may trigger various antimicrobial responses , such as production of antimicrobial peptides [55] and mucins , [56] as well as promote exfoliation of infected BECs . The existence of multiple pathways in BECs for triggering IL-6 responses could be an adaptation to avoid inactivation by UPEC . Several recent studies have suggested that host-adapted pathogens possess the intrinsic capacity to block NF-κB activation in macrophages and cultured human epithelial cells through release of toxins or proteases [13 , 57–60] . Hunstad et al . have recently identified several genes ( rfa , rfb , and surA ) in UPEC that contribute to suppressing the cytokine responses of BECs [14] . Our observations that primary BECs ( Figure 7 ) evoked a more modest IL-6 response to clinical UPEC CI5 compared to the laboratory E . coli strain and that the cAMP/CREB–mediated IL-6 response to UPEC CI5 in BEC lines was not striking compared to E . coli ORN103 ( pSH2 ) ( Figure 6A and 6B ) could be manifestations of this phenomenon . Conceivably , depending on the nature of the BECs , the UPEC CI5 strain is able to partially diminish one or both of the two TLR4-mediated signaling pathways leading to IL-6 secretion . Finally , because of the rapid emergence of multi-resistance among UPEC isolates , there is mounting interest in the development of alternate antimicrobial strategies . One approach is to bolster innate immune defenses in the urinary tract either before or during infection . In this regard , our finding that small molecule enhancers of intracellular levels of Ca2+ and cAMP are sufficient to trigger early and vigorous cytokine responses from BECs is of interest . There are available many compounds capable of modulating the intracellular levels of both Ca2+ and cAMP [31 , 61–63] . Judicious application of some of these agents for the treatment and prevention of urinary tract infections is a possibility that will require further examination .
|
In spite of frequent cross contamination by bacteria from the gut , urinary tract infections are relatively infrequent . Although much of the credit goes to cells lining the urinary tract , such as bladder cells , how this is achieved remains unclear . Human bladder cells display , on their surfaces , special molecules called Toll-like receptors , which sense the presence of bacteria and trigger the cells to release a variety of chemicals called cytokines . Cytokines contribute to the recruitment of phagocytic cells from the blood to the site of infection to clear bacteria . In this paper , we reveal that the Toll-like receptor–initiated intracellular signals leading to the production of cytokines by bladder cells involve the same pathway seen in other cells , as well as an additional and more rapid signaling pathway . Rapid production of cytokines by bladder cells will facilitate early clearance of bacteria . Additionally , possession of multiple signaling pathways by bladder cells for producing cytokines is advantageous because bacteria that infect the urinary tract have the capability to suppress certain signaling events that lead to cytokine production by bladder cells .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases",
"immunology",
"homo",
"(human)"
] |
2007
|
A Novel TLR4-Mediated Signaling Pathway Leading to IL-6 Responses in Human Bladder Epithelial Cells
|
Mining gene expression profiles has proven valuable for identifying signatures serving as surrogates of cancer phenotypes . However , the similarities of such signatures across different cancer types have not been strong enough to conclude that they represent a universal biological mechanism shared among multiple cancer types . Here we present a computational method for generating signatures using an iterative process that converges to one of several precise attractors defining signatures representing biomolecular events , such as cell transdifferentiation or the presence of an amplicon . By analyzing rich gene expression datasets from different cancer types , we identified several such biomolecular events , some of which are universally present in all tested cancer types in nearly identical form . Although the method is unsupervised , we show that it often leads to attractors with strong phenotypic associations . We present several such multi-cancer attractors , focusing on three that are prominent and sharply defined in all cases: a mesenchymal transition attractor strongly associated with tumor stage , a mitotic chromosomal instability attractor strongly associated with tumor grade , and a lymphocyte-specific attractor .
Despite their type-specific features , cancers share some common traits , or “hallmarks , ” related to , e . g . , the abilities of some cancer cells to divide uncontrollably or to invade surrounding tissues [1] . Furthermore , it has been recognized that gene expression signatures resulting from analysis of cancer datasets can serve as surrogates of cancer phenotypes [2] . Therefore , it is reasonable to hypothesize that computational analysis of rich biomolecular cancer datasets may reveal signatures that are shared across many cancer types and are associated with specific cancer phenotypes . Such rich datasets become publicly available at an increasing rate from many sources , such as The Cancer Genome Atlas ( TCGA ) . However , attempts to identify any such robust “bioinformatic hallmarks” of cancer have so far been largely unsuccessful . Gene signatures may occasionally be found to exhibit similarities across multiple cancer types . However , to our knowledge no algorithm has ever produced a set of nearly identical signatures after independently and separately analyzing datasets from different cancer types . There are various ways by which modules of co-expressed genes can be identified from rich datasets , some of which may be within the context of regulatory network discovery [3] . Clustering approaches can classify a selected set of genes into subsets each of which contains mutually related genes . Related techniques can also be used to classify samples into cancer subtypes [4]–[6] , each characterized by a set of characteristic genes . One of the most powerful computational approaches for this task has been nonnegative matrix factorization ( NMF ) [7] combined with consensus clustering [8] , resulting in successful subtype identification in several types of cancer . The main objective addressed by techniques such as NMF is to reduce dimensionality by identifying a number of metagenes jointly representing the gene expression dataset as accurately as possible , in lieu of the whole set of individual genes . Each metagene in NMF is defined as a positive linear combination of the individual genes , so that its expression level is an accordingly weighted average of the expression levels of the individual genes . The identity of each resulting metagene is influenced by the presence of other metagenes within the objective of overall dimensionality reduction achieved by joint optimization . By contrast , if the aim is exclusively to identify metagenes as surrogates of biomolecular events , then a fully unconstrained algorithm should be devised , without any effort to achieve dimensionality reduction , classification , mutual exclusivity , orthogonality , regulatory interaction inference , etc . We can consider , for example , a hypothetical case in which we have found a cluster consisting of a number of co-expressed genes in a rich gene expression dataset . We may wish to scrutinize and “sharpen” this co-expression trying to identify the “heart” ( core ) of the genes that are most strongly co-expressed in that case . In the absence of a defining phenotype , we can continue applying an unsupervised methodology , as follows: We can first define a consensus metagene from the average expression levels of all genes in the cluster , and rank all the individual genes in terms of their association ( defined numerically by some form of correlation ) with that metagene . We can then replace the member genes of the cluster with an equal number of the top-ranked genes . Some of the original genes may naturally remain as members of the cluster , but some may be replaced , as this process will “attract” some other genes that are more strongly correlated with the cluster . We can now define a new metagene defined by the average expression levels of the genes in the newly defined cluster , and re-rank all the individual genes in terms of their association with that new metagene; and so on . It is intuitively reasonable to expect that this iterative process will eventually converge to a cluster that contains precisely the genes that are most associated with the metagene of the same cluster , so that any other individual genes will be less strongly associated with the metagene . We can think of this particular cluster defined by the convergence of this iterative process as an “attractor , ” i . e . , a module of co-expressed genes to which many other gene sets with close but not identical membership will converge using the same computational methodology . The above description represents a simplified conceptual introduction of the computational methodology presented in this paper . Rather than using the average of the expression values in gene clusters of a particular size , the “attractors” are metagenes defined as weighted averages of all genes where each individual gene has a nonnegative weight , just like the metagenes defined using NMF [7] . We found that , given a rich ( loosely defined as containing at least 200 samples ) dataset represented by a gene expression matrix , such metagenes can be naturally identified as stable and precise attractors using a simple iterative approach . We experimentally verified these nice convergence properties without any exception after trying numerous times the method described in this paper on such rich datasets . This methodology is totally unsupervised , as it does not make use of any phenotypic association . As we show in this paper , however , once identified , a metagene attractor is likely to be found associated with a phenotype . We found that several attractor metagenes are present in nearly identical form in multiple cancer types . This provides an additional opportunity to combine the powers of a large number of rich datasets to focus , at an even sharper level , on the core genes of the underlying mechanism . For example , this methodology can precisely point to the causal ( driver ) oncogenes within amplicons to be among very few candidate genes . Importantly , this can be done from rich gene expression data , which already exist in abundance , without making any use of sequencing data . We identified several attractors , each of which has the potential to lead to corresponding testable biological hypotheses after scrutinizing their top-ranked genes and finding a putative underlying mechanism . For the purposes of this paper we present the general methodology for the benefit of the research community together with a listing of the attractors in six datasets from three cancer types ( ovarian , colon , breast ) . Here , we focus on a few interesting cancer-associated attractors that we found present in multiple cancer types . Particular emphasis is given to what we consider to be three key “bioinformatic hallmarks” of cancer , related to the ability of cancer cells to invade surrounding tissues; to divide uncontrollably; and the ability of the organism to recruit the immune system to fight cancer: a tumor stage-associated mesenchymal transition attractor , a tumor grade-associated mitotic chromosomal instability ( mitotic CIN ) attractor , and a lymphocyte-specific attractor .
Given a nonnegative measure of pairwise association between genes and , we define an attractor metagene to be a linear combination of the individual genes with weights . The association measure is assumed to have minimum possible value 0 and maximum possible value 1 , so the same is true for the weights . It is also assumed to be scale-invariant , therefore it is not necessary for the weights to be normalized so that they add to 1 , and the metagenes can still be thought of as expressing a normalized weighted average of the expression levels of the individual genes . See Materials and Methods for the choice of the measure . According to this definition , the genes with the highest weights in an attractor metagene will have the highest association with the metagene ( and , by implication , they will tend to be highly associated among themselves ) and so they will often represent a biomolecular event reflected by the co-expression of these top genes . This can happen , e . g . , when a biological mechanism is activated , or when a copy number variation ( CNV ) , such as an amplicon , is present , in some of the samples included in the expression matrix . In the following we use the term “attractor” for simplicity to refer to an attractor metagene , and the term “top genes” to refer to the genes with the highest weights in the attractor . The definition of an attractor metagene can readily be generalized to include features other than gene expression , such as methylation values . It can also be used in datasets of any objects ( not necessarily genes ) characterized by any type of feature vectors , with applications in other disciplines , such as social and economic sciences . The computational problem of identifying attractor metagenes given an expression matrix can be addressed heuristically using a simple iterative process: Starting from a particular seed ( or “attractee” ) metagene , a new metagene is defined in which the new weights are . The same process is then repeated in the next iteration resulting in a new set of weights , and so forth . In all gene expression datasets that we tried we found that this process converges to a limited number of stable attractors . Each attractor is defined by a precise set of weights , which are reached with high accuracy typically within 10 or 20 iterations . This algorithmic behavior with nice convergence properties is not surprising , because if a metagene represents co-expressed genes , then the next iteration will naturally “attract” other similarly co-expressed genes , and so forth , until there are no other genes more associated with the top genes than those genes themselves . Furthermore , the set of the few genes with the highest weight are likely to represent the “heart” ( core ) of the underlying biomolecular event . In support of this concept , the association of any of the top-ranked individual genes with the attractor metagene is consistently and significantly higher than the pairwise association between any of these genes , suggesting that the set of these top genes jointly comprise a proxy representing a biomolecular event better than each of the individual genes would . Indeed , related versions of the signatures identified by attractors in this paper have been previously identified in various contexts in individual cancer types , often intermingled with additional genes . However , the contribution of our work is that these signatures are found as pan-cancer biomolecular events , sharply pointing to the underlying mechanism . Therefore the top genes of the attractors will be appropriate for being used as biomarkers or for understanding the underlying biology . For example , one of the attractors that we identified ( the “mitotic chromosomal instability” attractor , described below ) has previously been found in approximate forms among sets of genes described generally [9] as “proliferation” or “cell cycle related” markers , while the actual attractor points much more sharply to particular elements in the structure of the kinetochore-microtubule interface . A reasonable implementation of an “exhaustive” search of attractor metagenes is to start from each individual gene as a seed ( “attractee” ) assigning a weight of 1 to that gene , and 0 to all the other genes . Each gene participating in a particular co-expression event will then lead to the same attractor when used as attractee . The computational implementation of the algorithm is described in Materials and Methods . We note that a dual method can be used to identify attractor “metasamples” as representatives of subtypes , and we can also combine such metasamples with the attractor metagenes in various ways to achieve biclustering , but this topic is not examined in this paper . We analyzed six datasets , two from ovarian cancer , two from breast cancer and two from colon cancer ( Supplementary Text S1 ) . In each case , we identified general ( Supplementary Table S1 ) and genomically localized ( Supplementary Table S2 ) attractors and we found that many among them appear in similar forms in all six datasets using particular merging and ranking criteria in each case ( Materials and Methods and Supplementary Text S1 ) . Following are descriptions of some of our results , starting with the three strongest multi-cancer attractors . This attractor contains mostly epithelial-mesenchymal transition ( EMT ) -associated genes . Table 1 provides a listing of the top 100 genes based on their average mutual information ( Materials and Methods ) with their corresponding attractor metagenes . The consistency of the attractor is established by the fact ( Supplementary Table S1 ) that there are many genes ( COL5A2 , COL1A2 , SPARC , CTSK , FBN1 , VCAN , AEBP1 , SERPINF1 ) that are among the top 50 genes in the attractors of all six datasets . The corresponding P value is less than 10−7 by permutation test ( Materials and Methods ) . Similar results are found in other solid cancer types in all cases that we tried . This is a stage-associated attractor , in which the signature is significantly present only when a particular level of invasive stage , specific to each cancer type , has been reached . Supplementary Table S3 demonstrates this phenomenon in three cancer datasets from different types ( breast , ovarian and colon ) that were annotated with clinical staging information , by providing a listing of differentially expressed genes , ranked by fold change , when ductal carcinoma in situ ( DCIS ) progresses to invasive ductal carcinoma; ovarian cancer progresses to stage III; and colon cancer progresses to stage II . In all three cases , the attractor is highly enriched among the top genes . Specifically , among the top 100 differentially expressed genes , the number of attractor genes included in Table 1 is 47 in breast cancer , 42 in ovarian cancer and 37 in colon cancer . The corresponding P values are 2×10−93 , 4×10−80 and 8×10−78 , respectively . This attractor has been previously identified with remarkable accuracy as representing a particular kind of mesenchymal transition of cancer cells present in all types of solid cancers tested leading to a published list of top 64 genes [10] , [11] . This list was generated using a supervised algorithm using association with tumor stage . Indeed 52 of these top 64 genes also appear in Table 1 ( P<10−114 ) , and furthermore all top 19 genes of Table 1 are among the 64 . We found that most of the genes of the signature were expressed by the cancer cells themselves , and not by the surrounding stroma , at least in a neuroblastoma xenograft model that we tried [11] . We also found that the signature is associated with prolonged time to recurrence in glioblastoma [12] . Related versions of the same signature were previously found to be associated with resistance to neoadjuvant therapy in breast cancer [13] . These results are consistent with the finding that EMT induces cancer cells to acquire stem cell properties [14] . It has been hypothesized that EMT is a key mechanism for cancer cell invasiveness and motility [15]–[17] . The attractor , however , appears to represent a more general phenomenon of transdifferentiation present even in nonepithelial cancers such as neuroblastoma , glioblastoma and Ewing's sarcoma . Although similar signatures are often labeled as “stromal , ” because they contain many stromal markers such as α-SMA and fibroblast activation protein , the fact that most of the genes of the signature were expressed by xenografted cancer cells [11] , and not by mouse stromal cells , suggests that this particular attractor of coordinately expressed genes represents cancer cells having undergone a mesenchymal transition . The signature may indicate a non-fibroblastic transition , as occurs in glioblastoma , in which case collagen COL11A1 is not co-expressed with the other genes of the attractor . We have hypothesized that a full fibroblastic transition of the cancer cells occurs when cancer cells encounter adipocytes [11] , in which case they may well assume the duties of cancer-associated fibroblasts ( CAFs ) in some tumors [1] . In that case , the best proxy of the signature [10] is COL11A1 and the strongly co-expressed genes THBS2 and INHBA . Indeed , the 64 genes of the previously identified signature were found from multi-cancer analysis [10] as the genes whose expression is consistently most associated with that of COL11A1 . The only EMT-inducing transcription factor found upregulated in the xenograft model [11] is SNAI2 ( Slug ) , and it is also the one most associated with the signature in publicly available datasets . We also found that the microRNAs most highly associated with this attractor are miR-214 , miR-199a , and miR-199b . Interestingly , miR-214 and miR-199a were found to be jointly regulated by another EMT-inducing transcription factor , TWIST1 [18] . The expression of the mesenchymal transition attractor indicates that the tumor is actively invasive at the specific sample site , so its prognostic value is cancer type and stage specific . As an example , we analyzed an oral squamous cell carcinoma dataset deposited in the Gene Expression Omnibus ( GEO ) under accession number GSE25104 . The corresponding Kaplan-Meier survival curve ( P = 0 . 0066 ) is shown in Figure 1 . This attractor contains mostly kinetochore-associated genes . Table 2 provides a listing of the top 100 genes based on their average mutual information ( Materials and Methods ) with their corresponding attractor metagenes , starting from CENPA , which encodes for a histone H3-like centromeric protein . The consistency of the attractor is established by the fact ( Supplementary Table S1 ) that there are many genes ( CENPA , DLGAP5 , KIF2C , CCNB2 , MELK , CCNA2 , KIF20A , HJURP , NUSAP1 , BUB1 , TTK , KIF11 , NCAPH ) that are among the top 50 genes in the attractors of all six datasets . The corresponding P value is less than 10−7 by permutation test ( Materials and Methods ) . Similar results are found in other solid cancer types in all cases that we tried . Contrary to the stage-associated mesenchymal transition attractor , this is a grade-associated attractor , in which the signature is significantly present only when an intermediate level of tumor grade is reached . Supplementary Table S4 demonstrates this phenomenon in three cancer datasets from different types ( breast , ovarian and bladder ) that were annotated with tumor grade information , by providing a listing of differentially expressed genes , ranked by fold change , when grade G2 is reached . In all three cases , the attractor is highly enriched among the top genes . Specifically , among the top 100 differentially expressed genes , the number of attractor genes included Table 2 is 40 in breast cancer , 38 in ovarian cancer and 27 in colon cancer . The corresponding P values are 4×10−74 , 3×10−69 and 3×10−49 , respectively . Consistently , a similar “gene expression grade index” signature [19] was previously found differentially expressed between histologic grade 3 and histologic grade 1 breast cancer samples . Furthermore , that same signature [19] was found capable of reclassifying patients with histologic grade 2 tumors into two groups with high versus low risks of recurrence . This attractor is associated with chromosomal instability ( CIN ) , as evidenced from the fact that another similar gene set comprising a “signature of chromosomal instability” [20] was previously derived from multiple cancer datasets purely by identifying the genes that are most correlated with a measure of aneuploidy in tumor samples . This led to a 70-gene signature referred to as “CIN70 . ” Indeed 31 of these 70 genes appear in Table 2 ( P<10−53 ) . However , several top genes of the attractor , such as CENPA , DLGAP5 , KIF2C , BUB1 and CCNA2 are not present in the CIN70 list . Mitotic CIN is increasingly recognized [21] as a widespread multi-cancer phenomenon . The attractor is characterized by overexpression of kinetochore-associated genes , which is known [22] to induce CIN . Overexpression of several of the genes of the attractor , such as the top gene CENPA [23] , as well as MAD2L1 [24] and TPX2 [25] , has also been independently previously found associated with CIN . Included in the mitotic CIN attractor are key components of mitotic checkpoint signaling [26] , such as BUB1B , MAD2L1 ( aka MAD2 ) , CDC20 , and TTK ( aka MSP1 ) . Also among the genes in the attractor is MKI67 ( aka Ki-67 ) , which has been widely used as a proliferation rate marker in cancer . Among transcription factors , we found MYBL2 ( aka B-Myb ) and FOXM1 to be strongly associated with the attractor . They are already known to be sequentially recruited to promote late cell cycle gene expression [27] to prepare for mitosis . Inactivation of the retinoblastoma ( RB ) tumor suppressor promotes CIN [28] and the expression of the attractor signature . Indeed , a similar expression of a “proliferation gene cluster [29]” was found strongly associated with the human papillomavirus E7 oncogene , which abrogates RB protein function and activates E2F-regulated genes . Consistently , many among the genes of the attractor correspond to E2F pathway genes controlling cell division or proliferation . Among the E2F transcription factors , we found that E2F8 and E2F7 are most strongly associated with the attractor . This attractor consists mainly of lymphocyte-specific genes with prominent presence of CD53 , PTPRC , LAPTM5 , DOCK2 , LCP2 and IL10RA . It is strongly associated with the expression of microRNA miR-142 as well as with particular hypermethylated and hypomethylated gene signatures [30] . There is also significant overlap between the sets of hypomethylated and overexpressed genes , suggesting that their expression is triggered by hypomethylation . Gene set enrichment analysis reveals that the attractor is found enriched in genes known to be preferentially expressed in differentiation into lymphocytes [31] . Table 3 provides a listing of the top 100 genes of the lymphocyte-specific attractor based on their average mutual information ( Materials and Methods ) with their corresponding attractor metagenes . The gene membership of the attractor provides hints about the underlying immune mechanism , which could be valuable towards generating hypotheses for potential immunotherapies such as adoptive transfer of lymphocytes . For example , the presence of the signal-transducing LCP2 ( aka SLP-76 ) gene , together with the adaptor FYB ( aka ADAP ) , suggests the formation of the SLP-76-ADAP adaptor module , which is known to regulate lymphocyte co-stimulation mediated by integrin ITGB2 ( aka LFA-1 ) , another prominent gene in the attractor [32] . We found that each of the above three main attractors under particular conditions is highly prognostic in breast cancer by analysing the METABRIC discovery breast cancer dataset [33] which includes both expression as well as survival data . The scope of the algorithm identifying attractor metagenes is different from that of other unsupervised methods , which are usually aimed at identifying subtypes or mutually exclusive clusters . Nevertheless , it is interesting to find to what extent other algorithms can produce multiple cancer signatures each of which appears in nearly identical form across different types . We applied three widely used methods , k-means clustering , principal component analysis and hierarchical clustering on the six cancer datasets used in this paper . In all cases , we listed the top fifty genes in each cluster and applied the same clustering algorithm as in the main text to find common genes among them and group them together . The results are shown in Supplementary Text S2 and Supplementary Tables S5 , S6 , S7 . We found that , in all cases , these well-established methods cannot identify multiple universal metagenes common in all six tested datasets . A biomolecular event , whether it is present in multiple cancer types or it is cancer specific , can be represented by a “consensus attractor metagene” after analyzing multiple datasets . To generate such consensus attractors , we use genes that were profiled by at least three of the six datasets , then rank individual genes in terms of their average mutual information ( Materials and Methods ) with the corresponding attractor metagenes across all datasets in which the gene was profiled . For example , Figure 5 contains scatter plots from four different rich breast cancer datasets connecting the mitotic CIN and estrogen receptor attractors . It has previously been reported [45] that breast tumors with high chromosomal instability are predominantly of the estrogen receptor negative phenotype . Although these scatter plots cannot be used for precise conclusions , they do suggest in all cases that ER-negative tumors have high mitotic chromosomal instability ( or equivalently that low chromosomal instability implies that the tumor is ER-positive ) . The reverse relationship , however , is not as clear .
Gene expression analysis has resulted in several cancer types being further classified into subtypes labeled , e . g . as “mesenchymal” or “proliferative . ” Such characterizations , however , may sometimes simply reflect the presence of the mesenchymal transition attractor or the mitotic chromosomal instability attractor , respectively , in some of the analyzed samples . Similar subtype characterizations across cancer types often share several common genes , but the consistency of these similarities has not been significantly high . By contrast , using an unconstrained algorithm independent of subtype classification or dimensionality reduction , we identified several attractors exhibiting remarkable consistency across many cancer types , suggesting that each of them represents a precise biological phenomenon present in multiple cancers . We found that the mesenchymal transition attractor is significantly present only in samples whose stage designation has exceeded a threshold , but not in all of such samples . On the other hand , the absence of the mesenchymal transition attractor in a profiled high-stage sample ( or the absence of the mitotic chromosomal instability attractor in a profiled high-grade sample ) does not necessarily mean that the attractor is not present in other locations of the same tumor . Indeed , it is increasingly appreciated [46] that tumors are highly heterogeneous . Therefore it is possible for the same tumor to contain components , in which , e . g . , some are migratory having undergone mesenchymal transition , some other ones are highly proliferative , etc . If so , attempts for subtype classification based on one particular site in a sample may be confusing . Existing molecular marker products make use of multigene assays that have been derived from phenotypic associations in particular cancer types . For breast cancer , biomarkers such as Oncotype DX [47] and Mammaprint [48] contain several genes highly ranked in our attractors . For example , most of the genes used for the Oncotype DX breast cancer recurrence score directly converge to one of our identified attractors: MMP11 to the mesenchymal transition attractor; MKI67 ( aka Ki-67 ) , AURKA ( aka STK15 ) , BIRC5 ( aka Survivin ) , CCNB1 , and MYBL2 to the mitotic CIN attractor; CD68 to the lymphocyte-specific attractor; ERBB2 and GRB7 to the HER2 amplicon attractor; and ESR1 , SCUBE2 , PGR to the estrogen receptor attractor . We envision , instead , a multi-cancer biomarker product that will include detection of the level of expression of each of the key attractor metagenes . These levels would need to be combined in different ways in different cancer types , but each of the metagenes would indicate the same attribute and the contributions of each component will be cleanly distinguished . Even though molecular marker genes in some existing products are already separated into groups that are related to our attractor designation , any improvement in diagnostic , prognostic , or predictive accuracy resulting from better such group designation and better choice of genes in each group would be highly desirable . We hope that the identification of the attractors of cancer , as presented here , will be valuable in that regard .
We chose the association measure between genes to be a power function with exponent a of a normalized estimated information theoretic measure of the mutual information [49] with minimum value 0 and maximum value 1 ( see “Mutual information estimation” below; more sophisticated related association measures [50] can also be used , but computational complexity will be prohibitive ) . In other words , , in which the exponent can be any nonnegative number . The value of is set to zero if the Pearson correlation between the two genes is negative . Each iteration defines a new metagene in which the weight for gene is equal to where is the immediately preceding metagene . The process is repeated until the magnitude of the difference between two consecutive weight vectors is less than a threshold , which we chose to be equal to 10−7 . At one extreme , if is sufficiently large then each of the seeds will create its own single-gene attractor because all other genes will always have near-zero weights . In that case , the total number of attractors will be equal to the number of genes . At the other extreme , if is zero then all weights will remain equal to each other representing the average of all genes , so there will only be one attractor . The higher the value of , the “sharper” ( more focused on its top gene ) each attractor will be and the higher the total number of attractors will be . As the value of is gradually decreased , the attractor from a particular seed will transform itself , occasionally in a discontinuous manner , thus providing insight into potential related biological mechanisms . We empirically found that an appropriate choice of ( in the sense of maximizing the strength of the attractor , as defined below ) for general attractors is around 5 , in which case there will typically be approximately 50 to 150 resulting attractors , each resulting from many attractee genes . An alternative to the power function can be a sigmoid function with varying steepness , but we found that the consistency of the resulting attractors was worse in that case . As mentioned in the Introduction , an attractor metagene can also be interpreted as a set of the top genes of the attractor , i . e . , a gene set that includes only the genes that are significantly associated with the attractor . One empirical choice for such a gene set would be to include only the genes whose mutual information ( or the z-score thereof ) with the attractor metagene exceeds a particular threshold . In fact , the attractor finding algorithm itself can be designed to discover “attractor gene sets , ” without assigning weights to genes . In that case , metagenes are defined as simple averages of the genes in a gene set , and each iteration leads to a new gene set consisting of the new set of top-ranked genes in terms of their association with the previous metagene ( gene set sizes can be constant or adaptively changing in various ways ) . We found , however , that such a method has the disadvantage of occasionally leading to attractors with significant overlap , which requires additional post-processing steps . Identified attractors can be ranked in various ways . The “strength of an attractor” can be defined as the mutual information between the nth top gene of the attractor and the attractor metagene . Indeed , if this measure is high , this implies that at least the top n genes of the attractor are strongly co-expressed . We selected n = 10 as a reasonable choice , not too large , but sufficiently so to represent a real complex biological phenomenon of co-expression of at least ten genes . For amplicons , n = 5 is sufficient to ensure that the oncogenes are included in the co-expression ) . We use these choices when referring to the strength of an attractor . The top genes of many among the found attractors are genomically localized . In that case the biomolecular event that they represent is often the presence of a particular copy number variation . In the cancer datasets that we tried , this phenomenon almost always corresponds to a local amplification event known as an amplicon . We therefore also devised a related amplicon-finding algorithm , custom-designed to identify localized amplicon-representing attractor metagenes , described below . To identify genomically localized attractors – almost always amplicons – we use the same algorithm but for each seed gene we restrict the set of candidate attractor genes to only include those in the local genomic neighbourhood of the gene , and we optimize the exponent a so that the strength of the attractor is maximized . Specifically , we sort the genes in each chromosome in terms of their genomic location and we only consider the genes within a window of size 51 , i . e . , with 25 genes on each side of the seed gene . We further optimize the choice of the exponent for each seed , by allowing to range from 1 . 0 to 6 . 0 with step size of 0 . 5 and selecting the attractor with the highest strength . Because the set of allowed genes is different for each seed , the attractors will be different from each other , but “neighbouring” attractors will usually be very similar to each other . Therefore , following exhaustive attractor finding from each seed gene in a chromosome , we apply a filtering algorithm to only select the highest-strength attractor in each local genomic region , as follows: For each attractor , we rank all the genes in terms of their mutual information with the corresponding attractor metagene and we define the range of the attractor to be the chromosomal range of its top 15 genes . If there is any other attractor with overlapping range and higher strength , then the former attractor is filtered out . This filtering is done in parallel , so elimination of attractors occurs simultaneously . Assuming that the continuous expression levels of two genes and are governed by a joint probability density with corresponding marginal and , the mutual information is defined as the expected value of . It is a non-negative quantity representing the information that each one of the variables provides about the other . The pairwise mutual information has successfully been used as a general measure of the correlation between two random variables . We compute mutual information with a spline-based estimator [51] using six bins in each dimension . This method divides the observation space into equally spaced bins and blurs the boundaries between the bins with spline basis functions using third-order B-splines . We further normalize the estimated mutual information by dividing by the maximum of the estimated and , so the maximum possible value of is 1 . We used Level 3 data when directly available , and imputed missing values using a k-nearest-neighbour algorithm with k = 10 , as implemented in R [52] . We normalized the other datasets on the Affymetrix platform using the RMA algorithm as implemented in the affy package in Bioconductor [53] . To avoid biasing attractor convergence with multiple correlated probe sets of the same gene , we summarized the probe set-level expression values into the gene-level expression values by taking the mean of the expression values of probe sets for the same genes . We used the annotations for the probe sets given in the jetset package [54] . To investigate the associations between the attractor metagene expression and the tumor stage and grade , we used the following annotated gene expression datasets . For stage association: Breast ( GSE3893 ) , TCGA Ovarian , Colon ( GSE14333 ) . For grade association: Breast ( GSE3494 ) , TCGA Ovarian , Bladder ( GSE13507 ) . For Breast GSE3494 we used only the samples profiled by U133A arrays . For Breast GSE3893 we combined two platforms by taking the intersections of the probes in the U133A and the U133Plus 2 . 0 arrays . For datasets profiled by Affymetrix platforms all the datasets were normalized using the RMA algorithm . For Bladder GSE13507 normalization was done as provided in the GEO . P values for gene set enrichment were evaluated with the cumulative hypergeometric distribution using the total number of genes in each dataset . The significance of the consistency of the mesenchymal transition and mitotic CIN attractors was evaluated as follows: Supplementary Table S1 contains 210 gene sets from six cancer datasets . Each of the gene sets contains 50 genes . The mesenchymal transition metagene has eight genes ( COL5A2 , COL1A2 , SPARC , CTSK , FBN1 , VCAN , AEBP1 , SERPINF1 ) common across all six datasets . The mitotic CIN metagene has 13 common genes ( CENPA , DLGAP5 , KIF2C , CCNB2 , MELK , CCNA2 , KIF20A , HJURP , NUSAP1 , BUB1 , TTK , KIF11 , NCAPH ) across all six datasets . To evaluate the significance of the consistency across the six datasets , we randomly generated 210 gene sets with the same sizes as those in the Table . In other words , we randomly selected 50 genes out of the 11 , 395 common genes to generate a random gene set . We created 210 such random gene sets , and then assigned them to six different datasets according to the settings in the Table . We then performed the clustering algorithm described in Materials and Methods . Each time , we counted the maximum number of genes common in all six datasets , and we repeated this process ten million times . This constitutes a conservative way of evaluating consistency , in the sense that for each random experiment we only record the maximum number of common genes in the gene set cluster , and we created random gene sets using only the common genes in the six datasets . In these ten million experiments , it never occurred that more than one gene was common in all six datasets . Therefore , the corresponding P value for both the mesenchymal transition metagene as well as the mitotic CIN metagene is less than 10−7 , and is in fact much lower than that given the large number ( 8 and 13 respectively ) of the common genes .
|
Cancer is known to be characterized by several unifying biological capabilities or “hallmarks . ” However , attempts to computationally identify patterns , such as gene expression signatures , shared across many different cancer types have been largely unsuccessful . A typical approach has been to classify samples into mutually exclusive subtypes , each of which is characterized by a particular gene signature . Although occasional similarities of such signatures in different cancer types exist , these similarities have not been sufficiently strong to conclude that they reflect the same biological event . By contrast , we have developed a computational methodology that has identified some signatures of co-expressed genes exhibiting remarkable similarity across many different cancer types . These signatures appear as stable “attractors” of an iterative computational procedure that tends to collect mutually associated genes , so that its convergence can point to the core ( “heart” ) of the underlying biological co-expression mechanism . One of these “pan-cancer” attractors corresponds to a transdifferentiation of cancer cells empowering them with invasiveness and motility . Another represents a mitotic chromosomal instability of cancer cells . A third attractor is lymphocyte-specific .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"systems",
"biology",
"biology",
"computational",
"biology",
"genetics",
"and",
"genomics"
] |
2013
|
Biomolecular Events in Cancer Revealed by Attractor Metagenes
|
Lymphatic filariasis , commonly known as elephantiasis , is a painful and profoundly disfiguring disease . Wuchreria bancrofti ( Wb ) is responsible for >90% of infections and the remainder are caused by Brugia spp . Mosquitoes of the genera Culex ( in urban and semi-urban areas ) , Anopheles ( in rural areas of Africa and elsewhere ) , and Aedes ( in Pacific islands ) are the major vectors of W . bancrofti . A preventive chemotherapy called mass drug administration ( MDA ) , including albendazole with ivermectin or diethylcarbamazine citrate ( DEC ) is used in endemic areas . Vector control strategies such as residual insecticide spraying and long-lasting insecticidal nets are supplemental to the core strategy of MDA to enhance elimination efforts . However , increasing insecticide resistance in mosquitoes and drug resistance in parasite limit the effectiveness of existing interventions , and new measures are needed for mosquito population control and disruption of mosquito-parasite interactions to reduce transmission . Mosquito insulin signaling regulates nutrient metabolism and has been implicated in reduced prevalence and intensity of malaria parasite , Plasmodium falciparum , infection in mosquitoes . Currently no data are available to assess how insulin signaling in mosquitoes affects the development of multi-cellular parasites , such as filarial nematodes . Here , we show that insulin receptor knockdown in blood fed C . quinquefasciatus , the major vector of Wb in India , completely blocks the development of filarial nematode parasite to the infective L3 stage , and results in decreased ecdysteroid production and trypsin activity leading to fewer mosquito eggs . These data indicate that a functional mosquito insulin receptor ( IR ) is necessary for filarial parasite development and mosquito reproduction . Therefore , insulin signaling may represent a new target for the development of vector control or parasite blocking strategies .
Lymphatic filariasis ( LF ) is one of the neglected tropical diseases and a major cause of permanent and long-term disability worldwide . The disease is caused by three species of nematode worms ( filariae ) –Wuchereria bancrofti ( Wb ) , Brugia malayi , and Brugia timori , which are transmitted by several mosquito species within the genera Culex , Anopheles , Aedes , and Mansonia . Humans are the exclusive host of infection with W . bancrofti , complicating efforts to study this parasite in the laboratory . The principal vectors in endemic areas are dictated by the habitat and geographical range of competent mosquito species . For instance , the southern house mosquito , Culex quinquefasciatus , is the major vector of Wb in India . India constitutes approximately 45% of the world’s LF burden where ~550 million people are at risk of infection , with 59 million infected and of which 19 . 6 million exhibit filariasis symptoms [1–2] . Worldwide , an estimated 120 million people across 55 countries are infected with Wb , leading to a loss of 5 . 9 million disability-adjusted life-years [3] . The availability of safe , single-dose , drug treatment regimens capable of suppressing microfilariae to very low levels has resulted in targeting of this mosquito-borne disease for global elimination [4] . The Global Program to Eliminate Lymphatic Filariasis was launched in 2000 with the principal objective of breaking cycles of transmission of Wb and Brugia spp . through annual mass drug administrations ( MDAs ) to entire at-risk populations . In India , a MDA program has been in effect since 2004 in districts where filariasis is endemic [2] . However , emerging challenges to this approach have raised questions regarding the effectiveness of using MDA alone to eliminate LF without the inclusion of supplementary vector control . The World Health Organization is now considering vector control as a critical component of LF reduction [5] . However , current strategies of vector control have their own limitations and challenges , such as increasing insecticide resistance among vectors , and new measures are needed for mosquito population control and disruption of mosquito-filarial worm interactions to reduce transmission [6] . As part of the transmission cycle , mosquitoes first ingest microfilariae ( mf ) in blood taken from infected human hosts . The ingested mf penetrate the midgut epithelium and move into the hemolymph within 1–2 h post blood meal where they molt to first instar larvae and invade the flight muscles . After two molts , the third-stage infective larvae ( L3 ) exit the flight muscles and lodge in the head cavity , from which they escape when the mosquito takes another blood meal [7] ( Fig 1 ) . Biological transmission of filarial worms is classified as cyclodevelopmental because the parasite undergoes development within the vector to become infective to the vertebrate host , but does not multiply [8–9] . Their successful transmission to another human host is entirely dependent on the ability of the mosquito vector to acquire and mobilize nutrients that sustain parasite growth and enable tissue repair and survival for the 10–14 days required for filarial nematode development [8] . Previous studies showed that mosquito flight muscle cells become devoid of glycogen granules following infection with Brugia spp . parasites [10–11] , suggesting carbohydrate metabolism is essential for nematode development . Carbohydrate metabolism and reproduction in female mosquitoes are regulated by insulin signaling . Mosquitoes , like most other insects , encode multiple insulin like peptides ( ILPs ) but only one receptor tyrosine kinase homolog of the insulin receptor ( IR ) [12] . Blood and sugar feeding by the yellow fever mosquito , Aedes aegypti , triggers medial neurosecretory cells in the brain to release ILPs into the hemolymph that bind to the IR and activate the insulin signaling pathway in different tissues [13–15] . Female Ae . aegypti decapitated within 1 h post sugar meal , to prevent release of endogenous ILPs , fail to store circulating carbohydrates , whereas a single dose of ILP3 directs circulating carbohydrates into glycogen and lipid stores to the same levels measured in intact females [13] . In blood fed Ae . aegypti mosquitoes , IR knockdown negatively impacts fecundity [15] , immune response [16] , and delays blood digestion in the midgut [15] . Some mosquito-borne pathogens are negatively affected by altered insulin signaling in mosquitoes . For example , IR knockdown in the African malaria mosquito , Anopheles gambiae , resulted in lower Plasmodium berghei loads [17] . Knockdown of two ILPs , ILP3 and ILP4 , in the Asian malaria mosquito , An . stephensi , also resulted in reduced prevalence and intensity of P . falciparum infection [18] . In transgenic An . stephensi , overexpression of Akt , a nexus of insulin signaling , in the midgut of heterozygous mosquitoes resulted in 60–99% reduction in the numbers of mosquitoes infected with P . falciparum , and parasite infection was completely blocked in homozygous transgenic mosquitoes [19] . Currently no data are available to assess how insulin signaling in mosquitoes affects the development of multi-cellular parasites , such as filarial nematodes . We hypothesized that insulin signaling maintains nutrient homeostasis in blood fed , Wb-infected C . quinquefasciatus , which aids egg and filarial nematode development , and that mosquito IR ( CqIR ) knockdown would disrupt these processes . Therefore , the aim of this study was to determine the effect of CqIR knockdown on the development of Wb and the reproductive physiology of C . quinquefasciatus .
This study was carried out in strict accordance with the recommendation of the Government of India institutional review board ( IRB ) for blood collection from patients . Venous blood was collected from consenting infected human volunteers ( adults over 18 years , both sexes ) in the Karimnagar district , Andhra Pradesh . The study was explained to the volunteers and oral consent was obtained . Written consent could not be obtained because the night clinic was in a rural area and most adults were uneducated and were unable to read and write . The IRB approved the use of oral consent , and the oral consent was documented by the health department authorities . The document contained name of the volunteer , father or husband’s name ( head of the household ) , date of birth , signs and symptoms of elephantiasis , prior MDA obtained , and any sickness four weeks prior to the night clinic . The information was read and signed by the village head . A night clinic was established so blood could be collected between 8–11 pm local time ( when the parasite is active in blood circulation ) into K2EDTA vacutainers ( BD Vacutainer , Gurgaon , India ) . Blood samples were kept on ice , and the number of viable mf was counted the next morning with Jaswant Singh and Bhattacharjee stain II ( JSB II ) [20] . Pupae of C . quinquefasciatus were field collected from a water retention pond in Hyderabad , India and identified to species [21] . Adults reared from these pupae were blood fed on the forearm of a consenting adult for egg production . Hatched larvae were fed ground Tetramin flakes . All stages were maintained at 12 h:12 h day-light cycle at 28°C and 80% relative humidity . Adults were fed on 10% sucrose solution ad libitum . Mosquitoes were raised for two generations in the laboratory before conducting experiments . Double-stranded RNA ( dsRNA ) for C . quinquefasciatus IR ( CPIJ005469 ) and a non-specific gene not present in C . quinquefasciatus , enhanced green florescence protein ( eGFP ) , was synthesized as described previously [15] . The CqIR PCR product generated with T7CqIRFwd-5’TAATACGACTCACTATAGGGAGCAGTTCAACACGAACCACGTC3’ and T7CqIRRev- 5’TAATACGACTCACTATAGGGCGTCATGCCAAAGTCGCCGAT3’ was used as a template for in vitro transcription , following the manufacturers’ protocol ( T7 Mega Script kit , Ambion ) . 50 females ( 24 h old ) were injected with either CqIR dsRNA or control eGFP dsRNA ( 2 μg/ 0 . 5 μl ) . Mosquitoes were kept on 10% sucrose solution for 5 days for recovery and on day 6 were fed mf-infected or uninfected blood through glass-jacketed artificial membrane feeders attached to a circulating water bath ( 37°C ) . Feedings were carried out in the evening after sunset . The feeders were covered with a black blanket to simulate night . Hemotek feeding membrane ( Hemotek ) was kept on human ankles under socks for a minimum of 2 hrs to absorb human odor and the authors exhaled into the cages periodically to provide CO2 stimulation . Total RNA was extracted from head , midgut , ovaries , and abdomen at 1 , 7 , and 10 days post blood meal using Trizol reagent ( Invitrogen ) following the manufacturer’s protocol , DNase ( Invitrogen ) treated , and re-purified with Trizol reagent . DNase-treated total RNA ( 1 μg ) was used for cDNA synthesis ( iScript cDNA synthesis kit , BioRad ) . All cDNA samples were diluted 10-fold and used as template for RT-PCR with primers for CqIRFwd- 5’GAACCACGTCGTCCGACTGCTCG3’ and CqIRRev- 5’TCGTCATGCCAAAGTCGCCGAT 3′ . RT-PCR was carried out under the following conditions: initial denaturation at 95° for 5 min , denaturation at 95° for 20 sec , annealing at 56° for 30 sec and extension at 72° for 30 sec , final extension at 72° for 10 min . Primers for Ae . aegypti actin were used as a control [15] . Five mosquitoes per treatment group ( CqIR and eGFP dsRNA injected ) were dissected 2–4 h post blood meal ( PBM ) to confirm the infectivity by presence of mf either in the gut lumen or in the hemolymph by microscope . Thirteen days PBM , an additional 6–10 mosquitoes per treatment were dissected to count mature L3 in thorax and head cavity . Experiments were repeated twice with different cohorts of mosquitoes fed infected blood . Data were analyzed by ANOVA with GraphPad software ( GraphPad Software Inc . La Jolla , CA ) . To examine fertility , five blood fed females per treatment per cohort were kept individually in small cages lined with moist paper towel . The number of eggs deposited by each female was counted 5 days PBM . To examine blood digestion , trypsin-like activity was measured in midguts dissected from a minimum of three females per treatment ( uninfected +eGFP dsRNA injected , infected+ eGFP dsRNA injected , infected + CqIR dsRNA injected ) at 24 and 48 h PBM as described previously [15] . Each treatment was replicated twice . Briefly , each midgut was transferred to 100 μl of 20 mM Tris ( pH 8 . 0 with 20 mM CaCl2 ) , sonicated , and then centrifuged ( 14 , 000 x g for 2 min ) . Midgut supernatants were frozen ( −80°C ) , and aliquots ( 0 . 05 equivalent ) were added to 100 μl of 4 mM Nα Benzoyl-L- Arginine-p-Nitroanilide ( BApNA ) for 10 min followed by measurement of absorbance at 405 nm using a BioTek plate reader . Activity was quantified based on trypsin standards ( bovine pancreas , Sigma T1426 ) . Ovaries collected from the aforementioned females were used for the in vitro assay of ecdysteroid production . Ovaries ( three pairs / 60 μl in a 1 . 7 ml polypropylene tube cap ) in duplicate ( 3 pairs per cap; 2 caps ) were dissected from different treatment groups ( uninfected + eGFP dsRNA injected; infected + eGFP dsRNA injected; infected + CqIR dsRNA injected ) at 24 and 48 h PBM and incubated in buffered saline [13] for 6 h at 27°C . Media ( 50 μl per cap ) were collected and stored at −80°C until assayed for ecdysteroid content by radioimmunoassay ( RIA ) , as described in Brown et al . [13] . Experiments were replicated twice ( each with two sets of three ovaries ) using different female cohorts . Mortality was also recorded daily throughout the experiment . Data were combined from both biological replicates and analyzed by ANOVA with GraphPad software . Culex quinquefasciatus Insulin Receptor , CPIJ005469
Reduced expression of the CqIR in C . quinquefasciatus tissues after dsRNA injection was confirmed by RT-PCR ( Fig 2A ) . CqIR transcription was silenced for up to 10 days PBM ( Fig 2A ) in head and midgut . In ovaries and abdomen ( including the fat body ) , CqIR transcript levels recovered 7 days PBM . CqIR knockdown had no impact on mf ingestion and movement from midgut to hemolymph of the mosquitoes ( Fig 2B ) . By 4 h PBM , most mf migrated to hemolymph from the midgut and were visibly motile in both control eGFP and CqIR dsRNA injected females . At day 13 , eGFP control mosquitoes had L3 in the head cavity with the exception of one L3 , which was still in the thorax . In CqIR knockdown mosquitoes , there were no L2 or infective L3 stage either in the thorax or in the head cavity ( Fig 2C ) . No melanized larvae were observed in the hemolymph or in the flight muscles of the thorax in the mosquitoes of either group . Given the collecting and rearing circumstances , limited numbers of C . quinquefasciatus females were available . In addition , only half or fewer of the CqIR dsRNA injected females were attracted to the artificial blood feeders . For these reasons , we fed only mf-infected blood to the CqIR dsRNA injected females to maximize the sample size for the following tissue assays . Trypsin enzyme activity was previously characterized in the midgut of blood fed C . quinquefasciatus [22] . It rises at ~4h PBM , peaks around 30–36 h PBM and then decreases until 72 h PBM . We assayed midguts at 24h PBM and found that eGFP control and un-infected mosquitoes had the highest level of trypsin activity . Parasite infection alone significantly decreased trypsin activity . Infected CqIR knockdown mosquitoes had the lowest trypsin activity . At 48 h PBM , trypsin activity was virtually undetectable in all three treatment groups ( Fig 3 ) . For blood feeding C . pipiens mosquitoes , ovary ecdysteroid production and hemolymph titers increase within a few hours PBM , peak 24–36 h later , and then decline rapidly [23] . We dissected and incubated ovaries in vitro from blood fed eGFP control and CqIR knockdown females to determine ecdysteroid production at 24 and 48h PBM . Wb infection did not affect ecdysteroid production by ovaries as comparable amounts were produced by ovaries from eGFP control females given an uninfected blood meal ( Fig 4 ) . However , ovaries from CqIR dsRNA injected females produced a significantly lower amount of ecdysteroids . Ecdysteroid production by ovaries taken from eGFP control ( uninfected ) females at 48 h PBM was variable–the amount in one sample was equal to that produced by ovaries at 24 h PBM , but the others were significantly lower . Control , uninfected eGFP females deposited an average of 92 eggs per female within 96 h PBM , whereas those infected with Wb produced an average of 58 eggs per female; 37% fewer . No eggs were deposited by CqIR dsRNA injected mf-infected females ( Fig 5 ) .
In this study , we provide insights into the role of insulin signaling in the reproductive and digestive physiology of C . quinquefasciatus , and how this influences the development of Wb , an important pathogen transmitted by this species . In eGFP control and CqIR dsRNA injected females given an infected blood meal , the mf moved out of the midgut and into the hemocoel within 2 h . Previous studies also suggested mf moved into the hemocoel within the first 6 h , with the majority ( ~90% ) in the first 2 h PBM [7 , 8] . There was no effect of CqIR dsRNA injection on mf ingestion or penetrating the gut epithelium to move to the hemocoel ( Fig 2A ) . However , at 13 days PBM , the CqIR dsRNA injected mosquitoes had no infective L3 larvae either in the thorax or the head cavity ( Fig 2B ) . Previous studies found that LF parasites elicit an immune response , e . g . , melanotic encapsulation , in the non-vector model species , Ae . aegypti ( Liverpool strain ) , but go undetected in natural vectors , such as C . quinquefasciatus [24 , 25] . We did not observe worm encapsulation or melanization in the CqIR dsRNA injected mosquitoes . The absence of L3 Wb in the CqIR dsRNA injected mosquitoes may be due to the death of L1 worms prior to reaching or entering in thoracic muscles or failure to develop to the L2 stage because of limited nutrients resulting from decreased trypsin activity during blood digestion . Dead filarial worms are easily overlooked , therefore histochemical stains will be used in follow-up studies to pinpoint exactly where and when Wb mortality occurs in CqIR dsRNA injected mosquitoes to better understand possible causes . Our data demonstrated that insulin signaling plays an important role in the activation of trypsin-like enzyme secretion in the midgut and ecdysteroid production by ovaries in Culex mosquitoes in a manner similar to that reported for Ae . aegypti . Ingestion of a blood meal by Ae . aegypti results in the biphasic release of trypsin-like enzymes by the midgut [26 , 27] . Late phase trypsin-like activity occurs between 12 and 30 h PBM [27] , and digests most of the blood meal . We previously showed that insulin and Target of Rapamycin ( TOR ) signaling interact to regulate the timing of late phase trypsin-like gene expression ( late trypsin and serine protease VI ) and blood meal digestion in Ae . aegypti [15] . In this study , Wb infected eGFP control mosquitoes had significantly lower trypsin activity compared to eGFP control un-infected mosquitoes , and CqIR knockdown and infected females had even lower trypsin activity ( Figs 3 and 5 ) . Similarly , Ae . aegypti infected with B . malayi had lower transcript levels of four serine proteases and sterol trafficking genes suggesting altered blood digestion/proteolysis in the presence of filarial parasites [8] . We suspect that digestion might be influenced because of increased immune response . When infected with parasite , the mosquito may be allocating more resources towards fighting off the infection instead of blood digestion . Aedes mosquitoes’ flight muscle cells become devoid of glycogen granules following infection with Brugia parasites [10 , 11] suggesting stored carbohydrates are mobilized during parasite development . Insulin signaling regulates carbohydrate and lipid storage in sugar-fed Ae . aegypti , as shown by IR knockdown [13] . Although we did not measure the lipid and carbohydrate levels in C . quinquefasciatus in this study , we noticed that IR knockdown females had less fat deposits than the control ( personal observation ) and the complete block of parasite development in CqIR knockdown females may in part be due to lower nutrient stores available to the developing nematode . Nutrients are limiting factors for both egg production by mosquito vectors and development of infective pathogens . The total number of eggs laid by female mosquitoes depends upon the mobilization of nutrients stored from the feeding larval stage or previous blood meals and those released by blood digestion [28] . A blood meal also stimulates ovaries to produce ecdysteroids , mediated in part by ILPs [29] , and subsequently both amino acids from the blood meal and ecdysteroids activate secretion of vitellogenin and other yolk proteins by the fat body for provisioning eggs to be used by developing embryos . Brugia and Dirofilaria nematode worms have an adverse effect on mosquito fecundity , with up to a 33% reduction in egg production [30] . A recent study of Ae . aegypti infected with B . malayi also supported these earlier findings [31] . Here we found that the ovaries of blood fed , Wb infected , eGFP dsRNA injected mosquitoes produced similar amounts of ecdysteroids in vitro as that of the uninfected controls ( Fig 4 ) . In contrast , the ovaries of CqIR knockdown females produced significantly lower amounts of ecdysteroids than that of the eGFP controls . Therefore , the complete shutdown of egg production in CqIR knockdown and infected mosquitoes likely was due to both lower levels of ecdysteroids and nutrients from blood digestion . Most studies on mosquito-filarial worm interactions are based on the Ae . aegypti- B . malayi system . Although relative ease of use in the laboratory for this system provides advantages , Ae . aegypti is not a natural vector of filarial worms [32] . The Ae . aegypti Liverpool strain was genetically selected for susceptibility to many pathogens , including B . malayi . In contrast , the results of our study were obtained with a native C . quinquefasciatus-Wb association , thus providing a better assessment of the role of insulin signaling on mosquito physiology and parasite worm development in the field . To further understand the mechanism of Wb developmental arrest or mortality in IR knockdown mosquitoes , we will examine the global metabolome and transcriptome in order to pinpoint the metabolites or gene products needed for the parasite to develop .
|
Lymphatic filariasis ( LF ) is caused by infection with nematodes of the family Filarioidea . 90% of infections are caused by Wuchereria bancrofti and the remainder by Brugia spp . In endemic countries , LF has a major social and economic impact with an estimated annual loss of $1 billion . Filarial infection can cause a variety of clinical manifestations , including lymphoedema of the limbs , genital disease ( hydrocele , and swelling of the scrotum and penis ) and recurrent acute attacks , which are extremely painful and are accompanied by fever . As one of the leading causes of global disability , LF accounts for at least 2 . 8 million disability-adjusted life year ( DALY ) . Mass drug administration ( MDA ) is used prophylactically on the community level where the infection is present to decrease disease transmission . These drugs have limited effect on adult parasites but effectively reduce microfilariae in the bloodstream and prevent the spread of microfilaria to mosquitoes . Use of mosquito population control strategies is supplemental to the core strategy of MDA . However , increasing insecticide resistance in mosquitoes and drug resistant nematode parasites are complicating elimination efforts and emphasizes the need for novel interventions for vector control and parasite transmission . Insulin signaling is a highly conserved signaling pathway that regulates growth and nutrient homeostasis in animals . Our previous work in Aedes aegypti mosquitoes showed additional roles of insulin receptor signaling in blood digestion and reproduction . The present data strongly supports our previous findings in a different mosquito species and further explores the role of mosquito insulin receptor in the development of the filarial nematode to the infective stage . This information is pertinent to ongoing efforts to control and eradicate filariasis because insulin signaling may represent a new target for the development of vector control or transmission blocking strategies .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"reproductive",
"system",
"body",
"fluids",
"enzymes",
"enzymology",
"parasitic",
"diseases",
"animals",
"nematode",
"infections",
"endocrine",
"physiology",
"serine",
"proteases",
"infectious",
"disease",
"control",
"insect",
"vectors",
"infectious",
"diseases",
"proteins",
"endocrinology",
"disease",
"vectors",
"insects",
"arthropoda",
"ovaries",
"biochemistry",
"mosquitoes",
"eukaryota",
"blood",
"anatomy",
"physiology",
"biology",
"and",
"life",
"sciences",
"proteases",
"species",
"interactions",
"insulin",
"signaling",
"organisms"
] |
2018
|
Insulin receptor knockdown blocks filarial parasite development and alters egg production in the southern house mosquito, Culex quinquefasciatus
|
All class I phosphoinositide 3-kinases ( PI3Ks ) associate tightly with regulatory subunits through interactions that have been thought to be constitutive . PI3Kγ is key to the regulation of immune cell responses activated by G protein-coupled receptors ( GPCRs ) . Remarkably we find that PKCβ phosphorylates Ser582 in the helical domain of the PI3Kγ catalytic subunit p110γ in response to clustering of the high-affinity IgE receptor ( FcεRI ) and/or store-operated Ca2+- influx in mast cells . Phosphorylation of p110γ correlates with the release of the p84 PI3Kγ adapter subunit from the p84-p110γ complex . Ser582 phospho-mimicking mutants show increased p110γ activity and a reduced binding to the p84 adapter subunit . As functional p84-p110γ is key to GPCR-mediated p110γ signaling , this suggests that PKCβ-mediated p110γ phosphorylation disconnects PI3Kγ from its canonical inputs from trimeric G proteins , and enables p110γ to operate downstream of Ca2+ and PKCβ . Hydrogen deuterium exchange mass spectrometry shows that the p84 adaptor subunit interacts with the p110γ helical domain , and reveals an unexpected mechanism of PI3Kγ regulation . Our data show that the interaction of p110γ with its adapter subunit is vulnerable to phosphorylation , and outline a novel level of PI3K control .
Class I phosphoinositide 3-kinases ( PI3Ks ) produce the lipid second messenger phosphatidylinositol ( 3 , 4 , 5 ) -trisphosphate [PtdIns ( 3 , 4 , 5 ) P3] and consist of a p110 catalytic and a regulatory subunit . The class IA catalytic subunits , p110α , β , and δ , are constitutively bound to p85-related regulatory proteins that link them to the activation by protein tyrosine kinase receptors . The only class IB PI3K member , p110γ , is activated downstream of G protein-coupled receptors ( GPCRs ) , and interacts with p101 or p84 ( also known as p87PIKAP ) adaptor subunits [1]–[3] . A tight complex of the p110γ catalytic subunit ( PK3CG ) with p101 was first discovered in neutrophils [1] . The p101 subunit ( PI3R5 ) sensitizes PI3Kγ for activation by Gβγ subunits of trimeric G proteins , and is essential for chemotaxis of neutrophils towards GPCR-ligands [1] , [4] , [5] . Mast cells do not express p101; however , they do express the homologous adaptor protein p84 ( [PI3R6] ) [6] , which shares 30% sequence identity with p101 . Both p101 and p84 potentiate the activation of p110γ by Gβγ , but the p110γ-p101 complex is significantly more sensitive towards Gβγ , and displays an enhanced translocation to the plasma membrane as compared with p110γ-p84 [7] . Although p84 is absolutely required to relay GPCR signals to protein kinase B ( PKB/Akt ) phosphorylation and degranulation [6] , its role is not completely understood: contrary to p110γ-p101 , p110γ-p84 requires additionally the presence of the small G protein Ras , and might operate in distinct membrane micro-domains [6] , [7] . Interestingly , genetic ablation of p110γ blocks high-affinity IgE receptor ( FcεRI ) -dependent mast cell degranulation in vitro and in vivo [8] . In part this is due to the fact that initial IgE/antigen-mediated mast cell stimulation triggers the release of adenosine and other GPCR ligands to feed an autocrine/paracrine activation of PI3Kγ , which then functions as an amplifier of mast cell degranulation . Interestingly , a substantial part of the observed PI3Kγ-dependent histamine-containing granule release ( ca . 40% ) was found to be resistant to Bordetella pertussis toxin ( PTx ) pretreatment [9] , [10] . Furthermore , although adenosine activates PI3Kγ via the A3 adenosine receptor ( A3AR; [ADORA3] ) , A3AR null mice are still sensitive to passive systemic anaphylaxis , and degranulation in A3AR−/− bone marrow-derived mast cells ( BMMCs ) upon antigen stimulation remains functional [11] , [12] . This and the strong degranulation phenotype of PI3Kγ−/− BMMCs suggest that GPCR signaling does not generate the full input to PI3Kγ-dependent degranulation , but a GPCR-independent activation mechanism for PI3Kγ has yet to be defined . Here we identify a mechanism that activates PI3Kγ independently of GPCRs: we demonstrate that ( i ) IgE/antigen complexes and extracellular Ca2+ influx activate PI3Kγ , ( ii ) PI3Kγ is operationally linked to the FcεRI specifically by PKCβ ( PRKCB ) , ( iii ) and that the phosphorylation of Ser582 located in the helical domain of p110γ by PKCβ leads to the dissociation of the p84 adapter to decouple phosphorylated p110γ from GPCR inputs . Further we characterize the p110γ-p84 interface , and delineate an activation process that seems to be conserved among class I PI3Ks .
A committed step in mast cell activation is the influx of extracellular Ca2+ by store-operated Ca2+ entry ( SOCE ) [13] . Thapsigargin , which inhibits the sarco/endoplasmic reticulum Ca2+ reuptake ATPase ( SERCA ) , causes depletion of Ca2+ stores , triggering SOCE . The latter achieves full-scale degranulation of BMMCs [14] . Surprisingly , BMMCs devoid of the p110γ catalytic subunit of PI3Kγ lost their responsiveness to thapsigargin and matched degranulation responses attained by wortmannin-pretreated cells ( Figure 1A ) . To investigate if thapsigargin-triggered , p110γ-dependent degranulation involved release of adenosine , BMMCs were preincubated with adenosine deaminase ( ADA ) ( Figure 1B ) to convert adenosine to inosine , which has a very low affinity for adenosine receptors . ADA attenuated degranulation induced by IgE/antigen but did not affect thapsigargin-stimulated degranulation in wild type cells and did not further attenuate residual degranulation in p110γ null BMMCs . Likewise , presence of ADA did not reduce the phosphorylation of PKB/Akt in response to thapsigargin via the p110γ-dependent pathway ( Figure 1C ) but did reduce phosphorylation of PKB/Akt in response to adenosine—illustrating that the added ADA removes adenosine quantitatively . PTx treatment of BMMCs ( Figure 1D ) blocked adenosine—but not thapsigargin-stimulated PKB/Akt phosphorylation . To exclude that autocrine/paracrine signaling to p110γ occurred through PTx-insensitive Gαq subunits to phospholipase β ( PLCβ [PLCB2 , PLCB3] ) , and a subsequent Ras activation by the Ras guanine nucleotide exchange factor RasGRP4 as described earlier in neutrophils [15] , we used platelet activating factor ( PAF ) to trigger cyclic AMP-responsive element-binding protein ( CREB ) phosphorylation . PAF was reported earlier to trigger a PTx-insensitive Ca2+ release from mast cells [9] , and induced here a robust CREB phosphorylation comparable to adenosine and IgE/antigen . In contrast , PAF failed to trigger phosphorylation of PKB/Akt by itself , and did not enhance signaling of IgE/antigen to PKB/Akt ( Figure S1 ) . Altogether , these results clearly illustrate that thapsigargin stimulates BMMCs via a PI3Kγ-dependent activation pathway , which operates separately from adenosine-induced activation of Gαi/o trimeric G proteins . PI3Kγ activation has been linked to ligation of GPCRs [1] , [16] , but not to elevated intracellular Ca2+ concentration ( [Ca2+]i ) . We therefore chelated extracellular Ca2+ using EDTA , and buffered intracellular Ca2+ with the cell permeable Ca2+ chelator 1 , 2-Bis ( 2-aminophenoxy ) ethane-N , N , N′ , N′-tetraacetic acid tetrakis ( acetoxymethyl ester ) ( BAPTA-AM ) . Extracellular and intracellular Ca2+ chelation prevented phosphorylation of PKB/Akt induced by thapsigargin or the Ca2+ ionophore ionomycin , while IL-3 and adenosine signaling to PKB/Akt remained unperturbed ( Figure 2A and 2B ) . Interestingly , the concentration of Ca2+ required to trigger PI3Kγ-dependent phosphorylation of PKB/Akt exceeded peak concentrations that are reached by GPCR stimulation: GPCR agonists release Ca2+ only from internal stores ( maximum [Ca2+]i<300 nM ) , while thapsigargin ( Figure 2C–2F ) and IgE/antigen trigger SOCE and elevate [Ca2+]i to µM concentrations [9] , [14] , [17] . Moreover , the correlation of maximally achieved [Ca2+]i after thapsigargin revealed a steep , switch-like activation of PKB/Akt . While Ca2+ release from internal stores triggered by Gαi-coupled GPCRs ( such as the A3AR stimulated by N6- ( 3-iodobenzyl ) -adenosine-5′-N-methylcarbox-amide [IB-MECA] ) is sensitive to PTx , thapsigargin-induced SOCE is not . Ca2+-induced activation of PI3Kγ only occurs after SOCE , and is therefore clearly separated from the GPCR/trimeric G protein/PI3Kγ axis . Protein kinase C ( PKC ) inhibitors ( Ro318425 , Gö6983 , Gö6976 ) targeting classical and atypical PKCs , and the inhibitor PKC412 , which mainly inhibits classical PKCs , all substantially blocked PKB/Akt phosphorylation in response to thapsigargin and phorbol 12-myristate 13-acetate ( PMA ) ( Figures 3A and S2A ) . Rottlerin , with a limited selectivity for PKCδ , had no effect on PKB/Akt activation . GPCR-dependent PI3Kγ activation by adenosine was resistant to all tested PKC inhibitors ( Figure S2B ) . The inhibitor profile suggested that a classical PKC activates PI3Kγ . While PKB/Akt activation by PMA and thapsigargin was blocked in PKCβ−/− BMMCs ( Figure 3B ) , signaling in PKCα−/− and PKCγ−/− BMMCs remained intact ( Figure S2C ) . Deletion of PKCβ eliminated phosphorylation of PKB/Akt on Thr308 and Ser473 completely , whereas a residual signal on Ser473 was observed after PI3K-inhibition by wortmannin . This may be explained by the observation that PKCβ2 can function as a Ser473 kinase [18] . Adenosine , IL-3 , and stem cell factor ( SCF ) -induced PKB/Akt activation was not affected by elimination of PKCβ ( Figure 3C ) , demonstrating that PKCβ does not relay adenosine signals to PI3Kγ , and is not required in cytokine and growth factor receptor-dependent activation of class IA PI3Ks in mast cells . The direct measurement of phosphoinositides in BMMCs confirmed that ablation of PKCβ or its inhibition eliminated production of PtdIns ( 3 , 4 , 5 ) P3 triggered by thapsigargin and PMA , but not adenosine ( Figure 3D–3F ) . Interestingly , the link between PKCβ and PI3Kγ seems to be transient in nature , as PMA stimulation triggers short lived PtdIns ( 3 , 4 , 5 ) P3 peaks ( Figure 3F ) . Impaired FcεRI-triggered degranulation has been reported in both p110γ−/− [9] and PKCβ−/− BMMCs [19] , and the sensitivity of degranulation to PKC inhibition fits the phospho-PKB/Akt output ( Figure 3G ) . This , combined with the similarity of p110γ and PKCβ null phenotypes in IgE/antigen-induced degranulation ( Figure 3H ) , suggests a direct link of PKCβ and PI3Kγ downstream of FcεRI . Co-expression of p110γ with tagged full length or truncated PKCβ2 ( Figure 4A ) revealed that only the catalytic domain fragment and a pseudo-substrate deletion mutant of PKCβ2 formed complexes with p110γ ( Figure 4B ) , suggesting that the presence of the pseudo-substrate in PKCβ results in a closed conformation that is unable to interact with p110γ . An in vitro protein kinase assay with recombinant PKCβ2 and glutathione S-transferase ( GST ) -tagged wild type p110γ or catalytically inactive p110γ ( KR; Lys833Arg mutant ) as substrate , showed that PKCβ robustly phosphorylated p110γ ( Figure 4C ) . The capability of p110γ to auto-phosphorylate [20] was not required in the process . Analysis of phosphorylated , catalytically inactive p110γ by liquid chromatography tandem mass spectrometry ( LC-MS/MS ) identified Ser582 as a target residue of PKCβ ( YESP[582]LKHPK; spectra in Figure S3 ) . Mass spectrometric multiple reaction monitoring ( MRM ) ( Figure S3C ) showed that Ser582 phosphorylation was absent in assays lacking PKCβ , or when PKC-inhibitor was added ( Figure 4D ) . Ser582 phosphorylation was also detected by MRM in PMA and IgE/antigen stimulated BMMCs ( Figure 4D , lower panel ) . Phospho-Ser582 site-specific antibodies ( see Figure S4 ) revealed p110γ Ser852 phosphorylation in PMA- , thapsigargin- , and IgE/antigen-stimulated BMMCs , but not in mast cells exposed to adenosine or IgE alone ( Figure 5A ) . Consistent with the requirement of Ca2+ mobilization , IgE/DNP-induced phosphorylation of p110γ on Ser582 was blocked by extracellular ( EDTA , EGTA ) and intracellular ( BAPTA/AM ) Ca2+ chelation ( Figure 5B ) . Although the extended peptide around Ser582 scores as a PKC substrate site , the core Arg-X-X-Ser582 sequence is a putative recognition site for several protein kinases ( scores are PKC>protein kinase A>calcium/calmodulin-dependent kinases [CAMK] ) . As mast cell activation is accompanied by a massive influx of extracellular Ca2+ , we assessed if CAMK could phosphorylate p110γ directly . In the presence of 32P-γ-ATP , recombinant CAMKII ( CAMK2 ) incorporated equal amounts of phosphate into free and p84-bound p110γ . In the same experiment , PKCβ preferentially phosphorylated free p110γ , illustrating a preference of PKCβ for p110γ surfaces obscured in the p84-p110γ complex . CAMK also substantially phosphorylated p84 , which was borderline in in vitro assays with PKC ( 12 versus 3 mol % ) ( Figure S5A ) . Probing the phosphorylation of Ser582 in vitro demonstrated that access to this site is blocked when p84 is bound to p110γ ( Figure S5B ) . In cellular assays stimulating BMMCs with thapsigargin , CAMK activation could be monitored using anti-phospho-CAMKII antibodies . While PKC inhibitors left CAMKII phosphorylation >50% intact , Ser582 and PKB/Akt phosphorylation were both reduced to background levels ( Figure S6 ) , even in a context favoring Ca2+-triggered responses . The above , and the fact that PKCβ−/− BMMCs showed a major reduction in phospho-Ser582 after PMA or IgE/antigen stimulation ( Figure 5C and 5D ) , illustrate that phosphorylation of p110γ is mainly mediated by PKCβ , and that other PKC isoforms ( see also Figure S2 ) and Ca2+-dependent kinases attribute to less than 18% of the observed overall signal . To evaluate if the Ser582 phosphorylation affected the intrinsic activity of p110γ , a phosphorylation-mimicking mutant ( Ser582Glu ) was produced . The activity of the p110γ Ser582Glu mutant was enhanced approximately 2-fold independently of the substrate used ( PtdIns ( 4 , 5 ) P2 , PtdIns , or auto-phosphorylation ) ( Figure 6A–6C ) . Ser582 is localized in the helical domain of p110γ , and it is interesting to note that helical domain mutants of p110α found in tumors display a similar increase in enzyme activity [21]–[25] . As mutations in the helical domain of p110α attenuate contacts with the p85 regulatory subunit [21] , we examined binding of p110γ mutants to the PI3Kγ adaptor subunit p84: the substitution of Ser582 with Glu and Asp , abrogated p110γ-p84 interactions in HEK293 cells and BMMCs , while the Ser582Ala replacement favored p110γ-p84 complex formation ( Figure 6D–6F ) . In line with this , PMA-induced phosphorylation of Ser582 in BMMCs was suppressed by the overexpression of p84 ( Figure 6G ) , which fits the very limited access of PKCβII to in vitro phosphorylate Ser582 in the p110γ-p84 complex ( reduced to 20% of phosphorylation of free p110γ ) ( Figure S5B ) . Most importantly , the correlation of phosphorylation of Ser582 on p110γ and the release of p84 could also be established in wild type BMMCs: when stimulated with PMA , the amount of p84 that could be co-precipitated with p110γ was reduced significantly , and was linked to Ser582 phosphorylation of p110γ . In the inverse co-immunoprecipitation , anti-p84-associated p110γ was reduced , and phosphorylation was below detection levels in the remaining p84-associated p110γ ( Figure S7 ) . The collected results are in agreement with a mechanism in which PKCβII-mediated phosphorylation of Ser582 and the interaction of p84 and p110γ are exclusive events , and in which PKCβ action displaces p84 from p110γ . In order to understand how p84 could mask Ser582 phosphorylation , and to map the p110γ-p84 contact interface , hydrogen deuterium exchange mass spectrometry ( HDX-MS ) was used . HDX-MS elucidated contacts of class IA p110δ with its p85 regulatory subunit , and the mechanism of action of cancer-linked mutations in p110α [25] , [26] . HDX-MS relies on amide hydrogen exchange with solvent at a rate dependent on their involvement in secondary structure and solvent accessibility . Following proteolysis , location and extent of deuterium uptake are analyzed by peptide mass determination . The primary sequence of p110γ was covered >90% by 202 peptide fragments ( Figure S8; Table S1 ) . Deuterium ( 2H ) incorporation into free and p84-complexed p110γ was analyzed at seven time points ( 3 to 3 , 000 s ) . Differences in 2H-exchange of free and complexed p110γ were mapped onto the crystal structure of p110γ lacking the N-terminal domain ( PDB ID:2CHX , residues 144–1 , 093 ) , to visualize conformational changes induced by p84 ( Figures 7A , S9 , and S10 ) . Peptides with highest decrease in 2H incorporation ( >1 . 0 Da ) clustered to the RBD-C2 linker , the C2-helical domain linker , and the helical domain . The 2H-incorporation in the presence of p84 is visualized as integrated average difference in exchange at all seven time points in Figure 7B , illustrating that the helical domain provides the dominant interface with p84 . Due to difficulties in producing free p84 , the contacts on p84 with p110γ could not be mapped . Interestingly , in the absence of the p84 subunit , the majority of peptides in the helical domain exhibited broad isotopic profiles ( HDX of peptide 623–630 , which is representative of peptides in the helical domain , is shown in Figures 7C and S11 ) . This type of profile known as type 1 exchange ( EX1 ) kinetics is indicative of concerted dynamic motions of a substructure in a protein , rather than the local fluctuations characteristic of EX2 kinetics [27] . The N-terminus of p110γ was shown to stabilize the p110γ-p101 heterodimer [28] . Expression of p110γ mutants lacking the first 130 amino acids ( Δ130–p110γ ) seemed to support this view , as association with p84 was lost ( Figure 7D ) . However , when truncated p110γ was N-terminally tagged with GST ( GST-Δ130-p110γ ) , binding of p84 was restored . Although we detected a small decrease in the 2H-incorporation in two N-terminal p110γ peptides ( 59–70 and 107–113 ) in the presence of p84 , this interaction seems to be dispensable . The N-terminus of p110γ instead has a role in stabilizing the intact catalytic subunit . The helical domain is the main location of interaction with p84; however , it appears that this interaction is vulnerable and easily broken , as a single phosphorylation at Ser582 is able to disrupt the contact .
The activation of PI3Kγ has been tightly linked to GPCRs-triggered dissociation of trimeric Gαi proteins , and has been shown to require the interaction of Gβγ subunits with p110γ and the PI3Kγ adaptor subunits p101 and p84 [1] , [2] , [4] , [16] , [29] , [30] . Moreover , GPCRs generate PI3K signals typically through PI3Kγ , thereby controlling extravasation of hematopoietic cells [31] , [32] , cardiovascular parameters [33] , [34] , and metabolic output [35] , [36] . Non-GPCR-mediated activation of PI3Kγ has not been reported so far , but it has been shown that phorbol esters and Ca2+ ionophores can modulate phosphoinositide levels in a variety of cells , including platelets [37] , adipocites [38] , fibroblasts [39] , and hematopoietic cells [40] . The proposed mechanisms have been diverse and involved protein tyrosine kinases and GPCR signaling . A recent finding that protein kinase D ( PKD ) can phosphorylate two distinct sites on the p85 regulatory subunit to control class IA PI3K activity [41] is an indication that PI3K control is more complex than anticipated . PI3Kγ has been shown to be a key element in enhancing IgE/antigen output by the release of adenosine . This process involves signaling downstream of Gαi-coupled A3AR , and is sensitive to PTx and ADA [9] . The resistance of thapsigargin-induced degranulation to ADA shown here , and the fact that PTx did not diminish the PI3Kγ-dependent , thapsigargin-induced phosphorylation of PKB/Akt , points to a novel mechanism of PI3Kγ activation , which is clearly distinct from GPCR action . This Ca2+-mediated PI3Kγ activation requires SOCE and [Ca2+]i >600 nM . In contrast , GPCRs yield phosphorylation of PKB/Akt in mast cells even in the absence of a change in [Ca2+]i . Furthermore , increases in [Ca2+]i triggered via GPCRs remain at levels incapable of engaging a Ca2+-dependent activation of PI3Kγ . Thapsigargin bypasses the signaling chain from IgE/antigen-clustered FcεRI to the activation of phospholipase Cγ ( PLCγ ) and inositol ( 1 , 4 , 5 ) -trisphosphate ( Ins ( 1 , 4 , 5 ) P3 ) production , and triggers SOCE by the depletion of Ca2+ stores . That thapsigargin requires functional PI3K activity to induce mast cell degranulation was first demonstrated using the PI3K inhibitors wortmannin and LY294002 [14] , but no link between [Ca2+]i rise and PI3Kγ activity was established previously . The fact that inhibitors for classical PKCs only prevented the thapsigargin- and PMA-induced phosphorylation of PKB/Akt , while the GPCR-mediated phosphorylation of PKB/Akt remained intact , points to a link between classical PKCs and PI3Kγ . Experiments on BMMCs lacking PKCα , PKCβ , and PKCγ , showed that only the PKCβ null cells lost the ability to activate PKB/Akt in response to thapsigargin or PMA stimulation . As the PMA- and thapsigargin-induced phosphorylation of PKB/Akt on Ser473 showed a partial resistance to wortmannin , and as it has been reported that PKCβ can directly phosphorylate Ser473 in the hydrophobic motif of PKB/Akt [18] , the effect of genetic and pharmacological targeting of PKCβ was also validated measuring PtdIns ( 3 , 4 , 5 ) P3 production directly . The lack of PtdIns ( 3 , 4 , 5 ) P3 production in thapsigargin or PMA-stimulated BMMCs treated with the PKC inhibitor PKC412 , and in cells devoid of PKCβ , is in agreement with a requirement of PKCβ upstream of PI3Kγ . That signaling from PKC to PI3K plays a role in mast cell degranulation is further supported by the close correlation of PKC inhibitor sensitivity of phosphorylated PKB/Akt and degranulation responses . Moreover , the loss of PKCβ or PI3Kγ results in a similar reduction of degranulation over a wide range of IgE/antigen concentrations . The results obtained here are in agreement with previous findings that mast cells and mice devoid of p85α/p55α/p50α [42] , [43] and p85β [44] remain fully responsive to IgE/antigen complexes . A previous report showed a biphasic activation of PI3K with PI3Kγ having an early role and PI3Kδ a later role downstream of FcεRI in murine mast cells [45] , [46] . A mechanistic link between PKCβ and the catalytic subunit of PI3Kγ was initially difficult to establish , as the direct PtdIns ( 3 , 4 , 5 ) P3 response to PMA-stimulation was transient ( main peak half life <1 min ) , and the two full-length enzymes interacted only weakly . The observation that truncated , activated forms of PKCβ formed stable complexes with p110γ , suggested that PKCβ must attain an open conformation to interact with p110γ , and that PKCβ binds to p110γ via its catalytic domain . This contact resulted in phosphorylation of Ser582 on p110γ , which could be detected both in vitro and in PMA or IgE/antigen-stimulated BMMCs by mass spectrometry . The PKCβ-mediated phosphorylation of p110γ was confirmed using site-specific anti-phospho-Ser582 antibodies . Stimuli like PMA , thapsigargin , and IgE/antigen complexes all required PKC to signal to PKB/Akt , which correlated with the phosphorylation of Ser582 on p110γ . Moreover , the phosphorylation of Ser582 on p110γ was sensitive to removal of extracellular Ca2+ , buffering of [Ca2+]i , and the genetic deletion of PKCβ . Adenosine stimulates PI3Kγ via GPCRs and PTx-sensitive trimeric G proteins [9] in a Ca2+-independent process , and did not yield a detectable phosphorylation of Ser582 . The increased turnover of PtdIns and PtdIns ( 4 , 5 ) P2 , and the increased rate of auto-phosphorylation displayed by p110γ with a phosphate-mimicking mutation ( Ser582Glu ) , suggests that a structural change in the helical domain of p110γ is sufficient to increase the catalytic activity independent of the presence of the p84 subunit . Previous work examining the activation of the class IA p110α , p110β , and p110δ catalytic subunits has shown that part of the activation mechanism occurs through a conformational change from a closed cytosolic form to an open form on membranes [25] , [26] . The helical domain of p110γ is exquisitely well placed to propagate conformational changes due to the fact that it is in contact with every other domain in p110γ . In the crystal structure of N-terminally truncated ( Δ144 ) p110γ , the side chain of Ser582 points inward [47] , [48] , and has to rotate to accommodate a phosphate . Our HDX-MS results showed a dynamic “breathing” motion in the helical domain in the free p110γ catalytic subunit that may allow for temporary exposure of Ser582 , enabling modification by PKC . HDX results showed that the p84 subunit slowed or prevented this dynamic motion , and this correlated with a decreased efficiency of phosphorylation by PKC in cells in the presence of p84 . Although Ser582 is not in a direct contact with the kinase domain , it is structurally linked to it: the heat repeat HA1/HB1 housing Ser582 , and the connecting intra-helical loop ( residues 560 to 570 ) , along with helix A3 ( 624–631 ) are in contact with helices kα9 and kα10 in the C-lobe of the kinase domain ( known as the regulatory arch [49] , [50] ) and could transduce a conformational change to the catalytic center of p110γ ( Figure 7C ) . The phosphorylated Ser582 and phosphorylation-mimicking mutants may activate lipid kinase activity by causing a conformational shift at this interface . It has been shown recently that this region of the p110γ kinase domain is critical in regulating lipid kinase activity , as phosphorylation of Thr1024 in the kα9 by protein kinase A ( PKA ) negatively regulates p110γ activity in vitro and in cardiomyocytes ( Figure 7C ) [51] . In contrast to the p110α-p85 heterodimer , stabilized by the N-terminal adaptor-binding domain ( ABD ) /inter SH2 domain interaction , the association of the p110γ subunit with its adaptor subunit is quite vulnerable , and the Ser582Glu mutant , but not Ser582Ala , abrogated the formation of a p110γ-p84 complex . HDX-MS revealed that the helical domain of p110γ was stabilized by p84 . Ser582 is located in the center of the p110γ-p84 contact surface , which explains how a change in charge ( Ser582Glu ) breaks the interaction with p84 , either by direct contact or by destabilization of the helical domain . Overexpression of p84 shields p110γ from a PMA-induced phosphorylation , and suggests that binding of p84 to p110γ , and Ser582 phosphorylation by PKCβ are mutually exclusive . This implies that the two activation modes of PI3Kγ—by GPCRs or PKCβ—are completely separated . At low [Ca2+]i , PI3Kγ is exclusively activated by Gβγ subunits . It has been demonstrated that a PI3Kγ adapter protein is absolutely needed for functional GPCR inputs to p110γ [6] . If the interaction of p84 with p110γ is blocked by the phosphorylation of Ser582 , p110γ is decoupled from its GPCR input ( for a schematic view of the process see Figure 8 ) . The PKCβ-mediated activation and phosphorylation of p110γ constitutes therefore an unprecedented PI3K molecular switch , which enables the operation of p110γ downstream of FcεRI signaling , and will elucidate cell type-specific activation processes in allergy and chronic inflammation .
BMMCs were derived from bone marrow of 8–12-wk-old C57BL/6J wild type , p110γ−/− , PKCα−/− , PKCβ−/− , and PKCγ−/− mice , and cultured and characterized as described in [9] . Animal experiments were carried out in accordance with institutional guidelines and national legislation . Human embryonic kidney 293 ( HEK293 ) cells were grown in DMEM supplemented with 10% HI-FCS , 2 mM L-glutamine , 100 units/ml penicillin , 100 µg/ml streptomycin . Sf9 cells were cultivated in IPL-41 medium ( Genaxxon Bioscience ) supplemented with 10% HI-FCS , 2% yeastolate , 1% lipid concentrate , 50 µg/ml gentamicin ( Invitrogen ) , and 2 . 5 µg/ml amphotericin B ( Genaxxon Bioscience ) . Detailed descriptions and references are available in Text S1 . PtdIns ( 3 , 4 , 5 ) P3 levels have been measured as described in [9] with some modifications . BMMCs were cultured for 2 h in phosphate-free RPMI medium/2% FCS at 37°C and 5% CO2 , followed by labelling with 1 mCi/ml [32P]-orthophosphate for 4 h . Cells were washed , stimulated , and lysed by the addition of chloroform/methanol ( 1∶2 , v/v , with butylated hydroxytoluene and carrier phosphoinositides ) . Lipids were extracted , deacylated , and separated by high-pressure liquid chromatography ( HPLC ) . PI3Kγ-His6 was incubated with PtdIns ( 4 , 5 ) P2-containing lipid vesicles ( PE/PS/PC/SM/PIP2 = 30/20/10/4 . 5/1 . 2; PIP2 final 5 µM ) , 10 µM ATP , and 4 µCi of [γ32P]-ATP in lipid kinase buffer ( 40 mM HEPES [pH 7 . 4] , 150 mM NaCl , 4 mM MgCl2 , 1 mM DTT [1 , 4-Dithio-DL-threitol] , 0 . 1 mg/ml fatty-acid free BSA ) for 10 min at 30°C . Alternatively , PtdIns/PS vesicles ( ∼200 µM each ) were used . Reactions were terminated by addition of 1 M HCl and CHCl3/MeOH . Lipids were isolated by chloroform extraction , separated by TLC and quantified on a Typhoon 9400 . Numeric results were tested for significance using a two-tailed Student's t test , ( paired or unpaired , as imposed by datasets ) . * or & , ** or && , and *** or &&& refer to p-values p<0 . 05 , p<0 . 005 , and p<0 . 0005 . * and & were used for comparison of different genotypes , stimuli an conditions as indicated . Calculations were carried out using Graph Pad Prism , Microsoft Excel , or Kaleidagraph software . Protein stock solutions ( 5 µl; Hsp110γ-C-His6: 30 µM; Hsp110γ-C-His6/Mmp84-C-His6: 35 µM ) were prepared in 20 mM Tris [pH 7 . 5] , 100 mM NaCl , 1 mM ammonium sulfate , and 5 mM DTT . Exchange reactions were initiated by addition of 25 µl of a 98% D2O solution containing 10 mM HEPES ( pH 7 . 2 ) , 50 mM NaCl , and 2 mM DTT , giving a final concentration of 82% D2O . Deuterium exchange reactions were allowed to carry on for seven time periods , 3 , 10 , 30 , 100 , 300 , 1 , 000 , and 3 , 000 s of on-exchange at 23°C , before addition of quench buffer . On-exchange was stopped by the addition of 40 µl of a quench buffer containing 1 . 2% formic acid and 0 . 833 M guanidine-HCl , which lowered the pH to 2 . 6 . Samples were then immediately frozen in liquid nitrogen until mass analysis . The full HDX-MS protocol can be found in Text S1 .
|
Phosphoinositide 3-kinases ( PI3Ks ) are involved in most essential cellular processes . Class I PI3Ks are heterodimers: class IA PI3Ks are made up of one of a group of regulatory p85-like subunits and one p110α , p110β , or p110δ catalytic p110 subunit , and are activated via binding of their p85 subunit to phosphorylated tyrosine receptors or their substrates . The only , class IB PI3K member , PI3Kγ , operates downstream of G protein-coupled receptors ( GPCRs ) . Recent work suggested that PI3Kγ also operates downstream of IgE-antigen complexes in mast cell activation , but no mechanism was provided . We show that clustering of the high-affinity IgE receptor FcεRI triggers a massive calcium ion influx , which leads to PKCβ activation . In turn , PKCβ phosphorylates Ser582 of the PI3Kγ catalytic p110γ subunit's helical domain . Downstream of GPCRs , p110γ requires a p84 adapter to be functional . Phospho-mimicking mutations at Ser582 disrupt the p84-p110γ interaction , and cellular Ser582 phosphorylation correlates with the loss of p84 from p110γ . Thus our data suggest that PKCβ phosphorylates and activates p110γ downstream of calcium ion influx , while simultaneously disconnecting the phosphorylated p110γ from GPCR signaling . Exploration of the p84-p110γ interaction surface by hydrogen- deuterium exchange mass spectrometry confirmed that the p110γ helical domain forms the main p84-p110γ contact surface . Taken together , the results suggest an unprecedented mechanism of PI3Kγ regulation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"enzyme",
"structure",
"signal",
"transduction",
"phosphoinositide",
"signal",
"transduction",
"inflammation",
"immunity",
"enzymes",
"allergy",
"and",
"hypersensitivity",
"immunology",
"signaling",
"pathways",
"biology",
"molecular",
"cell",
"biology"
] |
2013
|
PKCβ Phosphorylates PI3Kγ to Activate It and Release It from GPCR Control
|
Natural proteins often partake in several highly specific protein-protein interactions . They are thus subject to multiple opposing forces during evolutionary selection . To be functional , such multispecific proteins need to be stable in complex with each interaction partner , and , at the same time , to maintain affinity toward all partners . How is this multispecificity acquired through natural evolution ? To answer this compelling question , we study a prototypical multispecific protein , calmodulin ( CaM ) , which has evolved to interact with hundreds of target proteins . Starting from high-resolution structures of sixteen CaM-target complexes , we employ state-of-the-art computational methods to predict a hundred CaM sequences best suited for interaction with each individual CaM target . Then , we design CaM sequences most compatible with each possible combination of two , three , and all sixteen targets simultaneously , producing almost 70 , 000 low energy CaM sequences . By comparing these sequences and their energies , we gain insight into how nature has managed to find the compromise between the need for favorable interaction energies and the need for multispecificity . We observe that designing for more partners simultaneously yields CaM sequences that better match natural sequence profiles , thus emphasizing the importance of such strategies in nature . Furthermore , we show that the CaM binding interface can be nicely partitioned into positions that are critical for the affinity of all CaM-target complexes and those that are molded to provide interaction specificity . We reveal several basic categories of sequence-level tradeoffs that enable the compromise necessary for the promiscuity of this protein . We also thoroughly quantify the tradeoff between interaction energetics and multispecificity and find that facilitating seemingly competing interactions requires only a small deviation from optimal energies . We conclude that multispecific proteins have been subjected to a rigorous optimization process that has fine-tuned their sequences for interactions with a precise set of targets , thus conferring their multiple cellular functions .
Proteins engage in numerous protein-protein interactions , which together regulate the outcome of all biological processes in the cell . By some estimates , over a third of all mammalian proteins participate in two or more highly specific protein-protein interactions [1] . Proteins that can interact with a large number of partners play a central role in the modular organization of protein interaction networks [2] . Such proteins , usually referred to as protein hubs , tend to be more essential than others for cell survival [3] and usually exhibit slower rates of evolution [4] . Moreover , the comprehensive biological activity of these proteins typically requires them to recognize a precise set of targets in a specific way . For example , each subfamily of G protein regulators interacts with only a specific subset of G proteins [5] . Proteins with diverse binding capacity have also been termed multispecific proteins [6] , [7] . The central function of multispecific proteins within interaction networks imposes constraints on their amino acid sequences , especially in their protein-protein interfaces , i . e . , the regions that are used to mediate intermolecular interactions with various targets . There exist only a few studies that have characterized in great detail the molecular and structural features of multispecific protein interfaces [8]; this is mostly due to sparse representation of such protein-protein complexes in the Protein Data Bank ( PDB ) . A thorough understanding of atomic-level principles governing multispecific interactions is extremely important not only for the advancement of basic science but also for the design of new pharmaceuticals that modify protein-protein interactions . Furthermore , such molecular insights will provide critical feedback for systems biology research , which views protein-protein interactions from a high-level network approach [9] . Calmodulin ( CaM ) is a paradigm of a multispecific protein , with more than three hundred CaM targets identified to date [10] . CaM is the central player in the signaling pathways that control gene transcription , protein phosphorylation , nucleotide metabolism , and ion transport . This sensor protein translates the changes in concentration into activity of many downstream targets , including kinases , phosphatases , enzymes , and ion channels [11] . Remarkably , CaM targets display considerable variability in sequence and structure . CaM-binding regions within target proteins are generally rich in hydrophobic and positively charged residues . Nevertheless , no consensus CaM-binding sequence exists for all CaM target proteins ( Figure 1C ) . Recent structural studies have revealed that there are several binding modes accessible to CaM , allowing this protein to interact with its targets in a -saturated state ( 4 ions bound to CaM ) [12] , [13] , in a partially-saturated state ( 2 ions bound to CaM ) [14] , and in a -free state [15] , [16] . In the -saturated form , CaM usually binds to a stretch of amino acids that is unfolded in the absence of CaM and becomes helical upon interaction with the protein [11] . In this “conventional” binding mode , CaM undergoes a conformational change and embraces the target helix with its two globular domains , burying a substantial hydrophobic surface area and providing favorable hydrogen bond and salt bridge interactions with the target ( Figure 1A , B ) . -saturated CaM binds to its targets with high affinity , displaying values in the to M range [17] . This affinity is reduced at least 1000-fold in the absence of , allowing for quick dissociation of CaM from its targets when is depleted . The multitude of binding constraints placed on CaM during evolution is likely to have produced a sequence that may not be optimal for binding to any particular CaM target , but rather presents a compromise essential for interaction with a large number of partners . In this study , we employ a computational design approach [18] to understand how the compromises required for functional promiscuity [19] are achieved both on the level of amino acid sequences and on the level of binding energetics . First , we computationally “evolve” CaM to interact with single targets; second , we evolve this protein to bind to multiple partners simultaneously . Recently , a similar analysis was performed on twenty multispecific proteins , whose interactions with two to seven targets were considered [6] , [7] . In contrast to those works , we report a much more comprehensive investigation of a single multispecific protein , CaM . We examine interactions in sixteen different CaM-target complexes that exhibit the conventional binding mode . Using the structures of these complexes , we perform 697 separate CaM design calculations to obtain low energy CaM sequences optimal for either a single target or some combination of the targets . Rigorous quantitative and statistical comparisons of the designed CaM sequences and their energies allows us to draw conclusions regarding CaM evolution and to suggest strategies for the design of binders that are both promiscuous yet highly specific . In particular , we characterize the CaM binding interface by partitioning its residues into those that are critical for binding affinity and those that are important for multispecificity . Furthermore , we analyze the sorts of sequence compromises required to yield proteins with promiscuous interactions and show how this fits with past explanations for the ability of CaM to accommodate many targets . Finally , we examine the energetic compromises inherently crucial for multispecificity [20] , and we find that our results also shed light on the unexpected findings of previous experimental protein design research .
First , we assessed the similarity of our designed CaM interface sequences to the native CaM sequences . The number of mutations predicted in the single best CaM sequence , designed for interaction with one target , ranges from four to sixteen with a mean value of 9 . 5 ( Figure 3A ) . When two CaM targets are included in the design , the number of predicted mutations ranges from 3 to 13 with a mean of 7 mutations . The distribution of predicted mutations further shifts to the left when three CaM targets are incorporated into the design , exhibiting a mean of 6 mutations . Incorporation of all sixteen states in the design procedure resulted in only 4 mutations . Next , we compared the distribution of amino acids obtained from the one hundred CaM sequences designed for interactions with one , two , and three targets . This was done by calculating the Jensen-Shannon divergence ( JSD , see Methods ) between the evolutionary profile of CaM and the amino acid distribution obtained after CaM redesign . A JSD score of 0 corresponds to identical distributions , while a JSD score of 1 corresponds to completely discordant distributions . We henceforth refer to the JSD score as the “dissimilarity score” . A comparison of the hundred CaM interface sequences designed for one , two , and three targets ( Figure 3B ) showed the same trend as observed for the single best CaM sequences . The highest dissimilarity scores were obtained for single-state designs ( mean value of 0 . 48 ) , medium scores were obtained for two-state designs ( mean value of 0 . 37 ) , lower scores were obtained for three-state designs ( mean value of 0 . 35 ) , and the lowest score was obtained for sixteen-state design ( 0 . 24 ) . We next compared the hundred best CaM sequences designed for interactions with the various single targets . For each of the interface positions , we calculated the dissimilarity score between the distribution of designed amino acids and the evolutionarily-derived distribution ( Figure 4A ) . Our analysis revealed that , at some of the CaM interface positions , our design calculations predicted a distribution very similar to the evolutionary profile for the majority of the CaM-target complexes ( columns with lighter boxes , Figure 4A ) . On the other hand , at other positions , the design methods predicted amino acid distributions very different from the evolutionary profile ( columns with darker boxes ) . Among the 16 different CaM-target complexes , the average per-position dissimilarity score was very diverse and ranged from 0 . 276 to 0 . 741 ( mean of 0 . 48 ) , so that some structures inherently predict profiles much more similar to the evolutionary profile than others . These scores slightly decreased ( numbers in parentheses ) if we excluded from our analysis the CaM positions that belong to the common CaM binding interface but do not interact with the target in the particular CaM-target complex . We also noticed that the designed CaM sequences are more similar to the evolutionarily-defined CaM sequences for the targets that interact with a larger number of CaM residues . Figure 4B shows that there is an inverse correlation ( ) between the dissimilarity with the evolutionary profile and the number of the designed CaM positions that are in the binding interface for a particular CaM-target complex . In addition , not unexpectedly , the designed CaM sequences come out somewhat more similar to the native profile if the WT CaM sequence is predicted to be strongly compatible with the CaM-target complex structure . This is demonstrated in Figure 4C , which shows a correlation ( ) between the dissimilarity with the evolutionary profile and the energy of the WT CaM sequence in the context of a particular structure . Next , we quantified the correlation among the hundred best sequences designed for interactions with different single targets . This was done by calculating the dissimilarity score between all possible pairs of single-state designs at each of the design positions . This type of analysis allowed us to identify the CaM binding interface positions that , on the whole , exhibit similar amino acid identities in all CaM-target complexes ( affinity-defining positions: 19 , 36 , 71 , 72 , 92 , 109 ) and the positions that display much greater diversity among the single-state designs ( specificity-defining positions: 11 , 14 , 18 , 39 , 41 , 84 , 87 , 112 ) ( Figure 5A ) . In the evolutionary profile of CaM , the affinity-defining positions are occupied by hydrophobic residues , either Met , Leu , or Phe . The specificity-defining positions , on the other hand , are dominated by hydrophilic amino acids ( Glu and Gln ) and , in some cases , are occupied by Leu ( Figure 5B ) . The affinity- and specificity-defining positions are present in both the N- and the C-terminal domains of CaM and are also distributed evenly throughout the CaM structure ( Figure 5C ) . In addition , we could not detect any differential pattern in the way the targets interact with either class of CaM positions ( since the CaM targets do not exhibit distinctly conserved motifs; see Figure 1C ) . In an attempt to further understand the differences between interactions defining affinity and specificity in native CaM , we threaded the WT CaM sequence onto all sixteen selected CaM-target complexes and calculated the energetic contribution of each of the binding interface positions to the total energy . The energetic contributions at each position were further separated into intra- and intermolecular energies , corresponding to stabilization within CaM and between CaM and the target , respectively . We further averaged the per-position energetic contributions for the sixteen CaM-target complexes . We saw that there is a distinct difference in how the affinity- and specificity-defining positions stabilize the WT CaM-target complexes . This difference is especially striking for the intramolecular energy contributions ( Figure 5D ) . The six affinity-determining positions exhibit the highest intramolecular contributions among all positions , being crucial for stabilization of CaM in the target-bound conformation . The majority of the specificity-determining positions , on the contrary , exhibit higher than average , and sometimes even unfavorable , contributions to the intramolecular energy . However , most of these specificity-determining positions contribute more than average to the intermolecular energies , being more important for direct interactions with the target ( Figure 5E ) . We next investigated what happens to the energetic contributions in the CaM sequences designed for interactions with the single targets ( Figure 6 ) . This was done by computing the total energy contribution of each designed position first for the single best designed sequence and then for the WT CaM sequence , for each of the sixteen CaM-target complexes . The per-position energetic contributions were then averaged over the sixteen cases . Figure 6 shows that , at all design positions , the energetic contribution is either unchanged or is improved for the designed sequences compared to that of the WT CaM sequence . An unchanged value is observed at positions that are highly optimized for interaction with the target , including most of the affinity-defining positions . Large improvements in the energetic contributions from the WT to design are observed for positions where the WT energies were less favorable , including the majority of the specificity-defining positions . CaM needs to achieve a certain compromise to obtain a sequence compatible with binding each of the two targets . Comparison of the CaM sequences designed for interactions with each of the two single targets ( single-state designs ) and the combinations of these two targets ( two-state designs ) revealed that the compromise could be achieved via five different scenarios . This is demonstrated in Figure 7 using the examples of CaM-target complexes , corresponding to the PDB codes 2F3Y and 3BXL . In the most trivial scenario , CaM sequence profiles designed for the two single targets have an identical or very similar amino acid distribution at a particular position ( e . g . , position 145 in Figure 7B ) . This amino acid distribution remains the same when CaM is designed to interact with both of these targets ( “Kept same” in Figure 7A ) . In the second scenario , two different amino acid distributions are observed for the single-state designs . However , the sequence profile designed for both targets is similar to both of the two distributions resulting from the single-state designs , since it combines them in some form ( “Combined” in Figure 7A and e . g . , position 87 in Figure 7B ) . In the third scenario , two different amino acid distributions are observed for the single-state designs , while in the two-state design , one of these distributions dominates ( “Preferred one” in Figure 7A and position 124 in Figure 7B ) . In the fourth scenario , two different amino acid distributions are again observed for the single-state designs . In the two-state design , however , a new amino acid distribution appears; this distribution is significantly different from those observed for both single-state designs ( “New aa” in Figure 7A , position 18 in Figure 7B ) . In the fifth scenario , an identical amino acid distribution is observed for the single-state designs . Interestingly , a new amino acid distribution appears in the two-state design ( “despite same” in Figure 7A , position 14 in Figure 7B ) . This scenario , however , occurs only very infrequently throughout our design calculations . Expectedly , the affinity-determining positions in CaM ( 19 , 36 , 71 , 72 , 92 , 109 ) tend to exhibit the “Kept same” category of compromise , while the specificity-determining positions ( 11 , 14 , 18 , 39 , 41 , 84 , 87 , 112 ) tend to select the “Preferred one” category . We next investigated if the compromises required to achieve multispecificity bring the CaM sequence closer to its evolutionarily-derived sequence profile . For this purpose , we compared the amino acid distributions resulting from the single-state designs with those from the two-state designs , all in relation to the evolutionary profile ( Figure 8 ) . Here , we discarded the scenarios where the two-state design produced results similar to both single-state designs ( “Kept same” , “Combined” ) , since these scenarios do not result in changes relative to the evolutionary profile of CaM . Interestingly , for most of the designed positions , CaM sequences optimized for two targets were more similar to the evolutionary profile than those optimized for single targets ( “Benefit” in Figure 8A and position 112 in Figure 8B ) . In a few cases , no significant change was observed vis-a-vis the evolutionary profile ( “No Change” in Figure 8A , position 18 in Figure 8B ) , while in some cases the amino acid distribution becomes more different from the evolutionary profile compared to that of the single-state designs ( “Loss” in Figure 8A , position 14 in Figure 8B ) . It is interesting to see how the overall amino acid composition ( calculated for all 100 best sequences ) changes from the CaM interface sequences designed for interaction with a single target to the sequences designed for multispecificity ( either two-state or three-state design ) . Figure 9 shows several significant differences between the two situations . Methionine dominates the compositions of the CaM-binding interface for single-state designs . They become even more frequent when CaM is designed for interactions with two or three targets . In addition , we noted a significant increase in the number of Leu , Gln , Ser , Gly , and Val when introducing additional interaction constraints on the CaM sequence . On the other hand , all aromatic amino acids ( Phe , Trp , Tyr ) , as well as Arg , become significantly less abundant when more than one CaM target is considered in the design . In this study , we designed 100 CaM binding interface sequences for each of 697 design scenarios ( 1-state , 2-state , 3-state , and 16-state ) . We computed the energy of each of these sequences in the context of all sixteen structures of the CaM-target complexes . Each design scenario was assigned an energy value in each structure; this energy value was the minimum of the energies obtained by the 100 sequences designed in this scenario . We next analyzed how these energies vary as additional targets are either introduced into , or removed from , the design procedure ( Figure 10 ) . Note that the frequency histograms in Figure 10 are based on up to 7000 comparisons of energies between design scenarios . We denote by A , B , C , and D any four arbitrary CaM states , i . e . , complexes of CaM with different targets . Firstly , we asked how incorporating additional CaM-target interactions affects the stability of the newly incorporated CaM-target complex , as opposed to performing the same design without this complex ( Figure 10A ) . For example , denotes that energies in state A were compared for the sequence resulting from design in state B and the simultaneous design in states A and B ( A+B ) . As expected , adding a state ( A ) to the design procedure , when already designing for a different state ( B ) , yields a significant increase in the stability of the designed CaM sequence in state A ( with increase in stability for almost half of all such cases ) . Similar gains in the stability of a newly incorporated state ( A ) were observed in the transition from one state ( C ) to a total of three states ( A+B+C , middle panel of Figure 10A ) . On the other hand , when already designing for two states ( B+C ) , incorporating an additional state A ( A+B+C ) yielded much lower gains in stability for that state ( leftward shifted distribution , bottom panel ) . This is due to the fact that performing 2-state design for B+C already predicts a sequence somewhat compatible with A ( middle panel of Figure 10B ) . Next , we examined the necessity of actually including a particular state in the design process ( Figure 10B ) . For example , if two CaM-target interactions were very similar in nature ( due to relatedness of the targets ) , then simply designing for one of these interactions would suffice in stabilizing the other . We did not find this to be the case for our sixteen targets , as designing for one state ( B ) results in sequences that are highly unstable in state A ( B vs . A ) . Such sequences are sub-optimal for almost half of all cases ( top panel ) . However , designing for two states ( B+C ) or three states ( B+C+D ) yields sequences that are significantly more compatible with the binding of target A ( middle and bottom panels ) . Thirdly , we investigated the effect of incorporating other states into multispecific design on those states that are already included in the design process ( Figure 10C ) . Expectedly , we found that incorporating an additional state ( B ) into the design process ( top panel ) resulted in CaM sequences that are less optimized for interaction with the first target ( A ) . Incorporating two additional states ( B+C ) yielded sequences with an additional decrease in stability when interacting with target A . On the other hand , when already designing for two states ( A+B ) and adding a third state ( A+B+C ) , the resulting CaM sequences exhibit a smaller decrease in optimality for target A . Thus , overall , we found that a large decrease in stability occurs when incorporating one additional state , but adding a third state does not have the same effect ( top vs . bottom panels ) . Finally , since the WT sequence is optimized to bind all sixteen targets studied here , we expected it to posses sub-optimal stability in the complex with any particular single target . Indeed , our analysis showed that the WT CaM sequence , when threaded onto the structures of all sixteen CaM-target complexes , always obtains a substantially higher energy compared to that of sequences optimized for these structures ( Figure 10D , top panel ) . Note that a related phenomenon was also observed above for individual design positions ( Figure 6 ) . However , the relative sub-optimality of the WT sequence in a particular interaction ( with target A ) progressively decreases when compared to sequences optimized for interactions with two targets ( A+B , middle panel ) and three targets ( A+B+C , bottom panel ) . Thus , WT sequences seem to be most energetically similar to sequences optimized for multispecificity .
The CaM interface sequences that we designed to best interact with single targets have an average of 9 . 5 mutations , corresponding to a 52 . 5% wild-type recovery rate ( Figure 3A ) . Our WT recovery rates for single-state CaM designs are very similar to those observed , on average , when redesigning protein cores ( 51% ) [24] and somewhat lower than that observed in our previous study , where the interface of a very high-affinity protein-protein complex was redesigned ( 62% ) [25] . These results are reasonable , since CaM interactions with its targets are mostly conveyed by buried residues; the affinities of CaM-target complexes , while high , are not among the highest measured in nature . On the other hand , our WT recovery rates for single-state designs are considerably higher than those observed by Humphris et al . when redesigning the interfaces of twenty multispecific protein-protein complexes [6] . In many of their examples , however , a significant fraction of the redesigned positions do not interact with the target in each particular protein complex under design and are thus likely to mutate without any constraints . Moreover , we demonstrated that the WT recovery rate for design of the CaM interface is proportional to the number of residues directly interacting with the target ( Figure 4B ) . Having more interface residues results in the addition of intermolecular contacts to the network of molecular interactions [26] , better reproducing the environment within the native CaM interface . Hence , our higher WT recovery rates for single-state CaM designs , as compared to those reported by Humphris et al . , are easily explained by the high fraction of the designed CaM positions being found in direct interaction with the target for each CaM-target complex considered ( 85% or more for all but 2 of the complexes ) . Interestingly , CaM interface sequences designed using NMR structures as templates gave significantly higher dissimilarity scores with the CaM evolutionary profile ( 2BBN and 1SY9 in Figure 4A ) than those sequences obtained using X-ray structures as templates ( all others ) ; note that these structures also have the fewest of the commonly defined interface positions interacting with their respective targets . The lower rates of native sequence recovery in design calculations using NMR structures imply that these structures may be less optimal templates for protein design calculations , in agreement with recent findings by Schneider et al . [27] . When optimizing the CaM binding interface for two , three , or sixteen targets simultaneously , our WT sequence recovery rate increases from an average of 52% to an average of 65% , 70% , and 80% , respectively . These WT recovery rates are similar to those observed previously when redesigning multispecific proteins by considering several partners together [6] . Our high-level sequence analysis of the design predictions demonstrates that the native CaM binding interface sequence is not optimal for interaction with each target on its own but fits well the multispecific requirements imposed by nature . Moreover , our novel design procedure , which includes progressive incorporation of additional targets into the design , provides a plausible scheme for CaM evolution in nature . Specifically , when designing CaM to possess binding affinity to all 16 targets studied here , the predicted interface sequence is quite similar to that resulting from evolution ( Figure 3 ) . In theory , we expect the WT recovery rate for the CaM binding interface sequence to approach 100% if all native CaM targets were taken into account . Deviation from this number would result from inaccuracies in the energy function used for design ( see below ) , or possibly from other constraints that this technique does not currently incorporate , e . g . , sequence composition preferences for the organism . When evaluating our designed CaM interface sequences , we noticed that many of these sequences are more positively charged than the evolutionary profile of CaM ( Figure 9 ) . This increase in positive charge on the CaM interaction surface could , in principle , bring about a reduction in affinity between the redesigned CaM and its targets . Nonetheless , our previous experimental studies of CaM interactions with two separate targets revealed that carefully designed charge-reversal mutations in the CaM binding interface do not reduce CaM affinity to targets and , in some cases , even increase the affinity [17] , [28] . In addition , these charge-reversal mutations help to increase CaM binding specificity [28] . Still , it is also possible that our design calculations are slightly biased toward incorporating Lys and Arg residues , which have many atoms to participate in more interactions and a larger number of rotamers; hence , they may be chosen more often than other amino acids . The energy function and molecular models we used for CaM design might not realistically portray all atomic interactions , although they have been experimentally verified for many cases , e . g . , [17] , [21] . It has recently been pointed out that some inaccuracies in energy functions can be overcome by averaging the results of many protein design calculations [29] . In this work , we tried to minimize the effect of possible errors by designing 100 sequences compatible with each design scenario and by averaging the results obtained from all possible combinations of two- and three-state CaM designs . Additional sources of modeling errors include the use of both a fixed protein backbone and rigid amino acid side chains ( rotamers ) . Some contemporary research has attempted to overcome these limitations by permitting the backbone to be flexible [30]–[33] , the side chains to move more continuously [34] , or both [35]; however , introduction of additional flexibility is computationally expensive and hence would be incompatible with our high-complexity 700 design scenarios . In short , while our calculations could be inaccurate in some particular instances , overall they fit well with similar computational and experimental work and should be reliable in predicting general and unbiased trends in CaM evolution . The per-position analysis of amino acid compromises required for achieving multispecificity in CaM followed several scenarios , two of which are especially interesting ( Figure 7A ) . In the first situation , a new amino acid appears in the two-state design that is different from amino acids observed in both single-state designs ( “New aa” ) . This amino acid , while not optimal for interaction with each target on its own , was predicted to be the best compromise satisfying interactions with both targets . Interestingly , in the majority of cases where such a scenario was observed , the new amino acid was more similar to the evolutionary profile of CaM ( e . g . , position 112 in Figure 8B ) . This scenario demonstrates how the native CaM sequence has acquired its identity . In another interesting ( but rare ) scenario , we observed that the amino acid distribution in the two-state CaM design was different from that in both single-state designs in spite of the latter distributions being identical ( “despite same” , position 14 in Figure 7B ) . This scenario is likely to be due to correlated mutations . For example , positions 14 and 18 in Figure 7B are coupled to each other . Thus , in spite of the fact that Glu dominated position 14 in both single-state designs , the appearance of Glu18 in the two-state design forces the appearance of Arg at position 14 . In this work , we classified the CaM binding interface residues as either affinity- or specificity-defining [36] , [37] . Our predictions were derived solely from sequence comparisons , with affinity-determining residues being very similar to each other among all single-state designs and specificity-determining residues differing the most . Previous studies found that the residues that maximally contribute to protein-protein interactions ( hot-spots ) are also more evolutionarily conserved [38] and tend to be grouped into spatially distinct clusters with strong interactions within the clusters [39] , [40] . In agreement with these findings , the CaM interface positions that are most “conserved” among the designs ( affinity-determining ) are also very stabilizing for the native CaM-target complexes , and these six “hot-spot” positions are clustered into three pairs ( 19 and 36; 71 and 72; 92 and 109; see Figure 5C ) . Unexpectedly , the strong energetic contributions of the hot-spot residues were largely mediated by intramolecular interactions ( Figure 5D ) , meaning that the affinity-defining residues in CaM mostly stabilize it in the target-bound conformation . On the other hand , the specificity-determining residues often have an unfavorable effect on CaM intramolecular energies but provide favorable interactions with each particular target ( Figure 5E ) . Thus , the coupling between evolution and energetics is very strong in CaM , and the pattern of this coupling can even be used to infer that large conformational changes accompany target recognition by CaM . This finding is consistent with the population shift model [41]–[44] , which asserts that an unbound protein samples a multitude of conformations; the equilibrium is shifted towards the bound state upon addition of the binding partner . Our results suggest that the affinity-determining positions enable the transition to each of the bound CaM states , while the specificity-determining positions lock CaM into a target-specific conformation . We postulate that an analogous scenario should be detected for other multispecific proteins that undergo conformational changes upon binding . Finally , we also validated our positional classifications using the INTREPID web server for predicting functionally important residues ( based on evolutionary sequence conservation ) [45] . For the 142 CaM positions , the 6 affinity-determining residues were among the 14 ranked most important for function , while the 8 specificity-determining residues were ranked significantly lower than average . The latter is not unexpected , since these positions convey distinct favorable interactions with various targets and are hence not conserved at higher levels in the evolutionary hierarchy ( not shown ) . The energetic analysis of the WT and designed sequences in the context of all sixteen structures revealed a few interesting conclusions . Firstly , we demonstrate that , from an energetic perspective , the CaM interface is optimized for binding multiple partners but sub-optimal for interaction with each particular target ( Figure 10D , top vs . middle and bottom ) . This result is in accord with previous studies , which have shown that binding promiscuity results in weaker affinity toward targets [46] . Additionally , we find that designing the CaM interface for additional functions requires a notable tradeoff in stability that escalates as more functions are simultaneously added ( Figure 10C , top and middle ) . This finding is consistent with conclusions from mutational studies of enzymes , where function-stability tradeoffs were observed in positions that are highly constrained by the catalytic mechanism [19] , [20] . Nevertheless , the loss of binding stability associated with acquiring a second binding partner is only minor when balancing it with the huge gain in CaM's favorable interactions with this new target ( Figure 10C vs . Figure 10A ) . Finally , it is of great interest that , when gaining the ability to bind a third partner , the energetic penalty imposed on the interactions of CaM with its original two partners is not that great ( Figure 10C , bottom vs . top ) . This could explain why the transition from three-state to sixteen-state designs does not bring about a very large difference in predicted mutations ( Figure 3 ) . Furthermore , these results would suggest that the evolution of multispecific proteins may be subject to a phenomenon of positive feedback , where once a protein becomes somewhat promiscuous , it can be virtually uninhibited in the expansion of binding partners similar to the ones it already binds [47] . This phenomenon could partially contribute to the high connectivities of hub proteins ( such as CaM ) , which result in the scale-free nature of protein-protein interaction networks [48] . Comparison of the general amino acid composition of the CaM binding interface sequences designed for interaction with one or more targets provides valuable insight into the evolutionary processes resulting in the contemporary CaM sequence . For example , Met residues , so abundant in the CaM binding interface , were frequently postulated to be key to its ability to interact with multiple targets . Met possesses a long and flexible side chain that can , in principle , adjust for interaction with any target [49] , [50] . In agreement with these observations , we show that the methionine content increases as we introduce additional interaction partners in our design procedure ( Figure 9 ) . We found a number of similar cases where the progression from single-state to multistate design converges on a sequence composition more similar to that of the evolutionary profile . For instance , the reduction in Arg content in multistate designs might result from the need for CaM to satisfy salt-bridge interactions with a number of targets . These targets show different , yet mostly positive , charge distributions; hence an Arg would be more difficult to place without destabilizing one of the CaM-target complexes . The reduction in aromatic residue content might be due to the fact that these residues need to fit in the hydrophobic pockets between CaM and the target . Since such pockets could be located in different places for the different CaM-target complexes , it would thus be difficult to provide sufficient space for aromatic amino acids in all contexts . In such cases , the compromise sequences might replace the aromatic amino acids with hydrophobic residues , such as Leu , Met , or Val , whose content increased in the transition to multistate design . The results of our computational design experiments on CaM can provide useful strategies for the experimental redesign of any multispecific protein [47] . To improve the affinity of a promiscuous protein to a particular target , we should not touch the affinity-defining positions , since these positions are already highly optimized and attempts to improve them are likely to fail . On the contrary , the specificity-defining positions in multispecific proteins are usually occupied by non-optimal amino acids . For proteins that undergo a large conformational change upon binding , energetic improvements in the intramolecular interactions at these positions ( Figure 5D ) should result in enhanced affinity by stabilization of the protein in the target-bound conformation [28] , [51] . Improvement of the intramolecular energies , however , is not likely to bring about an increase in binding specificity if interactions with different targets are conveyed through the same binding mode [52] . Optimizing the charged positions for a particular target , on the contrary , is bound to increase the protein binding specificity . Such optimization was previously used to drive the correct assembly of 4-helix bundles [53] and to substantially increase CaM binding specificity [28] . In addition , proper placement of charged residues is likely to be used by proteins to prevent folding into non-native structures [54] and to determine substrate specificity for enzymes [36] . The energetic analysis of all of the designed sequences ( Figure 10 ) , in the context of the sixteen CaM-target complex structures , helps to explain our previous experimental results on substantially increasing CaM binding specificity [17] , [28] . In these experiments , we optimized CaM for interaction with a single target without incorporating an explicit negative design procedure , i . e . , considering CaM interactions with alternative , undesirable targets . Unexpectedly , in the majority of cases we observed a significant decrease in CaM affinity to these other targets . There has been some controversy if one should consider negative design when designing a protein to be compatible with certain conformations [17] , [54]–[59] , since , as a designer , one wants to prevent the constructed sequence from folding into an alternative conformation . Our present analysis ( Figure 10B , top ) shows that the optimization of twenty CaM binding interface residues for a particular target is sufficient for substantially increasing ( worsening ) the interaction energy with other targets . Nevertheless , the necessity of incorporation of negative design is highly dependent on the problem [47]; optimizing a large number of residues and considering more dissimilar states increases the chances that positive design will suffice . In conclusion , our simulations give valuable insights as to how a prototypical multispecific protein , CaM , has evolved in nature to recognize a large number of binding partners . We uncovered both sequence and energetic tradeoffs that are imposed by multispecificity . Specifically , as additional CaM targets were explicitly incorporated in the design procedure , the resulting sequences were more similar to the native sequence ( Figure 11A ) . Conversely , the energies with which these sequences bind the targets most closely resemble that of the WT sequence ( Figure 11B ) . These compromises are likely to represent authentic trends in the evolution of proteins with a large number of binding partners . Our analysis also uncovered two classes of CaM interface positions: the affinity-determining positions , which stabilize the intramolecular interactions; and the specificity-determining positions , which interact strongly ( but distinctly ) with the various targets ( Figure 11C ) . Our computational results will help in guiding future experiments on the redesign of CaM and other multispecific binders . Additional biochemical and structural studies of promiscuous proteins should be used to validate our findings and provide greater detail about the mechanisms employed by these proteins in achieving their diverse biological functions .
A thorough search of the PDB revealed 24 solved structures of CaM-target complexes . Of these , 16 were of high resolution ( less than 2 . 5 Å for X-ray structures ) and exhibited the conventional CaM-target binding mode ( Figure 1A ) . For each structure , the interface positions were determined as those that are within 4 Å of the respective target peptide . The CaM positions found in the interface for at least 75% of the 16 structures were defined as the common binding interface , 20 in total: 11 , 14 , 15 , 18 , 19 , 36 , 39 , 41 , 51 , 71 , 72 , 84 , 87 , 88 , 92 , 109 , 112 , 124 , 144 , and 145 . All CaM structures were drawn using PyMOL [60] , and sequence logos were generated using TeXshade [61] and WebLogo [62] . For the multispecific design , the goal was to predict the 100 CaM interface sequences that minimize the sum of total energies in the target structures of the respective design scenario ( i . e . , 1- , 2- , 3- , or 16-state designs ) . Thus , there were 16 single-state designs ( one for each CaM-target interaction ) , 120 two-state design scenarios ( one for each pair of the 16 CaM-target interactions ) , 560 three-state designs ( one for each threesome of interactions ) , and one design of all sixteen states; this yielded 697 design scenarios in all . In each design scenario , the energies of the multiple states were uniformly weighted; for full details , see [7] . For all energy calculations , we used the ORBIT protein design force field [21] with the parameters previously used for redesign of CaM-target interactions [17] . In all subsequent design calculations , all positions were allowed to mutate to all 20 amino acids except cysteine and proline . In addition , for all structures , the peptides were allowed to vary their side chain conformations . Amino acid rotamers were defined based on the backbone-dependent rotamer library of Dunbrack and Karplus [63] , with sub-rotamers added at one standard deviation around the mean value; native sequence rotamers were included as well . We used a combined algorithmic strategy for finding the lowest energy sequences , employing the tBMMF algorithm [7] , [64] and the HERO module of ORBIT [22] and then extracting the best hundred sequences from their aggregated output . Briefly , the tBMMF algorithm provides a framework for predicting successive low energy sequences compatible with multiple protein structures . Firstly , a probabilistic graphical model is built that simultaneously models multiple protein structures of the same molecule ( by requiring that the sequences predicted for the multiple structures be identical ) . Then , tBMMF iteratively performs energy minimization ( using max-product belief propagation ) within a particular sub-space of amino acid sequences in order to find the next lowest energy sequence . It then partitions this sub-space into two sub-spaces , such that subsequent low energy sequences can be readily determined; for full details , see [7] , [64] . Note that only the tBMMF algorithm was capable of efficiently handling the 560 three-state designs . For the single case of 16-state design , tBMMF did not converge or yield reliable results . Therefore , the search over the sequence space was performed using a Monte Carlo simulated annealing ( MCSA ) algorithm [23]; at each step , a sequence was evaluated in each of the 16 complexes by calculating its minimal conformational energy using belief propagation [65] . This MCSA algorithm was repeated 10 times , for 2000 sequence steps each , and the 100 top-scoring sequences were extracted . Although we have previously shown that MCSA is often less successful at finding low energy sequences than the tBMMF algorithm [64] , it seems to have performed reasonably well in this case . The native interface sequence was extracted from the CaM structures . Evolutionary profiles were obtained by downloading and parsing the homologous sequence hits from the HSSP ( Homology-derived Secondary Structure of Proteins ) database [66] for each of the 16 structures and concatenating these profiles , yielding over 2100 homologues for the 20 CaM interface positions . To quantitatively compare amino acid probability distributions ( for a particular design position ) , we use the symmetric Jensen-Shannon divergence ( JSD ) . JSD , or dissimilarity scores , were used to measure correlation either between design results and HSSP ( e . g . , Figure 2 ) or between various design scenarios ( e . g . , Figure 5 ) . The JSD score ranges from 0 ( identical ) to 1 ( “distant” distributions ) , so that lower JSD scores reflect higher similarity between distributions [7] , [67] . The JSD between distributions and is given by: ( 1 ) where is the average distribution , and ( 2 ) is the Kullback-Leibler divergence between distributions and . In all cases ( except where noted ) , the mean JSD from the evolutionary profile ( HSSP ) for a particular CaM-target complex was calculated by averaging the JSD from the HSSP profile for all 20 interface positions . To delineate CaM interface positions critical for either target affinity or target specificity , we compared the best sequences designed for interactions with the 16 single targets . This was done by calculating the JSD dissimilarity score between all 120 pairs of the 16 single-state designs at each of the design positions . Positions for which at least 50% of the pairs have a JSD dissimilarity were defined as affinity-determining , and those where at least 50% of the pairs have a JSD dissimilarity were labeled specificity-determining . For each CaM position , the results shown ( Figure 5A ) are for those pairs of structures for which the position interacts with the target in both structures; results were similar when considering all pairs of structures ( not shown ) . Per-position energy contributions ( e . g . , Figure 5D , E ) were calculated using the EANAL module of the ORBIT program . For a particular 2-state design scenario , the profile based on its 100 lowest energy sequences was compared to those designed for interactions with the same two single targets ( 1-state designs ) . The comparison was performed at each of the 20 design positions . For each position , the multistate sequence compromise was categorized ( Figure 7 ) based on a JSD comparison between the two 1-state designs and between the same 1-state designs and the 2-state design . We defined 5 intuitive categories: “Kept same” - the 1-state designs predicted similar results ( pairwise JSD ) and the 2-state design was similar to both of them ( both pairwise JSD ) ; “Combined” - the 1-state designs were dissimilar ( pairwise JSD ) , but the 2-state design was similar to both of them ( JSD ) ; “Preferred one” - 2-state design was similar to only one of the 1-state designs ( JSD ) ; “New aa” - the 1-state designs predicted dissimilar results ( pairwise JSD ) and the 2-state design was different from both of them ( both pairwise JSD ) ; “despite same” - despite the 1-state designs predicting similar results ( pairwise JSD ) , the 2-state design was different from both of them ( JSD ) . For positions where the 2-state design “preferred one” of the 1-state designs or chose a new profile altogether ( “New aa” , “despite same” ) , we quantified to what degree this affected the biological quality of the sequence results ( Figure 8 ) . To do this , we first calculated the per-position JSD scores comparing the 2-state profile to HSSP . Then , we constructed the profile resulting from averaging the two 1-state design profiles and calculated its per-position JSD scores from HSSP . For a particular position , the difference between these JSD values ( ) was used to define the effect of multistate compromise: “No Change” - ; “Benefit” - ; “Loss” - . Recall that lower JSD scores from HSSP indicate greater similarity to the evolutionary profile , so that a decrease in JSD is termed beneficial . We chose to represent the performance of the two 1-state design scenarios using their average profile since , barring any external information , the most logical procedure would be to simply combine these two profiles as a proxy to the low energy sequence space compatible with both targets . For all calculations , we show results for those pairs of structures for which the position interacts with the target in both structures; results were similar when considering all pairs of structures . To characterize the tradeoff in energetic stability required for promiscuity , we quantified the changes in sequence energy resulting from the inclusion or exclusion of additional target states in the multispecific design procedure ( Figure 10 ) . Recall that the design results in this paper are based on the 100 lowest energy sequences for each of the 697 design scenarios detailed above , yielding a total of sequences . Firstly , we calculated the energy of each of these sequences in each of the 16 target structures ( over calculations in total ) . To efficiently perform these calculations , we utilized belief propagation ( and Monte Carlo simulated annealing if the belief propagation algorithm did not converge , see [64] ) to calculate the lowest energy rotamer conformation of each such sequence threaded onto the structure of each CaM-target complex . For each structure , the energy of a particular sequence was normalized by the absolute value of the energy of the best sequence designed for that structure . Then , for each combination of design scenario and structure , we chose the sequence with lowest normalized energy when threaded onto the structure , among the 100 sequences designed for that scenario . This yielded the final normalized energies ( corresponding to 697 sequences16 structures ) utilized for plotting Figure 10 and Figure 11C . Now , denote by A , B , C , and D any four arbitrary CaM states , i . e . , complexes of CaM with different targets . For all 12 histograms in Figure 10 ( 3 rows4 columns ) , we enumerate all possible choices of the corresponding CaM states . For each such choice , we calculated the designated differences in normalized energy , and all resulting values were plotted in the respective histogram . For example , in the bottom panel of Figure 10 ( row 3 , column 2 ) , consider each of the 16 CaM-target complexes as state A . Then , consider all triples of other possible states as B+C+D . Finally , calculate the difference in normalized energy in state A , between the sequence resulting from the simultaneous design of B , C , and D and the sequence resulting from the exclusive design of A . This difference , necessarily positive , was one of the 7280 values ( 16 choices for A 455 choices for B+C+D ) used in creating this frequency histogram .
|
In nature , some proteins are more social than others , interacting with a large number of partners . These “promiscuous” proteins play key roles in cellular signaling pathways whose disruption may lead to diseases such as cancer . The amino acid sequences of such proteins must have evolved to be optimal for combined interactions with all natural partners . However , the evolutionary process leading to this promiscuity is not fully understood . We address this subject by predicting amino acid sequences that would be most compatible for interaction with each partner on its own and those most compatible for binding multiple proteins . We find that these two types of sequences are substantially different , the latter more closely resembling the natural sequences of promiscuous proteins . We also find that promiscuous proteins contain certain regions that are necessary for interfacing with all of their partners , while other regions convey specific interactions with each particular target protein . We analyze the tradeoffs required for such proteins to bind multiple partners and find that only some degree of compromise is typically needed in order to permit interactions that are seemingly antagonistic . We conclude that the simulations reported here mimic well the natural evolution of proteins that associate with multiple partners .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"computational",
"biology/macromolecular",
"structure",
"analysis",
"molecular",
"biology/molecular",
"evolution",
"biochemistry/protein",
"folding",
"biophysics/theory",
"and",
"simulation",
"molecular",
"biology/bioinformatics",
"computational",
"biology/macromolecular",
"sequence",
"analysis"
] |
2009
|
Tradeoff Between Stability and Multispecificity in the Design of Promiscuous Proteins
|
Recent imaging studies of mitochondrial dynamics have implicated a cycle of fusion , fission , and autophagy in the quality control of mitochondrial function by selectively increasing the membrane potential of some mitochondria at the expense of the turnover of others . This complex , dynamical system creates spatially distributed networks that are dependent on active transport along cytoskeletal networks and on protein import leading to biogenesis . To study the relative impacts of local interactions between neighboring mitochondria and their reorganization via transport , we have developed a spatiotemporal mathematical model encompassing all of these processes in which we focus on the dynamics of a health parameter meant to mimic the functional state of mitochondria . In agreement with previous models , we show that both autophagy and the generation of membrane potential asymmetry following a fusion/fission cycle are required for maintaining a healthy mitochondrial population . This health maintenance is affected by mitochondrial density and motility primarily through changes in the frequency of fusion events . Health is optimized when the selectivity thresholds for fusion and fission are matched , providing a mechanistic basis for the observed coupling of the two processes through the protein OPA1 . We also demonstrate that the discreteness of the components exchanged during fusion is critical for quality control , and that the effects of limiting total amounts of autophagy and biogenesis have distinct consequences on health and population size , respectively . Taken together , our results show that several general principles emerge from the complexity of the quality control cycle that can be used to focus and interpret future experimental studies , and our modeling framework provides a road-map for deconstructing the functional importance of local interactions in communities of cells as well as organelles .
All eukaryotic cells contain mitochondria , whose number and structure can vary substantially during the cell cycle [1] and between cell types [2] . One of the primary functions of mitochondria is the production of ATP through utilization of the electrochemical potential across the inner membrane ( ) [3] . However , this process also produces reactive oxygen species ( ROS ) , which can damage DNA and proteins [4] . The input-output relation between membrane potential and ATP production is highly nonlinear [5] , [6] , suggesting the importance of maintaining high membrane potential within an optimal range that results in high levels of ATP production with minimal ROS production . Indeed , mitochondria with low membrane potential are selectively degraded within autophagosomes [7] . Mitochondria are dynamic organelles that undergo frequent cycles of fusion and fission , leading to the formation of complex networks distributed heterogeneously across the cell [8] . The ultrastructure of the mitochondrial network varies significantly with cell type . Skeletal muscle cells contain long filamentous mitochondrial networks , while immune cells can contain large spherical mitochondrial aggregates [8] , [9] . In normal rat kidney cells , mitochondria switch from a fragmented to a hyperfused state during the transition [1] . During periods of fusion , mitochondria exchange both soluble and membrane-bound components . The time scale of this exchange can vary significantly depending on the component being exchanged and its location in the mitochondrion; soluble matrix components mix rapidly ( within seconds ) , while inner-membrane proteins mix on a time scale of tens of minutes [10] . Recent fluorescence microscopy studies have explored the relationship between and mitochondrial fusion and fission in H9c2 cells , a subclone of a cell line derived from embryonic BD1X rat heart tissue that exhibits many of the properties of skeletal muscle [11] , by simultaneously quantifying the levels of a -sensitive dye and of matrix-targeted photo-activateable green fluorescent protein ( PA-GFP ) [12] . By photo-activating the PA-GFP in only a small proportion of the mitochondria within a given cell , fusion events can be directly observed via the transfer of activated PA-GFP into a dark matrix compartment and the equilibration of . Fusion was observed to selectively involve mitochondria with high [12] . In addition , approximately 30% of fusion events were followed by a statistically significant difference in the membrane potentials of the two daughter mitochondria ( ) ; the mechanism behind the generation of this asymmetry in is currently unknown . Importantly , this difference has a large impact on the relative capacities for ATP and ROS production of the two daughter mitochondria [6] . Similar microscopy experiments on H9c2 cells revealed that the probability of fusion of two mitochondria in close contact was strongly dependent on whether one or both of the mitochondria were being actively transported at the time of contact [13] . This study also distinguished two different classes of fusion , one that allows complete mixing of matrix components and another that allows only partial mixing ( transient fusion ) [13] . This transient fusion usually occurs when two mitochondria on different microtubules come into contact , while complete fusion occurs when mitochondria are either stationary or if they are on the same microtubule [13] . This partial diffusion occurs mainly for membrane proteins , since diffusion of ions and soluble matrix proteins is very fast over the micron length scales of individual mitochondria , and suggests an absence of active mixing during fusion . In support of the physiological importance of mitochondrial dynamics , the inhibition of fusion or fission leads to a deterioration of overall mitochondrial health in many organisms [14]–[16] . Deletion of mitochondrial fusion proteins Mfn1 and Mfn2 in mice is embryonically lethal , and the mitochondria in these mutant cells are fragmented and dysfunctional ( depolarized with low ATP production ) [16] . In the budding yeast Saccharomyces cerevisiae , deletion of the mitochondrial fusion protein FZO1 results in fragmentation of the mitochondrial network and loss of respiratory function [17] . A unifying hypothesis is that mitochondrial fusion-fission and autophagy form a quality-control axis that maintains mitochondrial health by segregating dysfunctional mitochondria into a separate pool of mitochondria that are unable to fuse , and undergo autophagy [14] . Recent evidence suggests that mitochondrial motility plays a key role in each of the steps associated with mitochondrial quality control , including fusion , fission , biogenesis , and lysosomal function . Motility is affected by both normal changes in the cell's physiological state ( e . g . , during the cell cycle ) and in pathologies such as neurodegeneration [18] , mutations in proteins such as the mitofusins [16] and Parkin [19] , and calcium homeostasis pathologies [18] . Therefore , consideration of motility and the spatial distribution of the mitochondrial network is critical to address the effects of pathological changes on mitochondrial quality control . A previous computational study of the effects of fusion , fission , and autophagy on mitochondrial health in the presence of ROS-mediated damage demonstrated that selective fusion and autophagy can result in stable average health of the mitochondrial population [20] . Moreover , post-fission asymmetry in led to a further health enhancement , by using selective autophagy to ratchet against damage . The model was used to demonstrate that selective fusion could be important as an isolation mechanism that allows autophagy to recycle dysfunctional mitochondrial components . A different model of the effects of reducing mitochondrial fusion as the cell ages proposed two separate modes of mitochondrial damage [21]: stochastic damage similar to Ref . [20] and infectious damage , which was conjectured to be the main cause of the rapid depolarization observed after fission [12] . These studies were important first steps in modeling mitochondrial network dynamics . However , they did not consider the spatial distribution of mitochondria , preventing them from addressing whether cytoskeletal motility and mitochondrial density strongly affect quality control . Moreover , their artificial implementations of fusion in which all unfused mitochondria could interact , regardless of position , made it more difficult to connect model kinetics to experimental parameters . Here , we present a computational model that incorporates the spatial dynamics of mitochondrial networks into the quality control cycle of the mitochondrial population . Similar to Ref . [20] , we simplify the physiology of mitochondrial function to a health parameter mimicking . We show that the average health of the network is maximized when the fusion and autophagy processes have a common selective health threshold , suggesting a common mechanism of -selectivity by the fusion and autophagy machinery . We also demonstrate that as long as the density of mitochondria is high enough to preserve a critical level of fusion , quality control is sufficient to maintain a healthy mitochondrial population . Moreover , the effect of any motion dependence of fusion can be understood simply as a shift in the effective fusion rate . In the absence of an active segregation mechanism , the molecular mechanism underlying the generation of asymmetry in must be relatively discrete to benefit health , involving at most only a few tens of units that can be exchanged . Finally , we find that restricting the maximal rate of autophagy primarily affects the health of the network , while restricting mitochondrial replication ( biogenesis ) primarily affects the population size . Taken together , these results can guide the interpretation of experiments by revealing the processes for which cellular health has the greatest sensitivity .
We encompass the diverse range of mitochondrial network properties by simulating the dynamics of mitochondria that undergo autophagy , replication , fusion , fission ( Fig . 1 ) , and transport along cytoskeletal filaments ( Fig . 2 ) . Each simulation represents a cell initialized with a given number of mitochondria ( Methods ) . Similar to a previous study [20] , we define a quantity for each mitochondrion that we will refer to as its “health , ” which is a proxy for . The ATP-producing potential of a mitochondrion is correlated with its [5] , [6] . When is below 180 mV; ATP production only becomes nonlinear under stress , when the mitochondrion is hyperpolarized and respiration is slowed [5] . Our simulations thus most closely mimic cells in unstressed conditions , when lower is directly correlated with lower ATP production . Mitochondria lose their metabolic function over the course of days [22] , although the specific cause of this loss is unclear . We treat the health of a mitochondrion as being composed of binary health units ( HUs ) that are either undamaged ( 1 ) or damaged ( 0 ) ; in each simulation , is fixed for all mitochondria . We model the effects of this damage on each HU as a Poisson process with constant rate that switches an HU from undamaged to damaged ( Fig . 1 ) , mimicking the detrimental effects of ROS production on mitochondrial function . To determine the health of a given mitochondrion , we compute the sum of all the HUs and divide by the maximum number ( ) . The limit of a continuous health quantity that mimics continuous can in principle be realized by taking to be very large . In our simulations , each mitochondrion behaves as an independent entity unless it is fused to another mitochondrion , in which case the fused mitochondria behave as one entity . Although fragmented mitochondria can have a distribution of sizes [23] , we treat all networks as composed of individual mitochondria of a fixed size representing the average of this distribution . For computational simplicity , we represent each mitochondrion as a circle with radius ( Fig . 2A ) [13] . Mitochondria are too large to diffuse in a densely packed cell at any appreciable rate , and consequently rely on active transport by molecular motors along cytoskeletal filaments [24] . In our model , we treat active transport as persistent diffusion along a dense microtubule network that allows travel in any direction ( Fig . 2B ) . In our simulations , we restrict motion to only unfused mitochondria . Each unfused mitochondrion is treated as either stationary or attached to a cytoskeletal filament . An unfused mitochondrion can bind to a microtubule at rate from anywhere in the cell , and the direction of motion is chosen at random from to ( Fig . 2B ) . An attached mitochondrion moves at a constant velocity until it detaches from the filament at rate or until it contacts another mitochondrion , at which time fusion can occur ( Fig . 2C , Methods ) . Upon fission , the daughter mitochondria again behave as independent entities unbound to a cytoskeletal filament . We consider the mitochondria to be spatially distributed across a two-dimensional , square-shaped cell , a reasonable approximation of common experimental imaging conditions with cells spread on adhesive surfaces [25] . We incorporate the necessity for direct contact during fusion by determining a fusion probability for each pair of mitochondria based on whether they are touching ( Methods ) . The explicit incorporation of mitochondrial proximity in fusion dynamics is a key feature of our model that distinguishes it from previous models [20] , [21] , and allows us to address experimental observations such as the dependence of fusion on intracellular mitochondrial transport . Selectivity is incorporated by only permitting fusion between mitochondria whose health exceeds a fusion threshold . Fused mitochondria undergo fission at a fixed rate ( ) selected to reproduce the observed duration of fusion events in H9c2 cells [14] . In order to simulate the effects of asymmetry after fissioning , we exchange a set number of randomly chosen HUs between the two fissioning mitochondria . This strategy simulates partial diffusion of contents between the fused mitochondria that is interrupted by fission . It also helps to partially retain the original mitochondrial identity , as observed in fluorescence microscopy experiments [12] , [13] , [26] , [27] . In these “kiss and run” events , fused mitochondria retain their pre-fusion spatial extents post-fission despite large changes in their membrane potentials , suggesting that such partial diffusion along with retention of mitochondrial identity is a critical component of mitochondrial function . During autophagy , mitochondria are engulfed by autophagosomes and delivered to lysosomes for lysis [7] ( Fig . 2D ) . Autophagy is selective; only dysfunctional mitochondria with lower and hence lower ATP production potential are targeted to autophagosomes [7] . Therefore , we model autophagy by removing mitochondria whose health is less than a given autophagy threshold at a fixed rate , where we have assumed that the number of autophagosomes is not limiting for mitochondrial turnover . In accordance with experimental observations [12] , we do not allow large mitochondria to undergo autophagy , since autophagosomes have a maximum size limit that we take to be a single , unfused mitochondrion . The removal of mitochondria through autophagy is balanced by biogenesis , or replication [2] . Although increases in mitochondrial mass typically occur through the process of continuous growth followed by fissioning , for simplicity we model replication as discrete events that occur at rate involving the instantaneous doubling in size and then fissioning of one mitochondrion into two ( Methods ) . Given that the mitochondrial biomass ( population size ) in many cells is roughly constant [2] , we assume that there is a metabolic limit to the replication rate in which replication asymptotically approaches zero as the population size increases according to ( 1 ) where represents the typical number of mitochondria and represents the variability in mitochondrial number . This implementation of autophagy and replication allows the mitochondrial number to vary with the kinetics of fusion , fission , and autophagy , unlike previous models that have imposed a fixed mitochondrial population size [20] . In order to study how the mitochondrial population size and average health are affected by changes to the processes captured by our model ( Fig . 1 ) , we simulated the network dynamics within cells seeded with an initial population of 150 unfused mitochondria . In each simulation , the mitochondria were initially uniformly distributed within a two-dimensional cell whose geometry we approximate as square with sides . Except where indicated , each mitochondrion has 10 HUs , two of which are exchanged at the time of fission . The initial state of each HU was chosen randomly , giving mitochondria with a distribution of healths between 0 and 1 and an average across the population of . Damage alone in the absence of fusion , fission , or autophagy resulted in an exponential decrease in the average mitochondrial health ( Fig . 3A ) . The further inclusion of autophagy led to a monotonic decline in the total number of mitochondria in the cell ( Fig . 3B ) . However , the average health decreased more slowly than in Fig . 3A , since autophagy selectively removed the unhealthy mitochondria from the cell . This decline in mitochondrial population size was counterbalanced by sufficiently high rates of biogenesis ( Fig . 3C ) . In this case , the average steady-state health stabilized at a value close to the autophagy threshold . As reported in Ref . [20] , the average health can be substantially improved by selective fusion and asymmetric fission in conjunction with damage , autophagy , and replication ( Fig . 3D ) . The exchange of HUs that occurs during fission stochastically leads to an asymmetry in the healths of the fissioning mitochondria . The probability of asymmetry development is highest when the two fused mitochondria have comparable healths , and when the number of HUs exchanged is half the total number ( Fig . 4 ) . Randomly exchanging units is equivalent to exchanging units since the identities of individual mitochondria can be switched without affecting the simulation . This dependence on exchanged units is similar across fusion rates , and the peak occurs at in all cases studied ( Fig . 4 ) . Given that both fusion and autophagy have been observed to occur in a manner , we used our model to comprehensively determine the effects of and on health . For every value of , our simulations indicated that the steady-state health was maximized when ( Fig . 5A ) . When , there is a large fraction of mitochondria with health between and that are removed through autophagy even when they are still capable of fusion and hence improvement of their health could have occurred through ratcheting ( Fig . 5Bii ) . When , the consequences on health are more stark: the cell is unable to maintain mitochondria with health substantially above the autophagy threshold and hence they are incapable of fusion ( Fig . 5iii ) . For thresholds above 0 . 6 , the maximal average steady-state health was only slightly above the threshold values due to diminishing benefits from asymmetric exchange . Thus , for the remainder of our study , we set the two thresholds to be equal at 0 . 3 and studied other conditions that yielded optimal mitochondrial health . To examine whether mitochondrial density affects the efficacy of quality control , we performed simulations in which we changed density by varying the size of the cell in order to maintain similar levels of stochasticity due to mitochondrial numbers in all cases . Interestingly , health began to saturate when the mitochondria occupied of the cell area , and the further increase was less than 10% when the filling fraction , , was increased to 55% ( Fig . 6A ) . Typical network structures at each density exhibited a sharp transition around 5% from fragmented mitochondria to extended , fused networks ( Fig . 6B , C ) , and this transition was closely coupled to a sharp increase in the number of fusion events per unit time ( Fig . 6D ) . At low density , the mitochondria have a low probability of contacting each other , and hence are generally unable to exploit the benefits of fusion and fission before their health drops below . For the remainder of this study , we set the cell size to correspond to an intermediate filling fraction of , when health is relatively high and insensitive to changes in density . Given experimental observations in H9c2 and rat insulinoma INS-1 cells that fusion is more likely to occur between mitochondria that are being actively transported along cytoskeletal filaments [13] , we sought to determine whether this behavior would result in higher mitochondrial health compared with a uniform rate of fusion . While the purpose of intracellular mitochondrial transport is thought to be for supplying ATP on demand at locations such as neurotransmitter or hormone release sites , here we study the effects of such transport on quality control without considering other aspects of such motion related to spatially dependent ATP production . We scaled the fusion rate such that two proximal mitochondria would have a fusion rate of , , or depending on whether both , one , or neither mitochondria were in motion at the time of fusion ( Fig . 7A ) . We considered only values of , in which the fusion rate was decreased if one or both mitochondria were not moving . For the parameters in Fig . 3D ( corresponding to ) , approximately 70% , 25% , and 5% of the fusion events resulted from these three classes of motility pairs , with an overall fusion frequency of 168 events/hr . Lower values of ( more motion dependence ) resulted in a decrease in health ( Fig . 7B ) , though this was also coupled to a decrease in the frequency of fusion events and the degree of reduction in health was dependent on the fusion rate . Average health was relatively unaffected by a ten-fold decrease in the transport velocity , and then decreased at very slow velocities ( Fig . 7C ) . We postulated that these decreases could be explained solely on the basis of the decrease in the effective fusion rate , and scanned a wide range of values of and . As hypothesized , the average steady-state health was highly related to the frequency of fusion events ( Fig . 7D ) , demonstrating that fusion is the primary determinant of health maintenance , and that transport affects health through its indirect modulation of fusion frequency . Biogenesis and autophagy are limited by protein synthesis , import across the mitochondrial membranes , and the number of autophagosomes . To implement these constraints in our model , we enforced maximal rates of autophagy and replication ( Methods ) . We observed that increasing the maximal autophagy rate from very low values exerted a large impact on average health ( Fig . 8A ) , indicating that maintenance of autophagy above a critical level may be crucial for removing damaged mitochondrial components , thereby allowing biogenesis of less-damaged mitochondria and leading to a healthier mitochondrial population . Over a wide range of the maximal autophagy rate , the cell maintained a constant number of mitochondria , while the steady-state average health saturated at a value determined by the damage rate ( Fig . 8A ) . In contrast , a high average health was maintained over a wider range of maximum replication rates ( Fig . 8B ) . However , decreasing the maximal replication rate reduced the size of the mitochondrial population . The transition from a stable to a decreasing population size occured in a range coincident with the sharp change in health due to caps on autophagy ( Fig . 8A ) , illustrating the role of damage in constraining both health and population size when there are physiological limitations on autophagy and biogenesis . For the passive exchange mechanism in our model that generates asymmetry in , the effectiveness of segregation is determined not only by the fraction of HUs that are exchanged , but also by the total number of HUs per mitochondrion ( ) , which dictates the discreteness of the health parameter . To keep all other behaviors the same while changing the level of discreteness , we determined the average steady-state health by simultaneously varying while scaling the number of HUs exchanged during fission . In the absence of fusion , the proportion of exchanged HUs did not affect health , but average steady-state health decreased as increased ( Fig . 9A ) . This occurred because the distribution of health narrowed ( Fig . 9Bi , iv ) and hence the replacement of a dysfunctional mitochondrion by a healthy mitochondrion with health was less effective for health maintenance . Health sharply dropped at ; past this amount , the average health was substantially below and hence the replication rate was not large enough to compensate for the loss of mitochondria due to autophagy . This underscores the importance of a discrete health variable even in the absence of HU exchange . The relationship between the total number of HUs and the average steady-state health was more complex when fusion was allowed . Due to the fact that fused mitochondria are unable to undergo autophagy ( Methods ) , higher fusion rates meant that a greater fraction of the mitochondria were temporarily protected from autophagy and hence cells could maintain an average steady-state health just below the autophagy threshold at higher values of . Nevertheless , as increased , the average steady-state health again declined ( Fig . 9A ) , indicating that the efficacy of HU exchange decreased . Moreover , a sharp drop in health still occurred when the average health dropped below due to the same narrowing of the distribution of health unit values ( Fig . 9B ) . Thus , our simulations indicate that a greater degree of discreteness of mitochondrial health results in higher health due to selective removal of mitochondria associated with the extremes of the distribution of individual healths .
The physical proximity of the mitochondrial DNA ( mtDNA ) , RNA , and proteins to the ROS produced by mitochondria as a side effect of respiration makes them susceptible to damage via chemical interactions . For example , free radicals , including ROS , can cause deletions and base-pair substitutions in DNA [4] , potentially rendering proteins encoded by the mtDNA ineffective or even detrimental . The formation of aggregates of misfolded proteins may also exert a toxic effect on the organelles , burdening metabolism [28] . To address cellular capacities for reducing mitochondrial dysfunction , we have developed a model of network dynamics that incorporates mitochondrial motion and spatial heterogeneity in two dimensions . Although this system of interactions between , autophagy , fusion , fission , and cytoskeletal-mediated transport is complex ( Fig . 1 , 2 ) , the generality of our model has allowed us to address the relative importance of many potential factors regulating mitochondrial health . In the presence of selective autophagy , the benefits of the generation of asymmetry during the fusion/fission cycle substantially boost mitochondrial health ( Fig . 3 , 4 ) . During each cycle , the potential to create a healthier and a less healthy mitochondrion from a pair of mitochondria with average health leads to a increase in the average health of the population once autophagy removes the mitochondrion of lower health from the population . In addition , we demonstrated that the optimal operating point for maximal health occurs when the thresholds and are equal ( Fig . 5 ) , suggesting that the molecular mechanism underlying fusion selectivity may act in opposite manners on the autophagy machinery . Indeed , there is evidence that mitochondrial autophagy and fusion are both linked to the transmembrane protein OPA1 , suggesting that its concentration may positively and negatively affect fusion and autophagy , respectively [27] . Within the extensive parameter space of our model , we have extracted several simple rules that determine the region of optimal mitochondrial health . Average health is only weakly dependent on density as long as a minimal level of fusion is maintained ( Fig . 6 ) . Moreover , imposing a motion dependence on the fusion rate mainly rescaled the effective fusion rate and thereby adjusted health ( Fig . 7 ) . These results both suggest that the frequency of fusion events is a relevant metric for evaluating the consequences of perturbations to mitochondrial dynamics on health , and motivate its quantification in future experimental studies . Moreover , the cell may have limits on the absolute number of fusion events , due to the costs of expression for fusion and fission proteins . Our simulations demonstrate that the cell can enhance fusion rates by actively regulating mitochondrial density and motility . Illustrating the importance of autophagy and replication to health maintenance , our simulations also showed that limiting the maximal rate of autophagy to below the damage rate resulted in a sharp decrease in health , while limiting the replication rate caused a decline in population numbers without affecting health ( Fig . 8 ) . Where possible , we have used experimentally relevant parameters in our model , and our results are generally robust to changes of large changes of parameters . Moreover , specific details of a given experimental system such as experimentally measured kinetics can be easily incorporated to investigate differences in behavior [29] , [30] . Future experiments in a wide range of cell types will help to elucidate which behaviors are universal , for which our model will provide a framework for revealing the underlying mechanisms . The membrane potential is an experimental readout of the ability of mitochondria to produce ATP . In the interests of simplicity and generality , we have used the health parameter as a proxy for , which itself is a function of many components such as the electron transport chain , the number of ATPases utilizing , proton leaks across the inner mitochondrial membrane , and other geometric factors [31] . Although the health of a mitochondrion may be directly correlated with , it would be better represented as a discrete variable if it is a function of a discrete underlying quantity affecting , such as the healthy mtDNA copy number [32] . In the absence of an active mechanism for the development of asymmetry , our simulations indicate that the underlying quantity being exchanged during fission must be a unit that exists in small numbers ( Fig . 9 ) , such as a mtDNA molecule or a protein aggregate . By condensing “health” into a single parameter , we have systematically analyzed the effects of mitochondrial network dynamics . However , our current model does not address complementarity , the fusion-mediated exchange of different levels of various proteins and solutes between mitochondria , which has also been suggested as another potential benefit of mitochondrial fusion and fission [9] . In future work , this aspect could be explored by increasing the number of parameters that define the health of a mitochondrion . This study lays the framework for future simulations of the role of mitochondrial network dynamics in the quality control of mitochondrial health . In comparison with previous models [20] , [21] , which have proven useful for elucidating the systems-level importance of mitochondrial dynamics , we have specifically incorporated the need for spatial proximity and the role of motility in fusion , and thus have the capacity to study how the spatial extent and motility of the mitochondria affect quality control . Furthermore , we have explicitly modeled biogenesis as a separate metabolic process , allowing us to address how population size is stabilized in the presence of mitochondrial turnover . Although we have made simplifying assumptions such as a two-dimensional framework , a square cell geometry , and the absence of intracellular structures such as organelles , our model is easily extended to address more complex cell morphologies and structures that sterically exclude mitochondria . Future work will focus on specific cellular phenomena , such as the spatial heterogeneity of the cytoskeleton and autophagosomes . In addition , our model was implemented to mimic a relatively static cell that maintains a roughly constant number of mitochondria; future investigations will include the increase in mitochondria required for growing and dividing cells . Finally , mitochondrial turnover provides a model for how interactions among the members of a community can enhance collective fitness at the cost of sacrificing individual members , and a computational framework for studying such systems may similarly illuminate key features of system optimality .
For each set of parameters ( Table 1 ) , we carried out 25 independent simulations to approximate the size of a population of cells and the level of stochasticity present in a typical experimental study of mitochondrial network dynamics [27] . Each simulation represented a cell initialized with 150 mitochondria , an intermediate population size across different mammalian cell types [33] . The mitochondria were initially distributed randomly throughout the cell such that they did not overlap . The autophagy process removes a mitochondrion from the simulation; since fused mitochondria are not subject to autophagy in our simulations , our definition of an individual mitochondrion is the largest possible unit for autophagy , and varying that unit size would be similar to varying the mitochondrial population size or density ( Fig . 6 ) . The replication process selects a mitochondrion at random and generates a new mitochondrion with the same health as its progenitor . For the replication rate in Eq . 1 , we set ; our results were insensitive to changes in . Two strategies were tested for the placement of a new mitochondrion: randomly throughout the cell , or nearby the progenitor . In the latter case , a location within a certain radius of the progenitor is selected at random that does not overlap with other mitochondria . Both strategies converged on quantitatively similar results and hence we used the first strategy for simplicity . In most simulations , the frequencies of autophagy and replication were determined solely by the number of mitochondria below and the total size of the mitochondrial population , respectively . However , for the simulations in Fig . 8 , we tracked the cumulative number of autophagy and replication events in each cell and temporarily halted either process when more events had taken place in the preceding three hours than was permitted by the maximum allowed rates . This restriction was only lifted when the running event counter fell below the allowed number , thus preserving the stochastic nature of autophagy and replication while limiting the allowed number of events . Mitochondrial transport was simulated as a persistent diffusion process in which mitochondria bind to cytoskeletal filaments and move at a constant velocity until they either stochastically unbind from the filament or run into another mitochondrion or the edge of the cell . For simplicity , we ignored any organelles or other spatial features inside the cell , such as the nucleus , that might impede mitochondrial transport; such features would be straightforward to introduce in future extensions of the model . During each time step , each pair of touching mitochondria had the potential for fusion . For most simulations , the probability of fusion for touching mitochondria was . However , in Fig . 7 , we also maintained a record of whether the mitochondria came into contact via both , one , or neither being in motion in the previous time step , and scaled the probability by 1 , , and , respectively .
|
Mitochondria are the powerhouses of eukaryotic cells , oxidizing glucose to produce ATP . Most cells harbor tens to hundreds of mitochondria in a constant state of flux , in which they fuse with one another , undergo fission , import proteins to grow larger , and eventually are recycled by autophagy . These dynamic processes depend on the electrical potential that is maintained across the mitochondrial inner membrane and powers the production of both ATP and detrimental reactive oxygen species . How do mitochondria maintain high membrane potential in the face of damage due to reactive oxygen species ? Here , we develop a model to study how the reorganization of mitochondrial networks in space and time due to fusion , fission , and the experimentally observed development of membrane potential asymmetry after fission affect overall mitochondrial health . We show that health , which is a proxy for the mitochondrial membrane potential , is dominated by how density and motility affect the frequency of fusion events , and that several simple rules for the system kinetics lead to optimal quality control . This model predicts general behaviors that can be applied to specific studies of mitochondrial dynamics in a wide variety of cell types , and provides a framework for deconstructing complex organellar organization and their function in human disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"systems",
"biology",
"biophysic",
"al",
"simulations",
"biology",
"computational",
"biology"
] |
2013
|
Optimal Dynamics for Quality Control in Spatially Distributed Mitochondrial Networks
|
The SWR1 chromatin remodeling complex , which deposits the histone variant H2A . Z into nucleosomes , has been well characterized in yeast and animals , but its composition in plants has remained uncertain . We used the conserved SWR1 subunit ACTIN RELATED PROTEIN 6 ( ARP6 ) as bait in tandem affinity purification experiments to isolate associated proteins from Arabidopsis thaliana . We identified all 11 subunits found in yeast SWR1 and the homologous mammalian SRCAP complexes , demonstrating that this complex is conserved in plants . We also identified several additional proteins not previously associated with SWR1 , including Methyl-CpG-BINDING DOMAIN 9 ( MBD9 ) and three members of the Alfin1-like protein family , all of which have been shown to bind modified histone tails . Since mbd9 mutant plants were phenotypically similar to arp6 mutants , we explored a potential role for MBD9 in H2A . Z deposition . We found that MBD9 is required for proper H2A . Z incorporation at thousands of discrete sites , which represent a subset of the genomic regions normally enriched with H2A . Z . We also discovered that MBD9 preferentially interacts with acetylated histone H4 peptides , as well as those carrying mono- or dimethylated H3 lysine 4 , or dimethylated H3 arginine 2 or 8 . Considering that MBD9-dependent H2A . Z sites show a distinct histone modification profile , we propose that MBD9 recognizes particular nucleosome modifications via its PHD- and Bromo-domains and thereby guides SWR1 to these sites for H2A . Z deposition . Our data establish the SWR1 complex as being conserved across eukaryotes and suggest that MBD9 may be involved in targeting the complex to specific genomic sites through nucleosomal interactions . The finding that MBD9 does not appear to be a core subunit of the Arabidopsis SWR1 complex , along with the synergistic phenotype of arp6;mbd9 double mutants , suggests that MBD9 also has important roles beyond H2A . Z deposition .
Nucleosomes , the fundamental units of chromatin that consist of ~147 bp of DNA wrapped around a histone octamer , efficiently condense large eukaryotic DNA molecules inside the nucleus . At the same time , nucleosomes present a physical barrier that restricts the access of DNA-binding proteins to regulatory sequences . This physical constraint imposed by nucleosomes on DNA can be modulated to expose or occlude regulatory DNA sequences , and is thereby used as a mechanism to control processes such as transcription that rely on sequence-specific DNA binding proteins . Thus , enzymatic complexes that can remodel chromatin structure by manipulating the position and/or composition of nucleosomes are essential for proper transcriptional regulation and the execution of key developmental programs . All chromatin-remodeling complexes ( CRCs ) contain a DNA-dependent ATPase catalytic subunit that belongs to the SNF2 family of DNA helicases , along with one or more associated subunits [1 , 2] . There are four major subfamilies of CRCs: SWI/SNF , INO80 , ISWI , and CHD , all of which use the energy of ATP to either slide , evict , or displace nucleosomes , or to replace the canonical histones within nucleosomes with histone variants . One member of the INO80 CRC subfamily is the SWI2/SNF2-related 1 ( SWR1 ) chromatin remodeler , a multisubunit protein complex required for incorporation of the H2A variant , H2A . Z , into chromatin [3 , 4] . H2A . Z is a highly conserved histone variant found in all eukaryotes that plays important roles in regulating a variety of cellular processes , including transcriptional activation and repression , maintenance of genome stability and DNA repair , telomere silencing , and prevention of heterochromatin spreading [5–12] . Although loss of H2A . Z is not lethal in yeast [2 , 13] , H2A . Z is essential for viability in other organisms such as Tetrahymena [14] , Drosophila [15 , 16] , and mice [17] . Interestingly , H2A . Z-deficient plants are viable but display many developmental abnormalities such as early flowering , reduced plant size , altered leaf morphology , and reduced fertility [18–20] . The SWR1 complex that mediates H2A . Z incorporation into chromatin was first described in yeast and is composed of 13 subunits , including Swr1 , the catalytic and scaffolding subunit of the complex [3 , 4 , 21 , 22] . In mammals , the functional and structural homolog of yeast SWR1 complex is the SRCAP ( SNF2-related CREB-binding protein activator protein ) complex . This complex is composed of 11 of the same subunits found in yeast SWR1 and is likewise able to exchange H2A/H2B dimers for H2A . Z/H2B dimers in nucleosomes [23–26] . Intriguingly , higher eukaryotes possess an additional multi-subunit complex , 60 kDa Tat-interactive protein ( TIP60 ) , that has histone acetyltransferase ( HAT ) activity and can also mediate the deposition of H2A . Z into nucleosomes [27–30] . Furthermore , the yeast Nucleosome Acetyltransferase of histone 4 ( NuA4 ) complex , which is structurally related to SWR1 through the sharing of four subunits [31–35] , was shown to regulate the incorporation of H2A . Z into chromatin cooperatively with the SWR1 complex [34] . Many homologs of yeast SWR1 and animal SRCAP complex subunits have been identified in Arabidopsis thaliana including ACTIN RELATED PROTEIN 6 ( ARP6 ) , SWR1 COMPLEX SUBUNIT 2 ( SWC2 ) , SWC6 , PHOTOPERIOD-INDEPENDENT EARLY FLOWERING 1 ( PIE1 ) , and three H2A . Z paralogs: HTA8 , HTA9 , and HTA11 . Numerous genetic and biochemical experiments suggest that the SWR1 complex is conserved in Arabidopsis . For example , it has been recently shown that Arabidopsis SWC4 protein directly interacts with SWC6 and YAF9a , two known components of the SWR1 complex [36] . Additionally , protein interaction experiments have demonstrated that PIE1 interacts directly with ARP6 , SWC6 , HTA8 , HTA9 , and HTA11 [18 , 20 , 37 , 38] , suggesting that PIE1 may serve as the catalytic and scaffolding subunit of an Arabidopsis SWR1-like complex . Furthermore , functional characterizations of PIE1 , ARP6 and SWC6 have revealed that mutations in these genes have similar pleiotropic effects on Arabidopsis development , including a loss of apical dominance , curly and serrated rosette leaves , early flowering due to reduced expression of FLOWERING LOCUS C ( FLC ) , altered petal number , and reduced fertility [18 , 37–43] . Interestingly , genetic experiments revealed that the pie1 null phenotypes are more severe than those of arp6 , swc6 , hta9;hta11 , or hta8;hta9;hta11 ( h2a . z near-null ) plants [18 , 19 , 37–43] . The more dramatic phenotypes in pie1 plants suggest that PIE1 has additional functions outside of H2A . Z deposition by SWR1 , as previously proposed [5 , 6 , 18 , 19] . A recent report also showed that mutant plants null for pie1 and h2az exhibited early developmental arrest , dying shortly after germination [19] , further supporting the notion that PIE1 has H2A . Z-independent functions in Arabidopsis . On the other hand , genetic analyses of pie1;swc6 double mutant plants revealed that they had a phenotype indistinguishable from pie1 single mutants [38] , and arp6;swc6 plants displayed the same defects as either arp6 or swc6 single mutant plants [18 , 37] . These results further support the idea that PIE1 , ARP6 , and SWC6 act in the same genetic pathway and/or are the components of the same protein complex , but that PIE1 has additional functions . Despite the strong genetic and biochemical evidence that Arabidopsis contains many conserved subunits homologous to the components of the yeast SWR1 complex and mammalian SRCAP , the plant SWR1 complex has not been successfully isolated and characterized . Recently , Bieluszewski and colleagues used Arabidopsis SWC4 and ARP4 proteins , two subunits shared between yeast SWR1 and NuA4 , as well as mammalian SRCAP and TIP60 complexes , as baits to affinity purify their interacting partners from Arabidopsis cell suspension cultures [44] . These studies identified most of the subunits normally found in the SWR1 and NuA4 complexes , but it was not possible to determine whether this collection of proteins represented a single large complex or multiple complexes . Overall , it is not yet clear whether plants possess separate SWR1 and NuA4 complexes , SWR1 and TIP60-like complexes , or all three complexes [44] . The main goal of our study was to purify the Arabidopsis SWR1 complex in order to identify all of its components . To achieve this , we used the ARP6 protein , a subunit unique to SWR1 in other organisms , as bait in Tandem Affinity Purification ( TAP ) experiments . We performed three independent TAP experiments to isolate and identify ARP6-associated proteins . We identified all 11 conserved subunits found in yeast SWR1 and mammalian SRCAP complexes , demonstrating that Arabidopsis contains a bona-fide functional and structural homolog of these complexes . In addition , we identified several unexpected proteins that associated with ARP6 , including the plant homeodomain ( PHD ) - and Bromo domain-containing protein Methyl CpG-BINDING DOMAIN 9 ( MBD9 ) , and the PHD domain-containing proteins ALFIN-LIKE 5 ( AL5 ) , AL6 , and AL7 . Association of these proteins with the SWR1 complex is consistent with the results of a recent related study that used ARP6-MYC or -FLAG tag epitope purification followed by mass spectrometry to identify ARP6-interacting proteins ( Potok et al . 2019 , bioRxiv 10 . 1101/657296 ) . Genetic analyses revealed that mbd9 mutants showed phenotypic similarities to arp6 mutants , so we further explored a possible role for MBD9 in regulating H2A . Z incorporation into chromatin . We found that MBD9 is required for H2A . Z incorporation at a subset of the sites that normally harbor H2A . Z nucleosomes , and that these MBD9-dependent H2A . Z sites have distinct chromatin features . Furthermore , MBD9 strongly interacted with acetylated histone H4 peptides , H3 peptides mono- or dimethylated at lysine 4 , as well as H3 dimethylated at arginines 2 or 8 . We also found that MBD9 is not a core subunit of the Arabidopsis SWR1 complex and that double mutant arp6;mbd9 plants exhibited much more severe phenotypes than single arp6 or mbd9 mutants . These results collectively suggest that MBD9 targets the SWR1 complex to a subset of genomic loci but also has important functions beyond H2A . Z deposition .
To isolate the Arabidopsis SWR1 complex , we decided to use ARP6 protein as bait in Tandem Affinity Purification ( TAP ) experiments because ARP6 is exclusively found in the SWR1 complex in other organisms and is not shared by any other known CRCs [4 , 18 , 34] . For our purification experiments we used a GSrhino TAP-tag , which consists of two protein G domains , a tandem repeat of rhinovirus 3C protease cleavage site , and the streptavidin-binding peptide . This tag has been successfully used to purify several plant nuclear complexes , including SWI/SNF type CRCs [45] . Furthermore , the use of this tag provides high yield of purified proteins and specificity of purification . In addition , a list of 760 proteins that nonspecifically bind to this tag or the associated purification beads has been assembled from data on 543 GS-based TAP experiments [45] . We fused the GSrhino TAP-tag to either the N-terminal end ( N-TAP-ARP6 ) or the C-terminal end ( C-TAP-ARP6 ) of the genomic ARP6 coding sequence , containing the endogenous ARP6 promoter , and introduced the constructs into arp6-1 mutant plants to test for the ability of each transgene to complement a null arp6 allele . Using western blotting , we first detected the presence of the 67 kDa ARP6-TAP-tag fusion protein specifically in plants homozygous for the transgene and not in arp6-1 or wild-type ( WT ) plants ( Fig 1A ) . Next , we assessed the ability of the transgenes to rescue the morphological defects of arp6-1 plants . When grown in parallel , the transgenic plants appeared almost indistinguishable from WT plants , with more compacted , non-serrated rosette leaves as compared to arp6-1 mutants ( Fig 1B ) . The N-TAP-ARP6 and , to a lesser degree , the C-TAP-ARP6 transgenes are able to rescue the early flowering phenotype of arp6-1 plants ( Fig 1C ) as evident by the significantly higher number of rosette leaves at the time of flowering in wild type and transgenic plants compared to arp6-1 plants ( Fig 1D ) . Finally , all transgenic plants showed full complementation for the loss of apical dominance and fertility defects of arp6-1 mutant plants ( Fig 1E and 1F ) . Overall , we conclude that the N-TAP-ARP6 and C-TAP-ARP6 transgenes are fully functional and thus suitable for affinity purification . Since the N-TAP-ARP6 and C-TAP-ARP6 transgenes were fully functional , we proceeded with the affinity purification experiments using two independent N-TAP-ARP6 transgenic lines ( N-TAP 11–4 and N-TAP 1–2 ) and one C-TAP-ARP6 line ( C-TAP 10–2 ) . We followed the protocol described by Van Leene and colleagues to purify and elute the ARP6-TAP-tag interacting proteins and identified the eluted proteins by liquid chromatography coupled with tandem mass spectrometry ( LC-MS/MS ) [45] . All eluted proteins detected in our three TAP-tag experiments are listed in S1 Table . Using the database of non-specific binders of the GSrhino TAP-tag , we eliminated many proteins from this list as false positives and compiled a list of ARP6-interacting proteins . Among these proteins , we identified ARP6 , SWC2 , SWC4 , SWC6 , PIE1 , RuvB1 , RuvB2 , ACTIN1 , ARP4 , YAF9a , and H2A . Z proteins as Arabidopsis homologs of all 11 conserved subunits found in both yeast SWR1 and mammalian SRCAP complexes ( Table 1 ) . While we were able to detect HTA9 or HTA11 in all of our TAP-tag experiments , we never detected HTA8 , the third member of the H2A . Z family . This finding is perhaps not surprising considering the fact that HTA8 has the lowest expression in all tissues when compared to the other H2A . Z family members in Arabidopsis ( S1A Fig ) . We also did not detect the YAF9b protein ( Table 1 ) even though Arabidopsis YAF9a and YAF9b have been previously shown to act redundantly and are required for proper FLC expression [44 , 46 , 47] . In addition to H2A . Z , we identified three H2B histones in our TAP experiments: HTB2 , HTB4 , and HTB9 ( Table 1 ) . Since H2A . Z histones are deposited into nucleosomes as H2A . Z/H2B dimers , we sought to investigate whether specific H2A . Z proteins might have preferential H2B partners . If this is true , we would expect to see synchronized expression of specific H2A . Z/H2B pairs in various Arabidopsis tissues . Using publicly available microarray expression data [48] for the two H2A . Z and three H2B histones that we identified , we observed that HTA11 and HTB2 had highly similar expression profiles across tissues ( S1B Fig ) , while the expression of HTA9 matched very well with HTB4 expression and , to a slightly lesser degree , with HTB9 expression ( S1C and S1D Fig ) . These results indicate that the Arabidopsis H2A . Z histones may have preferential H2B partners when deposited as dimers into nucleosomes . Interestingly , in addition to known subunits of the SWR1 complex , we also reproducibly identified several other nuclear proteins as being associated with the SWR1 complex ( Table 1 ) . These include MBD9 , a protein with a methyl-CpG-binding domain and various chromatin-binding domains [49–51] , TRA1A and TRA1B , two subunits in the SPT module of an Arabidopsis SAGA complex [52] that are also homologs of the yeast NuA4 subunit Tra1 and the mammalian TIP60 subunit TRRAP [34 , 44] , and three members of a plant-specific Alfin1-like family ( AL5 , AL6 , and AL7 ) best known for their regulation of abiotic stress responses in Arabidopsis and ability to bind di- and tri-methylated lysine 4 of histone H3 ( H3K4me2/3 ) [53 , 54] . MBD9 , a methyl-CpG-binding domain-containing protein , was identified in all three TAP-tag experiments as an ARP6-interacting partner ( Table 1 ) . Previous studies have shown that mbd9 mutants flowered significantly earlier than WT plants , due to reduced FLC expression , and produced more inflorescence branches when compared to WT plants [50] , which are phenotypes also found in arp6 mutants [40] . We discovered that , in addition to the above-mentioned defects , mbd9 plants have serrated rosette leaves and a significantly increased number of flowers with extra petals ( S2 Fig ) , which are phenotypes also associated with the loss of ARP6 [18 , 39 , 40] . Furthermore , examination of the previously reported MBD9 enrichment pattern at the FLC locus revealed that MBD9 occupied the two FLC regions previously shown to also have the highest H2A . Z enrichment in WT plants [41 , 51] . Given that arp6 and mbd9 plants have similar phenotypes , that H2A . Z and MBD9 appear to occupy the same sites at the FLC locus , and that MBD9 co-purified with ARP6 in our TAP-tag experiments , we investigated whether MBD9 plays a role in the incorporation of H2A . Z into chromatin . We performed three biological replicates of chromatin immunoprecipitation coupled with high-throughput sequencing ( ChIP-seq ) using an H2A . Z antibody [41] on WT , arp6-1 , and mbd9-1 seedlings . At least 8 million high-quality , nuclear reads were aligned to the TAIR10 genome for each replicate , resulting in thousands of reproducible H2A . Z peaks for each genotype ( S3A Fig ) . Spearman correlation analysis indicated a high degree of overlap among H2A . Z ChIP replicates for a given genotype with the exception of mbd9-1 replicate 3 , which correlated moderately with other two mbd9 replicates ( S3B Fig ) . The average H2A . Z enrichment profile ( the average of all three scaled replicates ) across all gene bodies in WT plants showed the highest enrichment of H2A . Z just after the transcription start site , with decreasing enrichment toward the 3’ end , as expected ( Fig 2A ) . The pattern of H2A . Z enrichment across genes in arp6-1 showed a similar profile but with extremely reduced enrichment , while mbd9-1 plants had an intermediate level of H2A . Z enrichment between WT and arp6-1 ( Fig 2A ) . To analyze all of the regions normally enriched for H2A . Z , we identified peaks of H2A . Z enrichment that were present in at least two of the three H2A . Z ChIP-seq replicates in WT plants , and we then examined H2A . Z levels at these same sites in arp6-1 and mbd9-1 mutants . As observed for gene bodies , the 7039 sites reproducibly enriched for H2A . Z in WT were nearly depleted of H2A . Z in arp6-1 , while there was an intermediate H2A . Z depletion in mbd9-1 plants ( Fig 2B ) . Western blotting for H2A . Z on acid-extracted nuclear proteins isolated from WT , arp6-1 , and mbd9-1 showed a decrease in H2A . Z levels in both arp6-1 and mbd9-1 , suggesting that the lack of chromatin incorporation may lead to H2A . Z degradation ( S4A and S4B Fig ) . To test whether the observed reduction of H2A . Z incorporation might be due to either SWR1 components or H2A . Z genes being misregulated in mbd9 plants , we performed qRT-PCR experiments to measure the expression of PIE1 , ARP6 , MBD9 , SWC4 , SWC6 , and YAF9a , as well as HTA8 , HTA9 , and HTA11 genes , in WT , arp6-1 , and mbd9-1 plants . We found that the expression of these genes was not substantially altered and , therefore , unlikely to account for the observed depletion of H2A . Z levels in mbd9-1 plants ( S5 Fig ) . Using RNA-seq , a related study has also demonstrated that the expression of SWR1 subunit genes was largely unaffected in mbd9-3 plants ( Potok et al . 2019 , bioRxiv 10 . 1101/657296 ) . Overall , these results indicate that MBD9 is indeed required for proper H2A . Z incorporation into chromatin . To further confirm our results with respect to the role of MBD9 in H2A . Z deposition , we performed ChIP-qPCR experiments using WT , arp6-1 , mbd9-1 , and two additional mbd9 T-DNA alleles , mbd9-2 and mbd9-3 [50] . We first assayed H2A . Z abundance at two distinct regions of the FLC gene: the first and last exon ( regions 2 and 9 , respectively , as described in [41] ) . Regions 2 and 9 are the sites on the FLC gene where H2A . Z is most highly enriched in WT plants , and that enrichment is lost in arp6-1 mutant plants ( S6A Fig , [41] ) . We found that in plants homozygous for any of the three mbd9 alleles , the amount of H2A . Z at FLC regions 2 and 9 was reduced at least 2-fold when compared to WT plants ( S6A Fig ) , indicating that MBD9 contributes to H2A . Z deposition at the FLC gene . We also measured the H2A . Z abundance at ASK11 and At4 , two phosphate starvation response genes previously shown to have H2A . Z deposited in their chromatin [55] . We discovered that in mbd9 plants these genes were depleted of H2A . Z to similar levels as in arp6-1 plants when compared to WT ( S6B Fig ) . Taken together , our results indicate that MBD9 is required for proper H2A . Z levels at multiple Arabidopsis genomic loci and is , therefore , functionally related to the SWR1 complex . As MBD9 was previously reported to have histone acetyltransferase ( HAT ) activity in vitro and was found to associate with acetylated H4 [51] , we also examined the global levels of histone H4 N-terminal acetylation ( H4Ac ) in WT , mbd9-1 , and arp6-1 plants . At least 5 million high-quality , nuclear reads were aligned to the TAIR10 genome for each replicate , resulting in thousands of reproducible , H4Ac peaks for each genotype ( S3A Fig ) . Spearman correlation analysis indicated a high degree of overlap among H4Ac ChIP replicates for a given genotype with the exception of mbd9-1 replicate 3 , which correlated moderately with other two mbd9-1 replicates due to having the lowest levels of H4Ac among the replicates ( S3C and S7 Figs ) . If MBD9 is responsible for global acetylation of H4 , as previously reported [51] , we would expect to see a dramatic reduction of acetylated H4 in mbd9-1 mutants compared to WT . However , we found that the average genome-wide distribution of acetylated H4 was only modestly reduced in mbd9-1 plants , while arp6-1 plants had H4Ac levels similar to WT when examined across all gene bodies or at all sites enriched for acetylated H4 in WT plants ( Fig 2C and 2D ) . Western blot analysis using H4Ac antibodies on acid-extracted nuclear proteins from WT , arp6-1 , and mbd9-1 plants was consistent with the ChIP-seq findings that mbd9-1 plants have moderately reduced levels of H4Ac when compared to WT ( S4A and S4C Fig ) . Collectively , our ChIP-seq results demonstrate a major role for MBD9 in maintaining proper H2A . Z levels . MBD9 may also have a minor role in the acetylation of H4 , but this is likely to be indirect given that the protein does not contain an identifiable acetyl CoA-binding or acetyltransferase domain . The intermediate loss of H2A . Z in mbd9-1 plants compared with arp6-1 plants ( Fig 2A and 2B ) suggested that MBD9 may be required for incorporation of H2A . Z at only a subset of H2A . Z-enriched sites or it may be required for complete H2A . Z deposition at all sites . To identify genomic regions that require MBD9 for H2A . Z incorporation into chromatin , we quantified the normalized H2A . Z ChIP-seq read abundance from WT , arp6-1 , and mbd9-1 mutant plants across all of the H2A . Z-enriched regions that were reproducibly identified in the WT replicates . We performed DESeq analysis [56] to quantitatively compare WT to mbd9-1 and WT to arp6-1 H2A . Z levels at each site ( S2 Table ) . H2A . Z levels were significantly depleted in arp6-1 at nearly all of the H2A . Z sites , as expected for a mutation that disrupts the SWR1 complex ( Fig 3A ) . In contrast , out of the 7039 H2A . Z-enriched sites , we identified only 1391 sites that had reduced H2A . Z by at least 1 . 5-fold ( log2 fold change of at least 0 . 6 with an adjusted p value ≤ 0 . 05 ) in mbd9-1 compared to WT ( Fig 3B ) . To further examine the nature of the H2A . Z deposition defect in mbd9 mutants , we visualized H2A . Z enrichment and distribution across the 1391 sites that lose H2A . Z in mbd9-1 , which we refer to as MBD9-dependent H2A . Z sites . For comparison , we selected a similarly sized set of MBD9-independent H2A . Z sites ( 1505 sites with an average fold difference of less than 1 . 19 between WT and mbd9-1 , which is an absolute log2 fold change of less than 0 . 25 ) . This analysis revealed a drastic reduction in H2A . Z occupancy at each of the MBD9-dependent H2A . Z sites when comparing WT and mbd9-1 , but with maintenance of the same overall pattern of occupancy ( Fig 3C ) . The level of H2A . Z in mbd9-1 at MBD9-dependent sites was depleted to levels similar to arp6-1 . In contrast , the MBD9-independent H2A . Z sites showed equivalent profiles and occupancy levels between WT and mbd9-1 , but still showed reduced H2A . Z in arp6-1 ( Fig 3C ) . To further validate these results , we chose three MBD9-dependent and three MBD9-independent H2A . Z sites and performed qPCR experiments using the same H2A . Z ChIP material from WT and mbd9-1 plants previously utilized for ChIP-qPCR analysis of FLC , ASK11 , and At4 genes . We found that all three MBD9-dependent loci showed statistically significant depletion of H2A . Z in mbd9-1 when compared to WT ( S8A and S8B Fig ) , while none of the three MBD9-independent sites had significantly different levels of H2A . Z between mbd9-1 and WT plants ( S8C and S8D Fig ) . Thus , we conclude that MBD9 is required for proper H2A . Z deposition at a subset of the sites that this histone variant normally occupies . In order to understand why MBD9 is required for H2A . Z deposition at certain sites and not others , we first analyzed the genomic distribution of MBD9-dependent H2A . Z sites compared to the MBD9-independent H2A . Z sites . We found that the two sets are distributed similarly across the genome , with more than 80% of each set of coordinates localizing within genic regions ( S9A Fig ) . Both MBD9-dependent and MBD9-independent H2A . Z sites were found primarily at the 5’ end of protein-coding gene ( PCG ) bodies , with the MBD9-dependent sites found slightly more upstream of the MBD9-independent sites ( S9B and S9C Fig ) . PCGs nearest to the MBD9-dependent sites had reduced H2A . Z levels throughout the entire gene body in mbd9-1 and arp6-1 plants , whereas PCGs nearest to the MBD9-independent sites had similar gene body H2A . Z levels in WT and mbd9-1 , but still had reduced H2A . Z levels in arp6-1 ( S9C Fig ) . For the 1322 PCGs found nearest to the MBD9-dependent H2A . Z sites ( S2 Table ) there was no significantly overrepresented Gene Ontology ( GO ) terms identified using either of the two different GO analysis tools ( see methods ) , indicating that MBD9-mediated deposition of H2A . Z is not associated with functionally-related gene sets or a particular cellular pathway . We also examined various histone modification profiles at the two types of sites using publicly available ChIP-seq data from WT plants in order to discern any differences between H2A . Z sites that require MBD9 and those that do not . Interestingly , we found that in WT plants the average level of histone H3 lysine 9 acetylation ( H3K9Ac ) is higher at MBD9-dependent H2A . Z sites than it is at the sites that do not require MBD9 ( Fig 4A ) . However , no significant differences were found in the average enrichment of histone H3 lysine 18 acetylation or H3 lysine 27 trimethylation ( H3K18Ac and H3K27me3 , respectively ) between the two types of loci ( Fig 4B and 4C ) . On the other hand , we did observe less abundant enrichment of dimethylated lysine 9 of histone H3 ( H3K9me2 ) at MBD9-dependent H2A . Z sites ( Fig 4D ) . The same pattern was observed for H3 trimethylation at lysine 4 or lysine 36 ( H3K4me3 and H3K36me3 , respectively ) , with consistently lower levels of each mark at MBD9-dependent H2A . Z sites ( Fig 4E and 4F ) . To better understand the differences in chromatin profiles between the two sets of loci , ten different histone marks ( H2A . Z , H4Ac , H3K27me3 , H2AK121Ub , H3K9me2 , H3K9Ac , H3K18Ac , H3K27Ac , H3K4me3 , and H3K36me3 ) were plotted around a 2 kb window centered on the H2A . Z peak of either the MBD9-dependent or MBD9-independent sites ( S10A Fig ) . The heatmaps were clustered into 4 k-means clusters and three similar sets of chromatin states were observed in both MBD9-dependent and MBD9-independent loci . These include euchromatic regions ( S10B Fig ) , H3K27me3-rich sites likely regulated by PRC2 ( S10C Fig ) , and sites that have H2A . Z but are relatively depleted of the other examined histone modifications . The observed average differences in specific histone marks between the two sets of loci ( Fig 4 ) appear to be driven by increased levels of H3K9Ac at MBD9-dependent euchromatic regions ( S10B Fig ) , decreased levels of H3K9me2 and H3K4me3 in MBD9-dependent PRC2-regulated regions ( S10C Fig ) , as well as reductions in H3K9me2 and H3K36me3 at the MBD9-dependent regions relatively depleted of most histone modifications ( S10D Fig ) . Given the predominant gene-body localization of H2A . Z , the differences in histone modification levels between MBD9-dependent and -independent sites could simply reflect differences in expression levels of the underlying genes . However , using publically available RNA-seq data , we found that genes nearest to the sites in each category span a wide range of expression levels and are not significantly different from one another in terms of steady-state transcript levels ( unpaired t-test , p < 0 . 05 , S11 Fig ) . Thus , MBD9 may specifically recognize or be repelled by specific chromatin features , which could help guide the SWR1 complex to specific DNA sites . To examine which histone marks MBD9 may directly recognize , we performed a histone peptide array assay using the full-length MBD9 protein ( S12 Fig ) . We found that MBD9 most strongly interacted with histone H4 peptides that were acetylated at lysine ( K ) residues 12 , 16 , and 20 , or acetylated at K12 and K16 while being di- or trimethylated at K20 ( Fig 5 ) . The protein also bound to H4 peptides containing both K16Ac and K20Ac , or K12Ac and K16Ac along with K20me1 . This affinity of MBD9 for acetylated H4 is consistent with a previous report [51] . Interestingly , MBD9 interacted with H3 peptides that were mono- or dimethylated at K4 , but did not appear to bind the H3 peptide that was trimethylated at K4 . This lack of affinity for H3K4me3 could explain the relative depletion of this mark at MBD9-dependent H2A . Z sites ( Fig 4E ) . Other H3 peptides that showed significant binding of MBD9 include those dimethylated at arginines 2 and/or 8 ( Fig 5 ) . Collectively , these results suggest that MBD9 recognizes distinct combinations of histone modifications that are likely to influence its chromatin binding characteristics . To determine whether the MBD9 protein , with an estimated molecular mass of 240 kDa , is an integral component of the Arabidopsis SWR1 complex ( Fig 6A ) we performed size-exclusion chromatography ( SEC ) experiments on protein extracts from WT and mbd9-1 plants , followed by western blotting for ARP6 . This allowed us to define the native size of the complex and to determine whether this size changes in the absence of MBD9 , as would be expected if this protein were a stoichiometric component of the SWR1 complex ( Fig 6B ) , as previously demonstrated for the PIE1 subunit [41] . Using an ARP6 monoclonal antibody [40] , we detected ARP6 protein in its native form as a part of a multi-subunit complex with a molecular mass of ~800 kDa ( Fig 6C ) . When the SEC experiments were performed on mbd9-1 extracts , the ARP6 peak did not significantly shift and the estimated molecular mass of the native ARP6 complex in mbd9-1 plants was ~775 kDa ( Fig 6C ) in two biological replicates . These results strongly suggest that MBD9 is not a core component of the ARP6-containing SWR1 complex , but most likely interacts with it in a more transient manner . Alternatively , MBD9 may be a stable component of a minor subset of SWR1 complexes . Collectively , we discovered that MBD9 is required for proper H2A . Z deposition at many sites , but is not stably associated with the SWR1 complex . To investigate genetic interactions between MBD9 and ARP6 , we generated arp6-1;mbd9-1 double mutant plants . We have shown that single arp6-1 and mbd9-1 mutant plants have similar phenotypic defects ( S2 Fig ) and that both ARP6 and MBD9 regulate H2A . Z incorporation into chromatin ( Fig 2A and 2B ) . If these two proteins are subunits of the same complex or function exclusively in the same genetic pathway then double mutant plants should be phenotypically indistinguishable from single mutants , as previously shown for arp6;swc6 plants [18 , 37] . Instead , we observed that the double mutants displayed much more severe defects ( dwarf stature , deformed leaves , and drastically reduced fertility ) than the individual single mutants throughout development ( Fig 7 ) . Importantly , these phenotypes in the double mutant plants reverted back to those of each single mutant by introducing either the genomic ARP6 or MBD9 constructs into the double mutants ( S13 Fig ) , indicating that these defects were truly the result of simultaneous loss of ARP6 and MBD9 functions . To investigate whether the more severe phenotype in arp6-1;mbd9-1 plants is caused by further reduction of H2A . Z incorporation into chromatin , we performed a ChIP-qPCR experiment using H2A . Z antibody on WT , arp6-1 , mbd9-1 , and arp6-1;mbd9-1 plants at FLC regions 2 and 9 . We found that double mutant plants had similar levels of H2A . Z depletion at FLC when compared to arp6-1 ( S14 Fig ) . Collectively , our findings support the idea that MBD9 is not a core subunit of the SWR1 complex and suggest that this protein has additional functions outside of H2A . Z incorporation [57] .
Previous studies provided important evidence suggesting that Arabidopsis contains a SWR1-like complex that mediates incorporation of H2AZ into chromatin [6 , 18 , 20 , 36–44] . Using the SWR1-specific subunit ARP6 as bait , we successfully purified the Arabidopsis SWR1 complex and identified all 11 conserved subunits that are also found in the yeast SWR1 and mammalian SRCAP complexes . Recently , a similar study also used ARP6 as bait to isolate the Arabidopsis SWR1 complex and identified the same proteins as our study , with the addition of AL4 ( Potok et al . 2019 , bioRxiv 10 . 1101/657296 ) . These and our results suggest that the function and structure of the canonical SWR1 complexes that incorporate histone H2A . Z into nucleosomes have been well preserved over evolutionary timescales and may be found in all eukaryotes . TIP60 in animals is a single multifunctional complex that combines the subunits and functions of yeast SWR1 and NuA4 complexes [34] . It appears that this merger evolved as a result of the fusion of the two major scaffolding proteins , the Swr1 ATPase of the yeast SWR1 complex and the Eaf1 protein of the yeast NuA4 complex , into a single p400-like protein . This is based on the fact that p400 contains HSA , ATPase , and SANT domains , which are found separately in the yeast Swr1 and Eaf1 proteins [34 , 44] . Intriguingly , Arabidopsis PIE1 ( yeast Swr1 homolog ) also contains HSA , ATPase , and SANT domains , implying that PIE1 may also be an ortholog of p400 . In fact , Bieuszewski and colleagues originally hypothesized that PIE1 is a scaffolding component of an Arabidopsis TIP60-like complex [34 , 44] . Interestingly , TRA1A and TRA1B , two proteins that co-purified with ARP6 in our TAP experiments , were recently characterized as subunits of the SPT module of an Arabidopsis SAGA complex [52] and are homologs of the yeast NuA4 subunit Tra1 and the mammalian TIP60 subunit TRRAP , further suggesting an intimate functional relationship among Arabidopsis SWR1 and NuA4/HAT complexes ( Table 1 ) . Taken together , it is plausible that plants possess both the canonical SWR1 complex and an independent NuA4-like complex , as in yeast , and may also contain a TIP60-like complex , which is found only in animals . Future purification experiments using Arabidopsis PIE1 as bait are crucial to address the question of whether plants have two distinct PIE1-containing complexes ( SWR1 and a TIP60-like ) , and which subunits are shared between the two complexes . The existence of two different PIE1 complexes could also explain why the phenotype of pie1 mutant plants is distinct from that of h2a . z mutants or mutations in other SWR1 components [18 , 19 , 37–43] . Based on multiple studies in many model organisms , we now have a good understanding of how the SWR1 complex incorporates H2A . Z into nucleosomes [4 , 58 , 59] . However , several aspects of SWR1 biology are still poorly understood , including precisely how the complex is recruited to specific chromatin regions to deposit H2A . Z . In yeast , it has been shown that NuA4-mediated acetylation of specific nucleosomal sites is important for SWR1 targeting to chromatin and H2A . Z incorporation [31–35 , 60] . In addition , it has been proposed that Bdf1 , a bromodomain-containing subunit of the yeast SWR1 complex , recruits the complex to chromatin by recognizing acetylated H4 tails [31 , 61] . Supporting this notion , the loss of Bdf1 results in global reduction of H2A . Z in chromatin [62] . In plants , little is known about the mechanisms that target the SWR1 complex to specific chromatin loci . Recent results from the Jarillo group suggest that binding of the SWC4 subunit to AT-rich DNA elements in promoters of certain genes can recruit the Arabidopsis SWR1 complex to these chromatin regions to deposit H2A . Z [36] . However , only a subset of H2A . Z-enriched genes contain AT-rich elements in their promoters , which strongly suggests that additional mechanisms of SWR1 recruitment exist in plants . The same group has recently demonstrated that the YAF9a subunit , by interacting with acetylated histones , can recruit the SWR1 complex to a subset of SWC4 target genes [47] . What role may MBD9 play in SWR1 recruitment to specific chromatin loci ? In addition to a methyl-CpG-binding ( MBD ) domain , which is in fact thought to bind unmethylated DNA ( [63] , Potok et al . 2019 , bioRxiv 10 . 1101/657296 ) , MBD9 contains an acetyl lysine-binding bromodomain [64] , and two plant homeodomains ( PHD ) which may recognize methylated lysine and arginine residues [65 , 66] . We showed that in mbd9 mutant plants the level of H2A . Z incorporation is significantly reduced at a subset of H2A . Z-enriched regions ( Fig 3 ) and that these MBD9-dependent H2A . Z loci have distinct histone modification profiles relative to MBD9-independent H2A . Z loci ( Figs 4 and S10 ) . Specifically , H2A . Z sites that are dependent on MBD9 had higher levels of H3K9Ac and lower levels of H3K4me3 , H3K36me3 , and H3K9me2 . Using the histone peptide array assay , we demonstrated that MBD9 strongly interacts with acetylated H4 , suggesting that MBD9 may recognize specific H4 acetylations in the context of other modifications . The peptide array data also revealed that MBD9 has high affinity for both symmetrically and asymmetrically di-methylated arginines in the H3 N-terminus ( Fig 5 ) . The significance of this interaction is not yet clear as there are no currently available ChIP-seq data for these modifications in Arabidopsis , and the effects of methylated arginines on chromatin architecture are only poorly understood in plants [67] . The histone modification preferences of MBD9 defined by peptide array may at least partially explain some of the differences in chromatin profiles between H2A . Z sites that require MBD9 and those that do not . For example , the H3K4me3 depletion seen in a subset of MBD9-dependent sites ( S10C and S10D Fig ) may be explained by the relative preference of MBD9 for binding H3K4me1 or H3K4me2 , but not H3K4me3 . However , the binding preferences of MBD9 appear to be complex , and there are likely one or more features that more closely define the MBD9-dependent sites ( e . g . methylation of H3R2 and/or R8 ) . MBD9 was recently found to preferentially localize to nucleosome-depleted regions ( NDRs ) directly upstream of H2A . Z nucleosomes ( Potok et al . 2019 , bioRxiv 10 . 1101/657296 ) . In yeast , SWR1 also localizes to NDRs and it is proposed that while NDR localization serves as a general recruiting mechanism for SWR1 , it is the interaction between SWR1 components and the nucleosomes flanking these NDRs that define the actual sites of H2A . Z deposition [68 , 69] . Collectively , the current data point to a model whereby MBD9 recognizes nucleosomes with specific modification patterns , and perhaps specific DNA sequences , and interacts with SWR1 to effect its localization to specific genes . Although MBD9 was co-purified in all of our ARP6 TAP-tag experiments ( Table 1 ) , MBD9 appears not to be a core component of the SWR1 complex ( Fig 5 ) . Two possible conclusions about MBD9’s interaction with the SWR1 complex can be made based on these results . MBD9 may interact only transiently with components of the SWR1 complex , and is therefore detected in TAP-tag experiments as previously observed for transcription factors and cofactors that recruit Arabidopsis SWI/SNF and PRC2 complexes to specific chromatin sites [70–73] . Alternatively , MBD9 could be more tightly associated with only a subset of all SWR1 complexes in Arabidopsis . In that case , our size exclusion chromatography peak of SWR1 most likely would not show a significant mass reduction in mbd9-1 plants when compared to WT because the loss of MBD9 would affect only a minor fraction of SWR1 complexes . With regard to the synergistic phenotype of arp6;mbd9 mutants , this could result from a further reduction of H2A . Z deposition in the double mutant compared to arp6 , or it may be a manifestation of functions of MBD9 beyond H2A . Z deposition . These two possibilities are , of course , not mutually exclusive . While we did not observe additional H2A . Z reduction at FLC in arp6;mbd9 compared to arp6 , a related study found that in arp6;mbd9 double mutant plants the level of H2A . Z incorporation into chromatin is indeed further reduced genome-wide ( Potok et al . 2019 , bioRxiv 10 . 1101/657296 ) . However , this study also found that MBD9 strongly interacts with the ISWI family of CRCs , suggesting that MBD9 has additional nucleosome remodeling functions outside of H2A . Z deposition . While further experiments are needed to determine the precise nature of MBD9’s interaction with the SWR1 complex , it is clear that MBD9 is functionally associated with the SWR1 complex and is integral to the deposition of H2A . Z at a subset of loci . An important point to consider is whether the specific H2A . Z loss from chromatin in mbd9 mutants reflects a lack of targeting of SWR1 to MBD9-dependent H2A . Z sites , or whether the complex is targeted properly but H2A . Z is not retained in nucleosomes after deposition in mbd9 . It is indeed formally possible that MBD9 plays a role in H2A . Z retention in chromatin . While an absolute resolution to this issue awaits further experimentation , the most parsimonious explanation for the present observations is that MBD9 interacts with the SWR1 enzyme complex to influence where its reaction occurs , rather than acting on the product of the reaction ( an H2A . Z-containing nucleosome ) . The presence of H2A . Z in chromatin has been linked to both gene activation and gene repression , but how H2A . Z affects transcription in this context-dependent manner is not clear [74 , 75] . In addition , how the chromatin remodelers that deposit H2A . Z are recruited to specific chromatin loci is poorly understood . Our isolation of the Arabidopsis SWR1 complex identified unexpected proteins that co-purified with this complex , including MBD9 and three members of the plant-specific Alfin family , each of which contain known modified histone-binding domains . Based on our results and data from other studies , we propose that these SWR1-associated proteins are involved in the recruitment of the complex to chromatin to incorporate H2A . Z at specific loci . In this view , the H2A . Z landscape reflects the collective effects of inherent SWR1 targeting as well as SWR1’s association with a variety of adaptor proteins , such as MBD9 . With the identification of these SWR1-associated proteins , we can now start to address important mechanistic questions about the activity of SWR1 in plants and how MBD9 , and perhaps Alfin1-like family proteins , may modulate SWR1 functions .
Arabidopsis thaliana of the Columbia ( Col-0 ) ecotype was used as the wild type reference , and all mutant seeds are of the Col-0 ecotype . The arp6-1 ( SAIL_599_G03 ) , and mbd9-1 ( SALK_054659 ) , mbd9-2 ( SALK_121881 ) and mbd9-3 ( SALK_039302 ) alleles were described previously [40 , 50] . Seedlings were grown in either soil , half-strength Murashige and Skoog ( MS ) liquid media [76] , or on half-strength MS media agar plates , in growth chambers at 20°C under a 16 hour light/8 hour dark cycle . Plasmids containing the N-TAP-ARP6 , C-TAP-ARP6 , and gMBD9 constructs , driven under endogenous ARP6 and MBD9 promoters , respectively , were introduced into Agrobacterium tumefaciens GV3101 strain by electroporation . Plants were transformed with these constructs via the floral dip method [77] . Primary transgenic plants were selected on half-strength MS media agar plates containing 50 mg/L hygromycin and 100 mg/L timentin , and then transferred to soil . Two to three grams of sterilized WT seeds and T3 seeds homozygous for N-TAP-ARP6 and C-TAP-ARP6 constructs were germinated for 6 days in flasks containing 600 ml of half-strength MS media with constant shaking on rotating platform ( 80–90 rpm ) . After 6 days , the germinated seedlings were filtered to remove the excess liquid , and 50 grams of seedling tissue was frozen in liquid nitrogen and stored at -80°C . To construct the ARP6-TAP-tag , we fused genomic ARP6 sequence to the tandem affinity purification ( TAP ) GSrhino tag , recently developed for efficient affinity purifications of protein complexes in plants [45] . Gateway–compatible plasmids containing either a C-terminal TAP-tag ( pEN-R2-GS_rhino-L3 , [45] ) or an N-terminal TAP-tag ( pEN-L1-NGS_rhino-L2 , [45] ) were used to produce the C-TAP-ARP6 and the N-TAP-ARP6 constructs , respectively . To generate the C-TAP-ARP6 construct , a total of six primers were used in three overlapping PCR reactions to produce a ~4 . 7 kb attB PCR fragment . This PCR product contained ~4 . 1 kb of the genomic ARP6 sequence ( from -2040 bp upstream of the start codon to +2083 bp downstream from the start codon ) , ~600 bp of the TAP-tag sequence fused to the C-terminal end of the ARP6 gene , and attB adapters at the 5’ and 3’ ends of the PCR product for Gateway cloning . This PCR fragment was subcloned into pDONR221 gateway plasmid via BP recombination reaction using BP clonase II enzyme ( Invitrogen ) . The construct was verified by sequencing and further sub-cloned into the destination gateway plasmid pMDC99 [78] using the LR clonase II enzyme in LR recombination reaction ( Invitrogen ) . Similarly , the attB N-TAP-ARP6 construct was first produced using six PCR primers in overlapping PCR reactions containing the same genomic ARP6 DNA fragment as in the C-TAP-ARP6 , with the TAP-tag fused at the N-terminal end of the ARP6 gene . This PCR fragment was then sub-cloned into pDONR221 via BP reaction , verified by sequencing , and finally sub-cloned into the pMDC99 destination plasmid via LR reaction . To generate gMBD9 construct , which was used to transform arp6-1;mbd9-1 double mutant plants , we first PCR-amplified 11 , 311 bp of genomic MBD9 sequence ( from –1936 bp upstream of the start codon to + 9372 downstream from the start codon ) using gMBD9 sequence-specific primers with attB adapters at their 5’ ends . The attB PCR product was then sub-cloned into pDONR221 gateway plasmid via BP recombination reaction ( Invitrogen ) , verified by sequencing , and finally sub-cloned into the destination gateway plasmid pMDC99 [78] using LR recombination reaction ( Invitrogen ) . To clone the N-terminal Myc-MBD9 construct into pT7CFEChis vector ( Thermo Scientific ) , we designed two sets of primers: first set was used to PCR-produce an N-terminal Myc-tag fused in frame with the full length MBD9 cDNA , while second primer pair was used to PCR-amplify the full pT7CFEChis vector . The two PCR products were then used in a Gibson Assembly reaction ( New England Biolabs ) to clone Myc-MBD9 at an NdeI restriction cloning site in the pT7CFEChis plasmid to achieve a maximum expression of a fusion protein in this system . The cloning was verified by sequencing . To produce arp6-1;mbd9-1 double mutant plants , pollen from arp6-1 plants was used to manually pollinate mbd9-1 plants . Since ARP6 and MBD9 genes are both on chromosome 3 , we were only able to identify the F2 plants that were homozygous for one T-DNA allele and heterozygous for the other . We used F3 seeds from arp6-1/arp6-1;mbd9-1/+ plants to identify the double mutant plants . The proteins for western blot detection shown in Fig 1A were extracted from ~100 mg of 6-day old whole transgenic seedlings homozygous for the arp6-1 allele and either the C-TAP-ARP6 or N-TAP-ARP6 constructs by first making a crude nuclei preparation using Nuclei Purification Buffer [79] . The nuclei pellets were then resuspended in 2 volumes of 1x Laemmli’s sample buffer ( 125 mM Tris-HCl pH 6 . 8 , 4% SDS , 30% glycerol , and 1% β-mercaptoethanol ) prior to heating and loading on a gel . For ARP6 detection on fractions from SEC experiments ( see below ) , the eluted proteins were isolated by adding 20 μl of the StrataClean resin ( Agilent ) to 1 ml of each SEC fraction , incubating for 20 minutes at room temperature ( RT ) on a rotating platform , and then spinning down for 2 minutes at 5 , 000g at RT . The pelleted proteins were resuspended in 20 μl of 1x Laemmli’s sample buffer . The total proteins for western blot detection shown in S4 Fig were isolated from ~150 mg of young leaves using acid extraction protocol ( see below ) and the pelleted proteins were resuspended in 50 μl of 1x Laemmli’s sample buffer . The proteins were then separated on 4–20% Novex WedgeWell tris-glycine gel ( Invitrogen ) and transferred to Amersham nitrocellulose blotting membrane ( GE Healthcare ) . After blocking overnight in PBST buffer ( 137 mM NaCl , 2 . 7 mM KCl , 10 mM Na2HPO4 , 2 mM KH2PO4 , and 0 . 05% tween20 ) containing 5% non-fat dry milk , the blots were incubated with primary antibody ( 1:100 dilution for monoclonal mouse ARP6 antibody [40] , 1:2 , 000 dilution for peroxidase anti-peroxidase ( PAP ) soluble complex antibody ( Sigma-Aldrich ) that detects the TAP-tag , 1:1000 dilution for H2A . Z antibody [41] , and 1:500 dilution for H4Ac antibody ( Millipore 06–866 ) ) in blocking solution for 1 hour at RT . The blots were washed 3 times for 5 minutes in PBST . The ARP6 blot was then incubated with the anti-mouse horseradish peroxidase ( HRP ) -conjugated secondary antibody ( 1:2 , 000 dilution , GE Healthcare ) , while H2A . Z and H4Ac blots were incubated with anti-rabbit HRP-conjugated secondary antibody ( 1:3 , 000 dilution , GE healthcare ) . The ARP6 , H2A . Z , and H4Ac blots were washed 3 more times for 5 minutes in PBST , and all blots were then incubated with ECL detection reagents for 2 minutes ( Thermo Scientific ) . The ARP6 blots and PAP blot were exposed to Amersham Hyperfilm ECL ( GE Healthcare ) to detect protein bands , while H2A . Z and H4Ac blots were scanned for chemiluminescence signal using ChemiDoc MP imaging system instrument ( BioLab ) . Approximately 150 mg of young leaves from WT , mbd9-1 , and arp6-1 plants were ground to a fine powder , homogenized in 3 ml of histone extraction buffer ( 0 . 25 M sucrose , 1 mM CaCl2 , 15 mM NaCl , 60 mM KCl , 5 mM MgCl2 , 15 mM PIPES pH 7 . 0 , 0 . 5% Triton X-100 with protease inhibitors cocktail ( Roche ) and 10 mM sodium butyrate ) , filtered through the 70 micron strainer , and incubated for 15 minutes at 4°C on a nutator . After centrifugation at 4 , 500g for 20 minutes at 4°C , the pellets were resuspended in 500 μl of 0 . 1 M H2SO4 and incubated overnight at 4°C on a nutator . After centrifugation for 10 minutes at 17 , 000g , total proteins were precipitated from the supernatant with concentrated trichloroacetic acid to a final concentration of 30% . The protein pellets were washed twice with an acetone solution containing 0 . 1% HCl and then once with acetone . The protein pellet was then air-dried and resuspended in 50 μl of 1x Laemmli’s sample buffer . Myc-MBD9 fusion protein , cloned into the pT7CFEChis plasmid , was expressed using the 1-step human coupled in-vitro translation ( IVT ) kit ( Thermo Scientific ) following manufacturer’s recommendations . In parallel , a GFP protein ( pT7CFEcHis plasmid carrying GFP; included in the kit ) was expressed and served as a positive control for IVT reaction . The expression of GFP was confirmed by visualizing GFP on a fluorescence microscope ( Olympus ) using 2 μl of an IVT reaction , while the expression of Myc-MBD9 was confirmed by western blotting using 4 μl of an IVT reaction and an anti-myc antibody to detect the 245 kDa protein ( S12 Fig ) . Two MODified peptide array slides ( Active Motif ) were first briefly washed in PBST buffer and then blocked in 5% milk-PBST buffer for 1 hour at RT . After blocking , the slides were washed three times for 5 minutes in PBST . One slide was then incubated with 16 μl of the IVT Myc-MBD9 protein , while the other slide was incubated with 16 μl of the IVT GFP protein , both diluted in 8 ml of the binding buffer ( 50 mM HEPES pH 7 . 5 , 100 mM NaCl , 1 μM DTT , and 50% glycerol ) overnight at 4°C on a nutator . Next day , the slides were washed three times for 5 minutes in PBST buffer and then incubated with a primary anti-myc antibody ( 1:2 , 000 dilution , Abcam ) for 1 hour at RT . The slides were washed again three times for 5 minutes in PBST and then incubated with anti-mouse HRP-conjugated secondary antibody ( 1:5 , 000 dilution , GE Healthcare ) for 30 minutes at RT . The slides were incubated with ECL detection reagents for 2 minutes ( Thermo Scientific ) and scanned with the ChemiDoc MP imaging system ( BioRad ) to detect the spot signals . Two biological replicates of histone peptide array assays were performed for each protein . The chemiluminescence signal intensities of each spot on histone peptide array slides were quantified using The Array Analyze Software available from the manufacturer ( Active Motif ) and were assigned the values ranging from 0 ( minimum , no signal ) to 1 ( maximum , strongest signal ) . The strongest signal for each array was observed at spot “P21” where the myc-tag peptide is spotted given that an anti-myc antibody was used to detect Myc-MBD9 binding . Since MBD9 is not expected to interact with the H2B peptide ( spot “P4” on the array ) due to differences in the N-terminal amino acid sequences between the human H2B and Arabidopsis H2B [80] , the average signal intensity of the “P4” spot was subtracted from the average intensity values for each spot for each array . The average background intensities of the “P4” spots from two arrays ( S12A and S12B Fig ) were 0 . 1145 and 0 . 2605 , respectively . After the background subtraction , only the normalized average spot intensities with values higher than 0 . 3 were presented in Fig 5 , along with the spot intensities for H3K4me3 and H3K27me3 . The ARP6-TAP-containing protein complex was purified as described in [45] , with the following modifications: 1 ) instead of using a kitchen blender in a stainless steel wine cooler , 50 grams of frozen seedlings were ground with a mortar and pestle in liquid nitrogen , and 2 ) all washing steps of the IgG-Sepharose and streptavidin-Sepharose Poly-Prep columns were performed using a peristaltic pump at a flow rate of 1 ml/min at 4°C . On-bead digestion of the TAP-purified proteins was performed as previously reported [81 , 82] . Residual wash buffer was removed and 200 μl of 50 mM NH4HCO3 was added to each sample . Samples were reduced with 1 mM dithiothreitol for 30 minutes and alkylated with 5mM iodoacetamide in the dark for an additional 30 minutes . Both steps were performed at room temperature . Digestion was started with the addition of 1 μg of lysyl endopeptidase ( Wako ) for 2 hours and further digested overnight with 1:50 ( w/w ) trypsin ( Promega ) at room temperature . Resulting peptides were acidified with 25 μl of 10% formic acid ( FA ) and 1% trifluoroacetic acid ( TFA ) , desalted with a Sep-Pak C18 column ( Waters ) , and dried under vacuum . Liquid chromatography coupled to tandem mass spectrometry ( LC-MS/MS ) on an Orbitrap Fusion mass spectrometer ( ThermoFisher Scientific , San Jose , CA ) was performed at the Emory Integrated Proteomics Core ( EIPC ) [81 , 82] . The dried samples were resuspended in 10 μl of loading buffer ( 0 . 1% formic acid , 0 . 03% trifluoroacetic acid , 1% acetonitrile ) . Peptide mixtures ( 2 μl ) were loaded onto a 25 cm x 75 μm internal diameter fused silica column ( New Objective , Woburn , MA ) self-packed with 1 . 9 μm C18 resin ( Dr . Maisch , Germany ) . Separation was carried out over a 140-minute gradient by a Dionex Ultimate 3000 RSLCnano at a flow rate of 300 nl/min . The gradient ranged from 3% to 99% buffer B ( buffer A: 0 . 1% formic acid in water , buffer B: 0 . 1% formic in ACN ) . The spectrometer was operated in top speed mode with 3 second cycles . Full MS scans were collected in profile mode at 120 , 000 resolution at m/z 200 with an automatic gain control ( AGC ) of 200 , 000 and a maximum ion injection time of 50 ms . The full mass range was set from 400–1600 m/z . Tandem MS/MS scans were collected in the ion trap after higher-energy collisional dissociation ( HCD ) . The precursor ions were isolated with a 0 . 7 m/z window and fragmented with 32% collision energy . The product ions were collected with the AGC set for 10 , 000 and the maximum injection time set to 35 ms . Previously sequenced precursor ions within +/- 10 ppm were excluded from sequencing for 20s using the dynamic exclusion parameters and only precursors with charge states between 2+ and 6+ were allowed . All raw data files were processed using the Proteome Discoverer 2 . 1 data analysis suite ( Thermo Scientific , San Jose , CA ) . The database was downloaded from Uniprot and consists of 33 , 388 Arabidopsis thaliana target sequences . An additional sequence was added for the TAP-tagged bait protein . Peptide matches were restricted to fully tryptic cleavage and precursor mass tolerances of +/- 20 ppm and product mass tolerances of +/- 0 . 6 Daltons . Dynamic modifications were set for methionine oxidation ( +15 . 99492 Da ) and protein N-terminal acetylation ( +42 . 03670 ) . A maximum of 3 dynamic modifications were allowed per peptide and a static modification of +57 . 021465 Da was set for carbamidomethyl cysteine . The Percolator node within Proteome Discoverer was used to filter the peptide spectral match ( PSM ) false discovery rate to 1% . For each sample , 1 . 5 grams of 6-days-old seedlings ( with roots removed ) were used for ChIP-seq and ChIP-qPCR experiments . The ChIP-seq experiments were performed in biological triplicates on WT , arp6-1 , and mbd9-1 seedling tissues as described previously [83] , with the following modifications: 1 ) after the centrifugation of the nuclei in extraction buffer 3 , the pellets were resuspended in 210 μl of nuclei lysis buffer , 2 ) after aliquoting out 10 μl for input , the fragmented chromatin from each sample was split into two equal volumes ( 100 μl each ) and diluted with 1ml of ChIP-dilution buffer , and 3 ) one diluted half was incubated with 1 . 5 μg of the affinity-purified polyclonal H2A . Z antibody [41] while the other half was incubated with 5 μl of H4Ac polyclonal antibody serum ( Millipore cat . # 06–866 ) . The ChIP-qPCR experiments were performed in duplicates on WT , arp6-1 , mbd9-1 , mbd9-2 , mbd9-3 , and arp6-1;mbd9-1 seedling tissues as described previously [83] , with the following modifications: 1 ) after the centrifugation of the nuclei in extraction buffer 3 , the pellets were resuspended in 105 μl of nuclei lysis buffer , and 2 ) after sonication using a Diagenode Bioruptor , 5 μl of the fragmented chromatin was used for input , while 100 μl was diluted with 1 ml of the ChIP dilution buffer . The whole solution was then used for incubation with 1 . 5 μg of the affinity-purified polyclonal H2A . Z antibody . The ChIP and input DNA samples from the ChIP-qPCR experiment were analyzed by real-time PCR using the ACT2 ( At3g18780 ) 3’ untranslated region sequence as the endogenous control , and with primers that span the genomic regions of FLC ( At5g10140 ) , ASK11 ( At4g34210 ) , and At4 ( At5g03545 ) genes ( S4 Fig ) , and of AT2G34202 , AT1G62760 , AT5G28410 , AT4G15960 , AT1G76740 , and AT3G21240 genes ( S5 Fig ) . The sequences of the primers used in S6 Fig were previously described [41 , 55] , while the sequences of the primers used in S8 Fig are listed in S3 Table . Libraries were prepared starting with 500 pg of ChIP or input DNA samples using the Swift Accel-NGS 2S Plus DNA library kit ( Swift Biosciences ) according to the manufacturer’s instructions . All libraries were pooled and sequenced using single-end 50 nt reads on an Illumina NextSeq 500 instrument . Reads were mapped to the Arabidopsis thaliana genome ( TAIR10 ) using the Bowtie2 package [84] . Quality filtering and sorting of the mapped reads , as well as removal of the reads that mapped to the organellar genomes , was done as previously described [85] using Samtools 1 . 3 . 1 [86] . The filtered and sorted BAM files were converted to bigwig format as previously described [85] using deepTools 2 . 0 software [87] . Correlation plots for the different H2A . Z or H4Ac samples were computed and plotted using the “multiBamSummary” and “plotCorrelation” functions in the deepTools package . For visualization , for a given antibody , BAM files of each genotype were scaled to the same number of reads . This was done using the “samtools view -c” and “samtools view -s” commands to count the number of reads in a BAM file and to scale down the global read amounts in a BAM file , respectively . Three scaled , replicate BAM files of each genotype for H2A . Z were combined and converted to a single bigwig file for each genotype . The same was done for each genotype of H4Ac . Average plots and Heatmaps displaying ChIP-seq data were generated using the SeqPlots app [88] . Peak calling on ChIP-seq data was done by employing the “Findpeaks” function of the HOMER package [89] using the input ChIP-seq files as reference and the “-region” option to identify sites of read density enrichment . Called peaks were processed using Bedtools [90] to identify peaks called in at least one other replicate for a given genotype . This was done by keeping any peaks that overlapped by at least 50% between biological replicates . The retained peaks were concatenated and then merged together if they overlapped by at least 200 base pairs . The amount of H2A . Z reads in WT , mbd9-1 , and arp6-1 plants present in H2A . Z-enriched peaks reproducibly identified in WT plants was quantified using HTSeq’s htseq-count script [91] . Three replicates of counted reads for all three genotypes were then processed using DESeq2 [92] . MBD9-dependent and MBD9-independent peaks were determined from the DESeq2 results comparing wild type and mbd9-1 counted peaks . Peaks that had a log2 fold change of 0 . 6 or more and an adjusted p-value less than or equal to 0 . 05 were designated as MBD9-dependent H2A . Z sites . Peaks with an absolute log2 fold change less than 0 . 25 were designated as MBD9-independent H2A . Z sites . The PAVIS web tool [93] was used to determine the genomic distribution of H2A . Z ChIP-seq peaks . PAVIS annotates each peak to a genomic feature using the center of the peak . The “upstream” regions were defined as 2 , 000 bp upstream of the annotated transcription start site , and “downstream” regions were defined as 1 , 000 bp downstream of the transcription end site . Genes nearest to the MBD9-dependent and MBD9-independent H2A . Z sites were identified using the “TSS” function of the PeakAnnotator 1 . 4 program [94] as previously described [85] . Gene ontology ( GO ) analysis was carried out on gene lists from S2 Table using two different GO web tools: 1 ) the AgriGO GO Analysis Toolkit , with default parameters [95 , 96] , and 2 ) Gene Ontology enrichment analysis [97] . GO terms that had a false discovery rate ( FDR ) of 0 . 05 or less were considered significant . Publicly available RNA-seq FPKM values for genes nearest to the MBD9-dependent and MBD9-independent H2A . Z sites were plotted using ggplot2 [98] . Unpaired t-tests were used to determine whether FPKM values were significantly different between the two sets of genes , with P values less than or equal to 0 . 05 considered as statistically significant . We performed student t-tests to calculate the significance of fold changes in H2A . Z levels at three MBD9-dependent and three MBD9-independent sites between WT and mbd9-1 plants ( S8 Fig ) . Based on our ChIP-seq results , we only expect a reduction in H2A . Z levels at MBD9-dependent sites in mbd9-1 plants compared to WT . Therefore , we performed a one-tail t-test to calculate the significance of H2A . Z depletion at MBD9-dependent sites ( S8A Fig ) . At MBD9-independent sites , however , we expect that H2A . Z levels may vary in either direction in mbd9-1 plants versus WT . Thus , for MBD9-independent sites we performed a two-tail t-test analysis of qPCR results . Additionally , when the two-tailed t-test is performed on MBD9-dependent ChIP-qPCR results two out of three probed sites ( AT2G34202 and AT1G62760 ) still show statistically significant depletion of H2A . Z levels . Real-time PCR was performed on the Applied Biosystems StepOnePlus real-time PCR system using SYBR Green as a detection reagent . The 2-ΔΔCt method [99] of relative quantification was used to calculate the fold enrichment . The results presented for ChIP-qPCR experiments are averaged relative quantities from two biological replicates ± SD . Total RNA was isolated from 6-day old seedlings ( with roots removed ) using the RNeasy plant mini kit ( Qiagen ) . 2 μg of total RNA was converted into cDNA with Super-Script III first strand synthesis kit ( Invitrogen ) . The cDNAs were used as templates for real-time PCR and ran on StepOnePlus real-time PCR system ( Applied Biosystems ) using SYBR Green as a detection reagent . The 2-ΔΔCt method [99] of relative quantification was used to calculate the fold enrichment . The PP2A gene ( AT1G13320 ) was used as the endogenous control [100] . The primer sequences used to measure relative expression levels of PIE1 , ARP6 , MBD9 , SWC4 , SWC6 , YAF9a , HTA8 , HTA9 , and HTA11 in WT , arp6-1 , and mbd9-1 plants are listed in S3 Table . SEC was performed on the HiPrep 16/60 Sephacryl S-400 HR column ( GE Healthcare ) equilibrated with SEC buffer ( the same extraction buffer as described in [45] , but without NP-40 detergent ) . A mixture of protein standards ranging from 669 to 44 kDa ( GE Healthcare ) , resuspended in the SEC buffer , were run on the column to produce a calibration curve of molecular weights versus elution volumes . The slope equation of the calibration curve was then used to calculate the molecular weight of the peak ARP6 SEC fractions . Total protein extracts were isolated from 1 gram of the WT and mbd9-1 seedling tissues ( without roots ) using the same extraction buffer that was used for the ARP6-TAP-tag protein complex purification [45] . For each run , between 1 . 8 and 2 . 0 ml of the protein extract was loaded onto the column and 1 ml fractions were collected . For each sample , two biological replicates of the SEC experiments were performed and gave nearly identical results . Raw data from ChIP-seq experiments performed on young WT seedlings using antibodies against H3K4me3 ( GSM2544796 , [101] ) , H3K36me3 ( GSM2544797 , [101] ) , H3K9me2 ( GSM2366607 , [102] ) , H3K9Ac ( GSM2388452 , [103] ) , H3K18Ac ( GSM2096925 , [104] ) , H3K27Ac ( GSM2096920 , [104] ) , H3K27me3 ( GSM2498437 , [105] ) , and H2AK121ub ( GSM2367138 , [105] ) , were processed and analyzed with the same procedures as for our ChIP-seq data ( see above ) and used to generate the average plots presented in Fig 4 . The FPKM values from two different RNA-seq experiments ( GSM2752981 and GSM2367133 , respectively , [105 , 106] ) were used to compare expression levels in WT plants of genes nearest to the MBD9-dependent and MBD9-independent H2A . Z sites . All ChIP-seq data generated in this study have been deposited to the NCBI GEO database under accession number GSE117391 .
|
The histone H2A variant , H2A . Z , is found in all known eukaryotes and plays important roles in transcriptional regulation . H2A . Z is selectively incorporated into nucleosomes within many genes by the activity of a conserved ATP-dependent chromatin remodeling complex in yeast , insects , and mammals . Whether this complex exists in the same form in plants , and how the complex is targeted to specific genomic locations have remained open questions . In this study we demonstrate that plants do indeed utilize a complex analogous to those of fungi and animals to deposit H2A . Z , and we also identify several new proteins that interact with this complex . We found that one such interactor , Methyl-CpG-BINDING DOMAIN 9 ( MBD9 ) , is required for H2A . Z incorporation at thousands of genomic sites that share a distinct histone modification profile . The histone binding properties of MBD9 suggest that it may guide H2A . Z deposition to specific sites by interacting with modified nucleosomes and with the H2A . Z deposition complex . We hypothesize that this represents a general paradigm for the targeting of H2A . Z to specific sites .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"biotechnology",
"engineering",
"and",
"technology",
"brassica",
"dna-binding",
"proteins",
"fungi",
"plant",
"science",
"model",
"organisms",
"genetically",
"modified",
"plants",
"experimental",
"organism",
"systems",
"epigenetics",
"molecular",
"biology",
"techniques",
"plants",
"chromatin",
"genetic",
"engineering",
"arabidopsis",
"thaliana",
"research",
"and",
"analysis",
"methods",
"bioengineering",
"genetically",
"modified",
"organisms",
"artificial",
"gene",
"amplification",
"and",
"extension",
"animal",
"studies",
"chromosome",
"biology",
"proteins",
"gene",
"expression",
"histones",
"molecular",
"biology",
"nucleosomes",
"yeast",
"biochemistry",
"eukaryota",
"plant",
"and",
"algal",
"models",
"cell",
"biology",
"polymerase",
"chain",
"reaction",
"genetics",
"biology",
"and",
"life",
"sciences",
"plant",
"biotechnology",
"organisms"
] |
2019
|
Methyl-CpG-binding domain 9 (MBD9) is required for H2A.Z incorporation into chromatin at a subset of H2A.Z-enriched regions in the Arabidopsis genome
|
Bacterial microcompartments ( MCPs ) are protein-bound organelles that carry out diverse metabolic pathways in a wide range of bacteria . These supramolecular assemblies consist of a thin outer protein shell , reminiscent of a viral capsid , which encapsulates sequentially acting enzymes . The most complex MCP elucidated so far is the propanediol utilizing ( Pdu ) microcompartment . It contains the reactions for degrading 1 , 2-propanediol . While several experimental studies on the Pdu system have provided hints about its organization , a clear picture of how all the individual components interact has not emerged yet . Here we use co-evolution-based methods , involving pairwise comparisons of protein phylogenetic trees , to predict the protein-protein interaction ( PPI ) network governing the assembly of the Pdu MCP . We propose a model of the Pdu interactome , from which selected PPIs are further inspected via computational docking simulations . We find that shell protein PduA is able to serve as a “universal hub” for targeting an array of enzymes presenting special N-terminal extensions , namely PduC , D , E , L and P . The varied N-terminal peptides are predicted to bind in the same cleft on the presumptive luminal face of the PduA hexamer . We also propose that PduV , a protein of unknown function with remote homology to the Ras-like GTPase superfamily , is likely to localize outside the MCP , interacting with the protruding β-barrel of the hexameric PduU shell protein . Preliminary experiments involving a bacterial two-hybrid assay are presented that corroborate the existence of a PduU-PduV interaction . This first systematic computational study aimed at characterizing the interactome of a bacterial microcompartment provides fresh insight into the organization of the Pdu MCP .
Cellular organization has long been considered to be much simpler in bacteria than in eukaryotic cells . However , accumulating evidence indicates a higher-order organization in terms of cellular compartmentalization [1–3] and cell structure [4 , 5] . In particular , electron microscopy and higher resolution structural studies have demonstrated that some bacteria can form polyhedral capsid-like bodies that are 80 to 150 nm in diameter [6 , 7]; reviewed in [8–11] . These polyhedral inclusions , known as bacterial microcompartments , are widely distributed across nearly 20% of known bacterial strains [9 , 12 , 13] . We refer here to bacterial microcompartments as MCPs; they are sometimes referred to as BMC’s , but we reserve the latter name to refer to the family of shell proteins that comprise MCP shells . As opposed to membrane bound organelles characteristic of eukaryotic cells , MCPs are exclusively proteinaceous assemblies; they consist of a thin outer protein shell enclosing a metabolically active core of enzymes , earning them the status of bacterial organelles . MCPs fulfill diverse roles: enhancement of metabolic flux in their hosted enzymatic pathway [14] , confinement of toxic or volatile intermediates [15–17] and shielding of interior enzymes from reactions with reactive or competing molecules [18] . The founding member of the MCP family , the carboxysome , was first isolated 40 years ago [19] . Carboxysomes are present in some chemotrophic bacteria and probably all cyanobacteria [18 , 20 , 21] . The carboxysome serves as an organelle for carbon fixation through the encapsulation of two enzymes: carbonic anhydrase and ribulose-1 , 5-bisphosphate carboxylase/oxygenase ( RubisCO ) . Several other kinds of MCPs are found dispersed across the bacterial kingdom , where they carry out metabolic pathways different from carbon fixation . These include the Pdu and the Eut microcompartment from Salmonella [22–24] and E . coli [25 , 26] , which carry out the degradation of 1 , 2-propanediol and ethanolamine , respectively . These pathways rely on a similar mechanism: an initial substrate is first converted by a B12-dependent enzyme to give an aldehyde intermediate , which is sequestered long enough to be converted to less toxic metabolites , e . g . an alcohol and/or carboxylic acid . However , these three relatively well-characterized MCPs ( carboxysome , Pdu and Eut ) constitute only a subset of the entire MCP family . Recent computational and experimental studies delineate at least seven kinds of MCPs , all with different metabolic purposes [13 , 27–30] . The accepted three-dimensional model of the Pdu MCP and its encapsulated metabolic pathway is summarized in Fig . 1B . Though MCPs differ substantially according to their metabolic nature , they share a number of genomic and structural characteristics . In particular , most MCP proteins are encoded within operons , which consist of multiple paralagous genes coding for the shell proteins alongside the genes for the associated enzymes . Consistent with this shared genomic signature , diverse MCPs share a similar organization and structure . Typically , each shell protein sequence is comprised by a bacterial microcompartment ( BMC ) domain , or sometimes two such domains duplicated in tandem . The first high resolution structures of BMC proteins shed light on the structural organization of the MCP shell [9–11 , 28 , 31–35] . A few thousand copies of these BMC proteins self-assemble into cyclic hexameric units packed side-by-side in a layer forming the essentially flat facets of the roughly icosahedral structure ( Fig . 1A ) . The top and bottom sides of a BMC hexamer typically show distinctly different features: one face bears a central depression giving rise to a concave shape , whereas the other side is typically flatter and more polar in chemical character . Which of the two sides ( convex or flat ) faces inward to the MCP lumen is a question of key significance for MCP function [36–38] . Most often , the center of the hexamer is perforated by a narrow ( 4–7 Å ) hydrophilic pore that is thought to act as a canal for molecular transport [32 , 38–40] . In addition to the main BMC shell proteins , other minor proteins have been found to be essential to the formation or closure of the shell . These proteins , which our group recently coined the bacterial micrompartment vertex ( BMVs ) proteins , assemble into pentamers suspected to close the vertices of the MCP [37 , 41 , 42] ( Fig . 1A ) . Furthermore , a number of intriguing variations such as domain fusion , tandem duplication , circular permutation , or FeS cluster binding sites , have been revealed among the crystal structures of the paralogous BMC shell proteins [43–46] . Speculations on the roles of such variations support the idea that each type of BMC paralog has a defined role beyond simply assembling to form a physical barrier . Interactions between the shell proteins and the encapsulated enzymes are vital for MCP function . Recent studies on the assembly of the α-type carboxysome suggest assembly of this type of MCP is initiated from the interior; the formation of enzymatic seeds precedes acquisition of the shell [47 , 48] . However , the processes governing the interactions between the encapsulated enzymes and the shell proteins are complex and apparently divergent between different types of MCPs . Specific interactions have been demonstrated in a few cases using pull-down assays and other experiments [36 , 49] . Fan et al . [50] first showed that short sequence extensions present at the N-terminus of numerous enzymes exist to bind enzymes to the MCP shell . A subsequent study showed that the C-terminal region of an α-carboxysomal protein ( CcmN ) interacted with the shell in that system [49] . Though enzyme targeting mechanisms are presumed to be widespread across the MCP systems , only a few enzyme-shell protein interactions have been specifically identified . Characterizing these interactions would open new perspectives on MCP biology and applications in synthetic biology [51 , 52] . Some progress has already been made along these lines . Fluorescent proteins and other proteins have been successfully directed to MCPs by appending terminal targeting peptides [29 , 50 , 53–55] . Despite knowing the identities of a few interactions between enzymes and shell proteins , atomic level detail is lacking . Attempts to isolate and determine the structures of cognate complexes have been unsuccessful . This has prompted us to undertake a computational study to develop interaction models for an MCP system . The ever-increasing genomic and structural data available for MCPs provides an unprecedented opportunity to apply computational methods to characterize the molecular networks ruling these extraordinary supramolecular machines . Over the last two decades , a handful of methods exploring genomic data have been developed for predicting functional linkages between different proteins in a cell . Popular methods such as protein phylogenetic profiles [56 , 57] , gene fusion [58 , 59] , gene neighborhood [60 , 61] or a combination of these [62–64] , have been used extensively to make functional inferences about proteins . Indeed , one of our recent studies featured an adaptation of protein phylogenetic profile methods for investigating co-occurrence patterns in MCP operons , and led to an articulated classification of existing MCP pathways [13] . Here , we aim to characterize the molecular network of physical protein-protein interactions ( PPIs ) in a single MCP type , the Pdu system . In this case , strategies relying on genomic context have limited application due to the high similarity of the genomic patterns found for different proteins across the Pdu operons; essentially all of the MCP shell proteins and enzymes typically found in the Pdu operon are functionally linked according to genomic context , but only a subset engage in direct physical PPIs . Other computational strategies are therefore required to develop models for direct physical PPIs . Detailed sequence variations within protein families can be analyzed via phylogenetic tree-based approaches , and indeed methods based on mining of phylogenetic features have proven useful for predicting PPIs in multiple cases , as recently reviewed [65 , 66] . A non-exhaustive list of such methods includes the so-called mirror tree [67] , or its variant the tol-mirror [68] , which compares trees—one for each protein of interest—by computing the pairwise correlation of their underlying evolutionary distance matrices . Others explore the topological similarity of the trees , coined congruence by Vienne et al . [69] . All follow the co-evolution hypothesis , where interacting protein families are expected to exhibit similar phylogenetic trees with similar patterns of amino acid sequence divergence . In this work , we seek to identify new PPIs in the Pdu MCP with a coevolution-based machine learning algorithm . Specifically , we approach the PPI prediction problem within a binary classification framework: from the pairwise comparison of phylogenetic trees , coevolution features can be computed and subsequently mined by a decision tree classifier , a concept earlier described in Craig and Lio [70] . A group of PPIs that have been experimentally characterized recently in the Pdu system constitute a set of known positives for use as a “gold standard” for training the classifier . In the first part of this work , we design and train a Random Forests classifier to identify pairwise interactions of Pdu gene products , and then propose a model of the Pdu interactome . Following this genomic-based model , we further analyze selected predictions of PPIs and their probable binding modes via three-dimensional protein-protein docking calculation . We then provide new experimental data to support one of the predicted interactions .
We culled protein sequences from Pdu operons of 34 fully sequenced bacterial genomes , and collapsed them into 22 orthologous protein groups according to the canonical Pdu nomenclature [23] . For each of the 22 distinct protein families so identified , we inferred a phylogenetic tree from a multiple sequence alignment of its constitutive sequences . We refer to this as the ‘Pdu tree’ for that protein . Subsequently , for each pair of proteins seven co-evolution descriptors were computed from a comparison of their respective Pdu trees , following the general procedure depicted in Fig . 2 . Pairwise combinations of the 22 orthologous protein groups resulted in 231 unique pairs that needed to be classified . For this purpose , we used a Random Forests classifier [71] exploring the seven descriptors , which after a training and cross-validation phase exhibited an area under the receiver-operator ( ROC ) curve of 0 . 75 ( S1 Fig . , suppl . Data ) , thereby demonstrating a reasonably good discriminative power . We also assessed whether similar classification performance could be obtained with fewer descriptors than the seven initially employed . We evaluated the discriminatory power of the descriptors individually by ranking their accuracies in the context of an unsupervised analysis ( S2 Fig . ) . We found that the RF performs best when all seven of the descriptors are included in the classification analysis . Much of the signal can be recovered with just a few descriptors , but addition of subsequent descriptors does result in slight improvements in performance . When applied to the whole Pdu dataset , the classifier predicted a list of 109 positive PPIs along with their mean probabilities . To be conservative and increase the specificity of the classifier ( even if at the expense of the sensitivity ) , we removed the putative PPIs with a probability less than 0 . 7 , which reduced the final number of predicted PPIs to 51 ( Suppl . data ) . From these results we modeled the Pdu interactome as a molecular network of 51 interactions and 22 nodes . The resulting network model is presented in Fig . 3 . An analysis of this model showed that 15 of the 16 experimentally characterized PPIs could still be retrieved under a high specificity criterion , and that they yielded the highest probabilities , confirming the robustness of the method . Furthermore , the missing positive interaction , PduK-PduT , was initially predicted as positive by the classifier , but did not pass our 0 . 7 threshold . One striking feature of this model is the absence of a PPI connecting the PduX node to the network ( Fig . 3 ) . PduX is an enzyme involved in de novo synthesis of coenzyme B12 , an essential cofactor for enzymes of the Pdu pathway[72] . However , there is no evidence that PduX is directly associated with the MCP by any physical interaction [73] . Its tendency to occur within the Pdu operon ( typically at the end ) likely reflects an advantage of being under the influence of the Pdu promoter , rather than physical interaction with other MCP components . Two likely spurious findings involving interactions with PduF also appeared in our model , namely PduF-PduC and PduF-PduD . PduF is a propanediol/glycerol diffusion facilitator protein and is believed to be an integral membrane protein [74] , making its physical presence in the MCP unlikely . Finally , after exclusion of the “gold standard” interactions and the suspected spurious predictions , the final dataset consisted of 36 predicted PPIs , with an average node connectivity of 4 . 8 partners , which can be loosely compared to results obtained with other interactome studies across whole cellular systems in yeast [75] or in cell junction complexes [76] . One intriguing observation is the hyperconnectivity of certain specific nodes , such as PduA ( 11 PPIs ) , PduC ( 9PPIs ) and PduG ( 9 PPIs ) . The central position of the propanediol dehydratase large subunit PduC in the Pdu pathway makes it an essential piece of the interactome ( Fig . 1B ) . Likewise , PduG is the large subunit of the diol dehydratase-reactivating factor , which works in tight coordination with the propanediol dehydratase ( PduCDE ) . In a complex with PduH , PduG is believed to reactivate the dehydratase by exchanging its B12 cofactor , which becomes inactive during repeated catalytic cycling [73 , 77] . In our model , PduG was indeed predicted to interact with PduC and PduE but not with PduD . Although no structural information about this complex in Salmonella is available , crystal structures of highly similar homologs from Klebsiella oxytoca have been solved [22 , 78 , 79] . Studies with these same homologs demonstrated that the binding mechanism involved a subunit exchange between the dehydratase and the reactivase , where one PduH subunit is released from the reactivase and replaced by one PduD subunit [80] . Particularly notable in our interactome model is the number of PPIs in which PduA [73 , 81] , one of the most abundant shell proteins in the Pdu MCP , is predicted to be involved . Presently , PduP is the only enzyme in the Pdu MCP whose binding to individual shell proteins has been characterized . It was revealed that PduP interacts via its N-terminal region with PduA and PduJ , another major shell protein that shares high sequence identity ( 83% ) with PduA [36] . Other Pdu enzymes besides PduP are suspected to carry such N-terminal extensions [50] , but their shell protein partners have not been identified yet [54 , 82] . Sequence analysis as well as spectroscopic experiments on the PduP N-terminal segment show that it has a strong propensity to fold into an alpha-helical structure [36 , 55] . Here , we hypothesize that these structural features and associated binding mechanism are not specific to the PduP case , but that PduA ( or PduJ ) likely serves as a central binding hub for different enzymes carrying N-terminal extensions . To pursue this particular set of interactions further , we generated atomic models of predicted PduA-enzyme-tail complexes by molecular docking and analyzed their predicted modes of binding . Additionally , we analyzed the PduU-PduV case , the only PPI in which PduV was predicted . PduU was the first BMC shell protein from a non-carboxysome MCP whose three-dimensional structure was determined [44] . Its topology involves a circularly permuted BMC domain , and the existence of a six-stranded β-barrel capping the central pore of the hexamer makes it unique in the BMC protein family . Previous speculations about the peculiar beta barrel include a possible role in gating an unusually wide pore , but further data are lacking . Additionally , PduV is a Ras-like GTPase that has been implicated in MCP dynamics within the cell by Parsons et al . [82] . In this case , PduV is believed to reside outside the shell , as opposed to the other Pdu enzymes that appear to be sequestered in the MCP interior . To clarify how these two might interact , as predicted by our interactome model , we modeled the PduU-PduV complex with docking simulations and compared the result to control calculations involving non-interacting protein pairs . Of the 11 predicted interactions involving PduA , six include Pdu enzymes , namely PduC , PduD , PduE , PduG , PduP , and PduL ( Fig . 3 ) . As noted above , one of these interactions ( PduA-PduP ) has been demonstrated experimentally . Here we investigated whether enzymatic partners in addition to PduP are also able to bind PduA via terminal peptides , by attempting to model their presumptive binding modes computationally ( see Methods ) . As a first step , we searched for possible terminal peptidic extensions in the sequences of these six Pdu enzymes . Prediction of these extensions was done according to the method developed in Fan et al . [50] . The central idea is that enzymes that are targeted to the MCP exhibit extensions at their termini that are absent from homologous versions of the same enzyme that are not part of an MCP system . It is notable that among the six enzymes that are predicted by our model to bind to shell protein PduA ( or its close homolog PduJ ) , our computational analysis indicates that five carry recognizable sequence extensions ( PduC , D , E , P , L ) , ( as reported in [50] and [54] ) . In contrast , none of the 15 enzymes that do not have predicted interactions with PduA ( or PduJ ) exhibit recognizable terminal sequence extensions . Sequence comparisons between the N-terminal peptide tails did not reveal strong similarities ( less than 30% identity overall ) . However , ab initio predictions of their structures consistently modeled them as amphipathic α-helices . Experimental studies have already investigated the possible targeting of some of the Pdu enzymes; targeting by the N-terminal tail of PduP was noted above [50] . In the case of PduD , experiments showed that its N-terminal peptide can be used as a targeting signal , but there was no evidence that it would fold identically to the PduP peptide or that its interaction would be with PduA [54] . In these same studies , PduE was implicated as having such terminal extensions , but fusion of GFP to its respective peptides did not provide clear evidence for targeting . In the case of PduC , Parsons , et al . showed that that enzyme could direct other proteins to the MCP when fused genetically , though the presence of a terminal tail on PduC was not indicated [81] . Despite the mixed findings on terminal targeting peptides on different enzymes in different experimental protocols , the presence on several of the Pdu enzymes in our bioinformatics analysis of extended termini with predicted alpha-helical propensities , and the prediction here of interactions between those enzymes and the PduA shell protein , supported the idea that some of these peptides likely recognize the interior surface of the shell using similar binding modes . Since it was demonstrated that the targeting of PduP is mediated mostly by its terminal peptide segment [50] , we sought to characterize the binding mode of the various implicated enzymes by docking their N-terminal peptides onto the hexameric structure of the PduA shell protein; 18-amino acid terminal segments were used in all cases . The benefits of using a model of the terminal peptide instead of a complete protein are twofold: ( 1 ) to avoid spurious modeling of full-length proteins in the absence of close structural homologs , and ( 2 ) to substantially reduce the size of the search space to be explored by the docking algorithm . In earlier work we proposed a model of the PduP N-terminal extension bound to the concave face of a PduA hexamer ( proposed to be inward facing ) [35] . However , this model was generated with a rigid-body approach , where the PduP peptide had only flexible side chains . Here we push further the flexibility limits of the docking simulation by additionally allowing conformational degrees of freedom for the peptide backbone . To do so , we employed a two-stage docking approach: a rough search by Autodock Vina [83] of the binding site in the PduA hexamer with a rigid helical model of the peptide , followed by a second docking phase with the FlexPepDock protocol from the Rosetta suite [84] . In this second step , the peptide is placed in its start position according to Vina’s predictions; it is then simultaneously refolded and docked over the surface of the receptor . We applied this approach to the five identified PPIs and to a control case involving the N-terminal sequence from PduQ , an aldehyde dehydrogenase from the Pdu pathway that has no obvious targeting signal . In addition , the five peptides were alternatively docked on both faces of the PduA hexamer , with the expectation that meaningful results would have peptides docking to only one side of the PduA shell protein . Results of the peptide docking simulations are overlaid in Fig . 4 along with their energy scores and their buried surface areas . Remarkably , when docked onto the concave ( presumptively luminal ) face of PduA , all five peptides were predicted to bind the same binding cleft formed by the C-terminal segments of two adjacent PduA monomers in the hexamer ( Fig . 4A ) . Moreover , with the exception of PduL , FlexPepDock automatically folded the peptides into well-defined α-helical structures . In the case of PduA-PduP , the model is similar to the one initially proposed in Yeates et al . [35] , with a slight rotation and translation inside the cleft . Interestingly , the different peptides occupy the common binding cleft of PduA in different orientations: PduP and PduE have their N-termini pointing toward the pore , whereas PduC and PduD are docked in the opposite direction . The PduL peptide was also predicted to bind roughly the same region , but the flexible docking protocol did not automatically fold that peptide into a well-ordered alpha helix , leaving the veracity of the predicted binding mode in question in the case of PduL . In their computationally predicted bound configurations , most of the polar residues of the peptides are exposed to the solvent . A notable exception is an arginine recurrently found towards the center of each peptide , which is in all cases buried in the predicted interface and poised to form a salt-bridge with glutamate ( E36 ) of either one of the two monomers constituting the binding cleft ( Fig . 5A ) . The hydrophobic residues are oriented to interact with the C-terminal segment of PduA ( Fig . 5B ) . Various other docking calculations served as computational controls . In contrast to the results obtained for docking to the concave surface of the shell protein , docking of the peptides on the other ( flat ) side of PduA showed no consistent or compelling modes of binding . Those peptide models are instead scattered over the hexamer surface ( Fig . 4B ) . Moreover , comparison of the energy scores and buried surface areas in both docking cases shows that the peptides have a significantly better fit to the concave surface ( Fig . 4C ) . Another control consisted of docking the N-terminal 17 residues from PduQ ( which was not predicted to have a targeting tail ) following the same protocol . In the docking simulation the PduQ peptide partially folds into an α-helix , but does not seem to bind intimately in the canonical cleft ( Fig . 4A ) . An additional calculation involved the docking of N-terminal peptides onto a layer of three PduA hexamers packed side-by-side , to verify that potential binding modes at the interfaces between hexamers were not overlooked . This simulation exhibited similar binding modes to those found with a single PduA hexamer . Overall , these computational predictions and control calculations support the hypothesis that the interior surface of PduA serves as a hub for binding multiple enzymes with terminal extensions . The findings are largely consistent with previous experimental data , while painting a more detailed picture of how interior enzymes in the Pdu MCP interact with PduA , as predicted by our coevolution analysis . As an initial step in modeling a possible interaction between PduU and PduV , which was predicted by the coevolution analysis , a homology model had to be constructed for PduV . The PduV model was then docked into the crystal structure of the PduU hexamer using RosettaDock [85] ( see Methods ) . As a control , we ran two docking simulations under identical conditions on cases involving either PduU or PduV and non-interacting molecules: PduA-PduV , and PduU-ERA ( the homologous GTPase used as the template for modeling of PduV ) . A model of the PduU-PduV complex is proposed in Fig . 6A , along with statistics from the different docking simulations ( Fig . 6B ) . Compared to the two controls , the predicted interface between PduU and PduV achieved a better Rosetta energy score . Likewise , the PduU-PduV complex featured a better shape complementarity and larger buried surface than the controls . In this model , PduV is sitting on the axis formed by the PduU beta-barrel; this PduU protuberance is exclusively contributing to the interface and precludes any interaction between PduV and the main BMC domain of PduU . Most of the interaction surface on PduV is formed by the N-terminal region spanning residue 13 to 35 . This is consistent with preliminary results from Parsons et al . , where the first 42 amino acids from PduV were demonstrated to play a crucial role in PduV targeting to the MCP [82] . As a final control calculation , we investigated the binding mode of PduV after deleting the 17 N-terminal residues forming the β-barrel in the PduU hexamer . Here again , the model yielded worse interaction statistics than for the full-size PduU-PduV complex , supporting the model in which the β-barrel of PduU plays a crucial role in the interaction with PduV ( Fig . 6B ) . Preliminary experimental assays were carried out on the PduU-PduV pair in parallel with our computational analysis . The BacterioMatch II two-hybrid system was used to test for interactions between these two proteins . In this system , a reporter strain is co-transformed with appropriate target and bait fusion genes . A protein-protein interaction between the target and bait activates the transcription of HIS3 , an essential gene for histidine biosynthesis [86] , thereby increasing the expression of the HIS3 product to levels that are sufficient to allow growth on a selective medium lacking histidine and to overcome the effect of 3-amino-1 , 2 , 4-triazole ( 3-AT ) , a competitive inhibitor of the His3 enzyme . If a large number of colonies are obtained following co-transformation , an interaction between the target and bait proteins is indicated . When PduU and PduV were tested , the number of colonies obtained following co-transformation was comparable to that of a positive control with bait and prey proteins ( LGF2 and Gal11P ) that are known to strongly interact ( Table 1 ) . Results showed that PduU and PduV also interacted in reciprocal tests where their roles as bait and prey were reversed ( Table 1 ) . Negative controls showed that PduU or PduV alone did not confer 3-AT resistance ( Table 1 ) . The positive result with the UV pair was confirmed by streptomycin resistance of co-transformed E . coli which requires expression of a second reporter gene , aadA . This experimental confirmation of a PduU-PduV interaction supports the Pdu MCP interactome model developed in the first ( coevolution analysis ) part of our work , while the docking calculations reveal a plausible mode of binding between those proteins .
Proteins rarely carry out biological processes on their own . Instead , they typically participate with other proteins in the context of larger interaction networks . This is especially true for MCPs , where encapsulated pathways require coordination and spatial organization of their numerous components , from shell proteins to enzymes . Though structural studies of individual MCP components have paved the way to a better understanding of their assembly mechanism , a full comprehension of such metabolic systems requires investigation of their PPI networks . Unfortunately , experimental data for MCP protein complexes are still sparse , leading us to turn to predictive methods . Here , we used coevolution calculations and a binary classifier to predict pairwise PPIs in the Pdu MCP , and proposed a model of its interactome . Approaches using binary classifiers for coevolution-based PPI predictions have been developed by others . Comparable approaches have been successfully applied to E . coli [87] , and to the human genome [88] . Interpreting such networks is not a trivial task , considering that such methods are predictive in nature and can therefore include spurious predictions of PPIs or , a contrario , miss true interactions . Additionally , these methods cannot always distinguish direct ( i . e . physical binding ) and indirect ( functional ) correlations , a recurrent problem in coevolution studies that is illustrated here by the integration of PduF in our network . In order to mitigate the deficiencies of the computational methods we employed , a conservative approach was taken by considering only those predicted interactions that had the highest probability ( p≥0 . 7 ) . These cases were largely consistent with existing experimental data , where they were available . An example of a positive result is the agreement between our predictions and structural data relating to the reactivation mechanism of the diol dehydratase [80] . Extending on our predicted interactome model , we focused further analyses on PPIs emanating from the PduA shell protein node and involving Pdu enzymes ( Fig . 3 ) . Of these PPIs , five where identified as presenting an N-terminal extension , a characteristic of lumen-targeted enzymes . These N-terminal peptides , when docked onto a PduA hexamer , consistently bound the same cleft on the concave surface of the hexamer . Likewise , most of them folded into amphipathic α-helical structures , their hydrophobic faces oriented towards the C-terminal tail of the PduA shell protein , a region somewhat less conserved than the main BMC domain . These atomic details are depicted in Fig . 5 , where for example the PduA-PduD case is more clearly pictured . These results are consistent with experimental studies by Fan et al . , which demonstrated the necessity of the PduA C-terminal helix in PduP binding and the role of hydrophobic residues in that interaction [36] . An exceptional case during these docking simulations was the PduL peptide , which did not fold into an amphipathic helix . With regard to our inability to obtain a robust docking result with a PduL peptide , it is notable that the interior vs exterior location of PduL remains unclear in current models of the Pdu MCP . If it is interior , its enzymatic reaction ( depicted in Fig . 1B ) could internally recycle the coenzyme A used by PduP for the conversion of propionaldehyde to propionyl-coA . Indeed , a similar mechanism is used for HS-CoA recycling by the Eut MCP [89] and has been demonstrated for NAD+ recycling by PduQ [90] The results of our docking studies are of particular significance for the issue of sidedness of the MCP shell—i . e . which side of the shell proteins faces inward vs . outward . Previous arguments have suggested that the concave side of the shell protein faces into the MCP lumen [35 , 37 , 38] . Mutagenesis experiments by Fan et al . on the PduA C-terminal helix support that assignment [36] . In our present docking study , the consistent binding of the targeting peptides onto the concave side of the PduA hexamer , and the consistently better interface statistics compared to docking on the other side , strongly corroborate this idea . PduA and PduJ , two highly similar paralogs of the BMC shell protein , are the two most abundant shell proteins after PduBB’ in the Pdu system . As a consequence , they are suspected to play a critical structural role [73] . Indeed , while deletions of pduK , pduT or pduU do not affect the formation of the MCP , pduA mutations produce disrupted or enlarged shells [82 , 91] . Pull-down assays confirmed this architectural importance , where PduA was shown to interact with multiple other shell proteins [82] . Here we suggest that in addition to its transport and structural roles , PduA likely serves as a universal hub for a clique of cargo enzymes , attaching them to the shell via their N-terminal extensions . The highly similar shell protein PduJ is also predicted to interact with four of the same six enzymes associated with PduA . A possible interpretation is that the same clique of enzymes is able to bind both PduA and PduJ , some pairs being more thermodynamically favored than others . Another explanation would be that these PPIs are in fact exclusive , but that our approach is not sensitive enough to discriminate PPIs involving close homologs . Note that the absence of an available structure for PduJ prevented a comparison by computational docking . Whether PduA and PduJ have similar or different affinities for various enzymatic partners will require further investigations , including experimental studies . Attributing a special functional role to PduA ( or PduJ ) is consistent with the view that , though the multiple paralogous shell proteins in the MCP share a canonical BMC structure , each shell protein variant fulfills a specific task . For instance , tandem BMC proteins such as EutL are proposed to regulate the transport of metabolites via conformational changes and a gated pore [38 , 40 , 92 , 93] . The recent crystal structure of PduB , a EutL homolog , presents a view of a tandem domain shell protein from the Pdu system in a closed conformation [46] . Another apparently specialized shell protein is PduT , a tandem BMC domain shell protein that is suspected to bind an iron sulfur cluster in its central pore [39 , 45] . In this portrait of the Pdu family , the role of PduU remains to be elucidated . Here , we aimed to bring new clues by investigating the intriguing PduU-PduV case . Indeed , PduV is also poorly characterized compared to other Pdu components . Furthermore , from our predictions , PduV was the only enzyme exclusively interacting with a shell protein . The diverse docking simulations involving PduU and PduV all agreed with the existence of such a PPI , and predicted the N-terminal region of PduV binds directly to the PduU beta-barrel , consistent with recent experimental data on the importance of the N-terminus of PduV [82] . These predictions , coupled to our preliminary experimental data on a PduU-PduV interaction , fill a gap in understanding the role of the unique β-barrel in PduU . To conclude , the present study brings further insights into the organization of the Pdu MCP , and constitutes the first systematic computational effort to describe an MCP interaction network . The basis of this work is predictive , but we have investigated one of the predicted interactions experimentally as part of this investigation , with a positive result . Further experimental studies will be required to more fully evaluate the interactome model developed here . Application of the same approach to other types of characterized MCPs might be of equal interest and could reveal similar insights .
Protein orthologs were collected from 34 bacterial genomes in the KEGG database [94] and collapsed among the 22 types of MCP proteins known to be associated with the Pdu system: pduABCDEFGHJKLMNOPQSTUVWX ( Suppl . Data ) . Incomplete or erroneous annotations of the Pdu gene products were corrected after sequence comparison with the Pdu operon from Salmonella enterica LT2 , the best-characterized strain in terms of Pdu MCP . For each ortholog group , its corresponding protein sequences were aligned with MUSCLE [95] . The multiple sequence alignments were subsequently input in PhyML [96] for the construction of phylogenetic trees using the Maximum Likelihood method . Since some of the co-evolution descriptors also involve the Tree of Life of the 34 genomes studied , sequences of their respective 16S ribosomal RNA were submitted to similar treatment . For amino acid and nucleotide-based tree construction in PhyML , we used the LG [97] and HKY85 [98] substitution matrices , respectively . Additionally , distance matrices were calculated for each tree , where the distance between two leaves corresponds to the sum of the branch lengths separating them . Seven coevolution descriptors measuring the pairwise tree similarities have been defined . Of these , four are based on pairwise comparison of the distance matrices , as defined in the mirrortree approach , while three others reflect topological similarities ( Fig . 2 ) . In the former class of descriptors , the metrics correspond to the linear correlation coefficient between the two matrices in consideration , while in the latter , it involves the congruence index Icong as defined in Vienne et al [69] . Noteworthy is the fact that comparing two trees can be subject to artefacts and lead in some cases to spurious correlations if speciation events are not taken in account . For this reason , some of these descriptors involve comparisons of the individual proteins to the Tree of Life . Let A and B be the two MCP ortholog groups , mA and mB their respective matrices , tA and tB their trees , and ToL the Tree of Life of the 34 genomes . The parameter mirrorAB is the correlation between mA and mB , mirrorA is between mA and ToL , and mirrorB is between mB and ToL . The fourth descriptor , mirrorAB-ToL , involves an adaptation of the mirror tree , also known as tol-mirror [68] , which measures the correlation between mA and mB after removing the background similarity inherent to speciation events in the ToL . Since distances in the ToL are computed from a nucleotide-based substitution matrix , the distances in the ToL matrix have to be rescaled as proposed in [68] for proper comparison with the protein-based distance matrices . Topological descriptors are derived from the Icong index , defined as the probability that the Maximum Agreement Subtree ( MAST ) between two trees is arising by chance . Along the same idea , topological similarities were computed between tree A and ToL , tree B and ToL , and finally A and B ( topA , topB , topAB ) . We implemented a Random Forests ( RF ) classifier [71] from the Weka Library in Java [99] . Two classes were defined: pos for an interacting protein group pair and neg for those not interacting . Each of the ortholog group pairs sees its input vector of seven coevolution descriptors evaluated by the RF classifier . To classify a pair , its input vector is run through each decision tree of the forest and sees its mean probability attributed . The mean probability threshold for distinguishing the pos from the neg cases was set to 0 . 5 , where a probability ≥ 0 . 5 will classify the pair as pos . The dataset used for training the RF classifier—the “gold standard”—was derived from experimental data found in the literature on the Pdu MCP . Manual mining of this data led to a total of 40 pairs of Pdu proteins whose physical interactions ( or lack of interaction in many cases ) could be verified experimentally via binding assays [36 , 82 , 90 , 100 , 101] , complementation and expression studies [22] or crystallographic data [79] . An example of a verified non-interaction would be a direct binding experiment in which one protein component of a candidate pair failed to pull down the other . Among these , 16 are actual PPIs while the remaining 24 are non-interacting pairs . Of the 16 PPIs , 4 , 6 and 6 pairs fall within the categories of: shell-enzyme ( S-E ) interactions , shell-shell ( S-S ) interactions and enzyme-enzyme ( E-E ) interactions , respectively . Likewise , the non-interacting pairs include 12 S-E , 12 S-S and zero E-E interactions . Each of these cases was assigned a class according to the rules defined earlier . The reported AUC value ( 0 . 75 ) for the classifier was calculated after a 10-fold cross validation . In parallel , we also carried out a 5 -fold cross validation that yielded a comparable AUC ( 0 . 73 ) . The interactome was pictured as an undirected graph with the igraph library in R [102] . Nodes and edges were computed with a Fruchterman-Reingold layout [103] . While the structures of PduA and PduU are available in the PDB [104] , structural information on the specific enzymes believed to interact with the shell proteins is limited . A recent NMR structure of the PduP tail showed an alpha helical structure , consistent with sequence-based predictions . Similar data are not available for the tails of the other enzymes of interest . We elected to assume as little as possible about the various tail structures and to model ab initio the first 18 residues of each enzyme with the PEP-FOLD server [105] . PduV was not presumed or predicted to bind by way of a terminal extension , so a model of that intact enzyme fold was required for docking analysis . The structure of PduV is presently unknown . Therefore , to enable computational docking , we built a homology model with I-TASSER [106] by threading the sequence of PduV onto two structural templates from the PDB ( 3IEV_A and 3R9W_A ) . The final model achieved a TM-score of 0 . 76 , which is reasonable for further investigation by docking simulations [107] . For protein-peptide docking , our approach relied mainly on the Rosetta-based protocol FlexPepDock [84] . Its ability to simultaneously fold and dock allows full flexibility of the peptide . However , FlexPepDock sees its accuracy decrease when the starting peptide conformation has an RMSD higher than 5 . 5 Å compared to the native structure . Mindful of this constraint , we designed a two-step method for docking the N-terminal enzymatic peptides onto the PduA hexamer . The first stage is a coarse-grained search of the approximate binding mode by AutoDock Vina [83] . This model is subsequently refined by an ab initio FlexPepDock run , where the Vina model is used as an input coordinates file . Vina has been designed for small molecule docking , which allows a ligand flexibility up to 32 rotatable bonds only , a limit not existing in FlexPepDock . However , it can still be used efficiently when medium-sized ligands like peptides are treated as semi-rigid for predicting an approximate binding region . File preparation for AutoDock Vina included a configuration file specifying an exhaustiveness of 10 and a 27000 Å3 grid box encompassing the surface of the hexamer and centered on the pore . Coordinate files in PDBQT format were generated from the PduA crystal structure and the PEP-FOLD models of each peptide . For the peptides , rotatable bonds were defined for the side chains while Kollman United Atom charges were assigned to both the hexamer and the peptides . The pose computed by Vina with the lowest energy score was subsequently considered as the starting point for FlexPepDock . In this second stage , we ran 10000 simulations where the peptide was completely refolded and docked into the PduA hexamer . After ranking the 10 000 poses by lowest Rosetta energy , the top 500 poses were collapsed into clusters for which the internal RMSD was less than 2 . 5 Å . Finally , we picked the definitive model as the one with the lowest energy among the two most populated clusters . For the PduU-PduV case , we used a standard RosettaDock protocol where the input included coordinates of both partners in their unbound state , typically those from the PduU hexamer and and the PduV homology model . The number of simulations , the ranking , clustering and selecting methods were identical to the FlexPepDock procedure , while the allowed flexibility in this case is limited to the side chains . To test for interactions between PduU and PduV , the BacterioMatch II two-hybrid system ( Agilent technologies ) was used according to the manufacturer’s instructions with the following modification: co-transformation was carried out by using 30 ng each of the bait and prey vector . To construct the needed plasmids , pduU and pduV DNA sequences were amplified by PCR and then restricted and ligated into pBT for expression as fusions with the λcI protein , and into pTRG for expression as fusions with the RNAPα protein . Ligation reactions were used to transform E . coli XL1-Blue MRF’ . Plasmid DNA was purified using a Qiagen mini prep kit , and all clones were verified by DNA sequencing . Self-activation by each recombinant bait and prey was tested before the two-hybrid interaction assays to determine if the bait or prey was capable of activating the reporter cassette on its own . Determination of protein-protein interaction was carried out by co-transforming BacterioMatch II validation reporter competent cells using recombinant bait and target .
|
Many bacteria produce giant proteinaceous structures within their cells , which they use to carry out special metabolic reactions in their interior . Much has been learned recently about the individual components—shell proteins and encapsulated enzymes—that assemble together , thousands of subunits in all , to make these bacterial microcompartments or MCPs . However , in order to carry out their biological functions , these systems must be highly organized through specific protein-protein interactions , and such a higher level understanding of organization in MCP systems is lacking . In this study , we use genomic data and phylogenetic analysis to predict the network of interactions between the approximately 20 different kinds of proteins and enzymes present in the Pdu MCP . Then , we use computational docking to examine a subset of those that are predicted to involve enzymes bound to the interior surface of the shell proteins , and show that the results are consistent with recent experimental data . We further provide new experimental evidence for one of the predicted protein-protein interactions . This study expands our understanding of a complex system of proteins serving as a metabolic organelle in bacterial cells , and provides a foundation for further experimental investigations .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Exploring Bacterial Organelle Interactomes: A Model of the Protein-Protein Interaction Network in the Pdu Microcompartment
|
Analysis by liquid chromatography and tandem mass spectrometry ( LC-MS/MS ) can identify and quantify thousands of proteins in microgram-level samples , such as those comprised of thousands of cells . This process , however , remains challenging for smaller samples , such as the proteomes of single mammalian cells , because reduced protein levels reduce the number of confidently sequenced peptides . To alleviate this reduction , we developed Data-driven Alignment of Retention Times for IDentification ( DART-ID ) . DART-ID implements principled Bayesian frameworks for global retention time ( RT ) alignment and for incorporating RT estimates towards improved confidence estimates of peptide-spectrum-matches . When applied to bulk or to single-cell samples , DART-ID increased the number of data points by 30–50% at 1% FDR , and thus decreased missing data . Benchmarks indicate excellent quantification of peptides upgraded by DART-ID and support their utility for quantitative analysis , such as identifying cell types and cell-type specific proteins . The additional datapoints provided by DART-ID boost the statistical power and double the number of proteins identified as differentially abundant in monocytes and T-cells . DART-ID can be applied to diverse experimental designs and is freely available at http://dart-id . slavovlab . net .
Advancements in the sensitivity and discriminatory power of protein mass-spectrometry ( MS ) have enabled the quantitative analysis of increasingly limited amounts of samples . Recently , we have developed Single Cell Proteomics by Mass Spectrometry ( SCoPE-MS ) . SCoPE-MS uses a barcoded carrier to boost the MS signal from single-cells and enhance sequence identification [1 , 2] . While this design allows quantifying hundreds of proteins in single mammalian cells , sequence identification remains challenging because many lowly abundant peptides generate only a few fragment ions that are insufficient for confident identification [3 , 4] . Such low confidence peptides are generally not used for protein quantification , and thus reduce the data points available for further analyses . We sought to overcome this challenge by using both the retention time ( RT ) of an ion and its MS/MS spectra to achieve more confident peptide identifications . To this end , we developed a novel data-driven Bayesian framework for aligning RTs and for updating peptide confidence . DART-ID minimizes assumptions , aligns RTs with median residual error below 3 seconds , and increases the fraction of cells in which peptides are confidently identified . Multiple existing approaches—including Skyline ion matching [5] , moFF match-between-runs [6] , MaxQuant match-between-runs [7 , 8] , DeMix-Q [9] and Open-MS FFId [10]—allow combining MS1 spectra with other informative features , such as RT and precursor ion intensity , to enhance peptide identification . These methods , in principle , may identify any ion detected in a survey scan ( MS1 level ) even if it was not sent for fragmentation and second MS scan ( MS2 ) in every run . Thus by not using MS2 spectra , these methods may overcome the limiting bottleneck of tandem MS: the need to isolate , fragment and analyze the fragments in order to identify and quantify the peptide sequence . However not using the MS2 spectra for identification has a downside: The MS2 spectra contain highly informative features even for ions that could not be confidently identified based on spectra alone . This is particularly important when MS/MSed ions are the only ones that can be quantified , as in the case of isobaric mass tags . Thus , the MS1-based methods have a strong advantage when quantification relies only on MS1 ions ( e . g . , LFQ [11] , and SILAC [12] ) , while methods using all MS2 spectra can more fully utilize all quantifiable data from isobaric tandem-mass-tag experiments . DART-ID aims to use all MS2 spectra , including those of very low confidence PSMs , and combines them with accurate RT estimates to update peptide-spectrum-match ( PSM ) confidence within a principled Bayesian framework . Unlike previous MS2-based methods which incorporate RT estimates into features for FDR recalculation [13] , discriminants [14] , filters [15–17] , or scores [18 , 19] , we update the ID confidence directly with a Bayesian framework [20 , 21] . Crucial to this method is the accuracy of the alignment method; the higher the accuracy of RT estimates , the more informative they are for identifying the peptide sequence . The RT of a peptide is a specific and informative feature of its sequence , and this specificity has motivated approaches aiming to estimate peptide RTs . These approaches either ( i ) predict RTs from peptide sequences or ( ii ) align empirically measured RTs . Estimated peptide RTs have a wide range of uses , such as scheduling targeted MS/MS experiments [22] , building efficient inclusion and exclusion lists for LC-MS/MS [23 , 24] , or augmenting MS2 mass spectra to increase identification rates [14–19] . Peptide RTs can be estimated from physical properties such as sequence length , constituent amino acids , and amino acid positions , as well as chromatography conditions , such as column length , pore size , and gradient shape . These features predict the relative hydrophobicity of peptide sequences and thus RTs for LC used with MS [25–31] . The predicted RTs can be improved by implementing machine learning algorithms that incorporate confident , observed peptides as training data [15 , 19 , 32–35] . Predicted peptide RTs are mostly used for scheduling targeted MS/MS analyses where acquisition time is limited , e . g . , multiple reaction monitoring [22] . They can also be used to aid peptide sequencing , as exemplified by “peptide fingerprinting”—a method that identifies peptides based on an ion’s RT and mass over charge ( m/z ) [28 , 36–38] . While peptide fingerprinting has been successful for low complexity samples , where MS1 m/z and RT space is less dense , it requires carefully controlled conditions and rigorous validation with MS2 spectra [37–41] . Predicted peptide RTs have more limited use with data-dependent acquisition , i . e . , shotgun proteomics . They have been used to generate data-dependent exclusion lists that spread MS2 scans over a more diverse subset of the proteome [23 , 24] , as well as to aid peptide identification from MS2 spectra , either by incorporating the RT error ( difference between predicted and observed RTs ) into a discriminant score [14] , or filtering out observations by RT error to minimize the number of false positives selected [15–17] . In addition , RT error has been directly combined with search engine scores [18 , 19] . Besides automated methods of boosting identification confidence , proteomics software suites such as Skyline allow the manual comparison of measured and predicted RTs to validate peptide identifications [5] . The second group of approaches for estimating peptide RTs aligns empirically measured RTs across multiple experiments . Peptide RTs shift due to variation in sample complexity , matrix effects , column age , room temperature and humidity . Thus , estimating peptide RTs from empirical measurements requires alignment that compensates for RT variation across experiments . Usually , RT alignment methods align the RTs of two experiments at a time , and typically utilize either a shared , confidently-identified set of endogenous peptides , or a set of spiked-in calibration peptides [42 , 43] . Pairwise alignment approaches must choose a particular set of RTs that all other experiments are aligned to , and the choice of that reference RT set is not obvious . Alignment methods are limited by the availability of RTs measured in relevant experimental conditions , but can result in more accurate RT estimates when such empirical measurements are available [7 , 8 , 43] . Generally , RT alignment methods provide more accurate estimates than RT prediction methods , discussed earlier , but also generally require more extensive data and cannot estimate RTs of peptides without empirical observations . Methods for RT alignment are various , and range from linear shifts to non-linear distortions and time warping [44] . Some have argued for the necessity of non-linear warping functions to correct for RT deviations [45] , while others have posited that most of the variation can be explained by simple linear shifts [46] . More complex methods include multiple generalized additive models [47] , or machine-learning based semi-supervised alignments [48] . Once experiments are aligned , peptide RTs can be predicted by applying experiment-specific alignment functions to the RT of a peptide observed in a reference run . Peptide RTs estimated by alignment can be used to schedule targeted MS/MS experiments—similar to the use of predicted RTs estimated from the physical properties of a peptide [43] . RT alignments are also crucial for MS1 ion/feature-matching algorithms , as discussed earlier [5–10] , as well as in targeted analyses of results from data-independent acquisition ( DIA ) experiments [49–51] . The addition of a more complex , non-linear RT alignment model that incorporates thousands of endogenous peptides instead of a handful of spiked-in peptides increased the number of identifications in DIA experiments by up to 30% [52] . With DART-ID , we implement a novel global RT alignment method that takes full advantage of SCoPE-MS data , which feature many experiments with analogous samples run on the same nano-LC ( nLC ) system [1 , 2] . These experimental conditions yield many RT estimates per peptide with relatively small variability across experiments . In this context , we used empirical distribution densities that obviated assumptions about the functional dependence between peptide properties , RT , and RT variability and thus maximized the statistical power of highly reproducible RTs . This approach increases the number of experiments in which a peptide is identified with high enough confidence and its quantitative information can be used for analysis . The DART-ID program is freely available and can easily be run over the output of peptide search engines such as MaxQuant [7 , 8] .
Using RT for identifying peptide sequences starts with estimating the RT for each peptide , and we aimed to maximize the accuracy of RT estimation by optimizing RT alignment . Many existing methods can only align the RTs of two experiments at a time , i . e . , pairwise alignment , based on partial least squares minimization , which does not account for the measurement errors in RTs [53] . Furthermore , the selection of a reference experiment is non-trivial , and different choices can give quantitatively different alignment results . In order to address these challenges , we developed a global alignment method , sketched in Fig 1a and 1b . The global alignment infers a reference RT for the ith peptide , μi as a latent variable with value μik in the kth experiment . This can be related to the measured RT for peptide i in experiment k , ρik . ρ i k = μ i k + ϵ i k ( 1 ) where μik ≜ gk ( μi ) and ϵik is an independent mean-zero error term expressing residual ( unmodeled ) RT variation . As a first approximation , we assume that the observed RTs for any experiment can be well approximated using a two-segment linear regression model: g k ( μ i ) = { β 0 k + β 1 k μ i if μ i < s k β 0 k + β 1 k s k + β 2 k ( μ i - s k ) if μ i ≥ s k ( 2 ) where sk is the split point for the two segment regression in each experiment , and the parameters are constrained to not produce a negative RT and can be generalized to more complex monotonically-constrained models , such as spline fitting or locally estimated scatterplot smoothing ( LOESS ) . We chose this model since we found that it outperformed a single-slope linear model by capturing more of the inter-experiment variation in RTs , S2 Fig . Based on this model , we can express the marginal likelihood for the RT of the ith peptide in the kth experiment as a mixture model weighted by the probability of correct sequence assignment ( λik , the spectral posterior error probability ( PEP ) ) ; see S1 Fig for more details . P ( ρ i k | μ i k , σ i k , λ i k ) ︸ Likelihood ∝ ( 1 - λ i k ) × f i k ( ρ i k ∣ μ i k , σ i k ) ︸ PSM is correct + ( λ i k ) × f k 0 ( ρ i k ) ︸ PSM is incorrect ( 3 ) where fik is the inferred RT density for peptide i in experiment k and f k 0 is the null RT density . In our implementation , we let f i k ∼ Laplace ( μ i k , σ i k 2 ) and f k 0 ∼ Normal ( μ k , σ k 2 ) , which we found worked well in practice ( See S4 Fig ) . This framework is modular and can be easily extended to use different distributions . To account for the fact that residual RT variation increases with mean RT and varies between experiments ( S3 Fig ) , we model its standard deviation , σik , as a linearly increasing function of μi , Eq 7 . Using the vectorized likelihood function from Eq 3 and the priors described in Methods , we solve Eq 4 to infer the joint posterior distribution of all reference RTs ( and associated model parameters ) across all experiments: P ( a , b , β 0 , β 1 , s , μ ∣ ρ , λ ) ︸ Posterior ∝ P ( ρ ∣ a , b , β 0 , β 1 , s , μ , λ ) ︸ Likelihood Eq 3 P ( a , b , β 0 , β 1 , s , μ ) ︸ Prior ( 4 ) The inference described above infers all reference RTs , μ , from one global solution of Eq 4 . It allows the alignment to take advantage of any peptide observed in at least two experiments , regardless of the number of missing observations in other experiments . Furthermore , the mixture model described in Eq 3 allows for the incorporation of low confidence peptides by using appropriate weights and accounting for the presence of false positives . Thus this method maximizes the data used for alignment and obviates the need for spiked-in standards . Furthermore , the reference RT provides a principled choice for a reference ( rather than choosing a particular experiment ) that is free of measurement noise . The alignment process accounts for the error in individual observations by inferring a per peptide RT distribution , as opposed to aligning to a point estimate , as well as for variable RT deviations in experiments by using experiment-specific weights . The conceptual idea based on which we incorporate RT information for sequence identification is illustrated in Fig 1c and formalized with Bayes’ theorem in Fig 1d . We start with a peptide-spectrum-match ( PSM ) from a search engine and its associated probability to be incorrect ( PEP; posterior error probability ) and correct , 1-PEP . If the RT of a PSM is far from the RT of its corresponding peptide , as PSM1 in Fig 1c , then the spectrum is more likely to be observed if the PSM is incorrect , and thus we can decrease its confidence . Conversely , if the RT of a PSM is very close to the RT of its corresponding peptide , as PSM2 in Fig 1c , then the spectrum is more likely to be observed if the PSM is correct , and thus we can increase its confidence . To estimate whether the RT of a PSM is more likely to be observed if the PSM is correct or incorrect , we use the conditional likelihood probability densities inferred from the alignment procedure in Eq 3 ( Fig 1b ) . Combining these likelihood functions with Bayes’ theorem in Fig 1d allows us to formalize this logic and update the confidence of analyzed PSMs , which we quantify with DART-ID PEPs . To evaluate the global RT alignment by DART-ID , we used a staggered set of 46 60-minute LC-MS/MS runs performed over a span of 3 months . Each run was a diluted 1 × M injection of a bulk 100 × M SCoPE-MS sample , as described in Table 1 and by Specht et al . [2] . The experiments were run over a span of three months so that the measured RTs captured expected variance in the chromatography . The measured RTs were compared to RTs predicted from peptide sequences [30 , 31 , 34] , and to top-performing alignment methods [7 , 8 , 43 , 52] , including the reference RTs from DART-ID; see Fig 2a . All methods estimated RTs that explained the majority of the variance of the measured RTs , Fig 2a . As expected , the alignment methods provided closer estimates , explaining over 99% of the variance . To evaluate the accuracy of RT estimates more rigorously , we compared the distributions of differences between the reference RTs and measured RTs , shown in Fig 2b . This comparison again underscores that the differences are significantly smaller for alignment methods , and smallest for DART-ID . We further quantified these differences by computing the mean and median absolute RT deviations , i . e . , |ΔRT| , which is defined as the absolute value of the difference between the observed RT and the reference RT . For the prediction methods—SSRCalc [30] , BioLCCC [31] , and ELUDE [34]—the average deviations exceed 2 min , and ELUDE has the smallest average deviation of 2 . 5 min . The alignment methods result in smaller average deviations , all below < 1 min , and DART-ID shows the smallest average deviation of 0 . 044 min ( 2 . 6 seconds ) . Detailed alignment statistics can be visualized in both the graphical output of the DART-ID program and in the DO-MS visualization platform [54] . Search engines such as MaxQuant [7 , 8] use the similarity between theoretically predicted and experimentally measured MS2 spectra of ions to match them to peptide sequences , i . e . , peptide-spectrum-matches ( PSM ) . The confidence of a PSM is commonly quantified by the probability of an incorrect match: the posterior error probability ( PEP ) [21 , 55 , 56] . Since the estimation of PEP does not include RT information , we sought to update the PEP for each PSM by incorporating RT information within the Bayesian framework displayed in Fig 1c and 1d . This approach allowed us to use the estimated RT distributions for each peptide with minimal assumptions . The Bayesian framework outlined in Fig 1c and 1d can be used with RTs estimated by other methods , and its ability to upgrade PSMs is directly proportional to the accuracy of the estimated RTs . To explore this possibility , we used our Bayesian model with RTs estimated by all methods shown in Fig 2 . The updated error probabilities of PSMs indicate that all RT estimates enhance PSM discrimination , S5 Fig . Even lower accuracy RTs predicted from peptide sequence can be productively used to upgrade PSMs . However , the degree to which PSMs are upgraded , i . e . the magnitude of the confidence shift , increases with the accuracy of the RT estimates and is highest with the DART-ID reference RTs . We refer to the PEP assigned by the search engine ( MaxQuant throughout this paper ) as “Spectral PEP” , and after it is updated by the Bayesian model from Fig 1d as “DART-ID PEP” . Comparing the Spectral and DART-ID PEPs indicates that the confidence for some PSMs increases while for others decreases; see density plot in Fig 3a . Reassuringly , all PSMs with low Spectral PEPs have even lower DART-ID PEPs , meaning that all confident PSMs become even more confident . On the other extreme , many PSMs with high Spectral PEPs have even higher DART-ID PEPs , meaning that some low-confidence PSMs are further downgraded . Confidence upgrades , where DART-ID PEP < Spectral PEP , range within 1–3 orders of magnitude . The density plot in Fig 3a displays a subset of peptides with Spectral PEP > 0 . 01 and DART-ID PEP < 0 . 01 . These peptides have low confidence of identification based in their MS/MS spectra alone , but high confidence when RT evidence is added to the spectral evidence . To visualize how these peptides are distributed across experiments , we marked them with red dashes in Fig 3b . The results indicate that the data sparsity decreases; thus DART-ID helps mitigate the missing data problem of shotgun proteomics . Fig 3b is separated into two subsets , DART-ID1 and DART-ID2 , which correspond respectively to peptides that have at least one confident spectral PSM , and peptides whose spectral PSMs are all below the set confidence threshold of 1% FDR . While the PSMs of DART-ID2 very likely represent the same peptide sequence—since by definition they share the same RT , MS1 m/z and MS2 fragments consistent with its sequence—we cannot be confident in the exact sequence assignment . Thus , they are labeled separately and their sequence assignment is further validated in the next section . The majority of PSMs whose confidence is increased by DART-ID have multiple confident Spectral PSMs , and thus reliable sequence assignment . Analysis of newly identified peptides in Fig 3c shows that DART-ID helps identify about 50% more PSMs compared to spectra alone at an FDR threshold of 1% . This corresponds to an increase of ∼30–50% in the fraction of PSMs passing an FDR threshold of 1% , as shown in the bottom panel of Fig 3c . Furthermore , the number of distinct peptides identified per experiment increases from an average of ∼1000 to an average of ∼1600 , Fig 3d . Percolator , a widely used FDR recalculation method that also incorporates peptide RTs [13] , also increases identification rates , albeit to a lesser degree than DART-ID , Fig 3c and 3d . The visualizations in Fig 3a , 3c and 3d can be generated for user inputted data by the DO-MS visualization platform [54] . We observe that DART-ID PEPs are bimodally distributed ( S6 Fig ) , suggesting that DART-ID acts as an efficient binary classifier . Modifying error probabilities , however , does risk changing the overall false discovery rate ( FDR ) of the PSM set . To evaluate the effect of DART-ID on the overall FDR , we allowed the inclusion of decoy hits in both the alignment and confidence update process [55] . The results from this analysis in Fig 3e indicate that , as expected , the fraction of PSMs matched to decoys is proportional to the FDR estimated both from the Spectral PEP and from the updated DART-ID PEP . We encourage users of DART-ID to evaluate the results from applying DART-ID and other related methods on their datasets using this benchmark as well as the numerous quantitative benchmarks described in the subsequent sections . While we were motivated to develop DART-ID within the context of the SCoPE-MS method , we show in Fig 4 that DART-ID is similarly able to increase quantitative coverage in a label-free [57] and a TMT-labelled [58] bulk LC-MS/MS experiment . The DART-ID alignment performed differently between the label-free set ( 120 min gradients ) and the TMT-labelled set ( 180 min gradients ) Fig 4a , with slightly higher residuals for the longer gradient . The percent increase in confident PSMs , when using DART-ID PEPs instead of spectral PEPs Fig 4b , also fell into the expected range of 30–50% at 1% FDR . The increase in confident PSMs is shown in discrete terms in Fig 4c , where experiments in both the label-free and TMT-labelled sets receive thousands of more confident PSMs that can then be used for further quantitative analysis . These increases of confident PSMs , in both the SCoPE-MS and bulk LC-MS/MS sets , decreases the amount of missing data per run . In Fig 5a we show qualitatively that DART-ID can fill in many of these missing values on the protein level . On the level of experimental runs , as shown quantitatively in Fig 5b , DART-ID significantly reduces the amount of missing data and mitigates the stochasticity that is inherently to data-dependent MS methods . We next sought to evaluate whether the confident DART-ID PSMs without confident Spectral PSMs , i . e . DART-ID2 from Fig 3b , are matched to the correct peptide sequences . To this end , we sought to evaluate whether the RTs of such PSMs match the RTs for the corresponding peptides identified from high-quality , confident spectra . For this analysis , we split a set of experiments into two subsets , A and B , Fig 6a . The application of DART-ID to A resulted in two disjoint subsets of PSMs: A1 , corresponding to PSMs with confident spectra ( Spectral PEP < 0 . 01 ) , and A2 , corresponding to “upgraded” PSMs ( Spectral PEP > 0 . 01 and DART-ID PEP < 0 . 01 ) . We overlapped these subsets with PSMs from B having Spectral PEP < 0 . 01 , so that the RTs of PSMs from B can be compared to the RTs of PSMs from subsets A1 and A2 , Fig 6a . This comparison shows excellent agreement of the RTs for both subsets A1 and A2 with the RTs for high quality spectral PSMs from B , Fig 6b and 6c . This result suggests that even peptides upgraded without confident spectral PSMs are matched to the correct peptide sequences . We ran DART-ID on SCoPE-MS method development experiments [2] , all of which contain quantification data in the form of 11-plex tandem-mass-tag ( TMT ) reporter ion ( RI ) intensities . Out of the 10 TMT “channels” , six represent the relative levels of a peptide in simulated single cells , i . e . , small bulk cell lysate diluted to a single cell-level level . These six single cell channels are made of T-cells ( Jurkat cell line ) and monocytes ( U-937 cell line ) . We then used the normalized TMT RI intensities to validate upgraded PSMs by analyzing the consistency of protein quantification from distinct peptides . Internal consistency is defined by the expectation that the relative intensities of PSMs reflect the relative levels of their corresponding proteins . If upgraded PSMs are consistent with Spectral PSMs for the same protein , then their relative RI intensities will have lower coefficients of variation ( CV ) within a protein than across different proteins [59] . CV is defined as σ/μ , where σ is the standard deviation and μ is the mean of the normalized RI intensities of PSMs belonging to the same protein . A negative control is constructed by creating a decoy dataset where PSM protein assignments are randomized . For this and later analyses , we filter PSMs from our data into the following disjoint sets: where Spectra is disjoint from the other two sets , i . e . , Spectra ∩ DART-ID = ∅ and Spectra ∩ Percolator = ∅ . These sets of PSMs , as depicted in Fig 7a , are intersected with each other through a set of shared proteins between the three sets of PSMs . The protein CVs of the Spectra , DART-ID , and Percolator PSM sets , depicted in Fig 7b , show similar distributions and smaller CVs than those from the decoy set . In addition , Fig 7c shows agreement between the protein CVs of the Spectra and DART-ID PSM sets , as opposed to the CVs of the Spectra set and Decoy set . This demonstrates that the protein-specific variance in the relative quantification , due to either technical or biological noise , is preserved in these upgraded PSMs . The upgraded PSMs from the DART-ID set are not just representative of proteins already quantified from confident spectral PSMs , but when filtering at a given confidence threshold ( e . g . , 1% FDR ) , they allow for the inclusion of new proteins for analysis . As the quantification of these new proteins from the DART-ID PSMs cannot be directly compared to that of the proteins from the Spectra PSMs , we instead compare how the new proteins from DART-ID can explain the biological differences between two cell types—T-cells ( Jurkat cell line ) and monocytes ( U-937 cell line ) —present in each sample and experiment . The data was split into sets in the same manner as the previous section , as shown in Fig 7a , where the Spectra and DART-ID sets of PSMs are disjoint . We then filtered out all PSMs from DART-ID that belonged to any protein represented in Spectra , so that the sets of proteins between the two sets of PSMs were disjoint as well . To test whether or not DART-ID identified peptides consistently across experiments , we used principal component analysis ( PCA ) to separate the T-cells and monocytes quantified in our experiments . This PCA analysis in Fig 8a shows clear separation of T-cells and monocytes from both the Spectra and DART-ID PSM sets . If boosted peptide identifications were spurious and inconsistent , then the PCA analysis could not separate the cell types or cluster them together . In addition , relative protein ratios ( T-cells/monocytes ) estimated from the two disjoint PSM sets are in good agreement ( ρ = 0 . 84 ) ; see S7 Fig . While DART-ID2 PSMs are able to uncover entirely new proteins carrying consistent biological signal , on average these PSMs differ slightly from Spectral PSMs in purity , missed-cleavages , and missing data; see Fig 8b . However , the distributions of these features are largely overlapping , and the magnitude of these differences are relatively small; most spectra of DART-ID PSMs are still >90% pure , and have less than 16% missing data and missed cleavages . Of course the intended usage of DART-ID is not to separate these two groups of PSMs and analyze them separately , but instead to combine them and increase the number of data points available for analysis . Indeed , adding DART-ID PSMs to the Spectra PSMs doubles the number of differentially abundant proteins between T-cells and monocytes , Fig 9a , 9b and 9c .
Here we present DART-ID , a new Bayesian approach that infers RTs with high accuracy and uses these accurate RT estimates to improve peptide sequence identification . We demonstrate that DART-ID can estimate and align RTs with accuracy of a few seconds for 60 minute LC-MS/MS runs and can leverage this high accuracy towards increasing the confidence in correct PSMs and decreasing the confidence in incorrect PSMs . This principled and rigorous estimation of the confidence of PSMs increases quantification coverage by 30–50% , primarily by increasing the number of experiments in which a peptide is quantified . We validated the upgraded PSMs using methods for FDR estimation ( Fig 3e ) , cross-validation ( Fig 6 ) , intra-protein CV validation ( Fig 7 ) , and biological signal validation ( Fig 8 ) . All of these methods strongly support the reliability of DART-ID inferences . We encourage the use of these methods for benchmarking the application of DART-ID ( and any other related method ) on other datasets . DART-ID is applicable to any large set of LC-MS/MS analyses with a consistent LC setup . The more consistent the LC , the more powerful DART-ID is since its statistical power is proportional to the accuracy of RT estimates . Our SCoPE-MS and SCoPE2 runs have highly consistent RTs [1 , 4 , 60] and motivated us to develop DART-ID . However , we found ( show in Fig 4 ) that DART-ID performs similarly well with bulk LC-MS/MS runs of TMT-labeled and label-free samples . A principal advantage of DART-ID is that its probabilistic model naturally adapts to the RT reproducibility and obviates thresholds , e . g . , a threshold on RT errors . Rather DART-ID updates the confidence of each PSMs using a rigorous quantitative model based on empirically derived distributions of RT reproducibility . Thus , it adapts and controls for the reproducibility of the LC and the accuracy of the RT estimates as shown in S5 Fig . Another principal advantage of DART-ID is its ability to use all PSMs ( including those with sparse observations and low confidence ) to create a global RT alignment . This is possible because DART-ID alignment takes into account the confidence of PSMs as part of the mixture model in Eq 3 . This results in accurate RT estimates ( Fig 2 ) that are robust to missing data and benefit from all PSMs regardless of their identification confidence . If the LC and RTs of a dataset are very variable , one may extend the alignment model beyond Eq 2 to capture the increased variability . The two-segment linear regression from Eq 2 demonstrated here captures more variation than a single-slope linear regression . DART-ID , however , is not constrained to these two functions and can implement any monotone function . Non-linear functions that are monotonically constrained , such as the logit function , have been implemented in our model during development . More complex models , for example monotonically-constrained general additive models , could increase alignment accuracy further given that the input data motivates added complexity . While DART-ID is focused on aligning and utilizing RTs from LC-MS/MS experiments , the alignment method could potentially be applied to other separation methods , including ion mobility , gas chromatography , supercritical fluid chromatography , and capillary electrophoresis . The ion drift time obtained from instruments with an ion mobility cell are particularly straightforward to align and incorporate by DART-ID’s Bayesian framework . Another potential extension of DART-ID is to offline separations prior to analysis , i . e . , fractionation . RT alignment would only be applicable between replicates of analogous fractions , but a more complex model could also take into account membership of a peptide to a fraction as an additional piece of evidence . DART-ID is modular , and the RT alignment module and PEP update modules may be used separately . For example , the RT estimates may be applied to increase the performance of other peptide identification methods incorporating RT evidence [14–17] . One application is integrating the inferred RT from DART-ID into the search engine score , as done by previous methods [18 , 19] , to change the best hit for a spectrum , save a spectrum from filtering due to high score similarities ( i . e . , low delta score ) [21] , or provide evidence for hybrid spectra . Although DART-ID’s alignment is based on point estimates of RT , the global alignment methodology could also be applied to feature-based alignments [6 , 8–10] to obviate the limitations inherent in pairwise alignments .
The data used for the development and validation of the DART-ID method were 263 method-development experiments for SCoPE-MS and its related projects . All samples were lysates of the Jurkat ( T-cell ) , U-937 ( monocyte ) , or HEK-293 ( human embryonic kidney ) cell lines . Samples were prepared with the mPOP sample preparation protocol , and then digested with trypsin [2] . All experiments used either 10 or 11-plex TMT for quantification . Most but not all sets followed the experimental design as described by Table 1 . All experiments were run on a Thermo Fisher ( Waltham , MA ) Easy-nLC system with a Waters ( Milford , MA ) 25cm x 75μm , 1 . 7μm BEH column with 130Å pore diameter , and analyzed on a Q-Exactive ( Thermo Fisher ) mass spectrometer . Gradients were run at 100 nL/min from 5-35%B in 48 minutes with a 12 minute wash step to 100%B . Solvent composition was 0% acetonitrile for A and 80% acetonitrile for B , with 0 . 1% formic acid in both . A subset of later experiments included the use of a trapping column , which extended the total run-time to 70 minutes . Detailed experimental designs and mass spectrometer parameters of each run can be found in S1 Table . All Thermo . RAW files are publicly available online . More details on sample preparation and analysis methods can be found from the mPOP protocol [2] . Searching was done with MaxQuant v1 . 6 . 1 . 0 [7] against a UniProt protein sequence database with 443722 entries . The database contained only SwissProt entries and was downloaded on 5/1/2018 . Searching was also done on a contaminant database provided by MaxQuant , which contained common laboratory contaminants and keratins . MaxQuant was run with Trypsin specificity which allowed for two missed cleavages , and methionine oxidation ( +15 . 99492 Da ) and protein N-terminal acetylation ( +42 . 01056 Da ) as variable modifications . No fixed modifications apart from TMT were specified . TMT was searched using the “Reporter ion MS2” quantification setting on MaxQuant , which searches for the TMT addition on lysine and the n-terminus with a 0 . 003 Da tolerance . Observations were selected at a false discovery rate ( FDR ) of 100% at both the protein and PSM level to obtain as many spectrum matches as possible , regardless of their match confidence . All raw MS files , MaxQuant search parameters , the sequence database , and search outputs are publicly available online . Only a subset of the input data is used for the alignment of experiments and the inference of RT distributions for peptides . First , decoys and contaminants are filtered out of the set . Contaminants may be problematic for RT alignment since their retention may be poorly defined , e . g . , they may be poorly chromatographically resolved . Then , observations are selected at a threshold of PEP < 0 . 5 . Observations are additionally filtered through a threshold of retention length , which is defined by MaxQuant as the range of time between the first matched scan of the peptide and the last matched scan . Any peptide with retention length > 1 min for a 60 min run is deemed to have too wide of an elution peak , or chromatography behavior more consistent with contaminants than retention on column . In our implementation , this retention length threshold can be set as a static number or as a fraction of the total run-time , i . e . , ( 1/60 ) of the gradient length . For our data , only peptide sequences present in 3 or more experiments were allowed to participate in the alignment process . The model can allow peptides only present in one experiment to be included in the alignment , but the inclusion of this data adds no additional information to the alignment and only serves to slow it down computationally . The definition of a peptide sequence in these cases is dynamic , and can include modifications , charge states , or any other feature that would affect the retention of an isoform of that peptide . For our data , we used the peptide sequence with modifications but did not append the charge state . Preliminary alignments revealed certain experiments where chromatography was extremely abnormal , or where peptide identifications were too sparse to enable an effective alignment . These experiments were manually removed from the alignment procedure after a preliminary run of DART-ID . From the original 263 experiments , 37 had all of their PSMs pruned , leaving only 226 experiments containing PSMs with updated confidences . These experiments are included in the DART-ID output but do not receive any updated error probabilities as they did not participate in the RT alignment . All filtering parameters are publicly available as part of the configuration file that was used to generate the data used in this paper . Let ρik be the RT assigned to peptide i in experiment k . In order to infer peptide and experiment-specific RT distributions , we assume that there exists a set of reference retention times , μi , for all peptides i . Each peptide has a unique reference RT , independent of experiment . We posit that for each experiment , there is a simple monotone increasing function , gk , that maps the reference RT to the predicted RT for peptide i in experiment k . An observed RT can then be expressed as in Eq 1 . As a first approximation , we assume that the observed RTs for any experiment can be well approximated using a two-segment linear regression model as described by Eq 2 . This model can be extended to more complex monotonic models , such as spline fitting , or non-linear monotonic models , such as a logit function or LOESS . To factor in the spectral PEP given by the search engine , and to allow for the inclusion of low probability PSMs , the marginal likelihood of an RT in the alignment process can be described using a mixture model as described in S1 Fig . For a PSM assigned to peptide i in experiment k the RT density is P ( ρ i k | μ i k , σ i k , λ i k ) ︸ Likelihood ∝ 1 { ρ i k > 0 } ( ( 1 - λ i k ) × f i k ( ρ i k ∣ μ i k , σ i k ) ︸ PSM is correct + ( λ i k ) × f k 0 ( ρ i k ) ︸ PSM is incorrect ) ( 5 ) where λik is the error probability ( PEP ) for the PSM returned by MaxQuant , fik is the inferred RT density for peptide i in experiment k and f k 0 is the null RT density . In our implementation , we let: f i k ∼ Laplace ( μ i k , σ i k 2 ) f k 0 ∼ Normal ( μ k , σ k 2 ) ( 6 ) which we found worked well in practice ( See S4 Fig ) . However , our framework is modular and it is straightforward to utilize different residual RT and null distributions if appropriate . For example , with non-linear gradients that generate a more uniform distribution of peptides across the LC run [22] , it may be sensible for the null distribution to be defined as uniformly distributed , i . e . f k 0 ∼ Uniform ( RT m i n , RT m a x ) . Finally , to reflect the fact that residual RT variation increases with mean RT and varies between experiments ( S3 Fig ) , we model the standard deviation of a peptide RT distribution , σik , as a linear function of the reference RT: σ i k = a k + b k μ i ( 7 ) where μi is the reference RT of the peptide sequence , and ak and bk are the intercept and slope which we infer for each experiment . ak , bk and μi are constrained to be positive , and hence σik > 0 as well . To estimate all unknown parameters , we consider the joint posterior distribution of the experiment specific alignment parameters and the reference RTs given the observed retention times , P ( a , b , β 0 , β 1 , s , μ ∣ ρ , λ ) ︸ Posterior ∝ P ( ρ ∣ a , b , β 0 , β 1 , s , μ , λ ) ︸ Likelihood Eq 5 P ( a , b , β 0 , β 1 , s , μ ) ︸ Prior ( 8 ) where P ( a , b , β0 , β1 , s , μ ) are the prior distributions for all unknown alignment parameters and reference RTs and P ( ρ | a , b , β0 , β1 , s , μ ) is the likelihood , as determined by Equation . a , b , β0 , β1 , s are all K-vectors of alignment parameters for each experiment . μ consists of the reference RTs for every peptide . The priors for the Bayesian inference can be found in the . stan model files , and for the analyses in this paper , are as follows: where RTmean and RTsd are the mean and standard deviation of all RTs across all experiments , respectively . max ( RT ) is the maximum observed RT of all RTs across all experiments . These priors were chosen for groups of 60 min LC-MS/MS runs , and can be adjusted accordingly for different run lengths , gradient shapes , and groupings of runs with different run times . We compared the DART-ID alignment accuracy against five other RT prediction or alignment algorithms . As some methods returned absolute predicted RTs ( such as BioLCCC [31] ) and others returned relative hydrophobicity indices ( such as SSRCalc [30] ) , a linear regression was built for each prediction method . Alignment accuracy was evaluated using three metrics: R2 , the Pearson correlation squared , and the mean and median of |ΔRT| , the absolute value of the residual RT , and is defined as |Observed RT − Predicted RT| . We selected only confident PSMs ( PEP < 0 . 01 ) for this analysis , and used data that consisted of 33383 PSMs from 46 LC-MS/MS experiments run over the course of 90 days in order to produce more chromatographic variation . A list of these experiments is found in S1 Table . SSRCalc [30] was run from SSRCalc Online ( http://hs2 . proteome . ca/SSRCalc/SSRCalcQ . html ) , with the “100Å C18 column , 0 . 1% Formic Acid 2015” model , “TMT” modification , and “Free Cysteine” selected . No observed RTs were inputted along with the sequences . BioLCCC [31] was run online from http://www . theorchromo . ru/ with the parameters of 250mm column length , 0 . 075mm column inner diameter , 130Å packing material pore size , 5% initial concentration of component B , 35% final concentration of component B , 48 min gradient time , 0 min delay time , 0 . 0001 ml/min flow rate , 0% acetonitrile concentration in component A , 80% acetontrile concentration in component B , “RP/ACN+FA” solid/mobile phase combination , and no cysteine carboxyaminomethylation . As BioLCCC could only take in one gradient slope as the input , all peptides with observed RT > 48 min were not inputted into the prediction method . ELUDE [34] was downloaded from the percolator releases page https://github . com/percolator/percolator/releases , version 3 . 02 . 0 , Build Date 2018-02-02 . The data were split into two , equal sets with distinct peptide sequences to form the training and test sets . The elude program was run with the --no-in-source and --test-rt flags . Predicted RTs from ELUDE were obtained from the testing set only , and training set RTs were not used in further analyses . For iRT [43] , the same raw files used for the previous sets were searched with the Pulsar search engine [61] , with iRT alignment turned on and filtering at 1% FDR . From the Pulsar search results , only peptide sequences in common with the previous set searched in MaxQuant were selected . Predicted RT was taken from the “PP . RTPredicted” column and plotted against the empirical RT column “PP . EmpiricalRT” . Empirical RTs were not compared between those derived from MaxQuant and those derived from Pulsar . MaxQuant match-between-runs [7 , 8] was run by turning the respective option on when searching over the set of 46 experiments , and given the options of 0 . 7 min match time tolerance and a 20 min match time window . The “Calibrated retention time” column was used as the predicted RT , and these predicted RTs were related to observed RTs with a linear model for each experiment run . For DART-ID , predicted RTs are the same as the mean of the inferred RT distribution , and no linear model was constructed to relate the predicted RTs to the observed RTs . To compare the performance of the two-piece linear model for RT alignment against a simple linear model , we ran both alignments separately on the same dataset as described in the RT alignment comparison section . For S2 Fig , we used one experiment—180324S_QC_SQC69A—as an example to illustrate the qualitative differences between the two models . Panels b and c used all experiments from the set to give a more quantitative comparison . We update the confidence for PSM i in experiment k according to Bayes’ theorem . Let δik = 1 denote that PSM i in experiment k is assigned to the correct sequence ( true positive ) , δik = 0 denotes that the PSM is assigned to the incorrect sequence ( a false positive ) , and as above , ρik is an observed RT assigned to peptide i . At a high level , the probability that the peptide assignment is a true positive is P ( δ i k = 1 ∣ ρ i k ) = P ( ρ i k ∣ δ i k = 1 ) × P ( δ i k = 1 ) P ( ρ i k ) ( 9 ) Each term is described in more detail below: The confidence update depends on the global alignment parameters . Let θ consist of the global alignment parameters and reference RTs , i . e . β0k , β1k , σik and μi . If θ were known , then the Bayesian update could be computed in a straightforward manner as described above . In practice the alignment parameters are not known and thus must be estimated using the full set of observed RTs across all experiments , ρ . The PSM confidence update can be expressed unconditional on θ , by integrating over the uncertainty in the estimates of the alignment parameters: ( δ i k = 1 ∣ ρ ) = ∫ p ( δ i k = 1 ∣ ρ i k , θ ) p ( θ ∣ ρ ) d θ i k ( 10 ) Although we can estimate this posterior distribution using Markov Chain Monte Carlo ( MCMC ) , it is prohibitively slow given the large number of peptides and experiments that we analyze . As such , we estimate maximum a posteriori ( MAP ) estimates for the reference RTs μi , alignment parameters β0k , β1k , and RT standard deviation σik using an optimization routine implemented in STAN [62] . If computation time is not a concern , it is straightforward to generate posterior samples in our model by running MCMC sampling in STAN , instead of MAP optimization . This approach is computationally efficient but is limited in that parameter uncertainty quantification is not automatic . To address this challenge , we incorporate estimation uncertainty using a computationally efficient procedure based on the parametric bootstrap . Note that uncertainty about the alignment parameters β0k and β1k is small since they are inferred using thousands of RT observations per experiment . By contrast , the reference RTs , μi , have much higher uncertainty since we observe at most one RT associated with peptide i in each experiment ( usually far fewer ) . As such , we choose to ignore uncertainty in the alignment parameters and focus on incorporating uncertainty in estimates of μi . Let μ ^ i k and σ ^ i k denote the MAP estimates of the location and scale parameters for the RT densities . To approximate the posterior uncertainty in the estimates of μi , we use the parametric bootstrap . First , we sample ρ i k ( b ) from f i k ( ρ i k ∣ μ ^ i k , σ ^ i k ) with probability 1 − λik and f k 0 ( ρ i k ) with probability λik . We then map ρ i k ( b ) back to the reference space using the inferred alignment parameters as g ^ - 1 ( ρ i k ) and compute a bootstrap replicate of the reference RT associated with peptide i as the median ( across experiments ) of the resampled RTs: μ i ( b ) = median k g ^ - 1 ( ρ i k ( b ) ) , as the maximum likelihood estimate of the location parameter of a Laplace distribution is the median of independent observations . For each peptide we repeat this process B times to get several bootstrap replicates of the reference RT for each peptide . We use the bootstrap replicates to incorporate the uncertainty of the reference RTs into the Bayesian update of the PSM confidence . Specifically , we approximate the confidence update in Eq 10 as p ( δ i k = 1 ∣ ρ i k ) ≈ 1 B ∑ b = 1 B p ( δ i k = 1 ∣ ρ i k , μ i k ( b ) , σ ^ i k ) = ( 1 - λ i k ) ( 1 B ∑ b = 1 B f i k ( ρ i k ∣ μ i k ( b ) , σ ^ i k ) ( 1 - λ i k ) ( 1 B ∑ b = 1 B f i k ( ρ i k ∣ μ i k ( b ) , σ ^ i k ) + λ i k f k 0 ( ρ i k ) ( 11 ) This process is depicted in S8 Fig . In addition to updating the PEPs for each PSM , DART-ID also recalculates the set-wide false discovery rate ( FDR , q-value ) . This is done by first sorting the PEPs and then assigning the q-value to be the cumulative sum of PEPs at that index , divided by the index itself , to give the fractional expected number of false positives at that index ( i . e . , the mean PEP ) [56] . Reporter ion ( RI ) intensities were obtained by selecting the tandem-mass-tag ( TMT ) 11-plex labels in MaxQuant , for both attachment possibilities of lysine and the peptide N-terminus , and with a mass tolerance of 0 . 003 Da . Data from different experiments and searches are all combined into one matrix , where the rows are observations ( PSMs ) and the 10 columns are the 10 TMT channels . Observations are filtered at a confidence threshold , normally 1% FDR , and observations with missing data are thrown out . Before normalization , empty channels 127N , 128C , and 131C are removed from the matrix . Each column of the matrix is divided by the median of that column , to correct for the total amount of protein in each channel , pipetting error , and any biases between the respective TMT tags . Then , each row of the matrix is divided by the median of that row , to obtain the relative enrichment between the samples in the different TMT channels . In our data the relative enrichment was between the two cell types present in our SCoPE-MS sets , T-cells ( Jurkat cell line ) and monocytes ( U-937 cell lines ) . Assuming that the relative RI intensities of PSMs are representative of their parent peptide , the peptide intensity can be estimated as the median of the RI intensities of its constituent PSMs . Similarly , if protein levels are assumed to correspond to the levels of its constituent peptides , then protein intensity can be estimated as the median of the intensities of its constituent peptides . The previous steps of RI normalization makes all peptide and protein-level quantitation relative between the conditions in each channel . For the principal component analysis as shown in Fig 8a , data was filtered and normalized in the same manner as discussed previously . Additional experiments were manually removed from the set due to different experimental designs or poorer overall coverage that would have required additional imputation on that experiment’s inclusion . PSMs were separated into two sets , as described in Fig 7a: Spectra and DART-ID . PSMs in the DART-ID set belonging to any parent protein in the Spectra set were filtered out , so that the two PSM sets contained no shared proteins . Additionally , proteins that were not observed in at least 95% of the selected experiments were removed in order to reduce the amount of imputation required . Normalized TMT quantification data was first collapsed from PSM-level to peptide-level by averaging ( mean ) PSM measurements for the same peptide . This process was repeated to estimate protein-level quantitation from peptide-level quantitation . This data , from both sets , was then reshaped into an expression matrix , with proteins on the rows and “single cells” ( TMT channel- experiment pairs ) on the columns . As described earlier in the Results section , these samples are not actual single cells but are instead comprised of cell lysate at the expected abundance of a single cell; see Table 1 . Missing values in this expression matrix were imputed with the k-nearest-neighbors ( kNN ) algorithm , with Euclidean distance as the similarity measure and k set to 5 . A similarity matrix was then derived from this expression matrix by correlating ( Pearson correlation ) the matrix with itself . Singular value decomposition ( SVD ) was then performed on the similarity matrix to obtain the principal component loadings . These loadings are the left singular vectors ( the columns of U of SVD: UDUT ) . Each circle was then colored based on the type of the corresponding cell from annotations of the experimental designs . Our raw data was searched with both the PSM and protein FDR threshold set , in the search engine , to 100% to include as many PSMs as possible . Therefore , once PSM confidences were updated with RT evidence , we needed to propagate those new confidences to the protein level in order to avoid any spurious protein identifications from degenerate peptide sequences [63] . This is especially pertinent as many of the new DART-ID PSMs support proteins with no other confidently identified peptides , S9 Fig . Ideally we would run our updated PSMs back through our original search engine pipeline ( MaxQuant/Andromeda ) [7 , 21] , but that is currently not possible due to technical restrictions . Any interpretation of the DART-ID data on the protein-level was first run through the Fido protein inference algorithm [64] , which gives the probability of the presence of a protein in a sample given the pool of observed peptides and the probabilities of their constituent PSMs . The Python port of Fido was downloaded from https://noble . gs . washington . edu/proj/fido and modified to be compatible with Python 3 . The code was directly interfaced into DART-ID and is available to run as a user option . For the data in this paper , protein-level analyses first had their proteins filtered at 1% FDR , where the FDR was derived from the probabilities given to each protein by the Fido algorithm . We ran Fido with the default parameters gamma: 0 . 5 , alpha: 0 . 1 , beta: 0 . 01 , connected protein threshold: 14 , protein grouping and using all PSMs set to false , and pruning low scores set to true . In Fig 4 we evaluated DART-ID on two other third-party , publicly available datasets: iPRG 2015 [57] ( MassIVE ID: MSV000079843 ) , 12 label-free runs of yeast lysate , and TKO 2018 [58] ( ProteomeXchange ID: PXD011654 ) , 40 TMT-labelled runs of yeast lysate . Raw files were searched in MaxQuant 1 . 6 . 3 . 4 , against a UniProt yeast database ( 6721 entries , 2018/05/01 ) . The iPRG 2015 dataset was searched with cysteine carbamidomethylation ( +57 . 02146 Da ) as a fixed modification and methionine oxidation ( +15 . 99492 Da ) , protein N-terminal acetylation ( +42 . 01056 Da ) , and asparagine/aspartate deamidation ( +0 . 98401 Da ) as variable modifications . The TKO 2018 dataset was searched with TMT11-plex on lysine/n-terminus , cysteine carbamidomethylation ( +57 . 02146 Da ) as a fixed modification and methionine oxidation ( +15 . 99492 Da ) as a variable modification . Both searches were done with Trypsin specificity , and PSM/protein confidence thresholds were set at 1 ( 100% ) to obtain as many PSMs as possible . Searched data , configuration files , and DART-ID analysis results are available online . The DART-ID pipeline is roughly divided into three parts . First , input data from search engine output files are converted to a common format , and PSMs unsuitable for alignment are marked for removal . Second , we estimate the alignment parameters and reference RTs using an by finding the maximum of the posterior distribution ( Eq 4 ) . Initial values for the algorithm are are generated by running a simple estimation of reference RTs and linear regression parameters for fik for each experiment . Third , inferred alignment parameters and reference RTs are used to update the confidence for the PEP of a PSM . The model was implemented using the STAN modeling language [62] . All densities were represented on the log scale . STAN was interfaced into an R script with rstan . STAN was used with its optimizing function , which gave maximum a posteriori ( MAP ) estimates of the parameters , as opposed to sampling from the full posterior . R was further used for data filtering , PEP updating , model adjustment , and figure creation . The code is also ported to Python3 and pystan , and is available as a pip package dart_id that can be run from the command-line . DART-ID is run with a configuration file that specifies inputs and options . All model definitions and related parameters such as distributions are defined in a modular fashion , which supports the addition of other models or fits . Full instructions for using the Python program are available at https://dart-id . slavovlab . net . Code for analysis and figure generation is available at: github . com/SlavovLab/DART-ID_2018 . The python program for DART-ID , as well as instructions for usage and examples , are available on GitHub as a separate repository: https://github . com/SlavovLab/DART-ID . All raw files , searched data , configuration files , and analyzed data are publicly available and deposited on MassIVE ( ID: MSV000083149 ) and ProteomeXchange ( ID: PXD011748 ) .
|
Identifying and quantifying proteins in single cells gives researchers the ability to tackle complex biological problems that involve single cell heterogeneity , such as the treatment of solid tumors . Mass spectrometry analysis of peptides can identify their sequence from their masses and the masses of their fragment ions , but often times these pieces of evidence are insufficient for a confident peptide identification . This problem is exacerbated when analyzing lowly abundant samples such as single cells . To identify even peptides with weak mass spectra , DART-ID incorporates their retention time—the time when they elute from the liquid chromatography used to physically separate them . We present both a novel method of aligning the retention times of peptides across experiments , as well as a rigorous framework for using the estimated retention times to enhance peptide sequence identification . Incorporating the retention time as additional evidence leads to a substantial increase in the number of samples in which proteins are confidently identified and quantified .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immunology",
"experimental",
"design",
"database",
"searching",
"research",
"design",
"probability",
"distribution",
"mathematics",
"research",
"and",
"analysis",
"methods",
"information",
"retrieval",
"sequence",
"analysis",
"white",
"blood",
"cells",
"sequence",
"alignment",
"bioinformatics",
"animal",
"cells",
"proteins",
"t",
"cells",
"probability",
"theory",
"biochemistry",
"cell",
"biology",
"monocytes",
"proteomes",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences"
] |
2019
|
DART-ID increases single-cell proteome coverage
|
Ten years on from the finishing of the human reference genome sequence , it remains unclear what fraction of the human genome confers function , where this sequence resides , and how much is shared with other mammalian species . When addressing these questions , functional sequence has often been equated with pan-mammalian conserved sequence . However , functional elements that are short-lived , including those contributing to species-specific biology , will not leave a footprint of long-lasting negative selection . Here , we address these issues by identifying and characterising sequence that has been constrained with respect to insertions and deletions for pairs of eutherian genomes over a range of divergences . Within noncoding sequence , we find increasing amounts of mutually constrained sequence as species pairs become more closely related , indicating that noncoding constrained sequence turns over rapidly . We estimate that half of present-day noncoding constrained sequence has been gained or lost in approximately the last 130 million years ( half-life in units of divergence time , d1/2 = 0 . 25–0 . 31 ) . While enriched with ENCODE biochemical annotations , much of the short-lived constrained sequences we identify are not detected by models optimized for wider pan-mammalian conservation . Constrained DNase 1 hypersensitivity sites , promoters and untranslated regions have been more evolutionarily stable than long noncoding RNA loci which have turned over especially rapidly . By contrast , protein coding sequence has been highly stable , with an estimated half-life of over a billion years ( d1/2 = 2 . 1–5 . 0 ) . From extrapolations we estimate that 8 . 2% ( 7 . 1–9 . 2% ) of the human genome is presently subject to negative selection and thus is likely to be functional , while only 2 . 2% has maintained constraint in both human and mouse since these species diverged . These results reveal that the evolutionary history of the human genome has been highly dynamic , particularly for its noncoding yet biologically functional fraction .
“What proportion of the human genome is functional ? ” remains a contentious question [1]–[3] . In great part this reflects the use of definitions of ‘function’ that differ from the traditional definition that is based on fitness and selection ( see e . g . [4] for a discussion ) . For instance , equating functionality with annotation by at least one of the ENCODE consortium's biochemical assays [5] results in approximately 80% of the human genome being labeled as functional [1] , [6] . While this approach has the advantage of being empirical , it makes the definition of functionality dependent on the choice of experiments and details such as P value cutoffs . It is also questionable whether , for instance , introns should be classified as functional based merely on their transcription [2] , [4] . By contrast , evolutionary studies often equate functionality with signatures of selection . While it is undisputed that many functional regions have evolved under complex selective regimes including selective sweeps [7] or ongoing balancing selection [8] , [9] , and it appears likely that loci exist where recent positive selection or reduction of constraint has decoupled deep evolutionary patterns from present functional status [10] , [11] , it is widely accepted that purifying selection persisting over long evolutionary times is a ubiquitous mode of evolution [12] , [13] . While acknowledging the caveats , this justifies the definition of functional nucleotides used here , as those that are presently subject to purifying selection . This is of course not useful as an operational definition , as selection cannot be measured instantaneously . Instead , most studies define functional sites as those subject to purifying selection between two ( or more ) particular species . Studies that follow this definition have estimated the proportion of functional nucleotides in the human genome , denoted as αsel [14] , [15] , between 3% and 15% ( [3] and references therein , [16] ) . Since each species' lineage gains and loses functional elements over time , αsel needs to be understood in the context of divergence between species . The divergence influences the estimate of αsel in two ways . On the one hand , constrained sequence between closely related species , including lineage-specific constrained sequence , is harder to detect than more broadly conserved sequence because of a paucity of informative mutations , which reduces detection power . On the other hand , estimates of constraint between any two species will only include sequence that was present in their common ancestor and that has been constrained in the lineages leading up to both extant species' genomes , with the consequence that turnover of functional sequence leads to diminishing αsel estimates as the species divergence increases . Assuming that the first effect can be controlled for , higher estimates of sequence constraint that are obtained between more closely related species [15] , [17] are thus indicative of the turnover of functional sequence [15] . Here we understand turnover to mean the loss or gain of purifying selection at a particular locus of the genome , when changes in the physical or genetic environment , or mutations at the locus itself , cause the locus to switch from being functional to being non-functional or vice versa . Two previous studies have made quantitative estimates of the overall rate of turnover ( [15] , [17] , reviewed in [3] ) . The estimate by Smith et al . ( 2004 ) [17] was derived from an analysis of point mutations in alignments across a 1 . 8 Mb genomic region . While a high rate of turnover was inferred , the authors emphasised the preliminary nature of their work as a consequence of the limited amount of data available to them at that time . Later , Meader et al . ( 2010 ) [15] performed genome-wide analysis with a neutral indel model ( see [18] , here referred to as NIM1 ) to estimate the fraction , termed αselIndel , of human sequence that was constrained with respect to insertions or deletion mutations ( indels ) . This study also found a high rate of turnover , and estimated using two ad hoc heuristic approaches that 6 . 5–10% of the human genome is functional . Extrapolations using these data subsequently suggested that 10–15% of the human genome is presently functional [3] . NIM1 is a quantitative model describing the distribution of distances between neighbouring indels ( intergap segments; IGSs ) in neutrally evolving sequence , which provides an excellent description of the observed frequency of medium-sized IGSs . However , across whole genome alignments longer IGSs are strikingly overrepresented compared to this expectation under neutrality , presumably as a result of the presence of functional genomic segments under purifying selection in which indel mutations are unlikely to become fixed . By quantifying this overrepresentation it is possible to estimate αselIndel , the fraction of nucleotides contained within these functional segments . The model ( which also accounts for G+C content and sex chromosome-dependent mutational biases ) performs well for simulated data , and accurately identifies coding regions and ancestral repeats as highly conserved and neutrally evolving , respectively [15] , [18] . However , some concerns about the model's derivation and the quality of whole-genome alignments we used were subsequently brought to our attention , which motivated us to initiate this study . Here we present improved methods for the estimation of αselIndel and the inference of functional turnover , building on our previous approaches [15] , [18] . We apply these improved approaches to pairwise alignments between the genomes of diverse eutherian mammals , and we estimate that 7 . 1–9 . 2% of the human genome is presently subject to purifying selection , equating to 220–286 Mb of constrained sequence . We also take advantage of the additional high-quality eutherian genome sequences that have become available since our previous study to provide improved estimates of the rate of turnover of functional sequence in these species . Improvements in biological and biochemical annotation of genomic sequence mean that we can investigate turnover rates within particular classes of functional elements , such as coding sequences , DNase 1 hypersensitivity sites ( DNase HSs ) , transcription factor binding sites ( TFBSs ) , enhancers , promoters , and long noncoding RNAs ( lncRNAs ) . We find striking differences between the functional element classes; in particular constrained coding sequences are much more evolutionary stable than constrained noncoding sequences , and lncRNAs show the most rapid rate of turnover of all the noncoding element types .
We observe a strong negative correlation between estimates of αselIndel and the divergence of the two species being compared ( Figure 1 ) , consistent with substantial turnover of functional sequence and thus with earlier conclusions [15] , [17] , and inconsistent with simulation results under a scenario in which turnover is absent ( Figure 1A ) . To exclude the possibility that technical artefacts are driving this observation , we investigated ENCODE annotations in lineage-specific NIM1-constrained sequence . Specifically , we identified NIM1-constrained sequence that was not identified as pan-mammalian conserved by either the PhastCons [12] or GERP++ algorithms [19] , and found that such sequence is enriched for biochemically annotated sequences ( DNase HSs , TFBSs , and enhancers defined by the ENCODE consortium [5] ) ( Figure 2; Figure S4 ) . This is expected if functional elements , including these ENCODE functional classes , have been subject to evolutionary turnover , but is not expected if technical artefacts were causing the observations in Figure 1 . Furthermore , using low-frequency polymorphic indels from the 1000 Genomes project we could exclude the possibility that lower mutation rates in ENCODE functional regions were causing the observations . We therefore conclude that observations in Figure 1 reflect turnover of functional elements . A more detailed discussion on this issue is provided in Text S6 and Text S7 . To help describe and interpret the observations of turnover ( Figure 1 ) we propose a time-homogeneous model for sequence turnover on a genomic scale . We apply this model to specific sequence classes , such as protein coding genes or TFBSs , allowing us to discuss the rates of turnover for particular types of functional element . The model assumes that within a particular functional class both the total amount a of functional sequence and the rate b of turnover per nucleotide ( nt ) are constant , and that the turnover rate is the same for all nts in a class . Under this model the total amount of functional nts in any class remains constant over time , but the amount that is currently functional and retains homology to functional nts in the ancestral species at divergence d ( i . e . , the amount that was constrained and has not turned over in the course of evolution to the present ) is . We estimate the parameters a and b by fitting the model to observations using weighted linear regression ( Materials and Methods ) . Instead of the rate parameter b , we , equivalently , often refer to the turnover half life , d1/2 , which is defined as the divergence at which half the functional sequences in the class is expected to have turned over and is calculated as loge ( 2 ) /b . We express this divergence in time units corresponding to one expected nucleotide substitution per site in neutrally evolving sequence ( ‘divergence time’ ) . To convert this divergence to years , we apply a substitution rate of 2 . 2×10−9 per site per year [20] . This will be a more appropriate value for the human lineage , on which we focus , than on rodent lineages whose per-year substitution rate are substantially higher . The model is time-reversible , so that the same expression describes the amount of mutually constrained sequence between two extant species at divergence d , where d is calculated by adding the divergences along the two branches to their last common ancestor . Similarly , to convert d ( in years ) to the age of the most recent common ancestor , it should be divided by 2 . To calculate the divergence time we use ancestral repeats ( ARs , sequence derived from transposable elements whose insertion predates the species' last common ancestor ) as a proxy for neutrally evolving sequence , because they virtually all show the patterns of indel mutation expected under neutral evolution [18] . Our estimates of divergence using either ARs or synonymous sites as neutral proxy are concordant , hence our results are insensitive to the choice of putatively neutral sequence ( Figure S5 ) . We next used NIM1 to estimate the fraction of constrained sequence within coding and noncoding sequences ( Materials and Methods ) . Within protein coding sequence selective constraint is pervasive , as expected ( Figure 1B ) : 80–88% of human or mouse annotated coding sequence has been under selective constraint with respect to indels across eutherian evolution; slightly lower proportions were estimated under the NIM2 and for dog annotated coding sequences ( Figure S6; Text S8 ) . In contrast to protein coding sequence , estimates for the extent of constraint in noncoding sequence show a pronounced drop-off with increasing divergence ( orange filled circles in Figure 1B ) , an observation compatible with turnover occurring predominantly within the noncoding functional fraction of the genome . When applying the time-homogeneous turnover model to these data , we estimate the turnover rate parameter b for noncoding sequence at 2 . 48 turnover events per neutral substitution ( 2 . 26–2 . 71 , 95% confidence interval ) , equivalent to a turnover half life d1/2 of 0 . 28 ( 0 . 25–0 . 31 ) in units of divergence time , or 127 My ( 116–139 My ) in natural time units . The present estimate represents a slower turnover rate than a previous estimate of d1/2 = 0 . 19 ( 86 My ) made by Ponting et al . ( 2011 ) [3] with data from Meader et al . ( 2010 ) [15] . We observe a low yet significantly non-zero rate of turnover in coding sequence , b = 0 . 24 ( 0 . 14–0 . 33 ) events per neutral substitution , corresponding to d1/2 = 2 . 9 ( 2 . 1–5 . 0 ) , or in natural units 1300 My ( 950–2250 My ) . These estimates represent an average across the undoubtedly variable rates of turnover across different types of protein coding gene sequence . Nevertheless , under this simple model , we find that protein coding sequence is relatively evolutionarily stable , showing long-term conservation , so that assuming that protein coding sequences exhibit no turnover will often be justified ( e . g . [3] ) . By contrast , present-day constrained noncoding sequence is less stable , being relatively rapidly gained and lost in a lineage-specific manner . We next investigated whether various classes of functional element , identified in human primarily by the ENCODE project [5] , exhibit contrasting levels of constraint , and whether these constrained element classes show a propensity to turn over at different rates . Of the functional classes we considered , promoters , untranslated regions ( UTRs ) , DNAse HSs and TFBSs , enhancers and un-annotated sequences ( defined as sequences not within 50 bp of ENCODE DNAse HSs , TFBS loci , lncRNAs from [21] , Ensembl coding sequence , or UTRs ) all show intermediate levels of turnover ( Figure 3; Figure S7 , Figure S8 ) . LncRNA sequences show the highest level of turnover ( Figure 3; Figure S8 ) , and an even higher rate of turnover was inferred when the ENCODE-defined lncRNAs were used rather than the set from [21] ( Figure S9 ) . The fraction of sequence that the model inferred to be under present day constraint also varied across these categories , with intermediate fractions inferred for UTRs , DNAse HSs and TFBSs , and lower fractions for lncRNAs and enhancers . As expected , the lowest fractions were observed for un-annotated sequence; nevertheless , in absolute terms the amount of constrained sequence in this category is considerable ( 70 Mb , 45–85 Mb ) ( Figure 3 ) . Constrained sequence in this category may represent lineage-specific functional sequences that were not identified by the ENCODE project , for instance because of their function in tissues or developmental stages not investigated by ENCODE . Finally , transposable element-derived sequences show very small amounts of constraint , and as a result our methods have little power to detect turnover in this class . We next examined how constrained sequence in the human genome is distributed cumulatively for selected functional element categories . We do this by fitting the functional turnover model to the observed data and extrapolating to the present day . In this way we also infer the reciprocal quantities of sequence that , when comparing to another species or human ancestor at a particular divergence , are presently functional in human yet have lost ( or not gained ) constraint in the lineage leading to the ancestor or other species ( Figure 4 ) . We stress that this inference relies on the parsimonious yet not formally justified assumption that the total quantity of functional sequence in genomes remains constant over time and therefore across species , and within functional categories . With these caveats we estimate that 8 . 6 Mb ( 26% ) of constrained coding sequence has lost constraint ( and thus has turned over ) since the divergence of humans from monotremes approximately 228 million year ago ( AR divergence time 1 . 00 ) , while 200 Mb ( 79% ) of the constrained noncoding human genome is inferred to have lost constraint over the same period . DNAse HSs cover more indel constrained sequence at all divergence ranges than all other annotated noncoding sequence combined , implying that DNAse HSs are an abundant and informative biochemical marker of functionality outside protein coding regions . Enhancers also show a marked contribution towards the constrained human genome , while TFBSs , promoters , UTRs and lncRNAs contribute considerably less sequence once their overlap with other annotations is removed . Finally , about a quarter of sequence inferred to be presently under constraint is not present in any of the annotation categories we considered . In Figure 4 we sum up the quantities of constrained sequence estimated from independent NIM1 runs for different annotation types . If we make the assumption that the exponential decay model of functional sequence applies outside of the range of divergences we examined , then by extrapolating back to zero divergence we can estimate the total proportion of human genomes that is under present-day purifying selection with respect to indels . We perform this extrapolation across different annotation sets ( Table S6 ) . Although there is some variation in these estimates , we quote the estimate derived separately across multiple different annotation categories , namely coding sequence , DNase HSs , TFBS , Enhancers , unannotated sequence , and other sequence ( the latter consisting of promoter , UTR and lncRNA sequences ) . This is because this estimate allows the rate of turnover to vary across each annotation type , and thus is likely to be more accurate than the estimates that assume a single rate of turnover across the whole genome , or the whole noncoding genome . We therefore estimate that 8 . 2% of the human genome ( 253 Mb; 95% CI 7 . 1%–9 . 2% , 220–286 Mb ) is presently under purifying selection with respect to indels .
The question of what fraction of the human genome sequence are mutations preferentially purged owing to their deleterious effect has remained contentious ever since the first estimate was made in 2002 [22] . At that time it was not well appreciated that the amount of human constrained sequence that is also constrained in mouse is a minority ( 69 Mb; this study ) of all human constrained sequence , owing to the relatively rapid gain and loss of functional sequence in their two lineages since their last common ancestor . We find that NIM1-constrained sequence lacking evidence for pan-mammalian conservation is enriched for sequences with experimental evidence for biochemical activity , and we provide a detailed argument indicating that this is incompatible with the notion of technical artefacts causing the observed signature of turnover ( Text S6 ) . Extensive simulations indicating that estimates of constrained sequence are consistent across the divergence range we investigate further support this conclusion . Our estimate that 7 . 1–9 . 2% of human genomes is subject to contemporaneous selective constraint considerably exceeds previous estimates and falls short of others [3] , [23] . We have shown that our method's previous estimates for specific species pairs , as well as the calculation that suggested 10–15% of the human genome is currently under negative selection were inflated [3] , in large part owing to inaccuracies in whole genome alignments upon which our estimates were based . The problems associated with using whole-genome alignments could be circumvented entirely by instead using polymorphism data within a single species . However , this approach is technically highly challenging , and results have so far been controversial [16] , [24] , [25]; in addition this approach is not informative about functional turnover . Other published estimates [12] , [18] , [26] are lower because they , by design , were not sensitive to lineage-specific constrained sequence . Our current estimates have their own particular caveats . While our results show that turnover is a real and substantial effect , simulations show that NIM1 underestimates the true amount of mutually constrained sequence to an extent that shows some dependence on the divergence . While simulations and theory indicate that point estimates of constraint remain conservative , the possibility of an upward bias in the inferred rate of turnover cannot be excluded , which in turn could lead to upwardly biased extrapolations of present-day constraint . In addition , the assumptions of the turnover model , in particular that all elements within a class are subject to the same rate of turnover , clearly are only approximately valid . These potential sources of error are not reflected in our confidence estimates ( Table S6 ) . Our estimate that 7 . 1%–9 . 2% of the human genome is functional is around ten-fold lower than the quantity of sequence covered by the ENCODE defined elements [1] , [5] , [6] . This indicates that a large fraction of the sequence comprised by elements identified by ENCODE as having biochemical activity can be deleted without impacting on fitness . By contrast , the fraction of the human genome that is covered by coding exons , bound motifs and DNase1 footprints , all elements that are likely to contain a high fraction of nucleotides under selection , is 9% . While not all of the elements in these categories will be functional , and functional elements will exist outside of these categories , this figure is consistent with the proportion of sequence we estimate as being currently under the influence of selection . As expected , turnover has occurred least in protein coding sequence , and thus has been most concentrated on noncoding sequence ( Figure 4 ) . For example , of the 43 . 5 Mb of sequence annotated by the ENCODE project as being within a human TFBS peak and that we find to be constrained ( 19 . 3% of the total extent of ENCODE TFBS peaks ) , only a third ( 30 . 6%; 13 . 3 Mb ) is identified by NIM1 as being constrained in both human and mouse . A slightly higher proportion ( 45 . 6%; 19 . 8 Mb ) is constrained in human and dog , presumably reflecting these species' lower divergence . These estimates are in good agreement with previous experimental findings: for instance 23–41% of TF binding events have been found to be conserved across human , dog and mouse for four liver TFs [27] , while for two additional liver TFs , 7–14% of TF binding events are shared between human and mouse , and 15–20% between human and dog [28] . The phenomenon of turnover is well supported by both anecdotal evidence [27]–[29] and by broader studies of particular classes of elements , mostly TFBSs and enhancer elements [30]–[32] . The class of functional element inferred to turnover fastest was that of lncRNAs , again consistent with observations that most human lncRNAs are primate-specific and only 19% of lncRNAs are conserved over more than 90 My [33] . What our approach cannot clarify is to what extent the observed turnover at the sequence level amounts to different sequences encoding equivalent function [29] , [30] , or species-specific functional change [16] , [31] , [34] . Several lines of evidence , both from anecdotal [29] and broader [30] , [31] studies of TFBSs , indicate that a large fraction of sequence changes involving TFBSs preserve function . For example , some deeply conserved transcription factors have species-specific binding sites in the vicinity of orthologous genes [27] , [28] implying that despite their sequence divergence , the different DNA binding sites confer equivalent functions ( on orthologous genes ) in different lineages . Comprehensive studies of human and mouse embryonic heart enhancers found these to be weakly conserved [35] , [36] , despite human enhancers sequences largely driving expected tissue-specific expression in mouse embryonic heart tissue [36] . Another study found that two mammalian hypothalamic enhancers have no homolog across non-mammalian vertebrates , yet are still able to drive specific expression patterns in zebrafish neurons [37] . These findings are consistent with gene expression evolution being shaped predominantly by stabilizing selection on the expression level [38] , while evolution on the sequence level may involve an interplay between fixation of weakly deleterious mutations through drift , and weak positive selection on compensatory mutations [39] . However , not all TFBS turnover events are neutral or nearly neutral on the level of gene expression , and the fraction of such events that change gene expression may be substantial [31] . More generally , lineage-specific sequence is clearly a likely substrate for lineage-specific biology [16] , [34] , although adaptations to pre-existing functional sequence remain an alternative plausible mode for creating species-specific change [40] . Nevertheless , the sheer ubiquity of sequence turnover , and the clear potential for substantial regulatory change resulting from it , suggests that many aspects of noncoding human biology will not be fully recapitulated by orthologous sequence in eutherian model organisms , including mouse . Thus , our findings could provide a more quantitative basis for assessing the relevance of model organisms to specific questions of human biology .
We restricted our analyses to genome assemblies that have been sequenced at relatively high coverage , not using for example the 2-fold coverage assemblies of mammalian genomes [41] , to minimize the impact of sequencing and assembly errors . From the UCSC Genome Informatics website ( http://genome . ucsc . edu/ ) , we acquired softmasked versions of the following genome assemblies: human ( hg19 ) , mouse ( mm10 , mm9 , and mm8 ) , rat ( rn5 ) , cattle ( bosTau7 ) , dog ( canFam2 ) , horse ( equCab2 ) , guinea pig ( cavPor3 ) , rabbit ( oryCun2 ) , bushbaby ( otoGar3 ) , panda ( ailMel1 ) , and rhino ( cerSim1 ) . We also acquired a Ferret genome assembly ( M_putorius_furo_v1 ) produced by the Broad Institute . We softmasked the ferret genome assembly using RepeatMasker with carnivore repeat libraries [42] . When available , whole genome pairwise alignments were downloaded from the UCSC Genome Informatics website ( http://genome . ucsc . edu ) . Otherwise , we constructed alignments following UCSC's protocol [43] . Initial alignments were constructed with LASTZ ( http://www . bx . psu . edu/miller_lab/ ) , a derivative of BLASTZ [44] , and these alignments were subsequently chained and netted using tools from UCSC ( Table S1 for alignment parameterisations ) . We trimmed each of the whole genome alignments once we found that UCSC alignments contained a minority of poorly aligning sequence ( Figure S1 , Table S2 ) . Each alignment was rescored to generate a new substitution matrix using a log-odds ratio approach as described previously [45] . We did not impose symmetry on the scoring matrixes with respect to strand or species . We then used the generated substitution matrix , with gap penalties derived from the original alignments , to discard ( “trim” ) the maximal non-positively scoring terminal segments of the alignment blocks and any non-positively scoring inter-gap segments . Trimming removes terminal and internal alignment segments that are more likely to have arisen under a model of independent evolution than of evolution from a common ancestor . Subsequent analyses were carried out following the discarding of all trimmed sequence . We also excluded alignments that were led by sequence not mapped to chromosomes . We did not exclude non-reciprocally aligning sequence or sequence that lay within known indel hotpot locations as we found removing such sequence had relatively small effects on estimates of αselIndel ( Table S3 ) . The neutral indel model of Lunter et al . ( 2006 ) [18] ( NIM1 ) estimates the genomic fraction ( αselIndel ) of sequence constrained with respect to indels between a species pair . The model examines the distribution of IGSs from a set of whole genome pairwise alignments using a regression approach over a range of medium IGS lengths to estimate the parameters of a predicted geometric distribution of IGSs in neutral sequence . αselIndel in bp is then estimated by summing up the quantity x - 2K over all the long IGSs inferred to be in excess of predictions under neutral evolution . Here where x is the length of the overrepresented IGS , and K is the estimated mean spacing between indels ( “neutral overhang” ) . 20 equally populated G+C content bins are analysed separately to account , in part , for mutational variation that correlates with G+C content . The X chromosome is also analysed separately . A detailed description of the model is given in the original publications [15] , [18] . However , two theoretical issues of the model have not been described previously . These are: ( A ) that thresholding biases the expected lengths of the neutral overhang and , ( B ) that neutral segments are depleted from the background distribution due to the presence of constrained segments , changing the expected neutral distribution of IGS lengths; resolution of the two issues is described in Text S1 . Our implementation of the NIM1 differs from that of the preceding studies in the manner in which we calculate the bounds of the estimates . The previous approaches constructed the upper and lower bound estimates based on the uncertainty in the degree of clustering of functional elements . The lower bound estimate was derived assuming that functional elements are unclustered ( each overrepresented IGS contributes x - 2K bp towards the αselIndel estimate ) , while the upper bound was derived assuming a high degree of clustering ( each overrepresented IGS contributes x - K bp ) . In our revised approach , we construct a 95% confidence interval around the lower x - 2K bp estimate . The impact of this change on αselIndel estimates can be seen in the simulation study ( Table S5 ) . We made this conservative modification to the NIM1 for five reasons: Firstly , the previous upper bound estimate assumes an unrealistically high degree of clustering of functional elements . Secondly , only our modified estimate is always conservative under all the simulation scenarios , whereas the previous implementation of the NIM1 sometimes overestimates the true value of αselIndel ( Table S5 ) . Thirdly , altering the clustering of functional elements in the simulations actually has only a minor effect on the estimated quantities of constrained sequence ( Figure S11 ) . Fourthly , in addition to the clustering of functional elements , other parameterisations also influenced αselIndel estimates ( Table S5 ) , yet the uncertainty in the values of these parameters was not also incorporated into the NIM1 estimate . Instead , we now choose to incorporate the full extent of uncertainty into the simulations . Finally , by providing a 95% confidence interval for the αselIndel estimate of NIM1 , we have an estimate that is directly comparable to the NIM2 estimates . We have described above how NIM1 is used to estimate the fraction αselIndel of constrained bases within a genome G consisting largely of neutrally evolving sequence . To estimate αselIndel within a subset S⊆G that is not dominated by neutrally evolving sequence , for instance when estimating αselIndel within coding sequence , we instead estimate αselIndel within the subsets G and G\S; the difference between the resulting estimates is the estimate of αselIndel within S . We extracted ancestral repeat ( AR ) alignments from the trimmed whole genome alignments using RepeatMasker annotations to identify transposable element and repeat-derived sequence [42] . We then calculated the substitution rate for the alignments using the HKY85 model applied in the PAML package BASEML [46] . We also estimated synonymous substitution rates ( dS ) across protein coding regions for some species pairs . Estimates of dS for a species pair were made by calculating the median dS of all one-to-one gene orthologs in the Ensembl Compara database with dS<1 . Nucleotide substitution rates in AR sequences are very similar to estimates of the synonymous substitution rate ( dS ) ( Figure S5 ) , hence our results appear insensitive to the choice of neutral sequence standard . The time-homogeneous turnover model makes the following assumptions: for a particular class of functional elements , both the total amount of functional sequence and the rate of turnover are constant in time , and the turnover rate ( weighted by the length of the elements ) is identical for all elements in the class . Specifically , within a class of functional sites comprising a nucleotides , in a small time interval dt a number a b dt of sites dispense with function , while an identical number gain function . Note that to arrive at this result we make an “infinite sites” assumption , namely that the genome can be considered infinitely large compared to a; otherwise one would need to account for reversions back to functionality of neutral but previously functional material . Fitting the data to this model under the assumption of independent normally distributed errors in the observations provides estimates and error bounds on parameters a and b . Coding sequence for human ( hg19 ) , mouse ( mm10 ) , and dog ( canFam2 ) and UTR annotations for human ( hg19 ) were obtained from Ensembl version 72 ( http://www . ensembl . org/index . html ) . UTR sequence that overlapped coding sequence was not considered in the UTR analyses . Human ( hg19 ) PhastCons conserved elements were taken from the vertebrate PhastConsElements46way track downloaded from UCSC Genome Informatics ( http://genome . ucsc . edu/ ) . Human ( hg19 ) GERP++ conserved elements were downloaded from the Sidow laboratory website ( http://mendel . stanford . edu/SidowLab/downloads/gerp/ ) . Repetitive element annotations for all species were taken from RepeatMasker [42] . Other human ( hg19 ) annotations were taken from the ENCODE data available at UCSC Genome Informatics ( http://genome . ucsc . edu/ENCODE/ ) . Specifically , the TFBS data and DNase HS data were acquired from the ENCODE clustered merged sets ( wgEncodeRegTfbsClusteredV2 . bed and wgEncodeRegDnaseClusteredV2 . bed respectively ) . Promoter and enhancer elements were extracted from the ENCODE HMM Chromatin State segmentations tracks , and merged across these samples: wgEncodeBroadHmmGm12878HMM . bed . gz , wgEncodeBroadHmmH1hescHMM . bed . gz , wgEncodeBroadHmmHepg2HMM . bed . gz , wgEncodeBroadHmmHmecHMM . bed . gz , wgEncodeBroadHmmHsmmHMM . bed . gz , wgEncodeBroadHmmHuvecHMM . bed . gz , wgEncodeBroadHmmK562HMM . bed . gz , wgEncodeBroadHmmNhekHMM . bed . gz , and wgEncodeBroadHmmNhlfHMM . bed . We display the results from analysis of the set of Hangauer et al . ( 2013 ) [21] lncRNAs in Figure 3 . We also used the smaller set of ENCODE lncRNAs in Figure S9 .
|
Nearly 99% of the human genome does not encode proteins , and while there recently has been extensive biochemical annotation of the remaining noncoding fraction , it remains unclear whether or not the bulk of these DNA sequences have important functional roles . By comparing the genome sequences of different species we identify genomic regions that have evolved unexpectedly slowly , a signature of natural selection upon functional sequence . Using a high resolution evolutionary approach to find sequence showing evolutionary signatures of functionality we estimate that a total of 8 . 2% ( 7 . 1–9 . 2% ) of the human genome is presently functional , more than three times as much than is functional and shared between human and mouse . This implies that there is an abundance of sequences with short lived lineage-specific functionality . As expected , most of the sequence involved in this functional “turnover” is noncoding , while protein coding sequence is stably preserved over longer evolutionary timescales . More generally , we find that the rate of functional turnover varies significantly across categories of functional noncoding elements . Our results provide a pan-mammalian and whole genome perspective on how rapidly different classes of sequence have gained and lost functionality down the human lineage .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genomics",
"functional",
"genomics",
"genome",
"evolution",
"genetics",
"biology",
"and",
"life",
"sciences",
"comparative",
"genomics",
"computational",
"biology"
] |
2014
|
8.2% of the Human Genome Is Constrained: Variation in Rates of Turnover across Functional Element Classes in the Human Lineage
|
The oral microbiome–organisms residing in the oral cavity and their collective genome–are critical components of health and disease . The fungal component of the oral microbiota has not been characterized . In this study , we used a novel multitag pyrosequencing approach to characterize fungi present in the oral cavity of 20 healthy individuals , using the pan-fungal internal transcribed spacer ( ITS ) primers . Our results revealed the “basal” oral mycobiome profile of the enrolled individuals , and showed that across all the samples studied , the oral cavity contained 74 culturable and 11 non-culturable fungal genera . Among these genera , 39 were present in only one person , 16 genera were present in two participants , and 5 genera were present in three people , while 15 genera ( including non-culturable organisms ) were present in ≥4 ( 20% ) participants . Candida species were the most frequent ( isolated from 75% of participants ) , followed by Cladosporium ( 65% ) , Aureobasidium , Saccharomycetales ( 50% for both ) , Aspergillus ( 35% ) , Fusarium ( 30% ) , and Cryptococcus ( 20% ) . Four of these predominant genera are known to be pathogenic in humans . The low-abundance genera may represent environmental fungi present in the oral cavity and could simply be spores inhaled from the air or material ingested with food . Among the culturable genera , 61 were represented by one species each , while 13 genera comprised between 2 and 6 different species; the total number of species identified were 101 . The number of species in the oral cavity of each individual ranged between 9 and 23 . Principal component ( PCO ) analysis of the obtained data set followed by sample clustering and UniFrac analysis revealed that White males and Asian males clustered differently from each other , whereas both Asian and White females clustered together . This is the first study that identified the “basal mycobiome” of healthy individuals , and provides the basis for a detailed characterization of the oral mycobiome in health and disease .
Organisms residing in the oral cavity and their collective genome–the oral microbiome–are critical components of health and disease . Disruption of the oral microbiome has been proposed to indicate , trigger , or influence the course of oral diseases , especially among immunocompromised patients ( e . g . HIV-infected or cancer patients ) [1]–[3] . Although fungi , particularly Candida , are important components of oral microbiota and are influenced by the immune status and therapy of affected individuals , studies of oral microbiota have focused largely on the bacterial components . In the only oral microbiome study to date that included some fungal profiling , Aas et al . [4] reported the presence of Candida albicans and Saccharomyces cerevisiae in the subgingival plaque microbiota of HIV-infected patients . These investigators employed a PCR-based approach using the 18S rDNA primers ( that amplify Candida spp . and eight divergent fungal genera only ) to characterize the fungi present in the plaques . The approach used by this group provided only a limited snap shot of the fungal members of the microbial biome . To obtain a more comprehensive profile of the fungal microbiome ( mycobiome ) , in this study we utilized a novel Multitag Pyrosequencing ( MTPS ) approach to interrogate the fungal taxa in the oral cavity using universal internal transcribed spacer ( ITS ) primers , which have broad fungal specificity [5]–[13] . Using this approach , we characterized the “basal” mycobiome profile of 20 healthy individuals , and showed that across all the samples studied , the oral cavity contained 74 culturable and 11 non-culturable fungal genera . Among these culturable genera , 61 were represented by one species each , while 13 genera comprised between 2 to 6 different species; the total number of species identified were 101 . This is the first study that identified the “basal mycobiome” of healthy individuals , which provides the basis for detailed characterization of the oral mycobiome in health and disease .
Written informed consent was obtained from all participants in this study . Recruitment of study participants was performed according to protocol ( number 20070413 ) approved by the Human Subjects Institutional Review Board ( IRB ) of Case Western Reserve University , Cleveland , Ohio . Oral rinse samples were obtained from 20 healthy individuals after informed consent and following review of the IRB at Case Western Reserve University/University Hospitals Case Medical Center . The individuals were all from the Cleveland area and on standard Western diets . Summary demographic information of the study participants is provided in Table 1 . Self-reported ethnicities of study participants were classified based on the US Census criteria for classification of race ( http://www . census . gov ) , in which race has been classified as White ( including Hispanic , East Indian , or European ) , Black/African-American , Asian , Native American/Native Alaskan etc . Inclusion criteria were: >18 years of age , non-smoking , no recent antifungal use , and no clinical signs of oral mucosal disease . Exclusion criteria were: ( 1 ) a history of receiving medication or treatment with topical or systemic steroids , pregnancy , and ( 2 ) insulin-dependent diabetes mellitus ( IDDM ) . Concentrated oral rinse has been previously used to detect the presence of oral bacteria and fungi [14]–[16] . We selected oral rinse for our studies because: ( a ) it is relatively simple and noninvasive to collect , ( b ) safer to handle than other body fluids ( e . g . serum ) , ( c ) the oral cavity is the major entry point of microbes into the body and ( d ) oral rinse enables the collection of organisms from the dorsum of the tongue and the oral mucosal environment . Oral samples were collected at least 1 h after a meal , at approximately the same time ( 9–11 AM ) , to avoid contamination of samples with extraneous components and standardize the possible impact of variation in salivary flow rates . Study participants rinsed their mouth ( swish/gargle ) with 15 mL sterile phosphate buffered saline ( PBS ) for 1 min , and expectorated the contents of the mouth into a 50 mL centrifuge tube . The collected samples were centrifuged at 4000 rpm for 20 min at 4°C to separate the cells ( pellet ) from extracellular soluble components ( supernatant ) . The cell pellet was used for DNA extraction or stored at −80°C until the time of analysis . The first step in the mycobiome analysis was extraction of DNA from the cell pellet ( obtained from oral rinse samples , above ) followed by PCR analysis . Samples were extracted individually using the Fast DNA Spin Kit for fungi following manufacturer's instructions ( BIO 101; Vista , CA ) . Each extraction tube was agitated three times using a Fast Prep FP120 instrument at a speed setting of 5 for 30 s . Tubes were cooled on ice between agitations . The ITS1 region from DNA sample extracts was amplified in triplicate using primers with high specificity for ascomycete fungi ( fluorescently-labeled forward primer ITS1F ( CTTGGTCATTTAGAGGAAGTAA ) and unlabeled reverse primer ITS2 ( GCTGCGTTCTTCATCGATGC ) . The ITS primers were selected in this study to detect the presence of various fungi since these primers are able to detect consensus sequences present in a broad range of fungal organisms , and have been used for the detection of yeasts ( including Candida spp . ) , moulds and dermatophytes [5]–[13] . The reactions were carried out on ∼10 ng template DNA , in 20-µl ( final volume ) reaction mixtures consisting of 1×PCR buffer , 0 . 01% bovine serum albumin , 2 . 5 mM MgCl2 , each dNTP at a concentration of 0 . 25 mM , each primer at a concentration of 0 . 5 µM , and 0 . 5 U of AmpliTaq Gold DNA polymerase ( ABI , Foster City , CA ) . Initial denaturation at 94°C for 11 min was followed by 35 cycles of denaturation for 30s each at 94°C , annealing at 50°C for 30 s , and progressive extension at 72°C for 2 min . Following the 35 cycles there was a final extension time of 30 min to minimize artifacts induced by TAQ polymerase . Fungal PCR products were separated on the SCE 9610 capillary DNA sequencer ( Spectrumedix LLC , State College , PA ) using GenoSpectrum software to convert fluorescent output into electropherograms . Relative peak abundance of fungal amplicons was calculated by dividing individual peak heights by the total peak heights in a given electropherogram using a custom PERL script . Interleaved , normalized abundances were compared as stacked histograms using Microsoft Excel . Mean normalized abundance for each amplicon was calculated from the three PCR replicates of each sample , excluding means below 1% . Results were analyzed by visual inspection and Principal Coordinate ( PCO ) analysis using Multivariate Statistical Package ( MVSP , Kovach Computing Services , Wales , UK ) . Normalized abundance of each peak in the electropherogram was calculated with respect to the total peak area , since it is not possible to calculate absolute abundances with either the LH-PCR or MTPS technology . Mycobiome analysis was performed using multitag 454 pyrosequencing ( MTPS ) technique , which can be used for detailed characterization of nucleic acids and has the advantages of accuracy , flexibility , parallel processing , and easy automation potential [17] . In this technique , organism-specific sequences are amplified , and the PCR amplicons are converted to single-stranded DNA templates and immobilized onto streptavidin-coated beads . Next , an enzymatic cascade facilitates the simultaneous synthesis of complementary DNA . As each nucleotide is incorporated into the newly synthesized strand , a luciferase-dependent bioluminescent signal is generated , with the intensity of each signal being proportional to the number of incorporated nucleotides . The bioluminescent signal is detected and analyzed by the instrument in real time , with the resulting generation of a pyrogram that consists of a series of peaks whose temporal relationship and height reflect the DNA sequence . Specifically , we generated a set of 24 emulsion PCR fusion primers that contain the 454 emulsion PCR adapter , joined to a 7 base “barcode” along with the appropriate target primers . Specifically for this experiment , we used the A Adapter with a barcode and the ITS1F sequence for the forward primer and the B Adapter with the ITS4A sequence ( i . e . without a tag ) for the reverse primer . All sequences were read from the A Adapter side . Thus , each oral rinse sample was amplified with a uniquely barcoded set of forward and reverse rRNA primers and then up to 24 samples were pooled and subjected to emulsion PCR and pyrosequenced using a GS-FLX pyrosequencer . Data from each pooled sample were “deconvoluted” by sorting the sequences into bins based on the barcodes using custom PERL scripts . Thus , we were able to normalize each sample by the total number of reads from each barcode . Several groups have subsequently employed various barcoding strategies to analyze multiple samples and this strategy is now well accepted [18]–[21] . We developed a custom PERL script to demultiplex the MTPS data by sorting the sequences into bins based on the barcodes and the taxa in the samples , automatically blasting the pyrosequence data against Genbank ( 98% cutoff , which was sufficient for species identification ) . The annotations for each sequence were downloaded and a PERL script used to tabulate the taxa as a percentage of the total oral community in each sample . Fungal ITS sequences were compared with the Assembling Fungal Tree of Life ( AFTOL ) database using the BLAST interface of Web Accessible Sequence Analysis for Biological Inference ( WASABI ) as well as against the NCBI nucleotide database . Principal Coordinate ( PCO ) analysis has been recognized as a simple and straight-forward method to group or separate samples in a dataset , and has been used in disease-association studies [22] , [23] . In the current study , PCO was used to analyze the MTPS results using the Multivariate Statistical Package , MVSP ( Kovach , Wales , UK ) and SAS ( Cary , NC ) . The PCO analysis performs an Eigen analysis on the data matrix using a Brays Curtis distance metric . Graphically , PCO is a rotation of a swarm of data points in multidimensional space so that the axis with the greatest variance is the first principal component axis . The second axis orthogonal to the first is the second principal component and represents the second greatest variance of the data . The first two or three principal components generally account for most of the variance of the data . To compare the phylogenetic distribution between the various gender and race classes and confirm the results of PCO analysis , we used UniFrac significance test [24] , [25] . This test measures the probability that the designated classes are different based on the phylogenetic relatedness and the abundance of each cluster . We first clustered the MTPS reads using CD-HIT [26] to reduce the number of sequences used in the multiple alignment . These clusters were annotated by the number of reads in each sample and then aligned using KALIGN [27] . A tree file was constructed following the multiple alignment using PAUP and an environment file was defined for each sample using a custom PERL script . These two files were then loaded into the UniFrac online server ( http://bmf2 . colorado . edu/unifrac/index . psp ) . The analysis was weighted using the abundance of the sequence in each cluster . The P-value generated from the UniFrac significance test describes the degree of similarity between classes . Values reported were Bonferroni corrected for the number of sequences used in comparisons .
The demographic characteristics of the study participants were: 21–60 years of age , 8 females and 12 males , no history of smoking , insulin-dependent diabetes mellitus , or active medications; the self-reported ethnicities of study participants were White ( n = 8; 2 females , 6 males ) , Asian ( Han Chinese , Indian or Bangladeshi ) ( n = 10; 4 females , 6 males ) , or African-American ( n = 2 , both females ) . All participants were from the metropolitan Cleveland Ohio area , and were Faculty , staff or students at Case Western Reserve University ( Table 1 ) . Study participants were assumed to consume a varied “Western” diet . Analysis of complex population genetic structure to control for differences in continental ancestory was not performed for these studies . The ITS-based sequencing runs produced 39 , 226 sequence reads of which 36 , 155 contained identifiable tags . Of those , 34 , 049 sequences ( 94% ) were longer than 100 bases and were used in the analysis . Our analyses revealed an average of 1 , 702 sequences per sample with an average length of 248 bases ( Table 2 ) . A local copy of Genbank was searched using megablast and the highest hit ( 98% cutoff ) was compiled using the score for each sequence and the results were then tabulated using a custom PERL script ( see Supplemental Table S1 for sequence details ) . Figure 1 shows the genera identified in the oral rinse samples collected from the 20 participants that were ≥1% of the community in individual samples . Across all the samples studied , the oral cavity contained 74 culturable and 11 non-culturable fungal genera ( supplemental Tables S2 and S3 ) . Thirty-nine genera were present in only one person , 16 genera were present in two participants , 5 genera were present in three people , while 15 genera ( including the non-culturable genera ) were present in 4 or more people ( Fig . 1 ) . Among the culturable genera , 61 were represented by one species each , while 13 genera comprised between 2 to 6 different species; the total number of species identified were 101 . The number of species in the oral cavity of each individual ranged between 9 and 23 ( supplemental Tables S2 and S3 ) . More than 10 different genera with an abundance of >1% were detected in 70% ( 14/20 ) of the samples analyzed ( supplemental Tables S2 and S3 ) . When compared across all 20 individuals , four genera were present in the oral rinse of 10 or more study participants: Candida ( 15/20 ) , Cladosporium ( 13/20 ) , Aureobasidium ( 10/20 ) , and organisms belonging to the family Saccharomycetales ( 10/20 ) . Interestingly , a large percentage ( 36 . 1% ) of fungi belonged to non-culturable category . The minimum number of genera identified in a sample ( sample E3 ) was 3 ( Candida , 15%; Saccharomycetales , 18 . 1%; unculturable , 62 . 3% ) , while the maximum number of genera identified was 16 in sample A2 , which included Candida ( 5 . 5% ) , Dothediomycete ( 10 . 7% ) , Fusarium ( 4 . 1% ) , Aspergillus ( 3 . 9% ) , and Xylariales ( 7 . 7% ) ( Fig . 2 ) . The non-culturable genera/family detected in the oral samples included Glomus , Leptosphaeriaceae , Ascomycete , Basidiomycete , Ectomycorrhiza , Endophytic fungi , and Glomeromycete . To determine the basal fungal distribution in healthy individuals , we identified fungi that were present in at least 20% ( 4/20 individuals sampled ) of the study participants . This analysis revealed that 15 genera were present in ≥20% of the tested samples ( including non-culturable fungi , Fig . 3 ) . Among these samples , Candida species were the most frequently obtained genera , isolated from 75% of all study participants , followed by Cladosporium ( 65% ) , Aureobasidium and Saccharomycetales ( 50% for both ) . Other fungi that were present in the oral cavity of healthy individuals were Aspergillus ( 35% ) , Fusarium ( 30% ) , and Cryptococcus ( 20% ) . Fifty three percent ( 39/74 ) of the identified genera were observed only once in the tested samples . Analysis of the species distribution of the oral mycobiome revealed that 12 fungi were represented by two or more species in the oral rinse samples ( Supplemental Tables S2 and S3 ) . The highest number of species was detected for Aspergillus ( 6 species ) , followed by Candida ( 5 species ) , Cladosporium ( 4 species ) , Fusarium ( 3 species ) , and Penicillium ( 3 species ) ( Table 3 ) . Candida albicans was identified in 40% of the participants ( 8/20 ) , while the non-albicans Candida species indentified were: C . parapsilosis ( 15% ) , C . tropicalis ( 15% ) , C . khmerensis and C . metapsilosis ( in 5% of the subjects ) . Two species each of Alternaria , Cryptococcus , Ophiostoma , Glomus , Phoma , Schizosaccharomyces , and Zygosaccharmoyces were identified in the participants . To investigate whether there is an association between any of the subject demographics and changes in mycobiome , we performed PCO analysis and sample clustering followed by UniFrac analysis . Our analysis revealed that White and Asian males clustered differently from each other , whereas both Asian and White females clustered together ( Fig . 4 ) . UniFrac analysis of females , White males , and Asian males showed that each of these classes was significantly different from the other ( Table 4 ) , supporting the PCO clustering denoted by the circles in Figure 4 . These data suggest a trend of association between gender/ethnicity and the oral mycobiome .
In the current study , we demonstrated the presence of 74 culturable and 11 non-culturable fungal genera in the oral cavity of healthy individuals , with between 9 and 23 culturable species present in each person , representing a total of 101 species for all study participants . Our results demonstrate that the fungal component of the oral microbiome is not limited to a few species , principally Candida; rather it is represented by a large number of diverse fungi . The perception that fungi in the oral cavity are limited to only few species originated from previous studies that relied upon the use of culture-based methods or species-specific targeted PCR approach . In addition to Candida , other fungi previously reported in the oral cavity include S . cerevisiae , Penicillium , Geotrichum , Aspergillus , Scopulariopsis , Hemispora , and Hormodendrum [28]–[30] . In the current study , we used the pan-fungal ITS probes in conjunction with 454 pyrosequencing , which allowed us to identify oral fungi in a highly specific and sensitive manner . This real-time DNA sequencing method allows rapid analysis of sub-sequences within the ITS regions and comparison with nucleic acid sequence databases , thereby facilitating rapid and accurate species level identification of fungi [31] . Another reason for the successful identification of a large number of fungal species ( 101 species ) in the oral cavity is the ability of MTPS to perform concomitant analysis of multiple samples . Only two previous studies have investigated the profile of microbes present in the oral cavity of healthy individuals and both focused on the bacterial microbiome [32] , [33] . Aas et al . [32] used PCR amplification of 16S rRNA genes followed by sequencing to analyze nine oral sites from five clinically healthy subjects , and reported detection of 141 bacterial species across all 5 subjects , of which over 60% were non-culturable . The number of predominant species per individual ranged from 34 to 72 . In our study , we found a total of 101 fungal species across all 20 individuals , with the number of species per individual ranging between 9 and 23 . Diversity of fungal taxa has also been shown to exist in murine models [34] . Our results suggest that the distribution and profile of fungal species in the oral cavity of healthy individuals is complex , and similar to that of oral bacterial microbiome with respect to the number of species identified . While the fungal component of the oral cavity has not been investigated in healthy individuals , a previous study identified fungi present in the oral cavity of HIV-infected patients [4] . In this study , Aas et al . [4] analyzed sub-gingival plaque of 14 HIV-infected patients , and reported the presence of S . cerevisiae in 4 and C . albicans in 2 patients . No other fungal species were detected in analysis of 306 18S rDNA clones . In contrast , we found that the oral cavity of healthy individuals had 101 fungal species . The reason for this difference could be attributed to differences in: ( a ) sampling method – oral rinse versus sub-gingival plaque; oral rinse enables the collection of organisms from the dorsum of the tongue as well as from the ever changing oral mucosal environment as compared to the sub gingival biofilm plaque , ( b ) detection probe – 18S rDNA probes that detected Candida and eight other genera , versus the pan-fungal ITS1/ITS2 probe that could identify all fungi , and ( c ) sequencing technique – cloning of rDNA fragments followed by sequencing , versus real-time pyrosequencing . We also found that the distribution of fungal species varied greatly between different individuals . Similar variation was recently reported for bacterial microbiota by Nasidje et al . [33] , who analyzed the global diversity of the salivary microbiome in 120 healthy individuals and showed that it varied greatly within and between individuals . Our results from the PCO and UniFrac analyses showed some tendency for white males to cluster together , and Asian males to cluster together . However , given the small number of study participants in the current study , it is difficult to draw definite conclusions regarding association of the biome with gender and/or ethnicity . To our knowledge , this is the first study suggesting such trend , and needs to be confirmed in studies involving larger population sizes . The oral mycobiome of at least 20% of the enrolled individuals included the four most common pathogenic fungi – Candida ( present in 75% of the cohort ) , Aspergillus ( 35% ) , Fusarium ( 30% ) , and Cryptococcus ( 20% ) . While the abundance of Candida in these healthy individuals was not surprising , the actual percentage was higher than reported in earlier culture-based studies , where 40 to 50% of healthy individuals have been shown to contain Candida species in their oral cavity [35] , [36] . The high percentage of Candida detected in this study could be attributed to the use of the more sensitive ITS/pyrosequencing approach . Another interesting finding was the different types Candida species identified in the oral cavity of healthy individuals . The most abundant Candida species in this study was found to be albicans ( in 40% of the subjects ) , followed by C . parapsilosis ( 15% ) , C . tropicalis ( 15% ) , C . khmerensis ( 5% ) and C . metapsilosis ( 5% ) . These results are in agreement with those reported by earlier studies using culture-based as well as PCR-based analyses [35] , [36] . The presence of Aspergillus , Fusarium , and Cryptococcus isolates in the oral cavity of healthy individuals was unexpected , since these fungi have not been reported to be colonizers of the oral cavity . It is possible that the pathogenicity of these fungi is controlled in healthy individuals by other fungi in the oral mycobiome , as well as a functional immune system . It is possible that inter-dependent relationships may exist between components of the oral mycobiome , and need to be investigated using broader sampling and longitudinal studies . In our studies , we also identified 60 fungal genera that are ubiquitous in the environment ( present in plants , soil , air ) and not normally associated with infections . Among these genera , 39 genera occurred once among the 20 samples analyzed , 16 genera occurred with a frequency of 2 , while 5 genera occurred in three individuals . Due to their ubiquitous nature , the presence of these organisms in the oral cavities of healthy individuals was not surprising , which are most likely of environmental origin , from food and mouth breathing . In this study , non-culturable fungi represented a large percentage ( 36 . 1% ) of the organisms identified in the oral mycobiome of healthy individuals . This is the first study demonstrating the presence of non-culturable fungal organisms in the oral cavity , which may play important role in the oral milieu . The presence of non-culturable organisms has been reported for bacterial species in the oral cavity; about half the population of oral bacteria has been reported to be non-culturable . For example , Aas et al . [4] reported that of the 109 bacterial species identified in the oral cavity of HIV-infected patients , 60% were non-culturable . Although some studies suggested that non-culturable bacteria may be associated with oral disease and health [37] , [38] , inability to grow these non-culturable organisms renders it difficult to gain insight into their role in health and disease . Moreover , non-culturable organisms may exhibit antimicrobial resistance , which may be the underlying reason for failure to manage certain infections . The field of molecular ecology abounds with examples where molecular methods identify numerous taxa that play important ecological functions but cannot be grown in the lab . Therefore , it is critical to fully characterize ecological communities like the oral mycobiome to fully understand the functionality of that ecosystem . Ours is the first study that provides a snapshot of the oral mycobiome in various individuals , and addresses an important component of the Human Microbiome Project ( HMP ) . Since this is the first study that identifies the oral mycobiota , our findings complement the HMP ( focusing mainly on the bacterial component ) and support its stated goal of “generating resources enabling comprehensive characterization of the human microbiota and analysis of its role in human health and disease ( http://nihroadmap . nih . gov/hmp/ ) ” . Results from our study provides critical information that is likely to form the basis of further “hypothesis-driven” studies evaluating the oral mycobiome in terms of individual variabilities , longitudinal trends , and the effect of diet and geography , and studies focused on determining the association of oral mycobiota with health and disease . The clinical relevance for the presence of a diverse population of fungal species in the oral cavity is unknown . It is possible that the presence of a given fungal isolate ( e . g . Candida , Aspergillus , Cryptococcus , and Fusarium ) in an individual could be the first step in predisposing the host to opportunistic infections . In this regard , oral Candida colonization has been known to be a risk factor for Candida infections in immunocompromised patients [39] , [40] . Understanding the relationships between different fungal species as well as between fungi and other members of the oral microbiome will shed light on the pathogenicity of these organisms and may lead to the discovery of novel therapeutic approaches for the prevention and treatment of oral complications .
|
We characterized the fungal microbiome ( mycobiome ) of the oral cavity in healthy individuals . Our results demonstrate that the fungal component of the oral microbiome is diverse as revealed by the presence of 74 culturable and 11 non-culturable fungal genera in the oral cavity . A total of 101 species were identified , with between 9 and 23 culturable species present in each person . Fifteen genera ( which included four known pathogenic fungi and non-culturable organisms ) were present in ≥20% of the tested samples; Candida species were the most frequently obtained genera , isolated from 75% of all study participants , followed by Cladosporium ( 65% ) , Aureobasidium , Saccharomycetales ( 50% for both ) , Aspergillus ( 35% ) , Fusarium ( 30% ) , and Cryptococcus ( 20% ) . The remaining fungi detected in the oral wash samples represent organisms likely originating from the environment . This is the first study that identified the “basal mycobiome” of healthy individuals , and provides the basis for a detailed characterization of the oral mycobiome in health and disease .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"molecular",
"biology/bioinformatics",
"microbiology/microbial",
"physiology",
"and",
"metabolism",
"microbiology"
] |
2010
|
Characterization of the Oral Fungal Microbiome (Mycobiome) in Healthy Individuals
|
Escherichia coli bacteria respond to DNA damage by a highly orchestrated series of events known as the SOS response , regulated by transcription factors , protein–protein binding , and active protein degradation . We present a dynamical model of the UV-induced SOS response , incorporating mutagenesis by the error-prone polymerase , Pol V . In our model , mutagenesis depends on a combination of two key processes: damage counting by the replication forks and a long-term memory associated with the accumulation of UmuD′ . Together , these provide a tight regulation of mutagenesis , resulting , we show , in a “digital” turn-on and turn-off of Pol V . Our model provides a compact view of the topology and design of the SOS network , pinpointing the specific functional role of each of the regulatory processes . In particular , we suggest that the recently observed second peak in the activity of promoters in the SOS regulon ( Friedman et al . , 2005 , PLoS Biology 3 ( 7 ) : e238 ) is the result of positive feedback from Pol V to RecA filaments .
The SOS response in the bacterium E . coli encompasses many proteins involved in detecting and repairing DNA damaged by a variety of agents , such as UV radiation , or chemicals such as mitomycin and bleomycin [1] . A complex regulatory network , comprising both transcriptional and post-translational regulators , controls the concentrations and levels of activity of these proteins ( Figure 1 . ) The collective actions of this regulatory network are orchestrated so that the SOS response is commensurate with the magnitude of DNA damage [1] . Mutagenesis , such as the introduction of single-base substitutions in the DNA sequence , is not an inevitable consequence of DNA damage , but results from the action of specialized error-prone DNA polymerases that are part of the response [2] . This constitutes an extreme measure that might be useful for the cell only after very heavy DNA damage when DNA replication and repair cannot effectively proceed without it . While some mutations might benefit the offspring , the vast majority is harmful; therefore , the presence of error-prone polymerases should be tightly regulated to prevent their action at low doses of UV . Briefly , the sequence of events triggered by UV irradiation of E . coli is as follows: UV radiation damages the DNA by creating lesions that mechanically disrupt the process of DNA duplication by stalling the DNA-polymerase ( Pol III ) in a moving replication fork . This , in turn , results in the production of single-stranded DNA ( ssDNA ) gaps . These gaps are coated by the protein RecA [1 , 3 , 4] , forming long nucleoprotein filaments in which it assumes its active form , RecA* . RecA* , together with other proteins , is involved in the nonmutagenic filling in of ssDNA gaps via homologous recombination [5] , and it catalyses the cleavage of the transcriptional repressor LexA [6] and of the protein UmuD [7] , whose cleaved form—UmuD′—is necessary for mutagenesis [1] . The drop in the level of the transcription factor LexA , due to its cleavage , de-represses the regulon involved in the SOS response . This regulon comprises about 30 genes , including those encoding the mutagenesis proteins UmuD and UmuC , RecA , and LexA itself . Also part of the SOS regulon are genes encoding UvrA , B , C—a group of nucleotide excision repair ( NER ) proteins that locate and excise damaged regions from the DNA [8 , 9] . Mutagenesis in UV-irradiated E . coli cells is mainly the direct result of the activity of the error-prone DNA polymerase , Pol V [2] . Pol V consists of two units of UmuD′ and one unit of UmuC . It inserts several random base pairs in the DNA strand directly opposite a lesion , thus helping a replication fork to quickly bypass the lesion , after which Pol III can take over and continue replication . A distinct coordinated subnetwork of proteins centered on UmuD and UmuC controls the abundance , and thereby the activity , of Pol V ( Figure 1 ) . Even though the SOS response in bacteria has been studied for several decades , new discoveries continue to be made . Recent single-cell experiments measured the temporal dependence of the activity of LexA-regulated promoters [10] , which showed the following features: For low UV doses , the promoter activity peaks at about 10 min after the UV dose . This was also observed in bulk measurements of promoter activity averaged over a large population of cells [4] and can be attributed to the initial rapid drop in LexA levels after UV damage because of the activation of RecA , followed by a slow increase to its original level as the lesions are repaired by NER and the level of RecA* falls . More surprising was the observation that at higher doses of radiation , LexA-regulated promoter activity often had a second peak at about 30–40 min , sometimes even followed by a third peak at 60–90 min . This resurgence of the SOS response is puzzling because it indicates a temporary increase in RecA* levels at a time when the NER process is well under way and the number of lesions are already falling . This second peak ( but not the third peak ) was , however , absent in both ΔUmuDC null-mutants and mutants that have an uncleavable version of UmuD ( K97A ) [10] . The common element in both types of mutants is the absence of Pol V , which suggests that the second peak is related to mutagenesis . In this paper we propose a plausible mechanism for the appearance of this peak . We argue that E . coli bacteria can reliably measure the total amount of DNA damage . The ability of replication forks to bypass bulky lesions allows the cells to “count” the number of lesions they encountered over a fixed time interval ( the average lifetime of RecA* filaments ) . The result of this count , given by the instantaneous number of RecA* filaments , is then fed into the mutagenesis regulatory subnetwork , which—as we show below—is designed to time-integrate this input signal over a long interval ( 30–40 min ) and to abruptly turn on the Pol V if the integrated level of damage exceeds some critical threshold . The appearance of Pol V speeds up the bypass of lesions , and thus increases the rate at which new lesions are encountered by replication forks . We believe that this positive feedback from Pol V to the RecA* concentration is responsible for a temporary increase in the activity of SOS-regulated promoters 30–40 min after the radiation ( the second peak reported in [10] . )
The goal of this paper is to model temporal dynamics of the mutagenesis subnetwork of the SOS response system ( highlighted in yellow in Figure 1 ) for different doses and durations of UV radiation . This subnetwork is not isolated from the rest of SOS response , and therefore the model includes other parts of the entire E . coli regulatory network that interact with proteins involved in mutagenesis . Figure 1 shows the components of the SOS response that we quantify in our model . Different colored arrows correspond to different mechanisms of interactions between the nodes . An excellent earlier paper by Aksenov [11] contains a model of LexA-controlled transcriptional regulation coupled with the NER repair of lesions during the SOS response . Here that model is extended to incorporate the mutagenesis subnetwork . Full details of our model and parameter values are provided in the Methods section . We mathematically model the temporal dynamics of the density of UV-induced lesions , as well as concentrations of LexA , RecA* , unbound UmuD , unbound UmuD′ , UmuD–UmuD′ heterodimer , and Pol V , using a set of ordinary differential equations . Positive and negative terms in these equations represent different ways of production and consumption/degradation of the corresponding quantities . We do not explicitly simulate the creation and repair of individual lesions , nor do we simulate each replication fork moving along the DNA . Thus , our model ignores stochastic fluctuations . However , in later sections we do examine the effect of averaging over a population of cells in which various parameters , e . g . , the number of replication forks , vary from cell to cell . This provides an in silico comparison between single-cell and cell-culture measurements . We also treat all time delays , such as when a replication fork is stalled at a lesion , in a simplified manner , i . e . , we assume that these delays affect the RecA* level only via the average replication speed . Most parameters in our model have been fixed using experimental data . For example , the experiments in [3 , 4 , 12] allow us to fix the RecA*-mediated cleavage rates of LexA and UmuD . The model has a total of 18 parameters of which only three could not be fixed by experimental data . We have therefore scanned a range of reasonable values for these three , as described in a later section . Our model indicates four key features of the mutagenesis subnetwork in E . coli: 1 . A mechanism for measuring the local amount of damage , coupling the number of RecA* filaments to the current lesion density . 2 . A long-term “memory” used to time-integrate the RecA* signal and thus to determine whether the damage level remained high for a substantial time . This mechanism is based on slow accumulation of UmuD′ . 3 . Strong binding between UmuD and UmuD′ , which provides a highly ultrasensitive increase in unbound UmuD′ levels as its concentration exceeds that of its “inhibitor” UmuD . 4 . Positive feedback from Pol V to RecA* levels , which further increases the sharpness of the turn-on and turn-off of Pol V . This mechanism is also responsible for the second peak in activity of SOS promoters . In the subsequent sections , we discuss each of the above aspects in more detail . First we propose the following mechanism for the influence of the UV dose on the RecA* level . Consider a given replication fork proceeding on a DNA strand that has UV-induced lesions , as depicted in Figure 2 . The Pol III DNA-polymerase stalls at the first lesion , generating an ssDNA gap that is then covered with RecA . This RecA filament exists for an average time , denoted τRecA* , after which it disassembles . ( We assume that each filament disassembles independently with a rate that is not limited by other DNA damage–induced processes . ) During this time the replication fork may bypass the lesion and continue processing the DNA , leaving the first RecA filament behind . If the time the fork spends stalled at a lesion is sufficiently large or the lesion density is sufficiently small ( so that the time the fork spends traveling between lesions is large ) , then the first filament will disassemble before the fork reaches the next lesion and creates another filament ( as in Figure 2A ) . Therefore , in this case , there will be no more than one RecA* filament per replication fork at any time . On the other hand , if the stall time is small or the lesion density is large , the fork will reach a second lesion before the first filament disassembles and , as a consequence , there may be many RecA* filaments per fork existing simultaneously on the DNA ( as in Figure 2B ) . The RecA* level directly depends on the time a polymerase spends traveling between lesions , τmoving = 1/μν , where μ is the density of lesions on the chromosome , and v is the average speed with which Pol III processes DNA replication on undamaged DNA . This dependence can be quantified: one RecA* filament is produced every time the replication fork encounters a lesion . If the fork spends time τstalled at a lesion and time τmoving between lesions , then the rate of production of RecA filaments is given by the following formula: Further , the filament disassembly rate is , where Nfil is the number of RecA* filaments associated with the replication fork under consideration and τRecA* is the average persistence time of a RecA* filament . Because the rates of filament production and disassembly are much faster than all other processes we are interested in ( the transcription of SOS genes and the rate of NER repair ) [13] , we can assume that the number of RecA* filaments at any given time are such that the production rate equals the disassembly rate , i . e . , The total amount of RecA* , r* , is given by the above expression multiplied by LRecA*—the average length of a RecA* filament ( taking into account the finite probability of forming a filament at each lesion a fork encounters ) —and Nf , the total number of replication forks currently duplicating DNA in a cell , i . e . , After fixing the parameter values based on experimental data ( see Methods ) , this relation gives a RecA* level of approximately 100 nM for a fixed lesion density produced by a UV dose of 2 J/m2 , while it gives more than 400 nM for a UV dose of 50 J/m2 ( this neglects the effects of Pol V , which will be discussed later ) . The process shown in Figure 2 is thus a simple way for the cell to “count” the number of lesions on the DNA using a “memory , ” which is the finite existence time of a RecA filament . This is a short-time memory lasting only for a time τRecA* . However , the rate of UmuD′ production is proportional to the amount of RecA* , therefore the UmuD′ level is a measure of RecA* level integrated over time . Thus , UmuD′ accumulates if damage ( and therefore RecA* ) persists for a long time . In our model , with RecA* at its maximum possible level , the timescale for the UmuD′ level to exceed that of UmuD is about 15 min . For smaller UV doses , and therefore lower RecA* , this rise time can be more than 35 min . UmuD′ is an integral component of the error-prone polymerase Pol V . However , UmuD′ has to accumulate to a fairly high level before Pol V appears in any detectable quantities . The main reason for this is a strong physical interaction between UmuD and UmuD′ . The binding between them is stronger than that between UmuD or UmuD′ pairs; when UmuD and UmuD′ are mixed in equimolar concentrations , the heterodimer is found to be much more abundant than either homodimer ( UmuD–UmuD and UmuD′–UmuD′ ) [14] . This strong binding ensures that unbound UmuD′ homodimers required for Pol V formation appear in sufficient quantities only when ( and if ) the total concentration of UmuD′ exceeds that of UmuD . Figure 3 shows the equations we use to model the dynamics of UmuD , UmuD′ , and Pol V . These equations model the following processes: ( 1 ) LexA represses the production of UmuD ( βu , Ku ) ; here , we assume a Hill coefficient of 1 based on the fact that the upstream region of the UmuD promoter has only one LexA binding site [15]; ( 2 ) RecA* catalyzes the intermolecular cleavage of UmuD [16] ( of both free and heterodimer forms ) to produce UmuD′ at rate γu; ( 3 ) UmuD and UmuD′ form a heterodimer [14] with on- and off-constants given by Kf , and Kb; ( 4 ) ClpX degrades UmuD′ ( but not UmuD ) when it is in the heterodimer [17] , at rate γdd′; ( 5 ) All molecules are diluted by cell growth and division ( γdil ) . Pol V is composed of two units of UmuD′ bound with one unit of UmuC protein . Thus , the level of Pol V cannot exceed that of UmuC ( C ) , but for small amounts of UmuD′ it is proportional to u′2 . K controls how much of the UmuD′ homodimer is required to saturate the levels of Pol V . The UmuC concentration C for simplicity is assumed to be constant during the narrow time window where it matters ( i . e . , when u′ is nonzero ) . The qualitative aspects of the dynamics produced can be understood by looking at a simplified version of these equations: since RecA* levels change relatively slowly , first consider UmuD and UmuD′ levels at a fixed RecA* concentration , and thus a constant UmuD → UmuD′ cleavage rate γur* . If the heterodimerization is extremely strong , the time course of the total ( free + heterodimer ) UmuD′ ( = u′ + uhetd ) satisfies the following rate equation ( see Methods for the derivation from the equations in Figure 3 ) : Here utot = u + uhetd is the total concentration of noncleaved UmuD ( in free or heterodimer form ) . The first term , γur*utot , is the production of UmuD′ due to the cleavage of UmuD , the second term is the ClpX-dependent degradation of UmuD′ inside UmuD′–UmuD heterodimers , while the last term is the decrease in the concentration of UmuD′ due to cell growth and division ( the dilution term ) common for all proteins in the cell . With LexA and RecA* levels fixed , i . e . , γur* constant , we can calculate the steady-state levels of UmuD and UmuD′ from these equations and , hence , the condition for Pol V to be present , i . e . , when UmuD′ exceeds UmuD: > utot . Setting = 0 and = utot , we obtain the condition for > utot in the steady state: independent of UmuD production and degradation rates . Thus , Pol V abruptly appears once the RecA* level , and hence the value of γur* , crosses and stays above the required threshold for long enough to allow UmuD′ to accumulate and pass the UmuD level . This analysis also suggests that there would be a threshold minimum UV dose below which Pol V does not appear because the NER repair brings down DNA damage quickly enough to bring the level of RecA* below the amount required to satisfy Equation 3 . The behavior of replication forks at lesions ( described above ) naturally provides a positive feedback from Pol V to RecA* because Pol V reduces the stall time at the lesion , τstalled ( [2] estimates that Pol V bypasses lesions with 100- to 150-fold higher efficiency than Pol III ) . This is illustrated in Figure 4 . Initially , there is no Pol V; however , other “nonmutagenic” translesion synthesis polymerases , Pol IV and Pol II ( DinB or PolB ) , which are always present in the cell , ensure that even in the absence of Pol V the stalled replication fork could still bypass a lesion [18] at a rate we denote . In Figure 4 , this rate is slow enough that by the time the fork reaches the next lesion ( after a time + τmoving > τRecA* ) , the first filament disassembles . At a later time , when Pol V appears , the stall time reduces dramatically [2 , 19] . The scenario depicted in Figure 4 assumes the bypass rate is dominated by Pol V–assisted bypass ( for the more general treatment used in our model , see the Methods section ) . In this case , the reduction in stall time from to when Pol V appears is sufficient to allow the replication fork to reach a second lesion before the first RecA* filament disassembles . Therefore , the RecA* level rises when Pol V appears . When this rise is fast enough , which occurs for a large enough UV dose , this results in a second peak in LexA-controlled promoter activities , as shown in Figure 5 . Thus , the second peak is a natural consequence of the mechanism for setting RecA* levels represented by Equation 1 . This prediction of the model is confirmed by the recent single-cell fluorescence experiments of Friedman et al . [10] . They also found that the second peak was washed out when the signal was averaged over many cells , probably because of cell-to-cell variations . Among the parameters , which can vary between cells , is the number of replication forks . We find that averaging the LexA-controlled promoter activity predicted by our model over many cells with differing numbers of replication forks produces a curve with a single peak ( Figure 5 , red dashed line ) as observed in the experiments . The model reveals an almost digital response of Pol V levels to UV , which provides very tight control of mutagenesis . Figure 6 shows the predictions of our model for the time course of Pol V ( UmuD′2C ) for different UV doses . In these simulations , the cell is subjected to an instantaneous pulse of UV at the specified dose at time zero . The main features of this plot are: ( 1 ) the existence of a UV dose ( about 17 J/m2 ) below which the Pol V level is very low . Thus , with low damage , mutagenesis is virtually absent and DNA repair is error-free; ( 2 ) a sharp onset in the generation of Pol V at about 15–35 min for UV doses larger than 17 J/m2 . The time of onset is largely UV-independent at high doses; ( 3 ) a rapid turn-off of Pol V at variable times that increase with the UV dose . This plot confirms several points suggested by the analysis of the model in the previous sections . First , the existence of a minimum threshold UV dose below which no Pol V is produced is a consequence of the equations described in Figure 3 and , in particular , Equation 3 . The rapid onset and the later rapid decrease of Pol V is due to the combination of heterodimerization and the previously described positive feedback from Pol V to RecA* levels . We provide more evidence to support this conclusion in the next section . The above analysis uses a simplifying assumption that the binding between UmuD and UmuD′ is infinitely strong , so that the level of UmuD–UmuD′ heterodimer is simply given by min ( [UmuD′] , [UmuD] ) . The model can be used to examine the importance of the strength of this interaction in the mutagenesis response . Figure 7A illustrates the effect of decreasing this dissociation constant ( Kdd′ ) . It shows that a strong association is critical in setting the abruptness and positions of both the turn-on and turn-off points for Pol V . Another relevant protein–protein interaction is the binding between UmuC and UmuD′ homodimers to form Pol V ( K ) . This is one of the parameters for which experimental data are not available ( see Methods ) . However , Figure 7B shows that decreasing this dissociation constant makes the Pol V profile more “digital , ” i . e . , more step-like with the concentration being either zero or maximum most of the time . Decreasing the ClpX-dependent degradation rate of UmuD′ in the heterodimer , γdd′ , mostly delays the turn-off of Pol V without affecting its turn-on time ( Figure 7C ) . Figure 7D shows the effect of turning off the positive feedback from Pol V to RecA* . Clearly , this feedback , combined with strong heterodimerization , is a crucial ingredient in the rapid onset of Pol V . Without feedback , the Pol V level is an order of magnitude lower compared with when there is feedback . Another direct implication of Equation 1 is that the peak amount of RecA* saturates as the UV dose is increased . Indeed , as the density of lesions μ rises , τmoving = 1/μv decreases . According to Equation 1 , the RecA* level saturates once τmoving becomes much smaller than τstalled . Consequently , the height of the first peak of LexA-controlled promoter activity eventually saturates at high UV doses . During the second peak of promoter activity , the RecA* concentration rises again as τstalled drops due to the Pol V–assisted bypass of lesions . The height of the second peak also saturates , but at higher UV doses . For the parameters used in our model , the amplitude of the first peak of promoter activity reaches 90% saturation around 25 J/m2 , while that of the second peak reaches it around 48 J/m2 ( see Figure 8B . ) This prediction of our model is in agreement with the experimental data in Figure 4C of [10] , which show that the saturation of the second peak occurs at a higher UV dose than for the first peak . However , that data show the peak height averaged over a cell population . Therefore , to compare our model directly with the data , we show in Figure 8A the peak heights averaged more than 200 runs with varying Nf . The resultant peak height versus UV dose curves match the data of [10] satisfactorily with the exception of the first peak data point at 50 J/m2 , which is lower than the previous data points . One explanation could be the ambiguity in the averaging procedure because , especially at higher UV doses , the second peak may sometimes be large enough to outswamp the first one , and hence be counted as a first peak , raising the red curve . Note , however , that at the single-cell level our model will always show a monotonically increasing peak height as UV dose is increased . The behavior of our model also agrees with Figure 4A of [10] from which we conclude that the second peak of promoter activity starts to appear at a considerable frequency for UV doses between 10 and 20 J/m2 . The threshold of 17 J/m2 predicted by our model ( the same as the threshold for mutagenesis ) is consistent with this . The SOS response of bacteria to radiation is typically studied by exposing them to a very short burst of UV light and then following the repair of the DNA damage . However , in environments for which bacteria are evolutionarily adapted , there may be both short bursts of the UV radiation , similar to the experimental conditions imposed on them , as well as much longer spells of low intensity UV exposure . The latter type of perturbation might not be well-suited for in vivo experiments but is easily achievable in our in silico model . Adding a new term representing a continuous low rate of production of lesions ( see Methods ) gives rise to a stable steady state wherein the rate of NER repair equals the rate of creation of the new DNA damage . Figure 9A shows the typical response to a continuous UV dose , which is low enough that mutagenesis is never triggered; LexA and RecA* take about 60 min to reach a steady state . Experiments in which cells were exposed to continuous UV damage because of the presence of a constant amount of mitomycin C also indicate that the SOS response ( rates of LexA repressor synthesis and cleavage ) took 60 min to reach a steady state [3] , confirming this prediction of our model . We also simulated the response of our virtual cell to a pulse of UV radiation of a given integral intensity and duration that varied from 0 min to 300 min . Figure 9B separates the mutagenic and nonmutagenic regions of parameter space . Initially , the magnitude of the SOS response weakly increases with prolonging the duration of the pulse . This response was expected since very short pulses give the NER subsystem time to repair some lesions before replication forks encounter them; therefore , the average RecA* concentration is less than that for slightly longer pulses . The threshold for activating the mutagenesis subsystem reaches its minimum value for an ∼60-min pulse , and then it increases linearly throughout the duration of the pulse , indicating that the cell has reached the steady state in which mutagenesis is not triggered by the total intensity of the pulse , but rather by a sufficiently high rate of production of new lesions corresponding to a UV intensity per unit time of about 1 . 5 mW/m2 . This is an order of magnitude less than the typical solar UV intensity of 7–10 mW/m2 in Copenhagen at noon on a clear day in December . For comparison , the solar UV intensity in the tropics in similar conditions is more than 100 mW/m2 ( see http://www . temis . nl/uvradiation/UVindex . html ) .
When bacteria experience a large amount of DNA damage , their response has a mutagenic component that , it has been suggested , might afford some evolutionary advantage by altering the genome of offspring that would allow some of them to better survive high levels of the damage-inducing agents [20] . Precursors to an error-prone polymerase have also been implicated in slowing down DNA replication [21] , thereby allowing additional time for accurate repair processes to remove lesions from the DNA . This delay is immediately terminated once the error-prone polymerases are fully formed . However , this kind of evolutionary strategy would be harmful where there was no damage , or when it was sufficiently low that it could be quickly repaired by error-free mechanisms . Hence , mutagenesis must be tightly regulated . The main features of the mutagenic component of the SOS response system , according to published literature , are the following: ( 1 ) Mutagenesis is characterized by a sharp temporal onset and turn-off and threshold-like behavior as a function of UV dose . There is strong experimental evidence for this . For example , Rangarajan et al . [18] observed that in the absence of Pol II masking the effects , Pol V–assisted bypass rapidly appears about 45 min after the irradiation . Also , from Figure 4 of [21] we may conclude that the UmuD′ concentration becomes comparable to that of UmuD about 30 min after irradiation , irrespective of UV dose . This exactly matches the time at which Pol V appears in our model when UmuD–UmuD′ binding is very strong . ( 2 ) Mutagenesis gives rise to the second peak in activity of the SOS regulon . This is inferred from data in [10] that show this second peak is absent in mutants that lack UmuD or contain an uncleavable version of it . We constructed a network model of mutagenesis in the bacterial SOS response system to account for these features . Figure 10 summarizes the key aspects of the behavior of the system that emerged in our simulations . We demonstrated that strong binding between UmuD and UmuD′ is necessary for the sharp onset of mutagenesis and for its turn-off when UmuD′ again falls below UmuD ( see Figure 7A ) . Thus , initially , when levels of UmuD′ are low , almost all of the UmuD′ is sequestered in heterodimers so that no Pol V is generated . However , UmuD′ is being constantly produced by the cleavage of UmuD , whose production , in turn , is elevated due to the de-repression of its promoter . If the UV damage is large enough , eventually the concentration of UmuD′ rises sufficiently to exceed that of UmuD and allow the formation of Pol V . Additional control is afforded by the degradation of the UmuD–UmuD′ heterodimer by ClpX , which removes UmuD′ while freeing UmuD for further cleavage or dimerization . Although this degradation is not essential for the system's qualitative behavior , it substantially influences the turn-off time and rate ( Figure 7B ) . Indeed , without it , turn-off could be only realized by the reduction in UmuD cleavage rates due to DNA repair and would depend solely on the slower NER mechanism . In addition , Lon actively degrades UmuD homodimers and UmuC [17]; its physiological advantages are unclear . Including this mechanism in our model does not affect the system's qualitative behavior , provided the degradation rate is not too large . We suggested a simple mechanism by which the RecA* level can serve as a measure of the lesion density ( see Equation 1 ) . This mechanism relies on the possibility for RecA filaments to exist for some finite time after the replication fork has bypassed the lesion where the filament was created ( note that we assume that this happens whether the lesion was on the leading or lagging strand ) . This allows the replication fork to sample a stretch of DNA , thus counting the damage density that is then manifested in the RecA* level . A direct implication of this mechanism is that there is a positive feedback from the Pol V to RecA* levels ( see Figure 4 ) . The resulting temporary increase in RecA* levels due to the sudden appearance of Pol V is sufficient to explain the resurgence of the SOS response 30–40 min after irradiation , observed in the single-cell experiments of Friedman et al . [10] . In addition , this mechanism also explains their observation of saturation of the peak promoter activities , and hence RecA* levels , upon increasing the UV dose ( see Figure 8 ) . Note that the first peak in promoter activity is produced due to changes in the lesion density , and thereby τmoving , as NER swings into action , while the second peak is due to changes in τstalled , due to the action of Pol V . τmoving and τstalled both affect RecA* level in the same way , being symmetrically placed in the denominator of Equation 1 , but are influenced by different mechanisms . Of course , various parameters that we use in our model will vary from cell to cell in a population . Such stochasticity plays an important role in the observed behavior , probably only for those components that are present in low numbers in the cell . Therefore , we consider that stochasticity in the number of replication forks is likely to be the most important source of cell-to-cell variability for the SOS system . As a default we take this number , Nf , to be 2 . However , for comparing with data obtained from cell populations , we averaged several runs where Nf was allowed to vary between 1 and 3 ( see Figures 4 and 8 ) . Another component present in a relatively low concentration is UmuC , a variation of which is shown in Figure 7B . Figure 7C shows that the Pol V profile is quite sensitive to ClpX . Therefore , this might be another source of variability . As more directly observable predictions of our model , we offer the following: ( i ) Overexpression of ClpX should considerably reduce the Pol V concentration . At the other extreme , the absence of ClpX would lead to Pol V being turned off at a later time than in wild-type cells ( see Figure 7C ) . ( ii ) Overexpression of UmuC results in a flatter Pol V profile ( see Figure 7B ) , while a UmuC mutant should not be able to produce Pol V and hence should behave like the ΔUmuDC and uncleavable UmuD mutants studied in [10] . ( iii ) We find that some overexpression of UmuD ( up to a factor 2 ) , or the introduction of more UmuD before the UV pulse , causes an increase in the second peak height . Further , the peak occurs earlier , sometimes even swamping the first peak . However , removing the LexA repression of UmuD ( say by introducing UmuD on a plasmid with an unregulated promoter ) results in the vanishing of the second peak , except at particularly high UV doses , because UmuD′ is not formed fast enough to cross the UmuD level . There are alternative mechanisms by which UmuD and UmuD′ could affect LexA levels within the framework of the SOS model considered here . We discuss two mechanisms below that , unlike the mechanism we have concentrated on so far , could produce a second peak by causing a temporary rapid decrease of LexA , i . e . , a trough in the average promoter activity profile . ( 1 ) UmuD competes with LexA for the RecA* binding sites . Conceivably , high UmuD levels could prevent the access of LexA to them , thereby reducing its cleavage rate . This would create a trough ( not a peak ! ) in LexA-controlled promoter activity at the peak of UmuD concentration . An important observation in [10] requires this mechanism of competition: peak LexA-controlled promoter activity in the ΔUmuDC mutant appears to increase with increasing UV dose ( in the range 20–35 J/m2 ) rather than saturating , as in wild-type cells . Because of the absence of competition in the ΔUmuDC mutant , LexA is cleaved more and falls to a lower concentration than in wild-type cells , which , in turn , leads to a higher peak activity level of LexA-repressed promoters . However , if this competition is an important effect , the ΔUmuDC and uncleavable UmuD mutants studied in [10] should have different peak heights because the former would have no competition , while the latter should have more than wild-type ( since the uncleavable UmuD K97A retains the ability to bind to the RecA filament [22] ) . ( 2 ) UmuD was shown to preferentially bind to the beta-clamp subunit of DNA polymerase III while UmuD′ prefers to bind to the epsilon subunit [12] . Thus , by sequestering the beta-clamp , large levels of UmuD could possibly reduce the processivity of Pol III , which feeds back onto RecA* and LexA levels via τmoving in Equation 1 . Once again , the effect of this would be to produce a trough in LexA-controlled promoter activity . Later when the UmuD levels drop and the processivity of Pol II increases , it could lead to a second peak . To explore these postulated causes of the second peak in RecA* levels , we incorporated each of these two feedback mechanisms into our model . In the absence of Pol V → τstalled feedback , these mechanisms , alone or in combination , did not generate the second peak in promoter activity at a reasonable time ( within 40–50 min ) after irradiation . However , their presence did not interfere with the manifestation of this feature when the Pol V → τstalled feedback was included in the model . Therefore , they might well be operating in parallel . Friedman et al . also reported the existence of a third peak in the LexA-controlled promoter activities [10] . Unlike the second peak , the third one exhibited fewer fluctuations between individual cells . Indeed , its existence was previously mentioned in the Ronen et al . study , which used a signal averaged over many cells [23] . The generation of this third peak requires a mechanism that would increase the amount of RecA* at about 100–120 min after UV irradiation . As Friedman et al . suggest , one possible candidate is DinI , an SOS gene that is also repressed by LexA and induced in response to DNA damage . DinI is known to ( a ) stabilize already-formed RecA* filaments , ( b ) prevent RecA*-mediated cleavage of UmuD , and ( c ) leave the RecA*-mediated cleavage of LexA unaffected [22 , 24] . The first property would cause an increase in τRecA* and thus if DinI were to be generated , or to become sufficiently active , about 100–120 min after the initial damage , it would result in a rise in RecA* levels and a new peak in LexA-dependent promoter activity . Yoshimasu et al . [25] suggest that DinI coats RecA* filaments with a 1:1 stoichiometry . Therefore , its activity would become substantial only when its levels exceeded the RecA* levels in the cell . Since the RecA levels are high ( ∼7200 [4] ) even in the absence of damage , and increase further due to de-repression of the SOS regulon , it might take up to 100–120 min until the induced DinI levels would overtake the diminishing levels of activated RecA* . This speculation is corroborated by co-IP results presented in Figure 2B of [26] . A plausible evolutionary role for such delayed RecA* stabilization is that it would support a stable low-level SOS response when the cell is exposed to a persistent source of DNA damage . Overall , our model systematizes causes and effects in the best-known parts of the SOS response system in E . coli . It provides a framework for asking new questions about how ( and why ) the SOS response is organized . For instance , why is mutagenesis first initiated at such a late stage when only 10%–20% of the original lesions remain untouched by NER ? One hypothesis would be that mutagenesis is triggered in response to the presence of particular types of lesions that are less efficiently bypassed by other mechanisms . This can be tested by extending the model to incorporate different kinds of lesions with different feedback to the stall time . Another key aspect is the length of the time window during which Pol V is active , which is set by the degradation times of UmuD , UmuD′ , and UmuC . A more accurate determination of the regulation of these degradation times may shed light on effects of memory and why mutagenesis is at all initiated for severe DNA damage . In discussing this , however , one should keep in mind that mutagenesis may be designed to work primarily under a continuous source of DNA damage , and that the timing effects that we use to gain insight into the dynamics of the SOS response may be of secondary importance under typical real-world stresses . In any case , our study of the mutagenesis subnetwork suggests that its behavior is quite “digital , ” in the sense that it makes a very quick transition from a state where there is no mutagenesis to a state where Pol V is fully activated .
Our model of the SOS response defined by Equations 4–12 , below , builds upon excellent earlier work by Aksenov [11] . Equations 4 , 6 , and 7 are very similar to their counterparts in [11] , while Equations 5 and 8–12 are the new ones that we propose to describe mutagenesis and its feedback onto RecA* and LexA levels . The dynamics of LexA ( l ) level is modeled using the following equation: Here , the first term models the self-repression of LexA production . We assume a Hill coefficient of 1 [23] . The second term is for the cleavage of LexA by RecA* ( whose level is denoted by r* ) , while the third term is the degradation of LexA in nonirradiated cells . The RecA* level ( r* ) , in turn , is described by which is exactly the same as Equation 1 in the main text . In writing this equation for the RecA* level , we assume that the timescales involved in filament assembly and disassembly [13] are much smaller than those of transcriptional regulation and NER repair; therefore , a differential equation is not needed to describe the dynamics of RecA* . Denoting the density of lesions by μ and the speed with which Pol III moves on undamaged DNA by v , we get the following expression for τmoving: The density of lesions is not a constant; they are continuously being repaired by the NER mechanism , which we model as follows: Here we assume that the repair is limited by the number of lesions and not by the Uvr proteins , hence the repair rate is proportional to the lesion density . In taking the rate of repair per lesion , λ , to be a constant , we ignore the feedback from LexA to mRNA production from uvr genes; following [11] we assume that the repair is limited by UvrC , which is not repressed by LexA . We can write the overall lesion bypass rate 1/τstalled as the sum of two rates , bypass due to Pol V ( ) , and bypass due to all other mechanisms ( ) . Here P is the Pol V level and P/ ( Kp + P ) is the factor describing the concentration dependence of the Pol V–assisted bypass rate . This generalizes the situation described in Figure 4 . To describe the dynamics of the Pol V concentration , we must first consider concentrations of free UmuD ( u ) , free UmuD′ ( u′ ) , and the UmuD–UmuD' heterodimer uhetd: Finally , the Pol V level , which feeds back to RecA* and LexA levels via Equation 8 , is given by These are the equations shown and explained in Figure 3 . Our model is fully specified by 18 parameters . Below we divide them into two groups , the first whose values we could fix directly or indirectly from published literature , the second for whose values there is inconclusive or no published data . Parameters fixed from experimental data . ( 1 ) The repair rate by NER: λ = 0 . 035 min−1 , corresponding to a half-life of approximately 20 min as reported in [27] for cyclobutane pyrimidine dimers . ( 2 ) LexA concentration required for half-repression of LexA promoter: Ki = 270 nM , corresponding to an induction ratio of approximately 5 . 8 ( the relation between the two is K = 1 , 300/ ( I − 1 ) , where 1 , 300 nM is the LexA level in undamaged cells [4] ) , interpolated from induction ratios of 6 . 7 and 4 . 8 , measured at 30 °C and 42 °C , respectively [15] . ( 3 ) LexA concentration required for half-repression of UmuD promoter: Ku = 60 nM , corresponding to an induction ratio of approximately 22 . 7 , interpolated from induction ratios of 28 and 17 measured at 30 °C and 42 °C , respectively [15] . ( 4 ) Dilution rate: γdil = 1n ( 2 ) /60 = 0 . 012 min−1 , estimated average from a scatter plot of cell doubling time in [10] . Further , 60 min is the reported half-life of LexA in nonirradiated cells treated with chloramphenicol , i . e . , in the absence of production of LexA [4] , as well as from pulse-labeling measurements of LexA cleavage rates [3] . ) ( 5 ) Speed of Pol III on undamaged DNA: v = 1 , 000 bp s−1 [28] . ( 6 ) UmuC level: C = 200 nM [29] . ( 7 ) Number of lesions per unit UV dose: 50 per J/m2 . This corresponds to 250 lesions per E . coli genome for a dose of 5 J/m2 [30] . Thus , for E . coli the initial lesion density is given by μ ( t = 0 ) = 10−5 × D , where D is the UV dose in J/m2 . ( 8 ) Stall time in absence of Pol V: = 0 . 22 min . In [10] , the height of the first peak in LexA-repressed promoter activities reaches half-maximum at about 7–8 J/m2 . In our model , this occurs when τstalled = τmoving . At 7–8 J/m2 , μ = 7 − 8 × 10−5 hence τmoving ≈ 0 . 22 min and τstalled ≈ because the Pol V level is negligible . The chosen value is also consistent with an estimate of 10–12 s based on measurements of DNA synthesis in irradiated cells [30] . ( 9 ) Stall time in the presence of Pol V: = 0 . 022 min . Tang et al . [2] estimate that Pol V bypasses lesions with 100- to 150-fold higher efficiency than Pol III . However , in vivo the ratio of to will be smaller than that because , apart from Pol III , other polymerases such as Pol II also contribute to the bypass rate [18] . Szekeres et al . [19] find that in ΔumuDC cells , the frequency of replication past cis–syn T–T dimers , produced by 4 J/m2 irradiation , is approximately 40× lower than in cells containing a chromosomal copy of umuDC . In our simulations , we find that as long as , the dynamics do not depend much on the value of ; therefore we conservatively chose = 10 ( 10 ) Parameter determining RecA* levels: NfLRecA*τRecA* = 110 nM min . ( 11 ) LexA cleavage rate ( per nM of RecA* ) : γu = 8 . 8 × 10−4 nM−1 min−1 . These two parameters are together fixed so that the maximum rate of LexA degradation is corresponding to a half-life of about 1 . 5 min , chosen to match pulse-labeling measurements of LexA degradation rates in irradiated cells [3] and measurements in cells where LexA production was prevented by adding chloramphenicol [4] . Note that in our equations only the product of these two parameters , NfLRecA*τRecA* × γl , appears , so for the model simulations the individual value of each parameter is irrelevant as long as the product is preserved . However , we chose NfLRecA*τRecA* = 110 nM min to fix the maximum possible RecA* level , to 500 nM . This is a reasonable number , obtained by assuming that Nf = 2 and the maximum RecA* level is achieved when both these replication forks are stalled . Then , with each fork leaving an ssDNA gap of 900 nucleotides [11] , and given that each RecA* filament has one monomer per 3–5 nucleotides [4 , 5] , we obtain a maximum RecA* level about 500 nM . ( 12 ) UmuD cleavage rate ( per nM of RecA* ) : γu = 1 . 8 × 10−4 nM−1 min−1 . This was chosen to be approximately five times slower than the LexA cleavage , estimated from the following data: After a UV dose of 20 J/m2 , the half-life of the UmuD to UmuD′ cleavage is approximately 45 min [12] , while at the same UV dose , [4] reports a half-life for LexA of about 7–10 min . ( 13 ) Maximal LexA production rate: βl = 86 nM min−1 ( chosen so that in undamaged cells the level of LexA stabilizes to 1 , 300 nM [4] , for which it must satisfy the formula βl = 1 , 300 × γdil × ( 1 + 1 , 300/Kl ) . ) ( 14 ) Maximum UmuD production rate: βu = 47 nM min−1 ( chosen so that in undamaged cells the level of UmuD stabilizes to 180 nM [29] for which it must satisfy the formula βu = 180 × γdil × ( 1 + 1 , 300/Ku ) . ) ( 15 ) Constant involved in concentration dependence of Pol V–dependent bypass: Kp = 10 nM . The level of Pol V in irradiated cells ranges from 15–60 molecules for small UV doses [31] . We chose Kp such that this range of Pol V level will produce a substantial , but not saturating , contribution to τstalled . For parameters where data were unavailable or inconclusive , we have scanned a range of values around the following chosen defaults: ( 16 ) Binding constant of the UmuD–UmuD′ heterodimer: Kdd′ = kb/kf = 0 . 01 nM , chosen to be very strong ( kb and kf were individually chosen to be relatively large so that the heterodimer was , in practice , always in equilibrium with unbound UmuD and UmuD′ ) . ( 17 ) Degradation of UmuD–UmuD′ by ClpX: γdd′ = γdil = 0 . 012 min−1 . ( 18 ) UmuD′ level required for Pol V level to reach half maximum: K = 10 nM . For initial conditions , we have used the experimentally reported levels in wild-type cells: LexA = 1 , 300 nM [4] , RecA* = 0 ( naturally existing ssDNA , e . g . , lagging strand replication gaps , does not activate RecA in the absence of DNA damage [4] ) , UmuD = 180 nM and UmuD′ = 0 [29] . First , adding Equations 9 and 11 , we get Here , utot = u + uhetd is the total amount of UmuD . Similarly , adding Equations 10 and 11 and using the fact that when heterodimer binding is infinitely strong , uhetd = min ( utot , u′tot ) , we get Equation 2 in the main text To model the dynamics during exposures to pulses of UV radiation of finite ( possibly long ) duration , we modified Equation 7 as follows: where s is a source term , which is a nonzero constant when 0 ≤ t ≤ td . Here , td is the duration of the UV pulse and s × td is the total integral UV dose .
|
Ultraviolet light damages the DNA of cells , which prevents duplication and thereby cell division . Bacteria respond to such damage by producing a number of proteins that help to detect , bypass , and repair the damage . This SOS response system displays intricate dynamical behavior—in particular the tightly regulated turn-on and turn-off of error-prone polymerases that result in mutagenesis—and the puzzling resurgence of SOS gene activity 30–40 min after irradiation . In this paper , we construct a mathematical model that systematizes the known structure of the SOS subnetwork based on experimental facts , but which remains simple enough to illuminate the specific functional role of each regulatory process . We can thereby identify the interactions and feedback mechanisms that generate the on–off nature of mutagenesis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"eubacteria",
"computational",
"biology"
] |
2007
|
UV-Induced Mutagenesis in Escherichia coli SOS Response: A Quantitative Model
|
House dust mites are common pests with an unusual evolutionary history , being descendants of a parasitic ancestor . Transition to parasitism is frequently accompanied by genome rearrangements , possibly to accommodate the genetic change needed to access new ecology . Transposable element ( TE ) activity is a source of genomic instability that can trigger large-scale genomic alterations . Eukaryotes have multiple transposon control mechanisms , one of which is RNA interference ( RNAi ) . Investigation of the dust mite genome failed to identify a major RNAi pathway: the Piwi-associated RNA ( piRNA ) pathway , which has been replaced by a novel small-interfering RNA ( siRNA ) -like pathway . Co-opting of piRNA function by dust mite siRNAs is extensive , including establishment of TE control master loci that produce siRNAs . Interestingly , other members of the Acari have piRNAs indicating loss of this mechanism in dust mites is a recent event . Flux of RNAi-mediated control of TEs highlights the unusual arc of dust mite evolution .
House dust mites are ubiquitous inhabitants of human dwellings , and are the primary cause of indoor allergy [1] . Dust mites have an unusual evolutionary history , descending from a parasitic ancestor [2] . Parasite genomes are typically highly modified; possibly to accommodate genetic novelty needed to productively interact with a host [3 , 4] . The sequence of events leading to adoption of a parasitic lifestyle may require a period of genomic crisis to yield the rewired parasite genome . Dust mites represent an extreme case potentially experiencing a second round of genomic change to reacquire a free-living ecology . Transposable element ( TE ) activity is a major source of genome instability [5 , 6] . Silencing of TE activity in multicellular organisms is commonly achieved by RNA interference ( RNAi ) -based mechanisms , which employ small RNAs associated with Argonaute/Piwi ( Ago/Piwi ) proteins to target TE transcripts [7] . In many animals , the Piwi-associated RNA ( piRNA ) pathway is the primary RNAi-based defense [8 , 9] . In arthropods and vertebrates piRNAs are recognized as being roughly 23–32 nucleotides ( nt ) long , and unlike other small RNAs , such as microRNAs ( miRNAs ) and small-interfering RNAs ( siRNAs ) , they are not excised from double-stranded RNA ( dsRNA ) precursors by the RNase III enzyme Dicer [10] . piRNAs in Drosophila are generated in two collaborative pathways: Phased cleavage of transcripts by the RNase Zucchini ( Zuc ) and a “ping-pong” mechanism involving direct cleavage by Piwi proteins [11] . Ago/Piwi proteins may possess “slicer” activity which cuts transcripts base-paired with a bound small RNA 10 nt from the 5’ end of the small RNA [12] . Zuc-dependent piRNA biogenesis is initiated by Piwi-mediated slicing of targeted transcripts , which propagates in a 5’-3’ direction from the site of scission [13 , 14] . These piRNAs feed into the ping-pong where Piwi proteins collaborate to capture fragments of TEs and convert them to new piRNAs [15 , 16] . This leads to further production of Zuc-dependent piRNAs in an amplifying system [17] . As TE transcripts processed through the ping-pong pathway are products of slicing they exhibit 10 nt 5’ overlaps with cognate , antisense piRNAs [15] . In contrast , Zuc-dependent piRNAs are derived from single stranded RNA precursors , and while this process has been found to be dependent on initial slicing there are Drosophila cell types in which the ping pong system is absent that initiate zuc processing through the factor Yb [18] . Another feature of piRNA-mediated genome surveillance is the involvement of piRNA cluster master loci as sites of Zuc-dependent piRNA production [19] . These loci are composed of TE fragments , serving as catalogs of restricted sequences . Loss of master loci integrity compromises TE repression and causes sterility . Nematode piRNAs , while possessing a related role in controlling TEs , differ in that they are short ( 21nt ) cleavage products of small discrete transcripts [20] . Despite these differences in biogenesis , piRNAs in both species typically exhibit an “U” residue at the 5’ terminus . The exception is some ping-pong piRNAs , which instead have an “A” at the tenth position . While piRNA regulation of TEs is common in animals , it has been lost in several nematodes and platyhelminths [21 , 22] . In the nematode species , alternative mechanisms restrict TE mobilization involving Rdrp ( RNA dependent RNA polymerase ) and Dicer . Conversion of TE transcripts by Rdrp into dsRNA substrates of Dicer results in siRNA generation . Ago proteins then associate with nascent TE transcripts , recruiting chromatin modulators including DNA methyltransferase . This process , RNA-induced transcriptional silencing ( RITS ) is common in plants and fungi [23–25] . RITS-like mechanisms are found in animals as nuclear localized Ago and Piwi proteins can influence chromatin biology [26 , 27] . However , outside nematode clades , amplifying RITS mechanisms involving siRNAs have not been observed in vertebrates or other ecdysozoans–potentially due to absence of Rdrp [28] . One possible exception is chelicerae arthropods , a lineage where dust mites belong , which possess Rdrp proteins . RNAi pathways in chelicerates appear complex as they have both Rdrp and Piwi class Argonaute proteins , both of which appear to have roles in controlling TE’s [29–31] . Here we investigate the status of small RNA pathways in the dust mite to understand how RNAi biology might be structured in this highly-derived organism .
We obtained a genome sequence for the American house dust mite Dermatophagoides farinae using Illumina and PacBio platforms . The HGAP pipeline was used through PacBio SMRT analysis portal to filter and assemble PacBio reads , which resulted in 1 , 828 contigs producing a total length of 93 , 777 , 723 bp [32] . Then , Illumina reads were used to connect and extend the PacBio contigs using SSPACE scaffolding , which produced a 93 , 804 , 520 bp assembly in 1728 scaffolds [33] . After removal of bacterial contamination , the final contig number was reduced to 1706 with a N50 read length of 19 , 371 . The assembled and filtered final genome was ~92 Mb compared to a 53 Mb genome that was previously reported [34] . Using mRNA-seq datasets we annotated ~18 , 500 transcripts through the Cufflinks program [35] . 47% of the genic transcripts exhibited similarity to S . scabiei and/or D . melanogaster protein coding genes or to the NCBI conserved domain collection ( Materials and Methods ) ( S1 Table ) [34 , 36] . Ago/Piwi proteins were identified in the D . farinae genome using RNA-seq annotations and amino acid sequences of seven Ago and six Piwi proteins from Tetranychus urticae–the closest relative of D . farinae with a high-quality genome and experimentally supported annotations [29] . Eight confident Ago homologs were found ( Ago1-GenBank ID: KY794591 , Ago2-GenBank ID: KY794592 , Ago3-GenBank ID: KY794593 , Ago4-GenBank ID: KY794594 , Ago5-GenBank ID: KY794595 , Ago6-GenBank ID: KY794596 , Ago7-GenBank ID: KY794597 , Ago8-GenBank ID: KY794598 ) . Ago proteins from T . urticae , D . melanogaster , C . elegans , and Ascaris suum were compared to D . farinae Agos using amino acid sequences of Paz , Mid , and Piwi domains ( Fig 1A ) . Our phylogenetic analysis recovered two Ago family members likely involved in miRNA ( DfaAgo1 ) and siRNA ( DfaAgo2 ) pathways [37] . The remainder belong to a divergent clade specific to dust mites ( DfaAgo3-8 ) . Surprisingly , none of the Agos from D . farinae belong to the Piwi clade . We examined D . farinae Agos for the presence of slicer motifs . The DEDH slicer motif , which is common in metazoan Ago and Piwi proteins , was found in DfaAgo1 ( miRNA ) and DfaAgo2 ( siRNA ) . The divergent Agos have an uncommon DEDD catalytic motif ( S2 Fig ) . Orthologs containing a DEDD motif can be found in scabies ( S . scabiei ) , social spiders ( Stegodyphus mimosarum ) , and in C . elegans Ago family members of unknown function; which emphasizes the unusual nature of this Ago clade [36 , 38 , 39] . A dust mite small RNA library of nearly 400 million reads was generated to investigate whether piRNA-class small RNAs could be identified ( S1 Text ) ( S1 Fig ) . To accommodate the repetitive nature of piRNA targets all mapping events were captured for reads that mapped fewer than 100 times . An overall rate of ~80% mapping was observed with ~0 . 69% discarded due to mapping >100 times ( S1 Fig ) . Next an algorithm that determines small RNA read overlap probabilities in mapping data was used to characterize biogenesis of dust mite small RNAs [40] . When applied to either all mapping or mapping in discreet size ranges no clear bias for 10nt overlapping reads was uncovered , showing an absence of ping pong processing ( Fig 1B ) . Instead a strong signal seen in a register 2nt shorter than the length of read sizes . This is congruent with 2nt overhangs left by Dicer cleavage . Overlaps seen in dust mites starkly contrast with those seen in spider mites and Drosophila . In spider mites , ping pong signatures could be seen in longer reads ( 23-28nt ) and the dicer-associated 2nt register in shorter ( 20-22nt ) reads ( Fig 1C ) . Likewise , in RNAs sequenced from Drosophila female bodies a prominent ping pong signature is evident ( Fig 1D ) . siRNA processing was not evident when considering whole genome mapping , but could be seen in a group of Drosophila IDEFIX retroelements that had biased mapping of 21nt RNAs ( S3 Fig ) ( S3 Table ) . Drosophila endo-siRNAs are a relatively small proportion of total small RNAs , and are frequently produced from inverted repeat loci which are not captured by the overlap probability calculation used here [41] . Moreover , this highlights a correlation between the presence of Rdrp in spider mites and an expanded population of Dicer products . Together this shows a dramatic departure in the composition of dust mites small RNA populations relative to those in the Piwi protein possessing spider mite . The difference is even more stark when comparing dust mites to the more distantly related fruit fly . The configuration of RNAi in spider mites is likely ancestral due to the similarities to Drosophila , which is supported by clear orthology of spider mite Piwis to distinct Drosophila ping-pong partners: dmeAgo3 and Piwi/Aub ( Fig 1A ) . Thus , RNAi pathways have diverged in dust mites and appear to be dominated by siRNA-like Dicer products , and lack the signature of amplifying ping pong piRNAs [42] . To functionally characterize dust mite small RNAs , we sought to identify genomic loci that generate and/or are targeted by these transcripts ( S1 Fig ) . To achieve this , annotation of the dust mite genome was extended to find ncRNAs and repetitive elements using Repeatmasker . Additionally , non-miRNA , small RNA producing loci were annotated that exhibited >1000 read density and were longer than 200nt . The identities of regions were determined using blast2go [43] . Nearly a third of the loci were rRNA or mRNA . The remainder showed homology to either TEs or lacked similarity to known sequences . Together this permitted segmentation of the dust mite genome into mRNA , TE , rRNA , tRNA , snRNA , and unknown small RNA-mapping loci ( S2 Table ) . Small RNA reads were then mapped to these regions using multiple mapping conditions described above , as well as unique mapping . To ensure multi-mapping events were specific to loci groups , datasets were cleaned before mapping by removing reads that mapped to non-target genomic features ( S1 Fig ) . Multi-mapping alignments showed considerable enrichment at TEs relative to other classes , consistent with their repetitive nature ( Fig 2A ) . Both multi- and uniquely mapping TE reads also exhibited lower strand bias with only a single locus showing 100% bias after unique mapping ( S4 Fig ) . This is consistent with processing from dsRNA . Higher bias was seen at other loci , suggesting that some mapping events may be due to capture of RNA degradation fragments and not functional small RNAs . This was supported when overlap probabilities were calculated; which , with exception of TEs and mRNAs , did not show consistent processing signatures ( S5 Fig ) . This includes the unknown loci , suggesting these transcripts may be degradation products of uncharacterized ncRNAs and are not generally siRNA or piRNA class small RNAs . Closer inspection of per locus overlaps did show Dicer processing at a minority of loci ( S6 Fig ) . There was no clear ping pong processing at unknown loci . Small RNA mapping coverage was calculated per locus to understand siRNA production from TEs and mRNAs ( Fig 2B ) . On average , small RNA coverage was even across TE loci , while mRNAs had greater coverage at transcript 3’ ends . This pattern at mRNA loci is suggestive of cis-NAT siRNAs [44] . Depth of coverage at TE loci varies , showing that active targeting is occurring at a subset of loci . The absence of purely single-stranded small RNA producing loci that have homology to TE sequences suggests that dust mites also lack a Zuc-dependent piRNA-like pathway that is involved in genome surveillance . This does not rule out the existence of dual strand piRNA clusters; however , piRNAs produced from these loci are found to participate in the ping pong cycle , which we did not observe [45] . These data suggest that the piRNA pathway has been lost in dust mites and control of TE’s is likely under the purview of a siRNA-like pathway . To investigate the role of dust mite small RNAs in genome surveillance we compared the biogenesis of TE-associated small RNAs to those found in spider mites . The size distribution of genome-aligned dust mite small RNAs is unimodal with a peak at 24nt , versus a bimodal distribution in spider mites ( Fig 3A ) . When TE-mapping reads are examined , the 24nt sized RNAs in dust mite were enriched by 10% , while in spider mites only larger size range RNAs were found ( Fig 3B ) . In other locus groups , less read size bias was observed , consistent with the heterogeneity seen in strand bias and read overlap probabilities , further reinforcing that generally non-TE loci do not produce small regulatory RNAs ( S7 Fig ) . Next we looked at the 5’ nucleotide bias and found that dust mites TE siRNA reads have an equal prevalence of T and A residues versus spider mites where there was striking over representation of T ( Fig 3C ) . Then we examined per locus read size distribution and overlap probabilities to assess whether Dicer processed ~24 nt small RNAs are common across dust mite TE loci ( Fig 3D ) . All loci exhibited mapping of predominantly 24 nt reads , and in the most prevalent size ranges ( 23-26nt ) a clear pattern of overlaps could be seen that is consistent with Dicer processing ( Fig 3D ) . This contrasts with a similar analysis in spider mite where a ping pong signature was seen across all TEs . Together this suggests siRNAs are the main RNAi-based mode of controlling TEs in dust mites , accommodating the apparent loss of piRNAs . This is a clear departure from spider mites where stereotypical piRNAs target TEs . In the D . farinae genome we found three Dicers ( DfaDcr1-GenBank ID: KY794588 , DfaDcr2-GenBank ID: KY794589 , DfaDcr3-GenBank ID: KY794590 ) . DfaDcr1 is a close ortholog of Arthropod miRNA-producing dicer ( S8 Fig ) . The other two Dicer proteins are related to family members in other mites and lophotrochozoans , and are unrelated to Arthropod Dicer2 or nematode Dicer ( S8 Fig ) . Unexpectedly , DfaDcr1 possesses an ATP binding helicase domain , which is implicated for processing of long dsRNA ( S9 Fig ) [46] . The more divergent Dicers , DfaDcr2 and DfaDcr3 , lack both DUF283 and dsRNA binding domains , and have divergent PAZ domains ( S9 Fig ) [46–48] . Together this suggests that mites , and possibly other chelicerates , possess ancient Dicer biology present in basal protostomes that was lost both in nematoda and pancrustacea ( insects and crustaceans ) . To verify whether TEs are controlled by Dicer-produced siRNAs we sought to inhibit the activity of dust mite Dicer proteins . To generate loss of Dicer function we elicited RNAi against each Dicer by feeding mites cognate dsRNA ( Fig 4 ) . Dust mites tolerate being soaked for several hours in aqueous solution , which they can be observed to ingest after 30 mins ( Fig 4A ) . Small RNAs ( 20-27nt ) derived from dsRNA can be recovered from soaked mites ( Fig 4B ) . Knockdown of target genes can also be observed ( Fig 4C–4K ) . Depletion by RNAi of each DfaDcr protein resulted in derepression of multiple TEs ( Fig 4L ) ( S10 Fig ) . A strong effect was seen with loss of DfaDcr1 and DfaDcr2 function . The presence of processive helicase activity in DfaDcr1 suggests that long dsRNAs could be substrates . This combined with the lack of dsRNA binding motifs in DfaDcr2/3 suggests DfaDcr1 has a unique capacity to process dsRNA ( S9 Fig ) , and therefore it is unsurprising that it has a significant role in the control of TEs ( Fig 4L ) . Loss of DfaDcr2 showed a greater effect on TE expression compared to DfaDcr3 . How these atypical Dicer proteins function is unclear; however , residues in the DfaDcr3 PAZ differ significantly from those in DfaDcr2 PAZ suggesting non-overlapping roles in the metabolism of dust mite small RNAs ( S9 Fig ) . These results are consistent with reports that psoroptid mites are sensitive to dsRNA soaking , resulting in gene knockdown [49 , 50] . Investigation of RNAi in dust mites revealed loss of the piRNA pathway and replacement by siRNAs . This is similar to observations in nematodes and flatworms [21 , 22] . The loss of piRNA activity in dust mites , nematodes , and possibly in flatworms may be tolerated due to compensation by amplifying siRNAs produced by Rdrp [21 , 51] . The collective function of dust mite Rdrps , however , appears to be distinct from nematodes , as only processive versions are present , suggesting the de novo siRNA pathway may not be present in mites ( S11 Fig ) . Substantial Rdrp activity does appear to be present in dust mites; dsRNA soaking results in elevation of target mRNA when reverse transcription is carried out with random hexamers ( Fig 4E , 4G , 4I and 4K ) but not oligo dT ( Fig 4D , 4F , 4H and 4J ) . Increase of transcript abundance was not due to the presence of ingested dsRNA as the region cloned to generate dsRNA was distinct from the qPCR amplicon ( S10 Fig ) . Random priming will capture Rdrp products , while oligo dT will only hybridize to the initial transcript . For all the genes tested an elevation of cognate transcripts could be observed after random priming that were poorly recovered from Oligo dT primed cDNA . Dust mites differ from nematodes that lost piRNAs in the organization of siRNA producing loci . A key feature of piRNA biology is the cataloging of restricted sequences into master loci . In nematode lineages lacking piRNAs , master loci also appear to be absent [21] . This is not the case in dust mites ( Fig 5A ) . Three loci were discovered that span 62 kb , contain sequences from multiple varieties of TEs , and exhibit homology to 70% of TE mapped small RNAs ( Fig 5B ) . Two of the loci , ML-283 and ML-95 , appear to be generated by duplication; however , some sequence divergence indicates they are distinct loci . Similar regions could not be found in the S . scabiei genome [52] . Though , poor conservation is a characteristic of piRNA master loci [53] . The dust mite loci appear to be generated from a dsRNA precursor as both strands of the loci show similar rates of read mapping ( Fig 5A ) . We found a tendency for 2nt overhangs along with little evidence for nucleotide bias ( S12 Fig ) . The loci were inspected for common motifs using the meme suite [54] . Motifs recovered were primarily simple repeats with none being shared between loci suggesting dust mite master loci don’t possess elements like the Ruby motif which is central to directing piRNA transcription in C . elegans [55] . Following knockdown of each of the individual dust mite Dicers significant ( >80% ) reduction in siRNAs exhibiting homology to these regions was observed , indicating a dependence on the activity of all dust mite Dicers for biogenesis ( Fig 5C ) . Detection of the siRNAs was accomplished with a combination of oligonucleotide probes complementary to sites of highest small RNA density in the three master loci ( S1 Text ) . They also have homology to other regions of the genome , specifically TEs . Thus , the Dicer sensitive siRNAs include master loci derived primary siRNAs and potentially secondary siRNAs generated from processed TE transcripts . This is consistent with loss of TE control after knockdown of each Dicer ( Fig 4L ) . However , there is a clear difference in the magnitude of TE expression , which may point to roles for dust mite Dicer proteins outside the production of siRNAs and to involvement in targeting of TE transcripts . This could be similar to limiting of latent viral infection by Drosophila Dcr2 [56] . Next , we sought to characterize terminal moieties of master loci associated siRNAs through biochemical tests to gain greater insight into their biogenesis ( Fig 5D and 5E ) . The primary goal was to determine if the siRNAs had characteristics of Dicer cleavage: 5’-monophosphates and 3’-OH groups . β-elimination showed a shift to a lower molecular weight indicating an unmodified 2’OH; therefore , unlike Drosophila Ago2 endo-siRNAs or C . elegans Prg-1 associated small RNAs , dust mite siRNAs are not 2’-OH methylated ( 2’OMe ) ( Fig 5D ) [57 , 58] . Next , we identified groups on 5’ ends of small RNAs using the 5’ monophosphate specific terminator ribonuclease . After treatment , a 50% reduction in siRNAs could be observed ( Fig 5E ) . Degradation by terminator could be abrogated by prior treatment with calf intestinal phosphatase ( CIP ) . There is a noticeable lag in siRNA gel migration following CIP treatment , which is consistent with removal of 5’ phosphate groups and loss of charge . These results also reinforce the absence of a de novo siRNA pathway . Small RNAs produced by non-processive Rdrps in C . elegans have 5’ triphosphate groups . While treatment with terminator did not completely eliminate siRNAs there was no observable change in migration . If the remaining small RNAs were spared due to the presence of trisphosphate groups there would be shift towards a smaller molecular weight , relative to untreated . Together , dust mite master loci associated siRNAs appear to be Dicer products arising from a dsRNA precursor , possess the expected 5’-monophosphate , but differ from insect endo-siRNAs due to the absence of 2’-OMe groups . We were able to identify a dust mite gene with similarity to Hen1 methyltransferase proteins; however , inspection of potential open-reading frames revealed the absence of a common motif involved in recognition of 2 nt 3’ overhangs characteristic of Dicer products ( S12 Fig ) . This likely explains the lack of 2’-OMe groups on dust mite siRNAs . Extent of DNA methylation in CG widely varies across insect clades and can be as high as 40% in roaches , while other groups , like flies , show little evidence for this modification [59] . Here we investigated whether this epigenetic control mechanism is a component of TE control in dust mites , as the genomes of nematodes and platyhelminths that lack the piRNA pathway are frequently modified by cytosine methylation [21 , 60] . Dust mites differ from these organisms , as evidence for this modification seems minimal and it is not enriched at TE loci ( Fig 6A ) . Indeed , bisulfite sequencing showed potential CG and CHG methylation is underrepresented in TE sequences , despite these sites occurring at the same rate as other genomic loci . Furthermore , the overall rate of DNA methylation ( 0 . 5% ) was very low , suggesting this base modification is not a major feature of dust mite chromatin regulation . Moreover , we found a single DNA methyltransferase in the D . farinae genome , a Dnmt1 homolog ( Fig 6B and 6C ) . It is likely a pseudogene as it appears to be truncated and shows little evidence of expression . This further highlights the distinct , derived nature of small RNA-mediated genome surveillance in dust mites .
This work provides insight into the elaborate nature of RNAi in chelicerates , many of which appear to have both Piwi proteins and Rdrps [29 , 30 , 39] . Loss of the piRNA pathway in dust mites probably occurred in the parasitic ancestor . Inspection of the scabies mite genome similarly failed to uncover Piwi proteins ( S13 Fig ) [36] . Members of the divergent dust mite Ago family; however , were found . Indeed , a deeper inspection of scabies mite RNAi factors uncovered further similarities to dust mites ( Table 1 ) . Thus , absence of the piRNA pathway in dust mites is likely a consequence of descending from an ancestor that underwent dramatic genome changes , potentially during the acquisition of a parasitic life style . This highlights plasticity of RNAi pathways and how clade-specific biology might impact evolution of RNAi technologies . Dust mites exhibit a highly distinct RNAi biology , possessing both novel and ancient effectors that haven’t been studied in popular ecdysozoan model organisms . Indeed , there seems to be wholesale changes to the small RNAome of these organisms . Dicer produced siRNAs are an unusually common feature of the dust mite small RNA populations , comprising approximately three-fourths of all small RNA species . This contrasts with many other organisms where microRNA-class small RNAs are the archetype . Dust mite siRNAs are , at least in part , involved in genome surveillance . They target TE’s and depletion of Dicer proteins causes derepression of these elements . Control of TE’s is typically carried out by piRNAs in flies , from which dust mite siRNAs are distinct . A common feature of nearly all piRNAs is a “U” residue at the first position . We do not observe this in any subset of dust mite siRNAs . Furthermore , well-described modes of piRNA biogenesis found in Drosophila and C . elegans are absent in dust mites . Loss of piRNAs seems specific to psoroptidian mites , as they are clearly present in other Acari , like spider mites . The divergent nature of dust mite siRNAs is particularly apparent in the absence of 2’-OMethylation of siRNAs–a common feature of siRNAs and piRNAs in other organisms . Interestingly , scabies mites also lack the requisite Hen-1 protein [36] . Inspection of syntenic regions of the dust mite and scabies mite genomes showed rearrangements at this locus , potentially linking the loss of this activity to the evolution of Psoroptidia-specific Ago proteins ( S13 Fig ) ( Table 1 ) . The highly divergent RNAi pathways of dust mites provide an evolutionary perspective not only on the utility of small RNAs to acquire roles in genome surveillance , but also that the precise mechanism may not be that important . This is supported by relatively similar composition of classes of TE’s in spider mites , dust mites , and scabies mites ( S15 Fig ) . While similar classes were observed their locations and specific identities are distinct . Furthermore , this indicates that the collection of dust mite TEs analyzed in this study accurately represent the overall TE population . Flux of small RNA pathways correlates with evolutionary innovation; for example , higher arthropods lost Rdrp in favor of piRNA control of TE [61] . This also occurred when vertebrates diverged from basal chordates [62] . In both cases , loss of Rdrp accompanied innovation in body plan and sensory organs . In vertebrates , whole genome duplication occurred twice following descent from a Rdrp expressing chordate ancestor , affirming a period of genome instability [62] . TE mobilization may be fortuitous for adaptation , and dramatic evolutionary changes may require extreme events such as perturbation of surveillance mechanisms .
The dust mite genome was assembled using reads produced by PacBio and Illumina platforms . The initial assembly was generated by PacBio HGAP . Illumina reads were preprocessed in three steps before using them for extending PacBio contigs: a ) Using Trimmomatic [63] , from both ends of reads , nucleotides with base quality lower than 15 were removed . b ) Using FastUniq [64] , duplicate pairs were removed from the PE library , and c ) SOAPec [65] was used to correct read error [64 , 65] . Any initial genome sequence has bacterial contamination due , at least , to the presence of gut microbiota in DNA isolates . To remove bacterial DNA sequences from D . farinae genome sequence , 4 , 864 , 367 Bacterial genome sequences [66] were downloaded from RefSeq database at: ftp://ftp . ncbi . nih . gov/refseq/release/bacteria and a blast database was created using the sequences [66 , 67] . All the contigs were blasted against the created bacterial genome database to check bacterial contaminations in the sequenced contigs . Then the matched percentages were calculated for each of the contigs . If the matched percentages were higher than 10% of an individual contig length , the contig was considered as contaminated by bacterial DNA and was discarded . After this process , our final contig number was reduced to 1706 , N50 Read Length of 19 , 371 with the total length of 91 , 947 , 272 bp . Finally , a published dust mite genome [34] was compared to our assembled contigs using QUAST [34 , 68] . 79 . 3% bases of the reference genome could be aligned in the new assembly . Using available mRNA-seq datasets [34] , transcripts were identified by the Tuxedo suite . Initial mapping with Tophat was followed by transcript annotation with cufflinks [69] . Transcript similarity was estimated using Blast2Go . Total RNA isolated via the trizol method from bulk collected dust mites in order to capture life stages of D . farinae . Small RNAs were cloned from total RNA with an Illumina small RNA truseq kit , and sequenced on the Illumina NextSeq platform . The dataset was comprised of nearly 400 million reads . Quality of the sequenced library was assessed by FastQC tool and the small RNA reads were analyzed using a custom pipeline ( S1 Fig ) [70] . Mites collected with the salt bath method were suspended in a solution of dsRNA dissolved in nuclease free water ( S1 Text ) . After 6 hours , animals were washed in water and dried on filter paper . After that the animals were kept in 23°C with relative humidity of 80% . After two days , total RNA was extracted using trizol method and resolved in a 12 . 5% urea-polyacrylamide gel . When animals were fed unlabeled dsRNAs , RNAs were transferred to nylon membranes and subject to northern blotting as previously described ( S1 Text ) [56] . If radiolabeled RNAs were fed , gels were directly exposed to phosphoimager screens . 20 μg of total RNA was oxidized at room temperature in borax/boric-acid buffer ( 60 mM borax and 60 mM boric acid-pH 8 . 6 ) containing 80 mM NaIO4 for 30 min . β-elimination reaction was carried out for 90 min using 200 mM NaOH at 45°C . Following precipitation , RNA was resolved on a 12 . 5% urea-polyacrylamide gel , and subject to northern blotting as previously described [56] . 20 μg of total RNA was used for each of reaction . Terminator exonuclease ( epicenter ) was added to one tube and the tube was incubated at 30°C for 60 minutes . After that the reaction RNA was purified by organic extraction protocol [71] . In the second condition , 1 μl CIP ( Calf intestinal phosphatase , NEB ) was added and incubated at 37°C for 30 min . Terminator exonuclease was added followed by a second incubation at 30°C for 60 minutes . Precipitated RNAs were resuspended in loading buffer and resolved on a 12 . 5% urea-polyacrylamide gel , and subjected to northern blotting as previously described [56] . A Methyl DNA seq library was created with Illumina Methyl-seq TruSeq Kit from dust mite DNA recovered by organic extraction followed by precipitation . Using the Bismark algorithm [72] base converted dust mite genome indexes were used to determine the rate of cytosine methylation . Using coordinates from cufflinks ( mRNA ) , and RepeatMasker ( TE ) annotations , rates of methylation were determined for different genomic features . Reads were mapped uniquely and duplicated reads were discarded that resulting in an average 6X coverage depth [72] . Using bedtools , genomic regions that had >4 reads mapping were determined and the base conversion rate measured . Assembled genome was submitted under GenBank ID: NBAF01000000 . Small RNA bioSample accession number is: SAMN05441789 . Datasets of Bi-sulfite sequencing are deposited under the BioSample accession number: SAMN06891248 . Spider mite small RNA datasets used in the study can be accessed at GEO GSE32005 . Drosophila small RNA dataset using in the study can be accessed at GEO GSE83698 .
|
Investigation of small RNA populations in dust mites revealed absence of the piwi-associated RNA ( piRNA ) pathway . Apart from several nematode and platyhelminths lineages , piRNAs are an essential component of animal genome surveillance , actively targeting and silencing transposable elements . In dust mites , expansion of Dicer produced small-interfering RNA ( siRNA ) biology compensates for loss of piRNAs . The dramatic difference we find in dust mites is likely a consequence of their evolutionary history , which is marked by descent from a parasite to the current free-living form . Our study highlights a correlation between perturbation of transposon surveillance and shifts in ecology .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"sequencing",
"techniques",
"invertebrates",
"rna",
"interference",
"gene",
"regulation",
"invertebrate",
"genomics",
"animals",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"sequence",
"motif",
"analysis",
"epigenetics",
"molecular",
"biology",
"techniques",
"rna",
"sequencing",
"drosophila",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"small",
"interfering",
"rnas",
"genetic",
"interference",
"bioinformatics",
"gene",
"expression",
"mites",
"molecular",
"biology",
"animal",
"genomics",
"insects",
"arthropoda",
"biochemistry",
"rna",
"eukaryota",
"double",
"stranded",
"rna",
"nucleic",
"acids",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"non-coding",
"rna",
"organisms"
] |
2018
|
Rewired RNAi-mediated genome surveillance in house dust mites
|
Metabolic fluxes are a cornerstone of cellular physiology that emerge from a complex interplay of enzymes , carriers , and nutrients . The experimental assessment of in vivo intracellular fluxes using stable isotopic tracers is essential if we are to understand metabolic function and regulation . Flux estimation based on 13C or 2H labeling relies on complex simulation and iterative fitting; processes that necessitate a level of expertise that ordinarily preclude the non-expert user . To overcome this , we have developed SUMOFLUX , a methodology that is broadly applicable to the targeted analysis of 13C-metabolic fluxes . By combining surrogate modeling and machine learning , we trained a predictor to specialize in estimating flux ratios from measurable 13C-data . SUMOFLUX targets specific flux features individually , which makes it fast , user-friendly , applicable to experimental design and robust in terms of experimental noise and exchange flux magnitude . Collectively , we predict that SUMOFLUX's properties realistically pave the way to high-throughput flux analyses .
Metabolic fluxes describe the in vivo flow of organic matter through the biochemical reaction network , as defined by enzymes and transporters . An improved knowledge of metabolic fluxes is crucial if we are to understand how cells utilize nutrients , and how they regulate metabolism in the face of dynamic environmental conditions , or in stressed pathologic states [1–4] . Metabolic fluxes , as an emergent property of cellular systems , are prohibitively hard to predict using proteomics or metabolomics data , and are not , per-se , measurable . Hence , the task of assessing metabolic fluxes indirectly represents something of an analytic and mathematic tour-de-force . The most informative approach to estimate metabolic fluxes involves stable isotope labeling . Cells grown in the presence of 13C-enriched substrates incorporate heavy isotopes throughout their metabolic networks according to carbon fluxes and produce characteristic 13C patterns in metabolites and products . Some of these can be measured by mass spectrometry or nuclear magnetic resonance and can ultimately be used to deduce fluxes using two basic approaches . The first is global isotopomer balancing , which seeks to estimate all metabolic fluxes by iterative fitting [5–10] . The power of this approach is that it integrates all available data simultaneously in order to estimate metabolic fluxes across the entire system . The downside is that this approach is ill suited for high-throughput analyses as it necessitates quantification of all uptake and production rates , and analyzes each sample individually . In addition , the fitting procedure is mathematically cumbersome , and for complex or poorly calculable problems , can require extensive computation time . Finally , troubleshooting heavily relies on expert knowledge [8] . The alternative approach is to use flux ratio analysis , which focuses on the resolution of local fluxes , centered on metabolic nodes of particular interest [11–16] . For this purpose , flux ratio analysis adopts a targeted strategy in which relative ( fractional ) information on contributions from alternative pathways are calculated from a small subset of 13C-data using predefined analytic formulas . The advantage of this approach is that it is mathematically simple , rapid , well suited for large scale analyses [16] , and easily used by the non-expert user . However , this process suffers from the time-consuming procedure of deriving analytic formulas for each flux ratio of interest . These formulas , manually derived for each metabolic network , tracer , and environment , generally incorporate a mix of human intuition together with tacit assumptions regarding flux . Over the past 20 years , only a dozen have been derived to describe the central metabolism of microbes growing on single carbon sources . In practice , most experimental conditions cannot be addressed due to the lack of validated flux ratio predictors . In response to these limitations , automated tools have been developed to estimate flux ratios [17 , 18] , although , thus far , these have been limited to linear cases and consequently have failed to find any broader application . Here we present SUMOFLUX , a conceptually novel method to analyze , in a targeted fashion , flux ratios based on 13C-data . Our workflow circumvents concerns over the relevance and limitations of flux analyses by exploiting machine learning . A machine learning predictor is trained using in silico 13C-data , generated by surrogate modeling . The combination of surrogate modeling and machine learning permits the rapid estimation of flux ratios for virtually any metabolic network , label configuration , or available measurement . We now illustrate the proposed workflow for both canonical and novel flux ratios for central carbon metabolism . The speed and generality provided by machine learning makes SUMOFLUX particularly useful for optimizing experimental design , selecting metabolites to be measured , and merging data from several experiments . Moreover , we believe that the SUMOFLUX workflow provides a real prospect of high-throughput flux analyses .
In 13C-metabolic flux ratio analysis , the goal is to estimate a flux ratio of interest . Typically , this is a number that indicates the relative fraction of a specific metabolite flowing through a chosen reaction or pathway . Flux ratios are estimated based on a stoichiometric model , knowledge of the 13C-configuration of all of the relevant substrates , and the labeling patterns of metabolites as measured by mass spectrometry ( or nuclear magnetic resonance ) . We formulated the derivation of flux ratio estimates from 13C data as a nonlinear regression task to be solved using machine learning . By definition , the flux ratio of interest is the dependent variable that we aim to predict; measured 13C isotope labeling patterns of intracellular metabolites are the independent variables , or the input features for the algorithm . A random forest predictor [19] is then trained to build a functional relationship between the 13C data and flux ratios using a training dataset . To build a generalized predictor , the training dataset should comprise hundreds , if not thousands of representative examples for which a flux ratio and 13C data are available . Unfortunately , such a dataset is not accessible experimentally . First , because flux estimates are not amenable to direct measurement . Second , in the majority of cases it is impossible to select , or to construct , a cohort of cells with a phenotypic diversity that adequately represents the wide variety of fluxes and flux ratios that might exist . To overcome this fundamental problem , we have used surrogate modeling ( hence the term SUMOFLUX ) . We built , in silico , a synthetic cohort of representative data points . Each data point is defined by a complete set of fluxes that fulfill the stoichiometric constraints of the metabolic network . This allows us to calculate a ratio ( or any other derivative value ) for the fluxes of interest , and to simulate the 13C-labeling patterns of each metabolite , which is made possible because each flux distribution leads to a unique intracellular labeling pattern [20] . It is therefore possible to construct an in silico dataset comprising thousands of data points , with flux ratios spanning the feasible range , and corresponding metabolite 13C-labeling patterns . The synthetic in silico dataset is used to train , cross-validate , and then test the flux ratio predictor . The full SUMOFLUX workflow for flux ratio estimation consists of five steps ( Fig 1 ) . First , a reference dataset of several thousand flux-maps is sampled from that space of flux-maps that fulfills certain stoichiometric constraints ( mass balances ) of the metabolic network . Extracellular flux constrains can be further refined by the availability of substrates and a working knowledge of the major secreted products . Second , the 13C-labeling patterns of the metabolites included in the network are simulated independently for each reference data point . Label propagation is simulated using the existing algorithms [9] , given the 13C-label of the substrate ( s ) , and the map of the atom transition within the network . At this point , the simulated 13C data , does not , as yet , reflect actual measurements . Third , to capture measurement data , we select only those 13C features that are analytically accessible , and then superimpose noise values corresponding to those measured . Fourth , flux ratios of interest are calculated for all of the data points within the reference dataset , as the dependent variable in regression analyses . Fifth , we divide the reference dataset into independent training and test subsets , using the former to train a random forest with which to predict the calculated flux ratios from simulated 13C data . We then assess the predictor’s performance on the test dataset by calculating the mean absolute error of the predictions made . If the performance is insufficient ( e . g . mean absolute error MAE > 0 . 05 ) , we iteratively optimize our experimental strategy by changing the substrate label , or available measurements , then repeat the training . If the performance is judged to be satisfactory , we finally use the predictor to estimate flux ratios using real experimental data . To provide prediction intervals , we use quantile regression forests , which give a non-parametric and accurate estimates of conditional quantiles based on the full conditional distribution of the dependent variable [21] . The most time consuming aspect of the workflow is the simulation of 13C data in the reference dataset , which scales according to the number of samples and carbon atoms in the metabolites . For the model of central carbon metabolism , with 39 reactions and 21 measured metabolites and fragments ( S1 Fig , S1 and S2 Tables ) , 0 . 2 seconds are needed to simulate the labeling patterns for a single data point . Using a parallelization technique , this process can be accelerated to simulate the several thousand data points necessary for training and testing within a few minutes . Without parallelization , the simulation procedure for 20 , 000 data points takes ~1 . 5 hours , whereas the flux sampling and random forest training steps require less than a minute . Overall , the SUMOFLUX workflow requires information on the stoichiometry of the metabolic network , and the carbon atom arrangement for all of the metabolites within the network . The choice of 13C-tracer depends on the flux ratio of interest [8 , 22] , but in practice is primarily constrained by commercial availability and costs . Hence , it is quite common to test flux calculability using multiple configurations of tracers [23–25] , which can be easily accomplished using the SUMOFLUX workflow due to its rapid computational time . In the following sections , we demonstrate the performance , generality , and scalability of SUMOFLUX , as well as its versatility in terms of feature selection and experimental design . We chose to demonstrate SUMOFLUX by deriving estimates of the flux ratios for central carbon metabolism using the model organism , Escherichia coli . Its metabolic network includes the highly conserved pathways of glycolysis , the tricarboxylic acid ( TCA ) cycle , and the pentose phosphate pathway ( PPP ) . Furthermore , it includes alternative pathways such as the Entner-Doudoroff pathway and the glyoxylate shunt that conveys additional metabolic elements that might complicate flux estimates . As a reference , we considered the study of glucose metabolism in E . coli as described by flux ratio analyses using manually derived analytic equations [15] . We used our method to estimate five key flux ratios based on the labeling patterns measured by gas chromatography mass spectrometry ( GC/MS ) of proteinogenic amino acids upon silylation . We then sampled a reference dataset of 60 , 000 flux distributions using the E . coli central carbon metabolism network ( S1 Fig , S1 Table ) , and simulated the labeling patterns of 21 intracellular metabolites and their fragments ( S2 Table ) , assuming growth on either 100% [1-13C] glucose , or a mix of 20% [U-13C] glucose and 80% naturally labeled glucose . Several parameters had to be defined prior to predictor training . The performance of the random forest depends on the number of decision trees in the forest ( ntree ) , and the number of input features used at each tree node ( mtry ) . To choose these parameters we used five-fold cross-validation on the training dataset . We tested 16 combinations of ntree and mtry values for the five E . coli flux ratios . The combination of 100 prediction trees ( ntree ) with 20 mtry features delivered a good balance between predictor accuracy and computation time ( S2a Fig ) . These two parameters were then applied throughout the study . The number of simulated points used for training also influences predictor accuracy and computation time . Our tests demonstrated that ~10 , 000 simulated points were generally adequate in terms of generating a sufficiently accurate estimate of the key flux ratios in the E . coli dataset; thereafter , any further increase in the number of data points provided no tangible improvement in accuracy ( S2b Fig ) . We took these results into account when extracting the training sets for the predictors ( see Materials and Methods for details ) . We trained predictors for the five E . coli flux ratios on the simulated training dataset and then assessed their performance on an independent simulated test dataset . In all cases , the mean absolute error was < 0 . 1 ( Fig 2 , second column ) . For comparison , we also applied the analytic formulas manually derived for the E . coli study [15] ( S3 Table ) to the same simulated test dataset . For all tested flux ratios , SUMOFLUX outperformed the analytic formulas in terms of mean absolute error on the test dataset ( Fig 2 , fourth column ) . This possibly reflects the fact that the flux estimates for the test dataset were obtained through sampling of the entire solution space , and do not comply with some of the implicit simplifications and assumptions for the network , fluxes , and reaction reversibility , that are generally used to derive the analytic formulas [15] . For example , in calculating the fraction of oxaloacetate from phosphoenolpyruvate , the flux through the glyoxylate shunt was assumed to be zero , whereas in the test set it possessed a wide range of values ( Fig 2e ) . Furthermore , the analytic formula for estimating the malic enzyme flux ratio provides only a lower bound value ( Fig 2d ) . We also compared flux estimates generated by the two approaches using the real experimental 13C data from this study ( Fig 2 , third column , and S4 Table ) ; both produced concordant estimates ( Pearson correlation coefficient , PCC > 0 . 89 for all ratios ) . To further demonstrate the scalability and generality of SUMOFLUX , we applied the same approach and parameters to estimate four flux ratios using the GC-MS data for amino acids collected for 121 Bacillus subtilis transcription factor mutants grown on a mix of 80% [1-13C] glucose and 20% [U-13C] glucose [16] . Again , the random forest predictor outperformed the analytic formulas for the in silico test dataset ( S3 Fig ) . For three flux ratios , the two approaches provided consistent estimates for the experimental data ( PCC > 0 . 65 ) . However , the malic enzyme ratio could not be resolved with sufficient precision using either method . Presumably , the mix of tracers chosen was poorly suited to this task . In order to highlight the scope of SUMOFLUX applicability in context of global 13C flux analysis methods , we compared it with the classical 13C-metabolic flux analysis by global isotopomer balancing ( 13C-MFA ) approach , which seeks for a global flux solution that provides the best fit to the experimental data—measured metabolite labeling patterns and physiological parameters . We applied 13C-MFA to the data for the same eight E . coli strains and added glucose uptake rates [15] as an additional input . With INCA [9] , we calculated the best flux fit and flux confidence intervals using parameter continuation procedure ( S5 Table ) . SUMOFLUX and 13C-MFA differ in the demand of input information and produce different outcomes ( net fluxes vs . flux ratios ) . To compare , we calculated flux ratios from the net fluxes estimated by 13C-MFA and directly compared to SUMOFLUX results . Confidence intervals on flux ratios for 13C-MFA were obtained by repeating the optimization procedure 1000 times for each strain . Because it employs less input data , SUMOFLUX is expected to be worse than 13C-MFA . In general , however , the flux ratio estimates obtained with the two methods were in good agreement ( PCC>0 . 83 for all ratios calculated for the best fit to either [1-13C] data , [U-13C] data , or combined dataset , S4 Fig ) . Surprisingly , in several cases the confidence intervals of 13C-MFA flux ratio estimates were much larger than the prediction quantiles of SUMOFLUX and the accuracy of flux ratio estimates depended on the experimental dataset used for the fitting , perhaps pointing to the presence of inconsistent or overly noisy data that decrease the precision of 13C-MFA estimates . This example illustrates the complementarity of the two approaches . 13C-MFA provides global flux solutions , but in some cases the targeted approach performs better in resolving local fluxes . Collectively , the E . coli and B . subtilis results demonstrate that SUMOFLUX is broadly applicable to real experimental data with an accuracy that is comparable , if not better , than that of manually derived formulas . Even though there is no guarantee that a specific flux ratio can be accurately estimated for a given metabolic network , tracer , or experimental data , SUMOFLUX does allow for rapid verification and ad-hoc experimental design . Beyond the speed and ease with which predictors can be generated for calculating metabolic flux ratios from 13C data , SUMOFLUX offers the additional benefits of robust prediction , the option to vary and optimize experimental design , and the estimation of novel ratios that we explore in the next sections . Excessive measurement noise and underestimates of exchange flux of bidirectional reactions are two frequent causes of inaccurate flux estimates . We set out to assess their influence on SUMOFLUX by performing an in silico experiment using E . coli , varying the values of the superimposed measurement noise by up to 0 . 10 , i . e . 10-fold higher than that routinely obtained with careful peak integration . We also used exchange flux values of up to 100-fold that of the net flux value , i . e . a model approximation of full equilibration of the reactants . To exclude potential prediction accuracy differences arising from different training and test datasets , we used one set of flux vectors , divided into training and test subsets . Addition of four different noise levels and variation of four exchange flux magnitude values resulted in 16 datasets , which differ only in these two parameters . We again trained the predictors for each of the datasets using the training subset and calculated the mean absolute error ( MAE ) on an independent test subset . As a rule of thumb , we consider a ratio to be accurately predictable if the MAE < 0 . 05 . This criterion was met by the Entner-Doudoroff , glycolysis/PPP , and anaplerosis ratio predictors , within the normal ranges for noise ( ~ 0 . 01 ) and exchange flux ( ~10 times the net flux value ) ( Fig 3a ) . The other two tested predictors were less precise , and are better suited to the analysis of substantial flux changes . Alternatively , different tracers and measurement techniques could be tested , as outlined below , to achieve more accurate analyses . We also performed robustness analyses of the analytic equations , and found that only the formula for the Entner-Doudoroff pathway was sufficiently robust in terms of noise and flux exchange that were within the normal ranges ( S5 Fig ) . The remaining four formulas were either extremely sensitive to noise ( gluconeogenesis ratio ) , or were poorly suited to the entire range of flux maps . Another important aspect that can be assessed with this type of analysis is to what extent erroneous assumptions in the training dataset affect the accuracy of the flux estimates in the test dataset . We used the simulated data described above to test the effects of noise and exchange flux magnitude values separately . First , we fixed the exchange flux magnitude to 1 , and calculated the accuracy of the flux ratio predictors trained and tested on 16 combinations of train and test subsets with the four independently added noise levels . Notably , we observed that for all ratios underestimating the level of noise in the data was detrimental for the flux ratio prediction accuracies . On the contrary , overestimating the noise in the training data resulted in better flux estimates in the less noisy test datasets compared to the test data with the same levels of noise as in training ( Fig 3b ) . In practice , it is always desirable to only superimpose a realistic level of noise to in silico data , as the addition of noise inevitably decreases the prediction accuracy . However , it is advisable to adopt a conservative over-estimate noise to avoid overfitting . Adequate magnitude of exchange fluxes appears to be even more important for predictor accuracy . We observed that both under- and over-estimated exchange flux magnitude values resulted in lower accuracy compared to the accuracy on the test dataset corresponding to the training ( Fig 3c , diagonal values ) . Remarkably , in these simulated datasets we set one exchange value as upper bound for all reversible fluxes in the model , which presumably has a greater effect on prediction accuracy than an exchange flux magnitude of a single reaction . In biological systems , we expect large differences in the reversible flux magnitudes of different enzymes , and it is beneficial to include this information in the model , when available . One way to control whether the simulated data used for training and testing adequately represents the experimental data , is to compare the distributions of inter-quantile ranges of the flux ratio estimates for the experimental data to the ones of the test data . In our examples , the distributions of the inter-quantile ranges of the flux estimates for under-estimated noise or inappropriately estimated exchange flux estimates were significantly different ( p<0 . 01 , Wilcoxon-Mann-Whitney test , S6 Fig ) . In practice , it is advisable to compare the inter-quantile range distributions of the flux ratio predictions for in silico and experimental data , although statistical tests should be used with caution due to very different sample sizes . In summary , SUMOFLUX provides flux ratio predictors that are generally robust to noise and exchange fluxes , both of which are major confounding factors in labeling experiments . This robustness is dependent on flux ratio , labeling strategy , and the available measurements used for prediction , and can easily be assessed , if required , in each particular case . The glyoxylate shunt plays an essential role in bacterial adaptation to alternative carbon sources , such as acetate and fatty acids , as it replenishes the TCA cycle with C2 carbon fragments . Hence , this pathway has an important anaplerotic function besides phosphoenolpyruvate carboxylase . No analytic formulas were developed to resolve the relative contribution of the glyoxylate shunt due to the complexity of carbon rearrangement at this branch point , the additional complication introduced by multiple cycling in the TCA cycle , and the similarity of the labeling patterns of the relevant metabolites . Here , we opted to tentatively resolve this pathway using SUMOFLUX , and the 13C-data available from GC-MS analyses of protein-bound amino acids in E . coli [15] . Using the same simulated dataset described above , we trained two more predictors to estimate flux contributions to the formation of oxaloacetate , one derived from the glyoxylate shunt , and the other , from the TCA cycle ( Fig 4a ) . The accuracy of the predictors achieved for the in silico test dataset was acceptable ( MAE < 0 . 07 ) ( Fig 4b and 4c ) . Collectively , these two novel ratio predictors , and the one previously trained to estimate the anaplerotic reaction from phosphoenolpyruvate to oxaloacetate ( Fig 2e ) , allowed us to comprehensively assess the metabolic source of oxaloacetate . The prediction intervals of the estimates for the experimental data were in the range of 10% due to the difficulty of precisely resolving the glyoxylate shunt based on the available data . Nevertheless , the estimated values reflect those expected from the literature . Specifically , the differences between strains were consistent with their genotype ( Fig 4d , S7 Fig , S6 Table ) . The estimated glyoxylate shunt contribution for wild type bacteria was 16 ± 10% , with the highest glyoxylate shunt ratio ( 32 ± 15% ) estimated for the Δpgi mutant , which is consistent with other studies [26 , 27] . In contrast , both the double Δmdh Δsdh mutant and the ΔfumA mutant , in which the pathway from succinate to malate is disrupted , had an almost zero glyoxylate shunt and TCA cycle activity , with the major contribution to the oxaloacetate pool being the flux derived from phosphoenolpyruvate . The Δzwf mutant , with a compromised oxidative pentose phosphate pathway , exhibited the highest fraction for the TCA cycle flux ( 49 ± 18% ) , which reflects a compensatory response to ensure NADPH equilibrium via isocitrate dehydrogenase [26] . The glyoxylate shunt example shows how novel quantitative flux predictors can be rapidly generated using our approach . In the context of metabolic analyses , a priori experimental design aims at identifying the best settings from simulated data with which to accurately estimate the fluxes of interest . In global isotopomer balancing and fitting , numerical simulations have been frequently used to optimize tracer selection for one specific flux state , e . g . that of an unperturbed wild-type strain [23–25] . In targeted flux ratio analysis with manually derived analytic equations , simulation-assisted experimental design is not possible , as each equation is formulated for a specific experimental condition chosen by the researcher , and no simulation procedure is employed to assess its accuracy . In contrast , the speed and simplicity of SUMOFLUX facilitates the rapid testing of altered metabolic models , tracer choices , or data sets for the derivation of a flux ratio of interest . This enables us to systematically , yet rapidly , identify the optimal experimental strategy from those available . We demonstrated this feature of SUMOFLUX by testing different settings for the B . subtilis labeling experiment . Using the same reference flux dataset as above , we simulated the 13C metabolite labeling patterns for eight different glucose-labeling strategies . For each label , we simulated the measurements that could be obtained using four different measurement techniques: GC-MS analyses of amino acids , liquid chromatography LC–MS of intact intracellular metabolites , LC-MS/MS analyses of intact metabolites and their fragments [28] ( S7 Table ) , and all individual MS/MS traces used in multiple reaction monitoring of metabolites ( S8 Table ) . For each of the 32 experimental setups , we rapidly trained random forest predictors for the malic enzyme , gluconeogenesis , and glycolysis/PPP flux ratios , and assessed their performance in silico on the test dataset . To compensate for the different number of features , and avoid over-fitting , we introduced a feature selection procedure using cross-validation on the training dataset prior to training ( see Materials and Methods for details ) . As expected , flux calculability depends on the flux ratio of interest , the tracer , and the measurement platform ( Fig 5 ) . For the malic enzyme and glycolysis ratios , LC-based methods are preferable to GC-MS . Tracers such as [6-13C] , [5 , 6-13C] , or [4 , 5 , 6-13C] glucose offer the best overall accuracy ( MAE < 0 . 05 ) . Any of these tracers could be selected to quantify the three flux ratios in a single experiment . For specific flux ratios , the average error was reduced to about MAE 0 . 02–0 . 03 by selecting specific tracers . However , when taking into account the cost of tracers , a labeling experiment using 50% [U-13C] and 50% naturally labeled glucose might be seen as a compromise between prediction accuracy and cost . This analysis underlines the ease with which an experimental design , targeted to address specific biological questions , can be implemented using the SUMOFLUX workflow .
We have developed a generalized method for targeted analysis of 13C metabolic flux ratios , that builds on surrogate modeling ( SUMOFLUX ) , i . e . uses a synthetic dataset to train a machine learning predictor to estimate a given flux ratio directly from 13C-data . Synthetic datasets are constructed in silico on the sole basis of four easily accessible inputs: a stoichiometric model of metabolism , a list of possible metabolic substrates and their byproducts , a configuration of the 13C-substrate , and a list of measurable metabolites with measurement error . These inputs are sufficient to generate a representative synthetic dataset covering a broad range of fluxes and flux ratios . A random forest predictor is then trained on this dataset to capture the relationship between simulated 13C-data and the flux ratio of interest , that holds true for all of the simulated data points . Therefore , the same predictor can be used to estimate flux ratios for normal cells , as well as for knock-out mutants without the need for additional information on physiological parameters or their uptake/consumption rates . Due to the fact that the SUMOFLUX predictor targets only a single flux ratio at a time , it is very efficient in assessing calculability and eventually estimating flux values from real data . This feature is particularly relevant when tackling complex fluxes [29 , 30] , as dozens of different experimental designs can be trialed within a few hours . If necessary , measurement data from parallel experiments using different 13C tracers can be combined and passed as input features into the SUMOFLUX workflow . This approach has been proven to improve flux estimates in certain cases [31 , 32] . The surrogate modeling of cells grown in rich media with multiple substrates is made possible because carbon labeling experiments can be simulated for large or even genome-wide networks [33] , inclusive of all the key metabolic pathways . Overall , SUMOFLUX is generally applicable to virtually any combination of metabolic model ( organism ) , medium composition , isotopic tracer , or measurement technique . The crucial step in SUMOFLUX is the construction of the synthetic data used for predictor training . To obtain representative data , it is extremely important that the surrogate model be based on realistic assumptions of the metabolic network and experimental measurement accuracy . Prior knowledge can be integrated into the sampling procedure to limit the space of flux distributions and potentially improve the predictor’s performance . Network simplification and constraining bear some risks . The metabolic model should encode all possible metabolic reactions such that 13C-patterns can be correctly assigned to the underlying flux states . If a reaction is omitted from the surrogate model , the predictor will provide biased estimates . Although the omission of reactions from a model leads to better accuracy in silico , that step would only be justified if the reaction was proven to be inactive under all conditions tested , e . g . by biochemical assay or enzyme quantification . Unless such information is available , it is recommended that all reactions be included in the model in order to achieve robust predictor training . For similar reasons , it is equally important to provide a real-life or conservative error model of the measurement data . According to in silico testing , overestimating noise in the simulated dataset does not lead to overestimating predictor’s accuracy , on the contrary to underestimating noise . In our experience , a valid sanity check is to verify that the simulated data distribution covers the measured mass isotopomer fractions by comparing the distributions of simulated and experimental data . Another indicator of potential discrepancies between the simulated and experimental data is the difference between the distributions of inter-quantile ranges of the flux ratio predictions , which can be tested with a nonparametric test , such as Wilcoxon-Mann-Whitney . With these simple procedures , errors in the metabolic model , substrate composition , or experimental measurement can be detected . Targeted flux ratio analysis using SUMOFLUX is best suited to the assessment of flux ratio to a high accuracy , on selected metabolic nodes , or when mid to large throughput is necessary . High-throughput is made possible by the speed of the approach and by the fact that only 13C-data are required . Once trained , the predictor can be applied to estimate flux ratios for all tested samples simultaneously . A further optimization of experimental measurement time can be explored by including feature selection during training to identify the most informative as well as negligible 13C-features . SUMOFLUX complements the alternative global isotopomer balancing and iterative fitting method ( 13C-MFA ) , which requires measurements of uptake/consumption rates , and more detailed analyses of each dataset , but provides net flux estimates for all reactions in the model . Our short comparison with the data of 8 E . coli strains demonstrated that the two approach deliver consistent flux ratio estimates . In some instances , the confidence of SUMOFLUX estimates was better . Hence , it could be used before 13C-MFA to increase its performance . In this case , multiple flux ratios could be estimated independently to obtain experimental information on different degrees of freedom prior to applying global 13C flux analysis methods [34] . In principle , the concept of SUMOFLUX can be extended to isotopically non-stationary data . The simulation of dynamic 13C-data can be completed with the inclusion of metabolite concentrations in the sampling procedure with simulation of 13C dynamics at predefined time points to be matched in the experiment . The training of flux predictors from isotopically non-stationary data can use the same procedure outlined for stationary data , even though it is substantially more demanding because of the requirement to sample an increased number of degrees of freedom and measurable labeling features . However , it must be stressed that non-stationary labeling experiments are much more labor-intensive and data demanding , and can be performed only at low throughput [35] . For practical reasons , the traditional approach of flux estimation by both global [35 , 36] or local [37] iterative fitting is better suited to the analysis of small-scale non-stationary labeling experiments . Overall , the concepts underlying the proposed SUMOFLUX workflow are easily transferrable and can be applied alone , or in combination with other methods , to address different flux analyses questions . We believe that SUMOFLUX has the potential to become a core tool in the analysis of metabolic fluxes , and opens new possibilities for high-throughput flux profiling of a wide variety of metabolic systems .
Metabolic network with carbon atom transitions and the lists of input and output metabolites are defined by the user and are represented in the mat-file format required by the INCA software [9] . In order to reduce the dependency on the biomass vector coefficients , a separate output flux is defined for each of the biomass precursors , therefore biomass precursors are also added to the list of outputs . The substrates are defined as unbalanced compounds and do not participate in the stoichiometric equation system . In the flux sampling procedure , the definitions of net , exchange , forward and backward fluxes are used [38] . By default , the lower and upper bounds for reversible reactions are set to [-100 100] , for irreversible reactions to [0 100] , and the major uptake flux is set to 10 . First , the initial net flux solution is found by minimizing the sum of squared fluxes with stoichiometric constraints , inequality constraints on the output fluxes , and flux bounds using the MATLAB solver fmincon . Second , a cohort of net flux vectors is generated with Monte Carlo sampling by adding linear combinations of null vectors of the stoichiometric matrix with random coefficients to the initial flux solution . Third , for each net flux , an exchange flux value is randomly generated in the order of magnitude relative to the net flux defined by the user ( by default 1 ) , and forward and backward flux values are calculated accordingly . Optionally , to achieve uniform coverage of values for a particular flux ratio or set of ratios , the ratio range is split into segments ( for example , [0 0 . 1] , [0 . 1 0 . 2]… [0 . 9 1] ) , and the flux sampling procedure is repeated for each segment with the end points set as flux ratio constraints in the first step . The flux ratio of interest is calculated for each of the flux vectors with a formula defined by the user . Given the label of the substrate ( s ) and the list of metabolites and fragments , metabolite labeling patterns are simulated for each flux solution using the INCA software [9] . The INCA ‘simulate’ procedure is integrated into the SUMOFLUX workflow and is called internally for each of the sampled flux vectors . In case parallel computing is available , this procedure is parallelized . The measurement data is simulated by extracting the measured compounds from the simulated data matrix and adding uniform noise to the measurements ( 0 . 01 by default ) . After adding noise , the mass distribution vectors for each metabolite are normalized . Experimental data for E . coli and B . subtilis central carbon metabolism studies were downloaded from the supplementary materials available for the corresponding papers [15 , 16] . MATLAB code for SUMOFLUX and example scripts are available at http://www . imsb . ethz . ch/research/zamboni/resources . html All scripts are compatible with MATLAB 2013a ( MathWorks Inc ) .
|
Living cells adapt to ever-changing environments by regulating metabolic fluxes , the rates of nutrient flow through the metabolic network , to produce metabolites that are currently in demand . 13C-labeling techniques coupled with metabolic flux analyses are widely used to estimate metabolic fluxes and provide insights into cellular physiology and adaptation relevant in biological , biomedical and biotechnological applications . However , the existing methods are either computationally costly , or applicable to a limited amount of biological systems . Here , we combined surrogate modeling with machine learning to present SUMOFLUX , a generalized method for 13C flux ratio analysis . We validated our method by resolving canonical flux ratios in eight Escherichia coli mutants with known metabolic phenotypes and estimated a novel flux ratio for this dataset . We demonstrated scalability of SUMOFLUX and its application for experimental design by applying it to a cohort of 121 Bacillus subtilis mutants . SUMOFLUX , alone or in combination with global flux analysis methods , can be applied to resolve flux ratios in virtually any biological setup , and paves the way to high-throughput flux profiling .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"cell",
"physiology",
"carbohydrate",
"metabolism",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"chemical",
"compounds",
"pathogens",
"metabolic",
"networks",
"experimental",
"design",
"bacillus",
"microbiology",
"carbohydrates",
"glucose",
"metabolism",
"cell",
"metabolism",
"organic",
"compounds",
"glucose",
"research",
"design",
"prokaryotic",
"models",
"model",
"organisms",
"metabolites",
"network",
"analysis",
"molecular",
"biology",
"techniques",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"cell",
"labeling",
"computer",
"and",
"information",
"sciences",
"medical",
"microbiology",
"microbial",
"pathogens",
"chemistry",
"molecular",
"biology",
"metabolic",
"labeling",
"biochemistry",
"cell",
"biology",
"organic",
"chemistry",
"bacillus",
"subtilis",
"monosaccharides",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"metabolism",
"organisms"
] |
2016
|
SUMOFLUX: A Generalized Method for Targeted 13C Metabolic Flux Ratio Analysis
|
Cellular sophistication is not exclusive to multicellular organisms , and unicellular eukaryotes can resemble differentiated animal cells in their complex network of membrane-bound structures . These comparisons can be illuminated by genome-wide surveys of key gene families . We report a systematic analysis of Rabs in a complex unicellular Ciliate , including gene prediction and phylogenetic clustering , expression profiling based on public data , and Green Fluorescent Protein ( GFP ) tagging . Rabs are monomeric GTPases that regulate membrane traffic . Because Rabs act as compartment-specific determinants , the number of Rabs in an organism reflects intracellular complexity . The Tetrahymena Rab family is similar in size to that in humans and includes both expansions in conserved Rab clades as well as many divergent Rabs . Importantly , more than 90% of Rabs are expressed concurrently in growing cells , while only a small subset appears specialized for other conditions . By localizing most Rabs in living cells , we could assign the majority to specific compartments . These results validated most phylogenetic assignments , but also indicated that some sequence-conserved Rabs were co-opted for novel functions . Our survey uncovered a rare example of a nuclear Rab and substantiated the existence of a previously unrecognized core Rab clade in eukaryotes . Strikingly , several functionally conserved pathways or structures were found to be associated entirely with divergent Rabs . These pathways may have permitted rapid evolution of the associated Rabs or may have arisen independently in diverse lineages and then converged . Thus , characterizing entire gene families can provide insight into the evolutionary flexibility of fundamental cellular pathways .
Cells can respond to and shape their environments by taking up and releasing macromolecules . Both in- and outbound transport is facilitated by a network of membrane-bound compartments that represent one of the hallmarks of eukaryotic cells [1] . Traffic through the network is highly regulated , and a convergence of data from structural , functional , and evolutionary studies demonstrate that gene families encode proteins functioning as conserved specificity determinants for endocytic and exocytic compartments [2] , [3] . They do so primarily by controlling the formation , targeting and fusion of vesicles that transport cargo between compartments [4] . One family of key determinants are monomeric GTPases called Rabs , which function as molecular switches by interacting with membrane bilayers and diverse protein effectors in cycles controlled by GTP binding and hydrolysis [5] . A fundamental aspect of Rab function is that multiple Rabs , encoded as a gene family , are co-expressed within a single cell , and the individual family members are each targeted to a small subset of membrane compartments , where they interact with unique effectors [6] . In this manner , a cohort of Rabs can coordinately regulate a pathway consisting of sequential and distinct trafficking events [7] . It is therefore likely that gene duplications within the Rab family , followed by diversification into functionally-distinct variants , were key steps in the evolution of eukaryotic membrane complexity [8] , [9] . This model is strongly supported by the tendency of Rabs associated with specific organelles , across a wide range of eukaryotes , to be most closely related to one another . For example , the Rab1 clade has remained highly conserved at the sequence level among all eukaryotic kingdoms; the corresponding proteins , where they have been characterized , are all associated with traffic between the endoplasmic reticulum and cis-Golgi [10] . These experimental findings coupled with phylogenetic parsimony reveal that Rab1 was already a determinant for this step in an early eukaryotic ancestor . It follows that the analysis of lineage-specific expansions or losses in conserved Rab clades can provide insights into the range of evolutionary paths that have led to modern cells . Because of their central importance , Rabs have been studied in a variety of model and non-model organisms . As expected , the number of Rabs in an organism is generally related to organismal complexity [11] , [12] including tissue-specific expression in multicellular organisms [13] . However , a surprising result emerging from several recently-sequenced genomes is that some unicellular organisms express a cohort of Rabs whose number equals or exceeds the 63 Rabs in humans [14] . An extreme example is the parasite Entamoeba histolytica , whose genome encodes 91 Rabs [15] . The sheer number of Rabs hints at an interesting discordance between organismal simplicity and cellular complexity , the latter facilitated at least in part by extensive , lineage-restricted Rab expansions . However , the roles of most Entamoeba Rabs are unknown , and the large Rab number in Entamoeba could endow the cell with flexibility , rather than structural complexity , if different subsets of the Entamoeba Rabs are expressed under the widely disparate conditions that are cyclically encountered by a parasite . Ciliates have classically been recognized as unicellular organisms of great structural complexity including a wealth of membrane-bound compartments [16]–[22] . Recent sequencing of two ciliate genomes revealed that these organisms are similarly gene-rich [23] , [24] . The free-living ciliate Tetrahymena thermophila has more than 20 , 000 genes and preliminary analysis suggested that these included 70 Rabs , raising the same issues discussed above for Entamoeba [23] . To shed light on the nature and evolution of unicellular cell complexity , we have taken a whole genome approach by considering the entire set of Rabs in Tetrahymena , including phylogenetic analysis , expression profiling , and localization of GFP-tagged variants in living cells . While many of the cellular pathways illuminated by this study represent ancient , highly conserved functions , the unbiased nature of our survey also uncovered a surprising level of both evolutionary and cell-stage flexibility within Rab GTPases , key determinants of membrane traffic .
Beginning with amino acid sequences of human Rabs , we identified 56 Rabs in the ciliate Tetrahymena thermophila . This does not include seventeen putative Rabs that were reported in a previous survey [23] , because these were found to lack the specific C-terminal residues to which prenyl residues are attached in authentic Rabs . These 17 predicted genes should therefore be reassigned as Rab-like proteins [25] . We then analyzed all putative Tetrahymena Rabs , including 3 that were not detected in the previous survey , for the presence of five Rab-specific motifs: IGVDF , KLQIW , RFRSIT , YYRGA and LVYDIT [26] . While many of the Tetrahymena Rabs have diverged to various degrees at some of these consensus sites , these genes are nonetheless clearly more similar to Rabs than to any other class of small GTPase , i . e . , Ras , Rho/Rac , Sar/Arf or Ran ( Figure S1; GenBank ID numbers provided in Table S1 ) . Consistent with this , BLAST searches based on any of the 56 Rabs identified only other Rabs as best hits . Tetrahymena , although a single-celled organism , therefore has a somewhat larger set of Rabs than in D . melanogaster or C . elegans , and a much larger set than is present in model fungi ( S . cerevisiae or S . pombe ) or parasitic protists such as T . brucei or P . falciparum , the latter belonging to the clade most closely related to Ciliates , the Apicomplexans ( Table 1 ) . The number of Rabs is similar , however , to those reported for some other single-celled organisms , such as T . vaginalis . The large number of predicted Rabs suggests that some protists maintain networks of membrane compartments that are at least similar in complexity to those in multicellular organisms . However , no comprehensive analysis of the Rabs in any of these organisms has been reported , and the large number of Rabs could reflect alternative expression of paralogs to optimize a relatively simple network of compartments for changes in environmental conditions or lifestages . The transcript abundance of all Tetrahymena genes has been measured via whole genome microarrays , using mRNA samples derived from growing ( medium and high density ) and stationary cultures in growth medium , 7 successive time points during starvation , and 10 successive time points during conjugation [27] . This dataset is particularly useful since transcription plays a central role in the control of differential gene expression in Tetrahymena [28] . Using the publicly accessible database at http://tged . ihb . ac . cn , we found that the transcript abundance of all 56 Rabs is above background and , strikingly , 86% are in the highest-expressing class of genes in this organism [27] . Therefore , none is likely to be a pseudogene . Importantly , 94% of the Rabs are transcribed in growing cell cultures , indicating that membrane traffic in growing Tetrahymena involves more than 50 distinct Rab proteins . Thus the large number of Rab genes is likely to reflect a large number of pathways of membrane traffic that function concurrently in a complex cell , rather than a small number of pathways that are controlled by alternative Rab isoforms in a stage-specific fashion . This analysis also suggested that the majority of Tetrahymena Rabs could be meaningfully localized in cells from growing cultures . Both morphological as well as some molecular studies have revealed extensive changes in membrane traffic when Tetrahymena are starved [29]–[32] , and additional changes during conjugation including the elaboration of a cell-cell fusion zone and the creation of some , and loss of other , nuclei [33]–[35] . The expression data suggest that these changes involve two different phenomena with regard to Rab expression ( Table S1 ) . First , RabsD16 and D12 are not expressed in growing cultures but are greatly induced at specific stages in starvation or conjugation: the former at 6h of starvation , the latter most dramatically at 9h of starvation and at the beginning of conjugation ( Figure S2 ) . Secondly , many Rabs that are expressed in growing cells are redeployed in starvation or conjugation . In particular , many of the Tetrahymena Rabs that are expressed in growing cells are also expressed in starved or conjugating cells , and indeed most show even higher transcript levels in one of these conditions ( 65% in starved , and 27% in conjugating cells ) . However , because average translation efficiency is much greater in growing than in starved cells [36] , the changes in transcript levels under different culture conditions do not necessarily reflect corresponding protein abundance . We therefore looked for peaks of Rab expression within each growth condition , as an indication that some Rabs might be determinants of precise stage-specific structures in starvation or conjugation . Strikingly , 20 Rabs show a discrete >2-fold expression peak at a distinct time point during starvation ( Rabs D22 , D16 , D21 , all near S6 , i . e . , the 6h starvation timepoint ) or during conjugation ( C0 , i . e . , the beginning of conjugation: Rabs D29 , D7; C2 , i . e , the 2h conjugation time point: Rabs 11A , D26 , D33 , D24 , D7 , D23 , D31 , D38 , 21; C4: Rabs D35 , 11C; C6: Rabs 4A , D13; C14: RabD34; C16: Rabs D21 , D15; C18: Rabs D22 , D16 ) ( Table S1 ) . The preponderance of relative Rab expression peaks at C2 corresponds with extensive membrane restructuring during pair formation . In addition , many Rabs show a large expression peak at S0; however , this time point is difficult to interpret because it corresponds to a change in the culture medium . Figure 1 gives an overview of known or inferred pathways of membrane traffic in Tetrahymena , as well as those in a generalized mammalian cell . To identify Tetrahymena Rabs dedicated to specific pathways or structures , we expressed them individually as GFP-tagged proteins and determined the localization in living Tetrahymena . GFP-Rab expression was induced at the lowest level permitting clear visualization , and confirmed using Western blotting of Tetrahymena lysates ( Figure S3 , with two exceptions for Rabs expressed at low levels ) . For practical reasons , most Rabs were localized solely in growing cultures . In preliminary studies , we found evidence that Rabs whose endogenous expression was largely limited to starvation or conjugation were de-localized when expressed in growing cells ( Figure S2 ) , and therefore we did not include that small set of Rabs in this study . We captured time-lapse and/or full z-stacks of cells expressing each of the Rabs in living , immobilized cells . Movies of all Rabs described in this paper are publicly accessible at tetrahymenacell . uchicago . edu . A number of the Rabs were seen to associate with more than one structure ( Table S2 ) . In these cases we considered the stronger signal as primary , and the data below are organized based on these primary signals . In rare cases , we saw equally strong labeling of two structures . Ciliates like Tetrahymena can feed via the phagocytic uptake of bacteria or other particulate matter via a cavity at the cell anterior called the oral apparatus ( Figure 2A ) [37] , [38] . Bacteria are swept into the oral apparatus by cilia at the oral apparatus rim , and ingested at the base . Following ingestion , the bacteria are digested via phagosome maturation resembling that in mammalian cells , including fusion of phagosomes with both early and late endocytic compartments . The final stage of phagosome maturation , which has no precise equivalent in mammalian cells , entails docking and fusion with a structure at the cell cortex called the cytoproct , which results in the egestion of any residual material in the phagosome lumen . Egestion is followed by a burst of membrane retrieval , in a process ( documented by EM in another ciliate , Paramecium micronucleatum ) that results in the rapid appearance of endosomes in the cytoproct region . These endosomes appear to be transported anteriorly along cytoplasmic microtubules , where they may contribute membrane during the formation of phagocytic vesicles from the oral apparatus [39] . Remarkably , almost 1/3rd of the Tetrahymena Rabs were found to be associated with some aspect of the phagocytic pathway ( examples shown in Figure 2; full set in Figure S4 ) . Five of these are localized primarily to structures at or near the oral apparatus , while eleven are localized to phagosomes that can be unambiguously identified in cells that have ingested either fluorescent bacteria or India ink ( Figure S4 ) ( Table 2 ) . Some GFP-Rab fusion proteins are associated with the entire cohort of phagosomes in a cell ( Figure 2C ) , but many of the Rabs are associated with only a subset of the phagosomes ( Figure 2D ) , which may correspond to stages in maturation . Consistent with the specialization of the cytoproct as a zone of membrane fusion and retrieval , three of the phagosome Rabs are restricted to phagosomes that are localized at or near the cytoproct ( Figure 2E ) , and time-lapse movies of cells expressing these GFP-Rab fusions allowed us to directly visualize phagosome egestion and retrieval ( Videos S1 , S2 , S3 , S4 ) . TtRab7 , the Rab7 homolog , was previously associated with phagosomes in a proteomics study [37] . We found that this protein localizes both to lysosomes ( as defined by co-labeling with LysoTracker ) and to small puncta on the surface of phagosomes ( Figure 2G and 2H ) . TtRabD17 localizes to phagosomes in the vicinity of the cytoproct but also localizes strongly to the oral apparatus , suggesting the possibility of long-distance transport between these structures . Consistent with this , the live imaging of TtRabD3 provided a stunning illustration of transport over 40–50µM between late phagosomes , in the extreme posterior of the cell , and the cell anterior ( Figure 2F , Video S5 ) . Inferring from the EM studies in Paramecium , this transport is likely to involve endosomes generated during membrane retrieval following phagosome egestion , which are transported to the oral apparatus . The localization of RabsD17 and D3 suggests the possibility that some Rabs could contribute to coherence within a multi-step pathway by functioning at both early and late steps . Finally , one Rab is localized at the cytoproct itself ( Figure 2I ) . Tetrahymena has a classical endocytic pathway involving clathrin- and dynamin-dependent formation of small endocytic vesicles , which arise from hundreds of depressions at the plasma membrane called parasomal sacs [40] . These structures occur at regularly-spaced sites along well-defined ciliary rows , also called 1° meridians ( Figure 1A and 1B ) [41] . The endosomes arising from parasomal sacs can be labeled with the styryl dye FM1-43 [42] . There may also be other pathways of uptake from the plasma membrane , as revealed by residual FM1-43 uptake under conditions of clathrin or dynamin inhibition [42] , but this has not been rigorously examined . We identified endosome-associated GFP-Rabs as those Rabs showing significant colocalization with the styryl dye FM4-64 , after first confirming that FM4-64 labels the identical compartments as FM1-43 ( Figure S5A ) . In total , nine of the Rabs in Tetrahymena co-localized with FM4-64 ( examples in Figure 3; full set in Figure S6 ) . Only a subset of these Rabs were related to endocytic or endocytic/recycling Rabs from other organisms ( Figure 3E ) . The Rabs show a variable level of overlap with the endocytic tracer , consistent with the idea that multiple , sequential compartments are labeled ( Figure S5C ) . Moreover , the endocytic Rabs can be divided into three broad classes . The 1st group , a single Rab , labels what are likely to be parasomal sacs ( Figure 3B ) . The 2nd group labels small vesicles that are widely distributed throughout the cell ( Figure 3C ) . The 3rd group labels larger vesicles or tubular structures that are concentrated in the cell posterior , often in clumps . Based on previous studies , these are likely to represent a compartment to which FM1-43 gains access several minutes post internalization , and which may function as recycling endosomes ( Figure 3D ) . Tetrahymena secrete newly synthesized proteins via at least 3 recognized pathways beginning with translocation into the endoplasmic reticulum ( ER ) . The Tetrahymena ER is present as both a cortical as well as cytoplasmic reticulum , while the Golgi are present in multiple copies as small stacks of cisterna that are adjacent to mitochondria , close to the cell periphery [43] ( Figure 4A ) . The 3 known secretory pathways are functionally equivalent to known pathways in mammalian cells: a pathway of rapid constitutive secretion , a pathway of secretion via exocytic fusion of lysosomes , and regulated exocytosis from dense core granule-like bodies called mucocysts . However , the only exocytic sites or secretory carriers that have been directly visualized in Tetrahymena are the dense core granules , which dock and undergo exocytic fusion along both the 1° and intervening 2° meridians , at junctions between adjacent alveoli [44] . Tetrahymena has a single , very highly expressed member of the Rab1 clade , associated with ER-to-Golgi traffic . This association is robust in Bayesian and Neighbor Joining trees , but not Maximum Likelihood ( Figure S9 ) . TtRab1 shows a complex localization pattern , labeling mobile puncta that are primarily concentrated at the anterior end of the cell , and tubular or reticular structures both in the anterior and posterior of the cell ( Figure 4B ) . A 2nd , divergent Rab shows similar localization ( Figure S7B ) . Five Rabs , four of which belong to a conserved Golgi-associated clade , appear to localize to large puncta in a loose meridional array expected for Golgi stacks , and similar to that of a putative Golgi marker , Cda12p [45] ( Figure 4C; Figure S7C , S7D , S7E; Video S7 ) . Also consistent with the known location of Golgi , these puncta are positioned near but not at the cell periphery . TtRabD41 localizes to regularly-spaced puncta at both 1° and 2° meridians ( Figure 4Di ) and appears to co-localize with the secretory granule marker Grt1p [46] , in a pattern suggesting that this Rab may localize to granule docking sites ( Figure 4Dii and 4Diii ) . Since the Rab is also present at cortical sites between granules , it may also localize to unoccupied docking sites . The contractile vacuole functions to collect water from the cytoplasm and pump it to the cell exterior , to maintain osmotic balance . The complex structure includes a contractile bladder , from which water-collecting tubules extend into the cytoplasm ( Figure 5A ) [47] . Three divergent and unrelated Tetrahymena Rabs primarily labeled the contractile vacuole ( Figure 5F ) . TtRabD14 labels large vesicles that are clearly associated with the contractile vacuole but distinct from the central bladder , while TtRabD10 labels tubular extensions thereof ( Figure 5B and 5C ) . TtRabD2 shows diffuse labeling always centered on the contractile vacuole ( Figure 5D ) . In addition , several other GFP-Rabs show secondary contractile vacuole localization ( Table S2 ) . Active water pumping is required for all cells that lack cell walls and inhabit fresh water , and contractile vacuoles are accordingly found in diverse unicellular lineages . In Dictyostelium discoideum , a slime mold very distantly related from Ciliates , the Rabs associated with the contractile vacuole have been described [48] ( and refs therein ) . Based on phylogenetic analysis , none of the Tetrahymena contractile vacuole Rabs appears orthologous to any of the functionally related Dictyostelium Rabs ( Figure 5F ) . The remaining Tetrahymena Rabs , including proteins that localize to the nuclear envelope , cell cortex , ciliary basal bodies , and numerous other structures , are illustrated in Figure S8 . We used maximum likelihood , Bayesian , and neighbor-joining methods to understand the relatedness of the Tetrahymena Rabs to those in other lineages ( Figure S9 ) . Some of the bootstrap values for nodes linking the Ciliate sequences with those of other lineages were low , as expected in light of the known deep evolutionary divergence between the Alveolates ( Ciliates , Dinoflagellates , Apicomplexans ) and other lineages [49] . Nonetheless , these reconstructions produced strong evidence that 15 of the Tetrahymena Rabs align in clades with Rabs in animals and other lineages , and can therefore be considered as highly conserved . The conserved Tetrahymena Rabs fall within six predicted functional groups ( Figure 6 ) . Five of these clades were previously defined , by comparison of many eukaryotic genomes , to belong to a core set of eight whose wide distribution in existing eukaryotes implies that all may have been present in the last common eukaryotic ancestor [11] . The Tetrahymena Rabs in this group are predicted , based on sequence similarity , to be associated with ER-to-golgi traffic ( intermediate compartment ) ( 1 Rab ) , endocytosis ( 3 Rabs ) , endocytic recycling ( 5 Rabs ) , late endocytosis ( 1 Rab ) , and retrograde golgi traffic ( 4 Rabs ) . Tetrahymena does not appear to have representatives within three remaining core Rab clades . Two of these , including human Rabs 8 , 10 , 13 , 19 and 30 , are associated with Golgi traffic . The third clade is associated with regulated exocytosis , and includes human Rabs 3 and 27 . Strikingly , a single Tetrahymena Rab does not belong to the previously defined core set , but nonetheless aligns with a clade with representatives in both Opisthokonts ( H . sapiens , D . melanogaster ) and Amoebozoa ( D . discoideum ) ( Figure 6B ) . The animal Rabs in this group , which are all associated with transport of lysosome-related organelles , all share a WDIAGQE motif , which is also found in both the Dictyostelium and Tetrahymena members but not in Rabs from any other clade [50] . We propose that this clade , with representatives from at least three deeply divergent lineages , represents a previously unrecognized core Rab clade in eukaryotes . One feature that emerged from comparison of phylogenetic assignments and localization data was that some highly conserved Rabs failed to show clear association with the expected compartments . In particular , four of the putative endosomal Rabs showed little or no colocalization with FM4-64 . Three of these were localized to structures associated with the phagocytic pathway . Two of these , 11B and 4B , labeled highly mobile puncta at the oral apparatus , whose lack of overlap with FM 4-64 indicates that these are not vesicles pinching off from the plasma membrane ( Figure S4A , S4D ) . This is graphically reinforced by time-lapse movies of cells expressing TtRab11B-GFP , which show that labeled puncta , which are likely to be vesicles , are traveling toward the oral apparatus along a cytoplasmic fiber ( Video S6 ) , where they may be delivering bulk membrane required for phagocytosis [39] . TtRab4A labels vesicles that accumulate near the cytoproct , which is a zone where end-stage phagosomes fuse and eject undigested contents [51] . The fourth Rab in this group , TtRab31 , localized to puncta along cortical meridians ( Figure S8A ) .
Growing vs . starved Tetrahymena differ in pathways involved in feeding and secretion , while conjugation brings pronounced changes to the cell cortex , and to nuclear and other organelles [22] . We found that membrane traffic in starvation and conjugation chiefly involve the same set of Rabs that are expressed in growing cells . However , several Rabs are exclusively expressed in growth and/or starvation , some showing dramatic expression peaks at distinct stages . To understand the contributions of Rabs to cellular plasticity on an evolutionary timescale , one pertinent question is whether specific subsets of Rabs arose in Ciliates . Among the non-conserved Rabs , a subset of 14 aligned using all tree-building methods into seven clades containing only other Ciliate genes ( Figure 7 ) . Four of these clades are entirely composed of T . thermophila genes , suggesting four relatively recent paralogous expansions . In contrast , the remaining T . thermophila Rabs aligned in 3 clades with Rabs in Paramecium tetraurelia . These 14 Rabs may tentatively be considered as lineage-restricted . One interesting question is whether Ciliate-restricted Rabs are preferentially associated with pathways that arose in this lineage . Four of the currently recognized lineage-restricted Rabs localize to the oral apparatus , contractile vacuole , or cytoproct , which are structures that are likely to have undergone extensive elaboration , at a minimum , in the Ciliate lineage . The remainder localize to endosomes , the cell cortex , phagosomes , and as-yet undefined structures . The dataset also allowed us to ask whether structures or pathways that appear deeply conserved have conserved Rab determinants . In humans , 18 Rabs are associated with phagosomes , while 11 Tetrahymena Rabs were found on phagosomes . Only two of these Tetrahymena Rabs are orthologous to those in animals , TtRabs 32 and 7 ( Figure 2J ) . TtRab32 belongs to the Rab32/38 clade , whose members in animals are most strongly linked with transport of lysosome-related organelles . Similarly , Rab7 in both humans and Tetrahymena is associated with lysosomes as well as phagosomes . Therefore , most phagosomal Rabs in humans and Tetrahymena are unrelated , and the related Rabs are primarily associated with lysosomes . This implies either that mechanisms of phagosome maturation arose independently in animals and ciliates , or that this set of Rabs evolved under relatively few constraints in one or both lineages . In this regard , it is noteworthy that some phagosomal Rabs may have arisen within ciliates . The inference that phagosomes in Tetrahymena may have evolved separately from those in animals is consistent with the finding that phagosome-associated syntaxins in another Ciliate , Paramecium tetraurelia , form a lineage-restricted clade [61] . Similarly , we found no relatedness between Rabs associated with the contractile vacuole in Tetrahymena and in Dictyostelium . These results suggest that , in some cases , the similar cellular structures or pathways in different lineages do not primarily reflect constraints on an inherited ancestral pathway but rather parallel selective pressures leading via innovation to similar outcomes . This would depend , in the context of this paper , on Rabs that evolve novel functions , of which there are several good illustrations in the Tetrahymena cohort . First , several phylogenetically conserved Rabs were not associated with the predicted compartments , a finding that also underscores the importance of doing phylogenetic and localization analysis in parallel . Secondly , there is evidence for rapid Rab evolution within the lineage-restricted Rabs . Some lineage-restricted Rabs appear to have retained similar functions , judging by the cortical localization of TtRabs D41 , D29 and D35 . In contrast , TtRabs D39 and D4 are nearest neighbors in a single clade , but the former localizes to the cytoproct and oral primordium while the latter is found at endosomes . In summary , analysis of the large Rab gene family in Tetrahymena has provided an extensive set of new molecular markers for studies in this organism , and has provided insights into cellular and evolutionary aspects of membrane plasticity . Though Ciliate genes may be prone to undergoing fast evolution [54] , the observation that Ciliate and non-Ciliate protists show comparable ratios of conserved to divergent Rabs suggests that many of our observations are generalizable . This work sets the stage for functional analysis of informative Rab family members in this organism . More broadly , Rabs belong to a small set of proteins that , as products of large gene families , act as compartmental determinants of membrane traffic . The availability of sequenced genomes from divergent lineages highlights the need for combined approaches such as those taken here , to understand the consequences of gene family expansions and the evolutionary flexibility that is built into fundamental cell biological features , such as the complex network of membrane trafficking pathways that are crucial for homeostasis and signaling .
Beginning with amino acid sequences of known human Rabs ( accession numbers in [11] ) , we previously identified 70 putative T . thermophila Rabs by tblastn searches of the Macronuclear ( Mac ) genome , using the primary Tetrahymena hits as queries to detect additional Tetrahymena Rabs [23] . We subsequently annotated all genes using ESTs corresponding to 45 Tetrahymena Rabs for which ESTs became available during genome refinement [53] , aligning the ESTs with genomic sequences as well as the predicted mRNAs using Muscle ( MUltiple Sequence Comparison by Log-Expectation: http://www . ebi . ac . uk/Tools/muscle/index . html ) , and viewing the results using Seaview ( http://pbil . univ-lyon1 . fr/software/seaview . html ) . Confirmed mRNA sequences were translated ( including 5′ and 3′ UTRs ) in 3 frames to establish the ORF encoding the conserved Rab motifs and predict the start sites , which in general lay shortly upstream of the 1st conserved motif . Stop codons were identified as the 1st in-frame UGA ( the sole stop codon used in T . thermophila [62] . In this process , we identified 3 Rabs not detected in earlier work , adjusted several gene predictions , and disqualified 17 previously identified Rabs that were determined to be Rab-like . Forward and reverse primers ( Table S3 ) were designed to initiate from start and stop codons . All Rab genes were PCR amplified from genomic DNA isolated by phenol-chloroform extraction from strain CU428 . 1 , using Pfu Ultra polymerase ( Agilent Technologies , Santa Clara , CA ) , and confirmed by sequencing . GenBank ID numbers of all genes are listed in Table S1 . We used three different tree-building methods ( maximum likelihood [63] , Bayesian [64] , [65] and neighbor-joining [66] ) including all predicted Rabs from T . thermophila and H . sapiens and selected Rabs from D . melanogaster , A . thaliana , D . discoideum and P . tetraurelia . The basic tree topology presented in this manuscript is supported by all three distinct algorithms . In cases where detailed topological conclusions are supported by 2 out of 3 approaches , this is specified in the text . Protein sequences were aligned using Muscle [67] ( http://www . ebi . ac . uk/Tools/muscle/index . html ) , and gaps and the C-terminal hypervariable region of each Rab ( ∼30–80 residues ) were manually removed from the multiple alignment using Seaview [68] ( http://pbil . univ-lyon1 . fr/software/seaview . html ) . The output from alignment and gap removal were 167 total aligned sites , which were then used as input for bootstrap analysis using Seqboot from Phylip [69] ( http://evolution . genetics . washington . edu/phylip . html ) . 100 bootstraps were run . The bootstrapped outfile was used as the input for a maximum likelihood test using the Phyml program ( http://atgc . lirmm . fr/phyml/ ) , using the WAG substitution model . For Bayesian analysis , the bootstrapped outfile from Seqboot was used ( in Nexus format ) to run an analysis using MrBayes ( http://mrbayes . csit . fsu . edu/ ) for 100 , 000 generations , with sampling every 100 generations . For Tetrahymena Rabs that failed to align with any Rabs in H . sapiens , S . cerevisiae , or A . thaliana , we sought orthologs from other species using tblastn . The top-scoring putative homologs were then used in reverse BLAST searches to screen the Tetrahymena genome . If the original Tetrahymena Rab gene was the top-scoring hit in this reciprocal BLAST , the genes were considered potential orthologs and added to the large phylogenetic tree analysis . The chief outcome of this approach was addition of Paramecium genes to the phylogenies . Conserved Rabs were named according to their human ortholog . Divergent Rabs were numbered D1–D41 . The expression profiles of all Tetrahymena Rabs in growing , starved and conjugating cultures were downloaded from the whole genome expression database at http://tged . ihb . ac . cn/ [27] . The Gateway ( Invitrogen ) system was used to engineer each Rab as a GFP ( Green Fluorescent Protein ) fusion . First , PCR-amplified Rab genes were TOPO cloned ( Invitrogen ) into the pENTR-D-TOPO entry vector . CACC was added to each forward primer in order to allow directional cloning into pENTR-D . The pENTR clones were sequenced and the genes recombined using the Clonase reaction into the target Gateway-based T . thermophila expression vector pIGF-GTW , a gift from Doug Chalker [70] . We used site-specific mutagenesis to change the GFP gene in pIGF-GTW to the monomeric variant ( A206K ) [71] . Genes subcloned into pIGF-GTW are fused at their C-terminus , via a linker sequence , to the GFP gene , with the fusion under the transcriptional control of the cadmium-inducible MTT1 promoter [72] . When introduced into Tetrahymena , the vector is maintained as an extrachromosomal Mac plasmid and confers paromomycin resistance . TtRabD37 could not be cloned into pENTR; in addition , TtRab11C could not be expressed in Tetrahymena as a protein of the predicted size . T . thermophila strains were provided by Peter Bruns and Donna Cassidy-Hanley ( Cornell University ) ( CU428 . 1 ) and Eduardo Orias ( UC Santa Barbara ) ( B2086 ) . Cells were grown at 30°C in SPP media ( 1% proteose peptone , 0 . 2% dextrose , 0 . 1% yeast extract , 0 . 009% ferric EDTA ) with shaking . To reduce autofluorescence background in food vacuoles when imaging cells expressing GFP-tagged proteins , the cells were transferred to S medium ( 0 . 2% yeast extract plus 0 . 003% iron EDTA ) for 2h prior to imaging . For experiments requiring starvation ( also at 30° ) , cells were transferred to DMC , a one-tenth dilution of Dryl's ( 1 . 7 mM sodium citrate , 1mM NaH2PO4 , 1mM Na2HPO4 , 0 . 5 mM CaCl2 ) supplemented with an additional 0 . 1 mM MgCl2 and 0 . 5 mM CaCl2 . Transformation of GFP fusions was by conjugant electroporation [73] . Briefly , T . thermophila strains CU428 and B2086 , after 10h of conjugation , were combined with 20µg of DNA and electroporated; after 1d 100 µg/ml paromomycin sulfate was added to select for transformants , which were picked at 5 d . In general , the transformants were maintained at room temperature by weekly transfer in 24-well plates in SPP with paromomycin , and showed stable expression of the Rab-GFP for at least 6 weeks and often longer , depending upon the particular Rab . Transformants were stored in parallel in tube cultures ( in 2% proteose peptone , at room temperature , no paromomycin ) , and showed stable Rab-GFP expression for 2–3 months . The GFP-transgenes were induced with the lowest level of CdCl2 ( 0 . 25–1µg/ml , determined empirically for each strain ) that produced a discernible localization pattern . We analyzed 4 independent lines per transformation . The rare cases where differences were seen were due to variation in levels of what appeared to be diffuse cytoplasmic fluorescence . For imaging , transformants were grown overnight in SPP media and then transferred , 2h before imaging , to S medium at room temperature containing 0 . 25–1µg/ml CdCl2 . Cells induced in DMC required lower levels of CdCl2 ( 0 . 1–0 . 25µg/ml ) . Microscopy was performed at room temperature . For initial visualization , live GFP-Rab-expressing Tetrahymena cells were imaged on an Olympus DSU spinning disk inverted confocal microscope with a 100× objective . Cells were immobilized by mixing 1∶1 with either 6% polyethylene oxide ( PEO: MW = ca . 900 , 000 ) , or 6–10% poly ( ethylene glycol ) -polyalanine ( 2000–1500 ) diblock copolymer solutions that undergo sol-to-gel transition as the temperature increases [74] , both made in S media . Images were acquired with an EM-CCD Hamamatsu camera and SlideBook acquisition software . Unless indicated , figures are a projection of a z-stack of the entire cell or the cell mid-section . The Z projections and color channel merges , as well as adjustments to brightness and contrast , were made using the public domain NIH ImageJ program ( http://rsbweb . nih . gov/ij/ ) . Each image shown is representative of the majority of cells expressing that Rab-GFP . A fraction of the cells appeared to contain multivesicular phagosomes , which accumulated FM4-64 in cells exposed to this endocytic tracer ( see below ) and may be a response to stress . This fraction increased with time under the coverslip and was particularly evident in PEO-immobilized samples . Both for microscopy and for quantitative analysis , we chose cells with few or no such structures . Time-lapse and/or image stack movies of all Rabs , as well as additional images of some Rabs showing colocalization ( see below ) , are accessible at tetrahymenacell . uchicago . edu . Tetrahymena transformants were incubated for 5min with 5µM FM 4-64 dye [75] ( Invitrogen ) ( excitation/emission maxima ∼515/640 nm ) . The cells were then pelleted and resuspended in ∼40µl of supernatant , and immobilized as above for microscopy for up to 2h . In all cases , Rabs that showed colocalization with FM4-64 showed this pattern within 10 min . Cells were imaged on a Leica TCS SP2 AOBS laser scanning confocal microscope with LCS Leica confocal software with simultaneous capture in the green ( 495–520 nm ) and red ( 656–746 nm ) channels . Images shown are single slices for clarity . As above , adjustments in brightness and contrast were conducted in ImageJ . Phagosomes were labeled with india ink by incubating Tetrahymena for 1–2 h prior to imaging with 2 . 5% v/v india ink in S medium . The cells were simultaneously induced with CdCl2 ( for 2h ) to induce transgene expression . Alternatively , phagosomes were labeled by incubating Tetrahymena in S medium for 2h with E . coli expressing dsRed-Express2 [76] ( gift of R . Strack and B . Glick , U . Chicago ) ( adding 0 . 5% v/v of an overnight culture ) , with simultaneous induction of Rab-GFP expression . Cells were imaged using a Zeiss Axioplan2 upright widefield microscope with a Hamamatsu camera and Axiovision software ( using a 63× objective ) . Lysosomes and phagosomes were labeled in cells expressing TtRab7-GFP by addition of 0 . 05% ( v/v ) LysoTracker ( Invitrogen ) for 90min to cells in SPP at 22°C , in which Rab-GFP expression was induced for a total of 150min with 0 . 5µg/ml CdCl2 . The cells were pelleted and then resuspended in PEO for immobilization and microscopy . Cells were imaged using a Zeiss LSM-510 laser scanning confocal microscope with LSM 5 software . Overlap quantification ( e . g . , a Rab-GFP and FM4-64 ) was performed using NIH ImageJ , based on the method described in [77] . Briefly , pixels of interest were identified by generating a mask for each channel to eliminate background signal . Single focal plane images were used and we defined a threshold value in red and green channels separately that included only the brightest labeled structures . Overlap was defined as the percentage of total signal ‘intensity’ present in shared pixels . To recover intensity lost in the creation of the masks , we modified the binarized masks using subtract and invert functions in ImageJ to regain the green and/or red intensity values above threshold . To calculate overlap for the green signal we used the ratio of the average intensity of the green pixels in the AND mask over that of the green mask . We did the converse for the red signal . For each line , at least 4 cells were analyzed . Secretory granules ( mucocysts ) were immunostained with the 4D11 or 5E9 monoclonal antibodies as described in [78] , except fixation was at room temperature rather than on ice . Basal bodies were stained using the 20H5 anti-centrin antibody ( gift of Jeff Salisbury , Mayo Clinic ) , with Texas Red-conjugated goat anti-mouse 2° antibody ( Invitrogen ) . Immunostained cells were mounted with 0 . 1µM Trolox to inhibit bleaching and imaged the Leica SP2 laser scanning confocal microscope as described above . Whole cell lysates were prepared by 10% trichloroacetic acid precipitation ( ∼1×106 cells in SPP per sample ) and then dissolved in SDS-PAGE sample buffer . For western blots , 2×105 cell equivalents per sample were resolved by SDS-PAGE , transferred to 0 . 45µM nitrocellulose ( General Electric Water and Process Technologies ) , probed with polyclonal anti-GFP antibody at 1∶1 , 000 dilution ( Invitrogen ) and Alexa 680 anti-rabbit antibody ( Invitrogen ) at 1∶5 , 000 dilution , and imaged using the Li-Cor Odyssey protocol and scanner ( http://www . licor . com ) .
|
Single-celled organisms appear simple compared to multicellular organisms , but this may not be true at the level of the individual cell . In fact , microscopic observations suggest that protists can possess networks of organelles just as elaborate as those in animal cells . Consistent with this idea , recent analysis has identified large families of genes in protists that are predicted to act as determinants for complex membrane networks . To test these predictions and to probe relationships between cellular structures across a wide swath of evolution , we focused on one gene family in the single-celled organism Tetrahymena . These genes control the traffic between organelles , with each gene controlling a single step in this traffic . We asked three questions about each of 56 genes in the family . First , what is the gene related to in humans ? Second , under what conditions is the gene being used in Tetrahymena ? Third , what is the role of each gene ? The results provide insights into both the dynamics and evolution of membrane traffic , including the finding that some pathways appearing both structurally and functionally similar in protists and animals are likely to have arisen independently in the two lineages .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"evolutionary",
"biology/microbial",
"evolution",
"and",
"genomics",
"evolutionary",
"biology/genomics",
"cell",
"biology/membranes",
"and",
"sorting",
"evolutionary",
"biology/developmental",
"molecular",
"mechanisms",
"evolutionary",
"biology/morphogenesis",
"and",
"cell",
"biology"
] |
2010
|
Comprehensive Analysis Reveals Dynamic and Evolutionary Plasticity of Rab GTPases and Membrane Traffic in Tetrahymena thermophila
|
CRISPR-Cas9 technology can be used to engineer precise genomic deletions with pairs of single guide RNAs ( sgRNAs ) . This approach has been widely adopted for diverse applications , from disease modelling of individual loci , to parallelized loss-of-function screens of thousands of regulatory elements . However , no solution has been presented for the unique bioinformatic design requirements of CRISPR deletion . We here present CRISPETa , a pipeline for flexible and scalable paired sgRNA design based on an empirical scoring model . Multiple sgRNA pairs are returned for each target , and any number of targets can be analyzed in parallel , making CRISPETa equally useful for focussed or high-throughput studies . Fast run-times are achieved using a pre-computed off-target database . sgRNA pair designs are output in a convenient format for visualisation and oligonucleotide ordering . We present pre-designed , high-coverage library designs for entire classes of protein-coding and non-coding elements in human , mouse , zebrafish , Drosophila melanogaster and Caenorhabditis elegans . In human cells , we reproducibly observe deletion efficiencies of ≥50% for CRISPETa designs targeting an enhancer and exonic fragment of the MALAT1 oncogene . In the latter case , deletion results in production of desired , truncated RNA . CRISPETa will be useful for researchers seeking to harness CRISPR for targeted genomic deletion , in a variety of model organisms , from single-target to high-throughput scales .
CRISPR/Cas9 is a simple and versatile method for genome editing that can be applied to deleting virtually any genomic region for loss-of-function studies . Deletion requires the design of optimal pairs of single guide RNA ( sgRNA ) molecules that hybridise to sequences flanking the target region . While this approach is being employed for diverse applications , from single target studies [1–3] to high parallelized screening studies [4 , 5] , there presently exists no bioinformatic solution for selection of optimal pairs of sgRNAs . We present here a highly customisable design pipeline to address the needs of all such deletion projects , regardless of scale . CRISPR/Cas9 makes it possible to investigate the function of genomic elements in their endogenous genetic context . The Cas9 nuclease is recruited to desired genomic sites through its binding to an engineered , single guide RNA ( sgRNA ) [6] . Early studies focussed on protein coding genes , utilizing individual sgRNAs to induce small indel mutations in genomic regions encoding target proteins’ open reading frame ( ORFs ) . Such mutations frequently give rise to inactivating frameshift mutations , resulting in complete loss of function [7 , 8] . The delivery of a single sgRNA in such experiments is technically straightforward , and can be scaled to genome-wide , virally-delivered screens . CRISPR has also been brought to bear on non-coding genomic elements , including regulatory regions and non-coding RNAs , which have traditionally resisted standard RNA interference ( RNAi ) [2 , 9] . With some exceptions ( for example [10] ) , functional knockout of non-coding elements with a single sgRNA is not practical , because small indel mutations caused by single sgRNAs are less likely to ablate function to the same extent as in a protein-coding sequence . Instead , a deletion strategy has been pursued: a pair of sgRNAs are used to recruit Cas9 to sites flanking the target region [2 , 4] . Simultaneous strand breaks are induced , and non-homologous end joining ( NHEJ ) activity repairs the lesion . In a certain fraction of cases , this results in a genomic deletion with a well-defined junction [4] . Cas9 targeting is achieved by engineering the 5’ variable region of the sgRNA . This hybridises to a complementary “protospacer” region in DNA , immediately upstream of the “protospacer adjacent motif” ( PAM ) [11] . For the most commonly-used S . pyogenes Cas9 variant , the PAM sequence consists of “NGG” . A growing number of software tools are available for the selection of optimal individual protospacer targeting sequences [12–18] . The key selection criteria are ( 1 ) the efficiency of a given sequence at generating mutations , and ( 2 ) “off-targeting” , or the propensity for recognising similar , yet undesired , sites in the genome . Based on experimental data , scoring models for on-target efficiency have been developed , for example that presented by Doench et al [16] . At the same time , tools have become available for identifying unique sgRNA sites genome-wide , mitigating to some extent the problem of off-targeting [19] . However , few tools presented so far are designed for large-scale designs , and to the best of our knowledge , none was created to identify optimal sgRNA pairs required for deletion studies . To address this need , we here present a new software pipeline called CRISPETa ( CRISPR Paired Excision Tool ) that selects optimal sgRNAs for deletion of user-defined target sites . The pipeline has several useful features: first , it can be used for any number of targets in a single , rapid analysis; second , it returns multiple , optimal pairs of sgRNAs , with maximal predicted efficiency and minimal off-target activity; third , the user has control over the full range of design parameters . The pipeline is available as both standalone software and as a user-friendly webserver . In addition , we make available a number of pre-designed deletion libraries for various classes of non-coding genomic elements in a variety of species . Finally , using a quantitative deletion assay , we find that CRISPETa predictions are highly efficient in deleting fragments of a human gene locus , resulting in detectable changes to the cellular transcriptome . CRISPETa is available at http://crispeta . crg . eu .
To address the need for bioinformatic design of paired sgRNAs for genomic deletion , we created the CRISPETa pipeline ( Fig 1A and 1B ) . The guiding principles of CRISPETa are flexibility and scalability: the user has control over all aspects of the design process if desired ( otherwise reasonable defaults are provided ) , and the design may be carried out on individual targets , or target libraries of essentially unlimited size . The full set of user-defined variables , and their default values , are shown in Table 1 . The core objective of sgRNA design is the selection of optimal “protospacers” , defined as the 20 bp of genomic DNA sequence preceding the PAM sequence [11] . This is distinct from the sgRNA sequence itself , composed of the protospacer sequence and the constant , scaffold region ( Fig 1A ) . The CRISPETa workflow is divided into three main steps: target region definition , protospacer selection , and sgRNA pair prioritisation ( Fig 1B ) . Given a genomic target region or regions in BED format , CRISPETa first establishes pairs of “design regions” of defined length in which to search . Design regions may be separated from the target itself by “exclude regions” of defined length . The user may also specify “mask regions”: sgRNAs falling within the positive mask are prioritised , whereas those within the negative mask will be de-prioritized ( although not removed altogether ) . Positive masks might include regions of DNaseI-accessible chromatin , while negative masks may be composed of , for example , repetitive regions or compact chromatin . Using this information , the entire set of potential protospacers is defined . First , the design region sequence is extracted and searched for all possible 20mer sites followed by canonical S . pyogenes “NGG” PAM sites–candidate protospacers . These are considered with respect to two core metrics: their potential for off-target binding , and their predicted efficiency . Off-targeting , or the number of identical or similar sites with a given number of mismatches , is estimated using precomputed data for each genome . This strategy increases the speed of CRISPETa dramatically . We created off-target databases for five commonly-studied species , human , mouse , zebrafish , Drosophila melanogaster , Caenorhabditis elegans ( Table 2 ) , varying widely in genome size ( Fig 2A ) . The default off-targeting cutoff is set at ( 0:1 , 1:0 , 2:0 , 3:x , 4:x ) , that is , sequences having no other genomic site with ≤2 mismatches ( in our notation , “x” represents infinity ) . At this default , 78% of candidate protospacers are discarded in human , compared to just 29% in Drosophila , reflecting the relative uniqueness and compactness of the latter ( compare dark blue bars in Fig 2B ) . To estimate their efficiency in inducing double stranded breaks at their target sites , candidate protospacers are scored using the logistic regression measure of Doench et al[16] . This model was trained on experimental assays for 6085 and 1151 sgRNAs tiled across six mouse and three human genes , respectively . This score predicts sgRNA efficiency based on informative nucleotide preferences both within the core 20mer and in its immediate flanking nucleotides . Protospacers passing defined off-target and on-target thresholds are retained–henceforth referred to as “filtered protospacers” . In contrast to off-target filtering , efficiency score filters are more consistent across genomes , removing 60–70% of protospacers in the five genomes tested ( light blue bars in Fig 2B , Table 2 and Supplementary S8 File ) . Together , off-target and efficiency score filters eliminate 96% of candidate protospacers in human ( Fig 2B ) , but nevertheless yielding an average density of 6 . 6 usable protospacers per kilobase ( Fig 2A ) . Comparison across species shows that there is markedly lower density of usable filtered protospacer sequences in vertebrates compared to invertebrates ( Fig 2A ) . In general , and even after applying off-targets and score filters , the minimum deletion size constrained by sequence features alone is less than 150 bp for the majority of genomic regions ( Supplementary S7 File ) . In the final step , optimal sgRNA pairs are selected . First , all possible pairs of filtered protospacers are enumerated and ranked . Two ranking approaches are available: by combined efficiency score ( default ) , or by length of deleted region . Ranking by score will tend to result in pairs that are more evenly distributed throughout the targeting region , but with a heterogeneous distribution of deletion sizes . Conversely , ranking by length favours shorter intervals within the constraints of the targeting design . Short segments may be more efficiently deleted [20] , but will tend to be clustered into a smaller genomic region . The top-ranked pairs , up to a user-defined maximum of n , are returned for each feature . In principle , a single high-scoring sgRNA may end up contributing to many or all of the highest-scoring pairs . To control this process , the “diversity” measure is used to control the maximum fraction of pairs containing a single sgRNA sequence ( Table 1 ) . Finally , the user may specify constraints in sgRNA pair selection based on the plasmid construction method . Many plasmids employ the U6 promoter , which requires the sgRNA to commence with a “G” . For instance , the DECKO plasmid expresses two sgRNAs in tandem from U6 and H1 promoters , thus requiring the 5’ sgRNA to commence with G [4] . The “construction method” variable allows users to incorporate this constraint , specifically by ensuring that the first sgRNA commences with a natural or prepended G . CRISPETa returns a ranked series of paired sgRNA constructs for each target . Sequences are output in a format suitable for immediate ordering from commercial oligonucleotide synthesis services . Summary statistics and figures are produced for each design job . We tested the standalone pipeline using a set of 7000 human target genomic features compiled from a mixture of sources ( see Materials and Methods ) . At default settings , CRISPETa returns successful , full depth ( n = 10 ) designs for 68% of features , with a further 18% of partial depth ( 0<n<10 ) and 14% failures ( “default” in Fig 2C , Table 3 ) . We here define “full depth” to indicate the situation where all of n requested sgRNA pairs are successfully returned , and “partial depth” when the returned number is less than n . Performed on a workstation running CentOS6 , 86 . 6 Gb of memory and 12 CPUs ( Intel ( R ) Xeon ( R ) CPU E5649 @ 2 . 53GHz ) , the analysis took 44 minutes with a maximum RAM requirement of <100 MB . This benchmarking was repeated several times , in each case modifying a single parameter ( Fig 2C and Table 3 ) . As expected , strengthening the diversity requirement resulted in a drastic reduction of design success ( “diversity0” ) , while a complete relaxation ( “diversity1” ) did not produce a substantial gain . Some improvement was observed when relaxing off-targeting , but this benefit is negligible after “off1” ( allowing a single other match with two mismatches , ( 0:1 , 1:0 , 2:1 , 3:x , 4:x ) . As expected , increasing the paired score threshold has a strong effect on design depth , particularly after 0 . 6 ( “pScore0 . 6” ) ( the default is 0 . 4 ) . The most dramatic improvement was observed when the length of the design region was increased to 2000 bp , boosting the fraction of successfully targeted regions from 70% to 95% . Thus , by adjusting these parameters , the depth of library designs can be optimised for each target set . We next used CRISPETa to design knockout libraries for a variety of genomic element classes that cannot be targeted by traditional RNAi , either because their function is not thought to depend on RNA production ( eg ultraconserved elements [UCEs] ) [21] , or because their RNA product is too short ( eg microRNAs ) ( see Table 4 ) . We also created a collection of 3170 random intergenic target regions in human as a reference and for use as negative controls in screening projects . An example is shown in Fig 3A , created using the standard output of CRISPETa , where the IRX3 gene promoter and an upstream ultraconserved element ( UCE ) were targeted . The characteristics of these libraries are shown in Fig 3B and 3C and Table 4 . Overall , 68% of features could be targeted at full depth , with an additional 18% at incomplete depth . We observe considerable heterogeneity in the design success across classes , with protein-coding gene promoters reaching a full depth for 82% of cases , compared to 39% in random intergenic regions . We expect these differences arise from the former’s high sequence uniqueness ( decreasing off-target frequency ) and high GC-content ( increasing PAM density and predicted efficiency ) . To compare performance across species , we created designs targeting the entire annotated catalogue of microRNA genes in human , mouse , zebrafish , D . melanogaster and C . elegans ( Fig 3C ) . We observe considerably more efficient designs in non-mammalian species , likely reflecting their more compact , less repetitive nature . Nevertheless , at default settings we managed to create full or partial depth designs for 82% of human miRNA precursors , and this could likely be improved by altering design parameters . The entire set of designs are available for download from crispeta . crg . eu . Overall these results demonstrate the practicality of creating large-scale paired sgRNA knockout designs across diverse genomic element classes . We next evaluated the performance of CRISPETa designs in an experimental setting . As an model , we focussed on the MALAT1 locus , which expresses a potent oncogenic mono-exonic lncRNA [22] . In previous work , we managed to delete the MALAT1 promoter using pairs of sgRNAs delivered by a lentiviral vector , pDECKO [4 , 22] . We have created an updated version , pDECKO_mCherry ( hereafter referred to as “pDECKO” for brevity ) , carrying both antibiotic and fluorescence markers , into which designed sgRNA sequences can be rapidly cloned and expressed from independent promoters ( Fig 4A ) . We also developed a streamlined protocol for cloning these vectors , DECKO2 , described in detail in Supplementary S1 File . We selected two target regions: a conserved upstream element with enhancer-like chromatin modifications ( “enhancer” ) and a region of conserved exonic sequence ( “exon” ) ( Fig 4C ) . Each was submitted to CRISPETa , and from the resulting sgRNA designs we selected the three highest scoring pairs and one lower scoring pair ( details can be found in Supplementary S5 File ) . HEK293T cells , stably expressing Cas9-BFP , were transfected with pDECKO , and selected by antibiotic resistance for 6 days , after which their genomic DNA ( gDNA ) was extracted ( Fig 4D ) . In order to observe genomic deletion , we used two distinct PCR-based methodologies . The first , non-quantitative approach allows one to verify the correct size of deleted regions using primers flanking the target region ( Fig 4E , left panels ) . We used this “conventional” approach to genotype MALAT1 deletions in a previous study [4] . The second approach , which we call “quantitative CRISPR PCR” ( QC-PCR ) , allows one to estimate the deletion efficiency , in terms of percent of wild-type ( uncut ) alleles remaining in a cellular population ( Fig 4B and 4E , right panels ) . In tests using mixtures of wild type and deleted alleles , QC-PCR could accurately estimate known concentrations ( Supplementary S6 File ) . The primer configurations used by both approaches are shown as black arrowheads in Fig 4C and 4E . We used both conventional genotyping and QC-PCR to investigate target region deletion in gDNA of transfected HEK293T cells ( Fig 4D ) . Conventional PCR genotyping , using out-out primers , yielded amplification product sizes consistent with target site deletion for all pDECKO constructs , but not for control cells ( Fig 4E , left panels ) . QC-PCR analysis , using in-out primers , of independent biological replicates showed loss of ~40% of enhancer target sites for each of the four sgRNA pair designs targeting the enhancer region ( Fig 4E , top right ) . A non-targeted genomic region was not affected ( “Non-targeted” ) . Higher efficiencies were observed for the exon-targeting constructs , yielding >60% efficiency for the top two sgRNA pairs ( Fig 4E , bottom right ) . We did not observe a strong difference in the deletion efficiency between the four sgRNA pairs targeting the enhancer , although for the exon region , the lower-scoring two constructs displayed reduced efficiency . This underlines the value of using predicted efficiency scores in sgRNA selection , and supports the effectiveness of CRISPETa-predicted sgRNA pairs . We next sought to verify that the engineered deletions in the MALAT1 exon result in the expected changes to transcribed RNA . cDNA was generated from bulk cells treated with pDECKO vectors targeting MALAT1 exon . Given that not all cells have both alleles deleted , this sample should contain a mixture of RNA from both wild-type and mutated alleles . RT-PCR using primers flanking the targeted region amplified two distinct products , of sizes expected for wild-type and deleted sequence ( Fig 5A ) . TA cloning and Sanger sequencing of individual cDNA clones revealed a variety of deletion sites around the expected position within the MALAT1 exon ( Fig 5B ) . Therefore , targeted deletions by CRISPETa are reproduced in the transcriptome , and may be used in future dissect RNA functional elements .
We have here presented a versatile and scalable design solution for CRISPR deletion projects . To our knowledge , CRISPETa is the first tool for selection of optimal sgRNA pairs . A key feature is its scalability , making it equally suitable for focussed projects involving single target regions , and screening projects involving thousands of targets . The user has a large degree of control over the design process , enabling projects to be optimised for target regions with diverse sequence uniqueness and GC content . On-target efficiency is predicted using the latest , experimentally-informed design algorithm , while running speed is boosted by an efficient off-target calculation . A growing number of laboratories are adopting CRISPR deletion in their research for diverse applications , including modelling of human genetic disease [1] , functional dissection of enhancer elements [3] or insulators [23] , or loss-of-function studies on small or long noncoding RNAs [2] . In each case , it was necessary to manually design pairs of sgRNAs using available , single sgRNA design tools . There is clearly ample space to streamline this process . The second main application for CRISPR-deletion is for high-throughput loss-of-function screening studies , through the cloning of complex , pooled targeting libraries . These have enormous potential for the systematic identification of functional , non-coding genomic elements for the first time [3] . Manual design of paired sgRNAs for such projects is clearly out of the question . CRISPETa has been designed with both types of project in mind . The QC-PCR technique presented here now allows one to quantify and compare the efficiency of CRISPETa designs . For the 8 sgRNA pairs in two regions that we tested , deletion efficiencies of ~40–60% were consistently observed . Given that DECKO gives rise to an approximately equal mixture of heterozygous and homozygous mutants [4] , this would imply that over half of the cells in the mixture are being mutated . The induced deletions , when occurring within a transcribed region , are also observed in expressed RNA molecules . This is , to our knowledge , the first demonstration of the production of truncated RNA from an edited locus . It should be noted that our understanding of on- and off-target sgRNA efficiencies is evolving rapidly . The Doench score used here is trained on a limited number of protein-coding genes , and it is likely that its scoring algorithm will be further refined in the near future . We plan to incorporate such improvements into CRISPETa as they become available . Users who wish to omit the on-target filter , may simply set the on-target score thresholds to zero . Similarly , to remove off-target filters , users may set all mismatch settings to infinity . CRISPR enables us to study the function of non-coding genomic elements in their endogenous cellular context for the first time . The power of CRISPR-Cas9 genome-engineering lies both in its versatility , but also in its ready adaptation to large-scale screening approaches . The CRISPETa pipeline and experimental methods described here will , we hope , be useful for such studies .
The pipeline is outlined in Fig 1A . As input , CRISPETa requires a standard BED6-format file describing one or more target regions of the supported genomes . Presently these are hg19 ( human ) , mm10 ( mouse ) , danRer10 ( zebrafish ) , dm6 ( Drosophila ) , ce11 ( C . elegans ) . The webserver also directly accepts input as sequence , in FASTA format . Unstranded entries are assigned to the + strand , while those without identifiers are assigned a random ID . CRISPETa first defines design regions based on parameters g/du/dd/eu/ed ( see Table 1 for full list of parameters ) ( Fig 1A and 1B ) , and extracts their sequences using the BEDtools getfasta function . Design regions are searched for canonical PAM elements ( NGG ) using a regular expression . For every such PAM , a total of 30 nucleotides ( NNNN[20nt]NGGNNN ) are stored . Protospacers containing the RNA Pol III stop sequence ( TTTTT ) are removed . Next , candidate protospacers are searched against a precomputed , database-stored list of potential protospacers and their number of similar sequences with up to 4 mismatches , genome-wide ( see “Off-target analysis” section , below ) . Matches ( ≤2 mismatches ) to non-canonical ( “NAG” ) protospacers in any annotated protein-coding region are excluded from all analyses . By default , protospacers with one or more off-targets with ≤2 mismatch are discarded ( this cutoff can be modified by the user through parameter t ) . Remaining protospacers are then compared with the positive and negative mask BED files using BEDtools intersectBed . Candidate sequences not fully overlapping the positive mask file , or overlapping the negative mask by one basepair , are tagged as “disfavoured” . Next , 30mer regions encompassing remaining protospacers , including disfavoured ones , are assigned an efficiency score ( see below ) between 0 and 1 , and those above the score threshold ( controlled by parameter si ) are carried forward . Next , candidate sequences are assembled into pairs and filtered . For each target region , all possible pairs of upstream and downstream candidates are generated . If pairs are designed for DECKO cloning ( which utilizes the U6 promoter for the 5’ sgRNA gene , controlled by c ) , an additional step is applied: sgRNA pairs , where one of the pairs commences with G , are reordered as necessary such that the first sgRNA starts with G; for pairs where neither commences with G , an additional G is prepended to the first sgRNA[24] . It should be noted that this “DECKO construction” mode thus results in oligonucleotide libraries that vary in length by one nucleotide . A combined score for the resulting pairs is computed . By default , this is the sum of the two individual sgRNA scores , but users may choose to define the pair score as the product of individual scores ( parameter sc ) . Pairs are now filtered with a pair score threshold , and ranked first by mask score and then by pair score ( or , optionally , reverse ranked by distance , using parameter r ) . An optional “diversity” cutoff can be used to remove pairs such that no individual candidate sequence appears in more than a given fraction of returned pairs ( parameter v ) . Finally the program returns the top ranked pairs up to the maximum number specified by the user , n . CRISPETa is implemented in Python and available for download from git-hub and the CRISPETa web-server ( see availability below ) . All target sets and mask files were prepared in BED format , and obtained in April 2016 . Coding genes were obtained from the Gencode v19 annotation , filtered for the “protein_coding” biotype [25] . CTCF binding sites for GM12878 cells were downloaded from ENCODE data hosted in the UCSC Browser [26] . Enhancers were obtained from Vista [27] . Pre-miRNAs were obtained from miRBASE [28] . Disease-associated SNPs were obtained from the GWAS database ( http://www . ebi . ac . uk/gwas/api/search/downloads/full ) . Ultraconserved regions were obtained from UCNEbase [21] . For human positive and negative masks we used DNaseI hypersensitive sites identified through genome-wide profiling in 125 diverse cell and tissue types by the ENCODE consortium [29] and RepeatMasker repetitive regions [30] , respectively . To generate random intergenic locations , the entire span of all Gencode v19 genes ( both coding and noncoding , introns and exons ) , in addition to 100 kb up- and downstream , were subtracted . Random locations were selected within the remaining regions . Off-target analysis was performed using Crispr-Analyser [19] . We searched for all canonical PAM regions ( NGG ) in the genome and stored the 20nt that precedes each . Then using “search” and “align” options we obtained the number of off-targets with 0 , 1 , 2 , 3 and 4 mismatches for each unique 20mer . A second step was performed to remove protospacers with potential off-targeting in coding regions: for each genome , all 20mers followed by NGG were mapped using BWA mapper against sequences of 20nt followed either by NGG or NAG in all annotated protein-coding regions [31] . Those 20mers having alignments with ≤2 mismatches were removed in order to avoid potential off-targets in coding regions . Filtered 20mers were stored in a MySQL database . Precomputed files containing this information for various genomes can be directly downloaded ( see “CRISPETa availability” section ) . Downloadable files contain 6 comma-separated fields in this order: sequence of the sgRNA without the PAM sequence and the number of off-targets with 0 , 1 , 2 , 3 , and 4 mismatches for this sgRNA . These files can be used as input for CRISPETa-MySQL module to generate the MySQL database . CRISPETa uses the scoring method developed by Doench et al [16] , based on an experimentally trained logistic regression model employing 72 sequence features . The code was downloaded from http://www . broadinstitute . org/rnai/public/analysis-tools/sgrna-design-v1 . A test target set was assembled from 1000 randomly-selected elements from each of the individual target annotations , for a total of 7000 . Benchmarking analyses were run on a workstation running CentOS6 , 86 . 6 Gb of memory and 12 CPUs ( Intel ( R ) Xeon ( R ) CPU E5649 @ 2 . 53GHz ) . CRISPETa can be run through the web-server ( http://crispeta . crg . eu ) or locally . The software runs on python2 . 7 . In order to run CRISPETa locally two additional programs are required: BEDtools and MySQL . Source code to run locally can be found at git-hub ( https://github . com/guigolab/CRISPETA ) and also at the “Get CRISPETa” section of the web-server . Source code consist of two scripts: CRISPETA . py that execute the main pipeline described above , and crispeta_mysql . py that helps users to create the off-target MySQL database . Two other files can be found within the source code: func . py that contains all functions necessary to execute the two main scrips , and config . py that stores the information needed to login to MySQL . We used a modified version of our previously-described protocol for the creation of pDECKO_mCherry vectors expressing pairs of sgRNAs , DECKO2 [4] . A detailed protocol is available in Supplementary S1 File , as well as on CRISPETa webpage . Selected sgRNA pairs were converted to overlapping series of 6 oligonucleotides using a custom design spreadsheet ( available as Supplementary S2 File ) . All described plasmids are available from Addgene . org under plasmid numbers 78534–78545 . gDNA was extracted with GeneJET Genomic DNA Purification Kit ( Thermo Scientific ) and quantitative real time PCR ( qPCR ) from 1 . 6 ng of purified gDNA was performed using Lightcycler 480 SYBR Green master kit ( Roche ) on a LightCycler 480 Real-Time PCR System ( Roche ) . Primer sequences can be found in Supplementary S3 File . Target sequence primers ( TFRC_B out F / TFRC_B in R , Enhancer in F / Enhancer out R for enhancer , Exon in F / Exon out R for exon ) were normalised to primers GAPDH F/R amplifying a distal , non-targeted region . Another non-targeting primer set , LdhA F/R were treated in the same way . Data were normalised using the ΔΔCt method [32] , incorporating primer efficiencies . The latter were estimated using a dilution series of gDNA , and efficiency calculated by the slope of the linear region only ( Supplementary S4 File ) . We noted a decrease in efficiency at high template concentrations . For testing the accuracy of the QC-PCR method , we used genomic templates containing known proportions of a target allele from our previous study [4] . Genomic DNA from a heterozygous clone for TFRC gene of the human , diploid cell line HCT-116 , was used , combined with WT gDNA in controlled proportions .
|
CRISPR-Cas9 is a revolutionary biological technique for precisely editing cells’ genomes . Amongst its many capabilities is the deletion of defined regions of DNA , creating a wide range of applications from modelling rare human diseases , to performing very large knock-out screens of candidate regulatory DNA . CRISPR-Cas9 requires researchers to design small RNA molecules called sgRNAs to target their region of interest . A large number of bioinformatic tools exist for this task . However , CRISPR deletion requires the design of optimised pairs of such RNA molecules . This manuscript describes the first pipeline designed to accomplish this , called CRISPETa , with a range of useful features . We use CRISPETa to design comprehensive libraries of paired sgRNA for many thousands of target regions that may be used by the scientific community . Using CRISPETa designs in human cells , we show that predicted pairs of sgRNAs produce the expected deletions at high efficiency . Finally , we show that these deletions of genomic DNA give rise to correspondingly truncated RNA molecules , supporting the power of this technology to create cells with precisely deleted DNA .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biotechnology",
"genome",
"engineering",
"engineering",
"and",
"technology",
"synthetic",
"biology",
"genomic",
"library",
"screening",
"invertebrate",
"genomics",
"synthetic",
"bioengineering",
"crispr",
"genomic",
"databases",
"genome",
"analysis",
"mammalian",
"genomics",
"molecular",
"biology",
"techniques",
"synthetic",
"genomics",
"research",
"and",
"analysis",
"methods",
"bioengineering",
"synthetic",
"genome",
"editing",
"sequence",
"analysis",
"genomic",
"libraries",
"bioinformatics",
"biological",
"databases",
"molecular",
"biology",
"animal",
"genomics",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"sequence",
"databases",
"library",
"screening",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"computational",
"biology"
] |
2017
|
Scalable Design of Paired CRISPR Guide RNAs for Genomic Deletion
|
Mycetoma , one of the badly neglected tropical diseases , it is a localised chronic granulomatous inflammatory disease characterised by painless subcutaneous mass and formation of multiple sinuses that produce purulent discharge and grains . If untreated early and appropriately , it usually spread to affect the deep structures and bone resulting in massive damage , deformities and disabilities . It can also spread via the lymphatics and blood leading to distant secondary satellites associated with high morbidity and mortality . To date and despite progress in mycetoma research , a huge knowledge gap remains in mycetoma pathogenesis and epidemiology resulting in the lack of objective and effective control programmes . Currently , the available disease control method is early case detection and proper management . However , the majority of patients present late with immense disease and for many of them , heroic substantial deforming surgical excisions or amputation are the only prevailing treatment options . In this communication , the Mycetoma Research Center ( MRC ) , Sudan shares its experience in implementing a new holistic approach to manage mycetoma patients locally at the village level . The MRC in collaboration with Sennar State Ministry of Health , Sudan had established a region mycetoma centre in one of the endemic mycetoma villages in the state . The patients were treated locally in that centre , the local medical and health personals were trained on early case detection and management , the local community was trained on mycetoma advocacy , and environmental conditions improvement . This comprehensive approach had also addressed the patients’ socioeconomic constraints that hinder early presentation and treatment . This approach has also included the active local health authorities , community and civil society participation and contributions to deliver the best management . This holistic approach for mycetoma patients’ management proved to be effective for early case detection and management , optimal treatment and treatment outcome and favourable disease prognosis . During the study period , the number of patients with massive lesions and the amputation rate had dropped and that had reduced the disease medical and socioeconomic burdens on patients and families .
Mycetoma is a common neglected tropical disease , reported worldwide but endemic in many tropical and subtropical regions in what is known as the mycetoma belt and Sudan seems to have the highest endemicity [1 , 2] . It is a chronic granulomatous inflammatory disease caused by several true fungi or certain actinomycetes , and hence it is classified as eumycetoma and actinomycetoma respectively [3 , 4] . More than 70 organisms are incriminated in causing mycetoma [5 , 6] . It is believed that these causative organisms , which are soil inhabitants , are implanted in the subcutaneous tissue via traumatic inoculation [7 , 8] . Mycetoma usually spread to involve the skin , deep structures and bones leading to devastating destruction , deformities and disability [9 , 10] . Early localised disease is amenable to cure and good prognosis , however , the late advanced disease is characterised by high morbidity and can be fatal [11 , 12 , 13] . Mycetoma has serious medical , health and socioeconomic bearings on patients , families , and communities particularly in endemic areas [14 , 15 , 16] . To date , its global incidence and prevalence are not well documented as mycetoma is a neglected disease , not a notified or a reportable one . Furthermore , in most of the endemic regions , there is no proper disease surveillance system , especially in Sudan . Thus most of the reported cases are limited to anecdotal case reports and passive case detection [17 , 18] . Moreover , the disease susceptibility , resistance and route of infection are not well characterised [19 , 20] . Nevertheless , mycetoma is commonly seen in communities of poor hygiene and environmental conditions where population live in proximity to animals and their dungs . It is believed that thorn pricks and minor injuries are important routes of mycetoma infection . This is supported by the facts that mycetoma is seen more frequently in the feet of patients of low socioeconomic status , with poor hygiene and in villages with animals enclosures made of thorny trees [21 , 22] . Clinically , mycetoma starts as a small painless subcutaneous mass that gradually increases in size , then multiple sinuses with seropurulent discharge that contained grains of different colour and sizes develop [23 , 24 , 25 , 26] . Most patients present late with advanced disease and serious complications due to the painless nature of the disease , patients’ low health education level and lack of health facilities in endemic areas [27 , 28 , 29] . It affects all age groups , but children and young adults of low socio-economic status are affected most , leading to serious economic and social consequences [30 , 31 , 32] . The proper treatment of mycetoma depends on mycetoma type and disease extent . Numerous mycological and molecular tests are required to identify the causative organisms , and that include grain microscopy and culture , cyto-histopathological examinations and PCR identification [33 , 34 , 35 , 36 , 37 , 38] . Various imaging techniques such as X-ray , ultrasound , MRI , CT scans are required to determine the disease spread along the various body planes [39 , 40 , 41 , 42] . However , most of these tests and techniques are invasive , of low specificity and sensitivity , expensive for patients and health providers in endemic areas [37] . Currently , there is no point of care diagnostic test for mycetoma . Patients need to travel for long distances to regional centres to establish the diagnosis , and that is not always feasible due to their low socio-economic status , low health education , and roadblocks in particularly during the raining season . Patients need prolonged periods of management involving diagnosis , treatment both medical and surgical and regular follow up . Treatment may last at least one year for the minor lesions to resolve and several years for large lesions . Even after full recovery , patients need to be followed up closely for evidence of recurrence , which is not uncommon [43 , 44 , 45 , 46] . The currently available treatments for mycetoma are suboptimal and disappointing , characterised by low cure rate ( 28% ) and high patients’ follow up dropout ( 54% ) rate [47 , 48] . In general , actinomycetoma is treated by a combination of antibiotics , and for eumycetoma , a combination of antifungals and wide local surgical excision is needed . The available medication is not very effective , expensive , with many side effects and hence the high patients’ dropout rate [47 , 48] . The diagnostic tests and treatment are expensive that amount up to $2500 per year . Whereas to the annual income in Sudan is less than $400 per capita ( according to the UNDP , 2006 ) , that creates an enormous economic burden on the patients , their families , community and eventually the whole health system in the country [49] . It is interesting to note , worldwide , there are neither preventive or control measures nor programs for mycetoma [20] . The disease surveillance , especially in Sudan , is limited to targeted prevalence studies , case reports , and passive case detection . The absence of a standardised , centralised mycetoma surveillance system has far-reaching effects on how the existing interventions are delivered in a cost-effective and evidence-based manner . In summary , mycetoma is a very devastating endemic disease of the most underprivileged population whether socially , economically or in terms of development . Mycetoma patients tend to travel from distant remote parts of the country to central centres for the treatment . This causes high financial burden and delay in treatment initiation . Furthermore , the disabling nature of the disease hinders access to healthcare service for the majority of the patients . Thus health services decentralisation will improve the accessibility and equity of health services to patients and will directly drop the huge financial burden . With this background , this community-based study was conducted with the objective of applying a new holistic approach to the management of mycetoma patients at the village level . That included setting up a regional mycetoma centre with a telecommunication network , offering free of charge both medical and surgical treatment at the centre , training of medical and health staff on early case detection and management , community health education , improvement villages’ hygiene all these were based on available health system structure and minimum requirement .
Early case detection and management required house to house total coverage survey . That was conducted in 19 villages in the Eastern Sennar Governate . The data were collected by well-trained teams of medical officers , house officers , medical students , health care providers and community activists using a digital pre-designed validated closed-ended questionnaire in smart tablets . Computer Assisted Patients Identifier ( CAPI ) a computer application which was designed for this study was used . To validate the study questionnaire and the CAPI , a small pilot study was conducted before the data collection in a nearby village . The CAPI is a computer application predesigned for this study to collect data from the study villages and suspected patients . It was designed by an information and technology expert from the Faculty of Mathematical Sciences , University of Khartoum . It was used in computer tablets or smartphones . It can be used offline and online . CAPI was connected to the MRC , Data Centre system and the data analysis was performed spontaneously , the results can be displayed on Google maps , Fig 1 . The data collection questionnaire had included the suspected patient’s demographic characteristics; name , age , state and village localities , lesion site , the presence of mass , sinuses , grain colour , contact address , lesions photographs , the suspected patient’s locality geographic coordinates ( latitude , longitude , altitude ) and the neighbourhood photographs . All suspected mycetoma patients from the Governate were referred to Wad Onsa Mycetoma Regional Center ( WOMRC ) . The WOMRC was established in 2015 as a partnership between the Federal Ministry of Health , Ministry of Health , Sennar State , MRC and the local community to manage mycetoma patients locally in their region . The centre consists of small surgical operation complex , two wards , pharmacy , laboratory , ultrasound and out-patient suites and telemedicine facility connecting the WOMRC and the MRC , Fig 2 . At WOMRC centre , the patients were managed by the MRC mobile mission team with continuity of care provided by a surgical team from the regional Sennar Teaching Hospital and the resident doctor at the WOMRC in direct contact with the MRC team via the telemedicine facility . The diagnosis of mycetoma was established by careful clinical examination and lesion ultrasound examination by mobile ultrasound machine ( Paolus–UF-760AG ) conducted by the consultant radiologists . All mycetoma suspected patients underwent surgical excisions under general or spinal anesthesia by the consultants surgeons and surgical registrars . The histological examination of the surgical biopsies and grains culture were performed at the MRC in Khartoum as described previously [33 , 35] . Some patients with massive lesions were referred to the MRC for further assessment and management . All the investigations and treatment were provided free of charge . The Sennar Ministry of Health had provided free meals and transportation for the patients and their families The confirmed mycetoma patient’s information was entered into a predesigned patient management record . This included patient’s demographic characteristics , diagnostic tests results , management decisions , treatment received , follow-up and final patient treatment results . This information was regularly checked and updated throughout the patient management and follow-up period . A system for medicines and consumables procurement , delivery and storage was designed . The medicines included antifungals , antibiotics , analgesics , intravenous fluids and anesthetic medicines as well as the surgical and anesthetic consumables . The medicines were procured from the Central Medical Supply Corporate in Khartoum , shipped and stored at the WOMRC pharmacy at optimum conditions . They were dispensed by the local assistant pharmacist . Patients living in remote areas , of low socio-economic status and unable to attend the outpatient's clinic at WOMRC usually receive their medicines by a community activist who dispensed them use a toktoko ( large motorcycle ) . The patients’ information , the medicines doses and quantities were regularly registered . All these information and procedures were documented and regularly reported to the MRC . More than 300 care providers; medical assistants , nurses and public health officers were trained on different aspects of mycetoma , which included the disease causation , presentation , diagnosis and treatment , patients’ care , referral indications and system , community health education and disease advocacy . The instructional training methods included presentations , group discussions , clinical sessions and ultrasound diagnosis demonstration . Suspected patients’ referral card was designed and distributed to the trainee , Fig 3 . The improvement in knowledge , attitude and practice ( KAP ) of the trainees was assessed by pre and post-training tests S1 File . The training sessions were conducted at Singa Town , Sennar State capital , WOMRC and in various East Sennar Governate villages . Medical students from the local university; Sennar University , as well as the University of Khartoum and other health institutes were training at WOMRC , on early patients’ detection , referral and management and the CAPI use was conducted To gain the Sennar State political involvement and support , the training sessions were addressed by the Sennar State Governor and the Minister of Health , Eastern Sennar locality Governor and community leaders . Several meetings with the local villages’ leaders , villagers and community activists were conducted at the Wad Onsa village leader’s home , the village’s mosque and WOMRC to explain the study objectives and to gain their confidence and support . They were actively involved in the mycetoma advocacy and awareness activities . The local Red Crescent volunteers were trained in mycetoma advocacy and took an important role in improving the local environment and hygiene in the affected villages , Fig 4 . A toktok was donated by the MRC to a community activist at Wad EL Nimear village for transporting patients and their medicines between the different villages and for mycetoma advocacy . In appreciation of excellent , active and energetic involvement in mycetoma advocacy and awareness , three Mycetoma Ambassadors from the Sennar State were selected . The study Health Education Team was led by social workers from the Association for Aid and Relief , Japan , Khartoum Office , an active NGO in Sudan with several fine artists , musicians and community volunteers . The health care providers , community leaders and activists , school teachers and medical students from the local university were trained to conduct health education and advocacy sessions . Several health education sessions and activities were carried out for early active case detection were conducted . The sessions included small group discussion , school visits sessions , video films watching , and interactive open theatre drama , Fig 5 . “Mesaket Story” , a drama film documented a mycetoma patient journey from minor infection which was neglected till limb amputation was produced and was shown to more than 2000 individuals at WOMRC and other villages . Several campaigns to improve Wad El Nimear village environment , sanitation , and hygiene to reduce the mycetoma transmission risk factors such as thorns , sharp objects , animals dungs , were organised by the State Government , official local authorities , community leaders and activists , and Red Crescent volunteers in collaboration with the study team . The thorny trees and bushes , thorny animals enclosure , animals dungs , dirt and rubbish , were removed and burnt , Fig 6 . To improve the village hygiene , reduce the contact with the animals and their excreta and to eradicate the thorny cages , 72 modern animal enclosures were constructed outside Wad EL Nimear village . This project was conducted by a kind donation from an engineering company as its social reasonability activity . These new animals cages were distributed free of charge to the villagers , Fig 6 . In mycetoma , the foot is affected most , and traumatic inoculation of the causative organisms which are present in the soil is believed to be the route of infection . The habit of going barefooted in the villages and the minor trauma are considered the risk factors for mycetoma . To reduce these risk factors , the study team has distributed around 800 new shoes to the school pupils at Wad EL Nimear village to improve the personal hygiene and to reduce the risk of developing mycetoma . Forty students from the Department of Social Sciences , at the University of Khartoum , spent two weeks at Wad Onsa village studying the social background of the population in the affected villages in the study area and assessed their KAP to mycetoma and its socioeconomic impacts . They surveyed in depth ten villages in the locality . Opened ended questionnaire and focus group discussions were used to obtain the data . A Project Management Board was established headed by the Minister of Health and the senior health officials , Sennar State , the Sennar State Mycetoma control programme officer , local villages’ leaders and activists , local health care providers and MRC representative . The Board oversees the project implementation and update , problem sharing , analysis and solving . The Board has regular meetings to review the quarterly reports to provide advice and recommendations for improvement . The study ethical clearance was obtained from Soba University Hospital Ethical Committee to conduct the study . Informed consents were obtained from the leaders of the villages , informed written consents were obtained from State Ministry of Health and every suspected and confirmed patients . All medical records were anonymised
During the study period , 758 mycetoma suspected patients from the surveyed villages and other villages in Eastern Sennar Governate were seen at WOMRC . All of them had an ultrasound examination of the suspected lesions . Of them , 220 patients had ultrasonic evidence of mycetoma , and they underwent wide local surgical excisions ( 218 patients ) , and two patients had amputations . They were 134 males ( 60 . 9% ) and 86 females ( 39 . 1% ) . Their ages ranged between 2 and 70 years and age group 15–30 years was the most affected one . Most of them were students 68 ( 30 . 9% ) , housewives 46 ( 20 . 9% ) , farmers 35 ( 15 . 9% ) , ( Table 2 ) . The geographical distribution was uneven , but Wad El Nimear village had the highest prevalence , ( Table 2 ) . Most of the pateints ( 72 . 2% ) had short disease duration . Pain at the mycetoma site was not a common symptom in these patients; seen in only 39 patients ( 17 . 7% ) . Local trauma at the mycetoma site was reported in only 38 patients ( 17 . 3% ) . Most of the patients had no sinuses ( early lesion ) 142 ( 64 . 5% ) , and 72 patients ( 32 . 7% ) had black grains discharge from their sinuses , ( Table 2 ) . The foot 159 ( 72 . 2% ) and hand 59 ( 26 . 8% ) were affected the most . Less common sites were the back and gluteal one each , ( Table 1 ) . The majority of patients 139 ( 63 . 2% ) had small lesions less than 5 cm in diameter , 51 patients ( 23 . 2% ) had lesion between 5–10 cm in diameter , and only two patients ( 0 . 9% ) had lesions more than 10 cm in diameter , ( Table 2 ) . The lesions ultrasound examination findings were mycetoma in 202 patients ( 91 . 8% ) and foreign body granuloma in 18 patients ( 8 . 2% ) . The surgical procedures performed ranged from wide local excision 218 ( 99% ) to amputation 2 ( 1% ) . All patients had an uneventful postoperative recovery . The operatives findings included mycetoma lesions 192 ( 87 . 3% ) , foreign body granulomas with thorns 18 ( 8 . 2% ) , fibroma 2 ( 1% ) and others soft tissue masses . The diagnosis was confirmed by surgical biopsies histopathological examinations , and that showed evidence of eumycetoma in 189 patients ( 85 . 9% ) , foreign body granuloma 17 ( 7 . 7% ) , actinomycetoma 3 ( 1 . 4% ) and others 11 ( 5% ) . The latter included no-specific granuloma , neuromas and fibromas , ( Table 3 ) . Most of the patients were followed up at the WOMRC . Thirty-seven patients ( 16 . 8% ) developed recurrence , due to multifactorial factors which included massive lesion , patients’ non-compliance with treatment or other factors . Twenty-five patients ( 11 . 4% ) were lost to follow-up . Confirmed Mycetoma patients’ information was entered into the pre-designed patient’s management records . These records included full details of the patient’s demographic characteristics , diagnostic tests results , the management offered , follow-up and final patient treatment result . This information was regularly monitored and updated throughout the patient journey . All these data were systematically reported to the MRC in quarterly basis through two types of reporting format; hard copy and a digital one , the latter one was transmitted through a telemedicine facility at the WOMRC and MRC . The reported information was systematically entered in the pre-designed data analysis software for further analysis and systematically checked for information accuracy . Data from management teams , diagnostic services and inventory was crosschecked and discussed regularly to improve recording and reporting process . The current treatment of choice for eumycetoma is itraconazole in a dose of 400mg /day . It costs around 26 US$/day , that is not affordable by neither patients nor local health authorities , and hence the MRC managed to raise funds to procure and dispense itraconazole free of charge to patients at the WOMRC . A system for medicines procurement , delivery , storage and dispensing at the WOMRC was designed and tested during the study . A random sample of 218 individuals were tested before and after showing them “Mesaket Story” a drama film . The results showed improvement in their knowledge , attitude and practice and towards mycetoma , ( Table 4 ) . Several small group sessions were organised at different villages , schools , mosques and community clubs . 200 community activists , 50 Red Crescent volunteers and 500 school teachers were trained on mycetoma advocacy and awareness . The fine artists and musician had organised several interactive open theatre dramas . Different health education materials in different forms were used . Experts in watercolouring , oil painting and photography have greatly contributed to mycetoma advocacy and awareness through their production of high-quality paintings , photographs and videos captured from the studied Governate , Fig 7 . During the study , six patients with amputations received limb prosthesis donated by the Agent of Aid and Relief , Japan . This had remarkably improved the life quality .
The MRC records showed that the EL Gazeria , White Nile and Sennar States are the top endemic states in the country [28] . Despite been the third , Sennar State has been chosen for the present study due to the strong commitment of the political leaders , civil societies and communities leaders to support and implement the study and its research outcome . The local communities’ leaders were aware of the negative impacts of mycetoma on health and its socioeconomic bearings , and hence their response to the study team requests was swift and extremely positive . The concept of the village specialised mycetoma centre reported in this communication is a unique one . The WOMRC had delivered integrated medical and social services at the heart of an endemic area . The centre was established as a joint project between the Federal , Sennar State Ministries of Health and the local community , which by itself an exceptional initiative . This study demonstrated numerous positive impacts of the centre on the local communities . It provided local , decentralised mycetoma services in a location with bare minimal health service provision and has improved the local population health education and disease awareness . The telemedicine which links the MRC in Khartoum with the centre has facilitated the management and follow-up of patients , thus reducing the financial and geographic burdens on the patients and families , and also reduced the patients’ follow-up dropout rate . The dropout reported in this study ( 11 . 8% ) is less than that reported at the MRC ( 54% ) . Although the capital cost of the telemedicine setup is high in developing countries , however , in the long run , it is cost-effective for the patients , families and health authorities in mycetoma endemic regions . This is a unique experience that has not previously been reported for mycetoma or other neglected tropical diseases ( NTD ) and can be replicated for other endemic NTDs . The global lack of disease control or elimination programmes due to the unavailability of basic disease epidemiological characteristics has resulted in early case detection and management as the only available method to reduce the disease incidence and prevalence and its community impacts [28] . It is now evident that WOMRC has tremendous bearings on disease management by offering early case detection facilities and free , decentralised medical , health and advocacy services . That is evidenced the fact that many patients ( 63 . 2% ) with small lesions and patients ( 72 . 2% ) with disease duration of less than five years were seen at the centre and only two patients underwent amputation during the study period compared patients seen at the MRC [28] . Furthermore , that is supported by the improvement in disease awareness as evidenced by the KAP study results . The mycetoma onset and progress are usually slow and painless , affecting patients of low socio-economic and health education levels . Hence these patients are different from patients with other deadly infectious diseases , e . g . malaria , cholera , leishmaniasis , where patients have no other choice but to report early and follow medical instructions [50 , 51] . Moreover , mycetoma patients with early lesions differ from patients with large disabling mycetoma lesions . Early lesions are usually tiny and painless thus not interfering with their normal daily activities . Some patients consider it at this stage as a trivial or even normal event . These patients usually have many other more pressing social and economic problems than these tiny lesions , e . g . the short busy seasonal farming session , raising children in poor conditions and others . Most of the patients consider these early lesions are not a priority , and in fact , they believe that treatment will delay undertaking other urgent duties , and this explains the late presentation with massive lesions [28 , 31] . It is evident from this study that our holistic management has addressed many issues . The community engagement activities have led to active early case detection which is supported by the high number of patients with early disease seen at the WOMRC reported in this study . Such patients have a high chance of cure and were amenable to treatment with a good outcome [47 , 52] . The immediate access to free treatment at the village level has reduced patients’ delay in starting treatment , eliminated patients’ geographical and financial burdens , treatment interruption , and reduced the high follow up dropout rate . Treatment interruption can induce drug resistance . It is therefore vital to ensure sustainability and availability of the free mycetoma treatment services . The health system in Sudan consists of the three levels; the rural , regional and central levels . The medical assistants , the health care providers , are the backbone for the management of mycetoma patients in Sudan at the rural level . Most of them have poor surgical experience and used to operate on the mycetoma patients under local anesthesia and suboptimal conditions . This practice has led to the high recurrence rate which was documented in many reports [1 , 2 , 53] . Recurrent disease is usually associated with wide local disease spread . Hence it is difficult to cure and necessitate repeated surgical excisions , numerous deformities and disabilities [1 , 2] . At Sennar State , most of the medical assistants were successfully trained on the different aspects of mycetoma care , management and referral . A well-trained multidisciplinary team on mycetoma care was developed in Eastern Sennar region . It consists of a trained surgeon , surgical theatre attendants , anesthetic assistants , pharmacy assistant , ultrasound technicians , nurses , information technology expert , statistical clerks , community leaders and activists . This is an essential step in providing comprehensive and holistic management for the mycetoma patients . To date , the definitive route of infection in mycetoma is an enigma . However , it is clear that mycetoma incidence is high in areas of poor environmental conditions , among people with poor personal hygiene and people living in proximity to animals and their dungs and where thorns , dirt and mud prevail . Hence this study aimed to improve the living and hygienic standards of Wad EL Nimear village , one of the highly endemic villages in the locality . The local villagers were encouraged to improve their living conditions . To achieve this goal , many advocacy and awareness campaigns were conducted , and new modern hygienic animals’ cages were constructed and donated to them free of charge . In support of these measures , the local Governate authorities issued a law banning the presence of animals inside the village . The kind donation of the animals’ cages by the engineering company as a response to the intensive mycetoma awareness and advocacy in Sudan . In this study , the community leaders and activists were actively involved in conveying messages to their community in their own culture and traditions . This was important to accept these holistic disease management procedures . Likewise , the local villagers have actively engaged in promoting their health and improvement of local environmental conditions that believed to be the main source of transmitting mycetoma . In conclusion , the holistic and comprehensive management approach implemented in this study has improved the mycetoma patients’ quality of care in the studied endemic area . More early disease was detected and treated . The treatment interruption rate was reduced thus increasing the cure rate and decreasing the recurrence and hospitalisation rates . This will eventually lead to decrease in the amputation and disability rates . The results obtained from this study suggest that such a study can be expanded to other endemic areas in the country . The MRC , as a WHO Collaborating centre on Mycetoma , will communicate this experience to the WHO to share it with other mycetoma endemic countries and assist in better management , prevention and control of the disease .
|
Mycetoma enjoys all the neglected tropical diseases ( NTDs ) characteristics . It frequently affects the poorest of the poor in poor communities in remote regions . The affected population are of low socio-economic status , of low visibility and low political and social voice and hence they are neglected . The disease is considered as a social stigma in particularly among females and children thus they tend to hide it for prolong period and when they are compelled to seek medical care the condition is then at a late stage . The mycetoma patients have many financial constraints that hinder them from seeking medical and health care . In the remote mycetoma endemic areas , the health and medical facilities are meagre , and it is difficult for the patients to reach the regional health centres and thus the majority of patients present with late advanced disease . To overcome these treatment difficulties , the MRC had adopted this holistic management approach to decentralised the patient's care , improve the disease awareness and advocacy , provide free medical and surgical treatment locally at the village level , and to improve the affected villages environmental and hygienic conditions . In this communication , the MRC is reporting on this unique experience , discussing the advantages and difficulties faced it and suggesting recommendations to improve it to be adopted worldwide . Reviewing the medical literature revealed , no report on such management approached for mycetoma patients and thus it worth reporting it .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"mycetoma",
"tropical",
"diseases",
"patient",
"advocacy",
"geographical",
"locations",
"sudan",
"surgical",
"and",
"invasive",
"medical",
"procedures",
"health",
"care",
"signs",
"and",
"symptoms",
"neglected",
"tropical",
"diseases",
"fungal",
"diseases",
"africa",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"lesions",
"hygiene",
"people",
"and",
"places",
"socioeconomic",
"aspects",
"of",
"health",
"diagnostic",
"medicine",
"health",
"education",
"and",
"awareness"
] |
2018
|
A holistic approach to the mycetoma management
|
Helicobacter pylori’s ability to respond to environmental cues in the stomach is integral to its survival . By directly visualizing H . pylori swimming behavior when encountering a microscopic gradient consisting of the repellent acid and attractant urea , we found that H . pylori is able to simultaneously detect both signals , and its response depends on the magnitudes of the individual signals . By testing for the bacteria’s response to a pure acid gradient , we discovered that the chemoreceptors TlpA and TlpD are each independent acid sensors . They enable H . pylori to respond to and escape from increases in hydrogen ion concentration near 100 nanomolar . TlpD also mediates attraction to basic pH , a response dampened by another chemoreceptor TlpB . H . pylori mutants lacking both TlpA and TlpD ( ΔtlpAD ) are unable to sense acid and are defective in establishing colonization in the murine stomach . However , blocking acid production in the stomach with omeprazole rescues ΔtlpAD’s colonization defect . We used 3D confocal microscopy to determine how acid blockade affects the distribution of H . pylori in the stomach . We found that stomach acid controls not only the overall bacterial density , but also the microscopic distribution of bacteria that colonize the epithelium deep in the gastric glands . In omeprazole treated animals , bacterial abundance is increased in the antral glands , and gland colonization range is extended to the corpus . Our findings indicate that H . pylori has evolved at least two independent receptors capable of detecting acid gradients , allowing not only survival in the stomach , but also controlling the interaction of the bacteria with the epithelium .
Helicobacter pylori is a bacterium that has co-evolved with humans since the origin of the human species [1 , 2] . This intimate association with the human host has allowed it to develop a number of survival strategies to persist in one of the most challenging environments in the human body—the stomach . H . pylori’s survival relies on its ability to avoid the microbicidal effects of stomach acid . H . pylori can withstand acidic conditions for short periods of time due to its urease enzyme which degrades urea into ammonium and bicarbonate to buffer the cytoplasm and periplasm [3–6] . Another important strategy is to utilize chemotaxis to locate and swim to the gastric epithelium where the pH is near neutral due to the overlying protective mucus layer . The bacteria colonize a narrow niche within 25 microns of the surface of the gastric epithelium where they are either found actively swimming in the mucus or directly adhered to epithelial cells [7 , 8] . The attached bacteria utilize virulence factors to obtain essential nutrients from the host and replicate on the cell surface to form cell-associated microcolonies [9 , 10] . We recently reported that a subpopulation of cell-associated H . pylori is found deep in the antral glands in direct contact with gastric progenitor cells and stem cells [11] . The gland-associated H . pylori induce the expansion and proliferation of stem cells , alter stem cell gene expression , and lead to gland hyperplasia [11] . We hypothesize that the factors that control H . pylori’s ability to colonize the gastric glands will help explain H . pylori’s ability to persist long-term in the stomach and to cause gastric diseases . Despite living in the stomach , H . pylori is not an acidophile and swims away from hydrochloric acid ( HCl ) . The acid secreted into the stomach lumen by parietal cells in the corpus forms gradients that keep the bacteria close to the gastric epithelium [8] . Previous studies have reported that the chemoreceptor responsible for sensing HCl as a repellent is TlpB [12 , 13] . This chemoreceptor has been shown to detect auto-inducer 2 as a repellent as well [14] . We recently reported that TlpB can also sense chemoattractants , since it is a high affinity chemoreceptor for urea that allows H . pylori to sensitively detect and swim towards urea emanating from the gastric epithelium [15] . In this study , we initially proceeded to investigate how H . pylori may be sensing both a repellent and an attractant through TlpB . Using a previously developed videomicroscopy method that visualizes and films bacterial chemotactic responses to chemical gradients in real time [7 , 15] , we discovered that H . pylori mutants lacking TlpB ( ΔtlpB ) are not defective in detecting and swimming away from HCl gradients . Instead , we identified TlpA and TlpD as independent acid sensors with different sensitivities to HCl . We also found that TlpD allows H . pylori to chemotax towards less acidic and even basic pH environments , and this response is dampened by TlpB . Using a murine model of infection in the stomach , we discovered that the double mutant lacking TlpA and TlpD ( ΔtlpAD ) is about 100-fold defective in its ability to colonize the stomach compared to wild-type H . pylori . However , treatment with the proton-pump inhibitor omeprazole raises the gastric pH and partially rescues the ΔtlpAD mutant’s defect , allowing it to reach significantly higher bacterial numbers in the stomach . We also observed that omeprazole treatment promotes wild-type H . pylori’s colonization of the gastric glands and extends its range of glandular colonization from the antrum into the glands of the corpus . Despite the higher loads of ΔtlpAD H . pylori in omeprazole-treated animals , the mutant is unable to colonize the gastric glands to the same levels as wild-type , suggesting that these two chemoreceptors are important in establishing colonization deep in the gastric glands . Our study has identified two new roles for H . pylori’s chemoreceptors as acid sensors and demonstrate that H . pylori’s ability to detect and respond to the acid gradient is important for its localization within the stomach , its interaction with the glandular epithelium , and its survival in vivo .
H . pylori encounters many chemical gradients in the stomach that serve as cues to identify microniches that are safe for colonization . It must be able to integrate both repellents and attractants to optimize its chemotactic response . Since H . pylori’s TlpB chemoreceptor has been reported to detect HCl as a repellent [12 , 13] , and we found that it detects urea as a high-sensitivity attractant [15] , we wondered how H . pylori would respond if it were simultaneously exposed to both urea and HCl gradients emanating from one point source . We used our previously described microgradient chemotaxis assay [7 , 15] to record the swimming responses of a live culture of H . pylori ( strain PMSS1 ) exposed to a microscopic gradient of a mixture of 50 mM HCl and 5 mM urea emanating from the tip of a microinjection needle . Prior to the assay , H . pylori were grown in Brucella broth with 10% FBS ( BB10 ) , pH 6 . 7–6 . 8 to an OD600 of 0 . 3 . During culture , bacterial urease depletes urea from the surrounding medium ( S1A Fig ) and the pH remains at a range of 6 . 6–6 . 74 ( S1B Fig ) . A 270 μl volume of media with motile bacteria is placed onto a coverslip chamber and a microinjection needle is then rapidly inserted via a micromanipulator into the viewing field to produce a microscopic gradient of chemoeffectors . Using this method , we observed that H . pylori are attracted to the urea in the gradient until they reach a boundary approximately 60 micrometers away from the needle tip ( Fig 1A ) . Within this boundary is a zone of clearance avoided by the bacteria , representing a threshold concentration of acid that acts as a chemorepellant . When bacteria swim into this zone of clearance , they quickly stop , reverse direction , and swim away ( S1 Movie ) . As a negative control , we tested a chemotaxis null mutant ΔcheW H . pylori for its response to the same mixed microgradient . The density of swimming ΔcheW H . pylori remained constant throughout the field ( Fig 1A ) consistent with this mutant’s inability to respond to either the attractant or the repellent . These observations indicate that H . pylori is capable of simultaneously sensing and integrating multiple signals from one point source . To test whether the chemotactic response is dependent on the magnitude of these chemical gradients , we altered the concentration of acid within the needle while maintaining the same concentration of urea . We found that , indeed , the bacteria respond by increasing the distance of the boundary between repulsion and attraction and the needle tip as the concentration of HCl increases ( S2 Fig ) . Because H . pylori abundantly expresses a urease enzyme that degrades urea into ammonia to buffer its cytoplasmic and periplasmic pH , we wondered whether urease may play a role in acid sensing . We first tested a urease mutant ( ΔureAB ) for its response to a solution of urea and HCl . Unlike the WT H . pylori culture , which is depleted of urea due to urease activity , the ΔureAB culture contains levels of urea in the medium comparable to fresh media ( S1A ) . We observed that ΔureAB cleared the field of view completely ( S3A Fig ) , indicating that it was unable to sense urea because the surrounding urea in the medium interferes with chemotactic sensing as we had previously reported [15] but was able to sense acid . We further tested for ΔureAB’s response to a solution of 100 mM HCl in water ( no urea ) . We found that it cleared the field of view like wild-type ( S3B Fig ) . These results suggest that urease activity is not required for acid sensing . Next , we tested ΔtlpB H . pylori’s response to this mixed solution . As the urea and acid sensor , we predicted that this mutant would respond to neither compound and that its swimming behavior would be like that of ΔcheW H . pylori . We were surprised to find instead that ΔtlpB H . pylori rapidly swims away and clears from the field of view like ΔureAB ( Fig 1B , S3A Fig , and S1 Movie ) . This observation that the bacteria swim away from the urea and acid microgradient suggests that the ΔtlpB mutant is unable to detect urea as we had previously reported [15] but is able to detect HCl as a repellent . We proceeded to test ΔtlpB’s response to a solution of HCl in water without urea . We found that , indeed , ΔtlpB efficiently swims away from an acid gradient ( Fig 1C and S2 Movie ) . We plotted the bacterial density in the viewing field every fifth of a second before and after introduction of the needle tip injecting HCl , and by ten seconds post-exposure to the acid gradient , the ΔtlpB mutant has mostly cleared from the field of view ( Fig 1C ) . Interestingly , we note from the clearance curves ( Fig 1C ) that ΔtlpB clears from the field of view faster than wild-type upon acid exposure . This result suggests that TlpB does affect H . pylori’s response to acid since ΔtlpB has response kinetics different from wild-type . To verify that the ΔtlpB response to acid is not specific to a particular H . pylori strain , we constructed ΔtlpB mutants in other strains of H . pylori and tested them in the same assay . We found that the ΔtlpB mutants in all strains tested ( strains G27 , SS1 , PMSS1 ( a second independent clone ) , and 7 . 13 ) are still able to respond to an acid gradient like their wild-type counterparts by swimming away upon exposure to HCl and maintain the fast clearing phenotype ( S4 Fig ) . This result indicates that ΔtlpB’s response to HCl is not strain specific . For the rest of the experiments we use H . pylori PMSS1 as the strain background . Our finding that ΔtlpB responds to an HCl gradient suggests that other chemoreceptors may be acid sensors . We tested the response of mutants lacking each of the other three chemoreceptors , ΔtlpA , ΔtlpC , ΔtlpD , by videomicroscopy in the same assay . To simplify the comparisons of the movies showing the escape from acid of different mutants , we graphed the percent of bacteria present in the viewing field 4 seconds before and 10 seconds after exposure to the acid gradient ( Fig 2A ) . We found that each mutant was still able to respond and swim away from acid , indicating that either there is a novel unidentified chemoreceptor that senses acid or there are multiple chemoreceptors that can function redundantly in sensing acid . To test the hypothesis that there may be redundancy in acid-sensing chemoreceptors , we made mutants lacking two of the four chemoreceptors in all possible combinations . We discovered that of all the six combinations of chemoreceptor knock-outs , only the mutant lacking both TlpA and TlpD ( ΔtlpAD H . pylori ) lost the ability to respond to HCl gradients ( Fig 2B and S3 Movie ) . This result indicates that TlpA and TlpD function as acid sensors , each capable of compensating for the loss of the other in acid sensing . Furthermore , when we tested for the response of ΔtlpAD H . pylori to the solution containing a mixture of urea and HCl , we observed that the bacteria are attracted towards the needle tip with no zone of clearance ( Fig 2C ) . This result indicates that the ΔtlpAD mutant can detect and respond to urea as an attractant ( through TlpB ) , but is unable to detect the HCl as a repellent . This result suggests that TlpB is not sufficient to sense acid in our assay , and TlpA and TlpD are acid sensors that detect acid gradients . We next asked whether the response to acid is a response to low pH or a response specifically to HCl . To determine this , we tested the responses of WT , ΔtlpA , ΔtlpB , ΔtlpD , and ΔtlpAD to sulfuric acid ( H2SO4 ) and phosphoric acid ( H3PO4 ) . As with HCl , we observed that H . pylori responds to both of these acids as repellents and requires either TlpA or TlpD for the response ( S5 Fig ) . This result suggests that these two chemoreceptors allow H . pylori to sense and escape from conditions of low pH . Several chemoeffectors have been described for TlpA and TlpD . The TlpA receptor was reported to mediate attraction to arginine and bicarbonate [16] . The TlpD chemoreceptor was reported to mediate repulsion from conditions that induce low-energy in the bacterium [17] or conditions that promote oxidative stress [18 , 19] . These conditions may be triggered by low pH . One key difference between TlpA and TlpD is the location of these chemoreceptors within the bacterium . The TlpD chemoreceptor is cytoplasmic as it lacks a transmembrane domain [17] whereas the sensing domain of TlpA is periplasmic like that of the TlpB and TlpC chemoreceptors . Therefore , TlpD may be sensing changes in external pH indirectly . Given the difference in the location of these two chemoreceptors , we wondered if the two receptors may have distinguishable responses to the same HCl gradient . To assess if there is a difference in sensitivity for detecting HCl between the two chemoreceptors , we determined the threshold concentration of HCl necessary to elicit a response ( arbitrarily defined as a 60 microns clearance zone from the point source ) for each mutant . We empirically determined that 25 mM HCl loaded in the microinjection needle is the minimum concentration required to repel wild-type H . pylori from the point source by 60 microns ( Fig 3 ) . ΔtlpA H . pylori , like wild-type , responds to a minimal effective concentration of 25 mM HCl while ΔtlpD H . pylori requires 50 mM HCl to respond ( Fig 3 ) . It is worth noting , however , that the actual minimal effective concentration of HCl that H . pylori is capable of detecting is markedly below 25 mM , since only a minute volume of HCl on the order of 1 picoliter/minute is injected into the solution , and the culture medium surrounding the bacteria consists of Brucella broth with 10% fetal bovine serum , which has a buffering capacity that we empirically determined to be about 4 , 000 fold greater than water ( S6 Fig ) . Thus the gradient of free hydrogen ions would drop rapidly away from the needle tip . Since we observe the bacteria responding 60 microns away from the point source , the change in HCl concentration that the bacteria can sense is substantially less than the concentration of the solution in the needle . Indeed , when we expose the bacteria to a gradient starting from 15 mM HCl at the needle tip , we also see the bacteria being repelled but at approximately 30 microns away from the point source ( S7 Fig ) . Since H . pylori is exposed to both urea and HCl in vivo , we wondered whether urea sensing would affect the sensitivities of TlpA and TlpD in sensing acid . We exposed ΔtlpA and ΔtlpD to a gradient of a mixture of 5 mM urea and 50 mM HCl in the microinjection needle . We observed that both mutants are attracted to the urea but form a zone of clearance as they sense HCl like wild-type H . pylori . However , with urea present we could elucidate small differences in sensitivity because the attraction to urea highlights TlpD’s ability to detect lower concentrations of acid than TlpA . This is illustrated by the ΔtlpA mutant remaining at a farther distance away from the acid point source at the needle tip compared to ΔtlpD ( S8 Fig ) . This result indicates that urea sensing does not alter the sensitivity hierarchy of TlpD and TlpA in acid sensing . As chemotaxis is a result of detecting a change in the concentration of a chemical , we next sought to determine the smallest change in hydrogen ion concentration ( [H+] ) that H . pylori is able to detect and respond to . By comparing the pH of the Brucella broth culture medium in which the bacteria are grown with the pH of a buffer solution that elicits a repulsion response , we determined the smallest difference in [H+] that H . pylori is capable of detecting . To more accurately control the range of [H+] concentrations experienced by the bacteria , we loaded the microinjection needle with 1M phosphate buffer solutions with defined pHs made by combining different proportions of dibasic and monobasic phosphate solutions . We measured the pH of the Brucella broth medium that the bacteria are grown in to be pH 6 . 7 ( S1B Fig ) . We then tested wild-type bacteria’s response to buffer solutions ranging from pH 4 . 2 to 7 . 1 . As shown in Fig 4A , buffers with pH 4 . 2 , 6 . 0 , 6 . 3 , and 6 . 5 all elicited an escape response in the microgradient assay . The increase in [H+] between the solution released from the microinjection needle and the culture medium was approximately 63 μM , 800 nM , 360 nM , and 110 nM , respectively . For the lower pH measurements between pH 4 and pH 6 we also confirmed our results using citrate buffer , which buffers acidic pH more effectively than phosphate buffer ( S9 Fig ) . Phosphate buffer solutions of pH 6 . 6 and 7 . 0 , which locally changed the [H+] by + 35 nM or -100 nM , did not elicit a response , indicating that the differences were below the limit of detection of the receptors . Unexpectedly , a buffer solution with pH 7 . 1 , ( about -120nM change in [H+] ) elicited an opposite response with bacteria attracted and forming a swarm at the needle tip ( Fig 4B and S4 Movie ) . Thus , we discovered that H . pylori can sense and respond to small changes in both acidic and basic pH when the change in [H+] is greater than 100 nM . To further characterize wild-type H . pylori’s response to more basic environments , we also tested phosphate buffers with pH 7 . 25 and 9 . 2 . We found that H . pylori is attracted to basic pH solutions ( S10 Fig ) . By testing the single chemoreceptor mutants , we identified TlpD as the necessary receptor for attraction to higher pH ( Fig 5 and S5 Movie ) . The other receptors , including TlpA , were not necessary . To confirm that the TlpD-dependent attraction is a response to high pH rather than to a specific basic solution , we tested for H . pylori’s response to 40 mM sodium hydroxide ( pH 12 . 6 ) . We also observed H . pylori being attracted to sodium hydroxide , and the response was dependent on TlpD ( S11 Fig ) . Taken together these results suggest that TlpD mediates both repulsion from lower pH and attraction to higher pH while TlpA only detects and mediates repulsion to lower pH . We wondered whether the TlpB-dependent defect in acid sensing previously reported and our observation of faster clearance rates of ΔtlpB mutants compared to wild-type may be due to an effect of TlpB in modulating the sensitivity to pH of the other chemoreceptors . In order to see if lacking TlpB enhances the sensitivity to acid , we tested for the response of ΔtlpB H . pylori to low concentrations of HCl and found that the minimal effective concentration of HCl in the needle required to elicit an escape response is the same as that of wild-type at 25 mM HCl ( Fig 3 ) . We also tested the responses of the TlpA and TlpD chemoreceptors in the absence of TlpB by determining the sensitivities of the ΔtlpAB and ΔtlpBD double mutants to acid gradients . We found them to have the same sensitivity to HCl as ΔtlpA and ΔtlpD , respectively ( S7 Fig ) . However , when we tested for H . pylori’s response to higher pH , we noted that ΔtlpB’s attraction was always more pronounced than wild-type with each higher pH solution ( S10 Fig ) ; the attraction of ΔtlpB was faster , and the concentration of bacteria around the needle tip was denser than that of wild-type . This observation led us to hypothesize that TlpB may be modulating H . pylori's response to basic pH . We tested for the mutant’s response to a gradient formed by phosphate buffer at pH 7 . 0 where wild-type H . pylori does not show an attraction . The ΔtlpB mutant is still able to sense and respond to the solution at pH 7 . 0 ( -100nM change in [H+] ) ( S10 Fig ) . This result suggests that lacking TlpB increases the bacteria's sensitivity and attraction to higher pH , and thus TlpB may be modulating the pH responses mediated through TlpD . The mechanism by which TlpB influences pH sensing is unknown . We tested whether TlpB alters the expression levels of TlpA or TlpD , thereby resulting in a change in acid response kinetics . We performed a Western blot to assess expression levels of the chemoreceptors in the single knock-out mutants and ΔtlpAD H . pylori . We found that the expression levels of the remaining chemoreceptors were not affected by the loss of the TlpA , TlpB , or TlpD receptors ( S12 Fig ) . Our data show that TlpB is not sufficient for sensing acidic pH , but it does alter the kinetics of H . pylori’s response to changes in pH . Taken together , our data suggest that the mechanism of pH sensing in H . pylori is complex , involving multiple chemoreceptors . The difference in sensitivity between TlpA and TlpD and the role of TlpB in modulating pH responses allows H . pylori to discern even small changes in local pH gradients in the nanomolar range . This may have critical implications for H . pylori’s survival in the stomach . Given the importance of avoiding the acidic lumen of the stomach environment , we next investigated whether ΔtlpAD H . pylori would be able to colonize the stomachs of mice . We infected C57Bl/6 mice with either wild-type H . pylori strain PMSS1 or the isogenic mutants ΔtlpA , ΔtlpD , or ΔtlpAD . After two weeks of infection , a time point in which wild-type H . pylori has established stable colonization [7 , 20] , we harvested the stomachs and assessed colonization densities by colony-forming units ( CFU ) per gram of stomach tissue . We found that the ΔtlpA mutant was not defective in colonization with similar CFU counts as wild-type bacteria . However , the ΔtlpD mutant had a 10-fold defect as had been previously reported [20] , and the ΔtlpAD H . pylori mutant was about 100-fold defective in establishing colonization ( Fig 6 ) . We hypothesized that without the ability to respond to acid gradients in the stomach , the ΔtlpAD H . pylori is deficient in avoiding the microbicidal HCl . We wondered if pharmacologic inhibition of acid secretion in the stomach would improve ΔtlpAD H . pylori’s survival . We maintained two experimental groups to test this hypothesis . One group of animals was treated with the proton-pump inhibitor omeprazole for three days prior to infection to raise the gastric pH before infection ( S13 Fig ) , and the animals continued to receive omeprazole throughout the course of the 2-week infection . This experimental group allowed us to observe how the gastric pH experienced by the bacteria upon entering the stomach affects the bacteria’s survival in the stomach . The second group of animals was infected with H . pylori for a week before treatment with omeprazole throughout the final week of infection . This group represents the more common clinical scenario in which humans with an established H . pylori infection may take proton-pump inhibitors like omeprazole to treat conditions such as gastroesophageal reflux . We found that treatment of animals with omeprazole prior to or after the establishment of infection partially rescued ΔtlpAD H . pylori’s ability to colonize the stomach ( Fig 6 ) , while it did not affect the total number of wild-type bacteria . We previously reported that in this murine model of infection and in humans the majority of the bacteria reside in the overlying mucus layer . However , a subpopulation of H . pylori can be found deep in the gastric glands adhered to epithelia cells that make up the mid-glandular proliferative zone [7 , 11] . We wondered whether acid sensing would affect not only the overall fitness of the bacteria establishing colonization in the stomach , but also their ability to reach and colonize the epithelial surface of the gastric glands . To investigate gastric gland colonization , we used quantitative 3D confocal microscopy to determine the number of bacteria growing as microcolonies within the gastric glands in the antrum and corpus regions of the stomach . One of the main distinguishing features between antrum and corpus is the presence of acid-secreting parietal cell in the corpus glands . In control animals with normal acid secretion , the wild-type strain is found mostly colonizing the antral glands and the transition zone between the antrum and corpus [7 , 11] ( S14A and S14B Fig ) rather than the corpus glands ( S14C and S14D Fig ) . We found that in only two out of seven animals infected with ΔtlpAD H . pylori , the mutant was able to colonize the antral glands but at significantly lower densities than wild-type H . pylori ( Fig 7A , 7C and 7E ) . This result suggests that chemotaxis through TlpB and TlpC still allows some bacteria to reach the epithelium and colonize the glands . We know that one such signal is the chemoattractant urea sensed through TlpB [15] . The defect in gland colonization of the mutant , however , may be attributed to low bacterial numbers in both the mucus and the glands as it has a 100-fold defect in overall bacterial load compared to wild-type H . pylori ( Fig 6 ) . We performed a similar analysis of gastric gland colonization of ΔtlpAD H . pylori in animals treated with omeprazole after infection , since in these conditions the bacterial load in the stomach was comparable to that of wild-type ( Fig 6 ) . When we analyzed the antral glands of infected animals treated with omeprazole one week post-infection , we noted that the bacterial density of both wild-type and ΔtlpAD H . pylori in the antral glands increased significantly compared to those in control animals ( Fig 7B , 7D and 7E ) . Despite the increase in gland colonization , ΔtlpAD H . pylori does not reach the levels of wild-type in the antral glands of omeprazole-treated animals ( Fig 7E ) . These results suggest that omeprazole promotes the colonization of H . pylori in the gastric glands of the antrum , and chemotaxis through TlpA and TlpD is important for proper colonization of the antral glands . Using quantitative 3D-confocal microscopy , we also found that loss of stomach acidity through omeprazole treatment allowed H . pylori to extend its colonization to the corpus glands ( Fig 8 ) . Control animals normally do not have H . pylori in corpus glands ( Fig 8A ) . In our analysis , only one out of the seven control animals had detectable levels of H . pylori in the corpus glands ( Fig 8C ) . In the corpus , the gland-associated bacteria seen after omeprazole treatment were mainly concentrated in the neck region of the glands in the proliferative zone ( Fig 8B ) , and also were seen in close proximity to parietal cells ( Fig 8D ) . This result suggests that the acid in the stomach restricts H . pylori gland-colonization to the antral glands but that an increase in gastric pH allows H . pylori to extend its range to also colonize the epithelium of the corpus glands . We did not find ΔtlpAD H . pylori in the corpus glands of omeprazole-treated animals . These results from the omeprazole treatment experiments suggest that gland colonization is distinct from colonization of the mucus , and sensing through TlpA and TlpD may be necessary for localizing to and/or persisting in the corpus glands .
Our study has revealed that H . pylori evolved two independent chemoreceptors , TlpA and TlpD , capable of sensing and rapidly responding to acid gradients . The fact that H . pylori devotes at least half of its chemoreceptor repertoire towards acid sensing underscores the importance of this function for H . pylori’s survival in the stomach . Despite the same overall function , we found that there is a difference in the sensitivity of TlpA versus TlpD in detecting HCl . The cytoplasmic TlpD chemoreceptor appears to be more sensitive than the periplasmic TlpA chemoreceptor , and it is able to sense both lower pH as a repellent and higher pH as an attractant . H . pylori may have evolved a more sensitive acid sensor in its cytoplasm as it would be crucial to detect even small changes in cytoplasmic pH to ensure homeostasis . TlpD is the only chemoreceptor of the four that H . pylori possesses that is thought to be a soluble protein located in the bacterial cytoplasm and inner membrane [17] . It has been reported to be an energy sensor causing H . pylori to repel from electron transport inhibitors and other low energy-inducing environments [17] . Currently it is not known what are TlpD’s specific ligands and how TlpD may be detecting these ligands . TlpD contains a C-terminal chemoreceptor zinc binding domain ( CZB ) of unknown function [21] and does not contain a PAS domain commonly found in other chemoreceptors , making it challenging to identify the ligands that TlpD directly bind . It is possible that changes in intracellular pH may affect intracellular metabolism and the energy state of the bacterium , which would link TlpD’s ability to sense acid to its role as an “energy sensor , ” and consequently , an indirect sensor of acidic pH . Interestingly , it was recently published that changes to particular protein interactions with TlpD or the metabolic state of the bacterium alters the localization of TlpD in H . pylori [19] . This and another study , also reported that the sensing mechanism of TlpD may be linked to oxidative stress and iron limitation [18 , 19] . Specifically , upon metabolic stress or iron limitation , TlpD changes its localization from the poles of the bacterium to the cytoplasm . Also , in the absence of the interacting partners recently identified , TlpD changes its localization from the poles to the cytoplasm [19] . It is unclear , however , how TlpD’s localization to the poles or the cytoplasm affects its ability to sense and respond to particular chemoeffectors . There is no direct evidence that TlpD directly interacts or forms chemoreceptor arrays with TlpA , TlpB or TlpC to signal , and in fact TlpD has been shown to localize to the poles and be capable of transducing a chemotactic signal to the flagellar machinery in the absence of all other chemoreceptors [22] . Structural analysis of the chemoreceptor may reveal more insights into the mechanism of sensing through TlpD . TlpA has been reported to sense arginine and bicarbonate as attractants in H . pylori strains 26695 [16] and 700392 [23] , but its mechanism of sensing has not been further characterized . It has been proposed that chemotaxis towards a bicarbonate gradient in vivo may help H . pylori navigate towards a safe niche on the gastric epithelium where it is protected against the acid in the lumen . It is unclear whether TlpA is sensing bicarbonate directly or the basic pH that bicarbonate creates . While our studies here have shown that H . pylori is attracted to gradients of basic pH , we identify TlpD , not TlpA , as the chemoreceptor necessary for this attraction . It is possible that TlpA may be detecting bicarbonate specifically as a ligand rather than an increase in pH or we did not test the optimal basic pH that TlpA may be detecting . Further studies are needed to determine how TlpA is sensing acidic pH and whether TlpA may be involved in sensing basic pH as well . Our data surprisingly revealed that loss of TlpB does not result in defects in escaping from an acid gradient in the microgradient assay . However , we noted that the response kinetics of ΔtlpB H . pylori differed from that of wild-type ( Fig 1C ) . This suggests that lacking TlpB does have an effect on H . pylori’s ability to respond to acid even though TlpB is neither necessary nor sufficient to detect acid gradients . The difference between these recent results and previously published findings may be attributed to differences in the assays used to assess ΔtlpB’s responses to acid . Our assay generates and maintains a constant microscopic gradient from a point source and records the chemotactic behavior immediately after exposure to a gradient [7 , 15] . We observe bacterial responses within seconds , and the response is sustained for long periods of time ( we have tested it for as long as 10 minutes ) . However , our assay does not change the overall pH of the medium containing the bacteria because we inject minute amounts of acid ( with a flow rate on the order of picoliters per minute ) at very low pressure through a femtotip needle . A previously described assay that places H . pylori in an acidic environment for several minutes describes the formation of a barrier of bacteria at a region where the pH has been altered [12–14] . In this barrier assay the bacterial culture is infused with a 100 mM solution of HCl , which exceeds the buffering capacity of Brucella broth , and when mixed would decrease the pH of the bacterial culture from about pH 6 . 7 to pH 4 . 76 , as we determined empirically ( S6 Fig ) . Thus , the media is likely acidified when the chemotactic behavior is observed at about 5 minutes after exposure . Another assay used that has implicated TlpB’s role in acid sensing is a video chemotaxis assay where the bacteria are placed in chemical solutions of interest , such as acidic solutions [13 , 14 , 17 , 24] . In this assay , the bacteria are not exposed to a chemical gradient , but the assay measures motility behavioral differences in the presence or absence of a chemical as stops/sec or reversals/sec . An increase in stops or reversals per second indicates the detection of a repellent , but since a gradient is absent , directed movement cannot be assayed . An increase in reversals may therefore also represent the loss of an attractant or a change in the functioning of the chemosensory signal transduction pathway . The conditions in these two assays differ drastically from that of the microgradient assay with regard to the shape and steepness of the acid gradient as well as the time scale in which chemotactic responses are assessed . These other two assays may better assess H . pylori’s response when immersed in a more homogenous low pH environment such as the stomach lumen as opposed to an environment where there is a steep acidic pH gradient , such as across the gastric mucus layer . Perhaps TlpB is important for acid sensing in the acidic lumen . We do detect differences in the speed and sensitivity of acid sensing by the other receptors when TlpB is missing . Perhaps TlpB plays a role in acid sensing , for example , by changing the sensitivity of acid sensors at different baseline pH conditions or in other spatial and temporal conditions not replicated in the microgradient assay . The altered response kinetics of ΔtlpB compared to wild-type may be the resultant response from the integration of the remaining three receptors . Lacking TlpB may be changing the way the other receptors function in sensing acid as well as how other signals present in their environment alter responses to acid . While there is no evidence that TlpB and TlpD directly interact , chemoreceptors are known to form mixed arrays that transduce signaling responses to the flagellar motor [25 , 26] . It is possible that TlpB may directly interact with TlpD under certain conditions to dampen its response to basic pH . One intriguing speculation is that sensation and responses to urea through TlpB may also be coupled to sensing cytoplasmic pH through TlpD , through a yet unknown mechanism . Our observation that lacking TlpB enhances H . pylori’s sensitivity to detecting higher pH suggests this possibility . Our experiments elucidating acid sensing were all performed under conditions where urea was absent in the culture medium . Urea was only present when we deliberately introduced it in solution with HCl in the needle to investigate H . pylori’s response to multiple signals . This allowed us to pinpoint specifically the acid-sensing functions of the chemoreceptors . However , in vivo , H . pylori is exposed to both HCl and urea . While urea is a potent chemoattractant sensed through TlpB , which may have an effect on pH sensing , it is also the substrate for H . pylori’s highly expressed urease enzyme , whose activity certainly affects pH sensing . As a neutrophile , urease buffering activity is essential for H . pylori’s survival in vivo . In vitro studies have shown that H . pylori is able to survive under pH 1 conditions for several hours if the bacteria are in the presence of urea [27] . Many studies have been conducted to elucidate the intricate mechanism of how acidic conditions in the environment trigger a cascade of events resulting in an increase in cytoplasmic pH while the external pH remains acidic . Upon exposure to acidic pH conditions , urease assembles into a complex with a proton-gated urea channel embedded in the inner membrane [28–30] . Urea enters through this proton-gated channel to reach cytoplasmic urease where it optimally functions to degrade urea into ammonia and bicarbonate thereby raising the pH [31] . Based on the findings in these prior studies , we predict that in the presence of urea and urease H . pylori will be less responsive to an acid gradient in our microgradient assay . Future studies will need to integrate the role of urease and urea into acid sensing since it alters the intracellular and extracellular pH . It would be interesting to determine how urease activity might affect the sensitivities of TlpA and TlpD to acidic pH . This may explain why TlpD is more sensitive to pH since it detects changes in the cytoplasm . Our newer data show that TlpA and TlpD are the primary acid sensors in H . pylori that allow immediate response to HCl gradients and are important for stomach colonization in the presence of gastric acid . We find that a mutant lacking both chemoreceptors has a severe defect in establishing colonization in the murine stomach . It has been previously reported that TlpD is important for H . pylori survival and proliferation in the antrum [20] . We also found that ΔtlpD H . pylori has an approximate 10-fold defect in overall colonization of the stomach at two weeks post-infection ( Fig 6 ) . This could be due to the loss of chemotaxis towards other signals important for survival in addition to acid sensing . Indeed , we report here that TlpD also mediates attraction to environments of higher pH . However , TlpD does contribute to acid sensing in vivo , since deletion of TlpA has no effect on colonization in vivo , yet double deletion of TlpA and TlpD markedly worsens the colonization defect of the ΔtlpD H . pylori mutant ( Fig 6 ) . Interestingly , it has also been reported that tlpD is one of the most upregulated genes when H . pylori is exposed to acidic conditions in vivo [32] . These results indicate that chemotaxis through TlpD is most crucial for establishing colonization of the stomach , but that TlpA can compensate for the loss of acid-sensing through TlpD . Furthermore , we discovered that sensing through TlpA and TlpD is important for localizing properly to the gastric glands and that blocking acid secretion with omeprazole changes the distribution of gland-associated H . pylori in the stomach . We previously reported that bacteria growing as microcolonies in the gastric glands lead to inflammation in the regions of gland colonization . In addition , gland-associated H . pylori locally activate and induce proliferation of the stem cells in the infected glands leading to hyperplasia [11] . A previous investigation of the effects of omeprazole on H . felis distribution in the mouse stomach reported that acid suppression extends the colonization range of the bacteria into the corpus glands [33] . A similar finding was reported for mice infected with H . pylori and treated with omeprazole where the stomach glands were qualitatively found to contain less bacteria in the antrum than in the corpus [34] . Pulsed omeprazole dosing in gerbils has also been shown to alter the orientation of H . pylori within the mucus layer relative to the gastric epithelium , which could promote bacterial clearance by driving the bacteria closer to the lumen or perhaps improving the efficiency of antibiotics [35] . Our results support these previous findings that reducing gastric acidity through proton pump inhibitor treatment changes bacterial density and distribution in the gastric glands , extending their range within this niche . Our results further suggest that targeting chemosensation through TlpA and TlpD would interfere with overall stomach and gland colonization . H . pylori infection in humans most often leads to chronic inflammation in the antrum ( antral predominant gastritis ) . People with antral gastritis usually have no symptoms but may develop pyloric or duodenal ulcers as a consequence of the infection . People at risk for gastric adenocarcinoma are different in that they develop an anatomical pattern of gastritis that extends into the corpus . Corpus gastritis is associated with the loss of parietal cells , low acid secretion , high gastrin production and gastric atrophy ( multifocal atrophic gastritis ) [36] . These differences in the anatomical localization of the inflammatory changes in H . pylori infection has been ascribed to physiological differences in individuals , but could also reflect the anatomical site of infection of gland-associated H . pylori . That is , extension of the distribution of gland-associated bacteria from the antrum to the corpus glands may precede and contribute to the inflammatory and hyperplastic changes that lead to tissue pathology [11] . In our experimental system , gastric colonization of the corpus is promoted by proton pump inhibitor ( PPI ) treatment , suggesting that this could be a precursor step towards the development of multifocal atrophic gastritis and pre-neoplasia . Experiments in Mongolian gerbils , for example , have shown that PPI treatment can promote the development of adenocarcinoma [37] . In humans , the role of PPIs in atrophic gastritis remains controversial , but it is recommended that patients considered for long-term PPI therapy first be tested and treated for H . pylori infection [38 , 39] . Our study , and others , supports this clinical recommendation and also suggests that targeting pH sensing through TlpA and TlpD may be an effective way of disrupting H . pylori colonization in the stomach .
The previously published wild-type H . pylori strain PMSS1 [40] , ΔureAB and single knock-outs chemoreceptor mutants [15] were used for all experiments in this study except those strains noted in S4 Fig . H . pylori strains were either grown on Columbia blood agar plates or in Brucella broth supplemented with 10% fetal bovine serum ( BB10 ) at 37°C , 10% CO2 , as described previously [41] . The ΔtlpAB , ΔtlpAC , ΔtlpAD , ΔtlpBC , ΔtlpBD , and ΔtlpCD PMSS1 isogenic mutants were constructed by natural transformation with genomic DNA from the relevant single knock-out chemoreceptor mutant made in strain PMSS1 . The PMSS1 chemoreceptor mutants were verified by immunoblotting using an antibody that recognizes a conserved domain in all four chemoreceptors , courtesy of K . Ottemann [42] . All solutions were diluted or dissolved in sterile , distilled , and deionized water . Urea and sodium hydroxide solutions were made by dissolving ultra-pure urea or sodium hydroxide tablets into sterile , distilled and deionized water . One molar hydrochloric acid ( Sigma Aldrich , Saint Louis , MO ) was diluted into sterile , distilled and deionized water to create various concentrations of hydrochloric acid solutions to test . To obtain phosphate buffered solutions with different pHs , 1 M monobasic sodium phosphate ( NaH2PO4 ) pH 4 . 1 and 0 . 5 M dibasic sodium phosphate ( Na2HPO4 ) pH 9 . 0 were combined in various proportions until desired pH is achieved and confirmed by pH meter . To obtain citrate buffered solutions with different pHs , 1 M citric acid ( C6H8O7 • H2O ) pH 2 . 0 and 1 M sodium citrate ( C6H5O7Na3 • 2H2O ) pH 8 . 6 were combined in various proportions until desired pH is achieved and confirmed by pH meter . The buffering capacity and corresponding pH of various concentrations of HCl in BB10 was determined empirically ( S6 Fig ) . H . pylori cultures used for the assay were made by subculturing from a 16-hour overnight Brucella broth + 10% FBS ( BB10 ) culture with a starting OD600 of 0 . 15 . The subcultures ( also in BB10 ) were grown for 6 hours until they reach an OD600 of 0 . 3 before using in the microgradient assay . The pH of bacterial cultures prior to the microgradient assay ( spent media ) was measured at a mean of 6 . 65 +/- 0 . 04 SD ( S1B Fig ) . These were not different from the pH of BB10 kept in the incubator without bacteria measured at 6 . 6 , or freshly made BB10 ( pH 6 . 6–6 . 8 ) . The bacterial culture media for all strains tested , except ΔureAB , contain no detectable urea at the time of the assay due to bacterial urease activity during the subculture ( S1A Fig ) . This was determined by measuring the concentration of urea using the QuantiChrom Urea Assay Kit ( BioAssay Systems ) , a colorimetric assay that quantifies urea directly with a detection level of 13μM . Background was determined by quantifying urea in urease-treated BB10 . Twenty five milliliters of fresh BB10 ( predicted to contain 2–4 mM urea ) were incubated with 300 units of Jack Bean urease for 2 hours at 25°C . This amount of urease is calculated to liberate 0 . 3 millimole of ammonia in 1 min at 25°C , and therefore is sufficient to hydrolyze all the urea in the media in about one minute . To ensure complete urea degradation , we allowed the reaction to proceed for 2 hours before boiling the urease-treated media for 10 minutes to inactivate urease and filtering the media through a 0 . 2 μm filter to remove particulates . Two hundred seventy microliters of the subculture were placed into the center of a glass-bottom 35-mm dish ( MatTek ) contained by a ring of vacuum grease . The dish was placed above a 32x objective of a Zeiss Axiovert-35 inverted microscope equipped with phase-contrast optics and a heated stage ( 37°C ) . A Hammamatsu C2400 video charge-coupled-device ( CCD ) camera was used to record via an Argus-20 image processor onto Quicktime at 30 frames per second . A Femtotip II microinjection micropipette ( Eppendorf ) containing 8 μl of the test solution was inserted into or removed from the viewing field using a micromanipulator ( Eppendorf 5171 ) [7] . To create a microscopic gradient ( microgradient ) in the bacterial culture , a compensation pressure of 30 hPa was applied via the Eppendorf transjector 5246 to maintain a constant flow at 0 . 372 picoliters/minute from the tip . This pressure was determined empirically and selected because it did not physically affect the bacteria swimming near the micropipette tip while generating a stable gradient within the viewing field . H . pylori chemotactic responses were quantified from phase contrast video microscopy movies recorded at 30 frames per second using ImageJ software version 1 . 46r . Movies were analyzed starting from 4 to 10 seconds before micropipette insertion ( pre-injection ) to 10–30 seconds after micropipette insertion ( post-injection ) . After background subtraction and contrast adjustment , movies were then subsequently quantified or bacterial traces were obtained for visual depiction of swimming behavior in response to the gradient . For quantification , 0 . 2 second segments of the movies ( 6 frames ) were combined into Z-projections to generate traces of moving bacteria . Motile bacteria were detected by specifying a size between fifty to three hundred pixels and circularity values between 0 . 1 to 0 . 5 ( non-motile particles will be more circular than motile bacteria , which will be elliptical traces ) . These parameters were set in the Analyze Particle function in ImageJ and the output provided the number of motile bacteria per frame . The number of motile bacteria per frame is normalized to the average number of motile bacteria prior to needle entry . Responses are shown as either scatter plots , which display the percent bacteria per field from 4–5 seconds pre-injection to 10–15 seconds post-injection , or bar graphs , which display the percent bacteria per field at 4 second pre-injection and 10 seconds post-injection . Statistical significance between the percent bacteria per field pre-injection versus percent bacteria per field post-injection in the bar graphs or of the overall response in the scatter plots was assessed using a 2-way repeated measures ANOVA . To produce the images that depict the bacterial swimming behavior , longer traces were generated to clearly show the response . The time stated in the pre-injection panel indicates the first frame until 0 seconds used to generate the trace ( i . e . -1 . 5 sec means frames from 1 . 5 seconds prior to needle entry to 0 second when the needle enters were used to generate a 1 . 5 trace representing the pre-injection swimming behavior ) . The range of time stated in the post-injection panels indicates the frames used to generate the trace ( i . e . 18 . 5–20 sec means the frames from 18 . 5 seconds to 20 seconds post-injection were used to generate a 1 . 5 trace representing the post-injection swimming behavior ) . H . pylori cells were harvested from blood plates and lysed with 1x SDS sample buffer . Lysates were boiled for 5 minutes and then separated on a 10% SDS-PAGE gel . After transfer onto nitrocellulose membranes , H . pylori TlpA , TlpB , TlpC , TlpD were detected by blotting with rabbit anti-tlpA22 , an antibody that recognizes a conserved domain in all 4 chemoreceptors , courtesy of K . Ottemann [42] , followed by a goat anti-rabbit Alexa Fluor 660 . UreA , the small subunit of the abundantly expressed urease protein , was used as a loading control and detected with a mouse anti-UreA antibody followed by a goat anti-mouse Alexa Fluor 800 . The blot was then scanned with a Licor-Odyssey scanner at 700 and 800 nanometers . 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 9677 . Animals were euthanized by CO2 asphyxiation followed by cervical dislocation . 6 week-old female C57BL/6J mice were purchased from the Jackson Laboratory ( Bar Harbor , ME ) . Animals were infected intraorally by allowing the animals to drink a 5 microliter suspension containing 108 CFU of H . pylori grown in BB10 from a pipette tip [7] . One cohort of animals was treated with omeprazole at 400 μmol/kg/day via drinking water for seven days after seven days of infection . A second cohort of animals was treated with omeprazole ( 400 μmol/kg/day ) for three days prior to infection and then maintained on omeprazole for the duration of the two-week infection . The control cohort was given untreated water throughout the course of the two week infection . Animals were sacrificed at 2 weeks post-infection by CO2 asphyxiation . The stomach was harvested with the forestomach removed and discarded , opened via the lesser curvature , and laid flat onto filter paper . Luminal content was removed and the stomach was divided into halves that spanned the corpus to the antrum . One half of the stomach was weighed and mechanically homogenized for 30 seconds in 1 ml Brucella broth . Homogenized stomachs were serially diluted and plated for CFU counts . The data represent the number of CFU/gram of stomach . Bars represent the geometric mean . Statistical significance in the recovered bacterial load between strains was assessed by a Mann Whitney test . The efficacy of the omeprazole treatment was determined via a separate experiment where 5 mice were administered omeprazole in their drinking water for 3 days prior to sacrifice and 5 mice were given water for 3 days prior to sacrifice ( S13 Fig ) . To measure the pH of the stomach , the stomach was cut along the lesser curvature and splayed open . The stomach tissue along with its contents were placed at the bottom of a test tube ( with opening wide enough to fit a pH probe ) with the stomach lumen facing up . One milliliter of sterile water was added to the stomach in the tube . The pH was then measured using a pH meter . The other half of the stomach was fixed in a 2% paraformaldehyde and stained for confocal microscopy as previously described [7 , 11] . A custom made rabbit anti-H . pylori PMSS1 antibody and chicken anti-rabbit Alexa Fluor 488 antibody was used . DAPI ( 4 = , 6-diamidino-2-phenylindole ) and Alexa Fluor 594-phalloidin ( Molecular Probes ) were used for visualization of the nuclei and actin cytoskeleton . Samples were imaged with a Zeiss LSM 700 confocal microscope and z-stacks were reconstructed into 3D images using Volocity software ( Improvision ) . Number of bacteria per gland was determined with measurement functions in Volocity . All microgradient assay data comparing percent bacteria per field at 4 seconds pre-injection to percent bacteria per field at 10 seconds post-injection as well as the scatter plots of percent bacteria per field over time were analyzed via a 2-way repeated measures ANOVA test in the GraphPad Prism 7 software program . The p-value for the scatter plots indicates the significance of the time by group interaction via a 2-way repeated measures ANOVA . For animal experiments , statistical significance was assessed via a Mann Whitney test . Center values are geometric means and error bars represent standard deviation ( s . d . ) . n indicates the number of movies per condition or the number of animals used . NS indicates no statistical significance , * P < 0 . 05 , ** P < 0 . 01 , *** P < 0 . 001 , **** P < 0 . 0001 .
|
Helicobacter pylori is a bacterium that chronically infects the stomachs of 50% of humans , and infection can lead to serious diseases like peptic ulcers and stomach cancer . To survive , H . pylori rapidly senses acid and swims away to the protective mucus layer covering the stomach surface . The bacteria also burrow deep into the glands of the stomach through their narrow fissures and channels , and live in close contact with the cells lining the stomach . We report here that two H . pylori chemoreceptors , TlpA and TlpD , are the dominant acid sensors enabling H . pylori to discern and respond to minute changes in acid levels . H . pylori mutants lacking both TlpA and TlpD are unable to sense acid and are severely impaired in their survival in the murine stomach . By treating animals with omeprazole , a drug that blocks acid production , we restored the ability of the acid-sensor mutant to survive in the stomach . In addition , we found that blocking stomach acid production extended the range , distribution , and density of H . pylori living deep in the gastric glands . Our study provides new insights into H . pylori’s acid sensing machinery and how manipulation of acid gradients controls H . pylori’s localization and survival in the stomach .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"urea",
"swimming",
"pathology",
"and",
"laboratory",
"medicine",
"chemical",
"compounds",
"ureases",
"enzymes",
"pathogens",
"microbiology",
"enzymology",
"organic",
"compounds",
"biomechanics",
"biological",
"locomotion",
"helicobacter",
"stomach",
"bacteria",
"bacterial",
"pathogens",
"digestive",
"system",
"proteins",
"medical",
"microbiology",
"acids",
"microbial",
"pathogens",
"chemistry",
"pathogen",
"motility",
"gastrointestinal",
"tract",
"biochemistry",
"helicobacter",
"pylori",
"anatomy",
"organic",
"chemistry",
"hydrochloric",
"acid",
"virulence",
"factors",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"organisms"
] |
2017
|
Multiple Acid Sensors Control Helicobacter pylori Colonization of the Stomach
|
Microbes may maximize the number of daughter cells per time or per amount of nutrients consumed . These two strategies correspond , respectively , to the use of enzyme-efficient or substrate-efficient metabolic pathways . In reality , fast growth is often associated with wasteful , yield-inefficient metabolism , and a general thermodynamic trade-off between growth rate and biomass yield has been proposed to explain this . We studied growth rate/yield trade-offs by using a novel modeling framework , Enzyme-Flux Cost Minimization ( EFCM ) and by assuming that the growth rate depends directly on the enzyme investment per rate of biomass production . In a comprehensive mathematical model of core metabolism in E . coli , we screened all elementary flux modes leading to cell synthesis , characterized them by the growth rates and yields they provide , and studied the shape of the resulting rate/yield Pareto front . By varying the model parameters , we found that the rate/yield trade-off is not universal , but depends on metabolic kinetics and environmental conditions . A prominent trade-off emerges under oxygen-limited growth , where yield-inefficient pathways support a 2-to-3 times higher growth rate than yield-efficient pathways . EFCM can be widely used to predict optimal metabolic states and growth rates under varying nutrient levels , perturbations of enzyme parameters , and single or multiple gene knockouts .
Metabolic networks are shaped by evolution . In well-mixed , nutrient-rich environments , fast-growing bacteria are favored by natural selection . Such environments are commonly studied in laboratory settings , but natural environments are more diverse . In isolated ecological niches with limited resources , it is the total number of offspring cells , rather than fast growth , that determines evolutionary success . This puts a selection pressure on biomass yield ( biomass produced per amount of the limiting nutrient , e . g . glucose ) rather than on growth rate ( biomass produced per time and per cell biomass ) . Mechanistically , growth rate and yield might be expected to go hand in hand . It seems logical that a cell with a higher yield—i . e . one that can produce offspring from a smaller amount of nutrients—would also produce a larger number of offspring per time . However , in experiments we observe exactly the opposite; many fast-growing cells employ low-yield metabolic pathways ( e . g . yeast cells ( Crabtree effect ) and cancer cells ( Warburg effect ) [1] ) , and also many bacteria display a wasteful respiro-fermentative overflow metabolism and still attain high growth rates . Pure respiratory growth would give rise to a higher biomass yield per mole of glucose , but to lower growth rates . Since yield-inefficient metabolic strategies are widely observed , under various circumstances and in evolutionarily unrelated organisms , it has been suggested that growth rate and yield may be in conflict for physicochemical reasons . During evolution , such a conflict may lead to “tragedy-of-the-commons” situations in which yield-inefficient microbes gain an evolutionary advantage by over-exploiting shared resources [2–4] . The hypothesis of a general trade-off is supported by simple cell models in which high-yield pathways display lower thermodynamic forces or higher enzyme costs [5–7] . The rate/yield trade-off has been tested by lab-evolution experiments with fast-growing microorganisms , with varying levels of success . Growth rate and yield have been compared between different wild-type and evolved microbial strains [8–11] , but most studies found poor correlations between growth rate and yield . Novak et al . [9] found a negative correlation within evolved E . coli populations , indicating a rate/yield trade-off . A rare example of bacteria evolving for high yield in the laboratory was in the work of Bachmann et al . [12] . In their protocol , cells grow in separate droplets in a medium-in-oil suspension , simulating a fragmented environment , and offspring cells are mixed when the nutrients in the droplets have been depleted , and then resuspended . This creates a strong selection pressure for maximizing biomass yield . Indeed , the strains evolved towards higher yields at the expense of their growth rate , again indicating a trade-off between the two objectives . However , evidence from all these experiments may not be conclusive , because microorganisms may behave sub-optimally in the laboratory experiments . Thus , is the rate/yield trade-off universal ? We claim that the answer to this question lies in metabolism , especially in enzyme demand . At balanced growth , the relative amounts of all cell components remain constant in time , including the protein fraction associated with metabolic enzymes . If a metabolic strategy achieves a given biomass synthesis rate at a lower enzyme demand , the freed protein resources can be reallocated to other cellular processes that contribute to growth , and the cell’s growth rate can increase . Thus , a metabolic strategy will be growth-optimal if it minimizes enzyme cost at a given biomass synthesis rate [13] . In theory , the use of a high-yield flux mode affects the growth rate in two opposite ways . On the one hand , a high-yield mode achieves the same rate of biomass production at a lower glycolytic rate , and the lower enzyme demand in glycolysis allows for a higher growth rate . On the other hand , high-yield modes dissipate less Gibbs free energy [5] , which may slow down the reactions and must be compensated by higher enzyme levels , leading to lower growth rates [7 , 14 , 15] . The second effect may be obscured if another substrate , such as oxygen , provides additional driving force . When the first effect dominates , high-yield modes allow for a higher biomass production per enzyme invested , so yield and growth rate are maximized by a single flux mode . When the second effect dominates , it is low-yield modes that provide a growth advantage [6 , 13 , 16–18] , and there will be a trade-off: growth rate and yield are maximized by different flux modes , and there may be other modes in between that provide optimal compromises . In summary , a rate/yield trade-off in cells reflects a trade-off between enzyme efficiency and substrate efficiency in metabolism; and since the enzyme cost of a given pathway flux depends on external conditions , the occurrence of rate/yield trade-off will be condition-dependent as well . How can we describe this by models ? The specific growth rate μ for exponentially growing cells is given by the rate of biomass synthesis per cell dry weight and is typically measured in grams of biomass per gram cell dry weight per hour . The biomass yield YX/S is measured in grams of biomass per carbon mole of nutrient ( i . e . per 1/6 mole of glucose ) . If the carbon uptake rate vS were known , we could directly convert between yield and growth rate using this formula: Y X / S = μ v S . However , since carbon uptake , yield , and growth rate are tightly coupled , the changes in vS are hard to predict . Classic Flux Balance Analysis ( FBA ) places an upper bound on vS . If this is the only active flux bound , then maximizing biomass rate coincides with maximizing biomass yield , leaving no possibility for rate/yield trade-offs . Other constraint-based methods , such as FBA with Molecular Crowding [19] or Resource Balance Analysis [20] , account for enzyme costs . They can be used to explore the trade-off , but they are not fully quantitative because they ignore the kinetic and thermodynamic effects of varying metabolite concentrations ( see Discussion section for details ) . In [21] , a kinetic pathway model was used to directly compute the enzyme costs . Two variants of glycolysis , both common among bacteria , were compared by their ATP yields on glucose and by their ATP production per enzyme investment . At a given glucose influx , the Embden-Meyerhof-Parnas ( EMP ) pathway yields twice as much ATP , but was found to use more than 4 times as much enzyme than the Entner-Doudoroff ( ED ) pathway . This suggested that cells under yield selection should use the EMP pathway , while cells under rate selection should use the ED pathway instead . Aside from simple approximations [22 , 23] , the enzyme economics of other metabolic choices , e . g . respiration versus fermentation , and the resulting trade-offs , remain to be quantified . Here we combine a calculation of enzyme cost , based on kinetic models , with elementary flux mode analysis . Elementary flux modes ( EFMs ) describe the fundamental ways in which a metabolic network can operate [24–27] . Among the steady-state flux modes , EFMs are minimal in the sense that they do not contain any smaller subnetworks that can support a steady-state flux mode [24 , 25 , 27] . EFMs might be expected to have simple shapes in the network , but since biomass production requires many different precursors , biomass-producing EFMs can be highly branched . All biomass-producing EFMs are free of thermodynamically infeasible loops , and if the flux directions are predefined , the set of steady-state flux distribution is a convex polytope spanned by the EFMs . The EFMs of a metabolic network can be enumerated , and thermodynamically infeasible modes can be efficiently discarded [28 , 29] , but in practice an enumeration of EFMs may be impossible because of their large number . EFMs have a remarkable property , which makes them well-suited for studying rate/yield trade-offs: in kinetic metabolic models , the biomass production per enzyme investment is maximized by a vertex point of the flux polytope , and in models without flux bounds , all these vertices are EFMs [30–32] . The yield of an EFM , defined as the output flux divided by the input flux , is easy to compute and it is again an EFM that achieves the maximal yield among flux modes . Therefore , to find flux modes that maximize cell growth , we can enumerate the EFMs and assess them one by one; and to determine rate/yield trade-offs , we simply plot yields versus growth rates of all EFMs ( Fig 1 ( a ) ) .
To predict optimal metabolic fluxes and cell growth rates , we developed Enzyme-Flux Cost Minimization ( EFCM ) , a method for computing flux modes that realize a linear flux objective at a minimal enzyme cost . Constraint-based methods such as Flux Balance Analysis are entirely based on reaction stoichiometries . Some of them also use approximate enzyme costs , for instance the sum of absolute fluxes [33] or other linear/quadratic functions of the flux vector [19] . EFCM , in contrast , computes enzyme cost based on a given kinetic model . In our model , the flux objective represents biomass production , i . e . the production of small molecules and macromolecules that constitute the cell and do not explicitly appear in the network model . Below we argue that enzyme-optimal flux modes , with such a flux objective , are the ones that allow for maximal growth rates . To compute the maximal growth rate achievable , we use a kinetic model of metabolism , consider all possible flux modes , and compute for of them the optimal enzyme allocation pattern , i . e . the pattern that realizes the required fluxes at a minimal total enzyme investment . Enzyme Cost Minimization ( ECM ) is a method that finds optimal enzyme and metabolite profiles supporting a given flux distribution [34] . The ECM problem can be quickly solved using convex optimization , and the minimal enzyme cost of all EFMs can be computed in reasonable time ( a few minutes on a shared server , for models with ∼ 103 EFMs such as E . coli core metabolism ) . Knowing the enzyme investment per biomass production , we next compute the cellular growth rate . For each EFM , the enzyme demand per biomass production is translated into a mass doubling time ( i . e . the amount of time that metabolism would have to run in order to duplicate all metabolic enzymes assumed in our model ) . The mass doubling time can be translated into a cell growth rate by a semi-empirical formula ( see Methods and Figure 1 in S1 Text ) . Since EFCM does not impose any constraints on fluxes , the enzyme-specific biomass production—and thus growth rate—is maximized by elementary flux modes , regardless of the values chosen for kinetic parameters [30 , 31] . To see this , we consider all feasible steady-state flux modes , constrained to predefined flux directions and normalized to a unit biomass production rate . These flux modes form a convex polytope in flux space ( see Fig 1 ( b ) ) . The flux cost function is concave on this polytope [30] , or even strictly concave for some rate laws [32] , and so the minimal enzyme cost is achieved by a polytope vertex . In models without any active flux bounds , all these vertices are EFMs . Thus , to predict optimal flux modes , we need not scan all feasible flux modes , but can simply choose among EFMs . From our ECM calculations , we obtain the full spectrum of growth rates and yields of all EFMs . The rate/yield spectrum , a scatter plot between the two quantities , displays the possible trade-offs . We now focus our attention on flux modes that maximize growth at a given yield , or maximize yield at a given growth rate . Such modes , which are not dominated by any other flux mode in terms of growth rate and yield , are called Pareto-optimal . They represent optimal compromises between growth rate and yield . If we could evaluate the growth rates and yields for all metabolic states in the model ( including non-elementary flux modes ) , the resulting rate/yield points would form a dense , non-convex set . The border of this set , as drawn in Fig 1 ( a ) , is called the Pareto front . The EFMs on this front mark a selection of best compromises between growth rate and yield achievable in the model . By inspecting the rate/yield spectrum , we can tell whether there is an extended Pareto front or rather one metabolic state that optimizes both rate and yield . Even if growth and yield are positively correlated among all EFMs , the modes along the Pareto front will show a negative correlation whenever an extended front exists . Therefore , it is the size of the Pareto front that shows the extent of a rate/yield trade-off . While the yields are fixed properties of the EFMs , the growth rates depend on external conditions , and so does the rate/yield trade-off . We demonstrate this for a case study on E . coli bacteria , which have often been used for experiments on the rate/yield trade-off [9 , 35–37] and whose enzyme kinetics are relatively well studied . To study growth rates and yields in E . coli , we applied EFCM to a model of core carbon metabolism . Our model , a modified version of the model presented in [38] , comprises glycolysis , the Entner-Doudoroff pathway , the TCA cycle , the pentose phosphate pathway and by-product formation ( see Fig 2 ( a ) , and Section 2 in S1 Text ) . The biosynthesis of macromolecules ( “biomass” ) from small metabolites and cofactors is not explicitly described , but summarized in an overall reaction for biomass production . Reaction kinetics are described by modular rate laws [39] , and kinetic constants were obtained by parameter balancing [40] based on a large set of values reported in the literature ( see Section 1 . 1 in S1 Text ) . The yield of an EFM is defined as grams of biomass produced per mole of carbon atoms taken up in the form of glucose . EFMs that simultaneously use oxygen-sensitive enzymes ( pfl ) and oxygen-dependent reactions within the electron transport chain ( oxphos or sdh ) cannot be used by the cell . After discarding such EFMs , we obtained 568 EFMs that produce biomass under aerobic conditions and 336 under anaerobic conditions . 97 of these EFMs can operate under both conditions ( Fig 2 ( b ) ) . Statistical properties of the EFMs ( size distribution , usage of individual reactions , and similarities between EFMs and measured fluxes ) are shown in Figure 7 in S1 Text . If all EFMs required the same total enzyme amount at unit glucose uptake , growth rates and yields would be proportional . Alternatively , if all EFMs required the same total enzyme amount at a unit biomass production , all EFMs would have exactly the same predicted growth rate , regardless of yield . Instead of these naïve approximations , we can now use our kinetic model and the EFCM method to obtain the actual spectrum of possible growth rates and yields ( Fig 2 ( c ) ) . While the yields are constant properties of the EFMs , the growth rates depend on enzyme demands and therefore on kinetics and extracellular nutrient levels . As reference conditions , we chose [glucose] = 100 mM , [O2] = 0 . 21 mM . To visualize groups of similar EFMs , we used t-distributed Stochastic Neighbor Embedding ( t-SNE ) , a machine learning algorithm for nonlinear dimensionality reduction [41] . The algorithm found five major clusters of EFMs , which loosely correspond to metabolic strategies ( e . g . aerobic acetate-secreting EFMs ) . Since no kinetic information was used in t-SNE , we were surprised to find all EFMs with high growth rates in a single cluster ( see Figure 6 in S1 Text ) . To compare typical metabolic strategies , we focused on five EFMs with different characteristics and followed them across different external conditions and sets of kinetic parameters . We also show an experimentally determined flux distribution , called exp [42] ( for calculations see Section 4 . 1 in S1 Text ) . These focal EFMs are marked by colors in Fig 2 ( b ) and listed in Table 1 . Flux maps ( produced using software from [43] ) can be found in Section 5 . 3 in S1 Text . The first three focal EFMs are located on the Pareto front . max-yield , the EFM with the highest yield , does not produce any by-products nor does it use the pentose-phosphate pathway . max-gr ( whose flux map is shown in Fig 2 ( a ) ) has a slightly lower yield , but reaches the highest growth rate ( 0 . 739 h−1 ) in our reference conditions . It uses the pentose-phosphate pathway with a relatively high flux . In addition , we chose another EFM from the Pareto front ( denoted pareto ) with a growth rate and yield between the two extreme EFMs . Curiously , the EFMs along the Pareto front span only a narrow range of biomass yields ( 18 . 6—22 . 1 ) , so there is almost no rate-yield trade-off . This is not a trivial finding , and other choices of parameters or extracellular conditions can lead to broader Pareto fronts: in low-oxygen conditions , the trade-off between growth rate and yield becomes much more pronounced . To study by-product formation , we consider two other EFMs below the Pareto front: an anaerobic lactate-fermenting mode ( ana-lac ) with a very low yield ( 2 . 1 g/C-mol ) and an aerobic , acetate-fermenting mode ( aero-ace ) with a medium yield ( 15 . 2 g/C-mol ) . Interestingly , ana-lac has a ∼10 times lower yield , but it still reaches about one third of the maximal growth rate , thanks to the lower enzyme cost of pentose phosphate pathway and lower glycolysis , as compared to TCA cycle and oxidative phosphorylation ( per mol of ATP generated ) . This recapitulates a classic rate-versus-yield problem associated with overflow metabolism . Among all by-product forming EFMs , some acetate-producing EFMs have the highest growth rates , which might explain why E . coli , in reality , excretes acetate in aerobic conditions rather than lactate or succinate . Nevertheless , all by-product forming EFMs have lower growth rates than max-gr and are therefore not Pareto-optimal . Below we will see that this fact is subject to change when conditions are different , specifically at lower oxygen levels . To study how by-product secretion affects yield and growth rate in general , we focused on some major uptake or secretion fluxes and visualized these fluxes for all EFMs in the rate/yield spectrum ( Fig 3 ) . EFMs close to the Pareto front consume intermediate amounts of oxygen and do not secrete any acetate , lactate or succinate . Another group of EFMs ( shown in red in Fig 3 ( b ) ) consume slightly less oxygen , but secrete large amounts of acetate . Compared to pure respiration , these aerobic fermentation modes provide lower biomass yields . Other important fluxes are shown in Figure 8 in S1 Text . The growth rate achieved by a flux mode depends on environmental conditions and enzyme parameters . To study this quantitatively , we varied some model parameters and traced their effects on the rate/yield spectrum . Fig 4 ( a ) shows how lower oxygen levels affect the growth rate of oxygen-consuming EFMs . Lower oxygen levels need to be compensated by higher enzyme levels in oxidative phosphorylation , which lowers the growth rate ( Fig 4 ( b ) and Figure 16 in S1 Text ) . EFMs that function anaerobically , such as ana-lac , are not affected ( see Figure 18 in S1 Text for enzyme allocation ) . Therefore , a low oxygen level leads to a prominent rate/yield tradeoff , with a Pareto front spanning a wide range of growth rates and yields ( Fig 4 ( a ) ) . The effect of external glucose levels can be studied similarly ( Figures 12 and 16 in S1 Text ) : at lower external glucose concentrations , the PTS transporter becomes less efficient and cells must increase its expression in order to maintain the flux . This increases the total enzyme cost and slows down growth . Below a glucose concentration of 10−3 mM , the demand for transporter dominates the enzyme demand completely ( see Fig 5 ( b ) and Figures 17-18 in S1 Text for a breakdown of enzyme allocation ) . Since the PTS transporter is the only glucose transporter in our model , it is used by all EFMs , leading to a universal monotonic relationship between glucose concentration and growth rate . However , the detailed shape of the glucose/growth rate plot , known as the Monod curve [44 , 45] , depends on the PTS flux and on many other parameters that differ between EFMs ( see Section 3 . 3 in S1 Text ) ) . The performance of EFMs under high-glucose and low-glucose conditions is shown in Figure 19 in S1 Text . By varying the glucose and oxygen levels , we can screen a range of environmental conditions and obtain a two-dimensional Monod surface plot . The winning strategies , i . e . the EFMs with the highest growth rates can be depicted on this surface ( Fig 4 ( d ) and 4 ( e ) ) or in a glucose/oxygen phase diagram ( see Figures 13-15 in S1 Text , also for anaerobic conditions ) . More than 20 different EFMs achieve a maximal growth rate in at least one of the conditions scanned . To simplify the picture , we can focus on EFM features such as uptake rates and plot them on the Monod surface ( Fig 4 ( c ) –4 ( f ) ) . As expected , oxygen uptake ( Fig 4 ( d ) ) decreases when oxygen levels are low . This pattern occurs across the entire range of glucose levels , but the transition—from full respiration to acetate overflow ( Fig 4 ( e ) ) and further to anaerobic lactate fermentation EFMs ( Fig 4 ( f ) ) —is shifted at lower glucose levels . Interestingly , this transition disappears at extremely low glucose concentrations ( 0 . 1 μM ) , as the fully respiring pareto EFM exhibits the highest growth rate even at the lowest oxygen levels tested ( Figure 13 ( a ) in S1 Text ) . While glucose levels are relatively easy to adjust in experiments , it is difficult to measure oxygen levels in the local environment of exponentially growing cells . This has resulted in a long-standing debate about the exact conditions that E . coli cells experience in batch cultures [46–48] , and it makes it hard to validate our predicted transition from acetate fermentation to full respiration . Our model predicts that at a constant level of [O2] , E . coli will fully respire at low glucose levels and secrete acetate at high glucose levels ( see Fig 4 ) . A similar shift from pure respiration to a mixture of respiration and acetate secretion has been observed in chemostat cultures [49] , where higher glucose levels result from higher dilution rates . The choice of metabolic strategies does not only depend on external conditions , but also on enzyme parameters . As an example , we varied the kcat value of triose-phosphate isomerase ( tpi ) and traced changes in the rate/yield spectrum . Not surprisingly , slowing down the enzyme decreases the growth rate ( see Figure 20 in S1 Text ) . But to what extent ? Two of our focal EFMs ( max-gr and pareto ) are not affected at all , since they do not use the tpi reaction . All other focal EFMs show strongly reduced growth rates . To study this systematically , we predicted the growth effects of all enzyme parameters in the model ( equilibrium constants , catalytic constants , Michaelis-Menten constants ) by computing the growth sensitivities , i . e . the first derivatives of the growth rate with respect to the enzyme parameter in question ( see Section 4 . 2 in S1 Text , and supplementary data files ) . A sensitivity analysis between all model parameters and the growth rates of all EFMs ( or alternatively , their biomass-specific enzyme cost ) can be performed without running any additional optimizations ( Sections 4 . 3—4 . 4 in S1 Text ) . Growth sensitivities are informative for several reasons . On the one hand , parameters with a large impact on growth will be under strong selection ( where positive or negative sensitivities indicate a selection for larger or smaller parameter values , respectively ) . On the other hand , these are also the parameters that need to be known precisely for reliable growth predictions . The parameters of a reaction can have very different effects on the growth rate . For example , the sensitivities of the kcat and KM values of pgi are low , but the growth rate is very sensitive to the Keq value . To study the effects of a gene deletion , we can simply discard all EFMs that use the affected reaction: based on a precalculated EFCM analysis of the full network , we can easily analyze the restricted network without any new optimization runs . By switching off pathways , we can easily quantify the growth advantage they convey . Instead of studying pathways in isolation as in Flamholz et al . [21] , we can study their usage as part of a whole-network metabolic strategy . Fig 6 shows an analysis for two common variants of glycolysis , the ( high ATP yield , high enzyme demand ) EMP and the ( low ATP yield , low enzyme demand ) ED pathway , across different external glucose and oxygen levels ( see Section 3 . 4 in S1 Text ) . At low oxygen levels and medium-high glucose levels ( 10 μM—100 mM ) , cells profit strongly from using the ED pathway , and knocking it out decreases the growth rate by up to 25% . The EMP pathway provides a much smaller advantage ( up to 10% ) , and only in a narrow range of low-oxygen conditions .
Our case study on E . coli metabolism reinforces the notion that growth rate and biomass yield are not strictly coupled . Instead , their correlations across EFMs , and the extent of rate/yield trade-offs along the Pareto front , depend on details such as growth conditions and enzyme parameters . At high oxygen levels , growth-maximizing flux modes have an almost maximal yield and the Pareto front is very narrow . In contrast , under low-oxygen conditions the highest growth rates are obtained by low-yield strategies and a long Pareto front emerges ( Fig 4 ( a ) ) . It is not surprising that experimental results indicating rate/yield trade-offs were inconclusive and difficult to interpret . As shown in [9] , wild-type cell populations might be far from the Pareto front , and a selection for fast growth may push the populations and individuals closer to it . It would be interesting to study whether these results are in fact dependent on oxygen availability . EFCM predicts which flux modes are likely to be used by well-adapted cells . We expected that the EFM with the highest growth rate ( max-gr , in the standard conditions chosen in this study ) would coincide with the experimentally determined flux mode ( exp ) in the same conditions . However , this is not the case , and the two flux modes are not even very similar ( correlation r = 0 . 41 , see Figure 7 ( c ) in S1 Text ) . Our model predicts a much higher maximal biomass yield than the yield measured in batch cultures ( 18 . 6 vs 11 . 8 gr dry weight per carbon mole [51] ) , while the predicted growth rate is slightly lower ( 0 . 74 vs 0 . 89 h−1 ) . However , for the experimentally determined flux mode ( exp ) , we overestimate the yield ( 17 . 7 vs 11 . 8 [42] ) and underestimate the growth rate ( 0 . 41 vs 0 . 89 ) as well , so some of the discrepancies may be due to weaknesses of our model ( e . g . wrong kinetic parameter values ) rather than due to EFCM itself . The overestimation of yield ( which depends on network structure , not on kinetics ) may be caused by the fact that our model misses some waste products or additional processes that dissipate energy , or that our high-yield EFMs are kinetically unfavorable in reality . The underestimated growth rates may result from our simplistic conversion of enzyme costs into growth rates . However , we hope that these over- and underestimations occur consistently across EFMs and do not affect the qualitative results of this study . In contrast to the much simpler model by Basan et al . [49] , our model does not predict growth-rate dependent acetate overflow as observed in E . coli . In our standard aerobic conditions ( see Fig 2 and Figure 14 ( h ) ) in S1 Text , the winning mode , max-gr , is completely respiratory and produces no fermentation products . Only at low oxygen levels , EFMs with acetate overflow , such as aero-ace , become favorable ( see Fig 4 and Figure 15 ( e ) in S1 Text ) . This misprediction may depend on several factors: First , we may have underestimated the effective cost of oxidative phosphorylation ( oxphos ) , which becomes costly at lower oxygen levels , or we may have overestimated the oxygen availability . The oxygen concentration of [O2] = 0 . 21 mM , which we chose to represent typical laboratory conditions , may be inaccurate; oxygen availability may be as complex as in yeast , where it seems to diffuse too slowly to supply the mitochondria fully with oxygen [48] . Moreover , the affinity of the reactions to oxygen is not precisley known , so even a precise value of the oxygen concentration would not suffice . Second , the experimentally observed acetate production may result from additional , growth-rate dependent flux constraints like those employed by Basan et al . in their model . In our model , we did not impose any bounds on fluxes ( aside from normalizing the flux modes to unit per biomass production ) , and thus metabolic efficiency is maximized by an EFM . The growth rate does not even appear in the optimization . We account for it only later , when metabolic efficiency is translated into an achievable growth rate . Thus , it is possible that we miss some physiological constraints such as membrane real-estate [52] , changing biomass composition , or extracellular oxygen diffusion rates . Even without flux constraints , some EFMs mix respiration and acetate production , e . g . aero-ace . However , none of them corresponds exactly to the fluxes observed experimentally . Moreover , the measured relative rate of acetate production increases continuously with the growth rate , which cannot be captured by a single constant EFM . A usage of flux constraints in EFCM would be possible and would allow us , for example , to limit certain fluxes or to enforce some minimal flux , e . g . in ATP-consuming maintenance reactions . To screen all vertices of the flux polytope , one may build on the concept of elementary flux vectors [53 , 54] . However , the number of these vertices may become very large , and whenever flux bounds are changing ( e . g . as a function of growth rate ) , this would change the set of polytope vertices , and the entire calculation would have to be done for each growth rate . Third , it is also possible that the experimentally observed acetate secretion is simply not optimal . In adaptive laboratory evolution experiments [36 , 37] , the evolved strains grew about 1 . 5 times faster without a significant change in yield , but most of this increase could be explained by an increasing glucose uptake because the relative rates of acetate overflow did not change . Apparently , if acetate secretion is due to a glucose uptake constraint , this constraint can be bypassed by mutations and cells may be able to decrease acetate secretion while growing faster . In a recent comparison of seven E . coli wild-type strains [35] , three strains were found to secrete no acetate at all in aerobic conditions ( on glucose ) , but to use a fully respiratory strategy without any by-product secretion . Two of these fully respiring strains grew just as fast as the evolved strains from the adaptive evolution studies ( about 1 . 0/h ) , and significantly faster than the lab strain that we used for our reference flux data and for the stoichiometric model ( K-12 ) . Again , this finding raises questions about universal rate/yield trade-offs and supports our conclusion that the trade-off may almost disappear in high-oxygen conditions . Some variants of FBA manage to predict flux distributions with a suboptimal biomass yield by putting bounds on enzyme investments . An example is FBAwMC ( Flux Balance Analysis with Molecular Crowding ) , which relates fluxes to enzyme demands and limits the cytoplasmatic protein density [55] . However , these methods are insensitive to environmental conditions: the crowding coefficients assigned to reactions are constants , and metabolite concentrations are not considered at all . In [20] , Müller et al . ran a kinetic optimization ( which attempts to solve the nonlinear enzyme minimization problem directly ) and compared it to a linear approximation called satFBA . In this approximation , the constraints are exactly like in FBAwMC , except that the crowding coefficients of exchange reactions are divided by saturation values . The saturation values , numbers between 0 and 1 , account for the concentrations of external metabolites such as glucose and oxygen . For a small metabolic network ( comprising 5 reactions ) , satFBA yields the same qualitative predictions as a kinetic optimization ( and EFCM , for that matter ) , in particular with regard to the rate/yield trade-off . However , satFBA assumes that transport reactions are the only reactions affected by metabolite levels , whereas EFCM models the interplay between metabolite levels , enzyme efficiencies , and enzyme investments in all enzymatic reactions . It remains to be seen whether satFBA , with its single kinetic bottleneck , can reproduce complex predictions of EFCM like the ones shown in Fig 4 . Constraint-based whole-cell models such as Resource Balance Analysis ( RBA ) [56 , 57] or ME-models [58] treat protein production as a part of the cellular network and couple metabolic rates to production rates of the catalyzing enzymes . These methods differ from EFCM in three main ways: in the modeling of protein production , of catalytic rates , and of biomass composition and enzyme cost weights . ( i ) While RBA and ME model protein production in detail , EFCM is limited to metabolism: the partitioning between metabolic enzymes and ribosomes is captured by a formula that effectively converts enzyme cost into growth rate ( see Methods ) . ( ii ) In reality , enzymes often operate below their maximal speed ( i . e . the kcat value ) , at a catalytic rate called apparent kcat value [59] . This capacity utilization lower than 1 depends on metabolite levels and is quantified by the efficiency factors of ECM [34] . For each enzyme , the capacity utilization computed by EFCM varies across EFMs , but remains close to some typical value . These values , for different enzymes , span almost the entire range between 0 and 1 ( see Figure 11 in S1 Text ) . In a linearized variant of EFCM that assumes full capacity utilization , the growth rate would be overestimated and the growth differences between EFMs would be distorted . In fact , our predicted enzyme cost is between 1 . 4 and 4 . 7 times higher ( depending on the EFM considered ) than the ideal costs of enzymes operating at their maximal capacity ( see Figure 3 in S1 Text ) . RBA avoids this problem by replacing the kcat values by empirically determined , growth-rate dependent apparent catalytic rates . Constraint-based methods that ignore this effect [23 , 60] underestimate the actual enzyme demand , thus suggesting an “unused enzyme fraction” in cells [61] . We think that “unexplained enzyme fraction” would be a better term , because the enzyme amount predicted for fully efficient enzymes is an ideal value that would simply not suffice to catalyze the required fluxes in reality , given all thermodynamic and kinetic constraints [34 , 62] . ( iii ) In contrast to RBA and ME models , EFCM assumes a fixed biomass composition and fixed cost weights for the enzyme molecules . This means that cells , in EFCM , lack some strategic options that exist in RBA and ME models: to fine-tune the biomass composition towards a usage of “cheap” precursors , or to decrease the cost weights of proteins by cost-optimizing the production of limiting protein components such as iron . Again , these options would be hard to implement in EFCM because biomass composition is a defining part of the stoichiometric model , and any growth-rate dependent changes in biomass composition would also change the set of EFMs . Although efficient protein allocation may be important for fast growth [63] , there is empirical evidence that cells do not always minimize enzyme cost . Lactococcus lactis , for example , can undergo a metabolic switch that leads to big changes in growth rate , but involves no changes in protein levels [64] . These cells could , in theory , save enzyme resources while maintaining the same metabolic fluxes , but do not do so—possibly because their enzyme levels provide other benefits , e . g . anticipating metabolic changes to come . EFCM ignores such complex objectives: it describes fully optimal , but “short-sighted” cell strategies which define a lower bound on the enzyme demand . By considering secondary objectives , e . g . , a need for preemptive protein expression or safety margins to counter expression fluctuations , one would predict higher demands and lower growth rates . Our study has demonstrated that enzyme kinetics is a useful addition to constraint-based flux prediction ( see Section 1 . 4 in S1 Text ) ) . In contrast to the minimal model in [49] , our model was not fitted to recapitulate a specific known phenomenon , but was made to derive predictions ab initio in the spirit of “testing biochemistry” [65] . As long as in vivo kinetic constants are not precisely known , this harbours the risk of mispredictions . Curiously , for example , the EFMs with the highest predicted growth rates bypass upper glycolysis and use the pentose phosphate pathway instead . On the contrary , an ab initio approach allows modelers to recover empirical laws directly from cell biological knowledge , for example , the shape of Monod curves and Monod surfaces ( see Figure 15 and Section 4 . 6 in S1 Text for general simplified Monod functions ) . It allows us to compute quantitative effects of allosteric regulation or mutated enzymes ( see Figure 2 in S1 Text ) , the residual glucose concentration in chemostats ( see Figure 15 in S1 Text ) , and the trade-offs between metabolic strategies at different glucose levels ( see Figure 19 in S1 Text ) . The decomposition into EFMs also greatly facilitates calculating the epistatic interactions between reaction knockouts ( see Figure 2 ( f ) in S1 Text ) . Although yield-related epistatic interactions were previously computed using FBA ( see Section 3 . 5 in S1 Text ) , environment-dependent epistatic effects on growth rate have not been computed so far . EFCM could be applied to larger models and models with flux constraints , and other cost functions could be implemented ( see Section 1 . 6 in S1 Text ) . As a fully mechanistic method , it puts existing biochemical models and ideas about resource allocation to test and enables us to address fundamental issues of unicellular growth and cell metabolism , such as the trade-off between growth rate and biomass yield .
A metabolic state is characterized by cellular enzyme levels , metabolite levels , and fluxes . All these variables are coupled by rate laws , which depend on external conditions and enzyme kinetics . The EFCM algorithm finds optimal metabolic states in the following way . First , we enumerate the elementary flux modes of a network , which constitute the set of potentially growth-optimal flux modes . Then we consider a specific simulation scenario , defined by kinetic constants and external metabolite levels , and compute the growth rates for all EFMs . To determine the optimal metabolic state—a state expected to evolve in a selection for fast growth—we choose the EFM with the highest growth rate . The optimal state ( v , c , E ) can be found efficiently by a nested screening procedure ( Fig 1 ( b ) and 1 ( c ) ) . First , we consider all EFMs , normalized to a given biomass production rate vBM . To determine the relative enzyme demand of an EFM , we predefine vBM , scale our EFM to realize this production rate , and compute the enzyme demand by applying Enzyme Cost Minimization ( ECM ) , i . e . an optimization of metabolite levels c and enzyme levels E . ECM has recently been applied to a similar model of E . coli’s core carbon metabolism [34] . It assumes a given flux distribution ( in our case , an EFM ) and treats the enzyme concentrations as explicit functions of substrate and product levels and fluxes . Given a flux mode v , we consider all feasible possible metabolite profiles ln c , consistent with the flux directions and respecting predefined bounds on metabolite levels . For each such profile , we compute the enzyme demands Ei and the total enzyme mass concentration Emet = ∑i wi Ei ( in mg l−1 ) , where wi denotes the molecular mass of enzyme i in Daltons ( mg mmol−1 ) and enzyme concentrations are measured in mM ( i . e . , mmol l−1 ) . As a function of the logarithmic metabolite levels , Emet is convex; this allows us to find the global minimum efficiently . In the model , we use common modular rate laws [39] , for which the enzymatic cost in log-metabolite space is strictly convex ( Joost Hulshof , personal communication ) . The optimized enzyme cost is a concave function in flux space [30–32] . This combination of convexity and concavity allows for a fast optimization of enzyme levels and fluxes for each condition and set of kinetic parameters . We implemented ECM in the Network-Enabled Optimization System ( NEOS ) , an internet-based client-server application that provides access to a library of optimization solvers . The NEOS Server is available free of charge and offers a variety of interfaces for accessing the solvers , which run on distributed high-performance machines enabled by the HTCondor software . The NEOS Guide website ( https://neos-guide . org ) showcases optimization case studies , presents optimization information and resources , and provides background information on the NEOS Server . Using our online service , users can run EFCM for their own models , using different rate laws . With our E . coli model , the optimization for one flux distribution takes a few seconds , and for the complete set of all EFMs several minutes on a shared Dell PowerEdge R430 server with 32 intel xeon cores . Details can be found in Section 1 . 2 in S1 Text , and on the web page ( www . neos-guide . org/content/enzyme-cost-minimization ) . Following the approach of Scott et al . [66] , cell growth rates can be predicted from the demand for metabolic enzyme , divided by the rate of biomass production ( see Section 1 . 3 in S1 Text ) ) . A cell’s growth rate is given by μ = vBM/cBM , where cBM is the biomass amount per cell volume and vBM is the biomass production rate ( biomass amount produced per cell volume and time ) . If cell biomass consisted only of metabolic enzymes ( more precisely , of enzymes considered in the cost Emet ) , the enzyme-specific biomass production rate rBM = vBM/Emet , where cBM would be equal to the cellular growth rate . Since this is not the case , we convert between Emet and cBM using the approximation Emet/cBM = fprot ( a − b μ ) , where fprot = 0 . 5 is the fraction of protein mass within the cell dry mass and the parameters a = 0 . 27 and b = 0 . 2 h were fitted to describe the metabolic enzyme fraction in proteomics data , assuming a linear dependence on growth rate [66] . As shown in the S1 Text ( Equations 8–9 and Figure 1 ) , we obtain the conversion formula μ = a f prot v BM E met + b f prot v BM . ( 1 ) Note that the biomass flux vR70 in our model is set to 1 mM s−1 by convention , and the kcat of this reaction was set to a sufficiently high value so that it would never become a bottleneck ( see Figure 5 in S1 Text ) . By simple unit conversion we obtain vBM = 7 . 45 × 107 mg l−1 h−1 . As shown above , the total enzyme mass concentration is given by Emet = ∑i wi Ei in units of mg l−1 , so it requires no further conversion . The final formula for growth rates , with proper units , reads μ = 1 . 01 × 10 7 mg l - 1 h - 1 ∑ i w i E i + 7 . 45 × 10 6 mg l - 1 . ( 2 ) It shows that maximizing the growth rate μ is equivalent to minimizing the enzyme cost Emet . The link between biomass production , total enzyme mass concentration , and growth rate can also be understood through the cell doubling time . We first define the enzyme doubling time τ met ≡ ln ( 2 ) r BM = ln ( 2 ) · E met v BM , the doubling time of a hypothetical cell consisting only of core metabolism enzymes . Since E . coli cells contain also other biomass components , the real doubling time is longer and depends on the fraction of these other components within the total biomass . Furthermore , this fraction decreases with the doubling time , as seen in experiments [67] and as expected from trade-offs between metabolic enzymes and ribosome investment [66] . This leads to a constant offset in the final cell doubling time formula: T = 7 . 4 · τ met + 0 . 51 [ h ] = = 6 . 9 × 10 - 8 h l mg - 1 · ∑ i w i E i + 0 . 51 [ h ] . ( 3 ) The calculation of sensitivities between enzyme parameters and growth rate is based on the following reasoning . If a parameter change slows down a reaction rate , this change can be compensated by increasing the enzyme level in the same reaction while keeping all metabolite levels and fluxes unchanged . For example , when a catalytic constant changes by a factor of 0 . 5 , the enzyme level needs to be increased by a factor of 2 . The cost increase is given by Δ cost = ( k cat , old k cat , new - 1 ) ⋅[old enzyme cost] . Also for other parameters , the local enzyme increase can be simply computed from the reaction’s rate law . Instead of adapting only one enzyme , the cell may save some costs by adjusting all enzyme and metabolite levels in a coordinated fashion . However , the extra cost advantage is only a second-order effect and can be neglected for small parameter variations . Hence , the first-order local and global cost sensitivities are completely identical ( proof in Section 4 . 2 in S1 Text ) . Sensitivities to external parameters ( e . g . extracellular glucose concentration ) can be computed similarly . The growth sensitivities for a given EFM are computed by multiplying the enzyme cost sensitivities by the derivative between growth rate and enzyme cost .
|
When cells compete for nutrients , those that grow faster and produce more offspring per time are favored by natural selection . In contrast , when cells need to maximize the cell number at a limited nutrient supply , fast growth does not matter and an efficient use of nutrients ( i . e . high biomass yield ) is essential . This raises a basic question about metabolism: can cells achieve high growth rates and yields simultaneously , or is there a conflict between the two goals ? Using a new modeling method called Enzymatic Flux Cost Minimization ( EFCM ) , we predict cellular growth rates and find that growth rate/yield trade-offs and the ensuing preference for enzyme-efficient or substrate-efficient metabolic pathways are not universal , but depend on growth conditions such as external glucose and oxygen concentrations .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"cell",
"physiology",
"carbohydrate",
"metabolism",
"medicine",
"and",
"health",
"sciences",
"chemical",
"compounds",
"enzymes",
"oxygen",
"enzymology",
"carbohydrates",
"glucose",
"metabolism",
"organic",
"compounds",
"glucose",
"cell",
"metabolism",
"physiological",
"processes",
"enzyme",
"metabolism",
"enzyme",
"kinetics",
"enzyme",
"chemistry",
"proteins",
"chemistry",
"biochemistry",
"organic",
"chemistry",
"cell",
"biology",
"physiology",
"monosaccharides",
"secretion",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"metabolism",
"chemical",
"elements"
] |
2018
|
Metabolic enzyme cost explains variable trade-offs between microbial growth rate and yield
|
Whole-genome regression methods are being increasingly used for the analysis and prediction of complex traits and diseases . In human genetics , these methods are commonly used for inferences about genetic parameters , such as the amount of genetic variance among individuals or the proportion of phenotypic variance that can be explained by regression on molecular markers . This is so even though some of the assumptions commonly adopted for data analysis are at odds with important quantitative genetic concepts . In this article we develop theory that leads to a precise definition of parameters arising in high dimensional genomic regressions; we focus on the so-called genomic heritability: the proportion of variance of a trait that can be explained ( in the population ) by a linear regression on a set of markers . We propose a definition of this parameter that is framed within the classical quantitative genetics theory and show that the genomic heritability and the trait heritability parameters are equal only when all causal variants are typed . Further , we discuss how the genomic variance and genomic heritability , defined as quantitative genetic parameters , relate to parameters of statistical models commonly used for inferences , and indicate potential inferential problems that are assessed further using simulations . When a large proportion of the markers used in the analysis are in LE with QTL the likelihood function can be misspecified . This can induce a sizable finite-sample bias and , possibly , lack of consistency of likelihood ( or Bayesian ) estimates . This situation can be encountered if the individuals in the sample are distantly related and linkage disequilibrium spans over short regions . This bias does not negate the use of whole-genome regression models as predictive machines; however , our results indicate that caution is needed when using marker-based regressions for inferences about population parameters such as the genomic heritability .
Whole-genome regression ( WGR ) methods [1] are becoming increasingly used for analysis and prediction of complex traits , quantitative or categorical . These methods were first developed for prediction in plant and animal breeding ( e . g . , [2 , 3] ) . More recently , there has been an increased interest in using WGR methods for inferring the proportion of variance that can be explained by a linear regression on a marker panel , or ‘genomic heritability’ [4–6] . Prediction and inference are two different problems , and a model that may yield good ( e . g . , unbiased and precise ) estimates of parameters of interest may have a relatively poor prediction performance , and vice versa . Most of the methodological research in WGR methods was developed in animal breeding with a focus on prediction . Unfortunately , little is known about the inferential properties of estimates derived from WGRs models . For example , it is unclear whether the commonly used likelihood-based ( or Bayesian ) estimators of variance components or of genomic heritability estimate population parameters consistently [7] . Before copious marker information became available , genetic analysis ( e . g . , estimation of heritability ) was mainly based on mixed effects linear models applied to family data [8] . In the so-called infinitesimal model , relatedness due to kinship is assessed using pedigrees , and a central element of model specification is the assumption that genotypic values result from the small and additive effects of alleles at a large number of loci . A number of studies have investigated the quality of fit of the infinitesimal model to experimental [9 , 10] and to simulated family data [11] . Most of these studies have concluded that the additive infinitesimal model is a useful abstraction , except in situations involving a few large-effect non-additive loci . Therefore , at least at some operational level , with family information the distinction between the model that generated the data and the one used for analysis has not seemed critical . The availability of genotype information on a large number of loci has made possible to assess kinship among nominally unrelated individuals [9–13] . In this setting , and due to imperfect linkage disequilibrium ( LD ) between markers and quantitative trait loci ( QTL ) , the patterns of allele sharing at markers and at causal loci may be very different [6] . Hence , the distinction between the data generating process and the model used for data analysis , or instrumental model , must be made clearly: in the instrumental model marker genotype information is used in lieu of the causal genotypes that are at the basis of the classical model of quantitative genetic theory . Thus , clarifying the link between the parameters of the instrumental model ( e . g . , the genomic or SNP variance ) and classical quantitative genetic parameters ( e . g . , the genetic variance ) is essential . Yang et al . ( 2010 ) [4] proposed using the G-BLUP method [2] , a particular class of WGR , applied to data involving distantly related individuals , for estimation of the proportion of variance accounted for by a multiple-linear regression on common SNP . The proportion of unexplained genetic variance can be interpreted as ‘missing heritability’ , which conceptually can be attributed to imperfect LD between markers and QTL . Using a WGR approach Yang et al . ( 2010 ) found that approximately half of the heritability of human height was captured by common SNP . Other studies , e . g . , [6] , have corroborated Yang’s results using both simulated and real data . More recently , WGRs have been used for estimation in scenarios where all causal variants are assumed to be included in the marker panel , and various suggestions have been made with the purpose of obtaining inferences of genomic heritability that resemble more closely those based on pedigrees [5] . In the literature on genomic analysis of complex traits published so far [4–6 , 14] , genetic parameters have been defined based on the statistical assumptions of the instrumental model used for data analysis . This is so despite the fact that there is a key difference in the way genotypes and effects are treated in statistical models and in quantitative genetics theory . In the latter , inter-individual differences in genetic values are attributed to subject-to-subject differences on allele content at QTL [15–17]; therefore genetic variance stems from variation at QTL genotypes . In this framework , at any given time in a population , the effects of alleles on a trait ( e . g . , the average effects of allele substitution ) are fixed quantities , e . g . , [16] pp . 112–113 . On the other hand , in the instrumental regression models , genotypes are treated as fixed and variation stems from uncertainty about marker effects ( the so called ‘variance of marker effects’ ) . This key difference in the treatment of genotypes and of their effects has important consequences that we further explore in this article . An important contribution of this paper is to establish theory aiming at a precise definition of parameters arising in regression models using genomic data ( markers , sequence ) as explanatory variables . Our approach is framed within the classical quantitative genetics paradigm . We discuss how these “instrumental model parameters” relate to “structural parameters” of an underlying conceptual QTL model . We also present stylized cases that shed light into the interpretation of the parameters of the instrumental model . Finally , we discuss potential estimation problems and provide a limited set of simulations where some statistical properties of likelihood-based estimates are assessed .
In standard quantitative genetic theory [15–17] additive genetic values are linear functions of allele content at QTL . Concepts such as the additive effect of an allele in a population or narrow sense ( trait ) heritability are defined with reference to this framework . The standard quantitative genetic model ( hereinafter referred to as QTL-model ) assumes that a trait of interest measured on individual i ( yi; i = 1 , … , n ) is affected by alleles at q QTL . Hereinafter , for ease of presentation , we assume that all loci are bi-allelic and that genotype codes ( zij; j = 1 , 2 , … , q ) and phenotypes have been centered , so that E ( yi ) = 0 and E ( zij ) = 0 for all individuals and loci . The genetic value of an individual is defined as the expected phenotypic value given QTL genotypes , gi = E ( yi|zi ) , where zi = {zi1 , … , ziq} is a vector of genotype codes observed at the ith individual at each of the q QTL . The conditional expectation function maps from genotypes to expected phenotypic value . The genetic variance of a trait in the conceptual population is simply the average ( over individuals ) squared deviation of the genetic values from the population mean . In our setting , because E ( yi ) = 0 , the genetic variance becomes σg2 = EzEy|z yi|zi2 = Ezgi2 . Clearly , genetic variation stems from inter-individual differences in allele content at QTL ( this is what confers individuals different genetic values ) . In the previous section we argued that the statistical parameter σu2 agrees with the population parameter σg2 only under highly simplified scenarios . Another important question is that of whether estimators derived from potentially misspecified likelihoods ( either in likelihood-based or Bayesian estimates ) can estimate parameters consistently , meaning that that the estimator converges to the true value of the parameter asymptotically . This will happen , for instance , if the bias and variance of the estimator go to zero as sample size goes to infinity , e . g . , [24] ) . We focus our study on the G-BLUP procedure due to its widespread use and relative simplicity . In this method , phenotypes are regressed on markers using a linear model of the form yi = ∑jxijbj+εi where bj~iidN0 , σb2 and εj~iidN0 , σε2 , with marker effects and residuals mutually independent . The model implies the following marginal distribution of phenotypes y~N0 , Gσu2+Iσε2 where N ( . , . ) stands for a multivariate normal distribution , G is a genomic-relationship matrix and σu2 is a variance parameter . Maximizing the likelihood ( σu2 and σε2 are the unknown parameters ) function associated with above-expression yields maximum likelihood estimates of variance components and of the proportion of variance explained by the model: hu2 = σu2σu2+σε2 . Maximization of the likelihood function is facilitated using the eigenvalue decomposition of the G matrix; further details about this are given in S3 Text . Since G is computed using markers that are not necessarily QTL or that are in imperfect LD with QTL , the ( co ) variance patterns of additive effects and , consequently , the likelihood function , can be misspecified [6] . This potentially leads to large finite-sample bias and to inconsistency of estimates , meaning that the genetic parameters may not be well estimated even in large samples . To assess some statistical properties of maximum likelihood estimators ( MLE ) we conducted two simulations . In both , phenotypes were generated according to the additive QTL model of ( Eq 1 ) , yi = α′zi + δi , with QTL effects sampled from a zero-mean normal distribution and from an independent normal distribution with IID residuals , that is δi~N ( 0 , σδ 2 ) . Variance parameters were chosen to generate a trait heritability of 0 . 5 . In our first simulation markers and QTL were generated according to a stylized LD pattern . Our second simulation uses real human genotypes .
Above , we discussed conceptual problems emerging when genetic parameters are defined based on assumptions of statistical models that are not in line with fundamental quantitative genetic concepts . A second problem arises when estimation of these parameters is based on markers and not on QTL-genotypes . Under regularity conditions , maximum likelihood estimates are asymptotically unbiased ( [24] ) . However , consistency cannot be guaranteed unless the likelihood is correctly specified and , even if the estimator happens to be consistent , misspecified likelihoods can induce sizable finite-sample bias . The proportion of allele sharing at any given set of loci can be viewed as a random variable with expected value given by twice the kinship coefficient between the individuals ( derived from the full pedigree and with additional assumptions such as absence of mutation and no selection ) and random variation given by the effects of Mendelian sampling [32] . If a large number of markers segregate independently from the QTL the proportion of allele sharing at markers and at QTL can be very different ( e . g . , [6] ) . This is particularly important for pairs of individuals sharing short chromosome segments ( e . g . , distantly related individuals from populations with large effective population size ) . Under these conditions a likelihood , constructed based on proportions of allele sharing at markers , can be largely misspecified and consistency may not hold . Of course , this does not necessarily imply that estimates would be inconsistent or have noticeable finite-sample biases; it just poses a caveat: one needs to be careful with use of estimates derived from models based on misspecified likelihoods , as it is most probably the case for marker-based model for whole-genome inference . There are at least two cases where a likelihood function based on markers will not be largely misspecified . The first one is when patterns of allele sharing at markers and at QTL are very similar; this can occur if markers are in tight LD with QTL , and also with data from close relatives . A second case is when the components of genetic values that cannot be explained by markers ( ξi in expression ( 4 ) ) are IID , in which case the likelihood will correctly represent the ( co ) variance structure of the data . In these two cases the parameters of the instrumental model ( e . g . , hu2 ) could be inferred consistently . Unfortunately , in general there is no good reason to believe that the ξi′s are IID and therefore the likelihood function will typically be misspecified . Goddard et al . [33] suggested a way of computing the marker-based genomic relationship matrix which , the authors argue , would have the property that the expected value of the proportion of allele sharing at QTL ( GQTL ) given the realized proportions of allele sharing at markers ( GMRK ) is GMRK , that is E[GQTL|GMRK] = GMRK . However , even if this property were to hold , this does not imply that estimates of variance components derived using GMRK would be unbiased . Indeed , in each sample GQTL can differ from GMRK; if the differences in the proportion of allele sharing at markers and at QTL are large enough , misspecification of the likelihood function due to the fact that we use GMRK instead of GQTL , can induce a systematic bias . Recently Jiang et al . [34] examined large sample properties of REML estimators of variance parameters of a marker-based regression ( σε2 , σu2 ) and concluded that , under certain conditions , the REML estimators can converge in probability to the true value of the statistical parameters , this being the case even if the likelihood is misspecified . The proof of this result is based on three key assumptions: ( a ) the model used to estimate parameters , what we here call the instrumental model , includes all QTL genotypes plus a number of markers with no effect , ( b ) all the covariates used in the model ( i . e . , both QTLs and markers ) are mutually independent , and ( c ) the number of QLT is not too small relative to the number of markers . In this specific setting the authors prove consistency of the REML estimator of the error variance and the convergence in probability of the REML estimator of the statistical parameter σu2 . This result , which the authors verified in simulations , indicates some robustness of the REML estimator . On the other hand , in our study we detected several cases where misspecified likelihoods produced sizable finite-sample bias , this occurring in settings where ML estimates derived from correctly specified likelihoods were seemingly unbiased . However , there are important differences between the study of Jiang et al . [34] and ours . Firstly , Jiang et al . [34] focused on estimation of the statistical parameters of the instrumental model ( σε2 , σu2 ) while here we addressed estimation of parameters defined from a quantitative genetics perspective ( σε2 , σg2 ) and , as we discussed here , σu2 and σg2 coincide only under very stylized conditions ( e . g . , complete LE between loci , a condition assumed by Jiang et al . [34] which we did not adopt here ) . Secondly , in an attempt to resemble what one encounters with modern genomic data , our simulations used a ratio between the number of QTL relative to the total number of loci that is relatively small ( 200/50K in simulation 1 , and 5/305 in simulation 2 ) . Importantly , Jiang et al . [34] indicate that the asymptotic results presented in their study involve approximations that may not hold when the ratio of number of QTL relative to the total number of loci is close to zero . Thirdly , our simulations incorporated LD ( either LD blocks in simulation 1 or LD patterns realized in real human genotypes in simulation 2 ) ; therefore , in our case loci were not in mutual LE as assumed in Jiang et al . [34] . Fourthly , in the simulation used in [34] the ratio between the total number of loci ( both marker and QTL ) and sample size was 10; in our simulations this ratio was 50 in simulation 1 and approximately 60 in simulation 2 . We adopted these settings because with modern genomic data the ratio between the number of markers and sample size is expected to be large . In the light of the evidence presented here it seems clear that further studies are needed to characterize the finite sample properties of REML estimators derived using misspecified likelihoods in settings that resemble the conditions encountered in the analysis of real genomes . Simulation studies using real human genotypes , including the one presented here , indicate that estimates of hu2 based on the G-BLUP model incorporating both markers and all QTL genotypes are lower than the trait heritability of the trait ( h2 ) . This reinforces the idea that GBLUP may not lead to unbiased estimates of the genomic heritability ( hg2 ) because when all QTL are included in the marker panel hg2 = h2 . Speed et al . [5] argued that a main reason for incorrect estimation is that LD is ignored in the computation of the G matrix . These authors suggested an alternative method for computing G that resulted in estimates of genomic heritability closer to the simulated trait heritability . However , in a follow up discussion [35 , 36] presented alternative simulations scenarios where the method of Speed et al . [5] yielded biased estimates . All in all , this suggests that the appropriate choice of method used for computing G may depend on the genetic architecture of the trait , a feature that is typically unknown , even if attention is restricted to additive gene action only . From our perspective , the main problem does not reside in the manner the G matrix is computed but , rather , in the use of massive numbers of markers that are in LE with QTL . The use of such data increases the sampling variance of estimates and may cause large finite-sample bias and lack of consistency . We have argued that the instrumental parameter hu2 may not provide a good representation of the genomic heritability ( hg2 ) ; an important reason for the difference between these two parameters is that while LD plays a role in the determination of σg2 associations among marker genotypes are ignored in the definition of σu2 . An alternative approach that accounts for multi-locus LD , could be based on the sample-variance of the true genomic values [37] that is σ~g2 = n-1∑i = 1n xi'β-x¯i'β2 where x¯i represents the average genotype . Of course , marker effects are unknown; however as suggested by Sorensen et al . [37] one could infer σ~g2 in a Bayesian setting by evaluating σ~g2 using samples from the posterior distribution of marker effects . The parameter σ~g2 accounts for LD between markers in a very specific manner: σ~g2 becomes equal to the genomic variance parameter , σg2 = β'Σxβ ( see expression 7 ) , with Σx replaced by its method-of-moments estimator X′ Xn-1 . Indeed , when genotypes are centered , σ~g2 = n-1∑i = 1n xi'β2 = n-1∑i = 1nβxi xi'β = n-1β'∑i = 1nxi xi' β = β'n-1X'Xβ . Complex traits are possibly affected by large numbers of small-effect QTL and the analysis of such traits requires fitting a large number of variants concurrently using a WGR approach such as the one proposed by [1] . Close relatives share long chromosome segments and , under these circumstances , the patterns of allele sharing at markers and at QTL are very similar . This leads to high prediction accuracy and very small bias in genomic heritability estimates . When markers and QTL co-segregate , variable selection does not seem to be needed [6] . On the other hand with distantly related individuals , the addition of large numbers of markers that are in LE with QTL can lead to incorrect specification of genomic relationships and this can result in potential inconsistencies of estimates of genomic heritability . We do not question the use of WGR for analysis of complex traits and as a prediction machine . Rather , we warn about problems arising when these methods are used for inferences . In our opinion this problem has been overlooked and oversimplified , and further research is needed to understand if and under what circumstances WGRs such as the G-BLUP can be used to correctly assess the true proportion of variance that can be explained by a regression on markers in the population .
|
Whole-genome regression ( WGR ) methods are being increasingly used for inferring the proportion of variance that can be explained by a linear regression on a massive number of markers , called ‘genomic heritability . ’ However , the statistical assumptions involved in WGRs are somewhat at odds with important quantitative genetics concepts . We argue and show that the parameters of the statistical model used for data analysis typically bear a tenuous relationship with the quantitative genetic parameters of interest . We also study , using simulations , the extent of bias of likelihood-based estimates . We conclude that under certain circumstances estimates can have a sizable finite-sample bias; therefore , caution needs to be exercised when interpreting parameter estimates derived from WGR models .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Discussion"
] |
[] |
2015
|
Genomic Heritability: What Is It?
|
In the Old World , sandfly species of the genus Phlebotomus are known vectors of Leishmania , Bartonella and several viruses . Recent sandfly catches and autochthonous cases of leishmaniasis hint on spreading tendencies of the vectors towards Central Europe . However , studies addressing potential future distribution of sandflies in the light of a changing European climate are missing . Here , we modelled bioclimatic envelopes using MaxEnt for five species with proven or assumed vector competence for Leishmania infantum , which are either predominantly located in ( south- ) western ( Phlebotomus ariasi , P . mascittii and P . perniciosus ) or south-eastern Europe ( P . neglectus and P . perfiliewi ) . The determined bioclimatic envelopes were transferred to two climate change scenarios ( A1B and B1 ) for Central Europe ( Austria , Germany and Switzerland ) using data of the regional climate model COSMO-CLM . We detected the most likely way of natural dispersal ( “least-cost path” ) for each species and hence determined the accessibility of potential future climatically suitable habitats by integrating landscape features , projected changes in climatic suitability and wind speed . Results indicate that the Central European climate will become increasingly suitable especially for those vector species with a current south-western focus of distribution . In general , the highest suitability of Central Europe is projected for all species in the second half of the 21st century , except for P . perfiliewi . Nevertheless , we show that sandflies will hardly be able to occupy their climatically suitable habitats entirely , due to their limited natural dispersal ability . A northward spread of species with south-eastern focus of distribution may be constrained but not completely avoided by the Alps . Our results can be used to install specific monitoring systems to the projected risk zones of potential sandfly establishment . This is urgently needed for adaptation and coping strategies against the emerging spread of sandfly-borne diseases .
Globally , the number of vector-borne infections in humans and animals increases rapidly , meanwhile causing almost one third of all cases of emerging infectious diseases [1] . In the Old World , sandfly species of the genus Phlebotomus serve as vectors for sandfly-borne pathogens such as Leishmania , Bartonella and several viruses ( e . g . Phlebovirus , Vesiculovirus and Orbivirus ) [2]–[4] . Sandfly-borne diseases and in particular visceral leishmaniasis are a main public health concern [5] , which demands more attention in science and policy [6] . While the spatial distribution of leishmaniasis seems to expand in southern parts of Europe [7] , [8] , first cases of autochthonous origin are recently reported from Central Europe [9]–[11] , where this disease was not endemic in the past . The presence of sandflies as vectors is mainly regulated by the species' climatic requirements on temperature and humidity or soil moisture , respectively [3] , [12]–[15] . Temperature and humidity are also the main factors impacting the altitudinal structure of sandfly occurrences [16] . It is known that sandflies react very sensitive to wind speed and prefer breeding sites sheltered from wind [17]–[20] . Beyond that , high wind speed decreases or even excludes flight activity [17] , [21] . For the purpose of inferring geographic distribution for sandflies , the advantages of ecological niche models have been demonstrated on the example of Lutzomyia species ( Lutzomyia spp . ) in the New World [22] . For the first time , Peterson and Shaw [23] integrated climate change scenarios in order to project future distribution of Lutzomyia spp . in Brazil . Recently , range expansions for sandflies of the genus Lutzomyia have also been projected for North America in the face of climate change [24] . For Europe , surprisingly , only few studies estimated the risk of potential range expansions of sandflies in the face of climate change ( e . g . [25] , [26] ) . The need for such studies is supported by the first sandflies catches in Central Europe . P . mascittii has been caught in Austria on the frontier to Slovenia [27] . Furthermore , P . mascittii is reported from the “Upper Rhine Valley” in the outermost southwest of Germany near the French border [28] . P . perniciosus seems to be established in the German state of “Rhineland-Palatinate” [29] . These findings may either indicate spreading tendencies from Mediterranean regions or range expansion from small Central European refugial areas , which may have already been occupied by the species during the Holocene climate optimum about 6 , 500 years ago [30] . Possibly , sandflies has occupied more areas in the past than it was noticed . For Austria , establishment of sandflies in formerly non-endemic areas can be expected already by moderately increasing temperatures in the 21st century [25] . Recently , Fischer et al . [26] estimated potential temperature-derived establishment of sandflies in Germany by transferring the required temperature during their activity phase and annual mean temperature for persistence to the expected future climate conditions in Germany using data of a spatio-temporal highly resolved regional climate model . But up to now , projections of the current and climate-driven potential future distribution of Phlebotomus spp . which additionally consider species-specific dispersal ability are missing . As climate is expected to change rapidly in the 21st century , sandflies are forced to react promptly . Here , we close this gap of knowledge and hypothesise:
It is well known that species dispersal ability is dependent on the environment and varies strongly with landscape structure [53] . In order to make projections more realistic according to spatial characteristics , we used a least-cost path analysis based on graph theory [54]–[56] to determine the most likely way for Phlebotomus spp . to move across a spatio-temporally changing landscape . The path function indicates the least efforts ( “costs” ) for a species in moving through any particular cell in the respective landscape [57]–[59] . Least-cost path analyses are frequently used to determine potential dispersal pathways for mammals [60]–[63] but have also been applied to insects [58] .
AUC-values yielded in high scores for five species ( Table 1 ) . Binominal tests indicated that tests points are predicted better by the model than a random prediction with the same factional predicted area at the significance level p<0 . 01 . Similarity between current and projected climate is analysed by MESS-analysis . Highest similarity in projections is indicated for southern parts of Germany , the northernmost regions of Switzerland as well as for eastern and north-eastern parts of Austria . Instead , lowest similarity exists for alpine regions and northern Germany . However , our projections seem not to be biased by non-analogue climate . In general , climatic suitability can be expected to increase for all species in the 21st century ( Figure 3 and S1 , Table 3 and S1 ) . This is in accordance with the first part of our first hypothesis assuming increasing climatic suitability for the species in Central Europe . Nevertheless , we cannot completely verify the second part of our first hypothesis of more favourable conditions in the ( south- ) westernmost parts of Central Europe for species with current ( south- ) western focus of European distribution and the opposite for species which are currently distributed in ( south- ) eastern parts of Europe . Overall , projections based on the A1B scenario ( Table 3 ) represent higher suitability for species in comparison to projection of the B1 scenario ( Table S1 ) . Nevertheless , the spatial patterns of potential climatically suitable habitats remain to be the same for both scenarios . The detailed annotation of climatic suitability in the following refers to the A1B scenario . In general , projections hint on spreading tendencies for all studied Phlebotomus spp . to areas where they have not occurred so far in both scenarios ( Figure 4 and S2 ) . Nevertheless , sandfly species will not be able to become established in all climatically suitable areas of Central Europe according to the limited natural dispersal ability . The detected dispersal pathways show some differences between the two applied scenarios , in contrast to the modelled climatic suitability , where just temporal but no spatial variations are pointed out . The first part of our second hypothesis that species with current ( south- ) western focus of distribution are likely to disperse eastwards can be affirmed only for P . perniciosus but not for P . ariasi and for P . mascittii . We cannot confirm the part of the second hypothesis that the Alps will prohibit completely a northward spread for the species which are currently distributed in ( south- ) eastern European regions . However , it is very likely that the Alps will decelerate the range expansion .
Our aim was to determine future occurrences of five Phlebotomus spp . with spreading tendencies in the face of a changing climate . These sandflies serve as proven or assumed ( P . mascittii ) vectors of Leishmania infantum causing the leishmaniasis . Knowledge concerning the potential future presence of disease vectors is a first step towards an accurate and efficient risk assessment of vector-borne diseases [72] . Conventional static bioclimatic niche modelling can be extended by novel avenues for instance regarding species-specific abilities to disperse [73] . Therefore , we integrated species-specific dispersal pathways to the detected climatically suitable habitats . Within this study , we focus on active and natural dispersal of the species and excluded potential human assistance for range expansions , for instance via the transport of subtropical plants containing eggs or larvae in the moist substrate . However , these effects are not clearly understood and hence not included in this analysis . The results of this study represent the minimum range expansion of Phlebotomus spp . that is only related to active and natural movement in a changing environment without potential human-assistance . Our results suggest that the development of Central European climate will increasingly support suitable habitats for phlebotomine sandflies . This general trend will become even more pronounced in the second half of the 21st century . We project sandfly establishment in formerly non-endemic areas . This will additionally increase the risk of emerging sandfly-borne diseases in Central Europe such as leishmaniasis . Nevertheless , it is unlikely that sandflies will reach and occupy the provided climatically suitable habitats entirely . During the upcoming years , a spatial focus of surveillance regarding potential new-establishment of Phlebotomus spp . with current south- ( western ) focus of distribution should be directed for western parts of the German state “North Rhine-Westphalia” . These species may additionally occupy the regions around the “Lake Constance” ( southern part of Germany and northern part of Switzerland ) . Furthermore , Switzerland must be aware on potential sandfly occurrences around the river valley of the “Aare” and the “Lac dé Neuchâtel’ already within some years . In Austria , especially the south-eastern states “Carinthia’ , “Styria’ and “Burgenland’ seem to be at risk by new-infestations due to the spreading tendencies of P . mascittii and P . ariasi . In the case of sandfly species with current south ( -eastern ) focus of distribution ( P . neglectus and P . perfiliewi ) , the canton “Ticino” in southern Switzerland must be alert . According to our results , especially the regions around “Lago Maggiore” and “Lago di Lugano” should be monitored systematically . Furthermore , the Austrian regions neighbouring Slovenia and Hungary should be prepared for the establishment of these sandfly species in the near future . The main limitation in our projections refers to the accuracy of georeferencing maps of sandfly distribution . We have chosen an algorithm that is capable to cope with this source of uncertainty [48] . Expectedly , improved model projections would arise with geographically more accurate point data . However , the intension of this paper is rather to provide a methodological approach . Recently , it has been pointed out that pixel values of predictor variables in close proximity will be highly correlated , which would reduce the effect of inaccuracy in spatial data set of species occurrences [74] . Furthermore , when comparing the reports for cases of autochthonous leishmaniasis in regions that were considered as being non-endemic ( e . g . [9] ) with documented presence records of sandflies leads to the assumption that sandflies may be wider distributed than realised . However , as it is unknown which species acted in such regions as vectors , only documented presence records at the species level of the Phlebotomus spp . can reasonably be integrated . Assuming climate is generally considered to be suitable for the permanent establishment of populations , the presence of Phlebotomus spp . is additionally dependent on land cover e . g . forest , agriculture and urban areas [75] , [76] . In this study , we integrated altitudinal structures such as river valleys and mountain ridges in least-cost path analysis . In order to recalculate wind speed we additionally integrated land cover data as surface roughness to decelerate near-surface wind speed . Climate change may contribute to alterations in land use and cover , due to warming , changes of precipitation regimes and increases of climatic extreme events such as droughts or floods . This is likely to affect the spatial structure of agricultural systems . In addition to direct climatic impacts , these changes of land cover will additionally affect the spread and distribution of sandflies . However , land cover and land use changes depend on complex processes of decision making under specific political and economical conditions [77] and are hence difficult to project . Therefore land use and cover were considered to remain constant in this study . In general , biotic interactions such as predation or competition are crucial for species distribution [78] . The modifications of the ecological links or networks of an organism by climate change can substantially alter the realised niche of species population [79] . In Germany , P . perniciosus and P . mascittii do not co-exist at the same locations [29] . This can be a result of diverging invasion pathways or of competitive exclusion in the respective regions . Unfortunately , knowledge on biotic interactions of sandflies is scarce . Furthermore , one has to bear in mind that presence of phlebotomine sandflies is dependent on humans and their social factors , for instance living conditions [80] , [81] . However , all these factors become more important for species distribution on smaller spatial scales than applied in this study [82] , [83] . Concerning the least-cost analysis for species movement it has to be noted that the attributed costs are based on assumptions and/or preliminary observations and hence may not include all of the relevant factors [55] , [61] . For instance , it is questionable whether humans assist in the spread of sandflies . Nevertheless , in comparison to mosquitoes , direct human effects on dispersal are of minor importance . Furthermore , the species movement behaviour must not necessarily be optimal or well adapted in human-modified landscapes [53] . Especially dispersal behaviour of individuals between populations may differ from the general tendency at the metapopulation level [58] . Besides the effects of changes in long-term climatic conditions used in this study , extreme weather events are expected to increase in Europe [84] . This will influence organisms and ecosystems remarkably [85] , [86] . It has been shown that climatic variability in general [87] and extreme weather events such as floods particularly affect sandfly occurrences [88] . In order to integrate weather extremes in a satisfactory quality within climatic projections , a further downscaling of their spatial resolution to the local scale is required [89] . This is the only way to account for the contribution of weather extremes on disease vectors in risk analysis . Bioclimatic envelope models are powerful tools to envisage potential responses in species distribution to climate change from regional to global scales [82] , [90] . They can be seen as a useful first filter for approximations of the impact of climate change on the species distribution [82] , [91] . A well-adapted modelling approach is required to project climate change effects on species [44] . Therefore , we selected MaxEnt as algorithm , due to better performances in comparison with further presence-only and ( pseudo- ) presence-only algorithms ( see chapter Model runs for details ) . Results yielded in high model quality criteria , emphasised by threshold-dependent and independent criteria for Phlebotomus spp . In order to cope with the general uncertainty in species distribution modelling regarding the climatic evolvement [78] , we projected the climatic suitability based on two IPCC [37] scenarios ( A1B and B1 ) that best illustrate the respective storyline . We choose bioclimatic variables that are considered to be biologically meaningful variables for model input . Our projections of future climatic suitability refer to data of the European regional climate model CCLM , which is nested into the well-established global climate model ECHAM5 [35] . In comparison to their driving global models , regional patterns of climate change are projected more precisely , which enhances the quality of climate impact studies [36] . For instance , global climate models fail particularly in replication of observed wind speeds . Obviously , projections of changes in wind speed profit from downscaling to the regional level [92] . Furthermore , potential regions with non-analogue climatic conditions and where hence projections are inappropriate were excluded via MESS-analysis . Evidently , the consideration of the dispersal capacity of insects in a changing climate improves the quality of projections of species distribution [93] . Hence , we combined projected climatic suitability in the 21st century with dispersal ability for five Phlebotomus spp . We practiced least-cost analysis for future movement patterns by including temporarily stable ( elevation , landscape features ) and variable factors ( wind speed and development of climatic suitability ) . This allows integrating and combining expert knowledge on sandfly ecology and biology with statistical methods . In doing so , potential species-specific dispersal pathways can be pointed out . This offers the opportunity to distinguish between climatically suitable habitats that can be reached by invasive species and those that are not accessible to them . The proposed method for the detection of dispersal pathways can be applied to other invasive and mobile disease vectors in the face of climate change . For this purpose , however , species-specific cost surfaces have to be generated . The Alps , for instance , may not be seen as crucial natural barrier for the mainly human-assisted spread of invasive mosquitoes such as the Asian tiger mosquito ( Aedes albopictus ) . However , they are an efficient natural barrier for other species . Here , we provide a powerful methodological approach to extend conventional bioclimatic envelope modelling of disease vectors by species specific dispersal ability . Our findings promise more realistic projections concerning the vector species future distributions . We identify those Central European regions that are especially exposed to the emerging threat of disease vectors in the light climate change . For the modelling of hitherto neglected vector-connected risks , expertise from various scientific disciplines has been taken into account . Proactive monitoring activities and development of feasible adaptation strategies are required before the establishment of disease vectors may take place . Hence , those analyses help to focus control programmes on specific areas at risk on a regional scale . Due to the identification of spatial hot-spots for such activities , cost- and time-efficient surveillance strategies can be developed . This enables target-orientated counteractions directed against the suggested spread of disease vectors in time . Consequently , the risk of disease transmission in formerly non-endemic areas can be reduced . Once disease vectors such as sandflies are established , vector control and disease prevention have proven to be limited .
|
Growing evidence exists on the emergence of sandfly-borne diseases in the light of climate change . Determining the principle responses of phlebotomine sandflies to climatic changes supports our understanding of future regions that will be threatened by new-establishments of this important group of disease vectors . The aim of this paper is to combine projected climatic suitability for five Phlebotomus species in Central Europe ( Austria , Germany and Switzerland ) for different time-periods during the 21st century with their potential spreading capacity to disperse to climatically suitable areas . We indicate that the Central European climate will develop toward the preferred bioclimatic niche of the species , especially from mid-century onwards . Nevertheless , we also elucidate within this study that sandflies will hardly be able to occupy the whole areas that will provide suitable climatic conditions due to their limited natural dispersal ability . Our approach provides a framework to combine statistical modelling techniques with expert knowledge on species ecology . Indications of future occurrences of disease vectors may help to initiate surveillance systems in specific regions at an early stage of risk exposure . Hence , the threat of the climate-driven spatial extension of disease vectors and consequently of potentially emerging vector-borne diseases can be counteracted .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"geography",
"public",
"health",
"and",
"epidemiology",
"mathematics",
"environmental",
"sciences",
"ecology",
"earth",
"sciences",
"statistics",
"biology",
"public",
"health",
"physical",
"geography",
"veterinary",
"science",
"veterinary",
"medicine"
] |
2011
|
Combining Climatic Projections and Dispersal Ability: A Method for Estimating the Responses of Sandfly Vector Species to Climate Change
|
Hundreds of genes show aberrant DNA hypermethylation in cancer , yet little is known about the causes of this hypermethylation . We identified RIL as a frequent methylation target in cancer . In search for factors that influence RIL hypermethylation , we found a 12-bp polymorphic sequence around its transcription start site that creates a long allele . Pyrosequencing of homozygous tumors revealed a 2 . 1-fold higher methylation for the short alleles ( P<0 . 001 ) . Bisulfite sequencing of cancers heterozygous for RIL showed that the short alleles are 3 . 1-fold more methylated than the long ( P<0 . 001 ) . The comparison of expression levels between unmethylated long and short EBV-transformed cell lines showed no difference in expression in vivo . Electrophorectic mobility shift assay showed that the inserted region of the long allele binds Sp1 and Sp3 transcription factors , a binding that is absent in the short allele . Transient transfection of RIL allele-specific transgenes showed no effects of the additional Sp1 site on transcription early on . However , stable transfection of methylation-seeded constructs showed gradually decreasing transcription levels from the short allele with eventual spreading of de novo methylation . In contrast , the long allele showed stable levels of expression over time as measured by luciferase and ∼2–3-fold lower levels of methylation by bisulfite sequencing ( P<0 . 001 ) , suggesting that the polymorphic Sp1 site protects against time-dependent silencing . Our finding demonstrates that , in some genes , hypermethylation in cancer is dictated by protein-DNA interactions at the promoters and provides a novel mechanism by which genetic polymorphisms can influence an epigenetic state .
Causes of promoter DNA hypermethylation in cancer are unknown . Possibilities range from random events selected for to a model whereby an initial drop of transcription rate allows elimination of active chromatin boundaries and spreading of DNA methylation from “DNA methylation centers”- perhaps , repeat elements [1] . Alternatively , aberrant methylation might be caused by a repressor binding to the promoter and altering chromatin state to a closed configuration , which eventually causes abnormal methylation via recruitment of DNA methyltransferase [2] , [3] . Furthermore , methylation seeding and ongoing transcription as well as possibly transcription factor binding seem to be required for the protection of promoters against methylation [4] . Nevertheless , the details and order of events in these processes remain elusive . RIL is a ubiquitously expressed gene , which was originally identified as a candidate tumor suppressor [5] . Human RIL maps to chromosome 5q31 . 1 , a region frequently deleted in the malignant cells of patients with myelodysplastic syndrome ( MDS ) and acute myelogenous leukemia ( AML ) [6] , and appears to be a good candidate for one of the tumor-suppressor genes that reside in this area . Using methylated CpG island amplification [7] in the K562 cell line , we identified RIL as a novel gene aberrantly methylated in cancer . RIL CpG island is unmethylated in normal tissues , and is one of the most frequent targets for hypermethylation in various cancer cell lines and primary tumors . Hypermethylation of RIL correlates with loss of gene expression , which could be restored in methylated cell lines by 5-aza-dC . Moreover , RIL reexpression leads to a suppression of tumor cell growth and clonogenicity in soft agar as well as sensitization of cells to apoptosis [8] . Here , we describe a polymorphism in the RIL promoter that creates an Sp1/Sp3 binding site and protects against methylation in cancer .
In search for causes of aberrant hypermethylation of RIL in cancer , a ∼500-bp region around the transcription start site of the RIL promoter was sequenced , and a previously unknown polymorphic insertion was identified in the gene . In the population , 2 alleles of RIL exist , long and short , with a polymorphic area near a CGG repeat sequence adjacent to the transcription start site . The long allele is created by insertion of a 12-bp fragment ( CGGCGGCGGCTC ) and a substitution of T to G 3 bases upstream of the insertion site in the short allele ( Figure 1A ) . In silico analysis identified additional putative binding sites for 3 transcription factors which appear only in the long allele . To more fully characterize this phenomenon and determine the frequency of alleles , a total of 326 normal samples were sequenced , which included 227 normal colon and 99 normal blood samples . In the normal population , 45% of people are homozygous for the short allele , 44% heterozygous and 11% homozygous for the long RIL allele . We have also determined the frequency of the RIL alleles among 240 cancer specimen which included 113 colon cancers , 100 MDS and 27 AML samples . We found that 44% of those were homozygous for the short allele , 40% heterozygous and 16% homozygous for the long RIL allele . Differential methylation of the two RIL alleles was initially discovered after development of a COBRA assay . We initially noticed that methylated heterozygous tumors analyzed by COBRA seem to have preferential methylation of the short allele . After digestion of the COBRA products obtained from several homozygous and heterozygous tumors , we observed that among the 2 allele-specific bands , the predominant one is the lower band , which represents methylated short allele molecules ( Figure S1 ) . This led us to a suspicion that the short allele is more methylated than the long . For further studies , we decided to use pyrosequencing method , which is more quantitative in detecting methylation . To make sure that no amplification bias exists in detecting methylation , we designed all the assays primers either upstream or downstream of the insertion . We then tested the pyrosequencing assay A ( Figure 1A ) by studying mixed normal unmethylated with methylated DNA at known ratios , and found no bias in amplifying unmethylated vs . methylated DNA ( Figure S2 ) . Next , we studied methylation in 48 primary cancer samples ( 18 homozygous long and 30 homozygous short ) which included 44 colon cancer and 4 AML cases with this assay , and found that homozygous short tumors showed mean 51 . 7%+/−3 . 8% methylation density vs . mean 24 . 7%+/−4 . 5% methylation for homozygous RIL long tumors , and thus had a 2 . 1 fold increased methylation density ( P<0 . 001 ) ( Figure 1B ) . We also studied 34 MDS samples ( which included 11 homozygous long and 23 homozygous short ) using pyrosequencing assay B ( see Figure 1B ) and found similar results: homozygous short RIL samples showed mean 35 . 5%+/−5 . 4% methylation density vs . mean 14 . 7%+/−2 . 4% methylation for homozygous RIL long samples , P = 0 . 001 ) ( a 2 . 4 fold difference ) . To confirm the latter observations , we performed bisulfite sequencing of 10 methylated tumors with heterozygous alleles ( 6 colon cancer and 4 AML ) . These data showed that in 8 out of 10 samples studied , the short allele is the primary methylation target , whereas the long one is largely unmethylated ( Figure 1C ) . In the case of AML7 , both alleles were methylated , and in case C113 there was a mixture of unmethylated and methylated alleles with no difference in methylation . Furthermore , in C140 and AML5 many of the S alleles remain unmethylated; however this finding may well represent contamination from normal tissues in which both L and S alleles are largely unmethylated , or co-existence of unmethylated and methylated tumor cells . However , even in samples C140 and AML5 , the number of methylated clones is higher for the S allele . Using methylation of >30% of CpG sites within the clone as an arbitrarily chosen cutoff for dense methylation , the short alleles were 3 . 1 times more methylated than the long alleles ( p<0 . 001 by the Fisher exact test ) . Quantitatively similar results are obtained if the cutoff is 20% , 50% , or without using any cutoff ( not shown ) . Taken together , these data demonstrate unequivocally preferential methylation of the short alleles even when they coexist with the long alleles within the same tumor . We previously found that RIL methylation increases with age in normal colon [8] . To investigate whether this is influenced by the polymorphism , we studied 46 normal colon samples ( 17 homozygous short , 15 heterozygous and 14 homozygous long , obtained from individuals with comparable age ) using 5 pyrosequencing methylation assays , each covering several CpG sites around the transcription start site , as well as exon 1 and intron 1 regions ( Figure 1A , Table S1 ) . We found that homozygous long samples were significantly less methylated compared to homozygous short in all the regions studied , both upstream and downstream of transcription start site ( Figure 1D ) . Average methylation percent for the upstream region measured by assays A–F was 34 . 3±1 . 8 for homozygous short cases , 26 . 8±2 . 2 for heterozygous , and 19 . 4±1 . 1 for homozygous long cases and therefore the short alleles were 1 . 8 times more methylated in the normal colon . Finally , all samples showed increased methylation with age , independent of allele status ( R = 0 . 42 , p = 0 . 09 for homozygous short , R = 0 . 57 , p = 0 . 03 for heterozygous , R = 0 . 69 , p = 0 . 01 for homozygous long cases ( Figure 1D ) . We were puzzled by the finding of differential allele methylation and thought that one likely explanation for such a difference is lower expression of the short alleles predisposing them to methylation . We therefore investigated whether the expression levels between the alleles are different in vivo . First , we obtained and genotyped 96 primary EBV-transformed lymphoblastoid cell lines established from normal individuals and found that 42 of those were homozygous short , 42 heterozygous , and 12 homozygous long . We then analyzed 24 cell lines ( 12 long and 12 short ) for methylation status of RIL by pyrosequencing . We found variable levels of methylation among the cell lines , which was higher for short cell lines , although the difference was not statistically significant ( Figure 2A ) . Nevertheless , this difference was quantitatively similar to what we saw in normal ( non-malignant ) colon tissues . We studied RIL expression in these samples by quantitative PCR and found it to correlate well with methylation ( Figure 2B ) . We then excluded 11 cell lines with methylation above 15% and compared 8 unmethylated homozygous long with 5 unmethylated homozygous short cell lines and found no difference in expression ( Figure 2C ) . This led us to conclude that there is no difference in expression between the long and short alleles in a physiologic setting . We next hypothesized that the insertion polymorphism creates binding sites for a protein that protects the long allele against methylation . To determine whether nuclear proteins can bind to the polymorphic region of the long or short alleles , we first performed electrophoretic mobility shift assays ( EMSA ) with nuclear extracts from different cell lines using double-stranded radiolabeled oligonucleotides that corresponded to bp −69 to −37 ( L ) or −57 to −37 ( S ) relative to the translation start site ( see Figure 3A ) of the long and short RIL alleles , respectively . We found strong DNA-binding activity for the L probe in 3/4 cell lines tested ( PC3 , MDA-MB-231 , OVCAR5 ) , whereas no detectable DNA binding activity was observed with the S probe using the same nuclear extracts , even at longer exposures ( Figure 3B ) . Three major shifted bands were found as a result of this L probe DNA-binding activity . Specificity of those bands was confirmed by competition with unlabeled oligonucleotides ( not shown ) . To map the region responsible for binding , we synthesized shorter versions of the L probe of an equal size to the original S probe: L1 , corresponding to nucleotides −69 to −49; L2 , nucleotides −57 to −37; and L3 , nucleotides −63 to −43 ( Figure 3A ) . EMSA results with the nuclear extracts from MDA-MB-231 cells indicated that , while probe L2 had a strong DNA-binding activity resulting in appearance of 3 distinct bands , probes L1 and L3 had no detectable DNA-binding activity ( Figure 3C , first three lanes ) . To identify the nature of proteins bound to the long allele , we performed supershift assays using different antibodies . When the L2 probe was mixed with nuclear extracts from MDA-MB-231 , and then incubated with an anti-Sp1 antibody , the top band was diminished , and supershifted complexes were formed , indicating that the top band contained Sp1 ( Figure 3C ) . Incubation of the L2 probe with the nuclear extracts in the presence of Sp3 antibody completely abolished the two lower bands and resulted in supershifted bands ( Figure 3C ) . Incubation of the L probe with the nuclear extracts and both Sp1 , Sp3 antibodies resulted in almost complete disappearance of all three bands , whereas an Egr-1 antibody had no effect on the shifts . We confirmed these results using recombinant human Sp1 ( data not shown ) . The results of EMSAs strongly suggested that binding of Sp1 and Sp3 proteins occurs in the 3CTC-containing region of the inserted region of the long allele . First , the L1 and L3 probes were unable to show any DNA-binding activity , and both do not have an intact 3CTC region . Second , the S probe which was also not bound by factors in the nuclear extracts , contains a 2CTC repeat , and in fact has only a 2-nucleotide difference with the L2 probe , thus being a naturally occurring mutant of L2 ( Figure 3A ) . Finally , the 3CTC sequence has been previously shown to be a variant Sp1 site , important for regulation of several genes , including WT1 [9] . To map Sp1/Sp3 binding in this region , we generated a series of oligonucleotides with 3-bp mutations throughout the probe ( LM1-LM6 ) ; and oligonucleotides with 2-bp mutations restricted to the 3CTC region ( LM7-LM10 ) ( see Figure 3A ) . We then competed the L2 probe with a 20-fold excess of various unlabeled competitors . These experiments revealed poor competition capacity of all the probes where the 3CTC repeat is absent ( S , Egr1 , LM3-LM5 , LM7-LM10 probes ) , and efficient competition by all probes containing an intact 3CTC sequence ( Figure 3D ) . Identical results were obtained in OVCAR-5 cells ( data not shown ) . Next , we sought to determine whether the additional Sp1/Sp3 binding site in the long allele affects transcription in vitro . We generated three different luciferase constructs driven by allele-specific RIL promoter fragments containing the polymorphic region of −588 to +19 ( A ) , −217 to +19 ( B ) and −588 to +516 ( C ) . In a series of transient transfections using 4 cell lines for constructs A and B and 2 cell lines for construct C , no significant differences in luciferase activity between the allele-specific constructs were observed ( Figure 4A ) , suggesting that the additional Sp1 site had no substantial effect on RIL transcription . This was consistent with the earlier experiments in lymphocytes . To determine whether the additional Sp1/Sp3 binding site confers protection against DNA methylation in vitro , we used the same constructs described earlier . We initially stably co-transfected unmethylated RIL allele-specific constructs ( construct B ) with neomycin-containing plasmids into the mammalian cell lines RKO and NIH3T3 , and found that those maintained expression over 3 months , and were resistant to de novo methylation ( data not shown ) , as previously reported for multiple genes [4] . Furthermore , it has been previously shown using GSTP1 gene as a model that aberrant methylation at a given promoter requires pre-existing “seeds” of methylation which trigger silencing [4] . To trigger silencing in our system , we used methylation “seeding” of allele-specific transgenes using HpaII methylase , which methylates 9 . 3% of the total CpG sites in construct B ( 13 . 5% of CpGs of the RIL promoter fragment ) , and co-transfected those with neomycin-containing plasmids into NIH3T3 cells . The plasmid methylation status prior to transfection was validated by HpaII methylation sensitive restriction enzyme digestion ( Figure S3 ) . After neomycin selection , pooled clones were passaged and analyzed for luciferase expression and methylation . We chose to use pooled clones to avoid the problem of insertion site variegation effect of single clones . Moreover , such a strategy is commonly accepted and has been successfully used before in a similar study [10] . We measured copy number for every time point and adjusted expression levels of each construct accordingly . As shown in Figure 4B , luciferase readings for both plasmids increased initially , concomitant with stable integration and neomycin selection . However , beyond day 36 post-transfection , RIL expression seemed to be stable for long alleles throughout the course of the experiment , but showed rapid declines for the short alleles . A completely independent seeding/transfection experiment with the same cell line showed similar results ( Figure S4 ) . We then analyzed methylation status of the constructs after stable transfection with bisulfite sequencing using transgene-specific primers ( Figure 4C ) . We found that , at day 19 post-transfection , both constructs were equally methylated at HpaII sites with some evidence of spreading ( average methylation 20 . 3% and 19 . 6% for long and short construct , respectively ) . However , at day 57 , the RIL short allele construct remained methylated at the HpaII sites , and in addition , showed substantial evidence of methylation spreading ( average methylation 28 . 1% ) . By contrast , the RIL long allele construct was found to have demethylation , at HpaII sites , which coexisted with modest methylation spreading on other sites and showed lower levels of methylation overall ( average methylation 11 . 7% ) . Thus , the short allele construct was methylated 2 . 4-fold higher than the long allele construct on day 57 ( P<0 . 001 by Fisher exact test ) . Interestingly , new methylation in the short construct did not seem to be spreading to CpG sites directly adjacent to HpaII-methylated sites , but was rather observed randomly between those sites , a phenomenon that has been reported previously [10] . Furthermore , to more fully characterize methylation status of the transgenes , we designed plasmid-specific pyrosequencing assays and measured several time points ( Figure 4D ) . In general , our methylation and expression data agreed well . However , we must point out that the degree of loss of expression observed for the short alleles cannot be fully explained by methylation gains , thus factors other than methylation ( i . e . preceding chromatin changes ) might have contributed to the decline in luciferase activity , with methylation following this . In summary , the stably integrated RIL long allele construct is protected from methylation and time-dependent silencing , consistent with several previous observations that Sp1 sites confer protection against DNA methylation [11] .
Here we have identified a genetic variation in the promoter region of RIL that results from a presumed insertion of 12 bp around transcription start . This naturally occurring insertion results in a significant 1 . 8-fold lower methylation of the gene in normal samples , and a 2 . 1-fold to 3 . 1-fold lower methylation in cancer that we have traced to the creation of a new Sp1/Sp3 binding site . This finding is consistent with previous reports identifying Sp1 sites as regulatory DNA elements protecting CpG islands from methylation during embryogenesis [12]; [13] and can now be extended to hypermethylation in aging and cancer . This model is consistent with the concept proposed by Turker and others [14] , which suggests that methylation is a non-random process that results in a misbalance between DNA methylation-promoting events ( e . g . abnormal spreading of methylation in cis from adjacent “methylation centers” , repeat elements; and methylation “seeding” ) and DNA methylation preventing events ( i . e . , Sp1 elements in the promoters as well as other insulator sequences , and proteins bound to those sequences; as well as demethylation ) . Our study demonstrates that trans-acting factors can affect hypermethylation in aging and cancer , and establish a new concept in gene inactivation whereby genetic polymorphisms in protein binding sites result in variable susceptibility to epigenetic silencing . A relationship between genetic polymorphisms and aberrant methylation was recently described for MSH2 and was named heritable germline epimutation [15] . A similar relationship has also been described for MGMT [16] . In these cases , the exact mechanism of such predispositions is unknown . It is attractive to think that these heritable epimutations mechanistically represent polymorphisms that influence DNA-binding proteins , as in the case of RIL . It would now be of interest to determine how many genes hypermethylated in cancer are affected by such promoter polymorphisms . The mechanism of Sp1/Sp3 protection against silencing remains unclear . Our studies of RIL expression both in vivo in EBV-transformed cells ( Figure 2 ) and in vitro ( Figure 4A ) , using allele-specific reporter gene assays showed no difference in expression levels between the two alleles of RIL . Rather , our results with the in vitro seeding model are generally consistent with the idea that it protects against time-dependent silencing . However , the degree of methylation spreading observed in the short alleles cannot quantitatively explain the observed prominent drop in transcription over time ( Figure 4B–D ) , and we suggest that the effects of the additional Sp1/Sp3 site are to create local protection against a repressive chromatin environment ( e . g . histone based silencing ) which , secondarily , leads to the observed DNA methylation differences . There are also clear limitations to short term in-vitro assays in modeling in-vivo situations . The experiments described attempt to reproduce in a relatively short period of time an in-vivo situation which evolves over decades of human aging . Nevertheless , it provides an experimental validation of the in-vivo situation , and shows that the short alleles are more prone to time dependent silencing . Our findings are in line with previous in-vitro studies of the APRT gene that suggested that different Sp1 sites might have different functions; some involved in regulation of gene expression , whereas others in protection against DNA methylation [11] . In support of this model , recent data pointed out that Sp1 may have boundary activities and prevent heterochromatin spreading in yeast , which lack Sp1 homologues: targeting of Sp1 to transgenes in yeast cells unexpectedly revealed barrier activity , which was independent of a transactivation domain [17] . It is tempting to speculate that the Sp1/Sp3 site created by the polymorphism in RIL plays such a role . Identification of allelic variants of genes that have different susceptibility to methylation in aging and cancer is an important new link between human genetic variation and epigenetic silencing . In this study , we propose a plausible mechanism responsible for this unique biological phenomenon .
HCT116 , RKO and NIH3T3 cells were grown in high glucose DMEM ( Life Technologies , Gaithersburg , MD ) plus 10% fetal bovine serum ( FBS; Intergen , Purchase , NY ) ; OVCAR-5 , PC3 , MB-231 and the EBV-transformed cells were grown in RPMI 1640 media plus 10% FBS . All cells were grown in plastic tissue culture plates in a humidified atmosphere containing 5% CO2 at 37°C . Samples of primary colon cancers and primary leukemias were obtained from established tissue banks at M . D . Anderson Cancer Center and Johns Hopkins University . Of 44 primary colon cancer samples , 10 were stage II , 12 stage II , 7 stage IV , and 15 were of unknown stage . EBV-transformed lymphocyte cell lines established from normal Caucasian individuals were obtained from Baylor College of Medicine . All samples were collected from consenting patients according to institutional guidelines . DNA was extracted by using conventional phenol-chloroform method . Total cellular RNA was extracted with TriZOL ( Invitrogen , Carlsbad , CA ) according to the manufacturer protocol and resuspended in DEPC-treated water . Reverse transcription reactions were performed using MMLV-RT ( Roche , Indianapolis , IN ) on 2 µg of total RNA per reaction according to the manufacturer's protocol . Genotyping to determine RIL allele status was performed by genomic PCR using del7 ( TCCAGGCGCACAGGGAGC ) and del8 ( GCCTGAGCCGGACTCTGAGGA ) primers . PCR products were separated on 6% polyacrylamide gel and were classified as short , long , or heterozygous , depending on the size and number of bands . Several cases were verified by direct sequencing . Bisulfite induces deamination of unmethylated cytosines , converting unmethylated CpG sites to UpG without modifying methylated sites . Bisulfite treatment of genomic DNA was performed as described [18] . DNA was extracted using standard phenol-chloroform method . After extraction , 2 µg of DNA were used for bisulfite treatment . DNA was denatured in 0 . 2 N NaOH at 37°C for 10 min and incubated with 3 M sodium bisulfite at 50°C for 16 h . Bisulfite-converted DNA was purified using the Wizard cleanup system ( Promega , Madison , WI ) and desulfonated with 0 . 3 N NaOH at 25°C for 5 min . DNA was then precipitated with ammonium acetate and ethanol , washed with 70% ethanol , dried and resuspended in H2O . PCR reactions were carried in 50-µl reactions using the COBRA primers ( forward , GTTTATTAGGYGGAAGTTTTAGG and reverse , AACCAATCCAAACRCACAA ) are complementary to the RIL antisense strand . In each reaction , 2 µl of bisulfite-treated DNA were used , as well as 1 . 25 mM deoxynucleotide triphosphate , 67 mM Tris-HCl , pH = 8 . 8 , 16 mM ammonium sulfate , 10 mM β-mercaptoethanol , 0 . 1 mg/ml bovine serum albumin , 10 pmol of primers and 1 unit of Taq polymerase . All PCR reactions were performed using a hot start at 95°C for 5 min . After amplification PCR products were digested with the HpyCH4IV restriction enzyme ( New England Biolabs , Ipswitch , MA ) , which digests alleles that were methylated prior to bisulfite treatment . The digested DNA was separated in nondenaturing polyacrylamide gels and stained with ethidium bromide . The proportion of methylated versus unmethylated product ( digested versus undigested ) was quantitated by densitometric analysis , performed using a Bio-Rad Geldoc 2000 digital analyzer equipped with the Quantity One version 4 . 0 . 3 software . In case of bisulfite sequencing , restriction enzyme digestion step was omitted , PCR products were directly cloned into a TOPO-TA vector ( Invitrogen , Carlsbad , CA ) and individual clones were sequenced . To study methylation in normal colon , primary cancer samples and in 3T3 cells transfected with HpaII-seeded constructs , we used the pyrosequencing method [19] . For PCR , we used 2 µL bisulfite treated DNA , 1 . 25 mM deoxynucleotide triphosphate , 1 unit of Taq polymerase and the PCR buffer mentioned above , 10 pmol forward primer , 1 pmol reverse-universal primer , and 9 pmol universal biotinylated primer ( assays E , F , G ) . In assay A , we used 1 pmol forward-universal primer , 10 pmol reverse primer and 9 pmol universal biotinylated primer . In assay B , the reverse primer was directly biotinylated . In plasmid-specific assays , we used a two-step PCR approach using nested primers , and in the second step , we used 10 pmol forward primer , 1 pmol reverse-universal primer , 1 pmol reverse primer and 9 pmol universal biotinylated primer . All primer sequences and conditions are shown in Supplementary Table S1 . The final biotin-labeled PCR product was captured by Streptavidin Sepharose HP ( Amersham Biosciences , Sweden ) . PCR products bound on the bead were purified and made single stranded using a Pyrosequencing Vacuum Prep Tool ( Biotage , Sweden ) . The sequencing primers ( 0 . 3 µmol/L; Supplementary Table S1 ) were annealed to the single-stranded PCR product , and pyrosequencing was done using the PSQ HS 96 Pyrosequencing System ( Biotage , Sweden ) . Quantification of cytosine methylation was done using the provided software ( PSQ HS96A 1 . 2 ) . Nuclear extracts were prepared with the NE-PER nuclear extraction reagents ( Pierce , Rockford , IL ) , according to manufacturer's instructions . Briefly , cells were lysed in hypotonic CER buffer with added protease inhibitors and the cytoplasmic fraction was separated by centrifugation . The nuclear pellet was resuspended in hypertonic NER buffer with added protease inhibitors and incubated on ice for 40′ , vortexed every 10′ . For nuclear extracts , the soluble proteins in the lysate were separated by centrifugation . The protein content was quantified using the Bio-Rad protein assay . All EMSA probes were generated by annealing synthetic complementary oligonucleotides ( Invitrogen ) corresponding to the deletion/insertion region of the RIL promoter , sequences are shown in Figure 4A . 32P-end-labeled oligonucleotides ( 100 , 000 cpm ) were incubated for 30 min on ice with 5–10 µg of nuclear extract or , alternatively , with 200 ng of recombinant human Sp1 ( Promega ) in 15 µl of binding buffer containing 10 mM Tris-HCl , pH 7 . 5 , 0 . 5 mM EDTA , 0 . 5 mM dithiothreitol , 4% glycerol , 1 mM MgCl2 , 50 mM NaCl , and 1 µg of poly ( dI-dC ) . The binding reaction was incubated on ice for 30 min . Competition reactions were performed with a 20-fold molar excess of unlabeled double-stranded competitor DNA . For supershift analysis , the nuclear extracts were incubated with polyclonal anti-human Sp1 , Sp3 , Egr-1 and AP-2 antibodies ( Santa Cruz Biotechnology , Santa Cruz , CA ) for 1 h following the binding reaction with labeled probe . The DNA-protein complexes were separated on a 5% native polyacrylamide gel in 0 . 5× TBE buffer for 14 h at +4°C at 90 V . DNA from human colon cancer cells HCT116 and RKO was extracted as previously described , using phenol chloroform methods . The more 3′ located transcription start site ( Figure 1A ) was arbitrarily chosen as +1 . The 236-bp genomic fragment from HCT116 cells ( homozygous long ) and the 224-bp genomic fragment from RKO cells ( homozygous short ) spanning from −217 to +19 ( construct A ) were generated by PCR from genomic DNA using RIL-LUC1 ( CGGAGCTCTCTCTGAGAGCTGAGTGGGG ) and RIL-LUC2 ( GCAAGCTTCTGAGCCGGACTCTGAGG ) primers . The 607-bp genomic fragment from HCT116 cells and the 595-bp genomic fragment from RKO cells spanning from −588 to +19 ( construct B ) were generated by PCR from genomic DNA using RIL-LUC3 and RIL-LUC2 primers . Both RIL-LUC1 and RIL-LUC3 ( GGGAGCTCCCTTACTGGCCTCCACAAAC ) primers contained SacI restriction enzyme sites , whereas RIL-LUC2 primer contained HindIII restriction site . The genomic fragments from HCT116 cells ( homozygous long ) and from RKO cells ( homozygous short ) spanning from −588 to +516 ( construct C ) were generated by PCR from genomic DNA using RIL-LUC4 ( GTGCTAGCCCTTACTGGCCTCCACAAAC ) and RIL-LUC5 ( CCAAGCTTGGACCTGCGAGCAGACAAGCCTCATTTTGCCCCAGATCTTC ) primers . In construct C , the reverse primer contained additional intronic sequence to assure correct splicing of the exon 1 . The PCR products were digested with NheI and HindIII enzymes ( NEB ) and subcloned in the promoterless pGL3 basic vector ( Promega ) and , generating the allele-specific constructs A ( 236/224-bp ) , B ( 607/595-bp ) and C ( 1104/1092-bp ) . The constructs were confirmed by DNA sequencing . Plasmids were transfected using Lipofectamine2000 ( Invitrogen ) , according to manufacturer's protocol . The cells were grown in 6-well plates to 60% confluence for at least 18 h and then were transiently transfected using Fugene6 reagent ( Roche , Indianapolis , IN ) with 1 µg of a firefly luciferase reporter gene and 50 ng of the Renilla luciferase reporter gene , driven by the thymidine kinase promoter ( Promega ) . Cotransfection was performed by adding 1 µg of expression constructs to the DNA solutions . Luciferase activity was determined using the dual luciferase reporter assay system ( Promega ) in a microplate Monolight 3010 luminometer ( BD Biosciences , San Diego , CA ) according to the manufacturer's instructions . Normalization of transfection efficiency was based on cotransfected thymidine kinase Renilla luciferase activities . RIL allele-specific constructs ( construct B , ∼0 . 6 KB ) were methylated using HpaII methylase ( NEB ) according to manufacturer's protocol . Methylation was confirmed by resistance to restriction to HpaII enzyme ( Figure S3 ) . For stable transfections , allele-specific luciferase constructs were co-transfected with pcDNA3 . 1 neomycin-containing vector in 20∶1 ratio into 3T3 cells using Lipofectamine 2000 ( Invitrogen ) . Methylation analysis of transfected constructs was done by amplification of bisulfite treated DNA , using plasmid-specific primers ( F: GAGAGTTGAGTGGGGGTGT , R: TTCCATAATAACTTTACCAACAATACC ) followed by bisulfite sequencing of single clones . For pyrosequencing , these primers were used for the 1st step PCR , followed by second step with nested primers ( Supplementary Table S1 ) . One site was analyzed by the assay , which corresponds to CpG site # 9 in bisulfite sequencing ( Figure 4C ) . Q-PCR was used to quantify RIL mRNA in total RNA isolated from EBV-transformed lymphoblastoid cell lines and reverse transcribed into cDNA , and also to estimate relative plasmid copy numbers in the genomic DNA isolated from pooled transfected NIH3T3 cells . We used TaqMan Universal PCR Master Mix and ABI Prism 7000 Sequence Detection System ( Applied Biosystems , Foster City , CA ) . The primers and TaqMan probes were designed by using Primer Express software ( Applied Biosystems ) , except for RIL expression assay which was custom-designed and purchased from Applied Biosystems ( cat . No . Hs00184792_m1 ) . All probes were labeled with the 6-carboxyfluorescein fluorophore ( 6-FAM ) and a nonfluorescent MGB quencher . To normalize RIL expression , GAPDH was used as reference standards; to estimate the relative plasmid copy number per cell , murine beta-globin was used as endogenous reference . Primer and probe sequences used in the experiments were as follows: ( 1 ) RIL assay , custom-designed and purchased from Applied Biosystems , cat . No . Hs00184792_m1; ( 2 ) GAPDH assay , gapdh-284F , 5′-ATGGAAATCCCATCACCATCTT-3′; gapdh-340R , 5′-CGCCCCACTTGATTTTGG-3′; gapdh-307T ( MGB probe , FAM fluorophore labeled ) , 5-CGCAGTTGGGCACTT-3; ( 3 ) Luciferase assay , luc-161F , 5′-CACATATCGAGGTGGACATC-3′; luc-220R , 5′- GCCAACCGAACGGACATTT-3′; luc-184T ( MGB probe , FAM fluorophore labeled ) , 5-CTTACGCTGAGTACTTC-3; ( 4 ) Murine beta-globin assay , Mu-bglo-239F , 5′- AGGCCCATGGCAAGAAAGT-3′ , Mu-bglo-306R , 5′-GCCCTTGAGGCTGTCCAA-3′ , Mu-bglo-259T ( MGB probe , FAM fluorophore labeled ) , 5-ATAACTGCCTTTAACGATG-3 . All FAM probes were custom synthesized by Applied Biosystems . The primers were used at 900 nM and the probes at 100 nM concentrations . We used the RNA or DNA amount giving the linear range of response , typically CT range of 20–30 amplification cycles . We amplified each gene in separate reactions in triplicates , using a protocol recommended by manufacturer . PCR without reverse transcriptase was performed for each sample to control for the possible interference from gDNA contamination . The threshold amplification cycles ( CT ) at the normalized reporter signal minus the baseline signal level of 0 . 2 for RIL , GAPDH , Luciferase and murine beta-globin were determined and their differences deltaCT ( GAPDH-RIL ) and deltaCT ( murine beta globin-Luciferase ) were calculated . Statistical differences of bisulfite sequencing data were calculated using Fisher exact test . To analyze methylation levels across the CpG island , we calculated average±standard error of the mean for each region studied , and performed T-test using Microsoft Excel . To estimate correlation between average methylation density ( regions A–F ) and age for each region in normal colon samples , non-parametric two-tail Spearman test ( 95% confidence interval ) was used ( Graphpad Prizm 4 software ) . All EMSA experiments were done in triplicates and yielded identical results . Luciferase assays on cells were performed three separate times , and the results were expressed as average±standard error of the mean calculated using Microsoft Excel .
|
The factors that guide DNA hypermethylation in cancer are poorly understood . We identified the candidate tumor-suppressor gene , RIL , as a frequent methylation target in cancer . Here , we report on a 12-bp polymorphic sequence around its transcription start site that creates a long allele . Methylation analysis showed that , in aging colon , colon cancer , and leukemias , the short allele had 2 . 1–3 . 1-fold higher methylation than the long allele ( P<0 . 001 ) . Short and long alleles had similar expression levels in EBV-transformed cell lines . Electrophorectic mobility shift assay showed that the inserted region of the long allele binds Sp1 and Sp3 transcription factors . Transfection of RIL allele-specific transgenes showed no effects of the additional Sp1 site on transcription early on , but methylation-seeded constructs showed gradually decreasing transcription from the short allele with eventual spreading of de novo methylation . By contrast , the long allele showed stable expression over time as measured by luciferase , and ∼2–3-fold lower levels of methylation by bisulfite sequencing ( P<0 . 001 ) , suggesting that the polymorphic Sp1 site protects against time-dependent silencing . Our finding demonstrates that in some genes , hypermethylation in cancer is dictated by protein-DNA interactions at the promoters and provides a novel mechanism by which genetic polymorphisms can influence an epigenetic state .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"gastroenterology",
"and",
"hepatology/gastrointestinal",
"cancers",
"genetics",
"and",
"genomics/gene",
"expression",
"hematology/acute",
"myeloid",
"leukemia",
"molecular",
"biology/dna",
"methylation",
"oncology/gastrointestinal",
"cancers",
"genetics",
"and",
"genomics/epigenetics",
"oncology/hematological",
"malignancies",
"molecular",
"biology/chromatin",
"structure"
] |
2008
|
An Sp1/Sp3 Binding Polymorphism Confers Methylation Protection
|
Human Papillomaviruses ( HPV ) cause widespread infections in humans , resulting in latent infections or diseases ranging from benign hyperplasia to cancers . HPV-induced pathologies result from complex interplays between viral proteins and the host proteome . Given the major public health concern due to HPV-associated cancers , most studies have focused on the early proteins expressed by HPV genotypes with high oncogenic potential ( designated high-risk HPV or HR-HPV ) . To advance the global understanding of HPV pathogenesis , we mapped the virus/host interaction networks of the E2 regulatory protein from 12 genotypes representative of the range of HPV pathogenicity . Large-scale identification of E2-interaction partners was performed by yeast two-hybrid screenings of a HaCaT cDNA library . Based on a high-confidence scoring scheme , a subset of these partners was then validated for pair-wise interaction in mammalian cells with the whole range of the 12 E2 proteins , allowing a comparative interaction analysis . Hierarchical clustering of E2-host interaction profiles mostly recapitulated HPV phylogeny and provides clues to the involvement of E2 in HPV infection . A set of cellular proteins could thus be identified discriminating , among the mucosal HPV , E2 proteins of HR-HPV 16 or 18 from the non-oncogenic genital HPV . The study of the interaction networks revealed a preferential hijacking of highly connected cellular proteins and the targeting of several functional families . These include transcription regulation , regulation of apoptosis , RNA processing , ubiquitination and intracellular trafficking . The present work provides an overview of E2 biological functions across multiple HPV genotypes .
Papillomaviruses are non-enveloped small DNA viruses , of which over 140 types infect humans ( HPV ) . HPV are strictly epitheliotropic , with specificity for stratified epithelia of the skin ( cutaneous HPV ) or genital and oral mucosa ( mucosal HPV ) . They are either associated with asymptomatic infections or induce benign proliferative lesions , which have the potential to progress toward malignancy for the ‘high risk’ HPV ( HR-HPV ) . Although carcinogenic conversion occurs only in a minority of infections , mucosal HR-HPV are associated with almost all cervical cancers , and with 50% anogenital and 30% head and neck cancers [1] . In addition , growing evidence point to a role of some cutaneous HPV in non-melanoma skin cancer [2] . Therefore , from inapparent infections to cancers , HPV cover a large spectrum of diseases in humans [3] . The productive viral cycle both depends on and perturbs the differentiation of infected keratinocytes [4] , and HPV pathogenesis relies on complex interplay between early viral and host proteins . The carcinogenic conversion of HR-HPV-associated lesions proceeds from a deregulation of virus-host cross-talk , leading to over-expression of E6 and E7 viral oncogenes and to the accumulation of cellular genetic alterations . This long-lasting process culminates in the emergence of fully-transformed cells critically dependent on the immortalizing properties of the HR-HPV E6 and E7 proteins to drive continuous cell proliferation . The HPV E2 early protein is a pivotal factor of both productive and persistent infection . It provides the control of viral DNA transcription , replication and mitotic segregation through specific binding to the viral genome . Such activities are shared by all HPV and are mediated by E2 interactions with cellular transcription factors , mitosis-associated factors , and with the viral E1 helicase ( see [5] , [6] for review ) . As such , the E2 protein is mainly envisioned as a basic viral factor . Contrary to the E6 and E7 proteins , the involvement of E2 in the different features of HPV pathology is elusive . Indeed , only few studies demonstrated that E2 functions may differ between oncogenic HR-HPV and the Low-Risk HPV ( LR-HPV ) , which are always associated with benign hyperplasia . Some activities are specific of the HR-HPV E2 proteins , such as the induction of apoptosis or of a G2/M cell cycle arrest [7]–[9] . In addition , the HR-HPVE2 proteins induce genomic instability [9] , and E2 from cutaneous HPV8 exhibits intrinsic oncogenic potential when expressed in the skin of transgenic mice [10] , pointing to a role of E2 in the carcinogenic conversion of HR-HPV associated lesions ( see [11] for review ) . Given the major public health concern caused by genital cancers , the activities of viral early proteins have been far more extensively studied for mucosal HR-HPV than for other HPV . However , the variability of HPV-associated lesions indicates that the interplay among viral and host proteins may strongly differ . A global understanding of cell alterations generated by viral proteins according to the tropism and pathogenic potential is currently lacking . To make progress in this issue , we mapped the virus-host protein-protein interactions of the E2 proteins from 12 genotypes representative of HPV diversity . We selected HPV of different tropism specificity ( cutaneous: HPV1 , 3 , 5 , 8 , 9 or mucosal: HPV6 , 11 , 16 , 18 , 32 , 33 , 39 ) and with different pathogenic potential ( LR-HPV 1 , 3 , 6 , 9 , 11 , 32 or HR-HPV 5 , 8 , 16 , 18 , 33 , 39 ) . This selection spans over three clades of the typical HPV phylogeny based on the sequence of the L1 capsid protein [12]: α-types HPV 3 , 6 , 11 , 16 , 18 , 32 , 33 , 39; β-types HPV 5 , 8 , 9 and μ-type HPV 1 . Interaction mapping was performed by combining a large scale identification of E2 partners by successive yeast two-hybrid screenings and a cell-based interaction assay for the validation of protein-protein interactions . This work gives an overview of E2 biological functions across multiple HPV genotypes , and provides a comprehensive framework for understanding the role of E2 in HPV pathologies .
To provide a comprehensive assessment of E2-host Protein-Protein Interactions ( PPI ) , we mapped PPI of E2 from 12 HPV genotypes representative of HPV tropism and pathogenic potential: mucosal HR-HPV 16 , 18 , 33 and 39; mucosal LR-HPV 6 , 11 and 32; cutaneous HPV 1 , 3 , 5 , 8 and 9 . The 12 E2 proteins were used as baits in a mating-based yeast two-hybrid ( Y2H ) to screen a human keratinocyte ( HaCaT ) cDNA library . The number of diploid yeasts generated was systematically evaluated to be at least ten times higher than the library complexity . In order to obtain exhaustive Y2H datasets , successive screenings were performed with each of the E2 proteins . In total , the Y2H screen identified 251 distinct interactions involving 202 different cellular proteins . Few proteins were interacting with numerous E2 , indicating a low overlap of E2-PPI in Y2H . Indeed , only 27 proteins ( 13 . 4% ) were picked up with more than one E2 protein as follows: 16 with two E2 , seven with three E2 , and a single one with four , five , six or seven E2 proteins . Five proteins ( GPS2 , SFRS1 , AP3D1 , C1QBP and TP53 ) had been previously identified as E2 interacting proteins in the literature ( further referred to as Literature Curated E2-Protein Protein Interactions or LCE2-PPI ) . LCE2-PPI were extracted from the VirHostNet [13] , virusMINT [14] and PubMed databases ( Table S1 ) . For some of these interactors , PPI were detected in our Y2H screen with an E2 protein of a different HPV genotype than in LCE2-PPI ( Table S2 ) . These latter cases probably point to shared E2-PPI , which could be verified through their assessment with the series of 12 E2 proteins . The recovery of 5 out of 53 known E2 partners indicated a sensitivity of Y2H screening around 10% , which is in the range of previously described similar analyses [15] . This , combined with the high coverage of the HaCaT cDNA library reached in each screening , suggest a satisfactory sampling sensitivity . Overall , the Y2H screen led to the identification of 197 new potential cellular binding partners of at least one E2 protein . The Y2H screen applied to a wide spectrum of HPV genotypes was appropriate to get an overview of E2-PPI without bias toward the most studied E2 proteins , contrary to the LCE2-PPI datasets . However , the coverage of PPI detected by Y2H is estimated to be around 20% of total PPI [16] , highlighting a high false-negative rate inherent to this screening methodology [17] . We therefore speculated that , despite repetitive probing of the HaCaT cDNA library with each of the E2 protein , PPI detected with a subset of E2 might have escaped detection with the others , which would explain the low overlap of E2-PPI observed in the Y2H screens . Moreover , it was previously demonstrated that combining different methodologies is necessary to increase the robustness of PPI datasets [18] . A stringent validation strategy consists in the use of orthogonal PPI detection methods , as it ensures the discarding of false positive interactions generated in Y2H screens . We thus decided to challenge a subset of cellular proteins selected from the Y2H screen for pair-wise interaction with the whole set of the 12 E2 proteins using a mammalian cell-based orthogonal PPI detection assay . Such a strategy allows a comparative interaction analysis among the different HPV genotypes . We used a secondary High-Throughput Gaussia princeps luciferase-based Complementation Assay ( HT-GPCA , Figure 1A ) recently described [19] . Briefly , bait and prey proteins were expressed in 293T cells in fusion with two inactive fragments of the Gaussia princeps luciferase ( designated GL1 and GL2 ) , which restore a significant enzymatic activity when brought in close proximity by an interaction . The reconstituted Luciferase activity is estimated from a Normalized Luminescence Ratio ( NLR , Figure 1A ) . This assay has been recently benchmarked by using two positive reference sets of protein pairs known to interact , and a set of a priori non-interacting protein pairs [19] . It was determined that when setting a NLR threshold of 3 . 5 , there was only 30% false negatives ( known PPI not recovered in HT-GPCA ) and 2 . 5% false positives . A 3 . 5 NLR threshold was accordingly used to discriminate positive interactions in the present study . A high-confidence core set of 48 potential E2 partners was selected for validation from the Y2H dataset by keeping proteins identified at least three times in Y2H [20] . Assuming that potential false positives would be eliminated by combining two orthogonal methods , 54 proteins found only one or two times were additionally rescued for further validation in mammalian cells . This non-core set consisted in proteins known from other studies as E2 partners , proteins functionally relevant to E2 ( transcription , replication factors ) , or proteins related to potential E2 partners of the core set . In total , 102 proteins were selected , corresponding to 138 distinct Y2H interactions obtained through 1 , 135 sequenced PPI ( Figure S1 and Table S3 ) . We also increased the explored area by including 19 known E2 partners , which were used as positive controls herein referred as Gold Standards ( GS ) . Combined with the five known E2 partners recovered in our Y2H screen , the final list of GS comprised 24 cellular proteins . In total , 121 cellular proteins were to be validated for interaction with E2 , of which 97 represented novel potential partners of E2 . Before proceeding to HT-GPCA , we wished to ensure that fusion with a Gaussia fragment would not alter the folding and functionality of E2 in the GL2-E2 fusion proteins . To that aim , we assessed E2-dependent transcription of pTK6E2BS , containing six E2 binding sites ( E2BS ) upstream of the minimal TK promoter . The sequences of E2BS were designed to be optimal for the binding of a large panel of E2 [21] in order to homogenize E2 binding to this promoter . All GL2-E2 fusion proteins properly activated transcription , demonstrating that the E2 proteins were functional ( Figure 1B ) and thus that fusion of the GL2 tag at their N-terminus did not induce incorrect folding or localization . The relative accumulation of the E2 proteins was approximated by fusion with the Firefly luciferase protein ( Fluc-E2 fusion ) , so that their expression levels could be deduced from luciferase activity as previously reported [22] . Fluc-E2 fusion proteins accumulated to levels ranging from 5% ( HPV32 E2 ) to 35% ( HPV1 E2 ) of the Firefly luciferase alone , indicating variations in E2 accumulation levels ( Figure 1C ) . However , there was no correlation between steady-state levels and transcriptional activation ( Figure 1B and 1C ) , pointing to differences in the intrinsic transcriptional properties of the E2 proteins , thereby corroborating previous studies [23] , [24] . As for the GL2-E2 fusion proteins , the expression levels of the selected 121 cellular proteins expressed as GL1 fusions may vary . The heterogeneity in protein accumulation levels would potentially bring a degree of variability in HT-GPCA assay , that have to be taken into consideration for the comparative analysis of their interaction patterns . To evaluate the reliability and sensitivity of the HT-GPCA method applied to E2 , we first conducted a pilot experiment with the set of 24 gold standards ( GS ) , which covered about half of LCE2-PPI ( Table S1 ) . The results are displayed as Heat maps where the intensity of an interaction , based on the NLR , is represented by a color gradient from black ( no interaction ) to light blue ( strong interaction ) ( Figure 2A ) . As underscored in Figure 2 , the majority of E2-PPI in the GS set had been studied with a single E2 , mainly 16E2 . Thus our approach significantly broadens the scope of GS analysis . Out of the 39 studied known E2 interactions ( LCE2-PPI ) , 28 interactions ( 72% ) were recovered by HT-GPCA ( Figure 2A and Table S4 ) . For 7 of the 11 expected LCE2-PPI that failed to be recovered , the corresponding gold standard protein interacted with other E2 proteins , suggesting that the missing interactions represent HT-GPCA false negative interactions . We noticed that PPI with p53 ( TP53 ) were detected in HT-GPCA with all the E2 proteins , in contrast to previous studies showing that this interaction was restricted to mucosal HR-HPV E2 [8] . Such discrepancies might be due to increased sensitivity of HT-GPCA , and outline the need to combine different methods to improve the confidence of interactions datasets . A negative control interaction matrix provided a rough estimate of the false-positive rate of PPI detected by HT-GPCA applied to the E2 proteins . It consisted in cellular proteins randomly picked in the human ORFeome resource , a priori not interacting with E2 . Among this matrix of 120 PPI ( 12E2×10 proteins ) the false positive rate was 5 . 8% ( Figure 2B ) . Furthermore , the specificity of HT-GPCA was illustrated by using E2 proteins invalidated for interaction with BRD4 by mutation of Isoleucine 73 ( HPV16 ) or 77 ( HPV18 ) to Alanine [25] , [26] . Both mutated proteins , 16E2 I73A and 18E2 I77A exhibited an impaired binding to BRD4 , with a five fold decrease of NLR when compared to the wild-type proteins ( Figure 2C ) . Overall , these data demonstrate both the robustness and the sensitivity of HT-GPCA to detect pair wise interactions involving the E2 proteins . We then processed all selected cellular proteins , performing 1 , 452 ( = 121×12 ) tests . In total , 617 interactions ( 42% ) exhibited a NLR above 3 . 5 , thereby scoring positive in HT-GPCA . Of the 121 cellular proteins tested , 23 ( 19% ) did not engage detectable interaction with any of the E2 proteins ( Table S5 ) . Of note , virtually all of the 98 validated partners interacted with more than one E2 protein , highlighting a high overlap between E2 interactors . Comparison of the Y2H and HT-GPCA PPI datasets ( schematized in Figure S1 ) indicated that among the 138 interactions detected in the Y2H screen ( Y2H-PPI ) , 72 were validated in HT-GPCA , representing 53 cellular proteins . 38 Y2H-PPI , involving 27 cellular proteins , were not recovered in HT-GPCA but interactions were detected with different E2 proteins than in Y2H . As discussed previously , we assume that the corresponding non-recovered Y2H-PPI most probably represent HT-GPCA false negative interactions . Lastly , 28 Y2H-PPI were not validated and involved 22 cellular proteins that did not interact with any E2 proteins . These proteins were consequently discarded for further analyses . Altogether , these results point a 53% overlap between Y2H-PPI and HT-GPCA interactions . When considering the interactors , the recovery was 79% . PPI validation rate was higher with μ- or β-types E2 proteins ( HPV1 , 5 , 8 , 9 ) than with the α-types E2 ( HPV 3 , 6 , 11 , 16 , 18 , 32 , 33 and 39 ) , as reflected by brightness variations of the heat maps ( Figure 3A ) . Significantly , the overall NLR levels were not related to E2 accumulation levels , since 9E2 exhibited the highest interaction rate but was not the most accumulated . Conversely , 33E2 engaged the most interactions in the mucosal group , whereas it accumulated at low levels ( Figure 1C ) . These observations clearly argue that variations in E2 accumulation levels are not driving the differences observed by HT-GPCA , and therefore do not essentially alter this comparative interaction mapping . Differences in E2 interaction rates are more likely related to intrinsic characteristics of the proteins . Notably , the μ or β-E2 proteins contain the longest hinge regions ( > 122 amino acid , <79 for the others ) , which is an intrinsically disordered segment in E2 proteins . Their higher interaction rate is consistent with the notion that disordered regions are enriched in exposed interaction motifs [27] . The E2-host PPI profiles were gathered according to their similarities by agglomerative hierarchical clustering ( Cluster software ) . Matrix tree plots were generated from this analysis , and were used to build an E2-interaction dendrogram using the UPGMA ( Unweighted Pair Group Method with Arithmetic Mean ) method with the Euclidian distance function and a complete linkage method ( Figure 3B ) . Strikingly , in the interaction dendrogram , the E2 proteins segregated according to HPV tropism ( cutaneous and mucosal ) , and further clustered following pathogenic potential ( high-risk versus low-risk ) . We found a high correlation between the E2-interaction dendrogram and the phylogenetic dendrogram based on the E2 sequences ( Figure 3B ) , the Pearson correlation coefficient calculated from the distance matrices of the two dendrograms was 0 . 91 , with a p-value < 10−10 . Of note , using different parameters for the clustering analysis did not drastically affect the structure of the interaction dendrogram or the closeness to phylogenetic tree ( Figure S2 ) . This observation demonstrates the robustness of the interaction dendrogram generated by our approach . The E2 protein is mainly envisioned as a basic viral factor , essential for all HPV through its regulatory role of viral DNA transcription , replication and mitotic segregation . Only few studies demonstrated that E2 activities may differ according to HR or LR-HPV type [8] , [9] , [27] . We show in the present study that the E2 proteins engage different patterns of interaction with the host proteome depending on both the tropism and the HR or LR trait of HPV . Such interaction mapping may thus improve the understanding of cell alterations induced by E2 . The very first branching division in the E2-PPI dendrogram separates the β/μ from the α-types E2 , which essentially corresponds to a distinction between cutaneous ( β/μ-HPV ) and mucosal ( α-HPV except for HPV3 ) HPV . Within each group , the interaction profiles further clustered according to pathogenic potential . Now considering only the α group , we compared the interaction profiles of E2 from the genital HR-HPV 16 , 18 , 33 and 39 with those of the LR-HPV 6 and 11 , in order to extract interactions that may play a role for the life cycle of mucosal HR-HPV . Only one protein , GPS2 , which is an integral component of the NCoR complex ( Nuclear receptor Co-Repressor ) [28] interacted with all mucosal HR-HPV and not with the LR-HPV E2 proteins . Focusing on the most prevalent HR-HPV16 and 18 , we identified a series of cellular proteins differentially bound by either 16E2 or 18E2 compared to the LR-HPV E2 proteins ( table 1 ) . Six of these proteins were targeted by both 16E2 and 18E2 ( GPS2 , HSP5A , ARFIP2 , CDC20 , SPTAN , VPS52 ) , whereas the others were genotype-specific . It is noteworthy that the above-mentioned cellular partners interact with members of the μ/β-types HPV E2 proteins , suggesting that they could also take part to HPV pathogenesis in the context of cutaneous tropism . Of note , the CDC20 protein was recovered among the HR-specific partners interacting with both 16 and 18E2 , in line with the proposed role of this interaction in carcinogenic conversion associated with both genotypes [9] . Given that the mucosal HR and LR viruses infect distinct biological niches ( HPV16/18 infect mucosal transformation zones , while HPV6/11 infect the external genitalia ) , such discriminating interactions could result from the different proteome of infected tissues . They nevertheless might point to targets important for the life cycle of HPV16 or HPV18 genotypes , responsible for most of the genital cancers . Topological analysis of viral interaction networks can be informative with regard to the global impact of viral proteins in the host cell , as well as the dynamics of viral pathogenesis . To conduct such analysis , we built E2-host interaction networks with PPI scoring positive in HT-GPCA . The degree of a protein reflects the number of interactions it engages in the cell , and the degree distribution of a network gives a measure of its local dynamics . We studied the degree distribution of the E2-host network compared to that of a human interactome reconstructed from Human Protein Reference Database ( HPRD 2010 release 9 ) , including 39 , 100 binary protein-protein interactions . The cumulative plot of E2 and human interactomes relative to protein degree ( Figure 4A ) shows that 75% of proteins of the human interactome have a degree lower than eight ( estimated mean degree of the present human interactome ) , while for only 25% of E2 targets the degree is lower than eight . Such difference was found statistically significant by the Kolmogorov-Smirnov test ( 0 . 5 with a p-value < 0 . 002 ) . These results indicated that E2 proteins preferentially bind to highly connected proteins , also called hubs . The distribution of degree probability in both interactomes further substantiates a clear overrepresentation of high-degree proteins in the E2 interactome ( Figure 4B ) . Overall , these results show that E2 proteins preferentially target highly connected cellular proteins . Such findings indicate that E2 broadly impacts on host cells by interacting with key proteins involved in many pathways of the cellular network . This likely maximizes E2 effects on a wide range of cellular functions . The preferential targeting of central proteins was previously observed with other viral proteins from EBV , KSHV and HCV [29]–[31] . Indeed , the binding to hub proteins could be a general hallmark of viral proteins to hijack at a systemic level the cellular interactome . It is noteworthy that protein centrality has been correlated with the presence of disordered regions [27] . We may therefore speculate that the intrinsically disordered hinge region provides a platform for most E2 interactions with the host proteome , as discussed previously . We next analyzed the E2 interaction network from a functional point of view to get insights into the functions of E2 that could emerge . Gathering of E2-targeted cellular proteins based on their GO ( Gene Ontology ) terms with the DAVID bioinformatics base [32] indicated enrichment of E2 targets in the following functional families: regulation of transcription , regulation of apoptosis , RNA processing , ubiquitination processes and intracellular transport ( Figure 5 and details of the DAVID analysis are given in the Figure S3 ) . This analysis gives an overview of E2 biological functions across multiple HPV genotypes . A subset of E2 cellular targets was selected in order to provide further biological insight to some of the E2-host PPI identified from the HT-GPCA dataset .
The 12 ORF encoding for the E2 proteins were amplified from viral genomic DNA corresponding to the different HPV genotypes , cloned by the gateway recombinational cloning system ( Invitrogen ) into the entry vector pDON207 ( Invitrogen ) , and were listed in the ViralORFeome database [47] . The E2 ORFs were then transferred into gateway-compatible destination vectors pGBKT7-gw to generate E2-GAL4 DNA-binding domain fusion proteins for Y2H; pCI-Neo-FLuc-gw to generate Firefly luciferase-E2 fusions proteins for steady state levels measurement; pSPICA-N2-gw to generate proteins with amino acids 110 to 185 of the humanized Gaussia princeps luciferase in fusion with the N-terminus of E2 ( GL2-E2 fusion proteins ) for the High-Throughput Gaussia princeps Luciferase-based Complementation Assay ( see [19] for construct details ) . Entry gateway plasmids for cellular partners were obtained either by PCR amplification from clones recovered by Y2H or from the human ORFeome resource ( hORFeome v3 . 1 ) . The cellular ORF were transferred into gateway-compatible destination vectors pSPICA-N1-gw to generate proteins fused at the N-terminus with the amino acids 18 to 109 of humanized Gaussia luciferase ( GL1-fusion proteins ) . Mutagenesis of E2 proteins from HPV 16 and 18 was performed by PCR-directed mutagenesis method . The luciferase reporter ( pTK6E2BS ) driven by E2-responsive promoter contained 6 E2 binding sites upstream the minimal TK promoter . E2 BS sequences were as follows: ( aACCGTTTTCGGTtaaACCGTTTTCGGTt ) X3 , designed after the study of Sanchez et al [21] to be optimal for the binding of a large panel of E2 proteins . The polymerase III-directed Renilla Luciferase plasmid ( polIII-Ren ) used as an internal control of transfection contained a 100-mer nucleotide encompassing the human Histone H1 promoter upstream of the Renilla ORF ( hRluc ) . For yeast two hybrid screening , GAL4 DNA-binding domain-E2 fusion proteins , expressed from the pGBKT7 vector , were used to probe a human HaCaT cDNA library ( Clontech ) , cloned in fusion with the GAL4 transcription activation domain in pACT2 . Each independent screening was performed by mating pGBKT7-E2 transformed yeast strain AH109 ( MATa , trp1-901 , leu2-3 , 112 , ura3-52 , his3-200 , gal4Δ , gal80Δ , LYS2 : : GAL1UASGAL1TATA-HIS3 , GAL2UAS-GAL2TATA-ADE2 , URA3 : : MEL1UAS-MEL1TATA-lacZ ) with Y187 strain ( MATα , ura3-52 , his3-200 , ade2-101 , trp1-901 , leu2-3 , 112 , gal4Δ , met– , gal80Δ , URA3 : : GAL1UAS-GAL1TATA-lacZ ) transformed with the HaCaT cDNA library . Mating was performed 4 hr at 30°C on plates of non-selective rich YCM media . The number of diploid cells generated was systematically evaluated to be at least 10 times higher the HaCaT cDNA library complexity ( 2 . 5×106 , Clontech ) . Mated yeasts were grown on selective medium lacking tryptophan , leucine and histidine ( SD-W-L-H ) , and supplemented with 3-aminotriazol according to the basal autoactivation test previously performed ( see below ) . HaCaT cDNA sequences from positive colonies were PCR amplified and sequenced . Independent Y2H screens were repeated in the same way for each of the E2 protein until around 100 PPI could have been sequenced . Because bait constructs sometimes self-transactivate reporter genes , SD-W-L-H culture medium was supplemented with 3-aminotriazole ( 3-AT ) in the Y2H screenings . Appropriate concentrations of this inhibitor were determined by growing bait strains ( AH109 yeast strain transformed with each E2 bait ) on SD-W-H culture medium supplemented with increasing concentrations of 3-AT . Concentrations of 3-AT ranging from 5 mM ( for 33 , 39 , 18 , 11 , 5 and 8 E2 ) to 10 mM ( for 1 , 3 , 6 , 9 , 32 and 16E2 ) were sufficient to counter the weak transactivation observed . This falls into the range of Clonetech standards . A bioinformatic pipeline was developed to assign each IST to its native human genome transcript . First , ISTs were filtered by using PHRED at a high quality score , sequence was extracted based on a sliding window of 30 bases which is successively shifted 10 bases until the average quality value from the window falls . A 30 bases motif from pACT2 linker was searched , sequences downstream of this motif were translated into peptides and aligned using BLASTP against human protein sequence databases from Ensembl ( release 58 based on NCBI assembly 37 ) , Uniprot and primate EMBL . Low-confidence alignments ( E value > 10−10 , identity < 80% and peptide length < 20 amino acids ) , frameshifted and premature STOP codon containing sequences were eliminated . HEK-293T cells were seeded at 35 , 000 cells per well in 96-well plates . After 24 h , cells were transfected by linear PEI ( polyethylenimine ) with pSPICA-N2-E2 and pSPICA-N1-cellular protein constructs ( 100 ng each ) , for expression of the GL2-E2 and GL1-fusion proteins , where GL1 and GL2 are two inactive fragments of the Gaussia princeps luciferase . 10 ng of a CMV-firefly luciferase reporter plasmid was added to normalize for transfection efficiency . Cells were lysed 24 h post-transfection in 40 µL of Renilla luciferase lysis buffer ( Promega ) for 30 minutes . The Gaussia princeps luciferase activity was measured on 30 µL of total cell lysate by a luminometer Berthold Centro XS LB960 after injection of 100 µL of the Renilla luciferase substrate ( Promega ) . Firefly luciferase was measured on the remaining 10 µl lysate with Firefly luciferase substrate . Gaussia Luciferase activity was reported to Firefly luciferase activity for each sample , giving a normalized Gaussia luminescence . Each normalized Gaussia luciferase activity was calculated from the mean of triplicate samples . For a given pair of proteins ( A and B ) , the normalized Gaussia luminescence of cells coexpressing GL1-A+GL2-B proteins was divided by the sum of normalized Gaussia luminescence of each partner coexpressed with matched empty plasmid: GL1-A+GL2-B/ ( GL1-A +GL2 ) + ( GL1 + GL2-B ) . This gave a Normalized Luminescence Ratio ( NLR ) corresponding to the reconstituted Gaussia luciferase activity , thus reflecting the level of interaction between protein pairs . See [19] for further details on the method . Literature curated interaction ( LCI ) involving the E2 proteins were extracted from the VirHostNet [13] , virusMINT [14] and PubMed databases . Interaction data analyses were performed using the R statistics package . Raw NLR interaction data were separated into categories in order to minimize the dispersion of NLR values . Cut-off thresholds of each category were determined with the goal of maintaining the same frequency distribution across all categories . An Euclidian distance matrix was calculated from the data categories using the “dist” function from R . The interaction dendrogram was calculated using the “complete” ( UPGMA ) linkage method from the “hclust” function from R . E2 protein sequences were clustered using the “phylip” package [48] . Protein distances were calculated with the “prodist” program , using default parameters . The phylogenetic dendrogram was generated with the “neighbor” program using the UPGMA method and default parameters . Both interaction and phylogenetic dendrograms were generated using JavaTreeView [49] . A Pearson correlation coefficient was calculated with the “cor” function in R using the cophenetic distances between both interaction and phylogenetic dendrogram to determine the closeness of the two dendrograms , The label order for the intensity data was then randomly changed to generate 100 , 000 random dendrograms . The cophenetic distance matrix for these randomized dendrograms was compared to the cophenetic distance matrix from the phylogenetic dendrogram with a Pearson correlation ( “cor” ) function from R . The p-value was calculated based on the number of standard deviations the correlation between the interaction dendrogram and the phylogenetic dendrogram was from the mean of the distribution of the correlation between the random and the phylogenetic dendrogram . A Cumulative Density Function of the randomized dataset was compared to a normal distribution generated by the R function ‘rnorm’ using the same mean and standard deviation from the randomized dataset to check the normality of the data . The E2 interaction networks were generated with the cytoscape software [50] with interactions scoring positive in HT-GPCA ( NLR above 3 . 5 ) . The degree of each cellular protein in both E2 and HPRD-based human interactomes was extracted from cytoscape . To determine the overrepresented GO ( Gene Ontology ) terms in the interaction dataset and to evaluate the gathering of E2 targets by functional categories , we used the DAVID bioinformatic database [32] . P-values were generated by DAVID . 293T cells were plated at 35 , 000 cells per well in 96-well plates and transfected 24 h later by linear PEI with 25 ng of pTK6E2BS E2 responsive reporter plasmid , 10 ng of the polIII-Ren as internal control for transfection efficiency , and 100 ng of GL2-E2 fusion proteins or empty GL2 plasmid . To assess the effect of GTF2B , HeLa cells plated in 12-well plates were transfected by linear PEI with 100 ng of pTK6E2BS , 10 ng polIII-Ren , 100 ng of mCherry-fused E2 or mCherry expressing plasmids , and either 1 µg of GTF2B expressed from pCI Neo or of empty pCI Neo ( Promega ) . 30 h post transfection , cells were lysed in Passive lysis buffer according to manufacturer's instructions and luciferase activity was measured with Dual Glo Buffer ( Promega ) . Results are given as the mean of three independent tests ± SD ( errors bars ) . HaCaT cells grown in coverslip were co-transfected by linear PEI with expression plasmids for GFP-fused E2 proteins ( 3 µg ) and Cherry-fused cellular proteins ( 1 µg ) . 24 h post transfection , cells were fixed with 4% paraformaldehyde for 30 min , washed in PBS , and incubated with DAPI for 30 min . Cells were mounted with CitiFluor . Fluorescent Images were acquired using a ZEISS Apotome microscope . 7 , 500 HeLa cells were reverse transfected by INTERFERin ( Polyplus-Transfection ) with 1 . 75 picomole of a pool of four siRNA targeting GTF2B ( from Qiagen bank Human Whole Genome siRNA Set V4 . 1 ) , and plated in 96-well plates . 2 scrambled siRNA ( ref 1027310 , Qiagen ) were used as negative controls . 48 h later , 20 ng of Cherry-E2 expression plasmids were transfected by linear PEI along with the 25 ng of pTK6E2BS reporter and 10 ng of polIII-Ren as internal control for transfection efficacy and cell viability . 24 h post transfection , cells were lysed in passive lysis buffer according to manufacturer's instructions ( Promega ) . Firefly and Renilla luciferase were measured on a Berthold Centro luminometer to generate a Luciferase/Renilla ratio , each transfection was tested in triplicates with each bar representing the mean ± SD . Results are given as fold activation of TKE2BS by E2 in the presence of the siRNA , calculated relative to TKE2BS activity without E2 . P-values were calculated by a Student statistical test .
|
Over 100 types of human papillomaviruses are responsible for widespread infections in humans . They cause a wide range of pathologies , ranging from inapparent infections to benign lesions , hyperplasia or cancers . Such heterogeneity results from variable interplay among viral and host cell proteins . Aiming to identify specific features that distinguish different pathological genotypes , we mapped the virus-host interaction networks of the regulatory E2 proteins from a set of 12 genotypes representative of HPV diversity . The E2-host interaction profiles recapitulate HPV phylogeny , thus providing a valuable framework for understanding the role of E2 in HPV infection of different pathological traits . The E2 proteins tend to bind to highly connected cellular proteins , indicating a profound effect on the host cell . These interactions predominantly impact on a subset of cellular processes , like transcriptional regulation , apoptosis , RNA metabolism , ubiquitination or intracellular transport . This work improves the global understanding of HPV-associated pathologies , and provides a framework to select interactions that can be used as targets for the development of new therapeutics .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"systems",
"biology",
"virology",
"biology",
"microbiology",
"host-pathogen",
"interaction",
"genetics",
"and",
"genomics"
] |
2012
|
Large Scale Genotype Comparison of Human Papillomavirus E2-Host Interaction Networks Provides New Insights for E2 Molecular Functions
|
CD8+ T cell exhaustion represents a major hallmark of chronic HIV infection . Two key transcription factors governing CD8+ T cell differentiation , T-bet and Eomesodermin ( Eomes ) , have previously been shown in mice to differentially regulate T cell exhaustion in part through direct modulation of PD-1 . Here , we examined the relationship between these transcription factors and the expression of several inhibitory receptors ( PD-1 , CD160 , and 2B4 ) , functional characteristics and memory differentiation of CD8+ T cells in chronic and treated HIV infection . The expression of PD-1 , CD160 , and 2B4 on total CD8+ T cells was elevated in chronically infected individuals and highly associated with a T-betdimEomeshi expressional profile . Interestingly , both resting and activated HIV-specific CD8+ T cells in chronic infection were almost exclusively T-betdimEomeshi cells , while CMV-specific CD8+ T cells displayed a balanced expression pattern of T-bet and Eomes . The T-betdimEomeshi virus-specific CD8+ T cells did not show features of terminal differentiation , but rather a transitional memory phenotype with poor polyfunctional ( effector ) characteristics . The transitional and exhausted phenotype of HIV-specific CD8+ T cells was longitudinally related to persistent Eomes expression after antiretroviral therapy ( ART ) initiation . Strikingly , these characteristics remained stable up to 10 years after ART initiation . This study supports the concept that poor human viral-specific CD8+ T cell functionality is due to an inverse expression balance between T-bet and Eomes , which is not reversed despite long-term viral control through ART . These results aid to explain the inability of HIV-specific CD8+ T cells to control the viral replication post-ART cessation .
An effective CD8+ T cell response is required to eradicate or control intracellular pathogens . During the acute phase of an infection , pathogen-specific CD8+ T cells expand and differentiate into effector cells to clear the microbe . In the wake of antigen clearance , long-lived memory CD8+ T cells develop in order to launch an effective secondary response against future infections . Murine studies have indicated that the process of memory formation is highly regulated by the T-box transcription factors T-bet and Eomesodermin ( Eomes ) [1] , [2] , [3] . Although T-bet and Eomes are related transcription factors that show some expressional overlap , their functional roles are not entirely reciprocal . Whereas T-bet regulates the expression of effector functions , Eomes is thought to primarily dictate the expression of proteins to maintain a memory CD8+ T cell repertoire that effectively could expand in case of re-infection [4] , [5] , [6] , [7] . The current body of data thus suggest that the long-term fate of CD8+ T cell functionality and differentiation seems highly dictated by the expression ratio between T-bet or Eomes ( reviewed in [8] ) . Some viruses , including the human immunodeficiency virus type 1 ( HIV ) , evade the immune defense and develop into chronic infection . As a consequence , the pool of HIV-specific CD8+ T cells persists throughout the infection and become dysfunctional . This process has usually been referred to as CD8+ T cell exhaustion , which is characterized by a typical loss of different functions , including the ability to proliferate , kill target cells ( expression of cytotoxic molecules ) , and decreased IL-2 , TNF , and IFNγ production [9] , [10] . Initially , murine studies revealed that chronic lymphocytic choriomeningitis virus clone 13 ( LCMV-13 ) infection caused an up-regulation of PD-1 [11] and other inhibitory receptors , like CD160 , 2B4 , and Lag-3 , which cooperate to mediate CD8+ T cell dysfunction [12] . These findings were later extended to chronic human infections including HIV [13] , [14] , [15] , [16] , HCV [17] , [18] , [19] , [20] , and HBV [21] , [22] , [23] . In particular , HIV-specific CD8+ T cells have been studied with respect to dysfunctional characteristics , where seminal work have concluded that these cells in most individuals possess poor polyfunctionality [24] , [25] , and an immature/skewed maturation phenotype [26] , [27] . However , it remains unclear which transcriptional programming that governs the regulation of CD8+ T cell differentiation and exhaustion in HIV infection . Transcriptional networks have been related to CD8+ T cell exhaustion after LCMV-13 infection in mice [28] , and recently also in HIV-specific CD8+ T cells specifically expressing PD-1 [29] and CD160 [30] . Lately , Wherry and colleagues have elucidated in the murine model that an exhausted profile following LCMV-13 infection is associated with an inverse relationship between T-bet and Eomes [31] , [32] . Surprisingly , these studies showed that although T-bet caused terminal differentiation of CD8+ T cells , the transcription factor repressed expression of inhibitory receptors by direct binding to the promoter region of PD-1 . Eomes on the other hand was highly associated with expression of numerous inhibitory receptors . In a recent study , long-term non-progressors were shown to retain high T-bet expression within HIV-specific CD8+ T cells [33] . However , it is currently unknown whether the expression levels of T-bet and Eomes in human virus-specific CD8+ T cells are associated with the up-regulation of inhibitory receptors , poor polyfunctionality , and the skewed maturation phenotype described in HIV-infected individuals . To bring clarity to this , we examined the relationship between the expression of T-box transcription factors , markers of memory differentiation , and human viral-specific CD8+ T cell exhaustion at the single cell level . We provide evidence that HIV-specific CD8+ T cells in chronic infection largely possess highly elevated levels of Eomes , but lower T-bet expression . This relationship is associated with up-regulation of inhibitory receptors , impaired functional characteristics and a transitional memory differentiation phenotype . Importantly , these characteristics of HIV-specific CD8+ T cells remained stable despite suppressive ART for many years , and offer an explanation for the inability of CD8+ T cells to control viral replication post-ART cessation .
The Regional Ethical Council ( Stockholm , Sweden 2012/999-32 & 2009/1592-32 ) approved the study and all participants were provided with written and oral information about the study . Written informed consent was documented from all study subjects in accordance with the Declaration of Helsinki . All HIV-infected individuals were recruited from the Karolinska University Hospital Huddinge and Venhälsan at Stockholm South General Hospital ( Stockholm , Sweden ) . In total , 52 individuals with chronic untreated HIV infection and 12 HIV-infected individuals on ART for more than 10 years ( fully suppressed viral load for >8 years ) were enrolled in this study . Cell samples from 20 healthy controls were also collected from the Karolinska Institutet and Karolinska University Hospital Huddinge ( Table 1 ) . Out of the 52 individuals with chronic untreated HIV infection , 24 individuals were followed longitudinally from baseline ( median = 0 days before ART initiation ) and at 2 weeks , 4 weeks , 8 weeks , 12–16 weeks and 5–7 months post-ART initiation . All individuals initiated ART in chronic phase of infection and no one experienced virological failure ( i . e . >200 HIV RNA copies/mL after 6 months on therapy ) ( Table S1 ) . Peripheral blood mononuclear cells ( PBMCs ) were isolated from whole blood by Hypaque-Ficoll ( GE Healthcare ) density gradient centrifugation and cryopreserved in FBS ( Life Technologies ) containing 10% DMSO . For detection of HIV- and CMV-specific T cell responses , peptide pools ( 15-mers overlapping by 11 amino acids ) of HIV Gag-p55 ( JPT technologies ) and HCMV pp65 ( NIH AIDS Research and Reference Reagent Program ) were added to a final concentration of 1 µg/mL . All flow cytometry panels were tested on the HIV-infected and healthy control subjects within a 2-month time interval to avoid intra- and inter-individual differences of the flow analysis . The following antibodies were used: anti-CD3 APC-H7 ( Clone SK7 ) , anti-CD14 V500 ( Clone M5E2 ) , anti-CD19 V500 ( Clone B43 ) , anti-CD160 AF488 ( Clone BY55 ) , anti-CCR7 PE-Cy7 ( Clone 3D12 ) , anti-IFNγ AF700 ( Clone B27 ) , anti-TNF PE-Cy7 ( Clone MAb11 ) , anti-CD107a PE-CF594 ( Clone H4A3 ) , anti-HLA-DR BV605 ( Clone G46 ) ( BD Bioscience ) ; anti-2B4 PE-Cy5 . 5 ( Clone C1 . 7 ) , anti-CD45RO ECD ( clone UCHL1 ) ( Beckman Coulter ) ; anti-T-bet BV605 & -BV711 ( clone B10 ) , anti-CD4 BV650 & -BV785 ( Clone OKT4 ) , anti-PD-1 BV421 ( clone EH12 . 2H7 ) , anti-CD27 BV785 ( clone O323 ) , anti-CD38 APC ( clone HIT2 ) , anti-IL-2 BV605 ( Clone MQ1-17H12 ) , anti-Perforin BV421 ( clone B-D48 ) , anti-Granzyme A AF488 ( clone CB9 ) ( Biolegend ) ; anti-T-bet PE ( clone B10 ) , anti-Eomes EF660 ( clone WD1928 ) ( eBioscience ) ; anti-CD8 Qd565 ( Clone 3B5 ) , anti-CD4 PE-Cy5 . 5 ( Clone S3 . 5 ) , anti-Granzyme B PE-Cy5 . 5 ( Clone GB11 ) ( Life Technologies ) . LIVE/DEAD Aqua amine dye ( Life Technologies ) was used to discriminate dead cells and debris . MHC class-I tetramers and pentamers conjugated to PE ( Beckman Coulter and ProImmune ) were used to detect CD8+ T cells specific for HIV Gag SLYNTVATL ( SL9 ) /HLA-A*0201 , HIV Pol ILKEPVHGV ( IV9 ) /HLA-A*0201 and CMV pp65 NVLPMVATV ( NV9 ) /HLA-A*0201 . PBMCs were thawed and washed twice in R10 ( RPMI-1640 Medium AQmedia ( Sigma Aldrich ) containing 10% FBS , 50 IU/mL penicillin and 50 µg/mL streptomycin , 10 mM HEPES ( Life Technologies ) ) . Cells were counted on a Nucleocounter ( ChemoMetec A/S ) , resuspended to 2×106 cells/mL in R10 containing 10 U/mL DNase I ( Roche Diagnostics ) and rested for 5–6 hours at 37°C . U-bottom plates were plated with 5 µg/mL Brefeldin A ( Sigma Aldrich ) and overlapping HIV Gag-p55 and HCMV pp65 peptides or medium alone ( negative controls ) together with 2×106 PBMCs/well . When anti-CD107a was added at the start of the stimulation protocol [34] , monensin ( 0 . 7 µg/mL , BD Bioscience ) was also supplemented . The PBMCs were next transferred to V-bottom plates , washed in PBS containing 2 mM EDTA and stained with a LIVE/DEAD Aqua amine dye solution containing the extracellular antibodies . Cells were first incubated for 10 minutes at 37°C with an anti-CCR7 antibody and then for further 20 minutes in room temperature with other extracellular antibodies . PBMCs were washed in PBS:EDTA and fixed and permeabilized using the FoxP3 transcription factor buffer kit ( eBioscience ) . Monoclonal antibodies against T-bet , Eomes and other intracellular markers were incubated with the cells for 1 hour at room temperature in the dark . After further washing with Perm Wash solution ( eBioscience ) , the PBMCs were resuspended in PBS containing 1% paraformaldehyde PFA . All specimens were acquired on the flow cytometer within the next 7 hours . PBMCs were analyzed on a modified 4 laser LSR Fortessa ( BD Biosciences ) . In total , approximately 1 , 000 , 000 events were collected per specimen . Antibody capture beads ( BD Biosciences ) were used to prepare individual compensation controls after separate stainings with all antibodies used in the experiments . FlowJo 8 . 8 . 7 ( Treestar ) was used for flow cytometric gating analyses . Most manual gatings were based on fluorescence minus one ( FMO ) gating strategies like previously described [35] , [36] . A response was considered positive if the frequency of IFNγ producing cells were >0 . 05% of total CD8+ T cells after background reduction and twice the negative background . Plasma was extracted from whole blood and stored in −80°C before analysis of soluble factors . Concentrations of cytokines in plasma were analyzed by ELISA using the following reagents: Ready-Set-Go , eBioscience ( TNF , IL-6 ) , Matched Antibody Pairs , eBioscience ( IFNα ) and HS Quantikine , RnD Systems ( IL-12p70 ) according to the manufacturer's specifications . Experimental variables between two groups of individuals were analyzed using Mann-Whitney U test and Wilcoxon matched-pairs rank test . Correlations were assessed using non-parametric Spearman rank tests . Bonferroni corrections were applied to all cases where multiple testing was performed . One-Way ANOVAs followed by Kruskal-Wallis non-parametric Dunn's multiple comparison tests were used to analyze three groups or more . The large data set consisting of functional and inhibitory receptor frequencies from the Boolean combinations was analysed by principal component analysis ( PCA ) , an unsupervised statistical method for reducing data dimensionality while retaining the vital variation in fewer informative variables . The top 2 principle components ( PCs ) were plotted to visualize the trends such as clusters and outliers revealed by PCA . The Kolmogorov-Smirnov test was used to test the null hypothesis that the two groups of samples were drawn from the same distribution , where a large p-value suggests that the groups were drawn from the same distribution . All the statistical analyses were performed using GraphPad Prism 5 . 0 and R environment [37] . Permutation tests were analyzed using the data analysis program SPICE version 5 . 2009 [38] .
Previous studies have identified numerous inhibitory receptors , including PD-1 , CD160 , 2B4 , LAG-3 , CTLA4 , and Tim-3 , with increased expression patterns after chronic viral infections [13] , [14] , [15] , [16] , [39] , [40] . Due to low and partly negligible expression of LAG-3 , CTLA4 and Tim-3 on virus-specific CD8+ T cells ( data not shown and [16] ) , we here focused on studying the expression of PD-1 , CD160 and 2B4 on memory CD8+ T cells in HIV-infected individuals compared to healthy controls , and in relation to T-bet and Eomes expression . As expected , we found increased levels of CD8+ T cells mono- and co-expressing PD-1 , CD160 , and 2B4 in our cohort of 52 untreated HIV-infected individuals , compared to long-term treated ( >10 years ) HIV-infected subjects and healthy controls ( Figure 1A and Figure S1A ) . However , no significant differences in frequency of CD8+ T cells expressing the inhibitory receptors was observed between the long-term treated HIV+ subjects and healthy controls ( Figure 1A ) . We next examined whether the expression pattern of T-bet and Eomes was linked to mono- and co-expression of PD-1 , CD160 , and 2B4 within the total CD8+ T cell population . A close relationship of increased expression of all inhibitory markers was observed for T-betdim and Eomeshi CD8+ T cells ( Figure S1B ) , thus corroborating murine studies of LCMV-specific CD8+ T cells [31] . We used a gating strategy similar to that recently described on CMV-specific CD8+ T cells [41] , and gated on T-bethiEomesdim and T-betdimEomeshi CD8+ T cells ( Figure 1B ) . In all untreated HIV-infected subjects , the T-bethiEomesdim population contained significantly fewer cells with mono- and co-expression patterns of PD-1 , CD160 , and 2B4 , as compared to T-betdimEomeshi cells ( Figure 1B–C ) . By gating on PD-1+CD160+2B4+ CD8+ T cells , we found that the vast majority ( median = 91 . 7% ) of these cells were T-betdimEomeshi ( Figure 1D ) . Further corroborating this link , we discovered that CD8+ T cells expressing inhibitory receptors was significantly associated with a T-betdimEomeshi phenotype ( Figure 1E ) . Notably , similar associations were also observed for individuals on long-term ART and matched healthy controls ( Figure S1C ) . Together , these results indicate that an inverse expression pattern of T-bet and Eomes is highly associated with the up-regulation of several inhibitory receptors for total CD8+ T cells , independently of HIV infection status . As bulk CD8+ T cell expression of PD-1 , CD160 , and 2B4 was highly correlated with the T-betdimEomeshi population , we next examined whether this relationship could be attributed to HIV- and other human viral-specific CD8+ T cells in general . We therefore stimulated PBMCs with overlapping peptide pools spanning HIV Gag-p55 and HCMV-pp65 to identify IFNγ/TNF-producing HIV- and CMV-specific CD8+ T cells , respectively [36] , [42] . CMV-specific CD8+ T cell responses were selected as a control , since these responses have previously been shown to be highly polyfunctional and express less inhibitory receptors [16] , [43] , [44] . In Figure 2A , an example is shown for typical HIV- and CMV-specific IFNγ+/TNF+ responses . Notably , HIV-specific CD8+ T cells ( n = 50 ) were almost exclusively T-betdimEomeshi , while CMV-specific CD8+ T cells ( n = 46 ) could be detected in both the T-betdimEomeshi and T-bethiEomesdim populations in all untreated HIV-infected individuals with chronic disease ( Figure 2A–B ) . The median fluorescence intensity ( MFI ) of Eomes and T-bet , was also significantly different between HIV- and CMV-specific CD8+ T cells ( Figure S2A ) . Consistently , both the frequency and MFI of single and co-expression of PD-1 , CD160 , and 2B4 were elevated on HIV-specific compared to CMV-specific CD8+ T cells ( Figure 2C and Figure S2B ) . We further gated on PD-1+CD160+2B4+ HIV-specific CD8+ T cells and found that this population almost exclusively were occupied within the T-betdimEomeshi population ( Figure 2D ) . By employing a SPICE analysis on HIV-specific CD8+ T cells , we further confirmed that co-expression of the inhibitory receptors were significantly elevated in the T-betdimEomeshi population compared to T-bethiEomesdim cells ( P<0 . 001; Figure 2E ) . In further analysis we compared the relationship between expression of T-bet , Eomes , and inhibitory receptors among unmanipulated ( resting ) virus-specific CD8+ T cells . To this end , we directly examined HIV Gag SLYNTVATL ( SL9 ) and CMV pp65 NVLPMVATV ( NV9 ) specific CD8+ T cells in 5 HLA-A*0201+ donors , identified previously [45] , by HLA class-I tetramer ( tet ) analysis . In agreement with our previous results , the HIV/SL9-tet+ cells had significantly higher levels of Eomes , but lower T-bet expression than CMV/NV9-tet+ cells ( Figure 3A–B ) . The frequencies of CD8+ T cells mono- and co-expressing PD-1 , CD160 , and 2B4 were highly elevated among HIV/SL9-tet+ cells , compared to CMV/NV9-tet+ T cells ( Figure 3C ) . Combining both HIV/SL9-tet+ and CMV/NV9-tet+ cells in one correlation analysis , we found an inverse correlation ( P = 0 . 009 , r = −0 . 79 ) between the MFI of T-bet and PD-1+CD160+2B4+ co-expression ( Figure 3D ) . Similarly , a positive strong association ( P = 0 . 006 , r = 0 . 82 ) was observed between the MFI of Eomes and PD-1+CD160+2B4+ co-expression ( Figure 3D ) . Together , these results show that human virus-specific CD8+ T cell regulation of inhibitory receptors is associated with high expresison of Eomes , but low T-bet expression . Co-expression of PD-1 , CD160 , and 2B4 has been closely linked to limited T cell functionality in mice and humans [12] , [16] . Therefore , we next examined how differential expression of T-bet and Eomes was related to the functional properties of HIV- and CMV-specific CD8+ T cells in the context of these inhibitory markers . We assessed the expression pattern for IFNγ , TNF , IL-2 , CD107a and Granzyme B simultaneously for both HIV- and CMV-specific CD8+ T cell responses in 23 untreated individuals with chronic HIV infection . All individuals generated responses , except one with lack of an anti-CMV response . PD-1+ HIV-specific IFNγ+ CD8+ T cells were found to be T-betdimEomeshi , whereas PD-1− HIV-specific IFNγ+ CD8+ T cells were T-bethiEomesdim ( Figure 4A ) . The PD-1+ T cell response showed low co-expression of IFNγ with either TNF or Granzyme B . Interestingly however , we found that expression of CD107a together with IFNγ was more substantial in PD-1+ cells compared to PD-1− cells . PD-1− cells were more likely to simultaneously express IFNγ and TNF or Granzyme B together with CD107a . Using multi-parametric flow cytometry analysis , we combined the measurment of all inhibitory receptors ( PD-1 , CD160 , 2B4 ) together with the functional markers ( IFNγ , TNF , IL-2 , CD107a , Granzyme B ) using Boolean gating principles ( n = 256 T cell populations ) . The populations that were occupied in any HIV-infected subjects ( n = 200 ) were further assessed using PCA analysis to determine whether HIV- and CMV-specific CD8+ T cells could be differentiated based on the combination of these markers ( Figure 4B ) . A significant separation between the HIV- and CMV-specific responses was observed in terms of the PC1 and PC2 dimensions ( Figure 4B ) and confirmed by a Kolmogorov-Smirnov ( KS ) test , like previously described [46] . The KS test determines the probability that the two groups were drawn from the same distribution , where a P-value of 1 signifies that the groups were drawn from the same distribution . Particularly , the distributions of the PC2 scores for the two groups were shown to be derived from different distributions ( D = 0 . 60 , P<0 . 001 ) . Additionally , the PC1 scores also showed significant differences between the groups ( D = 0 . 46 , P = 0 . 01 ) . We next evaluated specific correlations between functional markers and the T-bet/Eomes axis . The MFI of Eomes in HIV-specific CD8+ T cells was negatively associated with the co-expression of IFNγ , TNF , CD107a and Granzyme B ( P<0 . 001 , r = −0 . 72 ) and positively associated with single CD107a production ( P = 0 . 004 , r = 0 . 57; Figure 4C ) . The MFI of T-bet in HIV-specific CD8+ T cells was primarily correlated to co-expression of IFNγ , TNF and Granzyme B ( P = 0 . 011 , r = 0 . 52 ) ( data not shown ) . Next , we assessed whether specific functional combinations were differentially distributed between the T-betdimEomeshi and T-bethiEomesdim HIV-specific CD8+ T cells by SPICE analysis . These analyses confirmed a polyfunctional diversity between the groups ( P<0 . 001 , permutation test ) and that single production of CD107a ( P<0 . 001 ) , or together with IFNγ ( P = 0 . 012 ) , was elevated for T-betdimEomeshi cells . In contrast , co-production of IFNγ , TNF and CD107a ( P = 0 . 027 ) and IFNγ , CD107a and Granzyme B ( P = 0 . 005 ) was expressed at higher frequencies by T-bethiEomesdim cells ( Figure 4D ) . We found similar associations for CMV-specific CD8+ T cells , showing that T-betdimEomeshi cells particularly expressed high single production of CD107a , or together with IFNγ , while diverse combinations of all cytokines/cytotoxins were elevated within the T-bethiEomesdim population ( Figure S3 ) . Altogether , our data suggests that increased expression of Eomes is linked to a profile of increased exhaustion for both HIV- and CMV-specific CD8+ T cells . In order to further assess the link between T-bet and Eomes with cytolytic functions , we recruited 5 HLA-A*0201+ donors with identified HIV Pol ILKEPVHGV ( IV9 ) - , HIV/SL9- , and CMV/NV9-specific CD8+ T cell responses . The effector functions were examined using tetramers or pentamers and cognate epitope stimulations . In agreement with the overlapping peptide stimulations , we found that resting HIV-SL9/IV9-tet+ cells showed increased expression of Eomes that was coupled to lower ex vivo expression of Granzyme B and perforin compared to CMV/NV9-tet+ cells . However , most HIV-SL9/IV9-tet+ cells were found to have high expression levels of Granzyme A ( Figure S4A–B ) . Correlation analysis confirmed strong associations between the frequencies of cytolytic markers ( perforin and Granzyme B ) with T-bet/Eomes MFI in virus-specific tet+ cells ( Figure S4C ) . Analysis on bulk CD8+ T cells further supported that perforin+ and Granzyme B+ cells were primarily T-bethi cells , while Granzyme A were expressed both within the T-bethi and Eomeshi compartments ( Figure S4D ) , thus clarifying the high Granzyme A content of HIV-tet+ cells . Cognate peptide stimulations additionally revealed that HIV-SL9/IV9-epitope specific CD8+ T cells , independently of whether they were bi- or monofunctional for IFNγ and/or CD107a , showed high expression levels of Eomes , but variable cytolytic content ( Figure S4E ) . Interestingly , IFNγ+CD107a− epitope-specific cells showed increased signs of perforin , Granzyme B and Granzyme A expression compared to IFNγ-CD107a+ and IFNγ+CD107a+ cells ( Figure S4F ) . These analyses further revealed that some HIV epitope-specific IFNγ-CD107a+ cells contained Granzyme A and B , but only in a minor fraction of the cells , which suggest that monofunctional CD107a+ cells might be highly exhausted ( Figure S4E–F ) . We further traced the expression of the inhibitory receptors to diverse memory phenotypes using CD45RO , CD27 and CCR7 in the untreated HIV-infected subjects . The composition of bulk PD-1+CD160+2B4+ CD8+ T cells was particularly elevated within the transitional memory ( TM; CD45RO+CD27+CCR7− ) phenotype compartment ( Figure 5A ) as previously described [16] . Consistently , increased co-expression of the inhibitory receptors was associated with a higher frequency of TM cells , but not terminally-differentiated effector cells ( Eff; CD45RO−CD27−CCR7− ) ( Figure S5A ) . PD-1+ and CD160+ cells were primarily found in the TM compartment , while 2B4+ cells were mainly effector memory ( EM; CD45RO+CD27−CCR7− ) and Eff cells . We next evaluated the phenotypic composition of T-bet and Eomes expressing cells and as expected found that T-betdimEomeshi expressing cells were enriched and strongly associated with a transitional memory phenotype ( Figure 5B and Figure S5B ) . Conversely , T-bethiEomesdim expression was associated with increased EM ( P = 0 . 032 , r = 0 . 30 ) and particularly Eff ( P<0 . 001 , r = 0 . 69 ) cell compartmentalization ( Figure S5C ) . In agreement with other studies , we next confirmed that the majority of HIV-specific CD8+ T cells ( median = 58% ) accumulated in the TM compartment ( Figure S5D ) [26] , [27] . CMV-specific CD8+ T cells on the other hand showed a balanced expression of T-bet and Eomes ( Figure 2B ) and were thus less encompassed to the TM compartment ( Figure S5E ) . Similar to before , these results were further confirmed using HIV/SL9- and CMV/NV9-tetramers . The HIV/SL9-tet+ cells were indeed enriched within the TM compartment , while CMV/NV9-tet+ cells displayed a higher proportion of EM cells ( Figure 5C ) . We next confirmed that expression levels of Eomes were positively associated ( P = 0 . 04 , r = 0 . 66 ) with the proportion of TM virus-specific CD8+ T cells , where as it was inversely correlated ( P = 0 . 01 , r = −0 . 76 ) with the frequency of EM virus-specific CD8+ T cells ( Figure 5D ) . These data suggest that HIV-specific CD8+ T cell expression of numerous inhibitory receptors is linked to a transitional differentiation phenotype , displaying a T-betdimEomeshi phenotype . Inflammatory cytokines have previously been demonstrated to inversely regulate the expression levels of T-bet and Eomes in memory CD8+ T cells [5] , [47] , [48] . Therefore , we next aimed to determine whether the T-betdimEomeshi or T-bethiEomesdim expression profile of bulk CD8+ T cells was associated with the levels of IL-12p70 , IFNα , TNF and IL-6 in untreated HIV infection ( n = 38 ) . Interestingly , we found tendencies of an inverse association pattern between T-bet and Eomes in terms of the inflammatory cytokine levels . While T-bet expression in general was positively associated with the cytokine plasma levels , Eomes instead was inversely associated with particularly the levels of IFNα ( P = 0 . 015 , r = −0 . 39 ) and TNF ( P = 0 . 01 , r = −0 . 41 ) ( Figure S6 ) . These data might implicate that bystander inflammation influence the bulk CD8+ T cell repertoire towards increased T-bet or Eomes expression in untreated HIV infection . HIV-specific CD8+ T cell compartmentalization within the T-betdimEomeshi or T-bethiEomesdim population was not associated with the levels of any inflammatory cytokine ( data not shown ) . To determine whether the viral load had a direct effect on the expression patterns of Eomes , T-bet and the inhibitory receptors , we analyzed samples collected before and at 2 , 4 , 8 , 12–16 weeks , and 5–7 months after ART initiation . These longitudinal analyses also allowed us to determine the kinetics of expression of these markers after succesful ART in 24 individuals . Before ART , the CD4 % was inversely associated ( P = 0 . 015 , r = −0 . 38 ) with the frequency of T-betdimEomeshi cells ( data not shown ) . Following initiation of ART , the frequency of T-betdimEomeshi cells in total CD8+ T cells progressively declined in most individuals . In contrast , the frequency of T-bethiEomesdim cells remained stable during the longitudinal assessment ( Figure 6A ) . The decreased frequency of T-betdimEomeshi cells was related to a similar decay of the PD-1+CD160+2B4+ population longitudinally after ART ( Figure 6A ) , where area-under-curve measurements confirmed a close longitudinal relationship between these populations ( P<0 . 001 , r = 0 . 71; Figure S7A ) . We next assessed whether the decay of the T-betdimEomeshi expressing CD8+ T cell population was a consequence of fading magnitudes of HIV-specific CD8+ T cells . However , the absolute decline of the T-betdimEomeshi population was more pronounced ( mean = 7 . 2% ) than the loss of HIV-specific cells ( mean = 0 . 76%; Figure 6B ) . Although the T-betdimEomeshi population was associated ( P = 0 . 003 , r = 0 . 55 ) with the levels of CD8+ T cell activation ( CD38+HLA-DR+ ) at baseline ( Figure S7B ) , the decay of immune activation was not associated ( P = 0 . 41 , r = 0 . 18 ) with the fading T-betdimEomeshi population longitudinally based on area-under-curve measurments ( data not shown ) . Instead , we found that the decay of T-betdimEomeshi cells was proportional to the recruitment of naïve CD8+ T cells after 6 months on ART ( P = 0 . 026 , r = −0 . 45; Figure 6C ) , indicating that the expression of Eomes in total CD8+ T cells might represent a balance between the frequency of naïve versus transitional memory CD8+ T cells . We next analyzed the expression pattern of inhibitory receptors on the HIV-specific CD8+ T cells longitudinally after ART initiation . The MFI of CD160 and 2B4 remained stable after ART initiation , while expression levels of PD-1 declined in a hierarchical manner from baseline until 6 months post-ART initiation ( P<0 . 001 ) . The pattern of PD-1 expression on HIV-specific CD8+ T cells was related to the decay in Eomes MFI after ART initiation ( Figure 6D ) . Despite this relationship between PD-1 and Eomes , we found no significant decline ( P>0 . 05 ) in the co-expression of PD-1 , CD160 , and 2B4 on HIV-specific CD8+ T cells from before and 5–7 months after ART initiation ( Figure 6E ) . This was linked to a steady T-betdimEomeshi expression that persisted over the entire study period ( Figure 6E ) . Notably , all individuals had fully suppressed viral load at 5–7 months , and in an attempt to clarify whether viremia had any effect on T-bet/Eomes or inhibitory receptors , we divided individuals into two groups based on whether the subjects had detectable or undetectable viremia after 12–16 weeks on ART . However , no significant differences were distingushiable between these groups in terms of HIV-specific T-betdimEomeshi compartmentalization ( P = 0 . 28 ) or PD-1 , CD160 , and 2B4 co-expression ( P = 0 . 11; Figure S7C ) . Similarly , despite >10 years on ART ( undetectable viral load for >8 years ) , the residual HIV-specific CD8+ T cells remained trapped within the T-betdimEomeshi compartment ( Figure 6F ) . Individuals on long-term ART showed tendencies of lower co-expression of the inhibitory receptors compared to individuals at baseline or on ART for 6 months ( Figure 6G ) . However , the co-expression frequencies of the anti-CMV response for the long-term treated individuals was significantly lower than the anti-HIV response both pre- and post-ART ( Figure 6G ) . Consistently , a majority of HIV-specific CD8+ T cell remained in the transitional memory compartment despite long-term therapy ( Figure 6H ) and showed similar functional characteristics , except increased Granzyme B expression pre-ART , before and after 6 months on ART ( Figure 6I ) . These results together suggest that viral control by ART for several years does not change the phenotype of HIV-specific CD8+ T cells , which in turn might be due to retained elevated levels of Eomes expression .
T-bet and Eomes regulate the differentiation process of CD8+ T cells following encounter with a foreign antigen . Despite the central role of T-bet and Eomes in differentiation fate decisions , little is known about their influence on human CD8+ T cell responses [49] . Here , we show that human bulk and virus-specific CD8+ T cells expressing high levels of Eomes and low levels of T-bet , express several inhibitory receptors , display a transitional differentiation phenotype , exhibit poor functional abilities , and increased immune activation . This phenotype was unique to HIV-specific CD8+ T cells as compared to CMV-specific CD8+ T cells , and aid in explaining the HIV-specific CD8+ T cell subset's inability to clear or control the chronic viral replication . Murine studies have demonstrated that T-bet and Eomes are centrally connected hub transcription factors that regulate exhaustion and memory fates of CD8+ T cells [28] . The negative regulation of inhibitory receptor expression by T-box transcription factors in CD8+ T cells was first demonstrated to involve direct binding of T-bet to the promotor region of PD-1 , and increased expression of other inhibitory receptors , like CD160 and 2B4 , in T-bet knock-out models [32] . Further analysis confirmed that T-bet deficiency in chronic LCMV infection also impact other neighbouring genes that are directly involved in CD8+ T cell differentiation and exhaustion , like Tigit , Itgam and IL-18Ra expression [28] . On the contrary , CD8+ T cell deficiency of Eomes leads to diminished PD-1 and Blimp-1 expression , as well as increased IFN-γ and TNF co-production following chronic LCMV infection , implicating that T-bet and Eomes regulate diverse sub-populations of CD8+ T cells [31] . In our study , T-bet and Eomes were expressed in distinct patterns and linked to the exhausted phenotype observed in the murine studies . The loss of polyfunctional HIV-specific CD8+ T cells was associated with increased expression levels of Eomes , and down-regulation of T-bet . These results confirm previous findings by Hersperger et al , demonstrating decreased expression levels of T-bet in chronic HIV progressors [33] . We also see tendencies that an increased fraction of viremic controllers ( VL<1000 ) show increased signs of T-bet expression , but due to the low number of these subjects ( n = 9 ) , no significant correlations were detected between VL and T-bet/Eomes compartmentalization of HIV-specific CD8+ T cells in this cohort . Nevertheless , given that recent studies have shown that T-bet expression is co-localized both in the nucleus and cytoplasm [50] , our results further suggest that HIV-specific CD8+ T cells in HIV chronic progressors might have high abundance of T-bet in the cytoplasm , leading to the inability to repress the expression of inhibitory receptors and induce cytolytic functions . However , a recent study by Ribeiro-dos-Santos et al concluded that loss of cytolytic potential of HIV-specific CD8+ T cells was a consequence of both T-bet and Eomes down-regulation in the chronic phase of HIV infection [51] . The discrepancies between this study and our results here are most likely due to the use of different techniques to observe the levels of transcription factors . While Ribeiro-dos-Santos et al isolated resting cell populations and used gene expression analysis to detect mRNA levels , we analyzed transcription factor protein expression on a single-cell level using flow cytometry . Both the peptide-stimulation protocol , as well as the tetramer staining for virus-specific T cells suggested that most HIV-specific CD8+ T cells had a T-betdimEomeshi phenotype , indicating that stimulation with HIV antigens per se did not up-regulate Eomes expression in our analysis . Furthermore , immunoblot analysis have previously confirmed that T-bet and Eomes expression in the nucleus and cytoplasm resembles what is distinguishable with flow cytometry [50] . The lack of cytolytic functions of HIV-specific CD8+ T cells in chronic progressors has previously been observed in numerous studies [52] , [53] , [54] . Interestingly , we found a close correlation between the expression intensity of Eomes and monofunctionality measured as upregulation of CD107a . Although the CD107a single positive cells might represent highly exhausted cells , HIV-specific CD8+ T cells have been reported to contain high levels of Granzyme A [43] , which partly was confirmed in our tetramer and cognate epitope stimulation assays . In addition , Eomes has been shown to also directly drive effector CD8+ T cell differentiation in T-bet knock-out mice [3] and therefore , we cannot conclude that Eomes high cells fail to deliver effector molecules . However , in experiments using effector ( virus-specific ) CD8+ T cells and target cells , the frequency of Granzyme B and perforin positive CD8 T cells is proportional to the percentage of target cells being lysed , while Granzyme A , -K and CD107a are not necessarily associated with specific lysis of cells [43] , [52] , [55] . T-bet has previously also been shown , in chromatin immunoprecipitation coupled with microarray analysis ( ChIP-ChIP ) , to bind and promote increased gene expression of Granzyme B and perforin in T cells [56] . These data therefore suggest that HIV-specific CD8+ T cells in chronic progressors do not contain the proper cytotoxic granules to kill virus infected cells potentially due to poor up-regulation of T-bet . The process of CD8+ T cell exhaustion is thought to be a consequence of increased antigen load , inflammation , and other events that drive the cells to an end-stage of their life cycle , where they lose the ability to induce effector functions . This hypothesis is supported by the fact that chronically high antigen levels cause T cell exhaustion during chronic viral infections [57] . In addition , HIV-infected individuals with low viral load generally have higher T cell polyfunctionality [25] and lower expression of inhibitory receptors on CD4+ T cells [58] . However , not all elite controllers show low expression of inhibitory receptors ( unpublished observations ) and despite viral control by ART , the functional chracteristics of HIV-specific CD8+ T cells are not fully restored [25] , [53] , [59] . Whether the state of exhaustion therefore is directly proportional to antigen burden and only due to chronicity is therefore not entirely clear . Instead , previous studies have shown that elevated HIV-specific CD8+ T cell expression of PD-1 occurs during early HIV infection and remains high in the chronic phase [60] , possibly due to immune activation and the inability of CD8+ T cells to rest ( M . R . B . /M . B . unpublished observations ) . We , and others [61] , have found associations between the state of immune activation ( CD38+HLA-DR+ ) and Eomes , indicating that Eomes is up-regulated early after HIV infection as a consequence of immune activation , sustaining the expression of inhibitory receptors . However , monocyte activation in terms of sCD14 was not associated with increased expression of Eomes ( data not shown ) , and no relationship between Eomes and CD38+HLA-DR+ CD8+ T cells was observed after ART initiation in longitudinal measurments . Whether the expression profile of increased Eomes and lower T-bet is a consequence , or cause , of chronic immune activation is therefore hard to determine . The associations between inflammatory cytokine levels and T-bet/Eomes expression , also suggest that the level of inflammation might influence the expression ratio between these transcription factors like previously observed in mice [5] , [47] , [48] . Nevertheless , our results strongly suggest at least that exhaustion of human viral-specific CD8+ T cells is not associated with T-bet up-regulation and thus terminal differentiation . These observations support previous studies showing a lack of association between markers of T cell exhaustion and terminal-differentiation ( KLRG1 and CD57 ) in chronic LCMV and HIV infection [62] , [63] . In agreement with these observations , our data suggest that less differentiated ( TM ) T cells show increased signs of exhaustion in comparison with EM T cells , possibly due to the reverse actions of T-bet and Eomes promoting diverse effector functions in the differentiation machinery . Further molecular delineation , and also identification of other transcription factors , that regulate T cell differentiation and exhaustion will therefore be informative for launching effective memory CD8+ T cell responses against persistent infections , like HIV , in future therapeutic settings . Despite over 10 years on therapy , residual HIV-specific CD8+ T cells showed high expression levels of Eomes , inhibitory receptors and an intermediate differentiation status . The sustained co-expression of inhibitory receptors on HIV-specific CD8+ T cells was surprising as other studies have distinguished decreased PD-1 and CD160 expression after ART administration [16] . However , the longitudinal follow-up in this study was shorter ( 6 months post ART initiation ) and after assesing the combined expression of PD-1 , CD160 , and 2B4 in individuals with >10 years on ART we found that these levels were lower . The longitudinal assessment of individuals initiating ART , nevertheless provided evidence that most HIV-specific CD8+ T cells showed elevated levels of inhibitory receptors and sustained expression levels of Eomes , compared to CMV-specific CD8+ T cells . Similarily , despite ART administration for 6 months , the HIV-infected subjects showed no significant improvement of functional characteristics compared to pre-ART . All of these individuals where however treated in chronic infection , and studies on patients initiating ART very early following infection would add valuable information on the dynamics of these responses . The persistent exhausted phenotype of virus-specific CD8+ T cells has also been observed in mice after antigen removal [64] and might be a consequence of unmethylated promotor regions of inhibitory pathways [65] . In addition to these data , our findings are supported by previous results showing that ART introduction or in vitro stimulations do not increase T-bet expression in HIV-specific CD8+ T cells [33] , potentially through the reverse actions of Eomes . Altogether these data suggest that the exhausted phenotype is imprinted in HIV-specific CD8+ T cells during chronic infection , and remains stable several years with undetectable viral load . In summary , we here show that HIV-specific CD8+ T cells retain an inverse expression pattern between T-bet and Eomes , which is highly associated with an exhausted phenotype . CMV-specific CD8+ T cells instead show increased signs of T-bet expression that potentially leads to the effector memory responses akin to those described by Hansen et al that control [66] or clear pathogenic SIV infection [67] . The sustained expression of Eomes and inhibitory receptors in/on HIV-specific CD8+ T cells after long-term ART further suggest that therapeutic strategies aimed at reinvigorating these responses might fail to elicit efficient responses to eradicate the viral reservoir . Future HIV vaccine or cure approaches most probably need to overcome this transcriptional barrier and induce sustained T-bet expression in order to clear virus infected cells .
|
CD8+ T cells display numerous traits of severe dysfunction in both treated and untreated HIV infection . Previous studies have demonstrated that HIV-specific CD8+ T cells in most individuals possess poor polyfunctionality , and an immature/skewed maturation phenotype . However , it remains unclear which transcriptional programming governs the regulation of CD8+ T cell differentiation and exhaustion in HIV infection . T-bet and Eomes represent two key transcription factors for CD8+ T cell differentiation and function , but surprisingly little is known about their influence of effector immunity following chronic viral infections in humans . In this study , we demonstrate that HIV-specific CD8+ T cells possess highly elevated levels of Eomes , but low T-bet expression . This differential relationship is linked to the up-regulation of several inhibitory receptors , impaired functional characteristics and a transitional memory differentiation phenotype for virus-specific CD8+ T cells . Importantly , these characteristics of HIV-specific CD8+ T cells remained stable despite suppressive ART for many years . These results implicate that reinvigoration of these cells might fail to elicit efficient responses to eradicate the viral reservoir .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immunology",
"microbiology",
"immunodeficiency",
"viruses",
"infectious",
"diseases",
"white",
"blood",
"cells",
"animal",
"cells",
"medical",
"microbiology",
"hiv",
"t",
"cells",
"microbial",
"pathogens",
"immune",
"response",
"cell",
"biology",
"clinical",
"immunology",
"immunity",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"viral",
"diseases"
] |
2014
|
T-bet and Eomes Are Differentially Linked to the Exhausted Phenotype of CD8+ T Cells in HIV Infection
|
Disordered protein-protein interactions ( PPIs ) , those involving a folded protein and an intrinsically disordered protein ( IDP ) , are prevalent in the cell , including important signaling and regulatory pathways . IDPs do not adopt a single dominant structure in isolation but often become ordered upon binding . To aid understanding of the molecular mechanisms of disordered PPIs , it is crucial to obtain the tertiary structure of the PPIs . However , experimental methods have difficulty in solving disordered PPIs and existing protein-protein and protein-peptide docking methods are not able to model them . Here we present a novel computational method , IDP-LZerD , which models the conformation of a disordered PPI by considering the biophysical binding mechanism of an IDP to a structured protein , whereby a local segment of the IDP initiates the interaction and subsequently the remaining IDP regions explore and coalesce around the initial binding site . On a dataset of 22 disordered PPIs with IDPs up to 69 amino acids , successful predictions were made for 21 bound and 18 unbound receptors . The successful modeling provides additional support for biophysical principles . Moreover , the new technique significantly expands the capability of protein structure modeling and provides crucial insights into the molecular mechanisms of disordered PPIs .
Intrinsically disordered proteins ( IDP ) , which have evolved to not adopt a stable structure under physiological conditions , are a departure from the traditional paradigm of structured proteins [1] . After initial recognition of their critical biological functions in the 1990s [1] , IDPs quickly gained attention as they were found to be abundant in genomes across all three kingdoms [2] . IDPs are known to be involved in many molecular recognition events . Particularly , it is estimated that 15–45% of protein-protein interactions ( PPIs ) are formed with IDPs [3] . A well-known example is the p53 tumor suppressor , which contains disordered regions that interact with dozens of partner proteins [4] . Due to the abundance and characteristic features of IDPs in PPI networks , including many critical signaling pathways , fully understanding the molecular mechanisms of PPI networks requires consideration of the role of interactions with IDPs . The binding mechanism of an IDP to a structured target protein , i . e . a disordered PPI , has drawn much interest in the context of binding rate constants , because disordered PPIs achieve high specificity and high dissociation rate constant simultaneously , which is an ideal characteristic for signaling pathways but difficult to realize with interactions of structured proteins [5] . It is generally accepted that binding precedes global folding of the IDP , although secondary structures in local regions may form before the interaction . In the model called the dock-and-coalesce [5] , a small segment of the IDP , which may be folded into secondary structure prior to binding , forms the initial contact with the ordered partner , followed by coalescence of the rest of the IDP into the bound conformation . This mechanism imparts both thermodynamic and kinetic advantages . Forming a binding interface out of segments leads to a large interface with fewer amino acids than a structured protein [2 , 6] and the binding affinity is accumulated from the affinities of each segment [5] . This allows IDPs to have high binding specificity , but the loss of entropy upon binding imparted by the flexibility makes the interaction reversible [7] . From a kinetic perspective , sequential binding of individual segments will have a much higher rate constant than a hypothetical situation in which a pre-organized IDP simultaneously makes all contacts with the ordered protein [5] . A computational method based on the dock-and-coalesce model was successful in predicting the binding rate constants of disordered PPIs [8] . Experimental structure determination of disordered PPIs using techniques such as X-ray crystallography and nuclear magnetic resonance ( NMR ) is challenging due to the flexible nature of IDPs and their tendency to form weak , transient interactions [9] . Indeed , not all IDPs form a single , stable structure when bound . Examples of these so-called “fuzzy” complexes are cataloged in FuzDB [10 , 11] . Along a similar line , pE-DB contains ensembles of conformations that can be adopted by an IDP [12] . Nevertheless , many proteins annotated as disordered in DisProt [13] do adopt a bound structure that can be experimentally determined . For PPIs of structured proteins , experimental structure methods can often be complemented by computational modeling of protein complexes ( docking ) [14] . However , current rigid-body and flexible docking methods ( which allow small conformational changes at the docking interface ) are not able to model disordered PPI prediction , because the required rigid structures are not available for IDPs . Among existing protein modeling techniques , peptide-protein docking methods would be the most similar to disordered PPI prediction . Approaches to peptide-protein complex modeling include template-based modeling ( TBM ) [15 , 16] , molecular dynamics ( MD ) [17–19] , small molecule docking [20 , 21] , protein-protein docking with flexibility [22–26] , and coarse-grained docking [27] . The characteristics of the docking and MD methods are compared in Table 1 . Several of the methods require knowledge of the binding site as input . Information about the binding site can be obtained experimentally or by using computational prediction of peptide binding sites [28–30] or protein binding sites [31 , 32] . More fundamentally , existing methods were developed and tested for binding short peptides of 2–16 residues , which is far shorter than the 10–70 residue IDPs that participate in disordered PPIs [2] , although some programs are able to accept peptides up to 30 residues in their web servers . To predict the tertiary structure of a disordered PPI , a method must solve two interdependent problems: the tertiary structure of the input sequence of the disordered protein and its binding location on the receptor protein . This is a difficult task as the conformational space to be explored for an IDP is enormous and grows with its length . Currently , no existing methods can dock a long disordered protein to its receptor protein . A totally new approach is required for predicting the structure of a disordered PPI involving commonly observed long IDPs . In this work , we describe the development of a novel computational method named IDP-LZerD , which is able to model for the first time the docked structure of long IDPs ( 15–69 amino acids ) . IDP-LZerD applies the biophysical principles of the dock-and-coalesce mechanism of IDP binding to model the structures of long IDPs . In the “dock” phase , small segments of the IDP are modeled in various conformations and docked globally to the ordered protein . Modeling and docking small segments is not only faster and easier but also consistent with the biophysical mechanism of small segments of the IDP binding sequentially . In the “coalesce” phase , the docked segments of neighboring regions of the IDP are found and combined into a complete structure of the disordered PPI . We found that correct bound conformations of the IDP were selected using scores evaluating docking with the receptor , which corresponds to the biophysical model that the conformation of an IDP is stabilized and determined by contacts with its receptor . In addition , the combination of the docking scores of multiple segments is analogous to the accumulation of the binding affinities of multiple segments [5] . Overall , we show that IDP-LZerD is able to yield docking models of a practical quality in a number of bound and unbound structures of PPIs involving long IDPs .
Secondary structure was predicted for each IDP using JPRED [44] , Porter [45] , SSPro [46] , and PSIPRED [47] . The secondary structure predictions were reasonably accurate ( Table 4 ) . If the predictions are considered correct when any of the four methods predicts the correct secondary structure , the accuracy is 86% . For 57% of residues , all four methods predicted the correct secondary structure . Even in the minority of cases where none of the methods predicted the correct secondary structure , fragments of all three secondary structure classes were created ( described below in Methods ) . The full sequence of a target disordered protein was divided into 9-residue windows with a 3-residue overlap . Fragment structures of each window were predicted using Rosetta Fragment Picker ( RFP ) [49] , which predicts structures based on the sequence profile [50] and predicted secondary structure [44–47] . RFP was configured to output 30 fragments for a window . Increasing the number of fragments chosen did not yield structures of a substantially lower root mean square deviation ( RMSD ) to the native structure ( S1 Fig ) . Fragment structure was predicted reasonably accurately: on average the largest backbone RMSD of 30 conformations for a window was 1 . 8 Å for the training set , 1 . 6 Å for the test set , and 1 . 8 Å overall ( S2 Table ) . For a sequence window , each of the 30 fragment structures was docked with the receptor protein using LZerD [35–37] . LZerD is a shape-based , rigid-body docking method with the advantage of a soft representation of the surface shape of a protein that accounts for some conformational change upon binding . Docked fragment poses were clustered and the top 4 , 500 cluster centers were selected ( see Methods ) . Ranking was performed using the sum of the Z-scores of two scoring functions , DFIRE [51] and ITScorePro [52] , named DI score . DI score was shown to perform better in docking pose selection than the individual scores ( S3 Table ) . The docking accuracy of fragments is summarized in the “All docked” columns in S2 Table . For bound cases , on average the worst ( largest ) of the minimum L-RMSD from all the windows in a target was 3 . 7 Å and 4 . 1 Å for the training and the testing set , respectively . For unbound cases , the values were slightly worse , 4 . 4 Å and 4 . 3 Å for the training and the testing set , respectively . Fragment structure and docking accuracy was further tested on an additional independent test set of 11 cases of 9-residue IDP complexes found in the database of eukaryotic linear motifs ( ELMs ) [53] ( Table 5 ) . The results are shown in Table 6 . The average fragment RMSD is 1 . 4 Å and the average minimum docked RMSD is 3 . 2 Å for both bound and unbound cases ( Table 6 ) , which are better than the results shown in S2 Table . Selection of docked fragments was successful for most of the training set complexes , with an average RMSD of 5 . 4 Å for bound and 6 . 5 Å for unbound ( “Selected docked” columns in S2 Table ) . On the testing set , the results are similar , 5 . 7 Å and 6 . 3 Å for bound and unbound cases . Exceptions included 2clt and 1bk5 , where poor selection of docked fragments prevented successful modeling in the subsequent steps . On the additional ELM-derived dataset , results were 4 . 9 Å for bound and 4 . 7 Å for unbound ( Table 6 ) , which are again comparable to the results on the testing and training datasets . Interestingly , as shown in Fig 2 , evaluating docking fit with DI score often identified fragments of a low RMSD . To understand the general trend , for each sequence window we compared the fragment RMSD distributions of the 30 fragment structures from RFP and the top 30 docked fragments by DI score . Out of 144 windows from the 28 cases in the training set , for 83 ( 57 . 6% ) windows the top 30 by DI score are either better ( p <0 . 05 by the Mann-Whitney U test ) or contained five or more fragments with an RMSD better than 3 . 5 Å ( considered because there were cases where all 30 fragments from RFP were below 3 . 5 Å RMSD and no further improvement is possible by the DI score choice ) . This indicates that the DI score is detecting the increased binding affinity of the correct conformation when bound in the correct location , analogous to induced fit upon binding . Docked fragments from each window were combined to form full-length IDP complexes , referred to as paths . First , we performed a pre-filtering of docked fragment pairs , which removes physically improbable pairs by considering mutual distances and angles; then , paths were assembled using an extend-and-cluster strategy ( see Methods ) . This procedure effectively reduced the search space from as many as 1041 to the order of 105 paths regardless of the length of the IDP ( S2 Fig ) . Overall , the combination process successfully produced low RMSD paths . Out of the fourteen IDPs in the training set , for eleven bound and eight unbound receptors , paths with a 6 . 0 Å or lower RMSD were constructed ( “Clustered paths” in S2 Table ) . Results were slightly worse for the testing set , an RMSD of below 6 Å was obtained for three bound and three unbound cases out of the eight IDPs . For a complex , up to 1000 paths were chosen for further refinement . Paths were scored using a linear combination of four terms ( Path Score ) : the energy score , representing the docking scores of fragments across all windows; the overlap score , evaluating how well the neighboring docked fragments fit into a continuous path; the cluster size , accounting for the consensus of docking poses; and the receptor score , which measures docking site consensus . Path Score selected more hits than any of the individual score components ( S4 Table ) . On average , the minimum RMSD of selected paths was 6 . 7 Å for bound and 8 . 0 Å for unbound in the training set and 7 . 5 Å and 8 . 2 Å for bound and unbound in the testing set ( S2 Table ) . As in the situation in the docked fragment selection ( Fig 2 ) , it was observed that Path Score selected many models with IDPs of correct conformation ( RMSD under 6 . 0 Å; Fig 3 ) . Out of the fourteen pairs of targets in the training set , in ten/eleven cases for bound/unbound at least one of the top 10 models by Path Score has a correct IDP conformation . For the testing set , in four out of eight cases for both bound and unbound Path Score selected a correct IDP conformation within the top 10 . These are again interesting results because Path Score mainly evaluates the binding affinity of a target IDP and its receptor , but also identifies IDPs of the correct conformation . Thus , in accordance with the biophysical mechanism , the binding affinity of the IDP is accumulated from the binding affinities of the individual segments and the conformation of IDPs is determined by binding . Selected paths underwent structure refinement using constrained molecular dynamics , which connects neighboring fragments in a path and relaxes the overall IDP structure . An initial structure of a path was created by averaging the positions of the overlapping atoms ( Fig 4A , purple ) . Multiple rounds of minimization were performed using tapering harmonic restraints to prevent excessive movement of fragments . Refinement improved the protein-like nature of the combined fragments in a path . Before refinement , only 50 . 4 ( 48 . 2 ) % of ligand Cα-Cα distances were between 3 . 75 and 4 . 0 Å in the training ( testing ) set , which was improved to 92 . 3 ( 95 . 8 ) % by the refinement ( S3A Fig ) with a small cost of deterioration of ligand RMSD ( L-RMSD ) for about half of the cases ( S3B Fig ) . In parentheses , results for the testing set are shown . Refinement improved both L-RMSD and rank for some models , including the first hit for Bcl2-like protein 1 ( Bcl2-L-1 ) and its antagonist ( BAD; PDB ID 2bzw; Fig 4A ) . Originally , the path was ranked at 14 with a L-RMSD of 4 . 40 Å , which improved to rank 1 with L-RMSD 3 . 75 Å by the refinement . Finally , refined models were re-ranked and selected using a composite score of DFIRE [51] , ITScorePro [52] , a molecular mechanics score [54] , and GOAP [55] ( Model Score ) . Model Score selected hits at a higher rank than the single scores ( S5 Table ) . Model Score has moderate overall correlation to L-RMSD but often selected acceptable models with low scores ( Fig 5 , left panel ) and successfully identified hits in many cases as we discuss in the next section . RMSD of IDPs only and L-RMSD of docked models only correlate for models with an L-RMSD less than 10 Å ( Fig 5 , right panel ) . Tables 7 and 8 summarize prediction results on the training and testing sets , respectively , listing the rank of the first acceptable model ( RFH ) ( the criteria for an acceptable model are shown in S1 Table ) and fnat . On the training set ( Table 7 ) , IDP-LZerD produced at least one hit within the top 1000 models for thirteen bound and eleven unbound targets , and Model Score ranked hits within the top 10 for ten bound and five unbound cases . Notably , the rank 1 model was a hit for four complexes ( three bound , one unbound ) . There was only one complex where no hits were produced for both bound and unbound ( 2c1t/1bk5 ) . On the testing set ( Table 8 ) , IDP-LZerD produced at least one hit within 1000 models for almost all of the targets: all eight bound and seven unbound targets , and one top 1 hit for both bound and unbound . These fractions of top 1000 hits are higher than on the training set . Hits were ranked in the top 10 for two bound and three unbound cases . The fraction of top 10 hits ( 2/8 , 25% , for bound cases ) is lower than for the result observed on the training set ( 10/14 , 71% ) , while higher for unbound cases ( 3/8 , 37 . 5% ) than the training set ( 5/14 , 35 . 7% ) . Interestingly , for most of the cases in both training and testing set results , the acceptable models have a high fnat , much higher than the 0 . 1 minimum for an acceptable model defined by CAPRI ( S1 Table ) . A high fnat indicates that binding positions of IDPs are well reproduced in the models . We also evaluated predictions in terms of the fraction of correctly placed ligand residues of the top 10 models ( BF10 ) . Unsurprisingly , the fraction is high for cases with hits ranked in the top 10 . What is more interesting is that there are cases where targets that do not have any hits within the top 10 nevertheless have substantial BF10 , which indicate largely correct models are ranked high . Such targets include 1wkw , 1l8c , 2o8g , and 1l2w from the bound targets and 1ipb , 4i9o , 1khx , and 1u2n from the unbound targets . Fig 6 shows examples of four bound and four unbound complexes with acceptable or better top 10 hits . The four bound cases shown , 1ycr , 2cpk , 3owt , and 1xtg , include two medium quality hits , with RMSD at the interface ( I-RMSD ) below 2 . 0 Å ( 1ycr and 2cpk ) , and the IDPs range in length from 15 to 59 amino acids . The four unbound cases , 4ah2 , 1ijj , 1l3e , and 1jya , have IDPs between 20 and 69 amino acids . In all these examples , binding sites of the receptor proteins were accurately identified and overall docking structures were well predicted; often , even the pitch of the helices was reproduced . These examples demonstrate that IDP-LZerD can successfully select and combine docked fragments to produce accurate top 10 models for IDPs , even for cases with well over 30 amino acids . We also tested if binding residue predictions of receptor proteins is useful to improve model selection ( Table 7 ) . We used BindML [56] , which predicts binding site residues from their mutation patterns . Models were first filtered by the agreement of binding residues to the BindML prediction ( S4 Fig ) ; then , the selected models were ranked by Model Score . Using BindML prediction ( Table 7; RFH-B ) did not make a large difference but slightly improved the model selection performance for 10 cases without worsening any cases . In this section we evaluated the impact of secondary structure prediction on the quality of final models in two ways . First , in S5 Fig we examined how the accuracy of the secondary structure of residues influenced the accuracy of the residue position ( Cα RMSD ) in the models . In the figure , for example , “HC” indicates cases where the native residue is helix and the modeled residue is coil . It turned out that correctly predicted helix residues ( class “HH” ) have lower mean Cα RMSD , e . g . are more accurate , than other classes ( one-way ANOVA p = 1 × 10−35 and Tukey’s range test ) . Next , in S6 Fig , we addressed the influence of the secondary structure prediction agreement on the Cα RMSD of residues . The X-axis shows the number of secondary structure prediction methods that agree ( e . g . consensus ) on the correct secondary structure of residues and the Y-axis is the Cα RMSD of residues in the models . Residues where none of the four secondary structure prediction methods predict the correct secondary structure ( consensus 0 ) have higher ( worse ) mean Cα RMSD than other residues ( one-way ANOVA p = 1 × 10−11 and Tukey’s range test ) . Thus , we see some influence of the accuracy of predicted secondary structure to the quality of the final model with statistical significance , but as seen from the figures , difference was not very large . In IDP-LZerD , the fragment generation procedure creates fragments of all three secondary structure classes even if none of the methods predict the correct class to minimize the impact of incorrect secondary structure prediction . To further examine performance of IDP-LZerD , we compared modeling results with other methods . While no other methods are designed to model complexes involving long IDPs , some peptide-protein modeling software can use relatively long peptides . We compared IDP-LZerD with CABS-dock [27] and pepATTRACT [26] , because as seen in Table 1 , these two do not require the binding site as input and the programs are available for us to run . The CABS-dock web server outputs 10 docking models for a peptide up to 30 amino acids while the pepATTRACT web server outputs 50 docking models and does not explicitly limit the length of the peptide . The performance was compared on the eleven bound and unbound complexes with IDPs up to 30 amino acids in Tables 2 and 3 . Within the top 10 , CABS-dock had hits for six bound cases and four unbound cases , pepATTRACT had hits for three bound cases and one unbound case , and IDP-LZerD had hits for seven bound and four unbound cases ( Table 9 ) . The longest IDP successfully modeled by CABS-dock was 26 amino acids and the longest IDP successfully modeled by pepATTRACT was 22 amino acids . In contrast , IDP-LZerD had top 10 hits for the longest IDPs in this table ( 27 amino acids; Table 9 ) in addition to even longer IDPs in the full dataset ( Tables 7 and 8 ) . Therefore , overall IDP-LZerD showed better performance than the two methods compared . In addition , we compared the performance of IDP-LZerD to the previously published results of MD-based peptide-protein modeling methods [17–19] . The protein-peptide complexes used in their literature range from 2–15 amino acids . Among their datasets , we ran IDP-LZerD on all cases with 11 or more amino acids and unbound receptors , for a total of eight cases ( S6 Table ) . IDP-LZerD produced acceptable models in the top 10 for five out of eight cases with a sixth case having an acceptable model at rank 306 ( Table 10 ) . For the two cases with no hits , 2am9 and 1b9k , paths with 5 Å RMSD were created in Step 3 ( Fig 1 ) but not selected for refinement . IDP-LZerD and AnchorDock produced the same number of hits , but the models produced by AnchorDock have a lower RMSD . The results indicate an advantage of MD over coarse-grained approach for short peptides . They also suggest a potential improvement of IDP-LZerD by employing MD for the initial fragment-docking step , although it would take significantly more computational time than the current procedure . In addition to the other successful cases , we chose four cases to discuss , which illustrate the usefulness of IDP-LZerD models . In some disordered PPIs , the IDP forms secondary structure in the bound form that is not seen in isolation . The interaction between β-catenin and Transcription factor 7-like 2 ( TCF7L2 ) , which is involved in the Wnt signal transduction pathway , is such an example . In isolation , TCF7L2 exhibits circular dichroism spectra consistent with 96% random coil and 4% β-sheet , indicating that it is intrinsically disordered [57] . In contrast , the crystal structure of the complex ( 1jpw ) shows a C-terminal helix ( residues 40–50 ) , which was correctly predicted by the secondary structure methods and many models by IDP-LZerD . For both bound ( 1jpw ) and unbound ( 2z6h ) receptors , the overall complex was well-modeled ( RMSD at the interface , I-RMSD: 2 . 85 Å for bound and 4 . 50 Å for unbound ) with the structure and location of the C-terminal helix and hotspot residue Leu48 ( full atom L-RMSD 1 . 43 Å ) predicted very well in the bound case ( Fig 7A ) . Interestingly , among 1000 docking models generated , Leu48 was the most frequent contact in both the bound and unbound cases , appearing in 956 models for bound and 944 models for unbound , compared to an average of 685 and 696 , respectively ( S7 Fig ) . There are two more experimentally verified hotspot residues in the IDP , Glu17 and Asp16 [57] . Glu17 was in contact with the receptor in both bound and unbound cases in more than the average number of models , but Asp16 did not stand out ( S7 Fig ) . The next examples are complexes between CREB-binding protein ( Cbp ) /p300 TAZ1 domain and its disordered regulator proteins , hypoxia inducible factor 1-α ( HIF-1α ) and its competitive inhibitor , Cbp/p300-interacting transactivator 2 ( CITED2 ) . HIF-1α and CITED2 are different lengths , have only 12 . 5% sequence identity , and bind differently to the Cbp/p300 TAZ1 domain ( in Fig 7B , the N-terminus of CITED2 is at the bottom right while in panel D the N-terminus of HIF-1α is at the bottom left . The TAZ1 domain is shown in the same orientation in all panels ) . Nevertheless , the IDPs share a conserved binding motif ( LPEL in CITED2 , LPQL in Hif-1α , referred as LPXL ) [58] . We docked two complexes: CITED2 with human TAZ1 ( bound , 1p4q ) and HIF-1α with mouse TAZ1 ( bound , 1l8c; unbound , 1u2nA ) . Because the human TAZ1 domain does not have an available unbound structure , we used its structure in complex with HIF-1α ( 1l3eB ) for the unbound case , which has a binding site RMSD of 5 . 11 Å to the bound form with CITED2 . Remarkably , the prediction was accurate not only for the bound ( Fig 7B ) , but also for the unbound case ( Fig 7C ) . Both leucines in the LPXL motif , Leu243 and Leu246 , were experimentally verified as hotspot residues by mutagenesis [59] , but differ in contact consensus among the 1000 models . Leu243 has above-average counts ( rank 11 , 814 models , average 679 for bound and rank 8 , 851 models , average 713 for unbound ) while Leu246 has below-average counts ( rank 36 , 571 models for bound and rank 40 , 486 models for unbound; S8 Fig ) . For the mouse homolog , the bound case had no model under 12 . 6 Å L-RMSD in the top 10 . The rank 16 model shown had L-RMSD 20 . 1 Å , but the LPXL motif is located roughly at the correct position ( Fig 7D ) . The unbound case had no model with L-RMSD under 10 . 4 Å in the top 10 . However , HIF-1α was bound to almost the right location in the rank 9 model ( Fig 7E ) , where the fraction of correctly placed ligand residues was 0 . 71 and the L-RMSD of the LPXL motif was 3 . 7 Å . In addition , the residue Leu795 , which was experimentally determined to be a hotspot residue [60] , has high contact consensus for both bound and unbound ( rank 5 , 911 models , average 734 for bound and rank 8 , 881 models , average 694 for unbound; S9 Fig ) in the final 1000 models . Thus , in these four models the IDPs were bound almost at the correct place with the LXPL motif predicted particularly well . Finally , we discuss two cases where predictions did not yield acceptable quality models . The first case is the complex between Bcl2-like protein 1 ( Bcl2-L-1 ) and Bcl2-associated Antagonist of cell Death ( BAD ) . While the bound receptor had an excellent result with a medium quality model at rank 1 ( 2bzw; Table 7 , Fig 4a ) , the unbound receptor ( 1pq0 ) had no hits . However , visual inspection of the top-ranked models shows that the rank 1 to 7 models have a correct IDP conformation and binding site; however , the IDP is rotated by 180° within the binding site ( Fig 4b ) . Thus , the scoring functions detected a region of affinity but lacked the specificity to distinguish the correct orientation . The last example is a complex between botulinum neurotoxin type A ( BoNT/A ) and the N-terminal SNARE domain of SNAP25 ( sn2 ) . BoNT/A causes paralysis by cleaving SNARE proteins which impairs neuronal exocytosis [61] . Using the bound receptor ( PDB ID: 1xtgA ) , the structure was correctly predicted at rank 5 ( Fig 6d ) . However , with the unbound receptor ( 1xtfA ) , no hits were found . In the rank 1 model of the unbound case , while the IDP shows a substantial registration shift , the model occupies 32 . 6% of the binding groove ( top in Fig 7F; measured by the number of receptor residues within 5 Å of both IDPs ) . Thus , even in cases where no hits are produced , the produced models are reasonable and capture characteristic binding modes of IDPs on their receptors .
The current study presents for the first time that PPIs with long IDPs can be modeled with reasonable accuracy . By taking advantage of the crucial observation that disordered proteins tend to bind in continuous segments , the procedure is not only more computationally feasible but also functions similarly to the biophysical mechanism of IDP association . The prediction by IDP-LZerD was successful for the majority of the complexes tested , including unbound cases . The study further observed that the correct conformation of IDPs are often identified by evaluating docking scores with receptor proteins . A major challenge in modeling IDP interactions is the existence of fuzziness , where the IDP continues to exhibit multiple conformations in the bound state [11] . Two cases in the dataset we used are listed as fuzzy complexes in the FuzDB [11]: 1g0v ( FuzDB ID FC0018 ) and 3wn7 ( FuzDB ID FC0076 ) . IDP-LZerD managed to obtain a rank 1 medium hit for 1g0v ( Table 8 ) , while for 3wn7 IDP-LZerD produced an acceptable model at a low rank . It is particularly challenging to predict complexes where an IDP binds with two or more regions separated by loop regions that do not have direct contact its receptor ( clamp complexes [10] ) , because IDP-LZerD is based on the assumption that each segment of the IDP is in contact with the receptor . There are several other potential areas of improvement for the method . Docking larger fragments in cases where the structure of the fragments can be predicted with confidence could improve accuracy . It is also interesting to employ a coarse-grained model such as CABS [62] for generating fragment conformations and for more efficient structure refinement . In addition , explicit consideration of receptor flexibility could improve performance , although the soft surface representation used by LZerD already accounts for some degree of receptor flexibility . A key feature would be the ability to handle phosphorylated residues , as IDPs are frequently sites of post-translational modification and some complexes . This would require consideration of the effect of phosphorylation on secondary structure in addition to modification of the docking and scoring protocols . Another potential area of improvement is to guide docking by considering known or predicted hotspot residues on both IDPs and receptor proteins . Methods that could detect hotspots include computational alanine scanning [63] or applying a statistical scoring function [51 , 52] on a per-residue basis . Alternatively , as we showed in the case studies ( S7 , S8 and S9 Figs ) some promise was shown that hotspot residues could be predicted by taking consensus binding sites from ensembles of docking models . Accurate detection of hotspot residues could also lead to improved performance for fuzzy complexes , particularly the clamp class where two or more stably bound regions of the IDP are separated by fuzzy regions . Disordered PPIs are involved in important roles in various pathways and diseases . Overall , the work opens up a new possibility of modeling disordered protein interactions , providing structural insights for understanding the molecular mechanisms and malfunctions of these interactions , which are difficult to obtain by both experimental means and conventional computational protein docking methods .
Protein complexes containing IDPs with diverse functions and lengths were selected for developing and testing IDP-LZerD . Candidate complexes were found from reviews of disordered protein complexes [2 , 6] . In addition , cases were found in databases of eukaryotic linear motifs ( ELMs ) [53] and fuzzy complexes ( FuzDB ) [11] . For each case , disorder was verified by searching the literature for experimental evidence and DisProt [13] for a corresponding entry ( if available ) . Each PDB file was visually inspected and the case was removed if the residues annotated as disordered were missing or phosphorylated . The remaining proteins were divided into a training set of 14 complexes ( Table 2 ) and a test set of 8 complexes ( Table 3 ) . For bound complexes , unbound structures of the receptor , which were solved without the IDP , were found by searching PDB entries of the same UniProt ID as the receptor protein ( s ) . Docking using an unbound structure of the receptor protein would be more similar to a realistic scenario where the bound structure is unknown . If no PDB entries had the same UniProt ID , PDB entries with 90–100% sequence identity were used . Gaps of up to 16 amino acids were rebuilt using MODELLER [43] . In addition to the bound and unbound dataset described above , an additional dataset of 9-residue intrinsically disordered region ( IDR ) fragments was constructed from the database of eukaryotic linear motifs ( ELM ) [53] . To select disordered fragments , 442 proteins with structures in the PDB were cross-referenced against DisProt [13] using the Uniprot [64] ID , yielding 26 candidate complexes . By manual inspection of the PDB files , cases were removed if they were redundant ( using PISCES [65] with a 25% sequence identity cutoff ) with the full-length training set ( Table 2; 5 cases ) , other proteins within the set ( 6 cases ) , phosphorylated ( 4 cases ) , only had one chain ( 2 cases ) , or had fewer than 9 residues around the ELM resolved ( 4 cases ) . In addition , cases were added using other chains ( 1 case ) or adjacent to the ELM and also annotated as disordered ( 5 cases ) . The 11 cases of 9-residue fragments are listed in Table 5 . For each IDP sequence , we provided four independent secondary structure predictions , from PSIPRED [47] , JPRED [44] , Porter [45] , and SSPro [46] , each of which was used separately to generate one fourth of the fragments output by Rosetta Fragment Picker ( RFP ) . RFP was configured to output 30 fragments for a window ( S1 Fig ) . RFP produces fragments of each secondary structure class in proportion to its confidence score . For predictions by Porter and SSPro , which do not output confidence scores , we used 0 . 67 for the predicted secondary structure class and 0 . 15 for the other classes . Thus , fragments of all three secondary structure classes are obtained even in cases where the secondary structure prediction has strong consensus for one class . From the Cα coordinates of a fragment produced by RFP , the full atom backbone and side-chains were constructed using Pulchra [66] and OSCAR-star [67 , 68] , respectively . LZerD is a shape-based , rigid body docking algorithm [35] . For two input protein structures , LZerD generates many docking poses by geometric hashing and evaluates docking models using a scoring function that considers surface shape matching . Surface shape complementarity is evaluated using a mathematical surface descriptor , 3D Zernike Descriptor ( 3DZD ) [69 , 70] . Since 3DZD controls the level of surface smoothness , some degree of protein flexibility is considered in LZerD . 50 , 000 docking models were generated by LZerD for each fragment structure . Docked fragments were clustered with an RMSD cutoff of 4 . 0 Å and cluster centers were chosen using the LZerD score . The cluster centers were scored with ITScorePro [52] and the top 1 , 000 scoring fragments were pooled for each of 30 fragments of a window . Out of the 30 , 000 ( 1 , 000*30 ) docked fragments for each window , 4 , 500 docked fragments with the lowest DI score ( the sum of the Z-scores of ITScorePro and DFIRE [51] ) were kept ( S3 Table ) . By choosing one docked fragment from each window , conformations of the full length IDP , referred to as paths , were created . Prior to the path search , distance and angle cutoffs ( Fig 8 ) were applied to remove physically improbable pairs of docked fragments . Distance cutoffs were determined heuristically from the observed distributions in IDPs in DisProt [13] ( S10 Fig ) . Docked fragment pairs from all pairs of windows were removed from consideration if they are too close , i . e . an atom distance less than 3 Å or fragment midpoint distance less than 6 . 5 Å for neighboring windows and 3 . 8 Å otherwise . Pairs were also removed if their midpoint residues are too distant , more than 18 . 5 Å times the separation between the windows ( e . g . 2 for windows A and C ) . Also , to ensure that fragments from neighboring windows can be connected in the refinement stage , pairs are removed if they do not satisfy the following criteria: the overlap residue distance ( min . 5 . 2 Å , max . 13 . 6 Å ) , the overlap atom pair distances ( max . 6 Å for all atoms or 10 Å for any atom ) , and the overlap angle ( cos θ ≥ 0 . 1 so that only smoothly connected turns are included ) . Paths were assembled by first combining allowed pairs of docked fragments from the first two sequence windows and clustering them with a cutoff of 4 . 0 Å . Paths were extended to three windows , clustered again , and this process was repeated until all windows were added . A path was evaluated by Path Score , a linear combination of the Z-scores of four component scores: energy score ( SE ) , overlap score ( SO ) , cluster size ( SC ) , and receptor score ( SR ) : S p a t h = w 5 w 1 Z ( S E ) + w 2 Z ( S O ) + w 3 Z ( - S C ) + w 4 Z ( - S R ) + ( 1 - w 5 ) min { Z ( S E ) , Z ( S O ) , Z ( - S C ) , Z ( - S R ) } ( 1 ) where Z represents the Z-score across all paths in one complex . The lowest Z-score among the scores was included as an additional term because in some cases a good model is only detected by some of the scores . SC and SR are inverted so that a low Path Score is a favorable score . SE is the average of the binding scores ( DI score ) of docked fragments in the path . SO is the average of mean square distance ( MSD ) between the overlapping residues between consecutive fragments: S O = ∑ n = 1 | W | - 1 M S D ( n ) | W | - 1 , w h e r e ( 2 ) M S D ( n ) = ∑ r = 1 v ∑ a ∈ A | | X n , r + l - v , a - X n + 1 , r , a | | 2 v × | A | ( 3 ) and W is the set of windows , A is a set of N , Cα , Cβ , and C atoms of an overlapping residue , v is the overlap size ( =3 ) , l is the window length ( =9 ) , and Xn , r , a is the 3D coordinates of atom a of residue r of the docked fragment for window n . The Cβ atom is included in the computation of SE to account for rotational as well as translational congruency . For glycine , a virtual Cβ was constructed . SC is defined as the number of members in the path’s cluster . SR for a path considers whether the IDP binds to surface regions of the receptor that are also bound by other paths . For a surface residue of the receptor , the number of paths that bind to the residue ( minimum heavy atom distance ≤ 5 . 0 Å ) was counted ( called the number of occupying paths , Nop ) , and SR of a path is defined as the sum of Nop of the binding residues of the path . Weights were trained using a grid search from 0 . 1 to 1 . 0 with an increment of 0 . 1 and ∑ 1 4 w = 1 . The weight for the lowest Z-score , w5 , was trained in a second grid search . Weights were chosen that maximized the minimum recall of the targets used ( S4 Table ) . Recall is the number of hits retrieved by a given score divided by the total number of hits . Hits were defined as paths having pooled RMSD ≤ 10 Å . The pooled RMSD for a given path is defined as ∑ n ∈ W d n 2 / | W | where W is the set of windows and dn is the backbone L-RMSD of the docked fragment for window n , computed using only residues present in the crystal structure . The final weights for w1 through w5 were 0 . 5 , 0 . 1 , 0 . 3 , 0 . 1 , and 0 . 3 ( Eq 1 ) , with minimum recall of 2 . 8% . In addition , to validate the trained weights , we further performed a 2-fold cross validation by splitting the training dataset ( Table 2 ) into two groups . The results are shown in S7 Table . The weights obtained by training on group 1 were 0 . 4 , 0 . 3 , 0 . 2 , 0 . 1 , and 0 . 1 . The minimum recall observed on the group 2 set when predicted by using these weights was was 2 . 4% ( obtained for 1j3hA ) . The weights obtained on group 2 were 0 . 3 , 0 . 2 , 0 . 4 , 0 . 1 , and 0 . 1 , and the minimum recall when applied to the group 1 targets was 1 . 1% , observed for both 1devA and 1khxA . Thus , the final weights used in this study and the minimum recall were not largely different from what was observed in the 2-fold cross validation . For each complex , the 1000 paths with the lowest Path Score were kept for refinement , described in the next section . The selected paths were refined using molecular dynamics simulation . FACTS implicit solvation [71] was used . For minimization , all atoms of the receptor were fixed . With the ligand under a harmonic constraint of 50 kcal/mol/Å2 , the complex was minimized using 100 steps of the steepest descent ( SD ) algorithm followed by 100 steps of the adopted basis Newton-Raphson algorithm ( ABNR ) . This was followed by four rounds of 100 steps of ABNR minimization with ligand constraints of 40 , 30 , 20 , and 10 kcal/mol/Å2 . Next , the constraints were only placed on the backbone atoms of the ligand . Three rounds of 100 steps of ABNR minimization were run with ligand backbone constraints of 10 , 5 , and 1 kcal/mol/Å2 . The final minimization round was 5000 steps of ABNR minimization with no ligand constraints . Finally , the structure was equilibrated for 40 ps using a 2 fs timestep , fixed hydrogen covalent bond lengths , and a harmonic constraint of 10 kcal/mol/Å2 on all Cα atoms . The molecular dynamics simulation protocol was performed using CHARMM [72] but will also run using the academic free version charmm and could be implemented using other standard molecular dynamics software that implements harmonic constraints . Refined models were re-ranked using Model Score , an integrated score of ITScorePro [52] , DFIRE [51] , a molecular mechanics score [54] , and GOAP [55]: S m o d e l = w 5 w 1 Z ( ITScorePro ) + w 2 Z ( DFIRE ) + w 3 Z ( MolMech ) + w 4 Z ( GOAP ) + ( 1 - w 5 ) min { Z ( ITScorePro ) , Z ( DFIRE ) , Z ( MolMech ) , Z ( GOAP ) } ( 4 ) The lowest Z-score among the scores was included as an additional term because in some cases a good model is only detected by some of the scores . Weights were trained using a grid search with increments of 0 . 1 and ∑ 1 4 w = 1 . The weight for the lowest Z-score , w5 , was trained in a second grid search . Weights were chosen that minimized the mean rank of first hit ( RFH ) across all complexes used ( S5 Table ) . Hits were determined following the CAPRI criteria [39] . RFH is the numerical rank of the first model with CAPRI classification of acceptable or higher quality ( S1 Table ) . The final weights for w1 through w5 were 0 . 1 , 0 . 2 , 0 . 3 , 0 . 4 , and 0 . 3 ( Eq 4 ) , with mean RFH of 11 . 3 . To further confirm the validity of the trained weights , we performed an additional 2-fold cross validation ( S7 Table ) . The weights obtained on the group 1 set were 0 . 2 , 0 . 4 , 0 . 1 , 0 . 3 , and 0 . 3 and the mean RFH observed on the group 2 set when the predictions were made using the group-1 weights was 16 . 5 . The second group weights were 0 . 4 , 0 . 1 , 0 . 1 , 0 . 4 , and 0 . 1 and the mean RFH observed on the group 1 set by using the second group weights was 16 . 4 . RFH values obtained from this 2-fold cross validation ( S7 Table ) were very similar to the values reported in Table 7 , which indicates that the final weights were reasonably trained and capture the score landscape of the docking models well: out of 28 targets , RFH results were either the same or within a difference of 5 ranks for 22 targets . Docking one fragment to a receptor structure takes 2–4 hours on a single CPU . Thus , the docking step ( step 2 in Fig 1 ) takes about 120 CPU hours for a small receptor with a short IDP and as many as 1000 CPU hours for a large receptor with a long IDP . The path assembly step ( step 3 ) takes between 3 and 9 CPU hours . Finally , the refinement step ( step 4 ) takes between 4 and 6 CPU hours per model . LZerD is available for download at http://www . kiharalab . org/proteindocking/lzerd . php . IDP-LZerD is available for download at http://www . kiharalab . org/proteindocking/idp_lzerd . tar . bz2
|
A substantial fraction of the proteins encoded in genomes are intrinsically disordered proteins ( IDPs ) , which lack a single stable structure in the native state . IDPs serve many functions including mediating protein-protein interactions ( PPIs ) . Such disordered PPIs are prevalent in important regulatory pathways , including many interactions of the tumor suppressor protein p53 . To elucidate the molecular mechanisms of disordered PPIs , obtaining tertiary structure information is essential; however , they are difficult to study with experimental techniques and existing computational protein-protein and protein-peptide modeling methods are unable to model disordered PPIs . Here we present a novel computational method for modeling the structure of disordered PPIs , which is the first of this sort . The method , IDP-LZerD , is designed to follow a known biophysical picture of the mechanism of how IDPs interact with structured proteins . IDP-LZerD successfully modeled the majority of disordered PPIs tested . This technique opens up new possibilities for structural studies of IDPs and their interactions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"molecular",
"dynamics",
"protein",
"structure",
"prediction",
"protein",
"structure",
"intrinsically",
"disordered",
"proteins",
"research",
"and",
"analysis",
"methods",
"protein",
"structure",
"determination",
"proteins",
"biological",
"databases",
"structural",
"proteins",
"chemistry",
"proteomics",
"biophysics",
"molecular",
"biology",
"physics",
"biochemistry",
"proteomic",
"databases",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"chemistry",
"macromolecular",
"structure",
"analysis"
] |
2017
|
Modeling disordered protein interactions from biophysical principles
|
Inferior olivary activity causes both short-term and long-term changes in cerebellar output underlying motor performance and motor learning . Many of its neurons engage in coherent subthreshold oscillations and are extensively coupled via gap junctions . Studies in reduced preparations suggest that these properties promote rhythmic , synchronized output . However , the interaction of these properties with torrential synaptic inputs in awake behaving animals is not well understood . Here we combine electrophysiological recordings in awake mice with a realistic tissue-scale computational model of the inferior olive to study the relative impact of intrinsic and extrinsic mechanisms governing its activity . Our data and model suggest that if subthreshold oscillations are present in the awake state , the period of these oscillations will be transient and variable . Accordingly , by using different temporal patterns of sensory stimulation , we found that complex spike rhythmicity was readily evoked but limited to short intervals of no more than a few hundred milliseconds and that the periodicity of this rhythmic activity was not fixed but dynamically related to the synaptic input to the inferior olive as well as to motor output . In contrast , in the long-term , the average olivary spiking activity was not affected by the strength and duration of the sensory stimulation , while the level of gap junctional coupling determined the stiffness of the rhythmic activity in the olivary network during its dynamic response to sensory modulation . Thus , interactions between intrinsic properties and extrinsic inputs can explain the variations of spiking activity of olivary neurons , providing a temporal framework for the creation of both the short-term and long-term changes in cerebellar output .
A multitude of behavioral studies leave little doubt that the olivo-cerebellar system organizes appropriate timing in motor behavior [1–3] , perceptual function [4–6] and motor learning [7–10] . Furthermore , the role of the inferior olive in motor function is evinced in ( permanent and transient ) clinical manifestations , such as tremors , resulting from olivary lesions and deficits [11–16] . Although the consequences of olivary dysfunctions are rather clear , the network dynamics producing functional behavior are controversial . At the core of the controversy is the question whether inferior olive cells are oscillating during the awake state and whether these oscillations affect the timing of the inferior olivary output [17–19] . The inferior olive is the sole source of the climbing fibers , the activity of which dictates complex spike firing by cerebellar Purkinje cells ( for review , see [20] ) . Climbing fiber activity is essential for motor coordination , as it contributes to both initiation and learning of movements [8 , 10 , 21–26] , and it may also be involved in sensory processing and regulating more cognitive tasks [27–30] . Understanding the systemic consequences of inferior olivary spiking is therefore of great importance . The dendritic spines of inferior olivary neurons are grouped in glomeruli , in which they are coupled by numerous gap junctions [10 , 31–33] , which broadcast the activity state of olivary neurons . Due to their specific set of conductances [34–40] , the neurons of the inferior olive can produce subthreshold oscillations ( STOs ) [41–43] . The occurrence of STOs does not require gap junctions per se [44] , but the gap junctions appear to affect the amplitude of STOs and engage larger networks in synchronous oscillation [10 , 16 , 42] . Both experimental and theoretical studies have demonstrated that STOs may mediate phase-dependent gating where the phase of the STO helps to determine whether excitatory input can or cannot evoke a spike [45 , 46] . Indeed , whole cell recordings of olivary neurons in the anesthetized preparation indicate that their STOs can contribute to the firing rhythm [42 , 43] and extracellular recordings of Purkinje cells in the cerebellar cortex under anesthesia often show periods of complex spike firing around the typical olivary rhythm of 10 Hz [17 , 47–49] . However , several attempts to capture clues to these putative oscillations in the absence of anesthesia have , so far , returned empty handed [19 , 50] . It has been shown that in the anesthetized state both the amplitude and phase of the STOs can be altered by synaptic inputs [10] . Inhibitory inputs to the inferior olive originate in the cerebellar nuclei and have broadly distributed terminals onto compact sets of olivary cells [51–53] . Excitatory terminals predominantly originate in the spinal cord and lower brainstem , mainly carrying sensory information , and in the nuclei of the meso-diencephalic junction in the higher brainstem , carrying higher-order input from the cerebral cortex ( Fig 1A ) [15 , 54 , 55] . In addition , the inferior olive receives modulating , depolarizing , level-setting inputs from areas like the raphe nuclei [55] . Unlike most other brain regions , the inferior olive is virtually devoid of interneurons [56 , 57] . Thus , the long-range projections to the inferior olive in conjunction with presumed STOs and gap junctions jointly determine the activity pattern of the complex spikes in Purkinje cells . How these factors contribute to functional dynamics of the inferior olive in awake mammals remains to be elucidated . Here , we combine recordings in awake mice–in the presence and absence of gap junctions–with network simulations using a novel inferior olivary model to study the functional relevance of STOs in terms of resonant spikes . We are led to propose a view of inferior olivary function that is more consistent with the interplay between STOs , gap junctions and inputs to the inferior olive . Rather than acting as a strictly periodic metronome , the inferior olive appears more adequately described as a quasiperiodic ratchet , where cycles with variable short-lasting periods erase long-term phase dependencies .
To study the conditions for , and consequences of , rhythmic activity of the inferior olive , we made single-unit recordings of cerebellar Purkinje cells in lobules crus 1 and crus 2 ( n = 52 cells in 16 awake mice ) in the presence and absence of short-duration ( 30 ms ) whisker air puff stimulation ( Fig 1B and 1C ) . In the absence of sensory stimulation , the complex spikes of 35% of the Purkinje cells ( 18 out of 52 ) showed rhythmic activity ( Fig 2A–2C; S1 Fig ) with a median frequency of 8 . 5 Hz ( inter-quartile range ( IQR ) : 4 . 7–11 . 9 Hz ) . Upon sensory stimulation , 46 out of the 52 cells ( 88% ) showed statistically significant complex spike responses . Of these , 31 ( 67% ) had sensory-induced rhythmicity ( Fig 2D–2F ) , which was a significantly larger proportion than during spontaneous behavior ( p = 0 . 002; Fisher’s exact test ) . The median frequency of the oscillatory activity following stimulation was 9 . 1 Hz ( IQR: 7 . 9–13 . 3 Hz ) . Hence , the preferred frequencies in the presence and absence of sensory stimulation were similar ( p = 0 . 22; Wilcoxon rank sum test ) ( Figs 2C , 2F and 3 ) . The duration of the enhanced rhythmicity following stimulation was relatively short in that it lasted no more than 250 ms . With our stringent Z-score criterion ( >3 ) , only a single neuron showed 3 consecutive significant peaks in the peri-stimulus time histogram ( PSTH ) . The minimum inter-complex spike interval ( ICSI ) across cells was around 50 ms , putatively representing the refractory period . We conclude that complex spikes also display rhythmicity in awake behaving mice , and that sensory stimulation can amplify these resonances in periods of a few hundred milliseconds , even though stimulation is not required for the occurrence of rhythmicity per se . The pattern of rhythmic complex spike responses that was apparent for a couple of hundred milliseconds after a particular air puff stimulus repeated itself in a stable manner across the 1 , 000 trials ( applied at 0 . 25 Hz ) during which we recorded ( Fig 4A and 4B ) . For example , the level of rhythmicity of the first 100 trials was not significantly different from that during the last 100 trials ( comparing spike counts in first PSTH peak , p = 0 . 824; χ2 test , or latency to first spike , p = 0 . 727 , t test ) . This strongly indicates that there is–in a substantial fraction of the Purkinje cells–persistent oscillatory gating of the probability for complex spikes after a sensory stimulus resulting in time intervals ( “windows of opportunity” ) during which complex spikes preferentially occur ( Fig 4B–4E ) . These windows of opportunity become even more apparent when sorting the trials on the basis of response latency: the first complex spikes with a long latency following the stimulus align with the second spikes of the short latency responses . Similarly , there are trials during which complex spikes appear only at the third cycle ( Fig 4C , seen as a steeper rise around trial no . 650 ) . The occurrence of spikes during later cycles , not predicated on prior spikes , argues against refractory periods or rebound spiking as the sole explanations for such rhythmic firing [58] and highlight the putative existence of network-wide coherent oscillations . Since sensory stimulation of the whiskers can trigger a reflexive whisker protraction [59–61] and complex spike firing is known to correlate with the amplitude of this protraction [61] , we examined the relation between periodic complex spike firing and whisker protraction . To this end , we further analyzed an existing dataset of simultaneously recorded Purkinje cells and whisker movements during 0 . 5 Hz air puff stimulation of the ipsilateral whisker pad . In line with our previous findings [61] , trials during which a single complex spike occurred within 100 ms of whisker pad stimulation had on average a slightly , but significantly stronger protraction ( from 6 . 1 ± 5 . 4° to 6 . 8 ± 5 . 3° ( medians ± IQR ) , n = 35 Purkinje cells , p = 0 . 033 , Wilcoxon-matched pairs test after Benjamini-Hochberg correction; S2A–S2C Fig ) . Our new analysis revealed that also the occurrence of a second complex spike was correlated with a stronger whisker protraction . This could be observed as a second period of increased protraction during trials with two complex spikes . When compared to the increase in trials with a single complex spike , this second protraction was highly significant ( p < 0 . 001 , Wilcoxon-matched pairs test after Benjamini-Hochberg correction; S2D Fig ) . The second complex spike was unlikely a mere reflection of stronger protraction following the first complex spike , as there was no difference in whisker protraction between the trials that had a complex spike during the first , but not the second 100 ms after stimulus onset , and the trials with the opposite pattern ( a complex spike during the second , but not the first 100 ms; p = 0 . 980 , Wilcoxon-matched pairs test after Benjamini-Hochberg correction ) . The rhythmic firing pattern of complex spikes was thus reflected in the behavioral output of mice . The existence of windows of opportunity for complex spike activity is compatible with the assumption of an underlying STO , and cannot solely be explained by rebound activity without invoking circuit-wide extrinsic mechanisms . To test the implications of assuming olivary STOs , we proceeded to reproduce a detailed network with a tissue-scale computer model of the inferior olive neuropil . The model is constituted by 200 biophysically plausible model cells [40 , 46 , 62] embedded in a topographically arranged 3D-grid ( Fig 5A–5C ) . It has the scale of a sheet of olivary neurons of about 10% of the murine principal olivary nucleus ( cf . [63] ) . The model was designed to test hypotheses about the interaction between intrinsic parameters of olivary neurons , such as STOs and gap junctional coupling , and extrinsic parameters including synaptic inputs during the generation of complex spike patterns . Each neuron in the model is composed of a somatic , an axonal and a dendritic compartment , each endowed with a particular set of conductances , including a somatic low threshold Ca2+ channel ( Cav3 . 1; T-type ) , a dendritic high threshold Ca2+ channel ( Cav2 . 1; P/Q-type ) and a dendritic Ca2+-activated K+ channel , chiefly regulating STO amplitudes , while a somatic HCN channel partially determines the STO period ( Fig 5B; see also Methods ) . The dendrites of each neuron are connected to the dendrites of , on average , eight nearby neighbors ( within a radius of three nodes in the grid , representing a patch of about 400 μm x 400 μm of the murine inferior olive ) , simulating anisotropic and local gap junctional coupling ( Fig 5C ) . As the inferior olive itself , our model has boundaries which have impact on local connectivity characteristics , such as the clustering coefficient , though these did not have significant impact on the average firing rate between edge and center cells ( p = 0 . 812 , comparing edge and center cells , Kolmogorov-Smirnov test; S3 Fig ) . The coupling coefficient between model cells varied between 0 and 10% , as reported for experimental data [45 , 64 , 65] . Sensory input was implemented as excitatory synaptic input , simulating the whisker signals originating from the sensory trigeminal nuclei that were synchronously delivered to a subset of model neurons . Additionally , a “contextual input” was implemented as a combination of inhibitory feedback from the cerebellar nuclei and a level setting modulating input ( Figs 1A and 5A ) . This contextual input is modeled after an Ohrstein-Uhlenbeck process , essentially a random exploration with a decay parameter that imposes a well-defined mean yet with controllable temporal correlations ( see Methods ) . The amplitude of the contextual input drives the firing rate of the model neurons , which we set around 1 Hz ( S4 Fig ) , corresponding to what has been observed in vivo [28 , 43 , 66] . Thus , our model network recapitulates at least part of the neural behavior observed in vivo due to biophysically plausible settings of intrinsic conductances , gap junctional coupling and synaptic inputs . Whether a model neuron at rest displays STOs or not is largely determined by its channel conductances . Activation of somatic T-type Ca2+ channels can trigger dendritic Ca2+-dependent K+ channels that can induce Ih , which in turn can again activate T-type Ca2+ channels , and so forth . This cyclic pattern can cause STOs that could occasionally produce spikes ( Fig 5B ) . In our model , the conductance parameters were randomized ( within limits , see Methods ) so as to obtain an approximate 1:3 ratio of oscillating to non-oscillating cells ( S5 Fig ) guided by proportions observed in vivo [43] . Sensitivity analysis with smaller ratios ( down to 1:5 ) did not qualitatively alter the results ( data and analyses scripts are available online in https://osf . io/9hmpy/ ) . In the absence of contextual input , model neurons were relatively silent , but when triggered by sensory input , as occurred in our behavioral data ( Figs 2 and 4 ) , STOs synchronized by gap junctions would occur for two or three cycles ( Fig 6A and 6B ) . Our network model confirms that gap junctional coupling can broaden the distribution of STO frequencies and that even non-oscillating cells may , when coupled , collectively act as oscillators ( S6 Fig ) [67] . Adding contextual input to the model network can lead to more spontaneous spiking in between two sensory stimuli . Compared to the situation in the absence of contextual input , the STOs are much less prominent and the post-spike reverberation is even shorter ( Fig 6C ) . Accordingly , despite the significant levels of correlation in the contextual input ( 10% ) , the periods between oscillations are more variable due to the interaction of the noisy current and the phase response properties of the network . In addition , in the presence of contextual input our model could readily reproduce the appearance of preferred time windows for spiking upon sensory stimulation as observed in vivo ( Fig 6D–6F , cf . Fig 4 ) . This was particularly true for the model cells that directly received sensory inputs ( Fig 6G–6H ) . Moreover , the observed rhythmicity in model cells as observed in their STO activity was in tune with that of the auto-correlogram ( Fig 6D–6I ) in that the timing of the STOs and that of the spiking were closely correlated ( cf . Fig 5B ) . It should be noted though that model cells adjacent to cells directly receiving sensory input showed only a minor effect of stimulation . Thus , even though the gap junction currents in the model were chosen as the ceiling physiological value for the coupling coefficient ( ≤10% ) [45 , 67] , these currents alone were not enough to trigger spikes in neighboring cells . Both directly stimulated model cells and those receiving only contextual input exhibited phase preferences , seen in the spike-triggered membrane potential average as well as in the spike-triggered average of the input currents ( Fig 6G–6I ) . Spike-triggered averages of membrane potentials for any cell showed depolarization followed by hyperpolarization . In contrast , trials in which no spike was generated showed a depolarization just before the occurrence of the input . Similarly , the average of the input showed a long-lived phase preference , not only for a hyperpolarization before the spike , but also a preference for a depolarization in the previous peak of the STO , more than 100 ms earlier . These results are in line in vitro experiments under dynamic clamp and noisy input [68 , 69] . Likewise , the model indicates that for short durations STOs can induce clear phase dependencies for spiking , which fades under the variation of period durations dependent on the trial-specific contextual input ( as seen in our data ) . Depolarizing sensory input delivered onto a subset of the model cells can reset the STO phase in oscillatory cells and create resonant transients in others ( Fig 7A and 7B; see also S7 Fig on the appearance of rebound firing ) . If a second stimulus is delivered during this short-lasting transient , the response probability is increased . As in most cells with resonant short-lasting dynamics , inputs delivered during different phases can cause phase advances or delays . Hyperpolarization advanced the phase between 0−π and delayed the phase between π−2π , whereas depolarization had roughly the opposite effect , in addition to phase advancements with spikes in later cycles between π−2π ( Fig 7C–7E ) . Thus , there is a mutual influence of synaptic inputs and STOs on periodicity . While STOs can lead to phase-dependent gating , synaptic input can either modulate or reset the phase of the STOs , generating variable periods that range between 40–160 ms for the chosen amplitude of the contextual input ( Fig 8; S6 Fig ) . The only means of settling the question about the prevalence of STOs in awake and behaving mice would be intracellular recordings of inferior olivary neurons , which remains a daunting experimental challenge . We therefore looked for a less invasive method that could read out , from indirect and infrequent complex spikes , the presence or absence of STOs . We have developed one such paradigm inspired by auditory studies [70 , 71] using a rhythmic gallop stimulus that we first applied to the network model ( Fig 7F ) . In the gallop paradigm , stimuli are applied in quick succession with alternating intervals , comparable to the putative period of the underlying oscillation . Enough stimuli should be applied such that after multiple presentations the stimuli sample a uniform distribution of phases . In the context of auditory stimuli , the standard gallop experiment involves different tones and is used to test perceptual separation of auditory streams . Such rhythmic stimuli can help indicate resonances or physical limitations of the system , and distinguish across possible models for this separation ( such as in neural resonance theory [71] ) . One possible mechanism of auditory stream separation is an underlying oscillatory process which resets in certain phases and is less responsive in others . According to the in vitro inferior olive literature [41–43] this behavior is to be expected , and hence , such a stimulus can help distinguish underlying processes . If spikes are modulated by an oscillatory process , the presence of spikes on a short interval should be able to predict , in the next interval , the absence or presence of spikes . Indeed , if the underlying process producing spikes has oscillatory components and a relatively stable period , the probability of spikes in each interval is systematically different , which would appear as asymmetric ratios of response in the different intervals ( Fig 7F ) . This can be inspected as the length of the empty and filled vertical bars representing ratio of probabilities of spiking for long or short stimulation intervals ( Fig 7G ) . Thus , if the period of the STO rhythm would be regular and cause phase-dependent gating , complex spike responses following each stimulus interval are expected to show preferences for the short or long window of stimulation; these preferences were indeed observed ( Fig 7G , left ) . However , this clear phase dependency only appears in the noiseless model scenario . After adding a moderate amount of contextual input , this dependency washes off , rendering the responses in the two windows more symmetric ( Fig 7G , right ) , with only a few cells ( 5/200 ) displaying significant ratio differences ( tested against bootstrap with shuffled spikes ) . In line with the experimental in vivo data ( e . g . , Figs 2 and 4 ) , the olivary spike rhythmicity in the network model was steadily present over longer periods , and for a wide range of contextual input parameters ( S5 Fig ) . In addition , it also comprised , as in the experimental data , variations in frequency and amplitude during shorter epochs ( Figs 8 and S6 ) . Analysis of the network parameters indicates that these latter variations in oscillatory behavior can be readily understood by their sensitivity to both the amplitude ( parameter 'sigma' ) and kinetics ( parameter 'tau' , temporal decay ) of the contextual input . Indeed , because of the underlying Ornstein-Uhlenbeck process , the generation of contextual input converges to a specified mean and standard deviation , but in short intervals the statistics including the average network STO frequency can drift considerably ( Figs 6C and 8 ) . Since relatively small differences in oscillation parameters such as frequency can accumulate , they can swiftly overrule longer-term dependencies created by periodically resetting stimuli , as an analysis of phase distributions shows ( Fig 9 ) . Thus , based upon the similar outcomes of the network model and in vivo experiments , we are led to propose that ( 1 ) the STOs in the inferior olive may well contribute to the continuous generation of short-lasting patterns of complex spikes in awake behaving animals , and that ( 2 ) the synaptic input to the inferior olive may modify the main parameters of these STOs . Note that in the absence of input , periodic rhythmic behavior should be the default behavior of oscillating cells . Thus , in all likelihood , even if the inferior olive oscillates endemically , sustained but variable input should induce highly contextual spike responses to variable periods and render the olivary responses quasiperiodic , rather than regularly periodic as observed in reduced preparations . In line with in vivo whole cell recordings made under anesthesia [10 , 43] our awake data support the possibility that the moment of spiking may be related to the phase of olivary STOs , especially during the period of several hundred ms following stimulation ( Figs 2 and 4 ) . As discussed above , a gallop stimulus would expose such an oscillatory process underlying the response probabilities . Four idealized scenarios about the expected results can be constructed , as follows: first , one can start with a complete absence of STOs , which would result in a response probability unrelated to stimulus intervals; second , it could be that there were STOs , but no phase-dependent firing ( to be expected if the STO amplitude is small ) , which would also lead to complex spike firing irrespective of stimulus intervals; third , there could be STOs , but each stimulus would evoke a phase-reset , which again , would not lead to interval dependencies; and fourth , there could be STOs in combination with phase-dependent gating , which would result in a clear dependency of complex spike firing on the previous interval length ( Fig 7G , left panel ) . It should be noted that the large majority of studies on inferior olivary physiology , especially in reduced preparations , found evidence for the fourth situation ( STOs + phase-dependent gating ) [41 , 43 , 67] . To study whether phase-dependent gating in conjunction with an underlying oscillatory process could shape complex spike response timing in vivo we applied both a 250 vs . 400 ms and a 250 vs . 300 ms gallop stimulation using air puffs to the whiskers . Using only trials with a CS in the previous trial to calculate the ratio of responses ( 'conditional firing' ) a slight bias could be observed in the 250 vs . 400 ms paradigm ( Fig 10A ) and to a lesser extent in the 250 vs . 300 ms ( Fig 10B ) . Analysis including all trials ( 'non-conditional' ) is included in S8 Fig and shows no significant bias for any of the cells tested . Hence , our in vivo data are in line with the results from the network model subjected to synaptic noise , and show that the timing of complex spike responses to sensory stimulation is biased but not strongly determined by STOs . Our experimental data provided evidence for phase-dependent complex spike firing during brief intervals , but gallop stimulation did not expose a strong impact of STOs on complex spike response probabilities . Therefore , we sought an alternative approach to study the impact–if any–of STOs on complex spike firing in vivo . We reasoned that , if a sensory stimulus triggers a complex spike response with a certain probability , higher stimulation frequency should result in a proportional increase in complex spike firing . In particular , stimulus frequencies that would be in phase with the underlying STO would be expected to show signs of resonance and result in disproportionally increased complex spike firing . However , over periods of tens of seconds the complex spike frequency was resilient to varying the stimulus frequency between 1 and 4 Hz ( linear regression = -0 . 02; R2 = 0 . 1 ) ( Fig 10C ) and did not show signs of resonance with any of the stimulus frequencies , as there were no frequencies at which the complex spike firing was substantially increased . Only a very high rate of sensory stimulation ( 10 Hz ) , commensurable with the average duration of windows of opportunity , could induce a mild increase in complex spike firing frequency , albeit at the cost of a highly reduced response probability ( average increase: 71 ± 64% corresponding to an average increase from 1 . 12 Hz to 1 . 92 Hz; n = 5; p < 0 . 05; paired t test ) . This examination indicates that the average complex spike frequency is robust and stiff to modulation over longer time periods , imposing a hard limit on the frequency with which complex spikes can respond to sensory stimuli , confirming recent reports on complex spike homeostasis [72] . As stimulus triggered resonances were not observed at any of the stimulation frequencies , we turned to a more sensitive measure for the detection of oscillatory components in complex spike firing . We developed a statistical model that extrapolated from frequencies inferred through inter-complex spike intervals and stimulus triggered histograms ( Fig 11A and 11B ) . We reasoned that phase-dependent gating would imply that the interval between the last complex spike before and the first one after sensory stimulation aligns to the preferred frequency . In contrast , if sensory stimulation would typically evoke a phase reset , as suggested by our network model ( Fig 7 ) , no such relation would be found . The method was applied only to Purkinje cells with highly rhythmic complex spike firing . For each of those , we calculated their preferred frequency in the absence ( Fig 11C ) or presence of sensory stimulation ( Fig 11D ) . We used that frequency to construct statistical models representing idealized extremes of phase-dependent ( oscillatory ) and -independent ( uniform ) responses . For the oscillatory component we employed an oscillatory gating model , where the timing of the first complex spike after stimulus onset would be in-phase with the ongoing oscillation . This model was contrasted to a linear response model in which sensory stimulus could evoke a complex spike independent of the moment of the last complex spike before that stimulus , apart from a refractory period . For each Purkinje cell , we compared the distribution of the intervals between the last complex spike before and the first complex spike after stimulus onset with the predicted distributions based on the linear model , the oscillatory model and nine intermediate models , mixing linear and oscillatory components with different relative weights ( Fig 11A–11E ) . For the two extreme models as well as for the nine intermediate models we calculated a goodness-of-fit per Purkinje cell . Overall , when using these relatively long periods ( 300 ms ) , the linear model was superior to the oscillatory model , although a contribution of the oscillatory model could often improve the goodness-of-fit ( Fig 11F–11H ) . Despite the apparent failure of the oscillatory model to fit the data , the data did show an oscillatory profile for many of the cells ( Fig 11E ) . This lends support to our observations that short-lived , but reliable , oscillations are apparent in complex spike timing , although they have little impact on the timing or probability of sensory triggered CS responses . Apart from the STOs , extensive gap junctional coupling between dendrites is a second defining feature of the cyto-architecture of the inferior olive [31 , 33 , 73] . Absence of these gap junctions leads to relatively mild , but present deficits in reflex-like behavior and learning thereof [10 , 74] . We analyzed the inter-complex spike interval times in Purkinje cells of mutant mice that lack the Gjd2 ( Cx36 ) protein and are hence unable to form functional gap junctions in their inferior olive [69] . In line with the predictions made by our network model ( Figs 5H and S6 ) , the absence of gap junctions did not quench rhythmic complex spike firing during spontaneous activity ( Fig 12A ) . In fact , the fraction of Purkinje cells showing significant rhythmicity in the Gjd2 KO mice was larger than that in the wild-type littermates ( Gjd2 KO: 38 out of 65 Purkinje cells ( 58% ) vs . WT: 15 out of 46 Purkinje cells ( 33% ) ; p = 0 . 0118; Fisher’s exact test ) , with their average rhythmicity being significantly stronger ( p = 0 . 003; Kolmogorov-Smirnov test ) , measured by Z-scores of side peaks ( Fig 12B ) . Indeed , the variation in oscillatory frequencies across Purkinje cells of the mutants was significantly less than that in their wild-type littermates in that the latency to peak times per Purkinje cell were less variable ( p = 0 . 0431; Mann-Whitney test; Fig 12C ) . This latter finding is at first sight contradictory to our findings in the network model , where we show that gap junctions promote more uniform firing rates through increased synchrony between neurons ( Fig 5H ) . These simulations were run in the absence of synaptic input , though . Addition of contextual input also creates more variability in the wild type cells ( S6B Fig ) . As the lack of gap junctions increases cell excitability [10 , 69] , it is likely that synaptic input has a larger impact in the absence of gap junctions , leaving less room for inter-cell heterogeneity . Overall , removal of gap junctions affected the temporal and spatial dynamics by increasing the stereotypical rhythmicity of complex spike firing . We made paired recordings of Purkinje cells in awake mice to study the temporal relations of their complex spikes during spontaneous activity . The cell pairs were recorded with two electrodes randomly placed in a grid of 8 x 4 , with 300 μm between electrode centers . For each pair of simultaneously recorded Purkinje cells , we made a cross-correlogram . The median number of complex spikes in the reference cell used for these cross-correlograms was 827 ( range: 74–2174 ) . Cell pairs showed coherent activity in that they could show a central peak and/or a side peak in their cross-correlogram ( Fig 13A–13C ) . The side peaks could appear at different latencies , similar to the range observed in auto-correlograms of single Purkinje cells ( cf . Fig 2B ) . Moreover , Purkinje cell pairs that did not produce signs of synchronous spiking in the center peak could still produce an “echo” in the side peak after 50–150 ms . Counter-intuitively , cross-correlograms of Purkinje cell pairs of the wild type mice showed less often a significant center peak than those of Gjd2 KOs ( WT: 51 out of 96 pairs ( 53%; N = 4 mice ) ; Gjd2 KO: 44 out of 61 pairs ( 72%; N = 7 mice ) ; p = 0 . 0305; Fisher’s exact test ) . In line with the more stereotypic firing observed in single cells in the absence of gap junctions ( Fig 12 ) , the strength of the center peak was on average enhanced in the mutants ( Z-scores of significant center peaks ( median ± IQR ) : WT: 3 . 47 ± 1 . 82; Gjd2 KO: 5 . 75 ± 5 . 58; p = 0 . 0002; Mann-Whitney test ) ( Fig 13D–13F ) . Instead , the side peak of Gjd2 KO Purkinje cell pairs was not stronger than that of WTs ( Z-scores of significant side peaks ( median ± IQR ) : WT: 3 . 01 ± 0 . 89; Gjd2 KO: 3 . 04 ± 1 . 52; p = 0 . 194; Mann-Whitney test ) , leading to a lower ratio between center and side peak ( mean ± SEM: WT: 90 . 70 ± 5 . 17%; Gjd2 KO: 72 . 86 ± 6 . 52%; p = 0 . 036; t = 2 . 143; df = 67; t test ) ( Fig 13E and 13F ) . Interestingly , the occurrence of side peaks in Purkinje cell pairs was unidirectional in approximately half the cell pairs ( WT: 47 out of 82 pairs with at least one side peak ( 57% ) ; Gjd2 KO: 25 out of 47 pairs ( 53% ) ; p = 0 . 714; Fisher’s exact test ) , which means that one of the neurons of a pair was leading the other , but not vice versa . As this was consistent in the Gjd2 KO as well as the WT Purkinje cells , these data could reflect traveling waves across the inferior olive , which , however , must have extrinsic sources [44 , 75] . Thus , the paired recordings are compatible with the findings highlighted above in that the presence of coupling can affect the coherence of STOs for short periods up to a few hundred milliseconds , while leaving the window for later correlated events open .
During spontaneous activity , Purkinje cells generally fire a complex spike roughly once a second , but this frequency can be increased to about 10 Hz by systemically applying drugs , like harmaline , which directly affect conductances mediating STOs in the inferior olive [41 , 81] . Since these drugs also induce tremorgenic movements beating at similar frequencies , it has been proposed that the inferior olive may serve as a temporal framework for motor coordination [11 , 82] . This oscillatory firing behavior of the olivary neurons may mirror limb resonant properties and act as an inverse controller , for example by dampening the dynamics of the muscles involved [83] . In line with previous recordings [22–24 , 28 , 61 , 72 , 76 , 78 , 84] , the current data indicate that only a small fraction of Purkinje neurons respond to sensory stimulation with a complex spike response probability larger than 50% . This probability falls substantially with increasing frequency of stimulation , as the overall spike frequency only marginally increases to high frequency stimulation . Even after applying different temporal patterns of sensory stimulation for longer epochs , we observed no substantial deviation from the stereotypic 1 Hz firing rate . Moreover , it should be noted that even if the frequency of underlying oscillations has bearing on the pattern of responses of the gallop stimuli , conditional dependencies should be expected for most STO frequencies , unless the ratio of the interval of gallop and the STO period has no remainder . Given the seemingly consistent frequencies predicted by PSTH's and autocorrelograms of single cells ( Figs 2 and 3 , but also seen in cross correlograms , as in Fig 13 ) , we chose gallop intervals with periods commensurate with a representative frequency of 8 Hz , each of which should sample different phases in the oscillation . If at all present , we should have observed conditional dependencies on at least a few cells . In our study , complex spikes remain as unpredictable as ever . Thus , regulatory mechanisms keep the complex spike rate relatively stable over longer time periods [72] . No resonance is exhibited , irrespective of an enduring powerful sensory stimulus in a variety of frequencies . Save few exceptions , the presence of a complex spike in an interval is compensated by the absence in another . Thus , it looks as if the complex spikes rearrange themselves in time in order to keep close to its proverbial 1 Hz frequency . It remains to be shown to what extent the mechanisms involved are intrinsic ( cell-dependent ) and/or extrinsic ( network-dependent ) . A possible candidate for setting the overall level of excitability through intrinsic mechanisms is given by Ca2+-activated Cl− channels , which are prominently expressed in olivary neurons along with Ca2+-dependent BK and SK K+ channels [85 , 86] . In addition , the olivo-cerebellar module itself could partly impose this regulation [86–88] . Indeed , the long-term dynamics within the closed olivo-cortico-nuclear loop may well exert homeostatic control , given that increases in complex spikes lead to enhanced inhibition of the inferior olive via the cerebellar nuclei [20] . The impact of such a network mechanism may even be more prominent when changes in synchrony are taken into account [89] . We propose that a closed-loop experiment conducted while imaging from a wide field , producing stimulation as a function of the degree of complex spike synchrony , could tease out conditional complex spike probabilities . Increasing our capability of predicting complex spikes is instrumental to elucidate the control of inferior olivary firing . The existence of temporal windows of opportunity for complex spike responses following sensory stimulation highlighted a potential impact of STOs on conditional complex spike gating [10 , 41 , 43 , 90] . Indeed , autocorrelogram peaks correlated well with interspike intervals following stimulation , arguing for an underlying rhythm . Complex spikes could appear in a particular window even when they were not preceded by a complex spike in a previous window during a single trial , arguing against a prominent role of refractory periods in creating rhythmic complex spike responses . Comparing actual firing patterns with statistical models mixing linear or oscillatory interval distributions indicated a potential impact of oscillations . The mild impact of the oscillatory component on explaining the data may in part depend on the assumption that cells have a well-defined frequency . In other words , a variable rebound time could offset the phase response by a couple of milliseconds , reducing the contribution of the oscillatory model , though phase preferences due to prior spikes may still occur ( i . e . , Fig 11E ) . Our biophysical model suggests that fluctuating inputs , such as those mediating inhibition from the cerebellar nuclei or those relaying depolarizing modulation from the raphe nuclei [91 , 92] , may induce variations in the oscillation period on a cycle-by-cycle basis ( Figs 8 and 9 ) . As these contextual inputs are absent or suppressed in decerebrate or anesthetized preparations , as well as in vitro , they may also explain why many earlier studies systematically encountered cells with well-defined STO frequencies [10 , 41 , 43 , 45 , 47 , 79 , 93 , 94] . In the network model , in which we mimicked the contextual input as an Ornstein-Uhlenbeck process with local variations but no long-term drifts of the mean [95] , the results agree well with the experimental observations in terms of synchronous firing , phase shifts , cross-correlogram peaks and side peaks , as well as overall firing frequency . Indeed , the absence of resonant responses over longer time windows and the inconsistency of individual olivary cells to fire on every trial or cycle indicate that the STOs are not regularly periodic , but rather quasiperiodic , while still being synchronous . Even though several lines of evidence suggest a role for STOs ( see above ) , we did not observe an unequivocal , significant conditional dependence of complex spikes in the gallop paradigm , as expected by a noiseless model . How can a system with rhythmic responses at least partially fail to be phase modulated by such periodic stimuli ? An attractive alternative explanation for rhythmicity might be the occurrence of high-threshold Cav2 . 1 P/Q-type Ca2+ channel-dependent rebound spikes ( S7 Fig ) [12 , 62] . If impulse-like input to the olive can evoke a spike , and if this spike produces a rebound spike some tens of milliseconds later , this could explain the alignment between the PSTHs and cross-correlograms . However , this argument cannot explain stimulus triggered spikes at the second or third window of opportunity , without an earlier spike as observed in Fig 4 . As the occurrence of the rebound spike is predicated on a prior spike , a spike in the second or third window without a prior spike cannot be explained by the rebound spiking phenomenon , at least not within the same cell . In other words , the spikes happening exclusively in the second ( or third ) window of opportunity cannot be the result of a previous spike in the same cell , unless there is a shared rhythm in the network . It is also conceivable that strong hyperpolarization that is synchronized with the complex spike rhythm could promote reverberating firing , but this is an extrinsic mechanism , discussed below . As they stand , our findings do not support the idea that the post-spike hyperpolarization is a prerequisite for the complex spike pattern observed . Multiple windows of opportunity could , according to our model , be enhanced by transient oscillations induced by resets relayed by gap junctions to the local olivary circuit . Apart from the almost complete absence of interneurons , the presence of STOs and the exclusive projection to the cerebellum , the abundance of dendro-dendritic gap junctions is another defining feature of the inferior olive . The absence of these gap junctions does not lead to gross motor deficits , but prevents proper acquisition and execution of more challenging tasks [10 , 16 , 74] , which is in line with the relatively minor impact found on complex spike activity in Gjd2 KO mice . At first sight , the effects of deleting gap junctions seem counterintuitive . Synchronous and rhythmic patterns are exacerbated , rather than diminished by the loss of gap junctions . However , the side peak of the auto-correlogram is significantly squashed , indicating that the gap junctions have a role in the increased coherence of the upcoming oscillation . Gap junctions do not only facilitate synchronization of coupled neurons , they also lower their excitability by increasing the membrane resistance [69] . Together , this results in less direct coupling , observed as reduced synchrony of direct neighbors [16 , 96] , and increased responsiveness to synaptic input . This leads to more long-range coherence and as a consequence gap junction networks may act as a “noise filter” , promoting short-range quorum-voting on phase ( a term coined by Winfree [97] ) . This effect is visible in our model as spikes are most likely to occur when excitation follows inhibition ( Fig 6H ) . This is in line with the finding that complex spikes of nearby Purkinje cells have a preference to fire together [72 , 98 , 99] . This concept also agrees with the possibility that coupled olivary neurons may control movements by dampening the dynamics of the muscles involved at an appropriate level [83 , 100] , as both the resonances and movement oscillations increase shortly after sensory stimulation in Gjd2 KO mice [16] . Network resonances are a pervasive feature of brain circuits and they can be induced by subthreshold oscillations of particular cell types [101 , 102] . In addition to the autochthonous dynamics of the inferior olive , reverberating loops through the circuit could help explain some features of complex spike firing , including the occurrence of complex spike doublets and side peaks in cross-correlograms . Such phenomena could be explained by "network echoes" , where complex spikes in one cycle would induce complex spikes in the next cycle [87 , 103 , 104] . The most obvious candidate loop to produce is that via the cerebellum and the nuclei of the meso-diencephalic junction [55 , 105] . The output of the inferior olive is mainly directed via exceptionally strong synapses to the Purkinje cells [106] . These Purkinje cells in turn inhibit neurons of the cerebellar nuclei that can show rebound firing after a period of inhibition [88 , 107] . This rebound activity can excite the inferior olive again via a disynaptic connection via the nuclei of the meso-diencephalic junction . While an isolated complex spike is unlikely to evoke such a rebound activity , a larger group of Purkinje cells could be successful in doing so [20 , 107 , 108] . The travel time for this loop ( around 50–100 ms ) has been indirectly assessed in the awake preparation [8 , 10 , 26] , and corresponds to the latency of the rebound firing in the cerebellar nuclei under anesthesia [87 , 88 , 107 , 109] . This implies that the travel times for the entire loop would be in the same order as found for the preferred frequencies of complex spike firing . Other , more elaborate loops involving for instance the forebrain may also exist [110] and could play an additional role in shaping complex spike patterns . A putative impact of reverberating loops on rebound activity could be a network phenomenon , as the impact of an isolated complex spike may not be sufficient to trigger this loop . This is in line with the reduced “echo” in the cross-correlograms of the Gjd2 KO mice and enhanced doublets following lesions of the nucleo-olivary tract as occurs in olivary hypertrophy [111] . Taken together , rebound spiking , STOs and reverberating loops all seem to promote in a cooperative manner complex spike rhythmicity at a time scale of about 200 ms . Through modeling , we found that not only the state of the inferior olivary oscillations determines which inputs are transmitted , but that these inputs also determine the state of the network . Thus , inputs from both the cerebellum and the cerebrum determine the probability of complex spike responses on a cycle-by-cycle basis providing a quasiperiodic framework to align synchronous groups . This sharply contrasts with a view in which the inferior olive is a clock with regular periodicity . A circuit-wide understanding of cerebellar resonances on the basis of such a mechanism could open a novel pathway to explore the cerebellar gating by other brain regions . The combination of delayed gap junctions and delayed inhibition , as found in the olivo-cerebellar loop [104 , 112] , can affect oscillatory behavior [113] . The interplay between STOs and delayed inhibition is therefore also relevant for other neural circuitries , for instance for creating filter settings for the perception of sounds with specific oscillatory properties [114–116] or orchestrating rhythmic movements as shown in the present study ( see also [117] ) . Well-coordinated movement sequences are not timed rigidly; they must be enacted flexibly and contextually . In order to catch a ball , or a prey , or to perform any other appropriately timed movement , it is essential to fine-tune the duration and onsets of multiple coordinated output systems . An inferior olive that responds contextually to time varying input by advancing and delaying cycles does not act as a rigid clock or metronome , but more contextually , as a ratchet-pole system , with the frequency of 'clicks' of the ratchet reflecting the recent history of applied torque . The properties we have encountered in this study are consistent with a 'ratchet-like' dynamics for the inferior olive , which integrates time-varying stimulus in a phase-dependent manner . According to this view , the inferior olive responds to all inputs ( sensory and otherwise ) , by producing phase changes that are informative about the recent history of input , and dictate the appearance of coherent complex spike waves arriving at the cerebellar cortex .
All experimental procedures were approved a priori by an independent animal ethical committee ( DEC-Consult , Soest , The Netherlands ) as required under Dutch law . Experiments were performed on 16 adult ( 9 males and 7 females of 25 ± 14 weeks old ) homozygous Gjd2tm1Kwi ( Gjd2 KO , formerly known as Cx36 KO mice [10] ) mice which were compared to 15 wild-type littermates ( 8 males and 7 females of 26 ± 13 weeks old; means ± sd ) . The generation of these mice has been described previously [118] . The data described in S2 Fig originated from previously published recordings in 35 wild-type mice [61] . All mice had a C57BL6/J background . The mice received a magnetic pedestal that was attached to the skull above bregma using Optibond adhesive ( Kerr Corporation , Orange , CA ) and a craniotomy of the occipital bone above lobules crus 1 and crus 2 . The surgery was performed under isoflurane anesthesia ( 2–4% V/V in O2 ) . Post-surgical pain was treated with 5 mg/kg carprofen ( “Rimadyl” , Pfizer , New York , NY ) and 1 μg lidocaine ( Braun , Meisingen , Germany ) . Mice were habituated during 2 daily sessions of 30–60 min . Extracellular recordings of Purkinje cells were made in the cerebellar lobules crus 1 and 2 of awake mice as described previously [28] . Briefly , an 8 x 4 matrix of quartz-platinum electrodes ( 2–4 MΩ; Thomas Recording , Giessen , Germany ) was used to make recordings that were amplified and digitized at 24 kHz using an RZ2 BioAmp processor ( Tucker-Davis Technologies , Alachua , FL ) . The signals were analyzed offline with SpikeTrain ( Neurasmus , Rotterdam , The Netherlands ) using a digital band-pass filter ( 30–6 , 000 Hz ) . Complex spikes were recognized based on their waveform consisting of an initial spike followed by one or more spikelets . A recording was accepted as that of a single Purkinje cell when a discernible pause of at least 8 ms in simple spike firing followed the complex spikes and when the complex spikes were of similar shape and amplitude throughout the recording . Sensory stimulation was applied as air puffs of 20 psi and 25 ms duration directed at the whisker pad ipsilateral to the side of recording . The stimuli were given in trains of 100 or 360 pulses either at regular or alternating intervals . During a recording , trains with different stimulus intervals were played in a random sequence . Whisker videos were made from above using a bright LED panel as backlight ( λ = 640 nm ) at a frame rate of 1 , 000 Hz ( 480 x 500 pixels using an A504k camera from Basler Vision Technologies , Ahrensburg , Germany ) . The whiskers were not trimmed or cut . Whisker movements were tracked offline as described previously [119] using a method based on the BIOTACT Whisker Tracking Tool [120] . We used the average angle of all trackable large facial whiskers for further quantification of whisker behavior . Of each Purkinje cell we computed the probability density function ( PDF ) of both its complex spike autocorrelogram and its distribution of intervals between consecutive complex spikes ( inter-complex spike intervals ( ICSIs ) ) . PDFs were calculated with an Epanechnikov kernel ( with finite support ) with a width of 10 ms . In order to exclude stimulus-induced alterations in complex spike firing , complex spikes detected between 20 and 200 ms after a stimulus were omitted from this phase of the analysis . PDFs were calculated from 0 up till 500 ms . The peak in the ICSI PDF was considered as the “preferred ICSI interval” and its strength was expressed as the Z-score by dividing the peak value by the standard deviation of the PDF . To understand the impact of Purkinje cell with little or no preference for specific ICSI intervals on the analysis , we chose to look both at the Purkinje cells with high and low Z-scores . Thus , we grouped Purkinje cells into high and low level Z-scores , using a threshold of 3 . Air puff stimulations frequently triggered double complex spike response peaks , suggestive of an underlying inferior olivary oscillation . For further analysis of the conditional responses , an estimate of the putative inferior olivary frequency was derived from the interval between these two response peaks . First , it was established for each Purkinje cell whether two peaks were present in the PSTH . To this end , we set a threshold for each of these two peaks . For the first peak , this was calculated by reshuffling the ICSIs over the recording followed by calculating a stimulus-triggered pseudo-PSTH , repeating this procedure 10 , 000 times and selecting the 99% upper-bound . We considered the first response peak to be significant if it crossed the upper-bound uninterruptedly for at least 10 ms . Since the second response peak typically is much smaller than the first one , we calculated a new threshold for the second peak by excluding the time-window for the first response peak . This window was set from the time of the stimulus until where the response probability drops to the average response frequency , the response frequency as expected if stimuli do not trigger complex spikes , following the significant ‘first’ responsive peak . In 5 out of 98 Purkinje cells , the PDF of the response rate between clear peaks remained above the average response rate , in which case we used the time point where the amplitude drop in the PDF was more than twice the difference between upper bound and average response probability . The rest of the bootstrap method was identical to that for the first response peak . Only peaks up to 0 . 5 s after the stimulus were included in the population analysis . In order to test whether the phase of the inferior olivary oscillations affected the complex spike response probability , we compared the complex spike intervals over an Air puff for each stimulus that triggered a complex spike . To this end , we analyzed the recordings of 25 Purkinje cells ( 10 WT and 15 Gjd2 KO ) that were measured previously in crus 1 and crus 2 of awake , adult mice . We included only Purkinje cells that displayed clear oscillatory complex spike firing indicated by the display of a secondary complex spike response peak , as evaluated according to the bootstrap method described above , and/or significant peaks in the ICSI histogram . Only stable recordings covering at least 500 stimuli at frequencies below 1 Hz were considered for this analysis . For each recording we compared two idealized statistical models of the observed ICSI distributions: an oscillatory model showing phase-dependent spiking and stable olivary oscillation frequencies and a uniform model lacking phase-dependencies . For the oscillatory model , we created complex spike probability functions for the pre-stimulus interval ( -300 to 0 ms ) based on the oscillatory period established either for the ICSI distribution or from the interval between the two complex spike response peaks . We fitted a sine wave with the observed frequency , having its peak at the moment of the first complex spike in the stimulus response window ( 20–200 ms after the air puff ) and derived spike probability levels during the pre-stimulus interval from these fits , with the trough representing zero probability . Frequency and amplitude of every cycle were kept constant for the whole recording . In the uniform model , we calculated the pre-stimulus spiking probability with a uniform distribution based on the complex spike frequency of each Purkinje cell . We did include a refractory period , being the shortest ICSI observed for each recording , to reflect the inability of consecutive complex spikes to occur with a very short time interval . Refractory periods were comparable between mutants and wild types; 49 ± 15 ms for Gjd2 KO cells and 50 ± 20 ms for WT cells . Subsequently , we constructed compound fits consisting of linear summations of the two models . One extreme was the oscillatory model and the other the uniform model and we considered nine intermediate combinations ( e . g . , 0 . 3 x the oscillatory model + 0 . 7 x the uniform model ) . Every compound fit was run for 10 , 000 times . The goodness of fit was computed as the absolute differences of every single run of the model with the actual ICSI distribution . The model networks used here are comprised of a topographical grid of 200 coupled cells , in a 10x10x2 lattice arrangement , which may resemble an area of about 400 μm x 400 μm of the inferior olive , for instance , the rostral portion of the dorsal lamella of the principle olive of the mouse . It is available online at https://github . com/MRIO/OliveTree , branch 'Warnaar' . For instructions on how to run the model and reproduce analysis , check README_Warnaar . txt' . Each cell within these networks was modelled according to the single cell model described in [46] , which is an elaboration of a previous model [62] with an added axon and modified fast sodium channel . Equations are provided in the appendix of that publication at ( https://doi . org/10 . 1371/journal . pcbi . 1002814 . s002 ) , and can be checked in the MATLAB functions IOcell and createDefaultNeurons in the codebase . The model includes three compartments ( soma , dendrite , axon hillock ) with 12 conductances . In addition to the ionic mechanisms , the dendrite of the model cell has a Ca2+ concentration state variable , which is related to the intrusion through the Cav2 . 1 channels . The main ionic conductances responsible for the oscillation are the somatic T-type Ca2+ and the Ca2+-activated K+ ( SK ) channels present in the dendritic compartment . The crucial parameters governing the emergence of subthreshold oscillations are randomized , reflecting the experimental facts that about one third of the cells oscillate endemically ( in vivo ) with intrinsic variations in oscillatory frequencies [43] . The behavior of the STO of the model cells in our network as a function of their parameters , for models with and without gap junctions , are included in S5 and S6 Figs . Cell parameters are found in S1 Table . Connectivity is created with the function 'createW . m' in the MATLAB codebase . Briefly , cells within a specified radius of each other were connected according to a probability function such as to ensure the specified mean degree in the network ( n = 8 ) , chosen to resemble the observed connectivity distributions reported in the literature . The connectivity parameters ( distance and average connection probability ) were chosen to match experimental values ( radius ≤ 120 μm ) and average connection probability ( ~8 neighbors ) . The procedure to obtain connectivity is as follows . First , pairwise distances between all cells are calculated . Then , a binary adjacency matrix is created by thresholding those distances within a specified radius . Thereafter , we assign a random number between 0 and 1 to each link from a uniform distribution . Finally , this matrix is made binary by comparing each entry with a probability so that the average number of connections approximates a given mean connectivity . This binary adjacency matrix is then multiplied by the mean gap junction conductance parameter . Finally , gap junction conductance values are then randomized by a uniform jittering of the conductance by 10% of their original value . The conductance of gap junctions was normalized with a saturating factor by difference of potential between the neighboring cells , according to [62] based on findings from [121] with the following function: gc¯ ( ∆V ) =gc ( 0 . 8e ( −∆V2/100 ) +0 . 2 ) FORMULA 1 Where ΔV is the voltage difference between the connected cells , gc is the nominal coupling and gc¯ is the effective coupling . Two inputs are given to the model , one emulating the sensory input from whisker pad stimulation and the other representing a stimulus-independent background reflecting diverse excitatory and inhibitory inputs to the inferior olive . The latter consisted of a continuous stochastic process with known mean and standard deviation with a relaxation parameter following the Ornstein-Uhlenbeck process [95] , succinctly described underneath . Only one subset of cells in the center of the network ( 40% of the cells in a mask spanning a radius of 3 cells from the center of the network ) representing efferent arborization , receives the “sensory input” , with “sensory” currents being delivered to the soma of modelled cells ( gAMPA = 0 . 15 mS/cm2 ) . “Sensory input” was modeled according to O’Donnel et al . [122] . The mask is represented in Fig 5A . The cells of the inferior olivary network most likely share input sources due to overlapping arborizations of efferent projections [123] . To represent both shared and independent input , we have modeled the current source in each cell as having an independent process and a shared process , with a mix parameter ( alpha ) of 10% input correlation shared by all the cells in the network . This level of correlation leads to a coherent background oscillation in the cells of the network , which is exacerbated in the presence of gap junctions ( S5 Fig ) . Ornstein-Uhlenbeck ( OU ) is a noise process that ensures that the mean current delivered is well behaved and that the integral of delivered current over time converges to a constant value [95] . The OU current is a good approximation for synaptic inputs originating in a large number of uncorrelated sources , where synaptic events are generated randomly and each event decays with a given rate ( τ ) . We use a recursive implementation according to the following recursive formula: ηi ( t+1 ) =ηi ( n , t ) *exp ( −δ/τ ) + ( 1/τ ) ( μ−ηi ( t ) ) +σ*√δ*ξi FORMULA 2 Where ηi ( t ) is the noise amplitude of neuron i at time t . The noise process is parameterized by τ , σ , μ where τ represents the synaptic decay time constant , δ is the integration step time for our forward Euler integrator , σ is the standard deviation of the noise process and μ is its mean . The random draw from a Gaussian distribution at every time step is represented by ξi . Neurons in the inferior olive receive broad arborizations , leading to input correlations across nearby neurons . In our model this is represented via a mixture of an independent process for each neuron nindependent and a shared process , nall , common to all the neurons in the network , parameterized by a mixing parameter α , called 'noise correlation': ni ( t ) =α*nindependent+ ( 1−α ) *nall ( t ) FORMULA 3 Simulation results throughout the article come from simulations with noise use an α where neurons share 10% of their noise input . Reported results are qualitatively robust to changes in this value ( S5 Fig ) . To examine the dependence of network dynamics on the characteristics of the incoming input , we computed the 200 neuron network sweeping a grid of the main input parameters ( τ , σ , μ , α ) of the Ornstein-Uhlenbeck noise process . The network response in terms of STO frequency , population firing rates , proportion of firing neurons was analyzed with respect to a grid of input parameters . For comparability of statistics and reproducibility of results , all results displayed in this article were obtained from a single random seed . We have tested the network with multiple seeds and the results are qualitatively indistinguishable . The parameters of the Ornstein-Uhlenbeck process were tuned such that the network emulating the wildtype network ( with gap junctions , WT ) produced an average frequency of 1 Hz and more than 95% of the model cells fire at least once every 5 seconds ( the parameter space for the network responses including STO , population firing rate and proportion of cells that fire within 3s is found in S5B Fig ) . The parameters to achieve these criteria are dependent on the total leak through the gap junctions . There are multiple methods to compensate the absent leak in the gapless network . In the present case , the network without gap junctions has been tuned to produce the same firing frequency as the network with gaps by increasing the membrane leak currents from 0 . 010 to 0 . 013 mS/cm2 . This results in a similar excitability but slightly lower STO frequency in the “mutant” . The average firing rate behavior of the network shows a linear relationship with the standard deviation of the OU process ( S4 Fig ) . For the present network with balanced connectivity and a single gap conductance of 0 . 04 mS/cm2 , the Ornstein-Uhlenbeck parameters are ( τ , σ , μ , α ) , μ = −0 . 6 pA/cm2 , σ = 0 . 6 pA/cm2 and τ = 20 ms . τ is a decay parameter that represents the synaptic decay times expected for olivary inputs , in this case chosen to emulate dendritic GABA according to Devor and Yarom [124] . Both synchrony and instantaneous frequency were estimated on the basis of a novel phase transformation of the membrane potential , which is more robust than the standard Hilbert transform , and can produce a linear phase response to the non-linear shape of the subthreshold oscillations [125] . This transformation improves the estimation of the momentary phase and compensates for the fact that ionic mechanisms induce different rates of membrane potential change at different phases of the oscillation . This phases analysis was conducted with the DAMOCO toolbox [126] . From the instantaneous phase , the instantaneous frequency is simply the inverse of the first order finite difference of phase . Synchrony across cells is estimated with the Kuramoto order parameter ( K ) : K ( t ) =|1N∑ei ( Ψ ( t ) −ϕn ( t ) ) | FORMULA 4 Where ϕn is the phase of each neuron , N is the number of neurons and Ψ is the phase average of all oscillators . To estimate a phase response curve of the stimulated neurons , first a “sensory stimulus” is delivered at a phase known to produce an action potential ( and resetting ) . The location of the first peak after stimulation is recorded . Subsequently , eight more simulations receive another stimulus , with same parameters as the resetting stimulus , but at different phases ( at incremental intervals of 2π/8 ) . The effect of that stimulation ( delay or advance ) on the next peak is recorded as a phase delta . Results are plotted in Fig 7A–7E .
|
Activity of the inferior olive , transmitted via climbing fibers to the cerebellum , regulates initiation and amplitude of movements , signals unexpected sensory feedback , and directs cerebellar learning . It is characterized by widespread subthreshold oscillations and synchronization promoted by strong electrotonic coupling . In brain slices , subthreshold oscillations gate which inputs can be transmitted by inferior olivary neurons and which will not—dependent on the phase of the oscillation . We tested whether the subthreshold oscillations had a measurable impact on temporal patterning of climbing fiber activity in intact , awake mice . We did so by recording neural activity of the postsynaptic Purkinje cells , in which complex spike firing faithfully represents climbing fiber activity . For short intervals ( <300 ms ) many Purkinje cells showed spontaneously rhythmic complex spike activity . However , our experiments designed to evoke conditional responses indicated that complex spikes are not predominantly predicated on stimulus history . Our realistic network model of the inferior olive explains the experimental observations via continuous phase modulations of the subthreshold oscillations under the influence of synaptic fluctuations . We conclude that complex spike activity emerges from a quasiperiodic rhythm that is stabilized by electrotonic coupling between its dendrites , yet dynamically influenced by the status of their synaptic inputs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"olives",
"cell",
"physiology",
"action",
"potentials",
"medicine",
"and",
"health",
"sciences",
"neural",
"networks",
"nervous",
"system",
"membrane",
"potential",
"junctional",
"complexes",
"electrophysiology",
"neuroscience",
"gap",
"junctions",
"network",
"analysis",
"plants",
"computer",
"and",
"information",
"sciences",
"animal",
"cells",
"fruits",
"cellular",
"neuroscience",
"eukaryota",
"cell",
"biology",
"anatomy",
"synapses",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"neurophysiology",
"purkinje",
"cells",
"organisms"
] |
2019
|
Quasiperiodic rhythms of the inferior olive
|
Visually evoked signals in the retina pass through the dorsal geniculate nucleus ( dLGN ) on the way to the visual cortex . This is however not a simple feedforward flow of information: there is a significant feedback from cortical cells back to both relay cells and interneurons in the dLGN . Despite four decades of experimental and theoretical studies , the functional role of this feedback is still debated . Here we use a firing-rate model , the extended difference-of-Gaussians ( eDOG ) model , to explore cortical feedback effects on visual responses of dLGN relay cells . For this model the responses are found by direct evaluation of two- or three-dimensional integrals allowing for fast and comprehensive studies of putative effects of different candidate organizations of the cortical feedback . Our analysis identifies a special mixed configuration of excitatory and inhibitory cortical feedback which seems to best account for available experimental data . This configuration consists of ( i ) a slow ( long-delay ) and spatially widespread inhibitory feedback , combined with ( ii ) a fast ( short-delayed ) and spatially narrow excitatory feedback , where ( iii ) the excitatory/inhibitory ON-ON connections are accompanied respectively by inhibitory/excitatory OFF-ON connections , i . e . following a phase-reversed arrangement . The recent development of optogenetic and pharmacogenetic methods has provided new tools for more precise manipulation and investigation of the thalamocortical circuit , in particular for mice . Such data will expectedly allow the eDOG model to be better constrained by data from specific animal model systems than has been possible until now for cat . We have therefore made the Python tool pyLGN which allows for easy adaptation of the eDOG model to new situations .
Visually evoked signals pass the dorsal geniculate nucleus ( dLGN ) on the route from retina to primary visual cortex in the early visual pathway . This is however not a simple feedforward flow of information , as there is a significant feedback from primary visual cortex back to dLGN . Cortical cells feed back to both relay cells and interneurons in the dLGN , and also to cells in the thalamic reticular nucleus ( TRN ) which in turn provide feedback to dLGN cells [1 , 2] . In the last four decades numerous experimental studies have provided insight into the potential roles of this feedback in modulating the transfer of visual information in the dLGN circuit [3–19] . Cortical feedback has been observed to switch relay cells between tonic and burst response modes [20 , 21] , increase the center-surround antagonism of relay cells [16 , 17 , 22 , 23] , and synchronize the firing patterns of groups of such cells [10 , 13] . However , the functional role of cortical feedback is still debated [2 , 24–30] . Several studies have used computational modeling to investigate cortical feedback effects on spatial and/or temporal visual response properties of dLGN cells [31–38 , 53] . These have typically involved numericallyexpensive dLGN network simulations based on spiking neurons [31–33 , 35 , 38] or models where each neuron is represented as individual firing-rate unit [36 , 37] . This is not only computationally cumbersome , but the typically large number of model parameters in these comprehensive network models also makes a systematic exploration of the model behavior very difficult . In the present study we instead use a firing-rate based model , the extended difference-of-Gaussians ( eDOG ) model [39] , to explore putative cortical feedback effects on visual responses of dLGN relay cells . A main advantage with this model is that visual responses are found from direct evaluation of two-dimensional or three-dimensional integrals in the case of static or dynamic ( i . e . , movie ) stimuli , respectively . This computational simplicity allows for fast and comprehensive study of putative effects of different candidate organizations of the cortical feedback . Taking advantage of the computational efficiency of the eDOG model , we here explore effects of direct excitatory and indirect inhibitory feedback effects ( via dLGN interneurons and TRN neurons ) on spatiotemporal responses of dLGN relay cells . In particular we investigate effects of ( i ) different spatial spreads of corticothalamic feedback and ( ii ) different corticothalamic propagation delays . Our analysis suggests that a particular mix of excitatory and inhibitory cortical feedback agrees best with available experimental observations . In this configuration an ON-center relay cell receives feedback from ON-center cortical cells ( ON-ON feedback ) , consisting of a slow ( long-delay ) and spatially widespread inhibitory feedback combined with a fast ( short-delay ) and spatially narrow excitatory feedback . Here the inhibitory and excitatory ON-ON feedback connections are accompanied by excitatory and inhibitory OFF-ON connections , respectively , following a phase-reversed arrangement [38] . For one this feedback organization accounts for the feedback-induced enhancement of center-surround antagonism of relay cells as observed in experiments [16 , 17 , 22 , 23 , 38] . Further , it seems well suited to dynamically modulate both the center-surround suppression and spatial resolution , for example , to adapt to changing light conditions [40] . Moreover , a longer thalamocortical loop time of ON-ON inhibitory feedback loop compared to ON-ON excitatory feedback may contribute to temporal decorrelation of natural stimuli [41] , an operation that has been observed accomplished at the level of dLGN in the early visual pathway [42] . At the same time , the rapid excitatory feedback may contribute to linking stimulus features by synchronizing firing of neighboring relay cells [10 , 19] . Previous experimental studies have focused on cat , monkey , and ferret dLGN , and the present model was adapted to neurobiological findings from cat . However , the last years have seen a surge of interest in mouse visual system , where new optogenetic and pharmacogenetic methods provide new tools for precise manipulation of identified neurons in the thalamocortical circuit [43–48] . Such data will expectedly allow for a detailed adaptation of the eDOG model to rodent dLGN , likely much better constrained by biological findings than what has been possible until now for cat . To facilitate this we have made the Python tool pyLGN ( http://pylgn . rtfd . io ) which allows for easy modification and evaluation of the eDOG model to new situations .
Spike responses of neurons in the early visual pathway are most commonly described in terms of receptive fields . Mathematically , the spatiotemporal receptive field is defined by an impulse-response function W ( r , t ) . This function describes the firing-rate response to a tiny ( δ-function ) spot positioned at r = 0 which is on for a very short time ( δ-function ) at t = 0 . If linearity is assumed , the response to any stimulus S ( r , t ) can be found by convolving the impulse-response function with the stimulus [39 , 49–52]: R ( r , t ) =∫τdτ ∫∫r′d2r′ W ( r−r′ , τ ) S ( r′ , t−τ ) , ( 1 ) or written more compactly R ( r , t ) = W * S . ( 2 ) Here S ( r , t ) is a spatiotemporal stimulus function describing , e . g . , the light intensity on a screen as a function of time and position . R ( r , t ) is the response of a neuron with its receptive-field center at r . The spatial integral goes over the whole visual field , i . e . , over all two-dimensional space . For mathematical convenience we have chosen the temporal integral to go from τ = −∞ to + ∞ . Since a stimulus input cannot affect the response in the past , it then follows that W ( r , τ < 0 ) = 0 . In Fourier space the convolution in Eq ( 2 ) corresponds to a product R ˜ ( k , ω ) = W ˜ ( k , ω ) S ˜ ( k , ω ) , ( 3 ) where R ˜ , W ˜ , and S ˜ are the Fourier transforms of the neural response R , the impulse-response function W , and the stimulus S , respectively . The tilde symbol ( ∼ ) will be used to denote the Fourier transform of any function throughout this paper . The function argument k is the wave vector which is related to the spatial frequency ν via |k| = 2πν . Correspondingly , the angular frequency ω is related to the temporal frequency f via ω = 2πf . With W ˜ and S ˜ known , the neural response can thus always be found by an inverse Fourier transform F - 1 { } , which entails an integral over temporal and spatial frequencies R ( r , t ) = F - 1 { W ˜ ( k , ω ) S ˜ ( k , ω ) } . ( 4 ) The response model in Eq ( 1 ) is an example of a descriptive model where the purpose is to summarize experimental data compactly in a mathematical form [51–53] . Here the aim is to find an appropriate impulse-response function , i . e . , spatiotemporal receptive-field function , that describes the measured neural response to different visual stimuli [50] . With this approach , however , limited insight is gained into how the neurons and neural circuitry in the early visual system provide such a receptive field . To address this question a mechanistic receptive-field model is needed . For a discussion of the difference between descriptive and mechanistic models in visual neuroscience , see [53 , 54] . In mechanistic LGN-circuit models the input from retinal ganglion cells have been described by descriptive models , see , e . g . , [36 , 37 , 39 , 52 , 55] . Likewise , in the present eDOG model the input from retinal ganglion cells is represented by the descriptive impulse-response function ( Eq ( 1 ) ) . Here a square grid of retinal ganglion cells with identical , spatially-localized receptive fields are considered ( see Fig 1 ) . The activity , i . e . , firing rates , of the neurons on the retinal ganglion cell layer then serves as input to the dLGN relay cell layer . This is represented by a spatiotemporal coupling-kernel function KRG , which reflects the direct synaptic input from retinal ganglion cells to dLGN relay cells . The coupling kernel , which is analogous to the descriptive impulse-response function in Eq ( 1 ) , is assumed to only depend on the relative distance between the cells in the visual field [52] . The response of a relay cell located at r is then given by [39 , 52]: RR ( r , t ) =∫τdτ ∫∫r′d2r′ KRG ( r−r′ , τ ) RG ( r′ , t−τ ) =KRG*RG , ( 5 ) where RR and RG are the firing-rate responses of relay cells and ganglion cells , respectively . The coupling kernel KRG ( r − r′ , τ ) denotes the strength with which the response of a ganglion cell , displaced by r − r′ from the relay cell , at time t − τ influences the response of the latter at time t . Note that , KRG ( r , τ < 0 ) = 0 due to causality . In Fourier space the relationship in Eq ( 5 ) can be written as W ˜ R S ˜ = K ˜ R G W ˜ G S ˜ , ( 6 ) where we have used the general relationship in Eq ( 3 ) . The key point here is that a descriptive model for the relay-cell impulse-response function W ˜ R now has a mechanistic interpretation . This relation is given as the product of the impulse-response function W ˜ G of the retinal ganglion cells and the coupling kernel K ˜ R G from the former cell type to the latter . In the eDOG model this approach is extended to include the various feedforward and feedback connections affecting the relay-cell response . The result is an expression for the relay-cell impulse-response function W ˜ R in terms of the impulse-response function W ˜ G of the retinal input and the coupling kernels connecting the neurons of the circuit . With such a mechanistic expression for W ˜ R , the response to any visual stimulus can be computed by means of the inverse Fourier transform in Eq ( 4 ) . Here we derive the impulse-response function for dLGN relay cells for the mechanistic eDOG model [39] . The complete circuit is shown in Fig 2 . In this figure each cell type correspond to a two-dimensional layer ( or population ) of identical cells . We will in the following focus on the dLGN relay cells with ON symmetry , but a similar model can be constructed for OFF-symmetry cells . These neurons receive feedforward excitation and indirect feedforward inhibition ( via intrageniculate interneurons ) from ON-center ganglion cells in retina . The relay cells further receive cortical feedback from both cortical ON cells and cortical OFF cells . In order to allow for easy exploration of the eDOG model and in particular effects of cortical feedback on relay-cell responses , we have developed an efficient , firing-rate based simulator of spatiotemporal responses in the early visual system . The simulator is named pyLGN and is written in Python . The design goals for pyLGN are to provide a software framework for studying the cortical feedback effects that is easy to use , extensible , and open . To facilitate usability , pyLGN has its own documentation page including installation instructions , several usage examples , and technical aspects ( http://pylgn . rtfd . io ) . To achieve extensibility , object-oriented programming is used , making it possible for the user to define new connectivity kernels and input stimuli . Lastly , to support openness pyLGN is both open-source and multi-platform . All calculations presented in this paper have been tracked using the Python software Sumatra [76] , which is an automated tracking tool for computational simulations and analysis . The source code for all presented simulations is available at ( https://github . com/miladh/edog-simulations ) .
As a first example application we focused on the effect of cortical feedback on the spatial response properties of dLGN cells . A specific focus was on so-called area-response curves , i . e . , responses to circular spots and patch-gratings as a function of stimulus size , which has received substantial experimental attention . In particular , studies have reported several effects of cortical feedback including sharpening of the receptive field by enhancing the center-surround antagonism of relay cells , increased receptive field center size with removal of feedback , and increased peak response to a an optimal diameter stimulus [17 , 22 , 23] . Other studies have reported a more diverse influence from corticothalamic feedback , including both facilitatory and suppressive effects on dLGN cell responses , and changes in the receptive-field structure [18] . Further , several studies have indicated that cortical feedback plays a role in extraclassical suppression in LGN [8 , 98 , 99] . A recent study confirms that extraclassical suppression indeed involves extraretinal mechanisms with a delay that is consistent with polysynaptic inhibition [100] . However , the early phase of the suppression is unlikely to involve cortical feedback , as the onset of the suppression is too fast , even though the cortical feedback may contribute in later phases of suppression . Our model demonstrated that cortical feedback can , depending on the feedback configuration , both enhance and suppress center-surround antagonism and both increase and decrease the receptive-field center size of relay cells . While the receptive-field center size decreases and the center-surround antagonism ( as measured by the suppression index αs ) increases with increased ( indirect ) cortical inhibitory feedback , the opposite is seen for excitatory feedback ( Fig 7 ) . These results support that a phase-reversed arrangement of the cortical feedback , where the ON-ON feedback is inhibitory while the OFF-ON feedback is excitatory , as suggested by data from [73] , is more effective to enhance the center-surround antagonism of relay cells as observed in experiments [16 , 17 , 22 , 23 , 38] . However , with this arrangement a reduction in response to the optimal diameter stimulus in the size tuning curves was observed in our model ( Fig 7 ) , in contrast to some experimental studies where an increase in response was reported [17] Here we also considered the more complicated mixed phase-reversed feedback situation with a spatially broad ON-ON inhibitory feedback ( combined with a corresponding OFF-ON excitatory feedback ) and a spatially narrow ON-ON excitatory feedback ( combined with a corresponding OFF-ON inhibitory feedback ) . Such a center-surround spatial organization of the feedback with excitatory bias to center and an inhibitory bias to the surround has been seen experimentally [3 , 16] . In our model studies such mixed feedback was seen to give increased center-surround antagonism compared to the situation with ON-ON inhibitory feedback alone . Further , this configuration could also both reduce the size of the optimal stimulus diameter , as well as increase the magnitude of the response to the optimal stimulus diameter ( Fig 7 ) . Correspondingly , with this configuration a sharper band-pass property of the spatial-frequency spectra was observed ( Fig 8 ) . As for the spatial response properties , the effects of ON-ON inhibitory and ON-ON excitatory feedback ( accompanied by the corresponding phase-reversed OFF-ON feedback ) are seen to be quite distinct . While delayed inhibitory feedback makes the impulse response more biphasic , the opposite is the case for delayed excitatory feedback ( Fig 11 ) . Likewise , while the temporal frequency tuning becomes sharper with delayed inhibitory feedback , it becomes blunter with delayed excitatory feedback ( Fig 13 , top row ) . These features , i . e . , increased biphasic index and sharper temporal frequency tuning , are maintained also for the case of mixed cortical feedback as long as the thalamocortical loop delay for the inhibitory feedback is much larger than for the excitatory feedback ( Figs 12 and 13 , bottom row ) . Such a mixed-feedback configuration is also found to be particularly suited to remove temporal correlation in the stimulus and thus reduce the temporal redundancy in the neural signals that are sent from dLGN relay cells to cortex ( Fig 14 ) . Our results concerning the spatial and temporal feedback effects suggests that a situation with a mixed organization of cortical feedback consisting of a slow ( long-delay ) and spatially widespread ON-ON inhibitory feedback , combined with a fast ( short-delay ) and spatially narrow ON-ON excitatory feedback may have particular advantages . Here the inhibitory and excitatory ON-ON feedback connections are accompanied by excitatory and inhibitory OFF-ON connections following a phase-reversed arrangement [38] . This specific prediction of a mixed organization of feedback could be tested experimentally , for example , by pharmacological inactivation ( or other means ) of specific feedback connections in the circuit . This feedback organization seems well suited to dynamically modulate both the center-surround suppression and spatial resolution , for example , to adapt to changing light conditions where the most efficient neural representation of the stimulus is expected to vary depending on the signal-to-noise ratio [40] . In particular , for high light levels ( i . e . , high signal-to-noise ) a band-pass like spatial spectrum ( as obtained with our model for certain parameter choices ) is expected to provide the most efficient coding , while for low light levels ( low signal-to-noise ) a low-pass spatial spectrum ( as obtained with our model for some other parameter choices ) seems better ( see Sec . 3 . 6 . 1 in [101] ) Further , a longer thalamocortical loop time of ON-ON inhibitory feedback compared to ON-ON excitatory feedback assures that temporal correlations in the natural visual stimuli are reduced in the relay-cell responses ( Fig 14 ) . This temporal feedback arrangement gives a large biphasic index ( Fig 12 ) which previously has been shown to provide temporal decorrelation of natural stimuli [41] , a feature that has also been seen in experiments [42] . An increased biphasic index in the presence of cortical feedback , has also been seen in other modeling studies [37 , 102] . In particular , a study using a predictive coding model , where the phase-reversed configuration emerged during training , reported that stronger biphasic response in LGN may result from ( predictive ) cortical feedback interactions [102] . While the slow inhibitory feedback is key for providing temporal decorrelation , the rapid excitatory feedback may have a role in linking stimulus features by synchronizing firing of neighboring relay cells to provide a strong input to cortical target cells [10 , 19] . Interestingly , a recent study found a large variation in axonal conduction times for corticothalamic axons , from a few milliseconds to many tens of milliseconds [84] . This suggests that differences in feedback delays indeed may have a functional role . It should be noted that the eDOG-model [39] on which pyLGN is based , assumes that the cortical feedback has a phase-reversed arrangement where each ON-ON feedback connection is accompanied by a phase-reversed OFF-ON feedback connection ( Eq ( 11 ) ) , i . e . , a push-pull arrangement as experimentally observed in [73] . An alternative is a phase-matched arrangement where relay cells receive feedback only from cortical cells with the same symmetry , including both the direct excitatory feedback and the indirect inhibitory feedback . However , such an arrangement is not only at odds with the observations in [73] , but also fails to explain the experimentally observed cortical-feedback induced increase in center-surround antagonism [38] . The aim of present work has been to introduce the simulation tool pyLGN and to provide a variety of example results to illustrate its use . The simulation tool is obviously not limited to the presented example applications , and here we discuss future applications of the tool , as well as its limitations . Studies over the last decades have shown that neurons in dLGN are not simple relays , and play much more important roles in information processing than previously appreciated [19 , 30 , 44 , 100] . ( With this in mind , referring to dLGN neurons as ‘relay cells’ could be misleading , and a renaming could be in order . If so , a more neutral term to use could for instance be ‘thalamocortical cells’ . ) Compared to the primary visual cortex ( V1 ) , i . e . , the next station in the early visual pathway , the dLGN has received relatively little attention from computational neuroscientists [113] . From a modeling strategy point of view , this is somewhat unfortunate as progress towards a mechanistic understanding of the function of the dLGN circuit seems more attainable given that ( i ) the dLGN circuit involves much fewer neuron types and is more comprehensively mapped out [1 , 30] , and that ( ii ) the dLGN has much fewer neurons making simulations computationally less intensive ( 18000 neurons in dLGN vs . 360000 neurons in V1 in mouse [114 , 115] ) . Further , the strong recurrent interactions characteristic for cortical networks ( which make them difficult to understand and analyze ) appear absent between the principal cells , that is , the relay cells , in the dLGN , even if key circuit network motifs such as feedforward and feedback interactions are present . Thus a focused and comprehensive effort on mechanistic modeling of the dLGN circuit would not only be of interest in itself , it would also likely be a very useful stepping stone for later attempts to model the visual cortex . While network simulations based on integrate-and-fire type neuron models ( e . g . , [96 , 110] ) and biophysically-detailed neuron models ( e . g . , [38 , 97] ) for entire dLGN nuclei are becoming computationally feasible with modern computers , there will still be a need for conceptually and mathematically simpler network simulation tools such as the present pyLGN tool based on the eDOG model . Such models are important to gain intuition about how the different circuit components may affect the overall circuit behavior , and will also be important for guiding the choice of the numerous , typically unknown , parameters in more comprehensive dLGN network simulations . Thus we envision that a future mathematics-based understanding of the dLGN circuit will be of a ‘multiscale’ nature and be based on a set of interconnected models at different levels of biophysical detail . The mouse seems particularly suitable as model animal since construction and testing of the multiscale models can be greatly facilitated by the ever more sophisticated techniques for controlling gene expression in mice as well as the possibility for optogenetic activation [114 , 116] . There are however differences in the way the mouse dLGN is organized compared to monkey and cat , which the current study has been based on , that should be considered [30] . For instance , although there are regions in the mouse dLGN clustered by morphology and ocular dominance , there are no clear laminations ( except for a core and a dorsal shell region ) , and it has been difficult to identify functionally distinct classes of thalamocortical cells [117] . With respect to the model presented in this work , the main challenge is the lack of experimental data on whether the cortical feedback in rodents is organized in a phase-reversed arrangement ( ON-ON feedback connections accompanied by OFF-ON connections ) as observed in cats [73] . Experimental recordings to characterize this will be of particular interest since several computational studies have based their model on this organization , and in one study where the biphasic responses in LGN was explained using a predictive coding model , this arrangement of feedback emerged during training of the network [37 , 93 , 102] . We envision that at a close and targeted collaboration between modelers and experimentalist , in particular direct assessment of predictions from mechanistic models in targeted experiments , holds great promise for unraveling mechanisms of visual information processing at the different levels of advancements in visual information processing from the earliest sensory systems to more complex computations of the higher cortical areas .
|
On route from the retina to primary visual cortex , visually evoked signals have to pass through the dorsal lateral geniculate nucleus ( dLGN ) . However , this is not an exclusive feedforward flow of information as feedback exists from neurons in the cortex back to both relay cells and interneurons in the dLGN . The functional role of this feedback remains mostly unresolved . Here , we use a firing-rate model , the extended difference-of-Gaussians ( eDOG ) model , to explore cortical feedback effects on visual responses of dLGN relay cells . Our analysis indicates that a particular mix of excitatory and inhibitory cortical feedback agrees best with available experimental observations . In this configuration ON-center relay cells receive both excitatory and ( indirect ) inhibitory feedback from ON-center cortical cells ( ON-ON feedback ) where the excitatory feedback is fast and spatially narrow while the inhibitory feedback is slow and spatially widespread . In addition to the ON-ON feedback , the connections are accompanied by OFF-ON connections following a so-called phase-reversed ( push-pull ) arrangement . To facilitate further applications of the model , we have made the Python tool pyLGN which allows for easy modification and evaluation of the a priori quite general eDOG model to new situations .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Discussion"
] |
[
"fourier",
"analysis",
"engineering",
"and",
"technology",
"signal",
"processing",
"social",
"sciences",
"vertebrates",
"neuroscience",
"animals",
"mammals",
"kernel",
"functions",
"mathematics",
"ganglion",
"cells",
"amniotes",
"interneurons",
"vision",
"research",
"and",
"analysis",
"methods",
"operator",
"theory",
"animal",
"cells",
"mathematical",
"and",
"statistical",
"techniques",
"cellular",
"neuroscience",
"psychology",
"eukaryota",
"cell",
"biology",
"cats",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"afferent",
"neurons",
"physical",
"sciences",
"sensory",
"perception",
"retinal",
"ganglion",
"cells",
"organisms"
] |
2018
|
Firing-rate based network modeling of the dLGN circuit: Effects of cortical feedback on spatiotemporal response properties of relay cells
|
Experimental recordings in hippocampal slices indicate that astrocytic dysfunction may cause neuronal hyper-excitation or seizures . Considering that astrocytes play important roles in mediating local uptake and spatial buffering of K+ in the extracellular space of the cortical circuit , we constructed a novel model of an astrocyte-neuron network module consisting of a single compartment neuron and 4 surrounding connected astrocytes and including extracellular potassium dynamics . Next , we developed a new model function for the astrocyte gap junctions , connecting two astrocyte-neuron network modules . The function form and parameters of the gap junction were based on nonlinear regression fitting of a set of experimental data published in previous studies . Moreover , we have created numerical simulations using the above single astrocyte-neuron network module and the coupled astrocyte-neuron network modules . Our model validates previous experimental observations that both Kir4 . 1 channels and gap junctions play important roles in regulating the concentration of extracellular potassium . In addition , we also observe that changes in Kir4 . 1 channel conductance and gap junction strength induce spontaneous epileptic activity in the absence of external stimuli .
Temporal lobe epilepsy , which has a specific clinical presentation of seizures arising from the hippocampus [1] , is a serious health risk to affected individuals . Extensive experimental and theoretical studies have shown that elevated K+ concentration in the extracellular microenvironment ( [K+]o ) may be linked to spontaneous epileptic seizure activity in the absence of external stimuli [2–6] . Astrocyte-mediated K+ buffering generally serves to maintain [K+]o , supporting normal neuronal electrical activities [7–10] . Mechanisms for K+ removal , regulated by astrocytes , can be broadly categorized as K+ uptake and K+ spatial buffering [11 , 12] . In the case of K+ uptake , excess K+ is absorbed primarily by astrocytes through Na+/K+-ATPase pumps and inwardly rectifying Kir4 . 1 channels or through Na+-K+-Cl cotransporters . K+ is spatially buffered via gap junctions , which redistribute excess K+ taken up from astrocytes in areas of excessive neuronal activity to astrocytes with lower K+ concentrations within the astrocytic network . Experimental data suggest that Kir4 . 1 channels play a prominent role in local uptake of extracellular K+ [8 , 13 , 14] , and many experimental studies have observed that dysfunction of local uptake by astrocytes or inactivation of astrocytic gap junction protein expression cause the generation or spread of seizure activity [15–20] . It was recently reported that down-regulation or dysfunction of K+ uptake channels ( Kir4 . 1 ) in astrocytes induces tonic-clonic seizures [7] . For example , low-intensity Schaffer collateral stimulation generates epileptiform activity in gap junction-deficient mice . Another experiment found that Tsc1 ( TSC: Tuberous Sclerosis Complex ) inactivation in astrocytes caused defects in astrocytic gap junction coupling and potassium clearance , leading to epilepsy in Tsc1GFAPCKO mice [21] . Furthermore , expression of the astrocytic gap junction proteins connexin 43 ( Cx43 ) and connexin 30 ( Cx30 ) is altered in epilepsy , and changes in gap junction communication have been observed in sclerotic hippocampal tissue of epileptic animal models [22] . Therefore , we investigated the dynamic mechanisms by which dysfunction of K+ uptake or gap junctions leads to changes in extracellular K+ concentration and resultant pathological epileptic discharges using a theoretical study with a computational astrocytic-neural network model . Recent studies have applied theoretical analysis approaches to explore the relationship between astrocytic modulation of [K+]o and the initiation or maintenance of epileptic seizure activity [23–32] . Cressman J . R . et al . provided a simplified astrocytic K+ uptake function that incorporates both passive and active local K+ uptake into a single sigmoidal response function that depends solely on extracellular K+ concentration [23–27] . Hübel studied the mechanism of extracellular K+ uptake in astrocytes using a phenomenological equation for the rate of K+ uptake , which depends entirely on extracellular K+ concentration [30] . Øyehaug L et al . modeled the mechanism of K+ uptake from the perspective of volume dynamics [28 , 29] , modeling the cotransporters ( Na+/K+/2Cl and K+/Cl- cotransporters ) and the Na+/K+-ATPase pump of astrocytes . However , these model systems did not adequately reflect the complexity of astrocytes . For example , in addition to cotransporters and the Na+/K+-ATPase pump , the inwardly rectifying K+ channels ( Kir4 . 1 ) in astrocyte membranes and inter-astrocyte spatial buffering also contribute to extracellular K+ removal . Sibille J et al . developed mathematical models of Kir4 . 1 channels that depend on both the intra-astrocyte K+ concentration and the astrocytic membrane potential [31 , 32] . In addition , they verified the prominent role of Kir4 . 1 channels in regulating [K+]o during repetitive stimulation using their mathematical model . It is worth mentioning that none of the aforementioned studies considered the relationship between Kir4 . 1 channel conductance and the occurrence of spontaneous epileptic seizures . Morphologically , previous studies of the extracellular K+ local buffer mechanisms in astrocyte-neuron networks have shown that astrocytes have a 1:1 quantitative relationship with neurons; however , experimental data show that one cortical neuron is typically surrounded by multiple astrocytes [20 , 33 , 34] . In support of this , in 2009 , Azevedo et al . reported that the ratio of neurons to glia is approximately 1:4 in the human cerebrum [33] . In light of these findings , we constructed a model of an astrocyte-neuron network consisting of a single compartment neuron connected to four surrounding astrocytes with Kir4 . 1 channels and Na+/K+-ATPase pumps . Extracellular potassium ions diffused in and out of the space between the neuron and astrocytes . In addition , we developed a novel configuration for astrocytic gap junctions , connecting two astrocyte-neuron network modules . The model function parameters of the gap junction are based on a nonlinear regression fit of a set of experimental data previously published [17] . The model simulation results were validated by comparison with published experimental data [17] . We conducted a series of simulations using the single astrocyte-neuron network models and the coupled astrocyte-neuron module network and showed that changes in Kir4 . 1 channel conductance and gap junction strength induce spontaneous epileptic activity in the absence of external stimuli .
To validate the role of Kir4 . 1 channels in regulating astrocytic and neuronal extracellular K+ concentration , a model was constructed representing astrocyte-neuron network modules consisting of a single compartment neuron connected to four surrounding astrocytes with Kir4 . 1 channels and Na+/K+-ATPase pumps ( Fig 1 ( a ) ) . Based on evidence from previous studies [23 , 25 , 26 , 30 , 31] , there exist several key factors mediating K+ concentration in the extracellular space: K+ currents across the neuronal membrane , Na+/K+-ATPase pumps for neurons and astrocytes , Kir4 . 1 channels in astrocytes , spatial diffusion of K+ in the extracellular space , and others ( illustrated in Fig 1 ( a ) ) . During an action potential , K+ is released by neurons at the extracellular space , followed by K+ flowing into adjacent astrocytes via the Na+/K+-ATPase pump and Kir4 . 1 channels . It is important to note that the Kir4 . 1 current is larger in magnitude than the outward Na+ current from the Na+/K+-ATPase pump , resulting in overall depolarization of the astrocytic membrane ( Fig 1 ( b ) ) . Effects of Kir4 . 1 channel blockage on extracellular K+ regulation were first tested by setting the Kir4 . 1 conductance to 45 . 0 pS for the normal ( control ) condition . Sine stimulus inputs ( the stimulus amplitude and frequency are 1μA/cm2 and 10 Hz , respectively; here , 10 Hz was used to verify our results in comparison with previous experimental recordings [35] ) with different durations , e . g . , 30 s , 10 s and 100 ms , were applied to the neuron when Kir4 . 1 channels were in the control state . Note here that a 100 ms stimulus resulted in a small undershoot and a very short recovery to the [K+]o baseline . As the duration increased , both the undershoot area and recovery time increased . Compared with 100 ms , the 10 s and 30 s stimulations required significantly longer times to recover to the baseline , as shown in Fig 1 ( c ) . This result is consistent with an experimental trend observed in vivo in the mouse hippocampus [35] . We next tested our model with Kir4 . 1 channel inhibition ( gkirA = 0 . 01 pS ) . We observed that extracellular K+ concentration displayed a more pronounced undershoot after long-lasting sine stimulation ( 10 Hz , 10 s ) . The maximum amplitude of extracellular K+ concentration for gkirA = 0 . 01 pS was higher than that for gkirA = 45 . 0 pS during a 10 s external stimulation input ( see lower panel in Fig 1 ( d ) ) . During the 10 s post-stimulus recovery period , extracellular K+ initially dropped to below the baseline levels before returning gradually to the resting-state equilibrium concentration , taking even longer to recover to the K+ baseline concentration , as shown in the lower panel of Fig 1 ( d ) . Furthermore , it was observed that in response to stimulus under control conditions , astrocytes responded with a rapid increase in intra-astrocyte K+ concentration with an accompanying high amplitude , while the baseline K+ concentration was low . On the contrary , the K+ concentration in astrocytes rose much more slowly to a lower peak concentration when Kir4 . 1 was blocked , as shown in the top panel of Fig 1 ( d ) . These computational results were all in agreement with previously recorded experimental results [35 , 36] . Additionally , the critical value of gkir was examined for the generation of spontaneous epileptic seizures in the absence of external stimuli . Model simulations suggested that spontaneous periodic epileptic activity is induced when gkir< = 7 . 0 pS without external stimulation ( Fig 2 ( b ) ) , characterized by prolonged ( 2–15 s ) interruptions in population spike generation . During these interruptions , neuronal firing is suppressed rather than desynchronized . Bikson et al . have observed this depolarization block in pyramidal cells during electrographic seizures in rat hippocampal slices ( Fig 1d [3] ) . Fig 2 shows the time sequences of neuronal and astrocytic membrane potentials , time courses of K+ concentration in the extracellular space and intra-astrocytes for gkirA = 45 pS and 5 . 0 pS . These results indicate that neurons fire normally only when the Kir4 . 1 channel conductance is adequate . A low Kir4 . 1 channel conductance ( Fig 2 ( b ) ) results in spontaneous periodic seizures in the astrocyte-neuron network module . Fig 2 also shows that spontaneous periodic epileptic firing is accompanied by higher amplitude oscillations of extracellular K+ concentrations and lower peak intra-astrocyte K+ concentrations compared with normal neuronal discharges . Fig 3 presents a new model in which astrocyte gap junctions connect two astrocyte-neuron network modules . This network model was constructed to examine the K+ spatial buffering mechanisms dominated by astrocytic gap junctions . In this model , individual neurons are surrounded by 4 astrocytes , and they share the same bath K+ concentration . During action potential generation , neurons release K+ into the extracellular space , and K+ enters adjacent astrocytes via the Na+/K+-ATPase pump and Kir4 . 1 channels . Subsequently , in astrocytes with higher concentrations , K+ may flow into adjacent astrocytes via inter-astrocytic gap junctions due to the electrochemical driving force . To validate whether this type of inter-astrocytic gap junction behaves the same way as those in experimental observations , the decay dynamics of extracellular K+ concentration ( [K+]o ) was examined in response to the sine stimulus trains ( 10 s , 20 Hz ) . We chose 20 Hz rather than the 10 Hz used above in order to compare the results in our model with previously published experimental results ( Fig . 5B in [17] ) . The results demonstrated that after the stimulus trains , [K+]o initially decayed very rapidly , followed by a prolonged phase of slower decay , often associated with an undershoot [37] ( Fig 4 ( a ) ) . To quantify the faster K+ decay phase , the time at which the [K+]o amplitude has decayed to 1/e of its initial maximal amplitude is defined as the decay time constant ( t1/e ) , providing a descriptive parameter for the decay process . Anke Wallraff [17] observed an inverse correlation between t1/e and [K+]o , which is well fitted by a power-law function ( Fig 4 ( b ) ) . Fig 4 ( b ) illustrates the experimentally observed fitted curve functions t1/e = 4 . 0591exp ( -0 . 0735[K+]o ) for normal gap junctions and t1/e = 4 . 4218exp ( -0 . 0624[K+]o ) for deficient gap junctions obtained from rat hippocampal slice recordings . Clearly , deficient gap junctions are characterized by a larger decay time constant , suggesting slower removal of K+ load [17] . Here , the proposed model simulated this process by applying the sine stimulus trains ( 10 s , 20 Hz , the stimulus amplitude is randomly chosen from 1 to 60μA/cm2 to induce different spiking patterns and potassium diffusion dynamics ) to neuron 1 ( N1 ) . The K+ concentration Kbath , the spatial diffusion coefficient ε and the Na+/K+-ATP pump strength ρ were set to different values to study their effect on the correlations between the diffusion time constant t1/e and extracellular K+ concentration , as shown in Fig 4c . The Kir4 . 1 channel conductance was set to 45 . 0 pS for Fig 1c ( for values of the other parameters , see Table 1 ) . By setting the strength of the gap junction to F = 20 . 0 and 0 . 01 , the model simulation reproduced the wild-type control ( Fig 4c ) and dko deficient ( Fig 4d ) gap junctions for neuron 1 ( N1 ) . We then used the nonlinear least-squares method to fit the relation curve of t1/e and [K+]o that was obtained by the coupled astrocyte-neuron network modules . The fitted curve function for normal gap junctions is t1/e = 3 . 8179exp ( -0 . 0623[K+]o ) ( left panel , Fig 4 ( c ) ) . The fitted curve function for deficient gap junctions is t1/e = 4 . 2055exp ( -0 . 0569[K+]o ) ( right panel , Fig 4 ( c ) ) . The simulated fitted curve functions are quantitatively similar to experimental observations . In addition , to validate the physiological significance of the model constructed in this section and the reliability of our numerical results , statistical analysis was conducted on the error between the experimental and numerical fits for different gap junction strengths , as shown in Fig 4 ( d ) . Here , the fit error is the mean absolute error ( calculations shown in S1 Text ) . The fit error between the numerical and experimental data is far less than 10% . Considering the experimental accuracy and other environmental factors , we believe that the gap junction-mediated coupled model presented herein is within the permissible error range . In addition , different strengths of gap junctions were applied between astrocytes , e . g . , F = 20 and 0 . 01 ( Fig 4 ( e ) ) . The results show that decreased gap junction strength induces higher [K+]o peaks , and if the strength of gap junctions experiences a strong decrease , [K+]o will have a larger undershoot and an increased baseline recovery time . The above simulations applied the exponential diffusion function observed in experimental data reported by Anke Wallraff [17] . While this is a reasonable method to validate our overall model , it does not necessarily justify modeling gap junction diffusion as an exponential function of intracellular K+ . Therefore , we performed a new set of numerical simulation experiments to address this concern . We chose several different types of diffusion functions for our model network . For example , a linear function ( Ji−j = F ( [K+]Ai − [K+]Aj ) ( the numerical simulation results and regression fitting curve of the network model are shown in S1B Fig , compared with our original model results S1A Fig ) and a threshold-nonlinear function ( Ji−j = F * ( 1 + tanh ( ( |[K+]Ai − [K+]Aj|—Kthr ) /KScale ) ) ) ( see S1C Fig ) . The fitting error ranges of the three functions are shown in S1D Fig . The simulation results suggest that the dynamic properties of potassium ion channel diffusion faithfully follow the exponential decay function . Numerous experiments have observed that the coupling-deficient astrocytic network ( with depression of connexin43 expression or connexin30 expression ) in the nervous system is related to generation of spontaneous seizures [17 , 22 , 38–40] . To verify the relationship between the dysfunction of the gap junctions and spontaneous epileptic discharges in the absence of external stimuli , we used the coupled two astrocyte-neuron network modules . Model simulations suggested that spontaneous epileptic activity is induced when F< = 10 . 0 pS without external stimulation ( Fig 5 ( a ) ) . Fig 5 ( a ) shows the time sequences of the neuronal membrane potentials ( VN1 ) and the K+ concentration in the extracellular space ( [K+]o1 ) for F = 20 . 0 , 8 . 0 and 0 . 001 . The neuron ( N1 ) exhibited normal spontaneous slow oscillation activity when the gap junction strength F was equal to 20 . 0 . However , the activity changed into spontaneous epileptic seizures after certain time periods when the value of F decreased . For example , the neuron ( N1 ) changed into spontaneous seizures firing at 50 s when F was 8 . 0 . Specifically , neurons rapidly enter spontaneous epileptic discharge with smaller gap junction strengths . As such , the neuron ( N1 ) exhibited spontaneous seizure firing at 50 s when F was 8 . 0 , while entering spontaneous seizure firing at 24 s when F was 0 . 001 ( Fig 5 ( a ) ) . This type of spontaneous epileptic seizure has been observed in various experimental recordings [2 , 41 , 42] . Fig 5 ( b ) summarizes the schema of astrocytic gap junction strength F and [K+]o , or t1/e when there is no external stimuli to the neurons , the Kir4 . 1 channel conductance is 45 . 0 pS and other parameters are fixed , as in Table 1 . [K+]o and t1/e for N1 gradually decrease as F increases . These relationships can be used to explain the extracellular K+ concentration requiring more time to recover to the baseline when the gap junction is deficient . The above results confirm that decreasing astrocytic gap junction strength induces spontaneous epileptic seizures in the astrocyte-neuron network in the absence of external stimulation . Furthermore , neurons will rapidly enter into spontaneous epileptic seizures with smaller gap junction strength .
Potassium clearance is widely accepted as a primary function of astrocytes . Several computational models have investigated the buffering mechanism for extracellular K+ accompanied by neural activities [23 , 25 , 26 , 28 , 30 , 31 , 43] . A model accounting for K+ concentration in the extracellular space and astrocytic compartments has been used to quantify involved astrocytic ion channels and transporters ( Na+/K+-ATPase , NKCC1 , etc . ) [28 , 29] . While NKCCl has been recently shown to not be involved in activity-dependent K+ clearance in hippocampal slices [15] , astrocytic Kir4 . 1 channels are crucial for the recovery of basal extracellular K+ levels and neuronal excitability during external stimulation [31] . To unravel how the acute role of astrocytes in [K+]o homeostasis induces spontaneous epileptic discharge and maintains normal electrophysiological activity in the astrocyte-neuron system , we constructed an astrocyte-neuron network module consisting of a single compartment neuron connected to four surrounding astrocytes with Kir4 . 1 channels and Na+/K+-ATPase pumps . Extracellular potassium ions were allowed to diffuse in and out of the space between the neuron and astrocytes . In addition , we have built a new model for astrocytic gap junctions , enabling connection of two astrocyte-neuron network modules . The model function parameters were based on a nonlinear regression fit based on a set of experimental data published previously [17] . The model simulation results were validated by comparison with the published results [17] . We first verified that spontaneous periodic seizures can be induced directly in the absence of stimuli when the astrocytic Kir4 . 1 channel conductance or the strength of gap junction-mediated K+ spatial buffering decreases to a certain level . We first simulated the specific dynamic characteristics of astrocytic Kir4 . 1 channel conductance inducing spontaneous epileptic activity in the absence of external stimulus input . To mathematically examine K+ dynamics in neurons , astrocytes and extracellular spaces , we constructed an astrocyte-neuron network module consisting of a single compartment neuron connected to 4 surrounding astrocytes . The results demonstrated that the K+ undershoot of the extracellular space increases after a period of stimulus in Kir4 . 1 knockouts [15 , 35 , 44] . We verified these results using the proposed computational model compared to controls and found that K+ concentration in the extracellular space exhibits an increased undershoot after a 10 s sine stimulus sequence , taking a longer time to return to the baseline in Kir4 . 1 channel-deficient conditions . Moreover , the baseline K+ concentration in astrocytes is higher and rises more slowly to a decreased peak value after an external stimulus train compared to controls . The high baseline K+ concentration in astrocytes is due to the normal function of the Na+/K+-ATPase pump . In particular , spontaneous epileptic activity was directly induced in the single astrocyte-neuron network module without external stimuli when the conductance of Kir4 . 1 channels was lower than a certain threshold , further demonstrating that down-regulation of astrocytic Kir4 . 1 channels is closely related to neuropathological hyper-excitability [8 , 45 , 46] . At the same time , our single astrocyte-neuron network correctly predicts that when there is a partial blockage of Kir4 . 1 channels in astrocytes , rapid instantaneous neuronal seizure activity is induced . Moreover , it is possible that astrocytic Kir4 . 1 channels modulate extracellular K+ as a means of regulating synaptic activity . The dynamic characteristics of astrocytic gap junction strength pertaining to spontaneous epileptic activity in the absence of external stimuli were first simulated in this study . Gap junctions in astrocytes appear to play a dual role: on the one hand , they counteract the generation of hyperactivity by facilitating the decrease of extracellular K+ levels; on the other hand , they constitute a pathway for energetic substrate delivery that fuels neuronal hyper-excitability [16] . Thus far , research involving the effects of astrocytic networks coupled with gap junctions on extracellular K+ spatial buffering is still in the biological experimentation stage . Previous studies have examined the role of astrocytic gap junctions in potassium buffering with respect to epileptic seizures [17 , 22] . We presented a new exponential function model for astrocyte gap junctions that connect two astrocyte-neuron network modules . To verify the validity of the proposed model of gap junctions , we compared two groups of data of the fitted exponential function curves of t1/e vs [K+]o obtained from numerical simulation data based on the coupled astrocytic-neural network modules and previous experimental recordings [17] . Both control conditions and deficient gap junction cases were compared while sine stimulus trains were applied to neuron 1 ( N1 ) . We discovered that the curve fitting error of numerical and experimental data was within a reasonable range . In addition , the numerical results validated the finding that gap junction-mediated K+ spatial buffering is slower in the absence of astrocytic gap junctions compared to wild-type rats . At the same time , the neuron entered a spontaneous epileptic state more quickly at weaker gap junction strengths . Finally , we summarized the schema of the astrocytic gap junction strength F and [K+]o or t1/e when there is no external stimulation to either of the neurons , and this relationship can be used to explain the observation that extracellular K+ requires more time to return to the baseline when the gap junction is dysfunctional . Indeed , our results reproduced experimental observations wherein expression-deficient astrocytic gap junction proteins ( Cx43 or Cx30 ) distorted neuronal information processing and the generation of spontaneous epileptiform events [16 , 47] . Our coupled astrocyte-neuron network modules predict that with partially blocked astrocytic gap junctions , neuronal seizure activity will gradually develop after a delay of tens of seconds . This finding could represent a protective mechanism to prevent K+ accumulation from reaching neurotoxic levels . These results can be used as a basis for performing further analysis of the characteristics of sodium , calcium , the power factor and other factors in the network , as well as the mutual influences among them . In this study , we considered the gap junction between one of two adjacent astrocytes . In vivo many glial cells surrounding the firing neuron will buffer potassium into adjacent glial cells that have lower concentrations . We have chosen several other different diffusion functions ( including an exponential function , a linear function and a threshold-nonlinear function ) to simulate experimental data ( Fig 4c and S1 Fig ) . The simulation results suggest that the dynamic properties of potassium ion channels faithfully follow an exponential decay function . Therefore , we chose the exponential function for our model and all other simulations . Moreover , we examined several new model network configurations to test inverse correlations between t1/e and [K+]o . First , we assessed gap junctions between astrocytes in the same neuron-astrocyte module ( S2A Fig ) . In the simulation , we reduced the gap junction weight to 6 . 5 in order to match previous experimental parameters . The simulation results of our novel model ( S2C Fig ) are indeed quite similar to our original model with only one gap junction in a two module network ( S2B Fig ) . Therefore , either there are more gap junctions with a weak weight between glial cells or fewer gap junctions with a strong gap junction weight , and the network model reproduced established experimental results . Choosing 1:4 as the ratio of neurons to glial cells might be closer to actual physiology [33] . Second , we scaled up the network to a relatively large network with more neurons and glia , as well as increased gap junctions ( same gap junction weight as the original model in S3A and S3B Fig ) . The simulation results show that scaled up large networks produced similar experimental decay curves as the two module small network ( S3D and S3E Fig ) . Hence , these results suggest that the gap junction form and basic network configuration are robustly represented in this model . It is well established that astrocyte dysfunction causes hyper-excitation and the generation or spread of seizure activity; dysfunctional astrocytes should be considered promising targets for new therapies . This study aimed to provide an in-depth understanding of Kir4 . 1 channels and astrocytic gap junctions in regulating the extracellular K+ microenvironment during epileptic seizures in order to facilitate construction of more accurate and dynamic models of neuron-astrocyte networks to improve recognition , forecasting and control of epileptic seizures .
It is assumed that neurons have different local architectures . Each excitatory neuron model has INa , IK , leak current IL [43] , after-hyperpolarization current IAHP . The H-H-type dynamic equations for the two neurons are as follows: CdVNdt=−INa−IK−IL−IAHP+Iext ( 1 ) where , C is the membrane capacitance . The voltage-gated currents INa and IK , the leak current IL , and the after-hyperpolarization current IAHP in Eq ( 1 ) are: INa=−gNam3h ( VN−VNa ) IK=−gKn4 ( VN−VK ) IAHP=− ( gAHP[Ca]i1+[Ca]i ) ( VN−VK ) IL=−gNaL ( VN−VNa ) −−gKL ( VN−VK ) −gCl ( VN−VCl ) ( 2 ) Where gNa and gK denote the conductances corresponding to the sodium and potassium currents , respectively . gAHP is the conductance corresponding to the after-hyperpolarization current . gNaL and gKL are the conductances corresponding to the sodium leak current and potassium leak current , respectively . VNa , VK and VCl denote the Na+ , K+ and Cl- channel reversal potentials , respectively . n , m and h are gating variables for sodium and potassium currents . gCa and VCa are the conductance and the reversal potential of calcium , respectively . [Ca]i corresponds to the intracellular calcium concentration , and the dynamics equation is: d[Ca]idt=−0 . 002gCa ( VN−VCa ) /{1+exp ( − ( VN+25 ) /2 . 5 ) }−[Ca]i/80 ( 3 ) The equations for the gating variables in Eq ( 2 ) are: dqdt=φ[αq ( VN ) ( 1−q ) −βq ( VN ) q] , q=m , n , h ( 4 ) αm=0 . 1 ( VN+30 ) /[1−exp ( −0 . 1 ( VN+30 ) ) ]βm=4exp[− ( VN+55 ) /18]αn=0 . 01 ( VN+34 ) /[1−exp ( −0 . 1 ( VN+34 ) ) ]βn=0 . 125exp ( − ( VN+44 ) /80 ) αh=0 . 07exp ( − ( VN+44 ) /20 ) βh=1/[1+exp ( −0 . 1 ( VN+14 ) ) ] The reversal potentials of Na+ and K+ and Cl- are given by the Nernst equation: VNa=26 . 64ln ( [Na]o[Na]Ni ) VCl=26 . 64ln ( [Cl]Ni[Cl]o ) VK=26 . 64ln ( [K]o[K]Ni ) ( 5 ) where [Na]Ni and [Na]o denote the sodium ion concentrations in the intra-neuronal and extraneuronal spaces , respectively , The reversal potential of the Cl- current is equal to a fixed value , that is , VCl = −81 . 93 mV . [K+]Ni and [K+]o denote the potassium concentration in the intra-neuronal and extraneuronal spaces , respectively . The [K]o value is continuously updated by the K+ currents across the neuronal membrane , Na+/K+-ATP pumps of neuron , K+ spatial diffusion [6 , 23 , 48] , Na+/K+-ATP pumps and Kir4 . 1 channels of four astrocytes . The electrical current through cell membrane can cause ionic concentration of the inside and outside cell changing . A electrical current I across a membrane is equal to ion flow per unit of time . Thus , the K+ concentration dynamics for the neurons and astrocytes and the extra-neuronal space are described as follows: d[K+]odt=JIK−2Jpump , N−2Jpump , A1−2Jpump , A2−2Jpump , A3−2Jpump , A4+JKir , 1+JKir , 2+JKir , 3+JKir , 4−Jdiff ( 6 ) d[K+]A , idt= ( −Jkir , i+2JpumpA , i ) vrate2 ( 7 ) d[K+]Ndt= ( −JIK+2Jpump , N ) vrate1 ( 8 ) Similar to the K+ dynamics , The [Na]o value is continuously updated by the Na+ currents across the neuronal membrane , Na+/K+-ATPase pumps of neuron [6 , 23] , Na+/K+-ATP pumps of four astrocytes . However , the main differences are that the pump exchanges two K+ for three Na+ ions , leading to the coefficient 3 in front of the pump term . In addition to sodium concentrations was added two constant leak terms JNaLA and JNaLN , Thus , the Na+ concentration dynamics for neurons and astrocytes and the extraneuronal space are modeled as follows: d[Na+]o , idt=JNa , N+3Jpump , N+3Jpump , A1+3Jpump , A2+3Jpump , A3+3Jpump , A4+JNaL , N+3JNaL , A ( 9 ) d[Na+]A , idt= ( −3Jpump , Ai−JNaL , Ai ) vrate2 ( 10 ) d[Na]Ndt= ( −JNa , N−3Jpump , N−JNaL , N ) vrate1 ( 11 ) Also , different terms in the expressions ( 6–11 ) are described as follows: Jpump , N=ρ ( 11 . 0+exp ( 25 . 0−[Na]Ni ) /3 . 0 ) × ( 11+exp ( 8−[K]o ) ) Jdiff=ε ( [K]o−kbath ) Jpump , Ai= ( 13 ) ρ ( 11 . 0+exp ( 25 . 0−[Na]Ai ) /3 . 0 ) × ( 11+exp ( 8−[K]o ) ) ( 12 ) where , ρ is the pump strength of Na+/K+-ATP , [Na]Ni and [Na]Ai are the sodium concentration for neuron and astrocyte , respectively , ε is the spatial diffusion coefficient of K+ , and Kbath is the K+ concentration in the largest nearby reservoir . The equation of the astrocyte membrane potential VA is: CAdVAdt=−Ikir−IAL ( 13 ) Where CA is the astrocytic capacitance , the leak current IIA=gIA ( VA-VIA ) . The inward rectifier Kir4 . 1 channels current Ikir in astrocyte depends on the membrane potential and the extracellular potassium ion concentration , the expression can be described as [31 , 32]: IkirA=gkir[K+]o ( VA−VkirA ) VkirA=Ekirlog ( [K+]o/[K+]A ) ( 14 ) Where gkir and Ekir are the conductance and the Nernst constant for the astrocyte Kir 4 . 1 channels , respectively . The K+ dynamics in two astrocytes coupled by gap junction is described as follows: d[K+]Aidt=−Jkir , i−2JpumpA , i−Jgap , k , ij ( 15 ) The electrical current through the cell membrane can cause the ionic concentrations inside and outside of the cell to change . An electrical current I across a membrane is equal to the ion flow per unit of time . We convert the electrical current I into the ionic flux J , which is computed from J = I/ ( C*γ ) , where C is the astrocytic membrane capacitance . Here , γ is the scaling factor that relates the ion flux to the membrane potential . where Jgap , k , ij is the potassium ions flow model mediated by astrocytic gap junction: Jgap , k={F/ ( θ ( Δi , jK ) ⋅e ( − ( θ ( Δi , jK-KThre ) /τ ) ) ) Δi , jK>KThre0Δi , jK<KThre ( 16 ) where F is the strength of gap junction between astrocytes . θΔijK is the K+ concentration gradient for two adjacent astrocytes . The threshold function θ ( x ) = x if x > 0 , and 0 otherwise . The astrocytic gap junction mediated K+ buffer allows astrocytic network to respond to changes of the extracellular K+ concentration . The values of the parameters used in the model are listed in Table 1 . In simulating the numerical results of gap junction dysfunction that induces extracellular K+ buffering delay , we assumed that neuron 1 ( N1 ) and neuron 2 ( N2 ) share the same potassium bath concentration ( in other words , neuron 1 ( N1 ) and neuron 2 ( N2 ) have the same value of Kbath in Eq ( 12 ) ) . The 4th-order Runge-Kutta method was used for numerical simulation with a time step of h = 0 . 01 ms . Additionally , parameter values used in the numerical simulation are shown in Table 1 , assuming no special emphasis .
|
Astrocytes are critical regulators of normal physiological activity in the central nervous system , and one of their key functions is removing extracellular K+ . In recent years , numerous biological studies have shown that astrocytic Kir4 . 1 channels and gap junctions between astrocytes act as major K+ clearance mechanisms . Dysfunction of either of these regulatory mechanisms may cause generation of K+-induced seizures . However , it is unclear how and to what extent these two K+-regulating processes lead to spontaneous epileptic activity . These questions were addressed in the present study by constructing novel single astrocyte-neuron network models and a coupled astrocyte-neuron module network connected by an astrocyte gap junction based on existing experimental observations and previous theoretical reports . Simulation results first verified that either down-regulation of astrocytic Kir4 . 1 channels or a decrease of the gap junction strength between astrocytes causes neuropathological hyper-excitability and spontaneous epileptic activity . These results imply that dysfunctional astrocytes should be considered as targets for therapeutic strategies in epilepsy .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"cell",
"physiology",
"medicine",
"and",
"health",
"sciences",
"neural",
"networks",
"nervous",
"system",
"astrocytes",
"membrane",
"potential",
"junctional",
"complexes",
"electrophysiology",
"neuroscience",
"macroglial",
"cells",
"epileptic",
"seizures",
"gap",
"junctions",
"research",
"and",
"analysis",
"methods",
"curve",
"fitting",
"epilepsy",
"computer",
"and",
"information",
"sciences",
"extracellular",
"space",
"mathematical",
"functions",
"animal",
"cells",
"mathematical",
"and",
"statistical",
"techniques",
"glial",
"cells",
"cellular",
"neuroscience",
"cell",
"biology",
"anatomy",
"synapses",
"physiology",
"neurons",
"neurology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"neurophysiology"
] |
2018
|
Astrocytic Kir4.1 channels and gap junctions account for spontaneous epileptic seizure
|
Comorbidity patterns have become a major source of information to explore shared mechanisms of pathogenesis between disorders . In hypothesis-free exploration of comorbid conditions , disease-disease networks are usually identified by pairwise methods . However , interpretation of the results is hindered by several confounders . In particular a very large number of pairwise associations can arise indirectly through other comorbidity associations and they increase exponentially with the increasing breadth of the investigated diseases . To investigate and filter this effect , we computed and compared pairwise approaches with a systems-based method , which constructs a sparse Bayesian direct multimorbidity map ( BDMM ) by systematically eliminating disease-mediated comorbidity relations . Additionally , focusing on depression-related parts of the BDMM , we evaluated correspondence with results from logistic regression , text-mining and molecular-level measures for comorbidities such as genetic overlap and the interactome-based association score . We used a subset of the UK Biobank Resource , a cross-sectional dataset including 247 diseases and 117 , 392 participants who filled out a detailed questionnaire about mental health . The sparse comorbidity map confirmed that depressed patients frequently suffer from both psychiatric and somatic comorbid disorders . Notably , anxiety and obesity show strong and direct relationships with depression . The BDMM identified further directly co-morbid somatic disorders , e . g . irritable bowel syndrome , fibromyalgia , or migraine . Using the subnetwork of depression and metabolic disorders for functional analysis , the interactome-based system-level score showed the best agreement with the sparse disease network . This indicates that these epidemiologically strong disease-disease relations have improved correspondence with expected molecular-level mechanisms . The substantially fewer number of comorbidity relations in the BDMM compared to pairwise methods implies that biologically meaningful comorbid relations may be less frequent than earlier pairwise methods suggested . The computed interactive comprehensive multimorbidity views over the diseasome are available on the web at Co=MorNet: bioinformatics . mit . bme . hu/UKBNetworks .
It has long been recognised that medical disorders frequently co-occur in the same individual [1] but the significance of comorbidity in revealing shared mechanisms of pathogenesis and outcome is a more recent realisation [2–4] . For a given disease or for a focused disease group , the exploration of comorbidities is largely hypothesis driven , together with the cautious selection and management of potential confounders [4–7] . The availablility of large health data sets with full multimorbidity information provides an unprecedented opportunity to understand the overall network of dependencies underpinning complex multimorbidities . These multivariate dependencies in turn become new targets for drug development and other therapies for multimorbid conditions , particularly relevant in aging societies [8–10] . However , the dissection of comorbidity relations is hindered by myriads of confounding factors [11–13] . Following the characterization from Bagley et al . [12] , epidemiological co-occurrences can arise through different routes: 1 ) shared genetic background , 2 ) disease interactions ( a disorder directly causes another ) , 3 ) common environmental cause and 4 ) different biases ( diagnosis artifacts , selection biases ) . Earlier diseasome-wide works focused on the exploration of shared genetic background ( 1 ) behind comorbidities [2 , 3 , 12 , 13] and the underlying molecular networks [3 , 14–18] . These works relied on pairwise comorbid relations partly controlled for potential confounding factors such as age ( for controlling with disease onset see e . g . [2] , for incidence-based control see e . g . [12] ) . However , these approaches do not address the issue of apparent comorbidity mediated by intervening associations with other diseases; a problem of indirect relations that is already attracting attention in other areas of network science [19] . In this paper we demonstrate that probabilistic graphical models ( PGMs ) in the Bayesian statistical framework provide a principled , unified solution for filtering such disease-mediated indirect relations , for correcting for potential external confounders and for coping with limitations and uncertainty of the data . Specifically , we construct sparse multimorbidity maps by applying PGMs for all diseases , i . e . for the whole diseasome . To our knowledge , this method has not been applied for the diseasome so far , despite the unique ability of PGMs to represent maximally sparse models , demonstrated on , for example genomic datasets [19–25] . Our diseasome-wide evaluations show that this approach efficiently scores and discriminates direct and disease-mediated indirect comorbidity relations and has resulted in a loss of more than 80% of comorbidity relations from prevailing pairwise methods . We made a more detailed investigation of BDMMs in the subset of psychiatric and metabolic disorders of the diseaseome . We focused on depression , which is a common psychiatric disorder with a complex neurobiological and psychosocial background [26 , 27] , with approximately 10% prevalence worldwide , and according to forecasts depression will be the largest contributor to the disease burden in the middle- and high-income countries by 2030 according to the World Health Organization [28–30] . Many epidemiologic studies have reported high comorbidity between mental illnesses [5 , 31] , which was partly explained by shared heritability between psychiatric disorders [32 , 33] . Less is known about the complex biopsychosocial mechanisms which underlie associations between somatic and common psychiatric disorders: depression frequently co-occurs with a wide range of somatic disorders , for example with migraine [34] , with other disorders causing chronic pain [35 , 36] , and with cardio-metabolic syndromes [37] . It has been also demonstrated that patients with depression have increased number of diagnosed disorders compared to non-depressed patients [5 , 38] and depression worsens the treatment outcome of the comorbid conditions [39] and is an independent predictor of increased mortality rate [40 , 41] . Therefore besides the exploration of further comorbidities of depression , it is equally important to discriminate its comorbidities as direct and indirect comorbidities . Such discrimination could reveal more specific pathophysiological subgroups of this heterogeneous condition and thus transform the power of genetic and epidemiological studies to advance precision medicine in psychiatry and metabolic disorders . We also evaluated the correspondence of BDMMs with molecular-level measures and relations , such as genetic overlap and the interactome-based association score for a depression-related subset of diseases . Focusing on depression we identified a direct multimorbid neighbourhood and confirmed that direct comorbidities correspond to direct relationships in the molecular interactome .
To explore the direct and indirect status of comorbidities in the diseasome and to investigate the effect of filtering disease-mediated , indirect comorbidity relations by BDMMs , we computed multiple pairwise measures . The prevailing non-systems based approaches usually use a pairwise measure such as odds ratio ( OR ) , Pearson’s correlation coefficient ( Φ ) or logistic regression to determine the epidemiologic relationship of disorders [2 , 3] . We computed these most often used statistical descriptors of comorbidity for all the investigated diseases ( see S2 Dataset for all computed pairwise measures for all possible disease pairs ) . We also computed BDMMs and cross-compared with the prevailing pairwise approaches [3 , 11 , 12 , 18] . The Fig 3 illustrates the sparsity of the direct ( systems-based ) map compared to a non-systems based network over the diseasome . The BDMM approach resulted in 320 direct connections ( BDMM edge posterior > 0 . 95 ) whereas applying the χ2 independence test 1714 disorder-disorder relations have significant p-values with a threshold of 0 . 05 after Bonferroni correction . We also investigated the sufficiency of the sample size on different approaches . Fig 4 shows characteristics of the most significant results in each approaches . We transformed the different scores to the [0 , 1] interval to make them comparable ( see Materials and methods ) . Descriptors and statistics are available on the web ( Co=MorNet: bioinformatics . mit . bme . hu/UKBNetworks ) , also as an interactive tool to visualise networks of direct and mediated connections of selected diseases . The examination of the filtering capacity of the BDMM approach confirms that the BDMM edge posteriors strongly differentiate the direct connections from the mediated ones , e . g . there are only a few disease pairs which have a BDMM edge posterior between 0 . 05 and 0 . 95 ( Fig 5 ) . For the definition of BDMM edge and BDMM ( structural ) association relations see Table S1 in S1 Appendix . On Fig 5 , we mark the direct comorbid connections with high BDMM edge posteriors ( > 0 . 95 ) in red , showing that all of them have significant Bonferroni-corrected χ2 p-values . Additionally , focusing on a subset of relations with BDMM edge posterior less than 0 . 95 , we examined the BDMM association posteriors versus the parametric association ( see Fig 6 ) . It confirms that almost all such strongly significant parametric association has high BDMM association posterior . Note , that relations with high BDMM edge posterior have high BDMM association posteriors as well . Furthermore , Fig 6 also illustrates that BDMM association posterior indicates many more structurally distant -presumably parametrically weaker- associations which cannot be inferred by parametric association tests such as Bonferroni-corrected χ2 p-values . To evaluate the genetic relevance of direct versus indirect comorbidity relations , we extended the epidemiological-level analysis with molecular-level approaches using a genetic overlap [3] and the interactome-based separation score [16] . This analysis included depression , metabolic syndromes and hypertension , see Fig 7 , S2 Fig and Table S3 in S1 Appendix . The genetic overlap measures for comorbidities were computed based on manually curated databases , using the DisGeNet [44] and NCBI PheGenI [45] ( for details , see S1 Appendix , for a related earlier work , see [2] ) . The interactome-based connectivity/separation scores for comorbidities were computed by the supplied method and data from Menche et al . [16] ( see S1 Appendix ) . Beside of the investigation of direct and indirect status of comorbidities and the cross-comparison of pairwise and BDMM measures , we performed a more detailed medical evaluation of BDMMs on psychiatric and metabolic disorders , especially on depression and its comorbidities .
First of all , results indicate that the UK Biobank dataset is sufficiently large for the construction of BDMMs for this variable set , as the BDMM edge posteriors are peaked at 0 and 1 ( see upper part in Fig 5 ) , indeed , with the thresholds 0 . 05 and 0 . 95 we can efficiently separate the statistically significant comorbid relations to direct ( co=morbid ) and indirect relations . Notably , BDMMs eliminated more than 80% of comorbidity relations as indirect ones ( 320 direct connections from 1714 candidate relations ) . Interestingly , a recent work using a priori disease categories for restricting comorbidity relations , controlling for confounding with incidence characteristics and using two independent data sets similarly reported nearly 90% elimination ratio [12] . Note that all co=morbid connections were also confirmed by standard statistical methods ( the right-lower quadrant is empty in Fig 5 ) , which implies the technical condition that the distribution of the BDMM posteriors is stable [48–50] . In Fig 6 we further evaluated the connections with significant Bonferroni-corrected χ2 p-values but below the 0 . 95 edge posterior threshold . This shows that most of these connections have high BDMM structural association posterior , which suggest that BDMM indeed filtered mediated and confounded relations . There are 31 connections which have a significant pairwise association score but no structural association ( BDDM structural association posterior < 0 . 1 ) . These disorder pairs rarely occur together in patients ( mean co-occurrence:8 . 45 , with standard deviation: 4 . 9 , and quantiles: 5 , 7 , 10 . 5 for the 25% , 50% and 75% respectively ) . In case of the χ2 test we applied Yates’ continuity correction but for such weak connections even with the large UK Biobank dataset the BDMM approach was not able to catch that weak dependency structure . Our results demonstrated that the interactome-based score provided similar maps as BDMM co=morbid diseases , in sharp contrast to the associative genetic overlap scores which followed the pattern of the pairwise disease relative risk ( see Fig 7 , S2 Fig and Table S3 in S1 Appendix ) . Note that the interactome-based score and co=morbidity are analogous as both use a systems-based approach on different levels ( molecular and epidemiologic level respectively ) . For detailed description of the molecular level methods and results see S1 Appendix and [2 , 16] . The high comorbidity between mood disorders and anxiety or stress related disorders is well known , and twin studies suggested that these comorbidities originated mainly from shared genetic risk factors [52 , 53] . Our results showed another expected aspect , namely that this relationship is independent of the order of the onset of these disorders . This observation is in line with a longitudinal study which showed that generalised anxiety disorder ( GAD ) and major depressive disorder ( MDD ) are strongly comorbid with an equal probability of GAD or MDD occurring first or simultaneously suggesting they might not be distinct disorders [54] . Although overlapping genetic risk factors for anxiety and depression have not yet been identified , common genetic vulnerability has been found for other comorbid psychiatric disorders [32 , 33 , 55] . Our BDMM further indicate strong and stable co=morbidity between depression , anxiety , stress , postnatal depression and nervous breakdown , pointing toward interactome-level overlaps; this reinforces the need to find potential common biological mechanisms [56] . Epidemiologic studies repeatedly report high comorbidity between depression and metabolic disorders [57] , depression and diabetes [58] , depression and cardiovascular disorders [37] , depression and hypertension [59 , 60] , and depression and obesity [61 , 62] . However , there have been several contradictory results , and this suggests a more complex relationship . Indeed , recent GWAS results found no shared genetic risk between these disorders and depression [55] . Nevertheless , based on the UK Biobank cohort data , obesity was co=morbid with depression while cardiovascular disorders , hypertension , high cholesterol and diabetes , including type 2 diabetes were only indirectly related to depression . When we excluded occurrences after the onset of depression the direct relationship between obesity and depression remained as expected but entirely new links with high posterior probability emerged suggesting a strong relationship between the consequences of metabolic syndromes and depression . Studies of the genetic relationship between obesity and depression suggest that atypical depression , characterised by increase in appetite and weight , is associated with genetic risk factors and polygenic risk scores of increased BMI and triglycerides , while typical depression , with decreased appetite and weight , show more similarities with other psychiatric disorders [63 , 64] . Thus , in line with our results comorbid obesity and metabolic disorders may identify a specific subtype of depression with a distinct biological background . Caution is required in inferring shared biology for co=morbid disorders given our current lack of knowledge about relevant GxE interactions . Although metabolic disorders are only co=morbid with depression when they precede it , a variety of non-biological associations of mediators may be at play . We cannot currently exclude the possibility that lifestyle factors , such as diet , physical activity and stress , or medication used to treat hypertension , hypercholesterolaemia and obesity may contribute to the later development of depression [65] . As a specific example , a previous study demonstrated that current psychological distress amplified the effect of genetic risk of high BMI [66] . Patients with increased genetic risk to become overweight showed worse physical outcome ( higher BMI ) , and quite probably more comorbid psychological symptoms , when life stress was present . Furthermore , it has been reported that statins , drugs with cholesterol-lowering effect , have antidepressant effect in patients with comorbid depression and coronary artery disease while the same drugs can have pro-depressive effect or no effect on depression when comorbidities and depression subtypes were not taken into account [67] . Migraine [34] , IBS [68] , FM [69] , and CFS [70] are highly comorbid with depression based on epidemiologic studies . It is therefore puzzling that they involve different etiological mechanisms . In addition , their symptoms often overlap making it difficult to apply diagnostic categories . We found that these disorders were not relevant when they occurred before depression but were highly co=morbid in the full analysis . The probable explanation is that in general , these disorders are related to consequences of depression and only specific subtypes of these disorders can be expected to have causal relations , e . g . shared biological background with depression . For example , a genetic risk score analysis demonstrated that migraine with comorbid depression was more genetically related to depression than to pure migraine , which suggests that migraine might develop as a consequence of different polygenic backgrounds [71] . Similarly , a large general population cohort study confirmed that FM , CFS and IBS increase the odds of depression and anxiety but that most patients who suffer from FM , CFS and IBS have no mood or anxiety disorder [72] . One of the main limitations is that all disorders were self-reported , although trained nurses evaluated and corrected all entries during face-to-face interviews . The second one is that the applied treatments or medications were not included in the analysis which could highlight comorbidities due to the side effect of treatments . We will address this problem in follow-up studies . Note that we only used a subset of the UK Biobank dataset selecting those participants who filled out the Mental Health Questionnaire and provided online dietary information , which may introduce confounding through selection bias . However , limiting our study to this subpopulation enabled us to test different definitions of depression and will allow us to connect this comorbidity network to relevant environmental risk factors . The use of large-scale health data sets , such as the UK Biobank dataset hold the promise of complementing and guiding the molecular level research of complex diseases . Adopting an intermediary approach between statistical association analysis and causal discovery we investigated the use of Bayesian networks in the Bayesian model averaging framework to explore direct probabilistic relations with respect to a given set of variables , i . e . to eliminate confounders and mediatory effects by a systems-based approach . We demonstrated the applicability of BDMMs , especially their principled capability of discriminating direct and indirect comorbidities . In summary , PGMs offer maximally sparse dependency models and utilize the omic nature of the epidemiologic data jointly modelling all the morbidities; while the Bayesian approach through posteriors provides an explicit representation for the uncertainties in a dataset . Thus the Bayesian direct morbidity maps provide sparse , systems-based , omic-wide perspectives . From a clinical perspective , our results also highlight that the direct and indirect subtypes of comorbidities support a finer biological interpretation , namely an interactome-based detailed interpretation using molecular mechanisms corresponding to direct relations , whereas genetic overlap using associative gene sets may only reflect indirect comorbidities . In addition , re-running the analysis by including only instances of disorders which preceded depression , we delineated comorbid disorders of depression with more refined causal roles that could specify subgroups of depressed patients with more homogenous background . The investigation of Bayesian direct morbidity maps also demonstrated , that even large-scale datasets such as UK Biobank , are still limited for non-ambiguous identification of complex dependency patterns such as multimorbidities [73] . However , the applied Bayesian statistical framework offers an automated , normative solution for the multiple hypothesis testing problem and the application of probabilistic graphical models in the Bayesian framework supports the versatile post-processing of the results and their efficient communication and sharing . The results of our research highlight the advantages of Bayesian systems-based modelling , which could be vital to cope with the growing heterogeneity of new health data sets containing full personal genetic information with high dimensional data about lifestyle , environmental factors and sequential decisions on drug therapies [8 , 74] .
In the present study we used the UK Biobank ( http://www . ukbiobank . ac . uk/ and [75] cohort where subjects’ chronic illness history together with onset age were ascertained by trained nurses during face-to-face interviews and were processed by experts resulting in 525 different disease categories . The investigated subset in this study consisted of 117 , 392 participants ( female: 64 , 320; male: 53 , 072 ) who provided the extended Mental Health Questionnaire ( http://biobank . ctsu . ox . ac . uk/crystal/label . cgi ? id=100060 ) and the online diet questionnaire data together with the extensive baseline dataset . We used the UK Biobank original disease categories with at least 1‰ prevalence in the selected subset , which resulted in n = 247 diseases including depression ( n = 6040 ) . In addition , we coded obesity in cases where BMI were equal or greater than 30kg/m2 , for further analyses . For statistical analysis , sex was included into the data set , and age was binned into 3 equal frequency categories with thresholds 60 and 68 years . Then we applied the different pairwise measures and logistic regression together with Bayesian systems-based modelling to compare the models computed on these datasets ( see below and in S1 Appendix ) . To investigate the effect of disease onset , self-reported disease onset data was used to filter the dataset . 6 , 040 patients affected by depression , provided onset data whose comorbid illnesses were eliminated if they occurred after the onset of depression . After removal of diseases with prevalence less than 1‰ the dataset contained 241 diseases . We extended the dataset with sex , age and BMI-based obesity . The data were analysed using same statistical methods as with the non-filtered dataset . To test the stability of comorbid relationships with depression we also used an alternative depression definition instead of self-reported depressive disorder . Depression and its severity was defined by the Mental Health Questionnaire data [46] , for definition see S1 Appendix . These alternative depression categories were analysed with Bayesian systems-based modelling . We applied text-mining and conventional statistical methods to explore comorbid relations , see S1 Appendix . For these computations we used in-house written R scripts together with the statistical programs included in the stats package of R [76] . To overcome the limitations of these conventional methods , we applied a Bayesian network Markov Chain Monte Carlo ( BN-MCMC ) method to explore the overall system of dependencies-independencies , visualized as an undirected graph with weighted edges [20 , 21 , 23–25 , 42] . The weighted edges correspond the a posteriori probabilities ( Pr ) of direct , nonmediated “co=morbidity” relations , the weights are in the [0 , 1] interval and the higher values show stronger relationship . The systems-based approach using Bayesian networks prunes the indirect , mediated connections between morbidities , thus resulting in a sparse co=morbidity map compared to pairwise association networks ( for detailed description of the method and for further types of dependency relations , see S1 Appendix and Table S1 in S1 Appendix ) .
|
Depression is one of the most common of psychiatric disorders and its causation is correspondingly multifactorial , complex and heterogeneous . It occurs in combination with a number of physical illnesses far more commonly than expected by chance . Such comorbidities may be important clues pointing to shared environmental and genetic risk factors and could identify different causal types of depression . However , a method is still needed to weed out statistically significant pairings that nevertheless arise through indirect routes involving comorbidities between other diseases . We examined the pairwise associations among 247 diseases of 117 , 392 participants recorded in the UK Biobank database . We found that the great majority of disease associations were indirect consequences of a sparse network of ‘direct’ comorbidities ( ‘sparse diseaseome’ ) constructed using probabilistic graphical models ( PGMs ) within the Bayesian statistical framework . In a depression-related subset of illnesses , we found that several pairwise associations of depression were indirect and due to their comorbidities with obesity which had a strong direct connection with depression . Furthermore , the direct comorbidities in a depression-related subset of disorders , but not the pairwise associations , strongly mapped onto an underlying molecular network ( ‘interactome’ ) suggesting that this approach significantly improved correspondence with molecular reality .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"body",
"weight",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"migraine",
"headaches",
"physiological",
"parameters",
"signs",
"and",
"symptoms",
"obesity",
"mood",
"disorders",
"genetic",
"epidemiology",
"epidemiology",
"hypertension",
"blood",
"pressure",
"mental",
"health",
"and",
"psychiatry",
"metabolic",
"disorders",
"clinical",
"genetics",
"diagnostic",
"medicine",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"depression",
"vascular",
"medicine",
"genetics",
"of",
"disease"
] |
2017
|
Comorbidities in the diseasome are more apparent than real: What Bayesian filtering reveals about the comorbidities of depression
|
Motility and phagocytosis are key processes that are involved in invasive amoebiasis disease caused by intestinal parasite Entamoeba histolytica . Previous studies have reported unconventional myosins to play significant role in membrane based motility as well as endocytic processes . EhMyosin IB is the only unconventional myosin present in E . histolytica , is thought to be involved in both of these processes . Here , we report an interaction between the SH3 domain of EhMyosin IB and c-terminal domain of EhFP10 , a Rho guanine nucleotide exchange factor . EhFP10 was found to be confined to Entamoeba species only , and to contain a c-terminal domain that binds and bundles actin filaments . EhFP10 was observed to localize in the membrane ruffles , phagocytic and macropinocytic cups of E . histolytica trophozoites . It was also found in early pinosomes but not early phagosomes . A crystal structure of the c-terminal SH3 domain of EhMyosin IB ( EhMySH3 ) in complex with an EhFP10 peptide and co-localization studies established the interaction of EhMySH3 with EhFP10 . This interaction was shown to lead to inhibition of actin bundling activity and to thereby regulate actin dynamics during endocytosis . We hypothesize that unique domain architecture of EhFP10 might be compensating the absence of Wasp and related proteins in Entamoeba , which are known partners of myosin SH3 domains in other eukaryotes . Our findings also highlights the role of actin bundling during endocytosis .
Entamoeba histolytica is the causative agent of amoebiasis disease in humans , a major public health problem in developing countries . Amoebiasis is the third-leading cause of deaths resulting from parasitic infections [1 , 2] . The ability of E . histolytica to phagocytose cells of the intestinal epithelia and the immune system is the major contributor to its pathogenesis [3 , 4] . Phagocytosis is associated with intensive cytoskeletal remodeling , which involves actin filaments , several actin-binding proteins , and myosins . Unconventional myosin I constitutes the largest class of the myosin superfamily [5] and has been associated with a variety of membrane-based motility processes such as pinocytosis , membrane ruffling , lamellipodia and filopodia as well as phagocytosis [6 , 7] , but the detailed mechanisms and pathways of the associations of myosin I with these processes remain to be studied . Even in simple eukaryotes like yeast and amoebazoa , myosin I exists in several isoforms , which carry out distinct functions [8–10] but the genome of E . histolytica has a single isoform of myosin I , namely EhMyosin IB [11] . EhMyosin IB has been observed to localize in phagocytic cups during erythrophagocytic processes [12] . Entamoeba cells overexpressing EhMyosin IB were found to display decreased phagocytosis due to a defect in the phagosome initiation process [13] . The SH3 domain of S . cerevisae myosin I ( Myo3 and Myo5 ) has been shown to interact with WASP-related proteins Vrp1 , Bee1p and Las17p and help in the recruitment of myosin I and Arp2/3 proteins at cortical actin nucleation sites [14] . In other amoebae such as Dictyostelium and Acanthamoeba , the SH3 domain of myosin I interacts with adapter protein like CARMIL and Acan125 respectively , that help to recruit and activate Arp2/3 [15] . Formation of multiprotein complexes via these interactions is essential for rearrangement of the actin cytoskeleton during endocytosis . WASP and other WASP-related proteins function downstream of several GTPases in FcϒR-mediated phagocytosis , leading to actin polymerization at the site of cup formation [16] . So far , only calcium binding proteins ( CaBPs ) , EhCaBP3 and EhCaBP5 have been shown to interact with EhMyosin IB , and they do not interact with its SH3 domain [12] . Crystal structure of the EhMySH3 domain provided details about the ligand preferences for interaction , [17] but no interacting partner for this domain has yet been identified . Phagocytosis has been found to be a key determinant of the virulence of E . histolytica facilitating colonization of gut and other vital organs during invasive amoebiasis . In a wide variety of eukaryotes , membrane phospholipids play a significant role in membrane trafficking , cytoskeletal rearrangements and cell surface receptor signaling [18] . Formation of the phospholipid PI3P , on the phagosome surface by PI3 kinases results in the recruitment of several PI3P-binding proteins including EEA1 and Hrs [19] . Inhibition of PI3P blocks phagosome maturation . PI3P is involved in endocytosis and phagocytosis in E . histolytica cells as well [20 , 21] . FYVE domains are considered biosensors for PI3P sites on the membranes [22] . The FYVE zinc finger domain has been hypothesized to associate only with phagosomes and not with fluid-filled pinosomes [22] . Consistent with this hypothesis , EhFP4 which is a FYVE domain containing Rho guanine nucleotide exchange factor ( Rho-GEF ) in E . histolytica , was found to be involved in phagocytosis . Several Dbl homology GEF proteins like EhGEF1 , EhGEF2 and EhGEF3 present in E . histolytica have been studied in detail and found to play an important role in phagocytosis , cell motility and migration , capping and chemoattractive response [23–27] while EhFP4 is the only FYVE family GEF studied so far . The genome of E . histolytica harbors 12 FYVE family Rho GTPase exchange factors ( EhFPs ) , of which EhFP 5 , 6 , 7 and 10 are highly expressed [21] . EhFP10 is the longest of the 12 FYVE family GEFs ( EhFPs ) and its gene expression has been detected in cultured clinical isolates and trophozoites , as well as in cysts [28 , 29] . In a transcriptome analysis , EhFP10 showed 2 . 7-fold upregulated expression in virulent E . histolytica compared to that in non-virulent Rahman strain of E . histolytica [29] . The presence of the EhFP10 protein in E . histolytica lysate was confirmed in a proteomic study by Marion et al . [30] . In the current work , we have identified a novel interaction between the SH3 domain of unconventional EhMyosin IB and the c-terminal domain of EhFP10 , and showed this interaction to regulate the phagocytic and pinocytic processes by affecting the cytoskeletal dynamics . The c-terminal domain of EhFP10 was found to regulate actin dynamics as reported for the APC basic domain and tau protein . Our results also suggested the participation of EhFP10 in pinocytosis along with phagocytosis in E . histolytica . So far , EhFP10 is the first FYVE domain containing GEF which has been found to be involved in pinocytosis . Moreover , the EhFP10 localization pattern makes it as a marker for distinguishing between pinosomes and phagosomes within E . histolytica cells . EhFP10 has a unique domain organization and no homologs across all genomes available in BLAST databases; i . e . , it was found only in Entamoeba species . Hence , this unique interaction could be of great advantage to the pathogen and may be responsible for its high motility and rate of phagocytosis .
The SH3 domain recognizes the PXXP motif in proteins . These motifs have been further classified into class I [ ( K/R ) xxPxxP] and class II [xPxxPx ( K/R ) ] types [31 , 32] . To identify proteins that interact with the c-terminal SH3 domain of EhMyosin IB , a proteomics data by Marion et al . was utilized [30] . The data was obtained from cells overexpressing myosin IB and undergoing active erythrophagocytosis [33] . The screening was done using “SH3 HUNTER” software [50] as described in Material and Methods . A FYVE family Rho-GEF , EhFP10 ( accession no . EAL46050 . 1 ) was predicted to harbor the maximum number of PxxP motifs ( Table 1 ) . Since no unconventional myosin I homologs have been reported to be involved directly or indirectly , in the regulation of Rho mediated cytoskeletal regulation , we proceeded further with EhFP10 in our current study . In a previous study , EhMySH3 was predicted to preferentially bind pseudo-symmetrical and class II ligands [17] . Hence , to confirm the type of ligand interactions preferred by the EhMySH3 domain , we selected a few peptide sequences from EhFP10 . Chosen ligands belonged to class I ( P4 ) , class II ( P1 ) , and an overlapping pseudo-symmetrical sequence belonging to both classes ( P2 ) , as predicted by “SH3 HUNTER’ software . Along with these , two PXXP peptides; P3 and P5 from a Rho GTPase , were also chosen ( Table 2 ) . All five peptides were chemically synthesized ( Biochain Inc . ) without any modifications . These peptides were tested for their ability to bind EhMySH3 by using Autolab Surface Plasmon Resonance ( SPR ) . The results showed that the P2 peptide , with a pseudosymmetrical ligand ( overlapping class I and class II motif ) from EhFP10 displayed highest binding response relative to other peptides , as was predicted in our previous study [17] . The weakest binding was observed with the P4 peptide ( class I ) ( Fig 1 ) . The binding affinity values of the peptides with the EhMySH3 domain could not be calculated accurately since the molecular mass of all of the peptides were about 1000 Da , close to the minimum detectable range for the Autolab instrument . Hence , on the basis of relative binding response , we proceeded with the high-response peptides , i . e . , P2 , P3 , P1 , and P5 , for co-crystallization with EhMySH3 . We were able to crystallize EhMySH3 in complex with the pseudo-symmetrical peptide P2 [KVAPPIPHR] . Co-crystallization attempts with other peptides resulted in the formation of the PEG-EhMySH3 complex instead of peptide-EhMySH3 . The EhMYSH3-P2 complex crystallized with two molecules of EhMySH3 and one P2 peptide per asymmetric unit ( Fig 2A ) , and with P2 binding to chain A of one EhMySH3 molecule and chain B of a symmetry-related EhMySH3 molecule ( Fig 2C ) . P2 was observed to make more electrostatic interactions with chain B than with chain A . P2 binds to EhMySH3 chain B in class II pattern , i . e . , [xPxxPx ( K/R ) ] . P2 residues A3 and P4 formed the first XP motif , which fit in the first consensus hydrophobic pocket made by residues Y9 and Y52 of EhMySH3; and I5 and P6 formed the second XP motif , which occupied the second hydrophobic pocket made by residues P49 and W36 of EhMySH3 . The c-terminal residue R9 of P2 was observed to interact with residue E34 of the n-Src loop present in the specificity pocket of EhMySH3 . In other available ligand-bound SH3 structures , terminal Arg or Lys residues interact with a residue of the RT loop , which forms the specificity pocket . In our case , though the electron density for the terminal R9 residue was not complete , after refinement we obtained better electron density near the n-Src loop , though there could be an alternate conformation near E18 in the RT loop , likely to be one of the probable orientations for residue R9 . Overall , we concluded that the EhFP10 peptide , P2 binds to PxxP motif of EhMySH3 in a class II peptide binding orientation and , due to its pseudo-symmetrical sequence , did so in a way resulting in the clustering of the two EhMySH3 domains ( Fig 2D ) . EhFP10 is a multi-domain protein , 876 amino acids long , sub-divided into n-terminal Dbl homology GEF domain ( 1–438 ) , followed by a FYVE domain ( 439–511 ) and a c-terminal domain ( 512–876 ) ( Fig 3A ) . Based on sequence analysis and conserved domain database analysis , the c-terminal domain of EhFP10 ( cterEhFP10 ) was predicted to be similar to the APC basic domain present in proteins of the adenomatous polyposis coli family and the microtubule-binding protein tau . Both APC basic and tau domains have been shown to bind microtubules as well as actin filaments [34] . They could also crosslink actin filaments and microtubules , in turn , mediate cytoskeleton crosstalk [35] . These observations and predictions suggest that the c-terminal domain of EhFP10 could associate with microtubules and actin filaments in E . histolytica . On analysis of the domain organization of all 12 FYVE family GEFs present in E . histolytica , EhFP10 was found to be the only GEF to have this unique c-terminal domain . The sequences of all Dbl-homology GEFs had highly conserved RhoGEF domain but different c-terminal domain . In a BLAST search across all databases , we could not find any GEF homologous to EhFP10 . EhFP10 gene is present in all Entamoeba species . Homologues of EhFP10 of non- pathogenic strain E . dispar and E . invadens ( pathogenic in reptiles ) , showed 95% and 35% identity with that of E . histolytica molecule respectively . Maximum differences were observed in the c-terminal domain ( S3 Fig ) . Peptide , P2 is located between 653–661 residues in the c-terminal domain of the EhFP10 molecule ( Fig 3A ) . To further confirm the EhMySH3-EhFP10 interaction , an in vitro GST pull-down assay with recombinant cterEhFP10 protein and EhMySH3 was performed . Here , cterEhFP10 protein got eluted along with GST-tagged EhMySH3 protein , and no cterEhFP10 protein was present with GST protein alone ( Fig 3B ) . SPR studies were also carried out where various concentrations of cterEhFP10 protein samples were passed over immobilized EhMySH3; these studies indicated a dissociation constant ( KD ) of 200 nm for EhMySH3-cterEhFP10 ( Fig 3C ) . EhMyosinIB has been shown to be enriched in the extending pseudopods of motile amoeba and to also participate in the closure of the phagocytic cup by associating with phagocytic machinery [13] . Since EhFP10 was found to be one of the interacting partners of EhMyosinIB , we proceeded to study the cellular distribution of EhFP10 in motile E . histolytica trophozoites . Localisation studies were carried out using cells expressing EhFP10 tagged at its N terminus with GFP ( GFP-EhFP10 ) . In immunostained cells , GFP-EhFP10 was found mainly in the cytoplasm in proliferating cells . It was also seen in some membrane invaginations and cup-like projections ( S4A Fig ) . The localisation showed a good correlation of GFP-EhFP10 with the untagged EhFP10 detected using EhFP10-specific antibody ( S4B Fig ) . To further look into the cellular distribution of EhFP10 in the dynamic cellular environment , time-lapse imaging was done using live proliferating amoebic trophozoites . GFP-EhFP10 was found to become enriched with time underneath the membrane at specific sites leading to ruffles and membrane protrusions . The level of enrichment initially increased with time at destined specific site near the membrane . Later the enrichment expanded along the membrane , and finally divided into two enriched sites leading to the simultaneous formation of two macropinocytic cups . The two cups then completed the macropinocytosis independently , usually one after the other . This pattern was recurrent and appeared to be the common mechanism of formation of the pinocytic vesicles . The GFP-EhFP10 protein was found to be associated with the cup from initiation until scission and also in internalized macropinosomes ( Fig 4A and S1 Video ) . To confirm the endocytic process to be pinocytosis , TRITC-dextran ( Sigma ) was added to the media , and endowed the media with red fluorescence . The presence of red fluorescence within internalized vesicles enriched with GFP-EhFP10 confirmed EhFP10 to be involved in fluid-phase macropinocytosis in E . histolytica . EhFP10 was present from initiation until scission of the macropinocytic vesicles as well as around internalized macropinosomes . A single macropinocytic event typically took 4–5 seconds to complete . ( Fig 5A and S2 Video ) . Since , phagocytosis plays a crucial role in amoebiasis disease pathogenesis , we were interested to determine whether EhFP10 also plays a role in phagocytosis in addition to macropinocytosis . GFP-EhFP10-overexpressing cells were utilized for this purpose . RBC membranes were stained with lipophilic , red fluorescent dye ( Thermo Fisher ) . EhFP10 was seen to be enriched at the site of pseudopod formation near the cell membrane as soon as RBC came in contact with E . histolytica cells , and it later became enriched in the extending pseudopods during the cup progression and was present until phagosome scission ( Fig 4B ) . The distribution of GFP-EhFP10 in the phagocytic process , however , was different from that for pinocytosis after the scission of the early endosome . Unlike macropinosomes , the phagosome membrane was not enriched with EhFP10 . An erythrophagocytic process took about 7–8 seconds to complete ( Fig 4B and S3 Video ) . To sum up , EhFP10 was found enriched in membrane ruffles and around the sites of both macropinocytosis as well as phagocytosis . It was enriched in the pseudopods during cup progression and was present until scission of the vesicle . A single macropinocytic event transpired in a shorter amount of time than did a single phagocytic event . This is the first report of the involvement of a FYVE family RhoGEF of E . histolytica in macropinocytosis . Previously FYVE family GEFs were hypothesized to be associated only with phagocytosis [22] . The difference in the localization patterns of EhFP10 in phagocytosis and pinocytosis; and the presence of EhFP10 in the membranes of early macropinosomes but not in those of phagosomes , established this protein as a potential marker to distinguish pinosomes from phagosomes in Entamoeba species . The SH3 domain of EhMyosin IB was shown , as described above , to interact with the c-terminal domain of EhFP10 . This interaction was further validated by measuring co-localization in images of immunostained actively proliferating amoebic cells overexpressing GFP-EhFP10 . EhMyosin IB was found to colocalize with EhFP10 at the pinocytic cups as well as at the macropinosomes ( Fig 6B and 6C ) . In cells undergoing phagocytosis , a similar association was observed between EhFP10 and EhMyosin IB ( Fig 6A and 6D ) . Since phagocytosis and macropinocytosis are actin-dependent processes , and since homologs of the c-terminal domain of EhFP10 are involved in actin assembly , we investigated the association of EhFP10 with actin within E . histolytica cells . Interestingly , in almost every cell where F-actin ( stained with TRITC phalloidin ) was observed , there was an enrichment of EhFP10 . The corresponding Pearson’s correlation coefficient was calculated to be between 0 . 5 and 0 . 6 , suggesting a likely association of EhFP10 with actin dynamics . However , in cells undergoing erythrophagocytosis , EhFP10 and EhMyosin IB colocalized at the proximal end ( tip ) of progressing pseudopods while actin was present both at the tip and the base of the cup . As the cups progressed towards closure , all three proteins were found to colocalize at the tip of the closing pinocytic and phagocytic vesicles ( Fig 6E ) . As the pseudopods fused , EhFP10 and EhMyosin IB moved more towards the tip of the closing pseudopods in the case of phagocytosis; while during pinocytosis , EhFP10 also resided at the membranes of closing pinosomes . In previous studies , myosin IB was reported to help in the pinching of the vesicle . On the basis of our localization studies , we suggest that EhFP10 also plays a role in this process . Neither EhFP10 nor EhMyosin IB were present in the internalized phagocytic vesicles ( Fig 6A ) . Images of immunostained fixed amoebic cells also confirmed the presence of EhFP10 throughout the pinocytic process including in macropinosomes , in contrast to the case for phagocytosis where EhFP10 was no longer observed during the separation of phagosomes from membranes . However , in early pinosomes , EhFP10 was found to colocalize more with actin than with myosin IB ( Fig 6C and 6D ) . These observations clearly pointed out the difference in molecular mechanisms governing macropinocytosis and phagocytosis and the involvement of EhFP10 and EhMyosin IB in both of the processes . APC basic and tau domains have been shown to be involved in actin dynamics [34] . Since EhFP10 has a c-terminal domain that shows similarity with APC basic and tau domains , a direct involvement of EhFP10 in actin dynamics was investigated . Purified recombinant cterEhFP10 domain was observed to bind and bundle purified rabbit muscle F-actin in centrifugation-based actin binding ( Fig 7A ) and bundling assays ( Fig 7C and 7D ) . In actin co-sedimentation assay at high-speed centrifugation of these proteins , a fraction of the cterEhFP10 was seen in the pellet along with F-actin , confirming its interaction with F-actin , and only cterEhFP10 ( i . e . , no actin ) was observed in the supernatant ( Fig 7A ) . The bundling assay was carried out using low-speed centrifugation . Here , in the absence of cterEhFP10 , F-actin stayed in the supernatant fraction , while in the presence of cterEhFP10 , almost all of the actin went into the pellet fraction ( Fig 7C ) . However , in both cases , the amount of cterEhFP10 that bound the actin remained almost constant for a fixed actin concentration . Bundles of F-actin were also visualized using TEM . As the concentration of the cterEhFP10 protein was increased , thicker actin bundles were observed ( Fig 7D ) ; specifically , actin filaments , which were around 4 nm to 7 nm in thickness , grew to as thick as 50 nm to 120 nm in diameter in the presence of 10 μM to 20 μM cterEhFP10 ( Fig 7D ) . The KD of cterEhFP10 for actin was calculated using SPR to be about 87 nM ( Fig 7B ) , indicating a stronger affinity of EhFP10 for actin than for EhMySH3 . Our data clearly showed EhFP10 to interact through its c-terminal domain with both F-actin and EhMyosin IB . It is likely that the nature of the interaction between these three proteins could either be competitive or cooperative . To test our hypothesis , actin binding and bundling assays were performed using a constant amount of EhMySH3 and various amounts of EhFP10 . EhMySH3 by itself did not show any binding to rabbit muscle filamentous actin in a centrifugation-based actin co-sedimentation assay . EhMySH3 was seen in the supernatant fractions while the filamentous actin was present in the pellet fraction ( Fig 8B ) . As expected , association of EhFP10 with EhMySH3 , led to some amount of the EhMySH3 to come along with the pellet fraction ( Fig 8A ) . The actin bundling assays were performed in the presence and absence of EhMySH3 at a constant cterEhFP10 concentration . Here , a greater percentage of actin was observed to be present in the supernatant fraction in the presence of EhMySH3 than in its absence . This result suggested that the interaction of EhMyosin IB with EhFP10 to be competitive in nature , and the presence of myosin IB to inhibit the actin-bundling activity by EhFP10 . As the concentration of EhFP10 was increased , it was able to overcome the inhibitory effect of EhMySH3 , as the majority of the actin filaments were seen in the pellet fraction ( Fig 8C and 8D ) .
Processes such as phagocytosis , macropinocytosis and cell movement are highly dependent on actin filaments for extension of membrane protrusions . Several myosins such as myosin VI , myosin V and myosin I have been found to be associated with the endocytic processes , actin assembly and maintenance of cortical tension [36–38] . Unconventional myosin I plays a significant role during endocytic processes—directly by interacting with actin , and indirectly by recruiting factors associated with nucleation , polymerization , and stabilization of actin filaments [39] . The SH3 domain located in the tail region of myosin I plays a prominent role in bridging the processes of actin assembly and endocytosis . The SH3 domain of myosin I isoforms in yeast interact with proline-rich regions of Vrp1 and recruits the Arp2/3 complex [40] . In amoebae , Dictyostelium myosin I recruits CARMIL via its SH3 domain and Acanthamoeba myosin I recruits Acan125 . Both these proteins , regulate Arp2/3 mediated actin dynamics [41 , 42] . Thus , in the present work , we set out to identify interacting partner ( s ) of the c-terminal SH3 domain of E . histolytica myosin IB , the only unconventinal myosin in this organism , to gain insight into the function of this important motor protein . On the basis of the crystal structure of EhMySH3 [17] and predictions by SH3 HUNTER , a FYVE domain containing guanine nucleotide exchange factor ( GEF ) , EhFP10 was predicted to possess multiple proline-rich motifs in its c-terminal domain , which could interact with the yeast myosin I SH3 domain . The known targets of the myosin I SH3 domain have all been shown to be multi-domain proteins with PRDs ( proline rich domains ) and to help in the formation of the Arp2/3-mediated endocytic complex . These observations led us to further investigate the potential for EhFP10 to serve as a binding partner of the EhMyosin IB SH3 domain . SPR studies and the co-crystal structure of peptide P2 from cterEhFP10 with the EhMySH3 domain , as well as a pull-down assay validated the proposed interaction . CterEhFP10 has a unique domain organisation , not present in any homologous GEFs . It has sequence similarity only with the APC basic domain from adenomatous polyposis coli protein and the tau protein domain , though it is very low . Both the APC basic and tau domains have been shown to bind actin filaments as well as microtubules and to each act as a molecular linker mediating actin-microtubule crosstalk [43 , 44] . Tau has also been shown to bind SH3 domains like Fyn and cSRC and to facilitate cSRC-mediated actin rearrangements in neuronal cells [45 , 46] . The EhFP10 c-terminal domain showed biochemical similarity with the APC basic domain and tau , as it was seen to interact with filamentous actin and form thick bundled actin filaments . Interaction of EhMySH3 with cterEhFP10 was observed to inhibit the actin bundling activity of EhFP10 . Increasing the concentration of cterEhFP10 was observed to overcome the inhibition and to result in an efficient bundling of the actin filaments . EhFP10 was found to be localized in the membrane ruffles and protrusion at the site of pinocytosis and phagocytosis in normal proliferating trophozoites and in cells induced for erythrophagocytosis , respectively . EhFP10 was present from initiation until the closure of the cup during both phagocytosis and macropinocytosis . However , EhFP10 was associated with the membrane encircling the newly internalized pinosomes , but disappeared after a while and was not present in newly formed phagosomes . This interesting observation states that there are differences in the regulation of these two endocytic processes . Previous studies have shown membrane ruffles to be regions of active actin polymerization and reorganization and these ruffles to eventually lead to the formation of macropinosomes . Macropinocytosis , like phagocytosis , is a process driven by the actin cytoskeleton . Formation and deformation of thick bundled actin filaments provide the force to push the membrane invagination and cell components inward , against the turgor pressure of the cell , apparently assisting in the endocytic cup progression . EhMyosin IB was found to colocalize with EhFP10 during macropinocytosis and phagocytosis in immunostained cells . EhMyosin IB mostly localized to the tip of the progressing cup while EhFP10 colocalized with both actin encircling the cup , and with EhMyosin IB . During Fc-receptor-mediated phagocytosis and macropinocytosis in macrophages , human Myo Ie ( also known as myosin IC ) , a homolog of EhMyosin IB , was shown to localize at the distal margin of the phagocytic cup and is believed to mediate a PI3K-dependent contractile activity for the closure of the cup aperture into an intra-cytoplasmic phagosome [47] . We hypothesize that in Entamoeba , EhMyosin IB and EhFP10 together are also involved in a similar kind of PI-3k-dependent contractile activity for endosome closure during endocytosis . As EhFP10 contains an FYVE domain , known to recognize phosphoinositide-3-phosphate ( PI-3P ) formed by the action of PI-3 kinases . Hence , EhMyosin IB and EhFP10 must act downstream of PI-3 kinases . Also , EhFP10 must be among the first molecules to be recruited to the endocytic cup as we showed in the live cell imaging results that it gets recruited to the membrane ruffles at an early stage , leading to endocytic cup formation . We hypothesize that EhFP10 reorganizes the actin filaments into thicker bundles , which act as a driving force for membrane invagination and cup formation . As the cup advances to closure , EhMyosin IB enriches into the tip of the growing cups; it interacts more with EhFP10 , which in turn inhibits actin bundling and promotes crosslinking of actin filaments , providing the required contractile activity leading to cup closure . However , further studies of myosin IB and EhFP10 along with other regulatory proteins like Arp2/3 , CaBP3 , CaBP5 , Rho GTPases , etc . remain to be carried out to obtain a complete overview of the pathway involved in the endocytic process . The known effectors of the SH3 domain of unconventional myosin I have been found to be WASP and WASP-related proteins which regulate the actin dynamics directly or indirectly via the Arp2/3 complex leading to actin nucleation . Several GTPases act upon these effectors and adaptor proteins like Vrp1 , Bee1p , Las17p , CARMIL , and Acan125 and this action leads to Arp2/3 dependent actin polymerization at the site of cup formation [39] . E . histolytica genome lacks the WASP/SCAR proteins but has WASP related proteins like WASH and MIM that possess VCA domain responsible for activating Arp2/3 complex mediated actin nucleation [48] . The ability of WASH and MIM proteins to promote actin nucleation and their involvement in phagocytosis still needs experimental validation [48 , 49] . As E . histolytica lacks homologue of WASP/WAVE which links phosphoinositides mediated rho pathway and actin nucleation , the existence of a unique regulatory pathway which connects both is possible . Hence , this highly expressing , unique RhoGEF EhFP10 , which can regulate actin dynamics via its c-terminal tail and possibly with downstream RhoGTPases , is of great importance and advantage to the pathogen . Our findings have revealed a novel mode of regulation of endocytic processes in highly motile phagocytosing gastric pathogens such as E . histolytica .
The primer sequences of EhMy1TD , GSTSH3 , EhGEFD , cterEhFP10 and NGFP-EhFP10 and their respective vector backbone and restriction site details are given in S1 Table . All of the genes were amplified by carrying out PCR ( Eppendorf ) using respective forward and reverse primers from the genomic DNA of Entamoeba histolytica strain HM1:IMSS . The PCR products and the vectors were double digested with their respective restriction enzymes ( Thermo Fisher ) . These digested and purified products were then ligated using T4DNA ligase ( Thermo Fisher ) and kept at 16°C for 16 hours . The ligation mixture was then transformed into E . Coli DH5α , plated on antibiotic-containing LB agar plates and kept at 37°C overnight . The colonies were screened for positive clones . The clone was further confirmed by double digestion of isolated plasmids and gene sequencing . The compositions of all the buffers used in purification are described in S2 Table . EhMySH3 was purified as previously reported [17] . EhMy1TD , GSTSH3 , and EhGEFD recombinant plasmids were transformed into BL21 ( DE3 ) and cterEhFP10 in the E . coli strain BL21 ( C41 ) . The secondary culture was grown at 37°C using 1% primary culture grown overnight from a single BL21 colony and was induced with 0 . 2 mM IPTG for 4 hours . Pellets were resuspended in their respective lysis buffer and sonicated following 4–5 freeze-thaw cycles . After sonication , cell lysate was centrifuged at 13 , 000 rpm for 30 minutes . The supernatant obtained was loaded onto a Ni-NTA column ( Sigma ) with Ni-sepharose ( GE healthcare ) . EhMy1TD , GSTSH3 , and EhGEFD proteins were eluted with their respective elution buffers following wash buffers . cterEhFP10 protein was co-expressed with a chaperone . To remove the chaperone , after the protein was bound to Ni-sepharose , it was incubated for 1 hour in incubation buffer ( 2 mM Mg-ATP , 50 mM Tris pH-7 . 5 , 300 mM NaCl , 10 mM MgCl2 , 5% glycerol , 10 mM imidazole and 5 mM β-ME ) and then was washed with a small amount of heat-denatured E . coli protein in the wash buffer . After 3–4 cycles of incubation followed by washes , cterEhFP10 protein was eluted in the elution buffer . Purified proteins were concentrated using centricon filter ( Amicon , Millipore ) and further purified by carrying out gel permeation chromatography using a Superdex G75 16/60 column and G200 10/300 columns ( GE Healthcare ) pre-equilibrated in the buffer . The peak fractions were pooled together and checked using SDS-PAGE ( S1 Fig ) . cterEhFP10 showed a regular degradation pattern soon after removal of the chaperone; hence concentrated protein was frozen immediately after purification and stored at -80°C . Most SH3-domain-binding sequences can be classified into three categories: PxxP type , class I type [ ( K/R ) xxPxxP] and class II type [xPxxPx ( K/R ) ] . To identify proteins that interact with the c-terminal SH3 domain of EhMyosin IB , a previously published proteomics data was utilized [30] . All of the listed proteins were screened for the presence of any of the three above-mentioned polyproline sequences by using the SH3 HUNTER web server [50] . From all the screened proteins , those having polyproline sequences similar to the ones recognised by yeast Myo3 and Myo5 ( unconventional myosins ) were shortlisted for further studies . Out of these , five peptides were selected for further studies , such that we had atleast one from each class of the SH3-binding polyproline sequences ( Tables 1 and 2 ) . For co-crystallization with all five synthesized peptides , EhMySH3 protein samples , each at a concentration of 30 mg/ml , were mixed with the respective peptides in different molar ratios and incubated overnight at 4°C . Crystallization trials were done with several commercial screens from Hampton Research and Molecular Dimensions . After four days at 4°C , small needle-like crystals appeared for the EhMySH3-P2 complex ( mixed with in a molar ratio of 1:2 ) in crystal screen II ( Hampton Research ) conditions 0 . 2 M ammonium sulphate , 30% PEG 8000 and 0 . 2 M ammonium sulphate , 30% PEG 4000 . Better crystals were obtained by macroseeding in 0 . 2 M ammonium sulphate and 30% PEG 8000 condition ( S2 Fig ) . Diffractable crystals were flash frozen using mother liquor as a cryoprotectant in the cold room since the crystals were very temperature sensitive . The X-ray diffraction data for EhMySH3-P2 were collected at the BM14 synchrotron beamline ( ESRF , Grenoble , France ) . The data sets were indexed and scaled using HKL2000 [51] . The structure was determined by molecular replacement using the PHASER MR Program of the CCP4 program suite [52 , 53] with the crystal structure of the c-terminal SH3 domain of myosin IB from E . histolytica ( PDB ID 5XGG ) [17] as a template for the EhMySH3-P2 peptide crystals . After a few cycles of manual building using COOT [54] , refinement was done using REFMAC5 [55] in the CCP4 suite followed by PHENIX refine [56] , resulting in a model with an R-factor of 0 . 19 ( and Rfree of 0 . 23 . Data statistics are given in Table 3 . Purified GSTSH3 and GST were incubated for one hour at 4°C with glutathione Sepharose ( GE Healthcare Lifesciences ) pre-equilibrated in equilibration buffer ( 50 mM Tris pH 7 . 4 , 150 mM NaCl ) . Then , it was washed thrice with the equilibration buffer and incubated with the same amount of purified cterEhFP10 for two hours at 4°C . Following 3–4 washes with wash buffer ( 50 mM Tris pH 7 . 4 , 300 mM NaCl ) , the proteins were eluted with 20 mM reduced glutathione , 50 mM Tris pH 7 . 4 , and 150 mM NaCl . The fractions were loaded onto a 12% SDS-PAGE gel and analyzed . G-actin was purified from rabbit muscle powder ( sigma ) [57] . For F-actin preparation , purified G-actin at a concentration of 23 μM was dissolved in G-buffer ( 20 mM Tris-Cl , 0 . 5 mM DTT , 0 . 2 mM ATP , 0 . 1 mM CaCl2and 0 . 1 mM NaN3 ) and was allowed to polymerize for 1½ hours at RT by adding to its solution an Imix solution ( 50 mM KCl , 2 mM MgCl2 , and 0 . 5 mM ATP ) . cterEhFP10 and EhMySH3 proteins of various concentrations were each added to separate samples of the mixture containing polymerized F-actin in a total volume of 200 μl and incubated for 30 minutes at RT . Each of the resulting mixtures was centrifuged at 40 , 000 rpm for two hours in a Beckman ultracentrifuge . Pellet and supernatant fractions were collected and analyzed using 10% SDS-PAGE followed by Coomassie blue staining . For bundling assays , cterEhFP10 and EhMySH3 proteins of various concentrations were each added to separate samples of the mixture containing polymerized F-actin in a total volume of 50 μl and incubated for 30 minutes at RT . Each mixture was centrifuged at 12 , 000 rpm for 10 minutes . Pellet and supernatant fractions were collected and analyzed using 10% SDS-PAGE followed by Coomassie blue staining . Purified alpha-actinin from rabbit skeletal muscle ( cytoskeleton Inc . ) was used as positive control and Bovine serum albumin ( SRL ) was used as negative control for both binding and bundling studies . The purified antigenic proteins ( EhMy1TD for mEhMyosin1B and EhGEFD for rEhFP10 Ab ) were dialyzed against PBS . Six mice were immunized subcutaneously with 100 μg of protein per mice per injection with an interval of two weeks between each injection . The first dose of the protein was emulsified with complete Freund’s adjuvant while the following doses were emulsified with incomplete Freund’s adjuvant . Following the immunization series , the serum of each mouse was stored in aliquots at -80°C . To check the titer of the polyclonal antibody , western analysis of the E . histolytica lysate was performed . E . histolytica strain HM1:IMSS trophozoites were grown axenically in TYI-S-33 medium [58] . The cells were maintained and grown in TYI-33 medium complemented with 15% adult bovine serum , 1X Diamond’s vitamin mix and antibiotic ( 125 μl of 250 U/ml benzyl penicillin and 0 . 25 mg/ml streptomycin per 90 ml of medium ) . Transfection was performed by carrying out electroporation as described previously [59] . Briefly , trophozoites in log phase were harvested and washed with PBS followed by incomplete cytomix buffer [10 mM K2HPO4/KH2PO4 ( pH 7 . 6 ) , 120 mM KCl , 0 . 15 mM CaCl2 , 25 mM HEPES ( pH 7 . 4 ) , 2 mM EGTA , 5 mM MgCl2] . The washed cells were then resuspended in 0 . 8 ml of complete cytomix buffer ( incomplete cytomix containing 4 mM adenosine triphosphate , 10 mM glutathione ) containing 200 μg of plasmid DNA and subjected to two consecutive pulses of 3000 V/cm ( 1 . 2 kV ) at 25 μF ( Bio-Rad , electroporator ) . The transfectants were initially allowed to grow without any selection . Drug selection was initiated after two days of transfection in the presence of 10 μg/ml G-418 for constructs with luciferase reporter gene . One million trophozoites growing in log phase were harvested by centrifugation at 280 g for 7 min at 4°C . The pellet was then washed with chilled PBS pH 7 . 2 and pelleted . For probing EhFP10 , washed cells were fixed with 3 . 7% pre-warmed paraformaldehyde for 60 min at room temperature . After the cells were fixed , they were again harvested at 280 xg for 7 min . In each of the two cases , the pellet was heated at 80°C for one minute , and then resuspended in 2x SDS dye without β-ME . The lysate was heated at 100°C for 7 minutes and centrifuged at 13000 rpm for 15 minutes to pellet down the debris . The supernatant was aliquoted and stored , and quantification was achieved by performing a BCA . E . histolytica trophozoites and transfectants were resuspended in incomplete TY1-33 medium and transferred onto acetone-cleaned coverslips placed in a Petri dish . The cells were allowed to adhere for 5 min at 37°C . The culture medium was discarded and the cells were fixed with 3 . 7% pre-warmed paraformaldehyde for 30 min . After fixation , the cells were permeabilized with 0 . 1% Triton X-100/PBS for 5 min , washed with PBS and then quenched for 30 min in PBS containing 50 mM NH4Cl . The coverslips were blocked with 1% BSA/PBS for 2 h , followed by incubation with primary antibody at 37°C for 1 . 5 h . The coverslips were washed three times with 1% BSA/PBS and then incubated with secondary antibody for 45 min at 37°C . Antibody dilutions used were mEhMyo1B at 1:100 , anti-GFP ( Sigma ) at 1:100 , rEhFP10 Ab at 1:150 , anti-rabbit/mice Alexa 488 , Alexa 556 and Pacific blue-410 ( Molecular Probes ) at 1:250 , TRITC-Phalloidin at 1:250 . Cells were further washed with 1% BSA/PBS twice and then PBS once and mounted on a glass slide using 1 , 4-diazbicyclo ( 2 , 2 , 2 ) octane ( DABCO ) ( 55 ) 2 . 5% in 80% glycerol . The edges of the coverslip were sealed with nail-paint to avoid their becoming dried out . Confocal images were made using an Olympus Fluoview FV1000 laser scanning microscope . Amoebic cells expressing EhFP10 tagged at its N-terminus with GFP , ( NGFP-EhFP10 ) were plated onto a 35 mm glass bottom dish and allowed to adhere to the dish . RBCs were labelled using CFDA SE dye ( Carboxyfluorescein diacetate succinimidyl ester , Thermo Fisher Scientific ) . RBCs were collected by pricking a finger with a needle and collecting the blood in PBS . The cells were washed with PBS twice followed by incubation in PBS containing 10 μM CFDA for 10 min at 37°C with intermittent tapping . The reaction was stopped by washing the RBCs with PBS and then the RBCs were kept in ice . Labeled RBC or TRICT-Dextran ( Sigma ) containing media was added and time-lapse imaging was done using a spinning disk confocal microscope ( Nikon A1R , Optics- Plan Apo VC606 oil DIC N2 , Camera- Nikon A1 , NA-1 . 4 , RI-1 . 515 ) . The temperature was maintained at 37°C with the help of a chamber provided along with the microscope . The images were captured at 500 ms intervals . The raw images were processed using NIS element 3 . 20 analysis software . All of the statistical analyses were done using GraphPad PRISMA 7 software using Student’s two-tailed t-test . Images of cells which were immunostained for all three proteins EhFP10 , myosin IB and actin proteins were used for quantitation . Data with P < 0 . 05 were considered to be significant .
|
Entamoeba histolyica is a highly motile human pathogen which eats the blood cells and immune cells by phagocytosis during progression of Amoebiasis disease . E . histolytica infections are a major concern in the developing countries . Myosins are motor proteins that move over actin cytoskeleton to drive the cellular processes . Unconventional myosins are a type of myosin which are different from myosin present in muscles , and are involved in regulation of membrane dependent processes crucial for cellular movement and endocytosis . In contrast to other eukaryotes , Entamoeba has only one unconventional myosin , Myosin IB which shows more similarity with metazoan myosins rather than amoeboid myosins . Myosin IB has been shown to be involved in phagocytosis . The exact role played by Myosin IB in the phagocytic process is still not fully understood . SH3 domain is present at the c-terminal tail of Myosin IB which has been found to interact with proteins that regulate the actin cytoskeleton in other organisms . In this work , we have identified EhFP10 as one of the interacting proteins of EhMyosin IB SH3 domain through a co-crystal structure and biophysical experiments . Our localisation studies demonstrated the involvement of EhFP10 in phagocytosis and pinocytosis . This is the first report of the involvement of a FYVE domain containing GEF in pinocytosis . We have also analysed that EhFP10 has a unique c-terminal domain not present in any other FYVE family GEFs in Entamoeba as well as in other organisms . Actin binding studies indicated that the c-terminal domain of EhFP10 binds to actin filaments and leads to formation of thicker actin bundles . Myosin IB interaction with EhFP10 inhibits the formation of actin bundles . Through our results , we could hypothesize that the presence of a unique GEF like EhFP10 could compensate for the absence of WASP proteins in E . histolytica which have been found to interact with the myosin I SH3 domain in other organisms and regulate actin dynamics during endocytosis . Our study reveals a rare interaction of a myosin with a GEF , which interact to regulate actin bundling . EhMyosin IB is different from other amoeboid myosins and lies between the metazoan and amoeboid myosins like human myosin IE . Hence , the findings have broader implication to completely understand the closure stage of a phagocytic and pinocytic cup .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"cell",
"motility",
"actin",
"filaments",
"cell",
"processes",
"parasitic",
"protozoans",
"guanine",
"nucleotide",
"exchange",
"factors",
"molecular",
"motors",
"actin",
"motors",
"protozoans",
"motor",
"proteins",
"contractile",
"proteins",
"actins",
"proteins",
"entamoeba",
"histolytica",
"phagocytosis",
"biochemistry",
"cytoskeletal",
"proteins",
"signal",
"transduction",
"eukaryota",
"cell",
"biology",
"protein",
"domains",
"myosins",
"biology",
"and",
"life",
"sciences",
"cell",
"signaling",
"organisms",
"signaling",
"molecules"
] |
2019
|
EhFP10: A FYVE family GEF interacts with myosin IB to regulate cytoskeletal dynamics during endocytosis in Entamoeba histolytica
|
Emerging pathogens are a major threat to public health , however understanding how pathogens adapt to new niches remains a challenge . New methods are urgently required to provide functional insights into pathogens from the massive genomic data sets now being generated from routine pathogen surveillance for epidemiological purposes . Here , we measure the burden of atypical mutations in protein coding genes across independently evolved Salmonella enterica lineages , and use these as input to train a random forest classifier to identify strains associated with extraintestinal disease . Members of the species fall along a continuum , from pathovars which cause gastrointestinal infection and low mortality , associated with a broad host-range , to those that cause invasive infection and high mortality , associated with a narrowed host range . Our random forest classifier learned to perfectly discriminate long-established gastrointestinal and invasive serovars of Salmonella . Additionally , it was able to discriminate recently emerged Salmonella Enteritidis and Typhimurium lineages associated with invasive disease in immunocompromised populations in sub-Saharan Africa , and within-host adaptation to invasive infection . We dissect the architecture of the model to identify the genes that were most informative of phenotype , revealing a common theme of degradation of metabolic pathways in extraintestinal lineages . This approach accurately identifies patterns of gene degradation and diversifying selection specific to invasive serovars that have been captured by more labour-intensive investigations , but can be readily scaled to larger analyses .
Understanding how bacteria adapt to new niches and hosts and thus emerge or re-emerge as a cause of infectious disease in human and animals is of critical importance to anticipating and preventing epidemic disease [1 , 2] . With the decreasing cost of genome sequencing , comparative genomics has become a rich source of insight into the origins and movement of bacteria in new pathogenic niches . However , translating whole genome sequence databases into mechanistic and functional insights remains a challenge . Early expectations were that pathogen evolution would be driven primarily by the acquisition of virulence factors . However , as whole-genome sequencing has become increasingly routine , a decidedly more complex picture has emerged [3 , 4] . A pattern of bacterial entrance to a new niche followed by adaptation through the loss of antivirulence loci and reduced metabolic flexibility is now recognised as a paradigm of the emergence of important human pathogens from non-pathogenic bacterial species [5–8] . These new niches can be the result of virulence factor acquisition providing access to a previously inaccessible niche in a so-called foothold moment [8] , or the emergence of new host niches driven by chronic disease [9–11] . While pathogen and host requirements for infection vary , there is increasing evidence of parallel evolution in bacteria adapting to the same or similar host niche . This is perhaps nowhere more evident than in the species Salmonella enterica . Salmonella enterica strains that cause disease in warm-blooded mammals lie on a spectrum from those that have a broad host range and cause self-limiting gastrointestinal infection , to those that are more restricted in host range , but cause systemic disease and are typically associated with higher mortality [11 , 12] . Host-restricted , extraintestinal variants of Salmonella enterica have evolved independently multiple times from gastrointestinal ancestors [13] , and show a greater degree of gene degradation compared to their generalist relatives [14–16] . There are common patterns in the genes that undergo pseudogenization in invasive Salmonella , most obviously an extensive network of genes required for anaerobic metabolism in the inflamed host [17 , 18] , a pattern with parallels in other host-adapting enteropathogens [5] . Identifying these signals of parallel evolution has been challenging , relying mainly on manual annotation and comparison of pseudogenes [17 , 18] . Detection of pseudogenes in particular relies on ad-hoc criteria to identify large truncations , deletions , or frameshifts [19 , 20] . It is rare that the same genes or complete pathways are pseudogenized in host-adapted species; rather interpretation has relied on identifying overrepresentation of independent pseudogenization events clustered in certain pathways [17] . If pseudogenization leads to pathway attenuation or inactivation , it seems likely that reduced selective pressure will lead to a higher incidence of detrimental mutation fixation in other genes in these pathways . Indeed , we have previously shown that functional variant calling , based on sequence deviation from patterns of conservation observed in deep sequence alignments , shows a similar functional signal in host-restricted Salmonella enterica serovar Gallinarum to pseudogene analysis [21] , identifying a larger cohort of genes where constraints on drift appear to have been lifted during host-adaptation . In previous work we developed delta bitscore ( DeltaBS ) , a profile hidden Markov model ( HMM ) based approach to functional variant calling [21] . The basic assumption of this approach is that variation in conserved positions of a protein sequence is more likely to affect protein function than variation in less conserved regions . This approach can integrate information about nonsynonymous mutations , indels , and truncations . We have previously shown that DeltaBS can successfully identify functional changes in genes that would be missed by standard pseudogene analysis [22] , and that a subset of genes in host-adapted strains appear to accumulate large DeltaBS values [21] . Additionally , others have observed similar changes in DeltaBS distributions during adaptation of Salmonella to a single immunocompromised host [10] . We generally assume that a large DeltaBS value is indicative of a decay in protein function , however a modest increase in DeltaBS associated with a phenotype may instead be indicative of diversifying selection . Here , we have leveraged these previous observations to identify signatures of mutational burden consistent with adaptation to an invasive lifestyle . We have developed a random forest classifier using DeltaBS functional variant calling [21] that can perfectly separate intestinal Salmonella serovars from host-adapted , extraintestinal serovars . We use random forest models because they perform well on datasets with few informative variables [23 , 24] , and the decision tree structure they employ has the potential to detect functional relationships ( i . e . epistasis ) between genes [25 , 26] . They have been applied successfully in the past to predict microbial phenotype using gene presence/absence data [27] , and SNPs already known to be associated with phenotype [28 , 29] . We show that these models produce interpretable signatures of host-adaptation , and furthermore that these signatures can be detected in strains of Salmonella associated with invasive disease in immunocompromised populations in sub-Saharan Africa .
The approach taken in this investigation is summarised in Fig 1 , and described below . We built our model using a collection of genomes from well-characterised reference strains of gastrointestinal and extraintestinal Salmonella serovars ( S1 Table ) , drawing on the extensive curation of orthology relationships performed by Nuccio and Bäumler [17] . These strains were originally characterised as “gastrointestinal” or “extraintestinal” based on common patterns of gene degradation , host restriction and clinical characteristics observed among the extraintestinal strains [17] , and we have employed this same categorisation our analysis . We scored the functional importance of sequence variation by comparing the protein coding genes of each serovar to profile HMMs from the eggNOG database [30] , designed to capture patterns of sequence variation typically seen in the protein coding genes of Gammaproteobacteria ( see Methods ) . For each genome , the functional significance of sequence variation within protein coding genes is quantified using the DeltaBS metric . Following scoring , a bootstrap sampling of genomes are used to train each decision tree . For each node in the tree , a random subset of genes are sampled , and the most informative gene from this set is chosen to split the data . For each node in the tree , the predictive utility of the selected gene ( variable importance ) is tested by calculating how well the gene separates the samples according to phenotype . We then employed random forests to identify the genes which were most informative of phenotype when viewed collectively . Random forests work by building an ensemble of decision trees designed to predict a characteristic of the samples [31] , in this case adaptation to an extraintestinal , or invasive , niche . For each node in the decision tree , the best gene of a random sampling from the training gene set is selected according to its ability to separate a randomly selected subset of samples by phenotype based on DeltaBS values . The process of building a random forest produces measures of variable importance that can be used to assess the relative utility of different genes in classification of Salmonella strains based on lifestyle . To obtain an indication of the proportion of the genome that shows patterns of unusual sequence variation associated with an invasive phenotype , we trained a random forest model on a set of 6 , 438 orthologous genes . Accuracy of the model was assessed using out-of-bag accuracy . This out-of-bag ( OOB ) measure of accuracy gives us an indication of how well each decision tree in the forest performs at predicting phenotype in a serovar it has never encountered before , using information on DeltaBS differences collected from other serovars . Next , we performed iterative feature selection to improve the performance of the model . This process involved repeated rounds of selecting the top 50% of predictors and re-training the model , until the model achieved perfect OOB predictive performance on the training dataset ( Fig 2A ) . When the full set of filtered orthologous genes was used to build a model , a subset of genes ranked much higher than the others in variable importance ( VI ) ( Fig 2B ) . We then saw a tailing off of VI , resulting in 4 , 721 orthologous groups either not being used in the model , or not improving classification accuracy ( as indicated by VI = 0 ) . This set of genes was discarded in the first round of feature selection , and 1 , 521 genes were discarded in the subsequent three rounds . The final model used 196 of the original 6 , 438 genes for prediction ( S2 Table ) . This model additionally achieved perfect classification accuracy on an independent set of genomes of the same serovars as our training data ( S1 Fig ) . We tested for overfitting using permutation tests , and for correlation bias [32] using a variety of alternative model building strategies , and found no evidence for either phenomenon in our model ( S1 File ) . We anticipated that the majority of informative genes identified in our study would be genes that showed functional degradation in invasive isolates but not in gastrointestinal isolates . Of the top predictors in our study ( N = 196 ) , 154 showed significantly greater mutational burden in extraintestinal strains compared to gastrointestinal strains ( Mann-Whitney U test , adjusted P-value < 0 . 05 ) , compared to 9 genes that showed significantly greater mutational burden in gastrointestinal strains . Of the genes that were more conserved in invasive isolates , one was the aldo-keto reductase yakC , which was deleted or truncated in all but one gastrointestinal strain and intact in all invasive strains . Another was the chaperone protein yajL , which appears to be important for oxidative stress tolerance [33 , 34] . Among the top predictors were several sets of genes belonging to the same operon ( S2 Table ) . Examples included the ttr , cbi and pdu operons , which are all required for the anaerobic metabolism of 1 , 2-propanediol [35] . These operons have previously been identified as key degraded pathways in invasive isolates [16–18] , and indicate the agreement of this method with other studies linking loss of gene function to host niche . Overall , a large proportion of the identified genes were involved in metabolism ( Fig 2C ) , consistent with the findings of similar studies [17 , 18] . Of the 167 central metabolism genes identified by Nuccio and Bäumler [17] as truncated or deleted in at least one extraintestinal serovar , only one of these was previously reported to be truncated in > 4 serovars . In contrast , we found that 20 of the 167 central metabolism genes were identified by our model as informative of phenotype , indicating that including signal from more subtle forms of loss of function improves our ability to detect parallelism across lineages of invasive Salmonella . Of the 13 genes reported to be frequently disrupted by Nuccio and Bäumler , our approach identified 9 . The other 4 were either not a match to profile HMMs in our database , or the truncation did not fall within the span of the model . Other major categories affected include proteins involved in cell wall and membrane function , perhaps suggesting changes affecting recognition by the host immune system , and signal transduction , suggesting some degree of consistent regulatory rewiring during adaptation to an extraintestinal niche . Information provided by multiple genes was often more informative of phenotype than a single gene individually , as was the case for fimD and fimH ( S2 Fig ) . FimD and FimH constitute central components of type 1 pili , and both are required for expression of normal fimbriae [36] . This demonstrates that our approach is capable of identifying epistatic relationships between genes , where a modification in function of one gene masks the functional status of the other . When examining individual genes that showed differences in mutational burden between invasive and gastrointestinal isolates , we found that most of these mutations had occurred independently , and had occurred at different sites in the protein . Using a permissive threshold ( DBS>3 ) , or a conservative threshold ( DBS>5 ) , there were close to twice as many deleterious , independent mutations in the genes of the invasive serovars than those of the gastrointestinal ( 476:910; 537:991 , respectively , see Methods ) . This phenomenon was even more pronounced when only mutations with DBS over the upper quartile were counted ( 249:612 , S3 Table ) . While the majority of genes identified appeared to be cases of gene degradation in invasive lineages , some genes showed more subtle signs of mutational burden , restricted to nonsynonymous changes of modest predicted functional impact . An example of this , Fig 2D , illustrates mutation accumulation in one of the top candidate genes , mrcB , encoding penicillin-binding protein 1b ( PBP1b ) . Not only does mrcB carry more mutations in invasive serovars compared to gastrointestinal serovars , the mutations have occurred independently in different positions within the protein . Penicillin-binding proteins are the major target of β-lactam antibiotics and are important for synthesis and maturation of peptidoglycan [37] . PBP1b in particular extends and crosslinks peptidoglycan chains during cell division . While PBP1b is not essential , it has been shown to be synthetically lethal when the partially redundant mrcA/PBP1a is deleted , and is important in E . coli for competitive survival of extended stationary phase , osmotic stress [38] , and—in Salmonella Typhi—growth in the presence of bile [39] . Bile is an important environmental challenge for Salmonella , particularly for extraintestinal serovars which colonize the gall bladder [40] . While there are more mutations in invasive than in gastrointestinal serovars , the mutations that occur in this protein are all amino acid substitutions of modest predicted impact . This suggests that sequence changes could result in a modification of protein function , rather than a loss , consistent with the importance of PBP1b for the survival of S . Typhi during a typical infection cycle [39] . To anticipate the performance of our random forest model on new data we computed out-of-bag ( OOB ) error . Because random forests train each decision tree on a random subset of the training data , OOB error can be computed by testing the performance of these trees on data they have not been trained on , providing inbuilt cross-validation [31] . In our case , perfect OOB classifications were only achieved by the fifth iteration of the model . The need for iterative improvement of the model came from difficulty in correctly classifying the reference strains for serovars Enteritidis and Dublin . This is reflective of their relatively recent divergence and niche adaptation compared to other serovars in the study ( S3 Fig , [18] ) . S . Gallinarum was classified much more readily than S . Enteritidis and S . Dublin , despite being closely related to both serovars , perhaps due to its host restriction . S . Enteritidis was initially mis-classified as invasive , indicating that it shares genomic trends with invasive lineages . Genomic analyses have indicated that the ancestor of S . Enteritidis previously possessed intact pathogenicity islands ( SPI-6 and SPI-19 ) , each encoding a type six secretion system [18 , 41] . These loci have been implicated in host-adaptation and survival during extraintestinal infection [42 , 43] , and it has been speculated based on their loss and other evidence that classical S . Enteritidis has been adapting towards greater host generalism with respect to its ancestral state [18] . This could explain the greater number of disrupted and deleted genes relative to other gastrointestinal serovars used in this study , and the difficulty in classifying it correctly . Conversely , S . Dublin was initially mis-classified as gastrointestinal . In previous studies S . Dublin has been shown to possess fewer pseudogenes than related invasive isolates [17 , 18] , suggesting a lower degree of host adaptation than other invasive isolates . Indeed , S . Dublin is more promiscuous in its host range , primarily infecting cattle [44] while still causing sporadic human disease [45] . It seems likely that a subset of informative genes identified in early iterations of the model may have been indicators of host restriction or generalism rather than broad extraintestinal adaptation . In recent years there have been reports of novel S . Typhimurium and S . Enteritidis lineages associated with invasive disease in sub-Saharan Africa [46–48] in populations with a high prevalence of immunosuppressive illness such as HIV , malaria , and malnutrition [49] . These lineages contribute to a staggering burden of invasive non-typhoidal salmonella ( iNTS ) disease , which is responsible for an estimated 3 . 4 million cases and circa 680 , 000 deaths annually [50] . Based on epidemiological analysis , high-throughput metabolic screening of selected strains , and analysis of pseudogenes it has been suggested that these lineages may be rapidly adapting to cause invasive disease in the human niche created by widespread immunosuppressive illness [11 , 46–48 , 51] . Two iNTS-associated lineages have recently been described within serovar Enteritidis [48] , geographically restricted to West Africa and Central/East Africa , respectively . Initial observations have demonstrated that a representative isolate of the Central/East African clade has a reduced capacity to respire in the presence of metabolites requiring cobalamin for their metabolism and has lost the ability to colonize a chick infection model [48] , suggesting adaptation to a new host niche . Similarly , two iNTS disease associated lineages have been described in serovar Typhimurium [47] , both members of sequence type 313 ( ST313 ) , generally referred to as Lineage I and II in the literature . Lineage II appears to have largely replaced Lineage I since 2004 , and it has been suggested this is due to Lineage II possessing a gene encoding chloramphenicol resistance [47] . Laboratory characterization of Lineage II strains has shown that they are not host-restricted [52 , 53] , but do appear to possess characteristics suggestive of adaptation to an invasive lifestyle [54–57] , though it is important to note that this is a complex trait and not easily quantified . Given the evidence of adaptation to an invasive niche in these lineages , we asked if genomic signatures of extraintestinal adaptation we had detected previously could be detected in iNTS disease associated lineages . To this end , we applied our predictive model trained on well-characterized extraintestinal strains to calculate an invasiveness index , the fraction of decision trees in the random forest voting for an invasive phenotype . First , we compared isolates from African iNTS-associated clades of S . Enteritidis ( N = 233 ) to a global collection of isolates generally associated with intestinal infection ( N = 100 ) [48] . Our model gave iNTS-associated S . Enteritidis strains a higher invasiveness index than the globally distributed isolates ( Fig 3A and 3B , S4 Table ) , indicating the presence of genetic changes paralleling those that have occurred in extraintestinal serovars of Salmonella . Similar gene signatures were only rarely observed in the global epidemic clade ( Fig 3C ) . These findings are consistent with the metabolic changes observed by Feasey et al . [48] in the Central/Eastern African clade compared to the global epidemic clade . In particular we found signs of gene sequence variation uncharacteristic of gastrointestinal Salmonella across a number of key genomic indicators , including tcuR , ttrA , pocR , pduW , eutH , SEN2509 ( a putative anaerobic dimethylsulfoxide reductase ) and SEN3188 ( a putative tartrate dehydratase subunit ) , all in pathways previously identified by Nuccio and Bäumler [17] as being involved in the utilization of host-derived nutrients in the inflamed gut environment . This indicates that our model is able to identify early signatures of adaptation , even in these recently emerged strains that still retain some capacity to cause enterocolitis [48] . To confirm this , we performed an additional comparison of S . Typhimurium ST313 isolates ( N = 208 ) , to global isolates from other STs , predominantly ST19 , associated with gastroenteritis ( N = 51 ) [51 , 58] . Similarly to iNTS associated S . Enteritidis isolates , S . Typhimurium ST313 isolates has a higher invasiveness index than isolates from other STs ( S4 Fig , S5 Table ) . Within ST313 , Lineage II scored higher than Lineage I , possibly suggesting differential adaptation to the extraintestinal niche . We found that there were in fact more degraded genes unique to Lineage I than Lineage II , but that these genes were assigned less weight in the model , so did not impact score as strongly ( S2 Fig & S3 Fig ) . Interestingly , ST313 has recently been shown not to be entirely restricted to Africa , with isolation reported in Brazil [59] and the UK [58] , associated primarily with gastrointestinal disease . We included a collection of UK ST313 strains [58] in our analysis , and found that their invasiveness index tended to be elevated compared to non-ST313 salmonellae , and intermediate between Lineage I and II , suggesting that this adaptation is not restricted to circulating African strains , as it can be seen in strains collected from other countries as well ( S5 Fig ) . This observation is consistent with the work of Ashton et al . [58] , who noted shared pseudogenes and phenotypic traits in UK and African ST313 isolates . This suggests our model is capturing features here associated with the ability to colonize an extraintestinal niche , rather than enter it in healthy individuals . In addition to the iNTS lineages we investigated , some other strains had unusually high invasiveness indices . Among the top scoring isolates outside of the African S . Enteritidis lineages are Ratin strains , a rodenticidal lineage used as commercial rat poison before the 1960s [60] . In S . Typhimurium , a clade containing strains DT99 , DT56 and U313 also scored highly . These strains appear to be adapted to birds , and DT99 and DT56 have been reported to be highly virulent in pigeons [12 , 61–63] . While the above data suggests that our model is detecting genetic changes associated with extraintestinal survival , it is difficult to infer directionality from large isolate collections . We have addressed this using a unique case of accelerated adaptation over the course of a single infection ( Fig 4 ) . We scored the invasiveness index of a collection of hypermutator S . Enteritidis isolates collected over a ten year period that were adapting to chronic systemic infection of an immunocompromised patient [10] . We found a significant positive correlation between invasiveness index and duration of carriage ( r = 0 . 96 , n = 6 , P = 0 . 002 ) . Additionally , there was a significant shift over time in the DeltaBS distribution for the genes in our model as compared to the rest of the genome ( P = 7 . 576e-05 , Mann Whitney U test ) . This suggests a specific change in selective pressure on genes inferred to be important for extraintestinal survival from established invasive serovars , and provides evidence for parallel adaptation .
In this study , we have demonstrated the insight to be gained by the layering of machine learning approaches to better understand niche adaptation in a bacterial pathogen . Firstly , profile hidden Markov models allow us to capture information on common patterns of sequence variation in protein families in order to understand the functional significance of specific mutations . Using data on the accumulation of functionally impactful mutations across the proteome as input , random forests then allow us to identify genes that display a difference in selective pressures between lineages with different phenotypes . Not only has this approach proved effective at identifying biological mechanisms behind bacterial niche adaptation , it has also allowed us to detect the emergence of new extraintestinal lineages by searching for these recurrent patterns of mutation accumulation in a way that allows the recognition of novel mutations as cases of the same underlying shift away from the sequence constraints a gene is usually subjected to . We believe this general approach will be broadly applicable to any pathogen where multiple lineages are adapting to the same niche , and will be able to detect signatures of adaptation that are missed by other methods .
High quality genomes for 13 well-characterised Salmonella enterica serovars were retrieved from the NCBI database ( accessions and serovar information can be found in S1 Table ) . The serovars were divided into gastrointestinal and extraintestinal serovars according to the classifications made by Nuccio and Bäumler [17] . Ortholog calls were also taken from the Supplementary Material of Nuccio and Bäumler [17] . A core gene phylogeny for the strains used to build the model was produced using RAxML [76] , based on a core gene alignment created in Roary [77] . Profile hidden Markov models ( HMMs ) for Gammaproteobacterial proteins were retrieved from the eggNOG database [30] . We chose this source of HMMs because it is publicly available , allowing for better reproduction of analyses , and we feel it provides a good balance between collecting enough sequence diversity to capture typical patterns of sequence variation in a protein , without sacrificing sensitivity in the detection of deleterious mutations , as we have observed with Pfam HMMs [21] . Each protein sequence was searched against the HMM database using hmmsearch from the HMMER3 . 0 package ( http://hmmer . org ) . The top scoring model corresponding to each protein was used for analysis ( N = 8 , 060 groups ) . Orthologous groups ( OGs ) with no corresponding eggNOG HMM , or more than one top model hit were excluded from further analysis ( N = 1 , 524 ) . If most genes in an OG had a significant hit ( E-value<0 . 0001 ) to the same eggNOG model , any genes within this OG that did not were assigned a score of zero , reflecting a loss of the function of that protein . These cases typically reflected a truncation that had occurred early in the protein sequence . Additionally , genes with no variation in bitscore for the match between protein sequences and their respective eggNOG HMM across isolates were excluded ( N = 188 ) . After this filtering process , 6 , 439 orthologous groups remained for analysis . Residue-specific DeltaBS ( as in Fig 2D ) was calculated by aligning orthologous sequences , choosing a reference sequence ( from S . Typhimurium ) , and substituting each variant match state and any accompanying insertions into the reference sequence and calculating the difference in bitscore caused by the substitution . The R package “randomForest” [78] was used to build random forest classifiers using a variety of parameters to assess which were best for accuracy . We used out-of-bag ( OOB ) error rate to measure the performance of the model [31] . Out-of-bag error is calculated automatically by the randomForest R package as the model is built . Briefly , calculations are performed as follows: as each decision tree is trained using a bootstrap sampling of the training genomes , a small number of samples are left aside to test the predictive accuracy of each decision tree on previously unseen samples . For each serovar , votes are collated and accuracy is calculated from only those decision trees that did not include the serovar in their training set . In this application , this step tests whether the genomic signatures of invasiveness captured by the decision trees based on some serovars are present in other serovars , and thus whether the model can detect adaptation to an invasive lifestyle in previously unseen lineages . OOB error rate , stabilised at 10 , 000 trees , so we chose this as a parameter for optimising the number of genes sampled per node ( mtry ) . mtry values of 1 , p/10 , p/5 , p/3 , p/2 and p ( where p = the number of predictors ) were tested , and we found that at mtry = p/10 , the number of genes that were either not incorporated into trees , or did not improve the homogeneity of daughter nodes when they were incorporated into trees ( as measured by mean decrease in Gini index , [79] ) stabilised at ~92% . Training the random forest classifier over five iterations took 55 seconds on a laptop computer . In order to assess how well this method would scale , we trained another model on a larger dataset of S . Enteritidis strains ( N = 677 ) using the same workflow and site of isolation as a proxy for phenotype , which took 28 minutes . To improve the performance of the model , we performed five model building and sparsity pruning cycles . For the first cycle , we built a random forest model using all genes that met the inclusion criteria , and performed sparsity pruning by eliminating all variables that had a mean Gini index ( variable importance ) of zero or lower ( meaning the gene was either not included in the model or did not improve model accuracy when it was ) . Four successive rounds of model building and sparsity pruning involved building a new model with the pruned dataset , then pruning the genes with the lowest 50% of variable importances . The resulting model had 100% out-of-bag classification accuracy . We also tested the accuracy of the full model on a collection of alternative strains related to the training dataset ( see S1 Table ) . Orthologs to the top genes identified by our model were identified using phmmer from the HMMER3 . 0 package ( http://hmmer . org ) . Additional notes on model building and testing are provided in S1 File . We tested the top 196 genes for the presence of independent mutations in each serovar by aligning each sequence to the profile HMM representing that protein family . Variation in each sequence with respect to a designated reference sequence from the set ( as selected by Nuccio and Bäumler , 2014 ) at each site in the HMM was identified and classified as either a mutation unique to a single serovar , or one shared among multiple serovars . Consecutive deletions or insertions with respect to the HMM consensus sequence were collapsed into single mutational events . Read data from Feasey et al . [48] and Klemm et al [10] was mapped to the reference genome S . Enteritidis P125109 . Reads from Okoro et al . [51] and Ashton et al . [58] were mapped to the reference genome S . Typhimurium LT2 . For samples in the Okoro study , if an isolate was sequenced using multiple runs , the most recent run was chosen for analysis . All reads were mapped using BWA mem [80] and regions near indels were realigned using GATK [81] . Picard ( http://broadinstitute . github . io/picard ) was used to identify and flag optical duplicates generated during library preparation . SNPs and indels were called using samtools v1 . 2 mpileup [82] , and were filtered to exclude those variants with coverage <10 or quality <30 . For tree building , a pseudogenome was constructed by substituting high confidence ( coverage >4 , quality >50 ) variant sites in the reference genome , and masking any sites with low confidence with an “N” . Insertions relative to the reference genome were ignored , and deletions were filled with an “N” . Pseudogenome alignments were then used as input to produce trees using Gubbins [83] to exclude recombination events , and RAxML v8 . 2 . 8 [76] to build maximum likelihood trees using a GTR + Gamma model . Samples with >10% missing base calls were excluded from the analysis . Sequences for the 196 genes of interest used in the random forest model were retrieved for each isolate and translated . These were then scored using their respective profile HMMs . Score data was collated , and any missing values were marked as ‘NA’ and imputed using the na . roughfix function from the randomForest R package [78] . This is a different approach used to that of the training dataset , due to the potentially lower quality of the sequenced genomes leading to gene absence due to low coverage rather than true deletion or severe truncation . The relationship between invasiveness ranking and phylogeny were visualised using Phandango [84] .
|
Researchers are now collecting a wealth of genomic data from bacterial pathogens , and this will continue to grow with the introduction of routine sequencing for disease surveillance . However , our ability to use this data to predict how changes in genome sequence lead to differences in disease is limited . Here , we have used machine learning to detect an enrichment in functionally significant mutations in genes associated with a shift in pathogenic niche . This approach captures convergence in functional outcomes that does not necessarily result in a convergence in sequence , facilitating the inclusion of rare variants of large effect in an analysis , and allowing for complex interactions between genes . We apply this approach to Salmonella , showing that we can detect changes associated with disease phenotype in emerging lineages associated with the HIV epidemic . This approach should be applicable to other bacterial species with lineages independently adapting to similar niches . We provide open-source implementations of both the predictive model , and the workflow used to build it .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"markov",
"models",
"engineering",
"and",
"technology",
"pathogens",
"microbiology",
"decision",
"analysis",
"bacterial",
"diseases",
"mathematics",
"management",
"engineering",
"enterobacteriaceae",
"bacteria",
"bacterial",
"pathogens",
"salmonella",
"typhimurium",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"infectious",
"diseases",
"decision",
"trees",
"sequence",
"alignment",
"bioinformatics",
"medical",
"microbiology",
"hidden",
"markov",
"models",
"microbial",
"pathogens",
"salmonella",
"enterica",
"salmonella",
"probability",
"theory",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"organisms"
] |
2018
|
Machine learning identifies signatures of host adaptation in the bacterial pathogen Salmonella enterica
|
The Fell and Dales are rare native UK pony breeds at risk due to falling numbers , in-breeding , and inherited disease . Specifically , the lethal Mendelian recessive disease Foal Immunodeficiency Syndrome ( FIS ) , which manifests as B-lymphocyte immunodeficiency and progressive anemia , is a substantial threat . A significant percentage ( ∼10% ) of the Fell ponies born each year dies from FIS , compromising the long-term survival of this breed . Moreover , the likely spread of FIS into other breeds is of major concern . Indeed , FIS was identified in the Dales pony , a related breed , during the course of this work . Using a stepwise approach comprising linkage and homozygosity mapping followed by haplotype analysis , we mapped the mutation using 14 FIS–affected , 17 obligate carriers , and 10 adults of unknown carrier status to a ∼1 Mb region ( 29 . 8 – 30 . 8 Mb ) on chromosome ( ECA ) 26 . A subsequent genome-wide association study identified two SNPs on ECA26 that showed genome-wide significance after Bonferroni correction for multiple testing: BIEC2-692674 at 29 . 804 Mb and BIEC2-693138 at 32 . 19 Mb . The associated region spanned 2 . 6 Mb from ∼29 . 6 Mb to 32 . 2 Mb on ECA26 . Re-sequencing of this region identified a mutation in the sodium/myo-inositol cotransporter gene ( SLC5A3 ) ; this causes a P446L substitution in the protein . This gene plays a crucial role in the regulatory response to osmotic stress that is essential in many tissues including lymphoid tissues and during early embryonic development . We propose that the amino acid substitution we identify here alters the function of SLC5A3 , leading to erythropoiesis failure and compromise of the immune system . FIS is of significant biological interest as it is unique and is caused by a gene not previously associated with a mammalian disease . Having identified the associated gene , we are now able to eradicate FIS from equine populations by informed selective breeding .
The Fell and Dales are related sturdy pony breeds traditionally used as pack animals to carry goods over the difficult upland terrain of northern England . Both breeds experienced near extinction during WWII , and the current populations are descended from very few animals . It is likely that this genetic bottleneck , together with the use of prominent sires , was responsible for the emergence of a fatal Mendelian recessive disease , FIS , which currently affects up to 10% of Fell and 1% of Dales foals ( data from UK breed societies ) . Both of these breeds are registered with the Rare Breeds Survival Trust due to their falling numbers , and important position in the UK's agricultural heritage . FIS was first described in 1998 as a unique syndrome in which affected Fell foals develop diarrhoea , cough and fail to suckle [1] . Despite an initial response to treatment , the infections persist and were shown to be due to a primary B-cell deficiency [2] associated with reduced antibody production , with tested immunoglobulin isotypes including IgM , IgGa , IgGb and IgG ( T ) being significantly reduced [3] . Paradoxically , circulating T-lymphocyte numbers are normal [4] . The reduced antibody levels in affected foals are consistent with an inability to generate an adaptive immune response , resulting in immunodeficiency once colostrum-derived immunoglobulin levels decrease at 3–6 weeks of age . This loss of maternally derived antibodies correlates with typical onset of FIS signs at 4–6 weeks . Concurrently , affected foals develop a non-hemolytic , non-regenerative progressive profound anemia [5] , in itself severe enough to cause death and the main marker for euthanasia decisions by vets . As a result of FIS , foals die or are humanely destroyed between 1–3 months of age , the disease being 100% fatal . In 2009 , this condition was reported in the Dales breed [6]; it is likely that the mutation has passed between the breeds given the similarity between them and the practice of occasional interbreeding . The clinical and pathological findings for FIS are compatible with a primary defect of genetic origin [1] , and this is supported by extensive genealogical studies [6] , [7] . FIS has a pattern of inheritance typical of an autosomal recessive disease , and the likely founder animal , which features in both the Fell and Dales studbooks , has been traced by pedigree analysis . Primary immunodeficiences , which include depleted levels of lymphocytes and/or immunoglobulins , have previously been reported in the horse . The recessive defect ‘severe combined immunodeficiency’ ( SCID ) , which is found in the Arabian breed , comprises a fatal deficiency in T- and B-lymphocyte numbers and function . The underlying lesion was found to be a 5 base-pair deletion in the gene coding DNA-dependent kinase , catalytic subunit DNA-PKCS [8] , a protein involved in V ( D ) J recombination required for adaptive immunity [9] . Like FIS foals , SCID foals have a markedly reduced thymus and have reduced numbers of germinal centers in secondary lymphoid organs [10]; unlike SCID foals however , FIS foals have apparently normal numbers of circulating T-cells [4] . Primary aggamaglobinemia is rare in horses and comprises of a complete absence of immunoglobulin and reduced peripheral B-lymphocyte levels , with normal T-lymphocyte activity [11] . In this respect there is a similarity to FIS , however primary aggamaglobinemia is only observed in males and is X-linked . Furthermore , profound anemia in combination with B-lymphopenia has not previously been reported in the horse or any other species , and as such FIS appears to be a unique disease process . Here we report the mapping and identification of the genetic lesion that causes FIS . An initial scan using microsatellite markers identified the chromosome region responsible . The opportune production of a SNP chip , which utilized the SNP variants generated during the sequencing of the equine genome ( http://www . broadinstitute . org/mammals/horse ) then allowed a confirmatory association scan . This was followed by re-sequencing of the implicated region in order to identify the causal mutation .
A genome-wide microsatellite scan was performed on 41 individuals taken from five pedigrees of Fell ponies in which FIS was segregating ( Figure S1 ) , using a panel of 228 markers ( Table S1 ) . The data were examined both for loss of heterozygosity and for linkage ( Table S2 ) . Only one microsatellite , at 30 . 25 Mb on ECA26 , showed a significant loss of heterozygosity ( χ2 = 7 . 15 , P = 0 . 028 ) and significant linkage ( LOD score = 3 . 29 at θ = 0 ) to the disease . The location of the lesion was confirmed and refined using a genome-wide association analysis with an equine SNP array ( Illumina EquineSNP50 Infinium BeadChip ) , which contains 54 , 602 validated SNPs . After applying quality control ( see Materials and Methods ) , data were available for 42 , 536 SNPs in 49 individuals ( 18 FIS-affected and 31 controls ) . To consider whether there was any population stratification among the samples , a multi-dimensional scaling plot of the genome-wide identity-by-state distances was performed ( Figure S2 ) ; there was no significant difference between the affected and control samples for the first two components ( P = 0 . 553 ) . In addition , a quantile-quantile plot ( Figure S3 ) to compare the expected and observed distributions of –log10 ( P ) , obtained by a basic association test , showed that there was little evidence of inflation of the test statistics ( genomic inflation factor λ = 1 . 04 ) , indeed the test statistics appear to be marginally depressed rather than inflated . No correction was considered necessary . Two SNPs on ECA26 showed genome-wide significance after Bonferroni correction for multiple testing ( Figure 1A ) . These are BIEC2-692674 at 29 . 804 Mb ( Praw = 2 . 88×10−7 ) and BIEC2-693138 at 32 . 19 Mb ( Praw = 1 . 08×10−6 ) . The associated region spanned 2 . 6 Mb from position 29 . 6 Mb to 32 . 2 Mb on ECA26 ( Figure 1B ) . In a subsequent fine-mapping phase , 62 additional SNPs within the region were genotyped on 13 FIS-affected samples . Several novel SNPs were identified ( dbSNP ss295469621-295469629 ) . In addition , two further microsatellites were also genotyped ( Table S1 ) . The homozygous affected haplotype was shared by these animals over a 992 kb segment ( Figure 2A ) . According to ENSEMBL gene prediction , fourteen genes lie in this interval ( Figure 2B ) . Five selected individual animals were re-sequenced over this critical region using sequence capture by NimbleGen arrays followed by GS FLX Titanium sequencing ( GenBank submission under study accession no . ERP000492 ) . The five individuals comprised one affected foal ( A13 on Figure 2 ) , the two obligate carrier parents , one apparently clear animal and one obligate carrier chosen for maximal homozygosity across the region . The FIS carrier status and familial relationships were confirmed for each individual by parentage verification . The last animal proved particularly useful in eliminating many potential causal variants . In total , eight verified SNPs were identified in the affected foal , narrowing the critical region to 375 , 063 bp ( ECA26: 30 , 372 , 557 – 30 , 747 , 620 bp ) . Coverage of this critical interval was increased from 92 . 9% to 98 . 4% using Sanger sequencing; none of the remaining gaps fell within 200 bp of protein-coding sequence . Only one variant , a SNP at 30 , 660 , 224 bp , segregated as expected for a causal recessive mutation within the five sequenced samples . In addition there was no evidence of DNA rearrangement , duplication or insertion/deletion seen in the affected foal ( Figure S4 ) . The segregating SNP was assessed for validity as an FIS marker in equine populations . Subsequently all 38 available affected foals ( 37 Fell , 1 Dales ) were shown to be homozygous for the affected allele and all 21 available obligate carriers were heterozygous . A selection of Fell and Dales samples which were submitted to the Animal Health Trust for parentage verification between 2000 and 2010 were anonymously screened for the affected allele: 82 / 214 ( 38% ) of the Fells and 16 / 87 ( 18% ) of the Dales were heterozygous for the lesion and no homozygous affecteds were discovered . These carrier rates are consistent with the approximate observed disease prevalence of 10% in the Fell and 1% in the Dales populations . In addition , a selection of horse breeds ( 184 individuals from 11 breeds consisting of Thoroughbred ( n = 29 ) , Appaloosa ( n = 8 ) , Arab ( n = 21 ) , Warmblood Sport Horse ( n = 17 ) , Lipizzaner ( n = 2 ) , Cleveland Bay ( n = 20 ) , Dartmoor Pony ( n = 19 ) , Icelandic Horse ( n = 8 ) , New Forest Pony ( n = 20 ) , Sheltand Pony ( n = 20 ) and Shire ( n = 20 ) ) which were considered unlikely to have interbred with either the Fell or Dales was genotyped and all proved homozygous wild-type .
The identification of a mutation that segregates 100% with the disease has enabled a diagnostic test to be developed and offered to breeders and owners , allowing them to avoid carrier-carrier matings , and consequently drastically reduce the numbers of FIS-affected foals born each year . A gradual reduction in the use of carrier animals will , over time , lead to a reduction in the affected allele frequency in the population , while conserving the gene pool as much as possible . In addition , other equine breeds that may have interbred with the Fell or Dales will now be screened for FIS carriers . The FIS-associated SNP falls within the single exon of the sodium/myo-inositol co-transporter gene ( SLC5A3 , also known as SMIT ) , which is a cell membrane transporter protein responsible for the co-transport of sodium ions and myo-inositol . This SNP is non-synonymous , causing a P446L substitution in SLC5A3; this amino acid residue ( equivalent residue 451 in the human protein ) is conserved in all 11 placental mammals for which high-coverage sequence is now available ( selection shown in Figure 3 ) . Similarly , this residue is conserved in other solute carrier family 5 ( SLC5 ) paralogs in the horse which share similar structural homology ( Figure 3 ) . The crystal structure of a bacterial homolog of SLC5A1 ( sodium/glucose co-transporter 1 ) has recently been elucidated [12] and shows this member of the SLC5 family to have 14 transmembrane helices; the structural conformations adopted during transport , and the precise positions of substrate binding during transfer are now being identified [13] . Alignment of the protein sequences of the SLC5 family suggests that P446 in equine SLC5A3 is located in a transmembrane helix which is involved in forming the substrate cavity [12] and which tilts during substrate transfer . The two prolines at positions 445 and 446 may be required for effective substrate binding by closing the substrate binding site after the substrate is bound . Proline residues introduce structural destabilisation into alpha helices and have long been obvious candidates for points of conformational change required for substrate binding and release [14] . Indeed , replacement of prolines in the transmembrane helices of transport proteins has shown that some residues are profoundly important in transport , affecting either substrate affinity or substrate movement [15] . SLC5A3 is an osmotic stress response gene , which acts to prevent dehydration caused by increased osmotic pressure in the extracellular environment . Dehydration causes the disruption of numerous cellular functions by denaturation of intracellular molecules and damage to sub-cellular architecture [16] . Extreme osmotic conditions are found in the kidney , although osmotic response mechanisms have also been found in numerous tissues , and in particular are critical for lymphocyte development and function [17] , [18] . The mechanism by which the osmotic stress response is mediated in mammals is not completely understood , but involves a signaling cascade comprising Rho-type small G-proteins , p38 Mitogen-activated protein kinase ( p38MAPK ) and the transcription factor , Nuclear Factor of Activated T-cells 5 ( NFAT5 ) [19] . NFAT5 directly stimulates the transcription of hyperosmolarity-responsive genes , of which SLC5A3 is one . These act to counterbalance the effects of extracellular osmotic pressure by transporting small organic osmolytes , such as myo-inositol , into the cell , thereby maintaining isotonicity with respect to extra-cellular conditions [20] . NFAT5 is the only known transcriptional activator of hyperosmolarity response genes , and was shown to be essential for normal lymphocyte proliferation and adaptive immunity [18] . Targeted knockout of NFAT5 in mice results in late gestational lethality whereas partial loss of function leads to defects in adaptive immunity and a substantially reduced spleen and thymus [18] . Furthermore , transgenic studies identify loss of T-cell mediated immunity as the prime deficiency ensuing from aberrant NFAT5 activity [21] , [22] . Similarly FIS-affected foals have markedly reduced thymus and spleen with a lack of germinal centers [1] , [23] . However , FIS disease immunopathology indicates that FIS foals have apparently normal circulating T-cell numbers with only peripheral blood B-lymphocyte numbers significantly depleted; currently , there are no data available indicating NFAT5 activity in specific B-lymphocyte functions . Studies are now required to demonstrate the functional differences between the Pro446 and Leu446 forms of the protein . This will be achieved by introducing this mutation into transgenic mice and assessing transport function . Further investigation into the physiological consequences of this mutation will then also be possible . In particular , it will be important to identify how the mutation leads to profound anemia and B-lymphopenia whilst neutrophils and T-cell numbers ( including CD4/CD8 ratio ) and function ( responses to mitogens PHA and Con A ) appear normal [4] . Importantly , it must be investigated whether there is a defect in T-cell function that is currently undetected or whether the antigen presenting function of B-cells is so suppressed that the T-cells cannot respond . In addition , it must be noted that the lymphoid organs in FIS foals have depleted thymus tissue and poor germinal centre development in spleen and lymph nodes , which suggests that there may be some unidentified T-cell dysfunction . Alternatively , any T-cell defects could be due to severe inflammatory responses in these very sick foals . SLC5A3 is not associated with any described mammalian disease , although a role in the pathogenicity of Down Syndrome is suggested [24] . The effect of loss of SLC5A3 activity has not been comprehensively studied , however SLC5A3 knockout mice die shortly after birth due to hypoventilation [25] , probably due to failure of the peripheral nervous system [26]; similarly FIS-affected foals have peripheral ganglionopathy [1] . There is relatively little literature on SLC5A3 function in hemopoietic or immunological tissues , in any species , although a role for osmotic control in developing cells is likely . Due to uncertainty regarding tissue distribution and function of SLC5A3 , it cannot be assumed that the profound anemia and severe loss of circulating B-lymphocytes in FIS is directly due to a functional change in SLC5A3 expression or function; formal proof that this is the case will entail functional studies . Whilst there is no doubt that the mutation in this gene is predictive of carrier or disease status , the mechanism by which this amino acid change could lead to the two described pathologies is speculative . It is , of course possible that the mutation site is close to another , as yet unidentified , mutation that is ultimately responsible for the severe hematological and immunological changes in homozygotes . However , all coding sequence within the critical region has been fully investigated and this is the only variant that segregates with the disease . We hypothesize that the phenotype seen in FIS-affected foals is either a result of partial loss or subtle alteration in SLC5A3 activity that has deleterious effects on B-lymphocyte and erythroid development but cannot discount the involvement of other genetic variants in the critical region . Whichever is the case , further analysis of FIS is justified as this genetically determined combination of immune phenotypes has not previously been reported in any other species .
Procedures were limited to the collection of blood by jugular venipuncture or hairs pulled from the mane or tail . Blood samples were taken as veterinary diagnostic procedures as all study animals were equine patients presenting with clinical signs suggestive of FIS or were healthy related or unrelated animals that were blood tested for anemia and/or B-lymphocyte deficiency . Many of the Fell and Dales ponies used in this study have been described previously [6] , [7] . Study animals were all equine patients presenting with clinical signs suggestive of FIS or were healthy related or unrelated Fell or Dales ponies that were blood tested for anemia and/or B-lymphocyte deficiency . Several FIS foals presented subsequent to euthanasia . Pedigree information was available for many of the Fell ponies ( Figure S1 ) , and these samples ( n = 41 ) were used in the linkage and homozygosity mapping analysis . An additional ten samples were added to these for the association study; these were isolated samples for which no pedigree information and/or samples from immediate family were available . Any adult Fell pony was eligible as a control for the association study . FIS diagnosis was based on breed , age of animal ( 4–8 weeks at presentation ) , and profound anemia with no other predisposing cause , and was confirmed on pathology . Specifically , this indicated severely reduced numbers ( or absence ) of germinal centers in spleen and regional lymph nodes . B-lymphocyte deficiency was also used for FIS diagnosis . Many accompanying clinical signs were also reported , primarily related to opportunistic infections , but these were not considered diagnostic alone . Blood samples were collected in EDTA collection tubes from all of the Fell pony individuals indicated in Figure S1 , and from a Dales foal and it's parents [6] . Genomic DNA was isolated from the samples using a Nucleon™ BACC Genomic DNA Extraction Kit . A panel of 228 markers , distributed as evenly as possible over the equine genome and described in Table S1 , was used . Two further markers , TKY1155 and TKY2012 , which were located in the implicated region , were subsequently genotyped . The genome scan was performed in multiplexes of three markers . Four PCR reactions , each utilising a different fluorescent dye , were pooled together post-PCR to form a panel of 12 markers for analysis . An 18 bp tail ( 5′-TGACCGGCAGCAAAATTG-3′ ) was added to the 5′ end of the forward primer and a complementary fluorescent labelling primer was included in the PCR reaction as a means of making the reactions more efficient and to reduce costs [27] . Amplification was performed in 6 µl volumes , using 2 . 5 pmol of reverse , 1 pmol of tailed-forward , 5 pmol of the labelled universal primer ( either 6-FAM , VIC , NED , or PET ) , 20 ng genomic DNA , 0 . 75 unit AmpliTaq Gold ( Applied Biosystems ) , 1× GeneAmp PCR buffer II ( Applied Biosystems ) , 1 . 5 mM MgCl2 , and 200 µM each dNTP . After denaturation at 94°C for 10 min , a 30-cycle PCR of 94°C for 1 min , 55°C for 1 min , and 72°C for 1 min , followed by 8 cycles of 94°C for 1 min , 50°C for 1 min , and 72°C for 1 min was performed , followed by a final extension at 72°C for 30 min . Genotyping analysis was performed on an ABI3100 ( Applied Biosystems ) according to the manufacturer's instructions . Genotyping data was analysed with GeneMapper version 4 . 0 ( Applied Biosystems ) ; alleles were assigned to pre-defined bins and automatically given an appropriate integer value . Mendelian inheritance was checked . We used 41 ponies ( 14 FIS-affected , 17 obligate carriers , 10 adults of unknown carrier status ) for which pedigree information and DNA was available , in the linkage analysis ( Figure S1 ) . A parametric linkage analysis was carried out using SUPERLINK v . 1 . 5 [28] assuming an autosomal recessive mode of inheritance . The disease allele frequency was estimated at 0 . 1 , with 100% penetrance . Pearson's chi2 test of independence was used to identify markers where homozygosity varied significantly between the cases and controls . An A×2 ( where A = number of alleles at a given locus ) contingency table with A-1 degrees of freedom was used . Expected and observed heterozygosity values were computed for cases and controls for all markers exhibiting a positive LOD score , using ARLEQUIN [29] . Statistical significance was assessed by calculating the one-tailed probability of the chi squared distribution . SNP genotyping on 51 genomic DNA samples was performed using standard manufacturer's protocols by Cambridge Genomic Services ( University of Cambridge , UK ) . The Illumina EquineSNP50 Infinium BeadChip , which contains 54 , 602 validated SNPs , was used; information on this array is available at http://www . illumina . com/documents/products/datasheets/datasheet_equine_snp50 . pdf . Quality control and genotype calling were performed using GenomeStudio v . 2009 . 2 ( Illumina Inc . ) . Samples with a call rate <95% were discarded ( n = 2 ) . We performed a basic case-control association analysis on the remaining 49 samples ( 18 affected and 31 controls ) . Analysis was performed with the software package PLINK [30] . SNPs with low minor allele frequency ( <0 . 02 ) or genotyping rate ( <90% ) were excluded; this left 42 , 536 SNPs for analysis . The presence of population stratification was assessed using multi-dimensional scaling ( Figure S2 ) and quantile-quantile plots were drawn to confirm that there was no over-inflation of the test statistics ( Figure S3 ) . A total of 62 polymorphic SNPs were studied in 13 affected individuals . A subset of these helped to delineate recombination breakpoints and these are identified in Figure 2 . Information regarding these SNPs can be found at http://www . broadinstitute . org/ftp/distribution/horse_snp_release/v2/equcab2 . 0_chr26_snps . xls . PCR amplification of the target sequence containing each informative SNP was performed in 12 µl volumes containing 20 ng genomic DNA , 0 . 75 unit AmpliTaq Gold , 1× GeneAmp PCR buffer II , 1 . 5 mM MgCl2 , 200 µM each dNTP , 10 pmol of reverse and of tailed-forward primer . A PCR program of 94°C for 10 min , followed by 30 cycles of 94°C for 1 min , 58°C for 1 min , and 72°C for 2 min , and then an extension of 72°C for 10 min was used . The PCR products were purified ( MultiScreen PCR96 filter plates; Millipore ) before sequencing in a 6 µl volume using 0 . 5 µl of 5× BigDye Terminator v3 . 1 ( Applied Biosystems ) , 5–20 ng PCR template , 1 µl of 1× BigDye sequencing buffer and 3 . 2 pmol universal sequencing primer ( Sigma-Aldrich ) . Templates >500 bp were also sequenced in the reverse direction . Sequencing was performed using cycle sequencing: 96°C for 0 . 5 min , 44 cycles of 92°C for 4 s , 58°C for 4 s and 72°C for 1 . 5 min . Purification was performed by isopropanol precipitation followed by sequencing on an ABI3100 according to the manufacturer's instructions . Sequences were viewed using STADEN [31] . This was performed at the Centre for Genomic Research ( University of Liverpool , UK ) . A region of 3 Mb ( ECA26: 28 , 942 , 655 – 31 , 942 , 655 Mb ) was selected for re-sequencing which encompassed the critical region . Custom tiling 385 k NimbleGen Sequence Capture arrays ( http://www . 454 . com/products-solutions/experimental-design-options/nimblegen-sequence-capture . asp ) which covered 92 . 9% of the target were designed from the horse reference sequence using standard repeat-masking algorithms . Five individuals were selected for re-sequencing consisting of one affected pony ( A13 in Figure 2 ) , its parents , one obligate carrier selected for maximal homozygosity over the region and one individual apparently homozygous wild-type . Sequencing was performed using GS FLX Titanium Series chemistry and assembled using Roche Newbler software v2 . 0 . 00 . An average 34-fold read depth was obtained . Sequence from each of the five sequenced animals was aligned to the EquCab2 reference sequence using the Artemis Comparison Tool ( ACT ) [32] to identify possible rearrangements or insertion/deletions ( Figure S4 ) . MySQL ( Oracle Corporation ) was used to interrogate the data . The critical region was narrowed using heterozygous variants in the affected foal and Sanger sequencing subsequently verified these . The narrowed critical region was then interrogated for variants that segregated as expected for a recessive mutation; putative causal variants were confirmed or disproved using Sanger sequencing .
|
Foal Immunodeficiency Syndrome ( FIS ) is a genetic disease that affects two related British pony breeds , namely the Fell and the Dales . Foals with FIS appear to be normal at birth but within a few weeks develop evidence of infection such as diarrhoea , pneumonia , etc . The infections are resistant to treatment , and the foals die or are euthanized before three months of age . The foals also suffer from a severe progressive anemia . Being a recessive condition , the disease is difficult to control without a diagnostic DNA test to identify symptom-free carrier parents . Within the last few years the horse genome has been sequenced , and this has allowed the development of tools to identify genetic mutations in the horse at high resolution . In this article we demonstrate the use of these new tools to identify the location of the FIS mutation . The presumptive causal lesion was then identified by sequencing this region . This has enabled us to develop a test that can be used to identify carrier ponies , allowing breeders to avoid FIS in their foal crop .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"animal",
"types",
"animal",
"genetics",
"immunology",
"lymphoid",
"organs",
"veterinary",
"diagnostics",
"veterinary",
"science",
"veterinary",
"medicine",
"biology",
"immune",
"system",
"veterinary",
"immunology",
"genetics",
"of",
"the",
"immune",
"system",
"large",
"animals",
"bone",
"marrow",
"genetics",
"genetics",
"of",
"disease",
"genetics",
"and",
"genomics",
"immunoglobulins"
] |
2011
|
Identification of a Mutation Associated with Fatal Foal Immunodeficiency Syndrome in the Fell and Dales Pony
|
Reversible infantile liver failure ( RILF ) is a unique heritable liver disease characterized by acute liver failure followed by spontaneous recovery at an early stage of life . Genetic mutations in MTU1 have been identified in RILF patients . MTU1 is a mitochondrial enzyme that catalyzes the 2-thiolation of 5-taurinomethyl-2-thiouridine ( τm5s2U ) found in the anticodon of a subset of mitochondrial tRNAs ( mt-tRNAs ) . Although the genetic basis of RILF is clear , the molecular mechanism that drives the pathogenesis remains elusive . We here generated liver-specific knockout of Mtu1 ( Mtu1LKO ) mice , which exhibited symptoms of liver injury characterized by hepatic inflammation and elevated levels of plasma lactate and AST . Mechanistically , Mtu1 deficiency resulted in a loss of 2-thiolation in mt-tRNAs , which led to a marked impairment of mitochondrial translation . Consequently , Mtu1LKO mice exhibited severe disruption of mitochondrial membrane integrity and a broad decrease in respiratory complex activities in the hepatocytes . Interestingly , mitochondrial dysfunction induced signaling pathways related to mitochondrial proliferation and the suppression of oxidative stress . The present study demonstrates that Mtu1-dependent 2-thiolation of mt-tRNA is indispensable for mitochondrial translation and that Mtu1 deficiency is a primary cause of RILF . In addition , Mtu1 deficiency is associated with multiple cytoprotective pathways that might prevent catastrophic liver failure and assist in the recovery from liver injury .
Transfer RNA ( tRNA ) is an adaptor molecule that converts genetic information into an amino acid sequence in protein synthesis . tRNAs contain a wide variety of modified nucleosides that are introduced post-transcriptionally [1 , 2] . In mammalian mitochondria , 22 subtypes of tRNAs encoded in mitochondrial DNA participate in the translation of 13 protein subunits of respiratory chain complexes in mitochondria . Fifteen species of modified nucleotides are found at 118 positions of bovine mitochondrial tRNAs ( mt-tRNAs ) [3] . A number of pathogenic point mutations associated with mitochondrial diseases are found in mt-tRNA genes [4–5] . Some of these mutations impair mt-tRNA modifications , leading to defective translation and mitochondrial dysfunction . In addition , a number of pathogenic mutations have been found in mt-tRNA-modifying enzymes , including MTO1 [6–8] , GTPBP3 [9] , MTU1 [10–13] , and PUS1 [14] , indicating that decreased modifications of mt-tRNAs caused by tRNA mutations and defective tRNA-modifying enzymes result in pathological consequences . Supporting these findings , the physiological roles of tRNA-modifying enzymes have been extensively studied in mouse models lacking mt-tRNA-modifying enzymes [15–16] . Unique to mitochondrial tRNA modifications , 5-taurinomethyluridine ( τm5U ) is present at the first position of the anticodon ( i . e . , the “wobble position” or position 34 ) of mt-tRNAs for Leu ( UUR ) and Trp , whereas its 2-thiouridine derivative ( τm5s2U ) is found at the same position of mt-tRNAs for Glu , Gln and Lys [17–18] . These modifications allow mt-tRNAs to recognize their cognate codons precisely and ensure accurate translation in the mitochondria . The enzymes mitochondrial tRNA translation optimization 1 ( Mto1 ) and GTP binding protein 3 ( Gtpbp3 ) are assumed to be responsible for τm5U formation [19] . In addition , mitochondrial tRNA-specific 2-thiouridylase 1 ( MTU1 ) catalyzes the 2-thiolation of τm5U to form τm5s2U [19] . The lack of a yeast homolog of MTU1 resulted in impaired mitochondrial translation activity and a severe respiratory defect [19] . Moreover , acute knockdown of MTU1 in HeLa cells reduced the oxygen consumption rate and resulted in a defective mitochondrial membrane potential [19] . Intriguingly , MTU1 has been implicated in the pathogenesis of reversible infantile liver failure ( RILF ) [10–13] , a life-threatening condition characterized by acute liver dysfunction during the first 2–4 months after birth . However , a majority of patients spontaneously recover and never exhibit another symptom [20] . Genomic analysis has identified a variety of autosomal recessive mutations , including substitutions , insertions and deletions , in the coding region of MTU1 in RILF patients [20] . These mutations in MTU1 are predicted to cause the loss of MTU1 activity , which subsequently triggers the pathological symptoms . Indeed , immortalized cell lines derived from RILF patients exhibit very low levels of MTU1 with a marked reduction of 2-thiolation levels in mt-tRNAs , leading to defective mitochondrial translation [21] . However , little reduction of mitochondrial translation has been observed in the fibroblasts of RILF patients [22] , which raised some discrepancies by different groups regarding the observed MTU1 functions . To reveal the physiological role of MTU1-mediated 2-thiouridine formation of τm5s2U in the regulation of mitochondrial protein translation in hepatocytes and its relevance in RILF , we generated multiple lines of conditional Mtu1 knockout mice and investigated the molecular functions of Mtu1 in these murine models .
To investigate the physiological role of Mtu1 , we generated constitutive Mtu1 knockout mice ( Mtu1-/- ) ( Fig 1A ) . However , no Mtu1-/- mice were obtained after multiple generations of breeding ( Fig 1B ) . The embryos at 9 days post-coitum ( E9 ) were isolated for morphological examination . While the morphology of Mtu1+/- embryos did not differ from that of Mtu1+/+ embryos , the size of Mtu1-/- embryos was strikingly small ( Fig 1C and 1D ) . To visualize the internal structure , the embryos were stained with platelet endothelial cell adhesion molecule-1 ( PECAM-1 ) . Mtu1+/+ and Mtu1+/- embryos exhibited organized blood vessel networks that surrounded well-developing tissues , such as the heart , brain and spinal cord ( Fig 1E ) . In contrast , neither blood vessel networks nor developmental stage-matched tissues were observed in Mtu1-/- embryos ( Fig 1E ) . Considering these morphologies , we concluded that the Mtu1-/- embryo died at a developmental stage as early as E7 . 5~8 . To avoid the embryonic lethality , we generated liver-specific Mtu1 knockout mice ( Mtu1LKO ) ( S1 Fig ) . The Mtu1LKO mice were viable and developed without any obvious morphological defects ( S1 Fig ) . We confirmed a 10-fold reduction in the Mtu1 transcript levels in the liver of Mtu1LKO mice compared to Mtu1Flox mice ( Fig 2A ) . The average body weight of Mtu1LKO mice did not differ from that of Mtu1Flox mice ( Fig 2B ) . Laboratory examinations of 3-week-old mice revealed significantly elevated plasma levels of lactate , aspartate aminotransferase ( AST ) and alanine aminotransferase ( ALT ) in Mtu1LKO mice ( Fig 2C–2E ) . In addition , the Mtu1LKO mice exhibited a high level of serum lactate dehydrogenase ( LDH ) and a low level of albumin ( ALB ) compared to the Mtu1Flox mice ( Table 1 ) . These results clearly indicate that liver injury occurs in the Mtu1LKO mice . Interestingly , the Mtu1LKO mice exhibited altered metabolism . The levels of total cholesterol ( T-CHO ) , high-density lipoprotein cholesterol ( HDL-C ) , amylase ( AMY ) , creatinine ( CRE ) and calcium ( Ca ) were significantly reduced in the Mtu1LKO mice ( Table 1 ) . In agreement with these biochemical results , a gene expression analysis revealed an increase in the expression levels of genes involved in glycolysis ( Glucokinase , Gck ) , gluconeogenesis ( Glucose-6-phosphatase catalytic subunit , G6pc ) , and lipid oxidation ( carnitine palmitoyltransferase 2 , Cpt2 ) in the Mtu1LKO mice ( Fig 2F ) . We next investigated liver injury in the Mtu1LKO mice at the histological level . The Mtu1LKO mice exhibited normal liver structures compared to Mtu1Flox mice ( S1 Fig ) . No obvious fibrosis was observed in the livers of Mtu1LKO mice by Masson trichrome staining ( Fig 3A and 3B ) . Interestingly , there was an infiltration of macrophages ( arrows in Fig 3C and 3D ) and spotty necrosis ( arrows in Fig 3E ) in the livers of Mtu1LKO mice . In addition , enlarged hepatocytes with karyomegaly and multiple nuclei were observed in the Mtu1LKO mice ( arrows in Fig 3F ) . These histological features of Mtu1LKO mice resemble the clinical features of RILF patients [20] . To clarify the molecular function of Mtu1 and its association with liver injury , we examined mt-tRNA modifications and mitochondrial translation in Mtu1-deficient hepatocytes . mt-tRNAGln , mt-tRNAGlu and mt-tRNALys were individually purified from liver tissues and subjected to mass spectrometry analysis . Interestingly , these mt-tRNAs were partially s2-modified even in the livers of Mtu1Flox mice; 40~70% of the mt-tRNAGlu , mt-tRNAGln and mt-tRNALys contained τm5s2U and s2U modifications , whereas the remaining 30~60% of the mt-tRNA contained τm5U modifications and unmodified U ( S2 Fig ) . In the livers of Mtu1LKO mice , the s2 modification was nearly absent in the three mt-tRNAs ( Fig 4A–4C ) . There was a trace of τm5s2-containing mt-tRNAs , but this result was most likely derived from non-hepatic cells . Notably , the level of τm5U in mt-tRNAGlu , mt-tRNAGln and mt-tRNALys remained unchanged despite of the loss of the s2 modification ( S2 Fig ) . In addition to mitochondrial s2 modification , we examined other sulfur-containing modifications , including mcm5s2U ( 5-methoxycarbonylmethyl-2-thiouridine ) modification in cytosolic tRNAs and ms2i6A ( 2-methylthio-N6-isopentenyladenosine ) modification in mt-tRNAs . Interestingly , the abundance of these modifications was up-regulated in the livers of Mtu1KO mice compared to Mtu1Flox mice ( S3 Fig ) . Next , we examined mitochondrial protein translation in primary hepatocytes isolated from Mtu1LKO and Mtu1Flox mice . The level of mitochondrial translation was markedly reduced in Mtu1-deficient hepatocytes compared with control cells ( Fig 5A ) . Interestingly , the degree of impairment in 13 mitochondrial proteins depended on their molecular weights . The translation of 9 mitochondrial proteins with molecular weights higher than 25 kDa ( equal to the molecular weight of COII/III ) was markedly impaired in Mtu1-deficient cells ( Fig 5A ) . On the other hand , the translation of the remaining 4 mitochondrial proteins ( ND6 , ND3 , ND4L and A8 ) with molecular weights lower than 25 kDa was only slightly increased in Mtu1-deficient cells ( Fig 5A ) . The impairment in mitochondrial translation resulted in a decrease in the steady-state level of mitochondrial proteins ( Fig 5B ) . Notably , the protein levels of NDUFB8 and MTCOI , which comprise Complexes I and IV , respectively , showed a marked reduction in the livers of Mtu1LKO mice ( Fig 5B ) . Accordingly , the formation of respiratory Complexes I~IV was impaired in Mtu1LKO mice ( Fig 5C and 5D ) . The disruption of mitochondrial translation consequently resulted in a broad and significant decrease in the activities of Complexes I , III and IV ( Complex I: 70% , Complex III: 67% , Complex IV: 52% versus Flox mice , Fig 5E ) . There was also a significant increase of citrate synthase activity in the livers of Mtu1LKO mice . These results clearly demonstrate that Mtu1 is indispensable for mitochondrial translation and respiratory activities . Because mitochondrial dynamics are closely coupled with mitochondrial proteostasis [23] , we examined mitochondrial morphology using electron microscopy . Striking mitochondrial enlargement and proliferation were observed in Mtu1-deficient hepatocytes ( Fig 6A ) . The average mitochondrial area in Mtu1-deficient hepatocytes was 4 . 3-fold larger than that in control hepatocytes ( Fig 6B ) . Notably , nearly all mitochondria in Mtu1-deficient hepatocytes exhibited aberrant cristae structures . The cristae were either abnormally swollen or lost in most of the mitochondria ( Fig 6C ) . Some mitochondria contained an inner vacuole with multiple layers of membrane , and some exhibited very low electron density ( Fig 6C ) . Disruption of mitochondrial function has been associated with the activation of the pathway related to mitochondrial proteostasis [24] . In agreement with the previous findings , there was a substantial increase in proteins related to proteostasis , such as LONP1 , AFG3L2 , DRP1 , MFN1 and PARKIN ( Fig 6D ) . Taken together , these results demonstrate that Mtu1-mediated mitochondrial translation is required for the maintenance of mitochondrial structures and proteostasis . The Mtu1LKO mice showed severe mitochondrial dysfunction in hepatocytes but maintained liver function . We investigated the molecular mechanism that enables Mtu1LKO mice to tolerate severe mitochondrial dysfunction in hepatocytes . Mitochondrial dysfunction often induces an increase in hepatic Fgf21 levels , which has been associated with compensatory signaling , such as mitochondrial biogenesis [25 , 26] . There was an approximately 32-fold increase in the Fgf21 levels in the livers of Mtu1LKO mice ( Fig 7A ) . Accordingly , peroxisome proliferator-activated receptor gamma coactivator 1-alpha ( Pgc1α ) , an upstream regulator of mitochondrial proliferation , was significantly up-regulated ( Fig 7A ) . Accordingly , there was an approximately 3-fold increase of the mtDNA copy number , which was associated with an approximately 4-fold increase in the mtDNA-encoded mitochondrial genes in Mtu1-deficient hepatocytes ( Fig 7B–7C ) . In addition , MTOR and ERK1/2 , which are effectors of FGF21 [27–28] , were also up-regulated at both the total protein level and phosphorylation level in Mtu1-deficient primary hepatocytes ( Fig 7D ) . Mitochondria are the major source for the production of reactive oxygen species , and its dysfunction has been associated with oxidative stress [29] . We examined how oxidative stress was managed in Mtu1-deficient hepatocytes . Surprisingly , there was a decrease in mitochondrial protein carbonylation , which is a byproduct of oxidative stress , in the livers of Mtu1LKO mice ( Fig 7E ) . In addition , we examined the levels of oxidative stress-related metabolites in the liver tissues ( S4 Fig ) . The glutathione ( GSH ) and cysteine levels in Mtu1LKO mice did not differ from those in Mtu1Flox mice . Interestingly , the level of glutathione disulfide ( GSSG ) , a marker of oxidative stress , trended toward a decrease in the livers of Mtu1LKO mice ( S4 Fig ) . These results prompted us to examine the gene profiles that are related to oxidative stress in liver tissues . These genes were dynamically changed in the livers of Mtu1LKO mice , indicating an active response to oxidative stress in the Mtu1LKO mice ( Fig 7F ) . Importantly , antioxidant genes , such as glutathione synthetase ( Gss ) , glutathione peroxidase ( Gpx2/3 ) and sulfiredoxin 1 ( Srxn1 ) , were significantly up-regulated . Together , these results show that the increase in mitochondrial biogenesis in combination with the adaptive scavenging of oxidative stress might compensate for the mitochondrial dysfunction and prevent catastrophic liver failure in Mtu1LKO mice . Finally , to evaluate the progression of liver injury in Mtu1LKO mice , we examined the histological , biochemical and genetic features of the livers at 16 weeks . The Mtu1LKO mice were alive and exhibited sustained liver function . Serum ALT and AST levels in 16-week-old Mtu1LKO mice , which were at the same levels as those of 3-week-old Mtu1LKO mice , were still higher than the levels in 16-week-old Mtu1Flox mice ( Fig 8A ) . Unlike the 3-week-old mice , the serum LDH levels in the 16-week-old Mtu1LKO mice did not differ from those of the Mtu1Flox mice ( Fig 8A ) . Similar to the histological features of 3-week-old mice , 16-week-old Mtu1LKO mice also exhibited enlarged hepatocytes with karyomegaly and spotty necrosis; however , no obvious fibrosis was observed ( Fig 8B ) . Biochemical examination of mtDNA-derived MTCOI in total tissue lysates and in purified respiratory complexes revealed that the protein levels of MTCOI remained at low levels in 16-week-old Mtu1LKO mice ( Fig 8C–8D ) . Nevertheless , it is worthwhile to indicate that MTCOI levels appeared to be slightly increased compared to 3-week-old Mtu1LKO mice ( Fig 5D ) . At the gene expression level , Mtu1LKO mice exhibited a marked up-regulation of Fgf21 at 16 weeks ( Fig 8E ) . Genes related to mitochondrial proliferation and suppression of oxidative stress , such as Pgc1a , Gpx3 and Srxn1 , also remained up-regulated in Mtu1LKO mice compared to age-matched Mtu1flox mice ( Fig 8E ) . Accordingly , the relative mt-DNA copy number was up-regulated in 16-week-old Mtu1LKO mice ( Fig 8F ) . Taken together , these results suggest that despite the lasting mitochondrial dysfunction from embryo to adulthood , Mtu1LKO mice were able to adapt to the liver injury and even exhibited some features of recovery in adulthood .
In the present study , we established hepatocyte-specific Mtu1-deficient mice that manifest the clinical symptoms of RILF . Mtu1 deficiency in hepatocytes resulted in a marked reduction in mitochondrial translation . The impairment in mitochondrial translation subsequently caused a broad decrease in respiratory activities and led to the disruption of membrane integrity . The mitochondrial dysfunction consequently induced liver injury . Our murine model provides mechanistic insights into the pathogenesis of RILF . Mutations in TRMU ( MTU1 ) have been implicated in the pathogenesis of RILF [20] . To date , 20 RILF patients carrying pathogenic mutations in the coding region of MTU1 have been reported [20] . In contrast to the genetic evidence , the mechanism of RILF remains controversial . Some studies have reported that Mtu1 deficiency alone does not always cause the decrease in mitochondrial translation in RILF patient-derived cell lines [21–22] . In contrast , we showed that Mtu1 deficiency resulted in a marked impairment in mitochondrial translation and respiration in yeast and human cancer cell lines [19] . This discrepancy was likely caused by different experimental conditions in these studies . Indeed , all these studies were performed in cells that originated from different species . Notably , none of the previous studies were performed in cells related to the liver , which is the symptomatic tissue in RILF . In the present study , we generated liver-specific Mtu1 knockout mice , which allow us to investigate the role of Mtu1 in hepatocytes for the first time . Consistent with our previous results , Mtu1 deficiency markedly inhibited mitochondrial translation in hepatocytes , resulting in a substantial decrease in respiratory complexes and activities . Thus , our results lead to a clear conclusion that Mtu1 is required for mitochondrial translation in hepatocytes . Why is Mtu1-mediated s2 modification crucial for mitochondrial translation ? A previous study showed that bacterial tRNAGln lacking the s2 modification was still capable of translating the GAA codon , but the decoding efficiency of s2-deficient tRNAGln was 4-fold lower than that of a fully modified tRNAGln [30] . Therefore , it is conceivable that the major role of the s2 modification at 34U is to facilitate base-pairing with A to increase decoding efficiency . In mice and humans , mitochondrial genes predominantly utilize the GAA , CAA and AAA codons , which are decoded by mt-tRNAGlu , mt-tRNAGln and mt-tRNALys , respectively ( S5 Fig ) . It is likely that mitochondria perverted the s2 modification during evolution for the optimal translation of biased codons . In addition , the codon usage might also influence the translation efficiency . In fact , the abundance of GAA , CAA and AAA codons is proportional to the length of mRNA , whereas the frequency of these codons did not differ among the mitochondrial genes ( S5 Fig ) . Therefore , s2 modification would be particularly required for the efficient translation of long mRNAs because of their high demand for their corresponding tRNAs . Despite the important role of s2 modification in regulating efficient translation , the s2 modification level only ranges from 40%~70% at steady state . It is likely that GAA , CAA and AAA codons would be preferentially and efficiently decoded by fully modified mt-tRNAs in wild-type mitochondria . However , the unmodified mt-tRNAs might still be partially functional because mitochondrial translation was still detectable in Mtu1-deficient cells . Supporting this speculation , a complete loss of the mcm5s2 modification , the cytosolic counterpart of τm5s2 , in a subset of yeast cytosolic tRNAs resulted in a lethal phenotype that was rescued by overexpressing unmodified tRNALys ( UUU ) [31] . The Mtu1LKO mice showed hepatic inflammation , necrosis , and elevated plasma levels of lactate and AST . These pathological features closely mimic the clinical symptoms of RILF [20] . However , compared with RILF patients , the phenotypes of Mtu1LKO mice were rather moderate . Notably , the Mtu1LKO mice did not exhibit hepatic fibrosis and nodule formation , which are frequently observed in RILF patients [20] . The genetic background of the Mtu1LKO mice might explain the mild hepatic injury . There is substantial evidence to suggest that mice with the C57BL6 background are resistant to hepatic fibrosis [32–33] . Despite this resistance , the Mtu1LKO mice still exhibited symptoms of liver injury , which further emphasizes the critical role of Mtu1-mediated mitochondrial translation in the development of RILF . While the hepatocyte-specific Mtu1 knockout mice were viable , the constitutive knockout mice were embryonic lethal at a very early developmental stage . In agreement with our results , the constitutive deficiency of murine Dars2 ( mitochondrial aspartyl-tRNA synthetase ) , which is a component of the mitochondrial translation machinery , is also embryonic lethal at a very early developmental stage [26] . These results suggest that efficient mitochondrial translation is indispensable for embryonic development . Importantly , lethality due to Mtu1 deficiency has been observed in RILF patients . Indeed , 6 of 20 patients died of acute liver failure between 1 and 8 months of age [20] . Three of the 6 patients carried homozygous mutations in either the translational start codon ( Met1Lys ) or the active site ( Asn96Ser ) . These mutations are predicted to cause a complete loss of Mtu1 or its enzymatic activity in the patients . In contrast , many of the surviving patients carried mutations in non-essential domains , which presumably cause a partial inhibition of Mtu1 activity . Taken together , these results suggest that the clinical progression of RILF might depend on the level of remaining Mtu1 activity in the tissues . Mtu1-mediated s2 modification requires a complicated enzymatic reaction that transfers the sulfur atom from cysteine to tRNAs [34] . Because cysteine metabolism is limited during neonatal development , it is proposed that cysteine availability might contribute to the development of liver failure in patients carrying pathogenic mutations [11 , 21] However , the cysteine levels in liver tissues of Mtu1LKO mice did not differ from those of Mtu1Flox mice . In addition , sulfur-containing tRNA modifications , including cytosolic mcm5s2U modifications and mitochondrial ms2i6A modifications , remained intact in Mtu1-deficient cells . Our results suggest that cysteine availability is likely not involved in the pathological phenotypes in our mouse model . The reversibility is the most interesting feature of RILF . After surviving the acute phase , the patients spontaneously recover without recurrence due to unknown mechanisms [20] . Mitochondrial biogenesis is a potential compensatory effect for mitochondrial dysfunction [26] . Indeed , we observed marked mitochondrial biogenesis in response to Mtu1 deficiency . Mitochondrial biogenesis is most likely activated by up-regulation of Pgc1α signaling . This result was consistent with the observation in Dars2 knockout mice , which also showed up-regulation of Pgc1α expression and mitochondrial biogenesis [26] . In addition to the mitochondrial effect , our study revealed a unique compensatory effect involving suppression of oxidative stress . Although mitochondrial dysfunction is usually associated with the generation of oxidative stress [29] , there was a moderate reduction in the stress levels and a marked increase in antioxidant gene expression in Mtu1-deficient hepatocytes . It is likely that the Mtu1 deficiency triggers oxidative stress due to severe mitochondrial dysfunction , but the adaptive scavenging response is strong enough to suppress the stress to a rather low level . Intriguingly , up-regulation of antioxidant genes in Mtu1LKO mice has been observed from the adolescent stage to the adult stage . The continuous activation of mitochondrial biogenesis and the suppression of oxidative stress might protect Mtu1LKO mice from catastrophic liver failure and maintain liver function in a tolerable condition from embryo to adulthood . From this perspective , treating RIFL patients with either cysteine or N-acetylcysteine might suppress oxidative stress and assist an early recovery . In addition to RILF , the Mtu1-mediated s2 modification has also been implicated in reversible infantile respiratory chain deficiency ( RIRCD ) [20] . RIRCD patients exhibit severe myopathy in the first months of life , followed by spontaneous recovery with some mild residual myopathy . The molecular mechanism underlying RIRCD is unknown; however , genetic analysis has revealed a single homoplasmic m . 14674T>C mutation in mitochondrial DNA that corresponds to mt-tRNAGlu [35–37] . Interestingly , there was a decrease in the s2 modification in mt-tRNAGlu in muscle samples with an m . 14674T>C mutation [21] . Similar to RILF , Boczonadi et al . reported that the decrease in s2 modification was not associated with the impairment of mitochondrial translation in fibroblasts and myoblast cells established from RIRCD patients [21] . Given the different regulatory mechanisms in immortal cells and intact tissues , it is likely that the m . 14674T>C mutation might also affect mitochondrial translation in the tissues of RIRCD patients . Further study using muscle-specific Mtu1-deificient mice may shed light on the molecular mechanism of RIRCD . In summary , we demonstrated that Mtu1-mediated s2 modification of mt-tRNA is indispensable for efficient mitochondrial translation and activities . Our study suggests that mitochondrial dysfunction due to Mtu1 deficiency is the primary cause of RILF . Our murine model is a valuable tool for understanding the molecular mechanism of RILF and for developing effective treatments .
Constitutive Mtu1 knockout mice were generated by crossing transgenic mice harboring exon 2 of the Mtu1 gene floxed by LoxP sequences ( Mtu1f/f mice ) with transgenic mice expressing Cre recombinase under the control of the CAG promoter ( CAGCre mice ) . Liver-specific Mtu1 KO mice were generated by crossing Mtu1f/f mice with transgenic mice carrying Cre recombinase under the control of the albumin promoter ( AlbCre mice ) . Mtu1f/f mice , CAGCre transgenic mice and AlbCre transgenic mice were backcrossed with C57BL6/J mice for at least seven generations to control the genetic background . Mice were housed at 25°C with 12 h light and 12 h dark cycles . Unless otherwise indicated , we sacrificed 3- to 5-week-old male mice for all experiments in Figs 1–7 . All animal procedures were approved by the Animal Ethics Committee of Kumamoto University ( Approval ID: A27-037 ) . Detailed information on genotyping can be found in the “Supplemental Methods” . Primary hepatocytes were isolated from Mtu1f/f and Mtu1LKO mice by perfusion of collagenase ( Worthington Biochemical Corporation , Lakewood , NJ ) following the manufacturer’s instructions . Isolated hepatocytes were cultured in high glucose DMEM ( Thermo Fisher Scientific , Waltham , MA ) supplemented with 10% fetal bovine serum ( Hyclone , GE Healthcare , NJ ) for 3 h . Subsequently , the culture medium was replaced with DMEM without serum for 14 h . All experiments using primary hepatocytes were performed within 24 h after isolation . Total RNA was isolated from fresh liver samples using TRIzol reagent ( ThermoFisher Scientific , USA ) following the manufacturer’s instructions . cDNA was synthesized from 100 ng of total RNA using the PrimerScript RT-PCR kit ( TAKARA , Tokyo , Japan ) and subjected to quantitative PCR ( SYBR Premix Ex Taq II , TAKARA ) using a 7300 Real Time PCR System ( Thermo Fisher Scientific , CA ) . The sequence information is included in the Supplemental Methods . The mitochondrial fraction was isolated from the livers of conditional Mtu1 knockout mice or HeLa cells using MOPS buffer as described previously . Ten micrograms of each sample were loaded into 12% SDS-PAGE gels and transferred to PVDF membranes . Mitochondrial and cellular proteins were detected using the proper antibodies as described in the supporting information . The data were analyzed using GraphPad Prism 6 software . An unpaired Student’s t-test was used to test the differences between two groups . A 2-tailed P-value of 0 . 05 was considered significant . The results are shown as the means ± S . E . M . Detail methods are provided in the “Supporting Information” .
|
Mitochondrial transfer tRNA ( mt-tRNA ) contains a variety of chemical modifications that are introduced post-transcriptionally . Three mt-tRNAs for Lys , Gln and Glu contain 5-taurinomethyl-2-thiouridine ( τm5s2U ) in their anticodons . It is known that the loss of 2-thiolation of τm5s2U is strongly associated with the development of reversible infantile liver failure ( RILF ) because pathogenic mutations of RILF were found in the MTU1 gene , which encodes an enzyme responsible for the 2-thiolation of τm5s2U . However , the molecular mechanism underlying RILF pathogenesis associated with a lack of MTU1 remains elusive . To understand the physiological function of MTU1 and its association with liver failure , we generated liver-specific Mtu1-deficient ( Mtu1LKO ) mice . Mtu1 deficiency abolished 2-thiouridine formation in the three mt-tRNAs . Loss of the 2-thiouridine modification resulted in a marked impairment of mitochondrial translation and abnormal mitochondrial structure . Consequently , the Mtu1LKO mice exhibited liver injury , which resembles the symptoms of RILF patients . Furthermore , mitochondrial dysfunction in Mtu1LKO mice induced mitochondrial biogenesis and suppressed oxidative stress . These findings elucidate the cellular and physiological functions of Mtu1 and provide a mouse model for understanding RILF pathogenesis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"transfer",
"rna",
"medicine",
"and",
"health",
"sciences",
"liver",
"oxidative",
"stress",
"animal",
"models",
"developmental",
"biology",
"model",
"organisms",
"mitochondria",
"bioenergetics",
"embryos",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"embryology",
"animal",
"cells",
"gene",
"expression",
"hepatocytes",
"mouse",
"models",
"biochemistry",
"rna",
"cell",
"biology",
"anatomy",
"nucleic",
"acids",
"protein",
"translation",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"biosynthesis",
"energy-producing",
"organelles",
"non-coding",
"rna"
] |
2016
|
Mtu1-Mediated Thiouridine Formation of Mitochondrial tRNAs Is Required for Mitochondrial Translation and Is Involved in Reversible Infantile Liver Injury
|
The mechanism underlying immune system recognition of different types of pathogens has been extensively studied over the past few decades; however , the mechanism by which healthy self-tissue evades an attack by its own immune system is less well-understood . Here , we established an autoimmune model of melanotic mass formation in Drosophila by genetically disrupting the basement membrane . We found that the basement membrane endows otherwise susceptible target tissues with self-tolerance that prevents autoimmunity , and further demonstrated that laminin is a key component for both structural maintenance and the self-tolerance checkpoint function of the basement membrane . Moreover , we found that cell integrity , as determined by cell-cell interaction and apicobasal polarity , functions as a second discrete checkpoint . Target tissues became vulnerable to blood cell encapsulation and subsequent melanization only after loss of both the basement membrane and cell integrity .
The discovery of Toll-like receptors and other categories of pattern recognition receptors has greatly enhanced our understanding of how the immune system recognizes different types of pathogens [1] , [2]; however , it is less clear why the immune system often turns its arsenal toward self-tissues . In fact , the same receptors that were originally found to bind specific types of pathogens are often involved in autoimmune diseases , making this issue more puzzling [3] . To understand this process , it is imperative to molecularly define the notion of the “immunological self” . Autoimmune-like responses are also observed in invertebrates . In Drosophila , degenerating internal tissues are subjected to hemocyte ( insect blood cell ) encapsulation , in which large , flat lamellocytes wrap up the target tissues in layers and melanize them via the phenoloxidase cascade . This process is called melanotic mass formation [4]–[6] . The same process occurs as part of the immunological defense against oversized pathogenic invaders , such as parasitoid wasp eggs , which are too large to be engulfed by the most abundant phagocytic hemocytes , the plasmatocytes [7] , [8] . Currently , about 100 genes have been found to display melanotic masses upon mutation or overexpression ( see FlyBase . org ) . These genes are seemingly unrelated , and the specific triggers of this autoimmune-like reaction are largely obscure . More than 30 years ago , Rizki and Rizki reported that the basement membrane ( BM ) appeared to serve as a barrier against hemocyte attack of self-tissues in Drosophila [9]–[11] . Whereas a same-species implant with the intact BM remained in the host , implants that had been mechanically damaged or pre-treated with collagenase to disrupt the BM triggered lamellocyte encapsulation [11] . Moreover , undamaged implants from sibling species did not induce lamellocyte encapsulation , whereas undamaged implants from distantly related species did , suggesting that hemocytes may recognize the molecular architecture of the BM of its own species . This interesting study raises several important questions as to which molecular component of the BM is crucial for blocking melanotic mass formation of self-tissues , and whether the BM is the sole surface feature for self-tolerance . Furthermore , their experiments were carried out in a sensitized genetic background , tu ( 1 ) Szts , in which the hemocytes were marginally hyperactive at a permissive temperature , which made it unclear whether the mass formation is caused primarily by defects in immune cells or target tissues . These questions have never been probed with genetic tools , largely due to the essential nature of the genes encoding the BM components collagen IV and laminin [12]–[16] . More recently , BM disruption was shown to act as a signal to recruit hemocytes to wound regions or to metastasizing tumors , providing further evidence for the BM-hemocyte relationship [17]–[19] . The BM is located on the basal side of epithelial tissues and serves multiple functions as a cell-supporting matrix , a tissue barrier , and ligands for cell surface receptors [20] , [21] . The composition of the BM varies between tissue types , but in general , the BM contains the following four major components: collagen IV and laminin , which together form a meshwork , and the proteoglycan Perlecan and Nidogen , which function in the scaffold . The BM is maintained by evolutionarily conserved cell surface receptors , such as integrin and dystroglycan [22] . Drosophila melanogaster has two collagen IV genes , Cg25C ( for α1 chain ) and vkg ( α2 ) , and four laminin genes , wb ( for laminin α1 , 2 ) , LanA ( α3 , 5 ) , LanB1 ( β ) , and LanB2 ( γ ) [16] , [23]–[25] . Collagen IV is thought to exist mostly as Cg25C/Vkg heterotrimers , and LanB1 and LanB2 form the common core of the two laminin trimers , laminin W and laminin A . Thus , the mutant phenotypes of these genes are very similar in their own categories , and absence of one subunit is known to prevent BM incorporation of the other ( s ) [14] , [15] , [26] . The major BM components are expressed and secreted predominantly by the fat body and hemocytes [12] , [26] , [27] , although laminins are also expressed in various other tissues [16] , [27] . Here , we genetically removed each of the major BM components using RNA interference ( RNAi ) followed by careful immunohistochemical analysis and examined their roles in melanotic mass formation . We discovered that lamellocyte encapsulation may be blocked by two separate and discrete self-tolerance checkpoints that operate in healthy target tissues . The first checkpoint involves laminin of the BM , and the second involves cell integrity as determined by cell-cell adhesion and apicobasal cell polarity .
To systematically investigate the relationship between the BM and the melanotic mass phenotype , we disrupted the BM using genetic approaches . We knocked down genes for the two collagen IV subunits and the four laminin subunits individually via UAS-RNAi using ubiquitous ( Act5C-GAL4 ) , inducible ( Hsp70-GAL4 ) , and tissue-specific GAL4 drivers ( HmlΔ-GAL4 , FB-GAL4 , and Cg-GAL4 ) ( summarized in Table S1 ) . Knockdown of any one of the six genes consistently induced black masses in the larvae with either Hsp70-GAL4 or Cg-GAL4 drivers ( Figure 1A and 1B; Table S1 ) . Knockdown of the genes for the BM receptor integrins ( scb for αPS3 and mys for βPS ) , Dystroglycan ( Dg ) , or its cytosolic adaptor Dystrophin ( Dys ) similarly induced black masses ( Figure 1A and 1B; Table S1 ) . Melanotic masses formed mainly in fat bodies and salivary glands . We analyzed the fat bodies of these larvae by immunostaining with the lamellocyte-specific L1 antibody . Pale brown-colored fat bodies ( dissected in early stages of melanin deposition ) from larvae in which collagen IV , laminin , or integrin had been knocked down by RNAi were encapsulated by a few lamellocytes ( Figure 1C–F ) . Black nodules ( dissected in late stages of melanin deposition ) recovered from these larvae were also L1-positive ( Figure S1A ) . Using confocal microscopy , we confirmed the complete disappearance or severe disruption of the BM in the fat bodies of these larvae ( Figure 1G–K ) . This immune response did not appear to be caused by pathogen infection , as the larvae did not induce the antimicrobial peptide gene Attacin-A ( Figure S1B; Text S1 ) . Thus , these data indicate that BM loss induced melanotic mass formation . To determine whether loss of the BM is a general feature of the melanotic mass phenotype , we examined various genes that had been previously associated with the melanotic mass using mutant or RNAi-treated larvae . Because we were primarily interested in the target tissues as opposed to hemocytes , we first excluded mutants that might be classifiable as “true blood cell tumors” , in which melanotic mass formation was due to hemocyte hyperactivation [4] . The following four genes were analyzed for the BM: spag [6] , krz [6] , mRpS30 [28] , and hyx [28] . We found that mutant or RNAi-treated larvae for these genes commonly had disrupted BMs in the fat bodies and that the fat bodies were positive for L1 ( Figure S1C and S1D ) . We also examined 30 other melanotic mass-associated genes; however , the RNAi-treated larvae did not reproduce black nodules with the available UAS transgenes and tissue-specific GAL4 drivers or exhibited early lethality with either ubiquitous or stronger GAL4 drivers , thus precluding further analysis ( Table S2 ) . We also analyzed hopTum larvae , in which the JAK kinase Hopscotch is constitutively active and melanotic mass phenotype is dominant at restrictive temperatures ( >25°C ) [29] . Although the melanotic phenotype of this mutant may fit the classification for the blood cell tumors [30] in that hemocyte numbers increase dramatically ( see Figure 2A and 2B ) , we sought to determine its target tissues . At 25°C , only the collagen IV level decreased severely , while at the restrictive temperature ( 29°C ) , both collagen IV and laminin were absent , and numerous lamellocytes attached to the fat body and the salivary gland ( Figure S1C and S1D ) . Altogether , these observations corroborated the evidence that BM-deficient tissues induce melanotic masses . To see whether lamellocyte encapsulation of BM-deficient tissues was a normal hemocyte reaction to abnormal self-tissue or rather due to an abnormality in the hemocyte itself [4] , we assessed the activation state of hemocytes first by counting the cells . The numbers of circulating hemocytes of some of the BM-deficient larvae were 2–2 . 5-fold higher than those of controls ( Figure 2A ) ; however , the numbers of lamellocytes and crystal cells , a third type of hemocytes that contain phenol oxidation enzymes as a crystal form in their cytosol [31] , did not increase significantly in any of the cases ( Figure 2A , S2A , and S2B; Text S1 ) . This result was in stark contrast to the results for hopTum ( Figure 2A ) , TollD , or cactus mutants , which harbor hyperactive hemocytes [5] , [30] . To inhibit hemocyte hyperproliferation displayed by some of the BM-deficient larvae , we expressed a dominant-negative allele for the JAK/STAT pathway receptor Domeless ( domeΔCYT ) in mys RNAi larvae [18] , [32] . Hemocyte numbers were restored to a normal level in these larvae; however , melanotic mass formation was not abrogated or reduced , indicating that the mass phenotype was not due to hemocyte hyperproliferation ( Figure 2A ) . We then examined the larval hematopoietic lymph gland , as robust lamellocyte differentiation in the lymph gland is a common feature of blood cell tumors [28] , [33]–[36] . The lymph glands of wild-type larvae rarely contained the L1-positive lamellocytes [37] ( Figure 2B ) . The lymph glands of collagen IV- or laminin-knockdown larvae occasionally contained 1–5 L1-positive cells , while lymph glands of integrin-knockdown larvae had 5–20 of these cells . The observed levels of activation may be expected for melanotic mass-forming larvae , but the levels were significantly different from that of hopTum in which the cortical zone of the lymph gland was filled with L1-positive cells [38] ( Figure 2B ) . We then examined sessile hemocytes , another source of lamellocytes upon wasp egg infection [39] . Collagen IV knockdown did not change the sessile hemocyte population , as analyzed by plasmatocyte-specific Eater-GFP ( Figure S2C ) . Finally , we knocked down the BM genes using the fat body-specific FB-GAL4 [28] to determine whether gene manipulation at the target site only , and not in the hemocyte or in the hematopoietic organs , still induced melanotic masses . Knockdown using any of the available UAS-RNAi transgenes singly did not induce melanotic masses , presumably due low knockdown efficiencies with FB-GAL4 driver ( Figure 2C ) ; however , knockdown of various combinations of the transgenes induced melanotic masses specifically in the fat body in the absence of a sensitized genetic background [11] . We obtained similar results following fat body-specific RNAi knockdown for integrins or Dystroglycan , which should act strictly in a cell-autonomous manner ( Figure 2C ) . Circulating hemocytes produced collagen IV and laminin , as reported previously [12] , [26] , [27] , but were not enclosed by a sheet of collagen IV or laminin outside of the cell ( Figure S2D; Text S1 ) , excluding the possibility that knockdown of these components in the fat body may have produced the mass phenotype by affecting the surface of the hemocyte rather than the target . Based on these results , we conclude that melanotic mass formation in BM-deficient larvae results from a normal immune response ( operationally defined by FB-GAL4 ) against altered self and is not ascribed to a failure in the immune system . To investigate which component of the BM is crucial in self-tolerance , we individually knocked down each of the BM genes using various GAL4 drivers and analyzed BM integrity and the melanotic mass phenotype . Since melanotic masses in the BM-deficient larvae formed mainly in fat bodies and salivary glands , we focused on these two organs . In fat bodies , collagen IV knockdown with Hsp70-GAL4 reduced not only collagen IV ( the fluorescence intensity was 5 . 23% of the control level ) but also laminin in the BM ( hereafter , BM laminin ) effectively ( 33 . 46%; Figure 1H and 1K ) . Similarly , laminin knockdown ( 11 . 56% ) nearly completely eliminated the BM collagen IV ( 0 . 91%; Figure 1I and 1K ) , indicating that these factors are structurally inter-dependent in this organ . In salivary glands , however , collagen IV knockdown with the same Hsp70-GAL4 eliminated only collagen IV ( 0 . 47% ) , while leaving 75 . 61% of laminin in the BM ( Figure 3B and 3D ) , allowing for separation of the two components in this organ . Laminin knockdown ( 6 . 71% ) effectively removed the BM collagen IV ( 18 . 02% ) in the salivary gland ( Figure 3C and 3D ) , as in the fat bodies . These results were consistent with the fact that laminin is the key component of BM assembly [14] , [20] , [21] . More importantly , collagen IV knockdown induced melanotic masses only in the fat body and not in the salivary gland , whereas laminin knockdown induced melanotic masses in both organs ( Figure 3E–G and L , the first four experiments; melanotic encapsulation often occurred regionally but not in the entire organs , which might be due to incomplete removal of the BM laminin ) . These results strongly suggest that BM laminin and possibly other factors that are tightly associated with laminin block melanotic mass formation , whereas collagen IV is not necessary for blocking this process as the collagen IV-deficient larvae did not form melanotic masses in the salivary gland ( Figure 3F and 3L ) . To further explore the self-tolerance checkpoint function of the BM components , we knocked down the collagen IV and laminin genes using HmlΔ-GAL4 and FB-GAL4 , which are active in hemocytes and fat bodies , respectively [28] , [40] . Laminin knockdown using these GAL4 drivers was inefficient . As for collagen IV knockdown , the changes in BM integrity induced by these two GAL4 drivers were generally the same as those with Hsp70-GAL4 ( Figure 3H–K ) , except that collagen IV knockdown specifically eliminated the BM collagen IV but left considerable laminin at the BM even in the fat bodies ( Figure S3A–D ) . These larvae did not form black masses , further demonstrating that collagen IV is dispensable in the BM for blocking melanotic masses ( Figure 3L , the last four experiments ) . We next examined Perlecan and Nidogen , the other two major components of the BM . Knockdown of the Perlecan-coding gene trol with Act5C-GAL4 did not induce black masses , indicating that Perlecan is not required in this process ( Figure S3E; Table S1 ) . Knockdown of Nidogen using available RNAi transgenes were unsuccessful; however , we found that Nidogen was completely absent in collagen IV-deficient , melanotic mass-free larvae , indicating that Nidogen is not required for blocking melanotic mass formation ( Figure S3F; Table S1 ) . Finally , we knocked down both trol and vkg , leaving laminin as the only major BM component , and found that melanotic masses were not formed ( Figure 3M and 3N ) . Thus , these results indicate that the BM laminin was sufficient to block melanotic mass formation against self-tissue . As an independent approach to determine BM function , we removed the BM by overexpressing Matrix metalloproteinase 2 ( Mmp2 ) . Mmp2 expression in the salivary gland severely disrupted BM integrity , as reported previously [18] ( Figure S4A and S4B ) , but contrary to our expectations , the larvae did not form melanotic masses . Hemocytes attached to the salivary gland , indicating that hemocyte access to the target cells was not blocked ( Figure S4C and S4D ) . We noticed that the salivary gland cells of these larvae looked similar to those of wild-type larvae ( Figure 4A and 4B vs . 4E and 4F ) , whereas the cells of laminin-knockdown larvae were dissociative and round ( Figure 4G and 4H ) , indicative of the loss of cell-cell adhesion . We further examined whether these melanotic mass-containing larvae had defects in cell polarity using apicobasal cell polarity markers [41] . In wild-type larvae , Cora and Dlg localized to the lateral and basal sides of cells ( Figure 4A and 4B ) . A similar pattern was observed in vkg trol RNAi larvae ( Figure 4C and 4D ) . In ptc>Mmp2 larvae in which Mmp2 is expressed in the salivary gland [18] , the lumen often twisted as it expanded , and cell arrangement occasionally became abnormal ( Figure 4F ) . Dlg diffused throughout the membrane , albeit weakly ( Figure 4F ) . Nevertheless , Cora was still excluded from the apical side ( Figure 4E ) , indicating that apicobasal polarity was partially maintained , perhaps due to the remains of the disrupted BM on the cell surface ( Figure S4B ) . In contrast , the salivary gland cells of laminin-knockdown larvae displayed complete loss of apicobasal polarity . Cora and Dlg were localized throughout the cell membrane and were more often lost from the membrane in these larvae ( Figure 4G and 4H ) . Cell-cell contacts were also severely disrupted . These results strongly suggest that loss of cell integrity , in addition to the loss of the BM , may be required for melanotic mass formation . In this report , we will subsequently use the term “cell integrity” to refer to the cellular aspects involving cell-cell adhesion and cell polarity . To examine the possible role of cell integrity as another checkpoint , we used melanotic mass-free AB1>mys-i larvae . Integrin knockdown with the salivary gland driver AB1-GAL4 [42] did not disrupt the BM but resulted in detachment of the BM from the salivary gland tissue ( Figure 4K ) . The cells lost both cell polarity and adhesion properties , and as a result , the BM appeared as a sack containing sticky balls ( Figure 4I–K ) . As expected , hemocytes were not detected on the surface of the salivary gland in these larvae ( Figure 4L ) . To explore this phenotype in more depth , we first mechanically sheared the salivary gland BM by pinching the AB1>mys-i , GFP larva at its anterior side with forceps ( Figure 5A and 5B ) . After two days , these larvae developed black masses in the salivary gland at a rate that was 12-fold higher than that of the pinched control larvae ( Figure 5C ) . Melanized salivary glands dissected from the wounded larvae were positive for L1 ( Figure 5D ) . Second , we enzymatically disrupted the salivary gland BM of ptc>mys-i larvae by overexpressing Mmp2 , as a means to more specifically manipulate the larvae . These larvae developed black masses , and the salivary glands dissected from the larvae were positive for L1 ( Figure 5E–G ) . In these experiments , ptc-GAL4 and mys-i27735 were used instead of the previously used AB1-GAL4 and mys-i33642 because the latter combination with UAS-Mmp2 caused severe growth retardation of salivary glands . The reproducibility of the knockdown phenotypes was confirmed using mys-i27735 ( Figure S5 ) . Third , we tried to wear out the BM by reducing levels of collagen IV in mys-i larvae . Black masses formed only in the fat bodies of FB>vkg-i , mys-i larvae ( Figure 2C ) : but due to the additional AB1-GAL4 driver , a few of FB+AB1>vkg-i , mys-i larvae developed black masses also in the salivary glands , and again , the salivary glands of those larvae were positive for L1 ( Figure 5J ) . Neither mys-i nor vkg-i alone formed black masses in the salivary gland with the same FB+AB1 GAL4 drivers ( Figure 5H and 5I ) . Taken together , our data demonstrate that cell integrity is an additional and discrete checkpoint for tolerance to self-tissues . We sought to define cell integrity in this system by knocking down genes known to be involved in cell-cell adhesion and cell polarity . Knockdown of either scrib , dlg , cora , FasIII , shg , or arm together with Mmp2 overexpression , however , did not induce melanotic mass formation , indicating that loss of any of these components at least singly did not affect cells sufficiently for disrupting the cell-integrity checkpoint function .
Based on the results of these studies , we propose that BM laminin on target tissues functions as a crucial component not only in BM assembly but as a tolerance checkpoint to self-tissues ( Figure 6 ) . As a self-tolerance checkpoint , the BM may either serve as a physical barrier or provide an inhibitory ligand for hemocyte receptors . We speculate that the latter is the case for several reasons . First , the heterospecific implantation experiments described above suggest that the hemocyte recognizes the BM structure of its own species [11] . Second , insect hemocytes are known to encapsulate a wide range of foreign materials , from parasitoid wasp eggs to synthetic beads , when injected into the hemocoel [37] , [43] , [44] . Encapsulation of parasites is faster than encapsulation of non-parasitic , heterospecific implants [discussed in 11] . Thus , hemocytes must be equipped with various cell surface receptors , including some as activating receptors with different binding spectra for pathogen-associated molecular patterns , and others as inhibitory receptors with narrow binding specificities to self-tissues . More specifically , laminin-coating of Sephadex beads inhibit melanotic encapsulation of the beads in mosquito hemocoel [43] . The outer surface of the Plasmodium oocyst appears to bind to mosquito-derived laminin upon passage through the midgut epithelium of the mosquito [45] , suggesting that the insect laminin may indeed serve as an inhibitory ligand to hemocytes of its own species . Furthermore , we have identified cell integrity as another checkpoint for self-tolerance . Self-tissues must lose both the BM and cell integrity in order to be subjected to lamellocyte encapsulation ( Figure 6 ) . The two checkpoints are indeed discrete and experimentally separable . In our experimental system , BM integrity or cell integrity could be disrupted in the salivary gland selectively via ptc>Mmp2 or AB1>mys-i , respectively . The salivary glands of these larvae are subjected to melanotic encapsulation if the tissues fail to acquire the state of self-tolerance from the other , remaining checkpoint . Therefore , multiple failures on sequential self-tolerance checkpoints elicit autoimmune responses , which is analogous to the mechanism of self-tolerance in the mammalian adaptive immune system [46] . In this respect , BM laminin is unique in that knockdown of laminin disrupted both BM integrity and cell integrity simultaneously . During pharyngeal tube formation of the C . elegans embryo , laminin is required for establishment of cell polarity of a group of precursor cells called the double plate [47] . Since this process occurs before the tissue BM is formed and the mutant phenotype is not shared by collagen IV or perlecan mutants , the authors concluded that this function of laminin is distinguished from its role in the BM . Our results indicate that the Drosophila laminin functions similarly and thus is unique among the BM components . Since cell integrity was identified as a discrete self-tolerance checkpoint , it is interesting that target tissue is not subjected to encapsulation during wound healing or developmental processes in which BM integrity is disrupted temporarily . Upon tissue damage , circulating plasmatocytes attach to the wound region as early as 5 min [17] , presumably to help remodel the tissue , including the disrupted BM , as hemocytes also produce BM components [12] , [27] ( Figure 3I ) . According to our model , as long as cell integrity remains intact , this checkpoint would warrant self-tolerance and block lamellocyte encapsulation in these cases , and thus , wound repair would proceed safely ( as in Figure 5C ) . Cancer metastasis is a more complex problem , as the hallmarks of cancer include BM degradation and loss of epithelial polarity [48] , [49] . Additional work will be required to probe this situation , but it is plausible that cell integrity may exert its checkpoint function normally through inhibitory cytokines , and prior to successful metastasis , cancer cells may have to obtain the ability to secret such cytokines independently of the state of cell integrity . While our results were obtained in the invertebrate Drosophila system , we think these findings have relevance to mammalian autoimmunity . In type I diabetes in humans and in a mouse model , leukocyte infiltration and subsequent β cell destruction have been shown to correlate with disruption of the peri-islet BM , suggesting that the BM may protect islets from autoimmune attack [50] , [51] . In Sjögren's syndrome , which mainly affects the exocrine tissues such as tear and salivary glands , increased degradation of the basal lamina is observed in labial salivary glands , and changes in laminin composition of the salivary acini correlates well with disease progression [52] , [53] . A clear causal relationship is still lacking , but it is tempting to speculate that at least some of the autoimmune diseases may be caused by alterations in BM integrity and/or cell polarity and loss of these features may be recognized by the immune system as “missing self” , similar to the way natural killer cells distinguish between self and nonself [54] , [55] .
The following strains were obtained from public stock centers: UAS-vkg-i ( 106812† ) , UAS-Cg25C-i ( 104536 , 28369 ) , UAS-LanA-i ( 18873 ) , UAS-wb-i ( 108020 ) , UAS-LanB1-i ( 23119 , 23121 ) , UAS-LanB2-i ( 42559 , 42560 , 104013† ) , UAS-trol-i ( 22642 ) , UAS-Ndg-i ( 13208 ) , UAS-mew- i ( 44890 ) , UAS-if-i ( 100770 , 44885 ) , UAS-scb-i ( 4891 , 100949 ) , UAS-ItgαPS4-i ( 37172 ) UAS- ItgαPS5-i ( 6646 , 100120 ) , UAS-mys-i ( 29619† ) , UAS-Itgβν-i ( 893 ) , UAS-Dg-i ( 107029 ) , UAS-Dys-i ( 106401 , 106578 ) , and UAS-krz-i ( 103756† ) from Vienna Drosophila RNAi Center; UAS-vkg-i ( 16858R-3 ) , UAS-trol-i ( 12497R-1 ) , UAS-Ndg-i ( 12908R-3 ) , UAS-mew- i ( 1771R-1 ) , UAS- ItgαPS4-i ( 16827R-2 ) , UAS- Itgβν-i ( 1762R-1 ) , UAS-mys-i ( 1560R-1 ) , UAS-mRpS30-i ( 8470R-4† ) , UAS-hyx-i ( 11990R-2† ) , UAS-dlg1-i ( 1725R-1 ) , and UAS-FasIII-i ( 5803R-1 ) from National Institute of Genetics , Japan; and Hsp70-GAL4 ( 1799 ) , Cg-GAL4 ( 7011 ) [56] , HmlΔ-GAL4 ( 30141 ) , ptc-GAL4 ( 2017 ) , AB1-GAL4 ( 1824 ) , Act5C-GAL4 ( 3954 ) , UAS-Dcr2 ( 24650 ) , UAS-Ser . mg5603 ( 5815 ) , UAS-GFP ( 4775 ) , LanB2MI03747 ( 37366 ) , hopTum ( 8492 ) , spagK12101 ( 12200 ) , UAS-LanA-i ( 28071 ) , UAS-trol-i ( 29440† ) , UAS-mys-i ( 33642 , 27735 ) , UAS-cora-i ( 28993 ) , UAS-dlg1-i ( 33620 ) , UAS-scrib-i ( 29552 ) , UAS-shg-i ( 32904 ) , and UAS-arm-i ( 35004 ) from the Bloomington Stock Center . The following stocks were obtained from private collections: Eater-GFP ( X ) from R . A . Schulz [57]; FB-GAL4 from R . P . Kühnlein [58]; and UAS-domeΔCYT , vkgG454 , and UAS-Mmp2 from T . Xu [18] . The RNAi constructs marked with † proved stronger in knockdown experiments than others and were used for further analysis including imaging . All flies were maintained at 25°C unless otherwise indicated on standard cornmeal and agar media . For RNAi knockdown using Hsp70-GAL4 , a single heat pulse of 30 min at 37°C was applied at 24 , 48 , 72 , or 96 h after egg laying ( AEL ) . After counting larval black nodules , 48 h and 96 h time points were chosen as the two optimal conditions for later experiments as 24 h yielded a lower rate of melanotic mass induction and 72 h resulted in a higher rate of lethality . For imaging the BM , a single heat-shock pulse was applied at 48 h AEL for Hsp70>vkg-i , and double pulses were applied at 48 h and 96 h AEL for Hsp70>LanB2-i104013;+/LanB2MI03747 for stronger knockdown . Larval density and stage were tightly controlled during culture due to the fact that melanotic mass formation was affected by overcrowding growth conditions . Ten virgin females of GAL4 strains were crossed to 5–7 males carrying different UAS-RNA-i strains . After 2 days , eggs were collected for 6 h ( approximately 40–60 eggs ) . Melanotic masses were counted at the wandering third instar stage by rotating the larvae under the dissection microscope . The percentages of larvae containing at least one melanotic mass in their bodies were determined after counting 3 or more vials ( n≥150 ) for each genotype or developmental time point . Wandering third instar larvae were dissected on a silicone pad in phosphate-buffered saline ( PBS ) using a pair of forceps . Larval organs were transferred immediately to 4% paraformaldehyde ( PFA ) and fixed for 15–30 min at room temperature . Samples were washed four times in PBS plus 0 . 1% Tween-20 ( PBST ) and incubated in a blocking solution of PBST plus 5% normal goat serum for 1 h ( PBST-NGS ) . Samples were then incubated with primary antibodies in PBST-NGS overnight at 4°C and then washed four times with PBST-NGS . Samples were then incubated with secondary antibodies alone or together with phalloidin-FITC in PBST-NGS for 2 h at room temperature . After four washes , samples were mounted in Vectorshield plus DAPI ( Vector Laboratories , Inc . , Burlingame , CA ) and were subjected to fluorescent microscopy ( Olympus BX40 ) or confocal microscopy ( Zeiss LSM 510 META ) . For junction staining of the salivary gland , PBS plus 0 . 3% Triton X-100 was used as the buffer . The following antibodies and reagents were used: mouse anti-collagen IV [59] ( Col IV ) ( 6G7 , 1∶100 ) , rabbit anti-LanB1 ( 1∶500; Abcam ) , rabbit anti-LanB2 ( 1∶500; Abcam ) , rabbit anti-Trol [60] ( 1∶2000 ) , rabbit anti-Ndg [61] ( 1∶2000 ) , mouse anti-Hemes [62] ( H2 ) ( 1∶500 ) , mouse anti-L1 [63] ( 1∶500 ) , mouse anti-Mys ( 1∶200; Developmental Studies Hybridoma Bank [DSHB] ) , mouse anti-Dlg 4F3 ( 1∶100; DSHB ) , mouse anti-Coracle C615 . 16 ( 1∶100; DSHB ) , anti-mouse IgG-Cy3 ( 1∶200; Jackson ImmunoResearch ) , phalloidin-FITC ( 1∶50 dilution of 400 µM stock; Sigma–Aldrich ) , and anti-rabbit and anti-mouse IgG conjugated to Alexa 488 or Alexa 546 ( 1∶200; Molecular Probes ) . For quantification of the BM fluorescence intensity , 400× magnified confocal images for the middle-marginal part of the salivary gland or region ‘6’ of the fat body were collected [11] . The fluorescence intensity was measured in 5–7 samples per genotype using the Image J software as described previously [26] . Wandering third instar larvae were washed in PBS and dried . A single larva was bled on a silicone pad in a 12-µl drop of PBS by ripping the epidermis with two fine forceps . The PBS/hemocyte drop was swirled gently using a micropipette tip and mounted on the Neubauer hemocytometer for counting . The total hemocytes were counted , and the lamellocytes were counted based on their characteristic morphology . For each genotype , at least 15 larvae were analyzed . In situ wounding was performed as described previously [18] with minor modifications . Third instar larvae were immersed in water on a silicone pad . Under the GFP/dissection microscope , larvae were gently pinched at the salivary glands ( with help of AB1>GFP ) using a pair of forceps ( Fine Science Tools , Cat . No . 11295-00 ) . Larvae were transferred to fresh cornmeal-agar media and incubated at 25°C . Dead or melanized larvae that were identified within the first 3 h were removed . After 48 h of wounding , non-pupariated larvae were dissected and analyzed for melanotic mass formation on the salivary gland .
|
Autoimmune diseases may be caused by failures in the immune system or by altered selfness in target tissues; however , which of these is more critical is controversial . To better understand such diseases , it is necessary to first define the molecular mechanisms that provide self-tolerance to healthy tissues . As a model system , we used Drosophila melanotic mass formation , in which blood cells encapsulate degenerating self-tissues . By manipulating basement-membrane components specifically in target tissues , not in blood cells , we could elicit autoimmune responses against the altered self-tissues . Moreover , we found that at least two different checkpoints for self-tolerance operate discretely in Drosophila tissues . This parallels mammalian immunity and provides etiological insight into certain autoimmune diseases in which structural abnormalities precede immune system pathology , such as Sjögren's syndrome and type I diabetes mellitus .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"genetics",
"biology",
"and",
"life",
"sciences",
"immunology",
"developmental",
"biology"
] |
2014
|
Basement Membrane and Cell Integrity of Self-Tissues in Maintaining Drosophila Immunological Tolerance
|
The genus Enterovirus of the family Picornaviridae contains many important human pathogens ( e . g . , poliovirus , coxsackievirus , rhinovirus , and enterovirus 71 ) for which no antiviral drugs are available . The viral RNA-dependent RNA polymerase is an attractive target for antiviral therapy . Nucleoside-based inhibitors have broad-spectrum activity but often exhibit off-target effects . Most non-nucleoside inhibitors ( NNIs ) target surface cavities , which are structurally more flexible than the nucleotide-binding pocket , and hence have a more narrow spectrum of activity and are more prone to resistance development . Here , we report a novel NNI , GPC-N114 ( 2 , 2'-[ ( 4-chloro-1 , 2-phenylene ) bis ( oxy ) ]bis ( 5-nitro-benzonitrile ) ) with broad-spectrum activity against enteroviruses and cardioviruses ( another genus in the picornavirus family ) . Surprisingly , coxsackievirus B3 ( CVB3 ) and poliovirus displayed a high genetic barrier to resistance against GPC-N114 . By contrast , EMCV , a cardiovirus , rapidly acquired resistance due to mutations in 3Dpol . In vitro polymerase activity assays showed that GPC-N114 i ) inhibited the elongation activity of recombinant CVB3 and EMCV 3Dpol , ( ii ) had reduced activity against EMCV 3Dpol with the resistance mutations , and ( iii ) was most efficient in inhibiting 3Dpol when added before the RNA template-primer duplex . Elucidation of a crystal structure of the inhibitor bound to CVB3 3Dpol confirmed the RNA-binding channel as the target for GPC-N114 . Docking studies of the compound into the crystal structures of the compound-resistant EMCV 3Dpol mutants suggested that the resistant phenotype is due to subtle changes that interfere with the binding of GPC-N114 but not of the RNA template-primer . In conclusion , this study presents the first NNI that targets the RNA template channel of the picornavirus polymerase and identifies a new pocket that can be used for the design of broad-spectrum inhibitors . Moreover , this study provides important new insight into the plasticity of picornavirus polymerases at the template binding site .
The family Picornaviridae contains 12 genera , and includes many human and animal pathogens ( reviewed in [1] ) . Among these is the genus Enterovirus which contains four human enterovirus species ( HEV-A , -B , -C , -D ) , three human rhinovirus species ( HRV-A , -B , -C ) , simian enterovirus , bovine enterovirus , and porcine enterovirus . The HEV species include poliovirus ( PV ) , coxsackievirus ( CV ) , echovirus , and several numbered enteroviruses ( EV ) . PV is the cause of poliomyelitis , which can lead to acute flaccid paralysis . Enterovirus 71 , a major cause of hand-foot-and-mouth disease , is also frequently associated with flaccid paralysis and is a growing concern due to major epidemics in Southeast Asia . Coxsackieviruses are the main cause of viral meningitis , conjunctivitis , herpangina , myocarditis , and pancreatitis . HRV infections manifest themselves in most cases as the relatively mild common cold , but can cause serious exacerbations in patients with asthma or chronic obstructive pulmonary disease ( COPD ) . Other well-known picornavirus genera are Hepatovirus , which contains hepatitis A virus , Aphthovirus , which contains foot-and-mouth disease virus ( FMDV ) , and Cardiovirus , which includes encephalomyocarditis virus ( EMCV ) , Theiler's murine encephalomyelitis virus , and the recently discovered Saffold virus ( SAFV ) , which , unlike the other cardioviruses , is a human-tropic virus . Currently , the toolbox to control picornavirus infections consists solely of vaccines against PV , hepatitis A virus , and FMDV . Prevention of diseases caused by the non-polio enteroviruses through vaccination seems unachievable given the great number of serotypes , with about 30 coxsackieviruses , 30 echoviruses , 50 numbered enteroviruses , and more than 150 rhinoviruses [2] . With this in mind , current efforts are aimed at developing antiviral compounds with broad-spectrum activity , targeting a wide range of viruses within a genus or ideally even multiple genera . No specific antiviral drugs have yet been clinically approved for the treatment of enteroviruses or any other picornavirus . Picornaviruses are non-enveloped RNA viruses with a single-stranded 7–8 kb RNA genome of positive polarity . Upon entry , the viral RNA is translated into a polyprotein which is proteolytically processed by viral proteases to release the structural proteins ( VP1-4 ) and the non-structural proteins ( 2A-2B-2C-3A-3B-3C-3Dpol and in some genera L ) as well as some stable precursors . The genome is replicated via a negative-sense RNA intermediate by 3Dpol , the viral RNA-dependent RNA polymerase ( RdRP ) . For this , 3Dpol uses the viral protein 3B , in this context usually termed VPg ( viral protein genome-linked ) , as a primer . The structure of 3Dpol has been studied extensively in the past decades . Crystal structures have been reported of 3Dpol of the enteroviruses CVB3 , PV , EV71 , HRV1B , HRV14 , and HRV16 , and of the aphthovirus FMDV [3–10] . These enzymes adopt a classical right hand architecture , with fingers , palm and thumb subdomains providing the correct geometrical arrangement of substrate molecules and metal ions at the polymerase active site for catalysis [11] . In addition , in all picornavirus RdRPs , the conformation of the right hand is closed by the connection of fingers and thumb sub-domains through the N-terminal portion of the protein and several loops protruding from fingers , named the fingertips . This connection leads to the formation of a completely encircled active site and largely restricts the inter-domain mobility . Three well-defined channels have been identified in the RdRP structures that allow access to the active site of the incoming rNTPs and the RNA template , as well as the exit path of the newly synthesized dsRNA product [11] . Seven conserved structural elements , termed motifs A–G , have been defined . Motifs A–E , located in the palm subdomain , are involved in nucleotide and nucleic acid binding as well as catalysis . Motifs F and G , located in the fingers , play a critical role in the binding of nucleoside triphosphates and RNA templates [12] . Given its critical role in genome replication , 3Dpol is regarded as an attractive target for antiviral drugs . Polymerase inhibitors can be subdivided into two classes . Most inhibitors identified so far fall under the class of nucleoside/nucleotide analogs , which compete with endogenous nucleotides and may act as chain terminators and/or mutagens . One example in this class is ribavirin which induces lethal mutagenesis [13] . The second class of polymerase inhibitors are the non-nucleoside inhibitors ( NNI ) which can have a variety of mechanisms of action , e . g . , stabilizing the enzyme such that necessary conformational changes cannot take place [14] . Frequently , these NNIs bind enzyme surfaces which are more variable between viruses and consequently NNIs are more likely to have a restricted spectrum of activity . Also , the barrier to resistance is lower due to the greater tolerance to amino acid substitutions . A few NNIs active against picornaviruses have been discovered , but for most of them , the mechanism of action remains poorly characterized [15–22] . In this study , we report a novel NNI , GPC-N114 that , unlike most NNIs , binds the core of 3Dpol and possesses broad-spectrum activity against enteroviruses and cardioviruses . GPC-N114 represents the first picornavirus 3Dpol inhibitor that targets the template-binding site and reveals a novel pocket in 3Dpol that is amenable for broad-spectrum inhibition .
We recently described a series of 5-nitro-2-phenoxybenzonitriles that inhibit in vitro enterovirus replication [23] . Further optimization of this class of molecules led to the identification of 2 , 2'-[ ( 4-chloro-1 , 2-phenylene ) bis ( oxy ) ]bis ( 5-nitro-benzonitrile ) , hereafter referred to as GPC-N114 ( Fig . 1A ) , with potent and selective in vitro antiviral activity against CVB3 . This small molecule inhibits CVB3 replication in multicycle CPE-reduction antiviral assay with a 50% effective concentration ( EC50 ) of 0 . 15 ± 0 . 02 μM ( Table 1 ) . To study its spectrum of activity , GPC-N114 was evaluated in similar multicycle antiviral assays with representatives of the different HEV and HRV species , as well as several other picornaviruses . All enteroviruses and rhinoviruses included in this study were found to be sensitive to the inhibitory effect of GPC-N114 , with EC50 values ranging from 0 . 1 μM to 1 . 7 μM ( Fig . 1B and Table 1 ) . GPC-N114 also inhibited the replication of EMCV ( strain mengovirus ) , a member of the genus Cardiovirus ( EC50 = 5 . 4 ± 0 . 49 μM ) . By contrast , FMDV and equine rhinitis A virus ( ERAV ) , two members of the genus Aphthovirus , proved insensitive to its inhibitory effect ( tested with concentrations up to 25 μM and 100 μM , respectively ) . To evaluate the antiviral effect of GPC-N114 in a single round of replication , CVB3 and EMCV , as representatives of the enterovirus and cardiovirus genus , were incubated in the presence of different concentrations of the compound and virus titers were determined at 8 h post infection ( p . i . ) . Cell viability assays performed in parallel showed that GPC-N114 was not toxic at these concentrations ( S1A Fig . ) . For CVB3 , the maximal inhibitory effect was obtained with concentrations of 3 μM and higher , although replication was not fully abrogated ( Fig . 1C ) . Replication of EMCV was fully inhibited at concentrations of 10 μM or higher ( Fig . 1C ) . Therefore , a concentration of 10 μM was selected for all further cellular assays . Detailed analysis of the kinetics of CVB3 and EMCV replication at this concentration showed that GPC-N114 strongly delayed virus replication ( Fig . 1D and S1B Fig . ) . Comparable results were observed for a panel of other enteroviruses ( EV71 , CVA21 , HRV2 , and HRV14 ) and cardioviruses ( EMCV and SAFV ) ( Fig . 1E and S1B Fig . ) . Again , no antiviral activity was observed against the aphthovirus ERAV ( Fig . 1E ) . Together , these results demonstrate that GPC-N114 exerts broad-spectrum in vitro antiviral activity , inhibiting the replication of a representative panel of viruses belonging to the enterovirus and cardiovirus genera from the picornavirus family . To identify the stage of the replication cycle that is targeted by GPC-N114 , we used CVB3 and EMCV subgenomic replicons in which ( a part of ) the P1 coding region has been replaced with the firefly luciferase gene . Luciferase expression thus allows the quantification of translation and replication of the viral RNA . At 2 h post-transfection , a time point when no viral RNA replication has taken place yet [24] , luciferase production from both replicons due to translation of incoming RNA was comparable in the presence and absence of GPC-N114 ( Fig . 1F ) , suggesting that GPC-N114 does not affect translation . However , at later time points , luciferase levels were lower in GPC-N114-treated cells in comparison with mock-treated control cells for both the CVB3 and the EMCV replicon . In agreement with the single cycle assays , the EMCV replicon proved more sensitive to the inhibitory effect of GPC-N114 than the CVB3 replicon . These results indicate that GPC-N114 targets the viral RNA replication stage . To determine whether the impaired replication in the presence of GPC-N114 is caused by a defect in the processing of the viral polyprotein , the viral proteins were visualized by metabolic labeling . CVB3-infected BGM cells were pulse-labeled with [35S]Met in the presence or absence of the compound between 5 . 5 and 6 h p . i . , a time when translation of cellular , capped mRNAs is shut off by the virus [25] . The labeled proteins were analyzed by SDS-PAGE and autoradiography . The quantity and pattern of the viral proteins was similar for treated and untreated samples ( S2 Fig . ) , demonstrating that GPC-N114 does not affect synthesis or processing of the viral polyprotein . By serial passaging of virus in the presence of increasing concentrations of GPC-N114 , we aimed to select for GPC-N114-resistant virus . For this , we used a protocol that previously allowed us to select virus variants resistant against a variety of compounds on different cell lines [26–28] . However , despite 32 passages on Vero cells and several attempts , no compound-resistant CVB3 isolates were obtained , indicating a high resistance barrier for this virus to GPC-N114 . Also no resistant viruses were isolated on BGM cells . The possibility that the inability to obtain resistant CVB3 was due to the incomplete inhibition of replication by GPC-N114 ( as indicated by single-cycle assays , though CPE was completely inhibited in multicycle assays ) , resulting in a too low selection pressure , seems unlikely because we also failed to isolate GPC-N114-resistant variants of PV1 ( of which replication was completely blocked by GPC-N114 in a single-cycle assay , S3 Fig . ) and E9 . In contrast , three independent virus pools of EMCV showed phenotypic resistance after as few as three or four serial passages in the presence of suboptimal concentrations of the compound . Viral RNA was isolated from these virus pools and the regions coding for the non-structural proteins were sequenced . Interestingly , the three virus pools each contained a single missense mutation compared to wild-type ( wt ) virus . These mutations were located either at nucleotide 7115 ( A7115G ) or 7127 ( A7127G ) resulting in amino acid changes M300V and I304V in the viral RdRP , 3Dpol , respectively ( Fig . 2A ) . To verify the causal relationship between these mutations and the resistance phenotype , the mutations were introduced into the EMCV cDNA clone pM16 . 1 and mutant viruses were generated . In the multicycle CPE-reduction assay , the EC50 dramatically increased from 5 . 44 ± 0 . 49 μM for EMCV wt to >100 μM for EMCV 3D-M300V and 3D-I304V , indicating that the mutants were resistant to the compound ( Fig . 2A and B ) . Accordingly , while the production of infectious virus particles by wt virus was strongly inhibited in a single cycle assay , virus titers of the mutant viruses after 8 h were unaffected by the presence of 10 μM GPC-N114 ( Fig . 2C ) . As expected , the mutations did not confer resistance to the cardiovirus inhibitor dipyridamole . These results demonstrate that the mutations in 3Dpol were indeed responsible for the resistance phenotype of the isolated virus pools . The X-ray structure of EMCV 3Dpol has been recently reported [29] . It shows that the amino acids M300 and I304 are both located in structural motif B both on the same face of helix α10 ( which is part of the palm subdomain of the polymerase ) , close to the active site ( Fig . 2D ) . Both residues are buried in the core of the protein and their side chains are involved in hydrophobic interactions that stabilize the local structure . Having identified GPC-N114 resistance mutations in EMCV 3Dpol , we set out to test whether the corresponding mutations in CVB3 ( i . e . , 3D-I296V and 3D-M300V ) would yield GPC-N114-resistant variants . Mutations were introduced in a CVB3 cDNA by site-directed mutagenesis and viruses were obtained by transfection of run-off transcripts in cells . The mutant viruses displayed similar increases in virus titer after 8 hours as CVB3 wt in the absence of compound , but did not show a compound-resistant phenotype ( S4A Fig . ) . Also mutation S299T in CVB3 3Dpol , which was previously shown to provide resistance against the 3Dpol inhibitor amiloride [19 , 22] , failed to confer resistance to GPC-N114 ( S4B Fig . ) . The location of the EMCV resistance mutations suggested that GPC-N114 targets the activity of 3Dpol . To verify this , biochemical assays were performed using purified EMCV 3Dpol in the presence or absence of GPC-N114 . Briefly , elongation of a dT15 primer on a poly ( rA ) template by recombinant enzyme was determined by measurement of the rate of incorporation of [3H]UTP . Elongation activity of EMCV 3Dpol on the poly ( rA ) template was inhibited by GPC-N114 in a dose-dependent manner ( Fig . 2E ) , which is in line with the resistance to the compound mapping to 3Dpol in EMCV . Using the same template-primer and metal ion ( i . e . , Mg2+ ) , GPC-N114 did not affect the elongation activity of the polymerase domain of the dengue virus ( DENV ) NS5 RdRP ( S5 Fig . , IC50 > 100 μM ) , demonstrating that the inhibition of EMCV 3Dpol activity was not due to an unspecific interference with these components of the assay . An inhibition constant ( Ki ) of 1 . 3 μM was calculated from the EMCV 3Dpol elongation activity measured at multiple UTP and inhibitor concentrations ( Fig . 2F ) . Fig . 2E shows representative Michaelis-Menten plots of the results acquired with EMCV 3Dpol . The data were fitted by non-linear regression assuming a competitive , noncompetitive , uncompetitive or mixed mode of inhibition . The model of noncompetitive inhibition provided the best fit , with the maximum velocity ( Vmax ) of the reaction declining with increasing concentrations of the compound . This indicates a noncompetitive mode of inhibition with respect to UTP ( and presumably NTPs in general ) . Markedly higher inhibition constants were observed for the two resistant mutants ( M300V , 26 . 9 μM; I304V , 10 . 3 μM ) as compared to that for the wt enzyme ( 1 . 3 μM ) ( Fig . 2F ) . Thus , GPC-N114 inhibited EMCV 3Dpol elongation activity , which could be counteracted by mutations M300V and I304V . Co-crystallization of 3Dpol in complex with GPC-N114 was attempted to study the binding of the compound to picornavirus polymerases . Co-crystals of EMCV 3Dpol with the inhibitor were not obtained , but the structure of the complex of CVB3 3Dpol with GPC-N114 was obtained by X-ray crystallography at 2 . 9 Å resolution ( Table 2 ) . The analysis of the difference electron density maps revealed the presence of extra densities to position the inhibitor within a pocket located at the bottom of the template channel , mimicking the position of the template acceptor nucleotide , the nucleotide which will base-pair with the incoming nucleotide ( Fig . 3A-C and S6 Fig . ) . The GPC-N114-binding pocket is in close proximity to the location of the mutated residues in resistant EMCV 3Dpol ( compare Fig . 2D and Fig . 3A ) . The compound bound in a conformation that perfectly fitted the shape of the pocket formed by residues L107 , E108 , L110 , D111 , T114 of motif G , R188 , Y195 , H199 of motif F , T294 , S295 , I296 of motif B and Y327 of motif A ( Fig . 3A and 3B ) . Most of these residues are conserved in picornaviral polymerases ( S7 Fig . ) . Among all contacts observed , the stacking interaction between Y195 and the central chlorophenyl ring of GPC-N114 appeared to be an important determinant of binding . This tyrosine is strictly conserved in enterovirus polymerases but not conserved in FMDV and EMCV . The first 2-cyano-4-nitrophenyl ring was oriented towards the entrance of the template channel , interacting with residues of motif G and B of CVB3 3Dpol , whereas the second 2-cyano-4-nitrophenyl ring appeared partially exposed to the solvent , occupying the position expected for the template acceptor base ( Fig . 3C ) . The interacting residues appear to be in direct contact with the RNA templates in all picornaviral 3Dpol–RNA complexes determined so far [4 , 12 , 30 , 31] . Structural comparisons between unbound and GPC-N114-bound CVB3 3Dpol revealed that the polymerase did not undergo any major conformational change upon binding of the compound . This preservation of the template cavity conformation extended to the interacting side chains , which maintained their native conformation upon GPC-N114 binding . We also solved the X-ray structure of the complex of CVB3 3Dpol with a second inhibitor of the GPC series , GPC-N143 , at 2 . 7 Å resolution ( S8A Fig . and Table 2 ) . This compound contains a central fluorophenyl ring instead of a chlorophenyl ring and displays similar antiviral activity as GPC-N114 ( S8A and S8B Fig . ) . As expected , the GPC-N143 binding site is identical to that of GPC-N114 , establishing almost equivalent interactions ( Fig . 3B , S6 Fig . and S8D Fig . ) . Thus , the crystallographic data demonstrate that GPC-N114 and GPC-N143 specifically interact with CVB3 3Dpol at the template channel without having major effects on the structure of the enzyme . All attempts to obtain the structure of EMCV 3Dpol–GPC-N114 complex were unsuccessful , however , the superposition of the CVB3 3Dpol–GPC-N114 complex onto the unbound EMCV 3Dpol structures allowed us to define the putative inhibitor binding site in the EMCV enzyme ( Fig . 4A ) . The overall shape of this site is similar in the two enzymes , but some important contacts are lost in EMCV ( e . g . , the stacking interaction between the side chain Y195 and the central ring of GPC-N114 is absent due to the substitution of this residue by alanine in EMCV ) . The model also shows that the major polymerase–inhibitor interactions are mediated by residues of the motif B loop ( β10-α10 ) ( Fig . 4A and S7 Fig . ) . The resistance mutations at EMCV 3Dpol residues 300 and 304 of helix α10 are close to the putative GPC-N114 binding site . However , only the long lateral chain of the EMCV wt residue M300 is able to establish a direct contact with the compound . The apo structures of the EMCV M300V and I304V polymerase mutants were obtained using the same crystallization conditions as for the wt enzyme [29] , yielding two different types of crystals for the mutants . The M300V mutant crystallized in the space group I4122 form with one 3Dpol molecule in the asymmetric unit ( a . u . ) and diffracted to 2 . 2 Å resolution . The 3Dpol I304V mutant crystallized in the space group C2 with six polymerase molecules per a . u . and diffracted to 3 . 2 Å resolution . Data collection and refinement statistics are summarized in Table 2 . Structural comparisons showed that the mutant polymerases were almost identical to the wt enzyme . The main differences were concentrated in helix α10 , which contains the mutated residues , and in the preceding loop B . The side chains of residues 300 and 304 participate in a large hydrophobic cluster linking α10 with helices α2 ( Y75 ) , α6 ( F194 , A195 , F199 ) , α7 ( I210 ) and α8 ( L241 , F255 and F256 ) , included in the inner fingers . The substitutions of the methionine and isoleucine side chains by the shorter valine in the mutants imply a weakening of these hydrophobic interactions . This could result in an increased flexibility of α10 , including the disruption of the last helical turn , and of the B-loop , which appears to have a key role in GPC-N114 binding in EMCV 3Dpol ( Fig . 4 ) . In addition , the substitution with valine may disrupt the interaction of residue 300 with the compound . Prompted by the identification of the binding site of GPC-N114 on CVB3 3Dpol , we analyzed the effects of GPC-N114 on CVB3 3Dpol in more detail . Consistent with the data obtained for EMCV 3Dpol , GPC-N114 inhibited CVB3 3Dpol elongation activity of the homopolymeric substrate poly ( rA ) /dT15 ( Fig . 5A ) . Again inhibition was noncompetitive with respect to UTP . Since the binding sites of GPC-N114 and the template on CVB3 3Dpol overlapped , we aimed to explore whether GPC-N114 competes with the RNA template-primer duplex for binding to 3Dpol . A possible method to test whether GPC-N114 precludes binding of the enzyme to RNA is by monitoring changes in fluorescence polarization signal [32] of a labeled symmetric heteropolymeric template-primer duplex ( sym/sub-U ) [33] . However , though GPC-N114 had an inhibitory effect in the elongation assays with the homopolymeric poly ( rA ) /dT15 template-primer ( Fig . 5A ) , the compound failed to inhibit the incorporation of NTPs using the sym/sub-U template-primer , the latter being in contrast with what we observed in the elongation assays with the homopolymeric poly ( rA ) /dT15 template-primer ( S9 Fig . ) . Therefore , an order-of addition assay was performed to compare the efficiency of GPC-N114 when added to the reaction mixture with 3Dpol before or after addition of poly ( rA ) /dT15 . If GPC-N114 would affect productive binding of the template-primer to 3Dpol , the inhibitor should display an improved potency when added to the reaction mix with 3Dpol before the template-primer as compared to after . When enzyme was pre-incubated with the inhibitor before adding the template-primer on average a 3-fold ( n = 3 ) reduction in the Vmax was observed ( Fig . 5B ) , which is consistent with the reported inactivation of picornavirus polymerases when they are incubated in the absence of nucleic acid [34] . Importantly , the IC50 value for GPC-N114 on CVB3 3Dpol was dramatically reduced ( on average 14-fold , n = 3 ) when the compound was added to the reaction mix before the template-primer ( Fig . 5B ) . Unfortunately , this assay is not able to discriminate between direct competition and the formation of a nonproductive ternary complex consisting of GPC-N114–template-primer–3Dpol . Nevertheless , the observation that GPC-N114 has a stronger inhibitory effect on 3Dpol activity when added before the template-primer is compatible with the overlapping binding sites of the compound and the template-primer on 3Dpol .
There is a great need for broad-range antiviral compounds to treat infections with enteroviruses , which includes important human pathogens such as poliovirus , coxsackievirus , enterovirus 71 , and rhinovirus . Here , we report a novel small molecule , GPC-N114 that exerts broad-spectrum anti-enterovirus activity and also inhibits members of the genus Cardiovirus , such as EMCV . Analysis of its mechanism of action revealed that GPC-N114 inhibits virus replication at the stage of RNA replication . GPC-N114-resistant enterovirus variants could not be obtained , but compound-resistant EMCV variants were readily selected in the presence of suboptimal concentration of GPC-N114 . These variants were found to carry mutations in the viral RdRP , 3Dpol . Consistently , the in vitro elongation activity of CVB3 and EMCV 3Dpol was inhibited by GPC-N114 , and the mutations identified in compound-resistant EMCV 3Dpol rendered this polymerase less susceptible to inhibition . GPC-N114 did not compete with incoming NTPs , but interfered with productive binding of the template-primer to 3Dpol in a primer elongation assay . This is in agreement with the crystallographic studies of the CVB3 3Dpol–GPC-N114 complex which revealed that the binding site of the compound is located at the junction of the palm and the fingers domains , and overlaps partially with the binding site of the template . A high barrier to resistance is a desirable feature for antiviral compounds since drug resistance represents a major problem in the treatment of viral infections . Notably , no GPC-N114-resistant enteroviruses ( CVB3 , PV1 , E9 ) were obtained after several attempts and a large number of serial passages of these viruses in the presence of suboptimal concentrations of the compound . Attempts to generate compound-resistant CVB3 by introduction of the mutations that correspond to those identified in GPC-N114-resistant EMCV ( i . e . , 3D-I296V and 3D-M300V in CVB3 ) , also did not result in a compound-resistant phenotype . The reason for the inability of enteroviruses to develop resistance against GPC-N114 remains to be established . A possible explanation is that mutations that would confer resistance to GPC-N114 also impair binding of the template-primer , thereby preventing replication . Although the exact reason remains to be determined , our data suggest that the compound-binding site in enterovirus 3Dpol lacks the conformational plasticity for resistance to develop , in contrast to most allosteric binding sites at the enzyme surface . In contrast , EMCV 3Dpol is sufficiently plastic to allow for substitutions that result in compound-resistance . The amino acid changes are conservative with all three amino acids having similar properties . Based on the constructed model , only the long side chain of M300 ( equivalent to I296 in CVB3 3Dpol , see Fig . 3B ) , would be in direct contact with the compound . In contrast , I304 seems to be located relatively far away of the GPC-N114 binding cavity . Both M300 and I304 are part of conserved structural motif B , which not only binds the RNA template in the active site but also participates in the mechanism of RNA translocation through fine movements of a loop located at the N-terminus of the motif , the B-loop [30 , 35] . Structural data from picornavirus polymerases [30 , 35] as well as other RdRPs [36] suggest that the B-loop is highly dynamic and that its dynamics can be regulated by external effectors [37] . The structures of the EMCV 3Dpol M300V and I304V mutants suggest that these mutations increase the flexibility of this loop , thereby affecting the binding site of GPC-N114 and probably losing a direct interaction with the shorter V300 side chain whilst retaining the ability to interact with the template-primer . EMCV 3Dpol and CVB3 3Dpol display a high structural similarity , but there is a major difference in the main interactions of these polymerases with GPC-N114 that could possibly underlie the difference in resistance emergence . In CVB3 3Dpol , the main interaction with GPC-N114 is mediated by Y195 . EMCV 3Dpol , however , possesses an alanine at this position which forms a weaker interaction ( S10 Fig . ) . Due to this weaker interaction , a flexibility increase in the compound-binding area induced by the B loop resistance mutations may result in a decreased binding of GPC-N114 to EMCV 3Dpol . For CVB3 3Dpol , a possible increase in flexibility induced by B loop mutations may not , or to a lesser extent , affect the binding of GPC-N114 due to its strong interaction with Y195 . GPC-N114 proved to be not a very robust compound in biochemical assays , for which the reason might be that it interferes with the binding of a by far larger molecule , the RNA template-primer . Thus , despite our efforts , it remains to be established whether GPC-N114 inhibits 3Dpol elongation activity by competing with the template-primer for binding to 3Dpol or whether binding of the compound results in the formation of an unproductive conformation of the polymerase-template complex . Nevertheless , both the crystallographic studies as well as the order-of-addition experiment strongly suggest that GPC-N114 affects binding of the template-primer by 3Dpol . Compounds that interfere with binding of the template-primer have been described for polymerases of some viruses , including DENV and HIV [38 , 39] , but not for any of the picornaviruses . The only known picornavirus polymerase inhibitors with a related mechanism of action are two compounds that target the incoming NTP binding site in the PV polymerase [15] . Clearly , the NTP binding site is close to , but distinct from , the GPC-N114 binding site . Hence , GPC-N114 is the first reported picornavirus polymerase inhibitor that binds the site of the template acceptor nucleotide . Polymerase inhibitors have demonstrated their importance for the treatment of other viral infections such as those caused by HIV , herpesvirus , HBV and HCV . Unfortunately , it was not possible to evaluate the antiviral effect of GPC-N114 in an animal model , since—despite extensive attempts—we have not been able to formulate GPC-N114 ( or GPC-N143 ) for administration . Notwithstanding this , the identification of a new druggable pocket in the polymerase through the characterization of the mechanism of action of GPC-N114 may be key in the ( in silico ) design of novel picornavirus polymerase NNIs .
Buffalo green monkey ( BGM ) kidney cells , HeLa R19 cells ( obtained from G . Belov , University of Maryland and Virginia-Maryland Regional College of Veterinary Medicine , US ) , baby hamster kidney 21 ( BHK-21 ) cells , and human rhabdomyosarcoma RD cells were grown at 37°C , 5% CO2 in Dulbecco’s Modified Eagle medium ( DMEM ) ( Gibco ) supplemented with 10% fetal bovine serum and antibiotics [28 , 40] . The Hela H cells , BGM cells , and Vero cells used for the multicycle CPE-reduction assays were maintained in Minimal Essential Medium supplemented with 10% FBS , 1% bicarbonate , and 1% L-glutamine [28] . The synthesis of GPC-N114 will be published elsewhere . Guanidine hydrochloride ( GuHCl ) and dipyridamole were purchased from Sigma Aldrich . Dipyridamole was dissolved in ethanol and GuHCl in water . The other compounds were dissolved in DMSO . GPC-N114 was used in a concentration of 10 μM unless stated otherwise . Coxsackievirus B3 ( strain Nancy ) was obtained by transfection of RNA transcripts of the full-length infectious clone p53CB3/T7 linearized with SalI into BGM cells [41] . EMCV , strain mengovirus , was obtained in a similar manner by transfection of RNA derived from cDNA clone pM16 . 1 into BHK-21 cells [42] . EV68 , EV70 , EV71 ( BrCr ) , E11 ( Gregory ) , CVA16 ( G-10 ) and CVA21 ( Coe ) were obtained from the National Institute for Public Health and Environment ( RIVM , the Netherlands ) . Poliovirus 1–3 Sabin reference strains were obtained from dr . J Martin ( NIBSC , UK ) . Human rhinoviruses 2 and 14 were supplied by Joachim Seipelt ( Medical University of Vienna , Austria ) . ERAV ( NM11/67 ) was a gift from David Rowlands and Toby Tuthill ( University of Leeds , UK ) . Saffold virus ( type 3 ) was described previously [43] . The FMDV strains O1 Manisa and A22 Iraq 24/64 were used [44] . For virus infections , virus was added to subconfluent cell layers and allowed to adsorb for 30 minutes , after which virus was removed and fresh ( drug-containing ) medium was added to the cells . At the indicated times p . i . the cells were freeze-thawed three times to release intracellular virus particles . Virus titers were determined by endpoint titration according to the method of Reed and Muench and expressed as 50% cell culture infective doses ( CCID50 ) /ml [45] . The antiviral activity of compounds was determined as published previously [11 , 44] . In short , the medium was removed from cells grown in 96-well plates and a serial dilution of compound in medium with 2% FCS was added after which the cells were infected with virus at the lowest MOI sufficient for full CPE development in infected , untreated cultures after three days of incubation at 37°C . To quantify the level of CPE by MTS assay , the medium was then replaced with AQueous One Solution Cell Proliferation Reagent ( Promega ) diluted in medium and incubated at 37°C for 1–2 h . Subsequently , absorbance at 498 nm was measured . The measured values were used to calculate the cell viability ranging from 0% ( full CPE with infected , untreated cultures ) to 100% ( uninfected , untreated cultures ) . The 50% effective concentration ( EC50 ) was defined as the concentration of compound that resulted in a 50% reduction in virus-induced CPE and was calculated using logarithmic interpolation . The subgenomic replicons of CVB3 ( pRib-LUC-CB3/T7 ) and EMCV ( pEMCV-FLuc ) encode the cDNA clones of the corresponding virus except that the P1 area encoding the capsid proteins has been ( partially ) replaced by the code for firefly luciferase [24 , 46] . The plasmids were linearized with MluI or BamHI , respectively . RNA transcribed from the linearized plasmids was used to transfect BGM cells plated in 24-well plates using the DEAE-dextran method as described previously [24] . After transfection , cells were treated with DMSO or compound and incubated at 37°C . At 2 , 4 , 6 , and 8 h post-transfection , cells were washed with PBS , lysed and luciferase levels were determined with the Luciferase Assay System ( Promega ) according to the manufacturer’s instructions . A similar procedure was followed for the PV1 subgenomic replicon assay pXpA-RenR-PV except that luciferase activity was measured with the Renilla Luciferase Assay System ( Promega ) , since this construct encodes the Renilla luciferase [47] . For metabolic labeling of proteins , BGM cells were grown in 24-well plates to subconfluence and infected with CVB3 at an MOI of 50 . At 5 h post-infection , the medium was replaced with 300 μl of methionine-free medium for 30 min . Subsequently , the cultures were pulse-labeled in methionine-free medium supplemented with [35S]Met in the absence or presence of 50 μM GPC-N114 for 30 min . At 6 h post-infection , cells were lysed with lysis buffer [50 mM Tris ( pH 7 . 4 ) , 150 mM NaCl , 1 mM EDTA , 1% Nonidet P-40 , 0 . 05% SDS] . Laemmli sample buffer was added to the lysates , boiled for 5 min and analyzed on a 10% polyacrylamide gel containing SDS . The gels were fixed in 30% methanol-10% acetic acid , rinsed in DMSO , fluorographed with 20% 2 , 5-diphenyloxazole in DMSO , dried , and exposed to Kodak XAR film . To obtain drug-resistant virus , three independent virus pools of CVB3 and EMCV were serially passaged in the presence of a concentration series of GPC-N114 on subconfluent to confluent Vero cultures in 96-well plates . The amount of virus used for infection was selected so that full CPE was visible after 3 days of incubation in the absence of compound . The concentration series encompassed 8 different concentrations in 2-fold dilutions starting from 25–100 μM . After 3 to 4 days , the amount of CPE was assessed and supernatant collected from the culture with the highest concentration of compound that still exhibited full CPE was used to perform the next round of infection . This procedure was repeated 32 times ( CVB3 ) or until there was a clear increase in the concentration required for inhibition of CPE ( EMCV ) compared to wt virus that was tested in parallel in each passage . Viral RNA was isolated from the resistant virus pools using the GenElute Mammalian Total RNA Miniprep Kit ( Sigma Aldrich ) and used for sequencing of the parts of the genome that encode the nonstructural proteins . The PCR product containing the mutations found in the virus pools generated in this procedure was digested with PinAI and NruI and cloned into pM16 . 1 to replace the corresponding wt sequence . The resulting infectious clones pM16 . 1-3D-M300V and pM16 . 1-3D-I304V were sequenced to confirm the presence of the mutations . The analogous mutations in CVB3 ( I296V and M300V ) were introduced into the infectious cDNA p53CB3/T7CVB3 by site-directed mutagenesis . The mutant clones were sequenced to confirm the presence of the mutations . The construction of pRibCB3/T7 with the S299T mutation was described previously [48] . Recombinant EMCV and CVB3 was generated as described above and the presence of the mutations was again confirmed by sequence analysis . Recombinant CVB3 3Dpol and the RdRP domain of DENV NS5 ( NS5-RdRP ) were obtained as previously described [49 , 50] . For expression and purification of EMCV 3Dpol , the coding sequence of 3Dpol ( polyprotein domain from amino acid position 1834 to 2293 ) was amplified on EMCV cDNAs ( pM16 . 1 wt , 3D-M300V , and 3D-I304V ) and cloned into pETG20A ( kindly provided by Dr . Arie Geerlof ) by Gateway recombination ( Life Technologies ) , downstream a cleavable Hexahistidine-Thioredoxin tag using a two-step PCR protocol . The protein was expressed in E . coli Rosetta ( DE3 ) pLysS strain ( Novagen ) at 17°C in Tur*best Broth ( Arcadia Biotech ) . The purification of the protein and the tag removal were performed in non-denaturing conditions as previously described [51] . The final size exclusion chromatography step was performed in 10 mM Hepes , 300 mM NaCl , 2mM dithiotreitol , pH 7 . 5 . The conditions for the polymerase elongation activity assays were adapted from conditions reported previously [10 , 50] . EMCV 3Dpol filter-binding polymerase activity assays were conducted in a mix of 50 μl containing 50 mM MOPS pH 7 . 0 , 10 mM KCl , 4 mM MgCl2 , 9% glycerol , 2 μM dT15 ( Invitrogen ) annealed to 350 nM poly ( rA ) ( GE Healthcare , average size 519 nt ) ( i . e . , a molar ratio dT15:poly ( rA ) of 5 . 7:1 ) , 1 μM EMCV 3Dpol , 50/100/200 or 500 μM UTP and 2 μCi [3H]UTP ( GE Healthcare , 35 . 5 Ci/mmol ) . The reaction mix excluding UTP was incubated with either DMSO or GPC-N114 with a constant concentration of 5% DMSO for 2 minutes after which the reaction was started by addition of labeled and unlabeled UTP . Samples of 10 μl were taken after 0 , 10 , 20 and 30 min of incubation at 30°C and added to 50 μl of 100 mM EDTA in 96-well sample plates to stop the reaction . Mixtures were then transferred onto glass fiber filter mats with DEAE active groups ( DEAE filtermat , Wallac ) using a Filtermat Harvester ( Packard Instruments ) . Filtermats were washed three times with 0 . 3 M ammonium formate pH 8 . 0 , twice with water , and once with ethanol after which they were dried and transferred into sample bags . Liquid scintillation fluid was added to the sample bags and incorporation of radiolabeled UTP was measured in counts per minute using a Wallac MicroBeta TriLux Liquid Scintillation Counter . For the DENV NS5 elongation assay the reaction mixture contained 50 mM HEPES pH 8 . 0 , 10 mM KCl , 10 mM DTT , 5 mM MgCl2 , 2 μM dT15 annealed to 350 nM poly ( rA ) , 0 . 4 μM DENV NS5-RdRP , 20 μM UTP and 2 μCi [3H]UTP . 10 μl samples were taken after 0 , 30 , 60 , and 90 minutes . Subsequently , the same procedure was followed as for EMCV 3Dpol . The CVB3 3Dpol reaction mixture contained 50 mM Tris pH 7 . 0 , 10 mM KCl , 0 . 8 mM MgCl2 , 9% glycerol , 2 μM dT15 annealed to 350 nM poly ( rA ) , 20 nM CVB3 3Dpol , UTP ( 2 , 5 , 10 or 20 μM ) and 0 . 5 μCi [3H]UTP . 10 μl samples were taken after 0 , 5 , 10 , and 15 minutes . For the order-of-addition experiment with CVB3 3Dpol , a reaction premix was assembled as above but without poly ( rA ) /dT15 , GPC-N114 and UTP . To this premix either a range of concentrations of GPC-N114 or 80 nM dT15 annealed to 14 nM poly ( rA ) were added . After an incubation of 1 min at RT , the other component was added to the mix followed by an incubation of several minutes at 30°C . The reaction was started by addition of UTP , and samples were taken after 0 , 3 , 6 , and 9 minutes . The data were fitted using non-linear regression using Graphpad Prism 5 . 0 . 3 . Ki values averaged from at least three independent experiments were statistically compared between EMCV 3Dpol wt and the 3Dpol mutants using a Student’s t-test . Recombinant CVB3 3Dpol was over-expressed essentially as previously described [49] , except that the expression was carried out at 20°C . Stored pellets were resuspended in a cold lysis buffer containing 50 mM Tris HCl pH 9 . 0 , 500 mM NaCl , 10 mM imidazole , 0 . 1% ( v/v ) Triton X-100 , 5% glycerol , 1 mg/ml lysozyme and 10 μg/ml DNase , and cells were lysed by sonication . Cellular debris was pelleted by centrifugation and the soluble fraction was filtrated and loaded into a 5 ml HisTrap HP column ( GE Healthcare ) . The protein was eluted using an imidazole gradient ( 10–500 mM ) in 50 mM Tris pH 8 . 0 and 300 mM NaCl . Protein fractions were dialyzed against 20 mM Tris pH 9 . 0 , 600 mM NaCl , 15% glycerol , 0 . 5 mM TCEP and further purified by size-exclusion chromatography on a Superdex 200 16/60 column ( GE Healthcare ) . Purified protein was concentrated to 5 mg/ml in the same buffer , using a Vivaspin concentrator ( 10 . 000 MWCO PES , Sartorius Stedim Biotech ) . The quality of the protein obtained was checked by SDS-PAGE . EMCV 3Dpol wt and mutants were expressed and purified as previously reported for EMCV 3Dpol wt [29] . Briefly , using the constructs described above , the polymerases were overexpressed in E . coli , and EMCV 3Dpol was purified by means of Ni-affinity ( HisTrap HP , GE Healthcare ) . After that , the His-tag was removed by TEV digestion , and size exclusion chromatography ( Superdex 200 16/60 , GE Healthcare ) . Crystals of isolated CVB3 3Dpol were obtained by the sitting-drop vapor-diffusion method at 16°C in 96-well plates ( MRC , Swissci AG ) using a Cartesian automated drop dispenser to mix 250 nl of 5 mg/ml protein solution with an equal volume of reservoir solution , containing 50 mM Tris pH 7 . 5 , 24 . 5% ( w/v ) glycerol , 1 . 29 M ammonium sulfate . Highly reproducible crystals of 25x25x25 μm in size appeared between 2–3 days . Protein/inhibitor complexes were prepared by soaking CVB3 3Dpol crystals in the inhibitor solution for 72h . The inhibitor solutions contained 100 μM GPC-N114 or GPC-N143 in 2 . 5% DMSO , 0 . 65 M ammonium sulfate , 12 . 5% glycerol and 25 mM Tris pH 7 . 5 . Crystals were then transferred to a cryo-protecting solution , containing 30% glycerol in the respective crystallization buffer prior to cooling by immersion in liquid nitrogen . The two datasets were collected at 100K by using synchrotron radiation at the European Synchrotron Radiation Facility ( ESRF , Grenoble ) on a Pilatus 6M S/N 60–104 detector at beamline ID 29 , λ = 0 . 977Å . Diffraction images were processed with iMOSFLM [52] and internally scaled with SCALA ( CCP4i [53] ) , achieving resolutions of 2 . 9Å and 2 . 7Å for the CVB3 3Dpol–GPC-N114 and CVB3 3Dpol–GPC-N143 crystals , respectively . Data collection statistics are given in Table 2 . Coordinates have been deposited at the PDB with accession codes 4Y2A and 4Y34 for the complexes CVB3 3Dpol–GPC-N114 and CVB3 3Dpol–GPC-N143 , respectively . Crystals of EMCV 3Dpol wt and M300V and I304V , mutants were obtained as previously described [29] . The data sets were collected at the ALBA-CELLS synchrotron light facility ( Barcelona ) on a Pilatus 6M Dectris detector at beamline XALOC λ = 0 . 977Å . Diffraction images were processed with iMOSFLM and SCALA , achieving resolutions of 2 . 2Å and 3 . 2Å for the M300V and I304V respectively ( Table 2 ) . Coordinates have been deposited at the PDB with accession codes 4Y2C and 4Y3C for the structures of EMCV 3Dpol-M300V and EMCV 3Dpol-I304V , respectively . The initial electron density maps of the CVB3 3Dpol–GPC-N114 and CVB3 3Dpol–GPC-N143 complexes were obtained after rigid-body fitting of the coordinates of the 3Dpol apoenzymes ( PDB id . 3DDK ) to the new unit cells ( Table 2 ) , using the program Refmac5 [54 , 55] . In both structures , the weighted 2|Fo|-|Fc| and |Fo|-|Fc| difference map showed the presence of extra densities that permitted the initial positioning of the inhibitor molecules . Inhibitor structure files were generated using “Online SMILES Translator and Structure File Generator” ( http://cactus . nci . nih . gov/translate/ ) and its respective CIF dictionaries were created by Grade [56] . Several cycles of automatic refinement , performed with Refmac5 , were alternated with manual model rebuilding using Coot [57] . The refinement statistics are summarized in Table 2 . The first electron density maps for both EMCV 3Dpol mutants were obtained after rigid-body fitting of the EMCV 3Dpol wt coordinates to the new unit cells . The PDB id 4NYZ , of space group I4122 , was used as a model for the M300V structure and coordinates 4NZ0 , of space group C2 , were used as a model for the I304V mutant . Noncrystallographic symmetry ( ncs ) restraints were initially applied to the six independent molecules present in the asymmetric unit of the C2 crystals using Refmac5 . However , the ncs restrained-refinement result was unstable because the presence of different conformations in many polymerase regions , especially in the surface-exposed loops , was affected by crystal packing interactions . Therefore , the variable regions were manually rebuilt using Coot , and the final cycles of automatic refinement with Refmac 5 , were performed in absence of ncs restraints . The refinement statistics are given in Table 2 .
|
Virus replication relies on multiplication of viral genomes by viral polymerases . For enteroviruses , a large group of clinically important human pathogens for which no antiviral therapy is available , this function is performed by 3Dpol , the RNA-dependent RNA polymerase . 3Dpol is therefore an attractive target for novel antiviral strategies . Most polymerase inhibitors identified today are nucleoside analogs , a class of compounds that exert broad-spectrum activity but often suffer from off-target effects . Non-nucleoside inhibitors on the other hand , in general have a more narrow spectrum of activity and are more prone to resistance development because in most cases they bind the surface of the enzyme which is less conserved and structurally more flexible . In this study , we present the identification of GPC-N114 as a non-nucleoside inhibitor of 3Dpol with broad-spectrum antiviral activity against both enteroviruses and cardioviruses , which also belong to the picornavirus family . Remarkably , it acts by targeting the RNA template-primer binding site in the core of 3Dpol , making GPC-N114 the first anti-picornaviral compound with this mechanism of action . Thus , the characterization of GPC-N114 has led to the identification of a novel drug-binding pocket in 3Dpol that can serve as a starting point for antiviral drug design .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
The RNA Template Channel of the RNA-Dependent RNA Polymerase as a Target for Development of Antiviral Therapy of Multiple Genera within a Virus Family
|
Important cellular processes such as migration , differentiation , and development often rely on precise timing . Yet , the molecular machinery that regulates timing is inherently noisy . How do cells achieve precise timing with noisy components ? We investigate this question using a first-passage-time approach , for an event triggered by a molecule that crosses an abundance threshold and that is regulated by either an accumulating activator or a diminishing repressor . We find that either activation or repression outperforms an unregulated strategy . The optimal regulation corresponds to a nonlinear increase in the amount of the target molecule over time , arises from a tradeoff between minimizing the timing noise of the regulator and that of the target molecule itself , and is robust to additional effects such as bursts and cell division . Our results are in quantitative agreement with the nonlinear increase and low noise of mig-1 gene expression in migrating neuroblast cells during Caenorhabditis elegans development . These findings suggest that dynamic regulation may be a simple and powerful strategy for precise cellular timing .
Proper timing is crucial for biological processes , including cell division [1–3] , cell differentiation [4] , cell migration [5] , viral infection [6] , embryonic development [7 , 8] , and cell death [9] . These processes are governed by molecular events inside cells , i . e . , production , degradation , and interaction of molecules . Molecular events are subject to unavoidable fluctuations , because molecule numbers are small and reactions occur at random times [10 , 11] . Cells combat these fluctuations using networks of regulatory interactions among molecular species . This raises the fundamental question of whether there exist regulatory strategies that maximize the temporal precision of molecular events and , in turn , cellular behaviors . A canonical mechanism by which a molecular event triggers a cellular behavior is accumulation to a threshold [3 , 4 , 12–14]: molecules are steadily produced by the cell , and once the molecule number crosses a particular threshold , the behavior is initiated . The temporal precision of the behavior is therefore bounded by the temporal precision of the threshold crossing . Threshold crossing has been shown to underlie cell cycle progression [3] and sporulation [4] , although alternative strategies , such as derivative [9] or integral thresholding [15] , oscillation [16] , and dynamical transitions in the regulatory network [8] , have also been investigated . Recent work has investigated the impact of auto-regulation ( i . e . , feedback ) on the temporal precision of threshold crossing [12 , 13] . Interestingly , it was found that auto-regulation generically decreases the temporal precision of threshold crossing , meaning that the optimal strategy is a linear increase of the molecule number over time with no auto-regulation [12] ( although auto-regulation can help if there is a large timescale separation and the threshold itself is also subject to optimization [13] ) . Indeed , even when the molecule also degrades , the optimal precision is achieved when positive auto-regulation counteracts the effect of degradation , preserving the linear increase over time [12] . However , in many biological processes , such as the temporal control of neuroblast migration in Caenorhabditis elegans [5] , the molecular species governing the behavior increases nonlinearly over time . This suggests that other regulatory interactions beyond auto-regulation may play an important role in determining temporal precision . In particular , the impact of activation and repression on temporal precision , where the activator or repressor has its own stochastic dynamics , remains unclear . Here we investigate the temporal precision of threshold crossing for a molecule that is regulated by either an accumulating activator or a degrading repressor . Using a first-passage-time approach [12 , 17–19] and a combination of computational and analytic methods , we find that , unlike in the case of auto-regulation , the presence of either an activator or a repressor increases the temporal precision beyond that of the unregulated case . Furthermore , the optimal regulatory strategy for either an activator or a repressor corresponds to a nonlinear increase in the regulated molecule number over time . We elucidate the mechanism behind these optimal strategies , which stems from a tradeoff between reducing the noise of the regulator and that of the target molecule , and is similar to the fact that a sequence of time-ordered stochastic events becomes more precisely timed with more events . These findings are robust to more complex features of the regulation process , including bursts of molecule production , more complex regulator dynamics , and cell division . Our results are quantitatively consistent with both the temporal precision and nonlinearity of the mig-1 mRNA dynamics of the migrating neuroblast cells in C . elegans larvae [5] . The agreement of our simple model with these data suggests that many molecular timing processes may benefit from the generic regulatory strategies we identify here .
To investigate the effects of regulation on temporal precision , we consider the timing variance σ t 2 as a function of the parameters k and K , or μ and K . The special case of no regulation corresponds to the limits k → ∞ and K → 0 in the case of activation , or μ → ∞ and K → ∞ in the case of repression . In this case , the production of X occurs at the constant rate α . Reaching the threshold requires x* sequential events , each of which occurs in a time that is exponentially distributed with mean 1/α . The total completion time for such a process is given by a gamma distribution with mean t ¯ = x * / α and variance σ t 2 = x * / α 2 [19] . Ensuring that t ¯ = t * requires α = x*/t* , for which the variance satisfies σ t 2 x * / t * 2 = 1 . This expression gives the timing variance for the unregulated process . In Fig 2 we plot the scaled variance σ t 2 x * / t * 2 as a function of the parameters k and K , or μ and K , for cooperativity H = 3 ( color maps ) . In the case of activation ( Fig 2A ) , the variance decreases with increasing k and K . This means that the temporal precision is highest for an activator that accumulates quickly and requires a high abundance to produce X . In the case of repression ( Fig 2B ) , the variance has a global minimum as a function of μ and K . This means that the temporal precision is highest for a repressor with a particular well-defined degradation rate and abundance threshold . Importantly , we see that for both activation and repression , the scaled variance can be less than one , meaning that regulation allows improvement of the temporal precision beyond that of the unregulated process . We have checked that this result holds for H ≥ 1 . To understand the dependencies in Fig 2 , we develop analytic approximations . First , we assume that H → ∞ , such that the regulation functions in Eqs 1 and 2 become threshold functions . In this limit , the production rate of X is zero if a < K or r > K , and α otherwise . The deterministic dynamics of X become piecewise-linear , x ¯ ( t ) = { 0 t < t 0 α ( t - t 0 ) t ≥ t 0 , ( 5 ) where t0 is determined by either a ¯ ( t 0 ) = K or r ¯ ( t 0 ) = K according to Eqs 3 and 4 . Then , to set α , we use the condition x ¯ ( t * ) = x * , which results in α = x*/ ( t* − t0 ) . Lastly , we approximate the variance in the first-passage time using the variance in the molecule number and the time derivative of the mean dynamics [13] . Specifically , the timing variance of X arises from two sources: ( i ) uncertainty in the time when the regulator crosses its threshold K , which determines when the production of the target X begins , and ( ii ) uncertainty in the time when x crosses its threshold x* , given that production begins at a particular time . The first source is regulator noise , and the second source is target noise . We estimate these timing variances from the associated molecule number variances , propagated via the time derivatives , σt2≈σy2 ( dy¯dt ) −2|t0︸regulator+σx2 ( dx¯dt ) −2|t*︸target , ( 6 ) where y ∈ {a , r} denotes the regulator molecule number . For the activator , which undergoes a pure production process with rate k , the molecule number obeys a Poisson distribution with mean kt . Therefore , the molecule number variance at time t0 is σ a 2 = k t 0 . For the repressor , which undergoes a pure degradation process with rate μ starting from N molecules , the molecule number obeys a binomial distribution with number of trials N and success probability e−μt . Therefore , the molecule number variance at time t0 is σ r 2 = N e - μ t 0 ( 1 - e - μ t 0 ) . For the target molecule , which undergoes a pure production process with rate α starting at time t0 , the molecule number obeys a Poisson distribution with mean α ( t − t0 ) . Therefore , the molecule number variance at time t* is σ x 2 = α ( t * - t 0 ) . Inserting these expressions into Eq 6 , along with the derivatives calculated from Eqs 3–5 and the appropriate expressions for α and t0 , we obtain σ t 2 x * t * 2 ≈ K x * ( k t * ) 2 + ( 1 - K k t * ) 2 ( activator ) , ( 7 ) σ t 2 x * t * 2 ≈ ( N - K ) x * N K ( μ t * ) 2 + [ 1 - log ( N / K ) μ t * ] 2 ( repressor ) . ( 8 ) As a function of kt* and K , the global minimum of Eq 7 occurs as kt* → ∞ and K → ∞ . The path of descent toward this minimum is given by differentiating with respect to K at fixed kt* and setting the result to zero , which yields the curve K = { 0 k t * < x * 2 k t * - x * 2 k t * ≥ x * 2 , ( 9 ) along which the variance satisfies σ t 2 x * t * 2 = { 1 k t * < x * 2 x * k t * ( 1 - x * 4 k t * ) k t * ≥ x * 2 , ( 10 ) where the first case comes from the fact that K must be nonnegative . In contrast , the global minimum of Eq 8 occurs at finite μt* and K: differentiating with respect to each and setting the results to zero gives the values K = e - 2 N , ( 11a ) μ t * = e 2 x * 2 N + 2 , ( 11b ) σ t 2 x * t * 2 = x * x * + 4 e - 2 N , ( 12 ) where we have assumed that K/N ≪ 1 ( see Materials and methods ) , which is justified post-hoc by Eq 11a . These analytic approximations are compared with the numerical results for the activator in Fig 2A ( white dashed line , Eq 9 ) and for the repressor in Fig 2B ( white circle , Eq 11 ) . In Fig 2A we see that the global minimum indeed occurs as kt* → ∞ and K → ∞ , and the predicted curve agrees well with the observed descent . In Fig 2B we see that the predicted global minimum lies very close to the observed global minimum . We have also checked along specific slices in the ( K , kt* ) or ( K , μt* ) plane and found that the analytic approximations generally differ from the numerical results by about 10% or less , despite the fact that the approximations take H → ∞ whereas the numerics in Fig 2 use H = 3 . The success of the approximations means that Eq 6 describes the key mechanism leading to the optimal temporal precision . Eq 6 demonstrates that the optimal regulatory strategy arises from a tradeoff between minimizing regulator and target noise . On the one hand , minimizing only the regulator noise would require that the regulator cross its threshold K with a steep slope d y ¯ / d t and therefore at an early time , meaning that the target molecule would be effectively unregulated and would increase linearly over time . On the other hand , minimizing only the target noise would require that the regulator cross its threshold only shortly before the target time t* , such that the target molecule would cross its threshold x* with a steep slope d x ¯ / d t , leading to a highly nonlinear increase of the target molecule over time . In actuality , the optimal strategy is somewhere in between , with the regulator crossing its threshold at some intermediate time t0 , and the target molecule exhibiting moderately nonlinear dynamics as in Fig 1C and 1D . Eqs 10 and 12 demonstrate that the timing variance is small for large kt*/x* in the case of activation , and small for large N/x* in the case of repression . This makes intuitive sense because each of these quantities scales with the number of regulator molecules: k is the production rate of activator molecules , while N is the initial number of repressor molecules . To make this intuition quantitative , we define a cost as the time-averaged number of regulator molecules , ⟨ a ⟩ = 1 t * ∫ 0 t * d t a ¯ ( t ) = 1 2 k t * , ( 13 ) ⟨ r ⟩ = 1 t * ∫ 0 t * d t r ¯ ( t ) = N μ t * ( 1 - e - μ t * ) , ( 14 ) where the second steps follow from Eqs 3 and 4 . We see that , indeed , 〈a〉 scales with k , and 〈r〉 scales with N . Thus , Eqs 10 and 12 demonstrate that increased temporal precision comes at a cost , in terms of the number of regulator molecules that must be produced . We test our model predictions using data from neuroblast cells in C . elegans larvae [5] . During C . elegans development , particular neuroblast cells migrate from the posterior to the anterior of the larva . It has been shown that the migration terminates not at a particular position , but rather after a particular amount of time , and that the termination time is controlled by a temporal increase in the expression of the mig-1 gene [5] . Since mig-1 is known to be subject to regulation [21] , we investigate the extent to which the dynamics of mig-1 can be explained by the predictions of our model . Fig 3A shows the number x of mig-1 mRNA molecules per cell as a function of time t , obtained by single-molecule fluorescent in situ hybridization ( from [5] ) . We analyze these data in the following way ( see Materials and methods for details ) . First , noting that the dynamics are nonlinear , we quantify the linearity using the area under the curve , normalized by that for a perfectly linear trajectory x*t*/2 , ρ = 2 x * t * ∫ 0 t * d t x ( t ) . ( 15 ) By this definition , ρ = 1 for perfectly linear dynamics , and ρ → 0 for maximally nonlinear dynamics ( a sharp rise at t* ) . Then , we estimate x* , t* , and the timing variance σ t 2 from the data . Specifically , migration is known to terminate between particular reference cells in the larva [5] , which gives an estimated range for the termination time t* . This range is shown in magenta in Fig 3A and corresponds to a threshold within the approximate range 10 ≤ x* ≤ 25 . Therefore , we divide the x axis into bins of size Δx , choose bin midpoints x* within this range , and for each choice compute the mean t* and the variance σ t 2 of the data in that bin . Fig 3B shows the average and standard deviation of results for different values of x* and Δx ( blue circle ) . The experimental data point in Fig 3B exhibits two clear features: ( i ) the dynamics are nonlinear ( ρ is significantly below 1 ) , and ( ii ) the timing variance is low ( σ t 2 x * / t * 2 is significantly below 1 ) . Neither feature can be explained by a model in which the production of x is unregulated , since that would correspond to a linear increase of molecule number over time ( ρ = 1 ) and a timing variance that satisfies σ t 2 x * / t * 2 = 1 ( square in Fig 3B ) . Furthermore , since auto-regulation has been shown to generically increase timing variance beyond the unregulated case [12] , it is unlikely that feature ( ii ) can be accounted for by a model with auto-regulation alone . Can these data be accounted for by our model with regulation ? To address this question we calculate ρ and σ t 2 x * / t * 2 from our model . For simplicity , we focus on the analytic approximations in Eqs 7 and 8 , since they have been validated in Fig 2 . In these approximations , since x ¯ ( t ) is piecewise-linear ( Eq 5 ) , calculating ρ via Eq 15 is straightforward: ρ = 1 − t0/t* , where t0 is once again determined by either a ¯ ( t 0 ) = K or r ¯ ( t 0 ) = K according to Eqs 3 and 4 . For a given ρ and cost 〈a〉/x* or 〈r〉/x* , we calculate the minimum timing variance σ t 2 x * / t * 2 . For the activator , we use the expression for ρ along with Eq 13 to write Eq 7 in terms of ρ and 〈a〉/x* , σ t 2 x * t * 2 = x * 2 ⟨ a ⟩ ( 1 - ρ ) + ρ 2 . ( 16 ) For the repressor , we use the expression for ρ along with Eq 14 to write Eq 8 in terms of ρ and 〈r〉/x* , and then minimize over N ( see Materials and methods ) to obtain σ t 2 x * t * 2 = e 3 27 x * ⟨ r ⟩ ( 1 - ρ ) 3 + ρ 2 . ( 17 ) Eqs 16 and 17 are shown in Fig 3B ( green solid and red dashed curves , respectively ) , and we see the same qualitative features for both cases: all curves satisfy σ t 2 x * / t * 2 = 1 at ρ = 1 , as expected; and as ρ decreases , each curve exhibits a minimum whose depth and location depend on cost . Specifically , as cost increases ( lighter shades of green or red ) , the variance decreases , as expected . Importantly , we see that at a cost on the order of 〈a〉/x* = 〈r〉/x* ∼ 10 , the model becomes consistent with the experimental data: both the low timing variance and the low degree of linearity predicted by either the activator or repressor case agree quantitatively with the experiment . This suggests that either an accumulating activator or a degrading repressor is sufficient to account for the temporal precision observed in mig-1-controlled neuroblast migration . Our minimal model neglects common features of gene expression such as bursts in molecule production [22] and additional sources of noise . Therefore we test the robustness of our findings to these effects here . First , we test the robustness of the results to the presence of bursts by replacing the Poisson process governing the activator production with a bursty production process . Specifically , we assume that each production event increases the activator molecule count by an integer in [1 , ∞ ) drawn from a geometric distribution with mean b [23 , 24] . The limiting case of b = 1 recovers the original Poisson process . The results are shown in Fig 4A for b = 1 , 3 , and 5 ( green solid , cyan dashed , and magenta dashed curves ) . We see that bursts in the activator increase the timing variance of the target molecule , as expected , but that there remain parameters for which the variance is less than that for the unregulated case , σ t 2 x * / t * 2 = 1 ( dashed black line ) . This result shows that even with bursts , regulation by an accumulating activator is beneficial for timing precision . We also recognize that whereas the activator can be assumed to start with exactly zero molecules , it is more realistic for the repressor to start with an initial number of molecules that has its own variability . We incorporate this additional variability into the model by performing stochastic simulations [25] of the reactions in Fig 1A and drawing the initial repressor molecule number from a Poisson distribution across simulations . The result is shown by the green dashed curve in Fig 4B . We see that the additional variability gives rise to an increase in the timing variance of the target molecule , as expected ( compare with the green solid curve ) . However , for most of the range of degradation rates , including the optimal degradation rate , the variance remains less than that of the unregulated case , σ t 2 x * / t * 2 = 1 ( dashed black line ) . This result indicates that the benefit of repression is robust to this additional source of noise . Then , we test the robustness of the results to our assumptions that the activator undergoes pure production and the repressor undergoes pure degradation . Specifically , we introduce a degradation rate μ for the activator , and a production rate k for the repressor , such that either regulator reaches a steady state of k/μ . The blue curves in Fig 4A and 4B show the case where the increasing activator’s steady state k/μ is twice its regulation threshold K , or the decreasing repressor’s steady state k/μ is half its regulation threshold K , respectively . In both cases , we see that the timing variance of the target molecule increases because the regulator slows down on the approach to its regulation threshold . Nonetheless , we see that it is still possible for the variance to be lower than that of the unregulated case . The red curves show the case where the regulator’s steady state is equal to its regulation threshold . Here we are approaching the regime in which threshold crossing is an exponentially rare event . As a result , the variance further increases , to the point where it is above that of the unregulated case for the full range of parameters shown . These results demonstrate that the benefit of regulation is robust to more complex regulator dynamics , but only if the regulator still crosses its regulation threshold at a reasonable mean velocity . Finally , we test the robustness of the results to a feature exhibited by the experimental mig-1 data: near the end of migration , cell division occurs ( Fig 3A , black data ) . One daughter cell continues migrating ( Fig 3A , dark blue data ) , while the other undergoes programmed cell death [5] . To investigate the effects of cell division , we perform stochastic simulations , and at a given time td we assume that the cell volume V is reduced by a factor of two . For each simulation , td is drawn from a Gaussian distribution with mean t ¯ d and variance σ d 2 determined by the data ( Fig 3A , black ) . At td , we reduce the molecule numbers of both the regulator and the target molecule assuming symmetric partitioning , such that the molecule number after division is drawn from a binomial distribution with total number of trials equal to the molecule number before division and success probability equal to one half . We also reduce the molecule number threshold K by a factor of two because it is proportional to the cell volume via K = KdV , where Kd is the dissociation constant . Fig 4C shows the dynamics of the mean molecule numbers of the activator ( green solid ) and its target ( blue solid ) , or the repressor ( red dashed ) and its target ( blue dashed ) . We see that the activator , repressor , and target drop in molecule number at division but that the abruptness of the drop is smoothed by the variability in the division time . The smoothing is more pronounced in the cases of the repressor and the target because the molecule numbers of these species are smaller at division . Thus , for either the activator or repressor mechanism , we see that the experimentally observed variability in division time is sufficient to smooth out the dynamics of the target molecule number , consistent with the experimental data in Fig 3A . Additionally , we see in Fig 4D that the timing variance of the target molecule in both the activator and repressor cases is similar to that without division in the region of the experimental division time . This further indicates that either model remains sufficient to account for the low experimental timing variance , even with the experimentally observed cell division . Taken together , the results in Fig 4C and 4D show that the key results of the model are robust to the effects of cell division .
We have demonstrated that regulation by an accumulating activator or a diminishing repressor increases the precision of threshold crossing by a target molecule , beyond the precision achievable with constitutive expression alone . The increase in precision results from a tradeoff between reducing the noise of the regulator and reducing the noise of the target molecule itself . Our minimal model is sufficient to account for both the high degree of nonlinearity and the low degree of noise in the dynamics of mig-1 in C . elegans neuroblasts , providing evidence for the hypothesis that these cells use regulated expression to terminate their migration with increased temporal precision . These results suggest that regulation by a dynamic upstream species is a simple and generic method of increasing the temporal precision of cellular behaviors governed by threshold-crossing events . Why does regulation increase temporal precision , whereas it has been shown that auto-regulation ( feedback ) does not [12] ? After all , either regulation or positive feedback can produce an acceleration in molecule number over time , leading to a steeper threshold crossing . The reason is likely that positive feedback relies on self-amplification . In addition to amplifying the mean , positive feedback also amplifies the noise . In contrast , regulation by an external species does not involve self-amplification . Therefore , the noise increase is not as strong . The target molecule certainly inherits noise from the regulator ( Eq 6 ) , but the increase in noise does not outweigh the benefit of the acceleration , as it does for feedback . Future work could investigate the interplay of regulation and feedback , as well as active degradation of the target molecule , especially as mig-1 is thought to be subject to feedback and degradation in addition to external regulation [5 , 21] . Our finding that regulation increases temporal precision is related to the more basic phenomenon that a sequence of ordered events has a lower relative timing error than a single event [19 , 26] . Specifically , if a single event occurs in a time that is exponentially distributed with a mean τ , then the relative timing error is σ/τ = 1 . For n such events that must occur in sequence , the total completion time follows a gamma distribution with relative timing error σ / τ = 1 / n , which decreases with increasing n . Thus , at a coarse-grained level , the addition of a regulator can be viewed as increasing the length of the sequence of threshold-crossing events from one to two , and thus decreasing the timing error . This perspective suggests that the timing error could be decreased even further via a cascade of regulators . Although we have demonstrated that our findings are robust to complexities such as bursts and cell division ( Fig 4 ) , our model neglects additional features of regulated gene expression such as transcriptional delay . Transcriptional delay has been shown to play an important role in regulation [27 , 28] and to have consequences for the mean and variance of threshold-crossing times [29] . If a delay were to arise due to a sequence of stochastic but irreversible steps , then we conjecture that the relative timing error would decrease with the number of these steps , due to the same cascading mechanism mentioned in the previous paragraph . However , it has been shown that if there are reversible steps or cycles within a multistep process , then the first passage time distribution can approach an exponential as the number of steps becomes large [26] . In this case the timing statistics would be captured by our simple model , which assumes single exponentially distributed waiting times . Future work could explore the effects of transcriptional delay in more detail . Finally , we have shown that the mig-1 data from migrating neuroblasts in C . elegans are quantitatively consistent with either the accumulating activator or diminishing repressor model , but the data do not distinguish between the two models . A direct approach to search for a distinction would be to use genetic knockout techniques to screen directly for regulators of mig-1 and their effects on its abundance . A less direct approach would be to more closely investigate the effects of the cell division that occurs during migration . For example , we assumed in this study that the volume fraction is equal to the average molecule number fraction in the surviving cell after division . However , if they were found to be unequal for either mig-1 or its regulator ( s ) , then the concentrations of these species could undergo an abrupt change after division , which may have opposing consequences for the activator vs . the repressor mechanism . Future studies could use these or related approaches to more concretely identify the role of gene regulation in achieving precise timing during cellular processes .
We compute the first-passage time statistics t ¯ and σ t 2 numerically from the master equation following [12] , generalized to a two-species system . Specifically , the probability F ( t ) that the molecule number crosses the threshold x* at time t is equal to the probability Py , x*−1 ( t ) that there are y regulator molecules ( where y ∈ {a , r} ) and x* − 1 target molecules , and that a production reaction occurs with rate f± ( y ) to bring the target molecule number up to x* . Since this event can occur for any regulator molecule number y , we sum over all y , F ( t ) = ∑ y = 0 Y f ± ( y ) P y , x * - 1 ( t ) , ( 18 ) where Y = {amax , N} . The repressor has a maximum number of molecules N , whereas the activator number is unbounded , and therefore we introduce the numerical cutoff a max = k t * + 10 k t * . The probability Pyx evolves in time according to the master equation corresponding to the reactions in Fig 1A , P ˙ a x = k P a - 1 , x + f + ( a ) P a , x - 1 - [ k + f + ( a ) ] P a x , ( 19a ) P ˙ r x = μ ( r + 1 ) P r + 1 , x + f - ( r ) P r , x - 1 - [ μ r + f - ( r ) ] P r x . ( 19b ) The moments of Eq 18 are ⟨ t m ⟩ = ∑ y = 0 Y f ± ( y ) ∫ 0 ∞ d t t m P y , x * - 1 ( t ) , ( 20 ) where t ¯ = 〈 t 〉 and σ t 2 = 〈 t 2 〉 - 〈 t 〉 2 . Therefore computing t ¯ and σ t 2 requires solving for the dynamics of Pyx . Because Eq 19 is linear in Pyx , it is straightforward to solve by matrix inversion . We rewrite Pyx as a vector by concatenating its columns , P → ⊤ = [ [ P 00 , … , P Y 0 ] , … , [ P 0 , x * - 1 , … , P Y , x * - 1 ] ] , such that Eq 19 becomes P → ˙ = M P → , where M = [ M ( 1 ) M ( 2 ) M ( 1 ) M ( 2 ) M ( 1 ) ⋱ ⋱ M ( 2 ) M ( 1 ) ] . ( 21 ) Here , for i , j ∈ {0 , … , Y} , the x* − 1 subdiagonal blocks are the diagonal matrix M i j ( 2 ) = f ± ( i ) δ i j , and the x* diagonal blocks are the subdiagonal matrix M i j ( 1 ) = - [ k ( 1 - δ i a max ) + f + ( i ) ] δ i j + k δ i - 1 , j or the superdiagonal matrix M i j ( 1 ) = - [ μ i + f - ( i ) ] δ i j + μ ( i + 1 ) δ i + 1 , j for the activator or repressor case , respectively . The δ i a max term prevents activator production beyond amax molecules . The final M ( 1 ) matrix in Eq 21 contains f± production terms that are not balanced by equal and opposite terms anywhere in their columns . These terms correspond to the transition from x* − 1 to x* target molecules , for which there is no reverse transition . Therefore , the state with x* target molecules ( and any number of regulator molecules ) is an absorbing state that is outside the state space of P → [12] . Consequently , probability leaks over time , and P → ( t → ∞ ) = ∅ → . Equivalently , the eigenvalues of M are negative . The solution of the dynamics above Eq 21 is P → ( t ) = exp ( M t ) P → ( 0 ) . Therefore , Eq 20 becomes 〈 t m 〉 = V → ⊤ [ ∫ 0 ∞ d t t m exp ( M t ) ] P → ( 0 ) , where V → ⊤ is a row vector of length x* ( Y + 1 ) consisting of [f± ( 0 ) , … , f± ( Y ) ] preceded by zeros . We solve this equation via integration by parts [12] , noting that the boundary terms vanish because the eigenvalues of M are negative , to obtain ⟨ t m ⟩ = ( - 1 ) m + 1 m ! V → ⊤ ( M - 1 ) m + 1 P → ( 0 ) . ( 22 ) We see that computing t ¯ = 〈 t 〉 and σ t 2 = 〈 t 2 〉 - 〈 t 〉 2 requires inverting M , which we do numerically in Matlab . We initialize P → as Pax ( 0 ) = δa0 δx0 or Prx ( 0 ) = δrN δx0 for the activator or repressor case , respectively . When including cell division , we compute t ¯ and σ t 2 from 50 , 000 stochastic simulations [25] . The dynamics of the mean regulator and target molecule numbers are obtained by calculating the first moments of Eq 19 , ∂ t y ¯ = ∑ y x y P ˙ y x and ∂ t x ¯ = ∑ y x x P ˙ y x , where y ∈ {a , r} . For the regulator we obtain ∂ t a ¯ = k or ∂ t r ¯ = - μ r ¯ in the activator or repressor case , respectively , which are solved by Eqs 3 and 4 . For the target molecule we obtain ∂ t x ¯ = 〈 f ± ( y ) 〉 , which is not solvable because f± is nonlinear ( i . e . , the moments do not close ) . A deterministic analysis conventionally assumes 〈 f ± ( y ) 〉 ≈ f ± ( y ¯ ) , for which the equation becomes solvable by separation of variables . For example , in the case of H = 1 , using Eqs 1–4 , we obtain x ¯ ( t ) = { α t - ( α K / k ) log [ ( k t + K ) / K ] ( activator ) ( α / μ ) log [ ( N + K e μ t ) / ( N + K ) ] ( repressor ) . ( 23 ) Eq 23 is plotted in Fig 1C and 1D . When including cell division , we compute the mean dynamics from the simulation trajectories ( Fig 4C ) . To find the global minimum of Eq 8 , we differentiate with respect to kt* and K and set the results to zero , giving two equations . kt* can be eliminated , leaving one equation for K , 1 2 log N K = 1 - K N ( 24 ) This equation is transcendental . However , in the limit K ≪ N , we neglect the last term , which gives Eq 11 . To derive Eq 17 , we use ρ = 1 - t 0 t * = 1 - log N / K μ t * ( 25 ) where the second step follows from r ¯ ( t 0 ) = K according to Eq 4; and , from Eq 14 , ⟨ r ⟩ = N μ t * ( 1 - e - μ t * ) ≈ N μ t * ( 26 ) where the second step assumes that the repressor is fast-decaying , μt* ≫ 1 . We use Eqs 26 and 25 to eliminate μt* and K from Eq 8 in favor of ρ and 〈r〉 , σ t 2 x * t * 2 ≈ x * ( e N ( 1 - ρ ) / ⟨ r ⟩ - 1 ) ⟨ r ⟩ 2 N 3 + ρ 2 . ( 27 ) For nonlinear dynamics ( ρ < 1 ) we may safely neglect the −1 in Eq 27 . Then , differentiating Eq 27 with respect to N and setting the result to zero , we obtain N = 3〈r〉/ ( 1 − ρ ) . Inserting this result into Eq 27 produces Eq 17 . To estimate the time at which migration terminates in Fig 3A , we refer to [5] . There , the position at which neuroblast migration terminates is measured with respect to seam cells V1 to V6 in the larva ( see Fig . 4D in [5] ) . In particular , in wild type larvae , migration terminates between V2 and the midpoint of V2 and V1 . This range corresponds to the magenta region in Fig 3A ( see Fig . 4B , upper left panel , in [5] ) . Under the assumptions of constant migration speed and equal distance between seam cells , the horizontal axis in Fig 3A represents time . To compute ρ for the experimental data in Fig 3A according to Eq 15 we use a trapezoidal sum . For the choices of x* and t* described in the text , this produces the ρ values in Fig 3B .
|
Cells control important processes with precise timing , even though their underlying molecular machinery is inherently imprecise . In the case of Caenorhabditis elegans development , migrating neuroblast cells produce a molecule until a certain abundance is reached , at which time the cells stop moving . Precise timing of this event is critical to C . elegans development , and here we investigate how it can be achieved . Specifically , we investigate regulation of the molecule production by either an accumulating activator or a diminishing repressor . Our results are consistent with the nonlinear increase and low noise of gene expression observed in the C . elegans cells .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"invertebrates",
"cell",
"motility",
"engineering",
"and",
"technology",
"caenorhabditis",
"cell",
"cycle",
"and",
"cell",
"division",
"signal",
"processing",
"cell",
"processes",
"animals",
"noise",
"reduction",
"animal",
"models",
"developmental",
"biology",
"regulator",
"genes",
"caenorhabditis",
"elegans",
"model",
"organisms",
"systems",
"science",
"mathematics",
"experimental",
"organism",
"systems",
"gene",
"types",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"gene",
"expression",
"life",
"cycles",
"nonlinear",
"dynamics",
"eukaryota",
"cell",
"biology",
"cell",
"migration",
"genetics",
"nematoda",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"larvae",
"organisms"
] |
2018
|
Temporal precision of regulated gene expression
|
Nutrient availability is an important environmental variable during development that has significant effects on the metabolism , health , and viability of an organism . To understand these interactions for the nutrient copper , we used a chemical genetic screen for zebrafish mutants sensitive to developmental copper deficiency . In this screen , we isolated two mutants that define subtleties of copper metabolism . The first contains a viable hypomorphic allele of atp7a and results in a loss of pigmentation when exposed to mild nutritional copper deficiency . This mutant displays incompletely penetrant skeletal defects affected by developmental copper availability . The second carries an inactivating mutation in the vacuolar ATPase that causes punctate melanocytes and embryonic lethality . This mutant , catastrophe , is sensitive to copper deprivation revealing overlap between ion metabolic pathways . Together , the two mutants illustrate the utility of chemical genetic screens in zebrafish to elucidate the interaction of nutrient availability and genetic polymorphisms in cellular metabolism .
Proper maternal nutrition is critical for early embryonic development . The Dutch Famine Study examined the consequences of nutrient deprivation on developmental outcome during severe food shortages near the end of the Second World War and clearly demonstrated that inadequate nutrient availability during human gestation increases the likelihood of developmental anomalies [1] . From these initial observations arose the well-recognized link between maternal folate supplementation and the suppression of neural tube defects [2] . Despite overwhelming epidemiologic data indicating the benefits of folate and other nutrient supplementation we do not fully understand the genetics of predisposition to these abnormal developmental phenotypes when faced with suboptimal nutrient levels . There are several large difficulties in the study of these processes in mammals that have prevented faster progress . The first is that the genetics of mammals has been cumbersome . The second , and more important , is that development of placental animals occurs in utero making rapid detection of developmental phenotypes difficult . Finally , controlling the level of nutrient available to the developing embryo cannot be done with precision as it depends both on the genetics of the mother and the embryo as well as maternal nutrition . Copper is an essential nutrient which when absent results in severe developmental abnormalities . This is most clearly illustrated by Menkes disease ( OMIM #309400 ) , a rare X-linked disorder of copper metabolism . Patients with Menkes disease have an array of symptoms including seizures , neurodegeneration , hypopigmentation , and lax skin which result from decreased copper incorporation into critical enzymes such as dopamine-β-hydroxylase and lysyl oxidase [3] , [4] . This usually fatal disease is caused by mutations in a copper transporter , ATP7A ( NM_000052 ) , which resides in the secretory pathway and is responsible for transport of copper into this compartment . The Menkes gene product is also responsible for placental copper transport . While patients complete in utero development apparently normally , it is clear from biochemical studies at birth that there are significant defects that arise from gestational copper deficiency [5] . In order to study the effects of developmental copper deprivation our lab has previously created a zebrafish model of severe copper deficiency [6] . High doses of the cell permeable copper chelator neocuproine cause embryonic zebrafish to exhibit a Menkes-like phenotype with neurodegeneration , hypopigmentation , and connective tissue defects . Isolation and cloning of the mutant calamity , which shared these same characteristics , revealed a loss-of-function mutation in the zebrafish orthologue of ATP7A ( NM_001042720 ) . In this current study we expand this model to study the effects of induced genetic alterations on the developmental response to mild copper deprivation . We describe two mutants sensitive to nutritional copper deficiency that illustrate the potential power of this approach to overcome the limitations of studying gene-nutrient interactions in vertebrate organisms and that define combinations of loss-of-function mutations of known ion homeostatic pathways that result in aberrant development .
In order to elucidate the molecular genetics of copper metabolism we performed a forward genetic screen for zebrafish mutants with enhanced sensitivity to subthreshold copper deficiency . To control copper levels zebrafish embryos were treated with the cell permeable copper specific chelator neocuproine which has been previously shown to cause a copper-deficient phenotype including loss of pigmentation and notochord defects at a dose of 1 to 10 µM due to loss of cuproenzyme activity [6] , [7] . Prior to screening , a subthreshold dose of 100 nM neocuproine was determined to cause no alteration in pigmentation in wild-type , haploid embryos . We then used this concentration of neocuproine to screen clutches of haploid embryos derived from F1 carriers of ENU-induced mutations . One half of each clutch was placed in 100 nM neocuproine at 3 hours post fertilization ( hpf ) and allowed to develop until 48 hpf when clutches were screened for loss of melanin pigmentation in 50% of the embryos ( Figure 1A ) . Only those clutches which had loss of pigmentation at 100 nM neocuproine but contained at least some pigmentation when untreated were scored as mutant . In this pilot screen we examined 700 F1 females and found five potential mutants . Seven hundred mutagenized haploid genomes at an estimated single locus mutation rate of 1 . 1×10−3 represents approximately a 65–70% coverage of the genome [8] . Of the five potential mutants , four were confirmed as true mutants as defined by the transmission of the neocuproine sensitive phenotype to the offspring . One of these mutants fit the “ideal” criteria ( no defect in vehicle and complete loss of pigment in 100 nM neocuproine in 50% of the haploid clutch ) as illustrated in Figure 1A and subsequent analysis revealed important insight into the intersection of genetics with sub-optimal copper nutrition in early development . A second mutant reveals a role for proton transport in copper metabolism . The final two mutants were similar in phenotype to the first but full analysis has not been completed . The first mutant isolated from the screen displayed normal melanin pigmentation when untreated but completely lost all melanin upon treatment with 100 nM neocuproine ( Figure 1B , C ) . Crossing this mutant with calamityvu69 ( cal ) which bears an inactivating mutation in the copper transport protein atp7a resulted in partial non-complementation . The compound heterozygote had no melanin in the developing retinal pigment epithelium ( RPE ) and normally distributed mild hypopigmentation over the rest of the body ( Figure 1D ) . Based on the partial non-complementation we tentatively assigned this mutant as an allele of calamity , designated gw71 . The second mutant has a phenotype that is independent of neocuproine . Named catastrophe , this mutant has normally distributed melanocytes that are small and punctate ( Figure 1E ) . Catastrophe ( cto ) is homozygous lethal at about 3 days post fertilization ( 3 dpf ) . The heterozygotes have no overt phenotype . In addition , cto homozygotes display sensitivity to copper deficiency by losing all melanin pigmentation in 100 nM neocuproine ( Figure 1F ) . Crossing cto with calvu69 results in complete complementation ( Figure 1G ) including the observation that the double heterozygote is not more sensitive to neocuproine than calvu69 heterozygotes ( data not shown ) . Thus , we continued our analysis on the basis that cto identifies a new locus involved in copper metabolism . Chromosomal localization using the early pressure parthenogenesis method [9] placed the mutation in calgw71 ( referred to below as gw71 ) near the centromere of chromosome 14 , the known location of atp7a . Combining this data with the partial non-complementation , we hypothesized that this mutant represented a hypomorphic allele of atp7a and confirmed this by direct sequencing of the mRNA . Mutant atp7a was cloned and displayed 100% identity with the published atp7a sequence ( NM_001042720 ) with the exception of a single base change present in both mutant clones , T3182G , which results in a single , non-conservative amino acid substitution , I1061S ( Figure S1A ) . This mutation is located in a region highly conserved in copper transporting ATPases and exchanges a hydrophobic amino acid for one that is polar and hydrophilic ( Figure 2A ) . This single amino acid change results in significant depletion of the full-length protein in mutant embryos ( Figure 2B ) . To verify that this was the causative mutation in gw71 , we performed an in vitro activity assay for the protein using wild-type and mutant atp7a . Fibroblasts from patients with Menkes disease which lack functional ATP7A were transfected with tyrosinase in combination with either wild-type or mutant zebrafish atp7a created via site-directed mutagenesis of the wild-type cDNA . These fibroblasts were then treated with increasing doses of neocuproine , fixed , and stained for tyrosinase activity using L-DOPA . Activity is dependent on both atp7a and tyrosinase cDNAs ( Figure S1B and S1C ) . In contrast to zebrafish mutant embryos , equal amounts of wild-type and mutant Atp7a were obtained via transfection in these fibroblasts ( Figure 2D ) . L-DOPA staining of cells expressing mutant cDNA was only mildly reduced when compared with wild-type ( Figure 2D vs . E ) indicating that the mutant retains some transport function . Overnight treatment with 25 nM neocuproine resulted in complete loss of tyrosinase activity in fibroblasts transfected with mutant , but not wild-type , atp7a though a reduction in staining was observed with wild-type ( Figure 2F , G ) . These data suggest that this single mutation in atp7a not only affects steady-state protein levels but is also capable of reducing the functional capacity of the protein , leading to sensitivity to copper deficiency . The I1061S mutation is located in the intracellular loop which comprises the ATPase domain of the transporter ( Figure 2H ) . Dimitriev et . al . have previously performed NMR spectroscopy on the homologous domain of the Wilson disease copper transporter , ATP7B , in the presence and absence of bound ATP and have derived from the resulting chemical shift data the residues important for ATP binding and hydrolysis [10] . We mapped the same region of Atp7a onto their model by sequence alignment ( 64% consensus , 49% identical ) to better understand the potential effect of this mutation on protein function . The mutation in calgw71 lies five amino acids away from a critical ATP binding residue , E1064 , which is highly conserved from yeast to humans ( Figure 2A and Figure S1D ) . While a mutation of a critical residue would be expected to significantly alter ATP binding or hydrolysis , a non-conservative mutation in the region of a critical residue might be expected to only slightly alter ATP binding/hydrolysis through minor shifts in regional structure . Because the gw71 allele is homozygous viable , we were able to examine several post-embryonic roles for atp7a . Adult homozygous mutant zebrafish placed in varying doses of neocuproine did not display an overt sensitivity phenotype ( data not shown ) . However , further study revealed a maternal effect of this mutation on embryonic copper metabolism . Homozygous mutant embryos derived from heterozygous females had a normal quantity and distribution of pigmentation that was partially sensitive to 100 nM neocuproine which abolished pigment in the retinal pigment epithelium ( RPE ) and reduced pigment over the body of the fish ( Figure 3A , B ) . In contrast , homozygous mutant embryos derived from homozygous mutant females had no pigment in the RPE and reduced pigment over the body; treatment of these embryos with 100 nM neocuproine completely abolished pigmentation throughout the embryo ( Figure 3C , D ) . The effect of the mother's genotype on the embryonic phenotype indicates that though not overt , the adult homozygous mutant does have defects in copper metabolism demonstrated by a nutrient-deficient state in the offspring . Thus the sensitivity of the embryo to neocuproine is due not only to aberrant embryonic copper metabolism , as the embryos from heterozygous mothers are sensitive to copper deficiency , but also to a deficient maternal loading of copper into the egg as the phenotype is exacerbated by maternal homozygosity . The importance of optimal copper nutrition during development is further illustrated by the presence of vertebral skeletal defects in homozygous mutants . Homozygous mutant embryos were stained at 21 dpf with alcian blue/alizarin red to reveal bone and cartilage respectively . These were compared with wild-type syngeneic age-matched controls raised in the same manner . The wild-type fish had straight vertebral columns along the entire length with long , straight bony processes extending from each vertebra ( Figure 3E ) . In contrast , homozygous gw71 fish displayed variable vertebral defects , most often a significant warping of the bony structures in the caudal-most region of the column caused by irregular length of vertebrae and defects in the joint angles ( Figure 3F ) . In addition the bony processes were also shortened and bent . Consistent with the observations in embryos that the mutation in gw71 brings the homozygous embryo close to , but not over , a threshold for copper deficiency , the persistent skeletal defects in the juvenile fish were not fully penetrant . Whereas wild-type fish had no vertebral defects ( n = 26 ) , a significant number ( 38% , n = 60 ) of the gw71 fish contained defects ( Figure 3G ) . Incomplete penetrance of the defect in the homozygous mutant fish could be attributed to either separate subtle genetic interactions or to variable nutrient availability . We hypothesized that if the penetrance of the defects were based on nutrient availability then reducing the nutrient levels would worsen the defects and increase the penetrance and vice versa . We thus took gw71 mutant embryos and placed them in either normal egg water or egg water supplemented with 100 nM neocuproine or 500 nM CuCl2 from 3 to 51 hpf ( 48 hour exposure ) . In addition , two separate groups of embryos were treated with neocuproine from 16 to 64 hpf and from 30 to 78 hpf to determine if there was a window of developmental time critical for the genesis of later defects . At 21 dpf the larvae were stained with alcian blue/alizarin red and scored for the presence or absence of vertebral defects ( Figure 3G ) . Untreated wild-type embryos ( not shown ) or wild-type embryos treated with 100 nM neocuproine from 3–51 hpf had no perceptible skeletal defects . Thirty-eight percent of gw71 embryos had skeletal defects and this number was not significantly affected by treatment with 100 nM neocuproine or 500 nM CuCl2 from 3–51 hpf . However , there was a 50% increase in the number of skeletal defects in gw71 embryos treated with 100 nM neocuproine from 16 to 64 hpf . The larvae treated with 100 nM neocuproine from 30 to 78 hpf died approximately 8 dpf from an unidentified cause . These results indicate an increasing sensitivity to mild copper deprivation as the embryo develops in the first 16–72 hrs . Further experimentation with smaller , more discrete treatment times might allow the determination of any developmental window required for the effects of copper on vertebral axis formation . In addition to the presence of vertebral skeletal defects in fully ossified skeletons , larvae at earlier stages of development displayed hyperossification of vertebrae adjacent to defects in the vertebral column ( Figure 3H , I ) . Normal zebrafish bone ossification begins rostrally and generally proceeds caudally with the exception of the caudal fin vertebrae [11] . In gw71 this pattern is maintained ( arrowhead in Figure 3H ) except for areas containing defects ( arrow in 3H ) . The defects affected the joints between vertebrae and had differing degrees of connective tissue bulges which partially stained with alcian blue indicating the presence of some cartilaginous tissue in these defects ( Figure 3H arrowhead ) . Before mapping the catastrophe mutant it was important to determine the extent of the defect in copper metabolism . The loss of pigmentation in the mutants could result from toxicity in a “two-hit” model whereby the mutation damages melanocytes and the drug acts to further affect these already sick cells . Therefore we examined the sensitivity of the mutant to another copper-dependent process–notochord formation . Notochord formation requires the action of the cuproenzyme lysyl oxidase and its family members . Both reduction in lysyl oxidase levels and copper chelation result in wavy , distorted notochords [12] . Placing cto mutants in 2 µM neocuproine at 3 hpf resulted in wavy notochords in the mutant embryos at 24 hpf while having no effect on heterozygous or wild-type embryos ( Figure 4A , B ) . This experiment indicates that the mutation in cto causes a global defect in copper metabolism and is not limited to melanocytes . The mutation in cto was localized to chromosome 7 and further mapping reduced the region of interest to an approximately 1 Mbp region between markers z21519 and z43308 ( Figure 4C ) . It was possible to assemble a nearly complete BAC contig between these markers using database BAC sequences ( www . sanger . ac . uk/Projects/D_rerio/ ) . This contig was scanned for potential genes using the FGENESH program ( www . softberry . com ) and comparing to the Ensembl database ( www . ensembl . org ) . A list of candidate genes was generated from this comparison . To further refine the list , a database of zebrafish insertional mutants was scanned for mutants displaying a similar melanocyte phenotype [13] . Approximately 6 mutants in this database had punctate melanocytes , 5 of which had insertions in genes encoding subunits of the vacuolar ( H+ ) ATPase ( Atp6 ) ( NM_199620 ) . As the critical region in cto contained the d subunit of the V0 complex of the vacuolar ATPase we cloned and sequenced this cDNA in the catastrophe mutants . A single base pair change C406T present in the mutant resulted in a premature stop codon , Q136X ( Figure S2 ) . Sequence alignment with the human sequences ( NM_004691 ) revealed a highly conserved protein sequence ( 94% identical ) that most closely aligned with the d1 subunit ( Figure 4D ) . Further database searches did not reveal a second d1 subunit in zebrafish . The significant identity between the human and zebrafish protein sequences allowed us to use an antibody directed against human ATP6V0D1 to examine the steady state levels of protein . We hypothesized that the early stop codon would result in a significant decrease in protein levels . Indeed , in 48 hpf embryos there is a near total reduction in Atp6v0d1 protein as compared with wild-type embryos ( Figure 4E ) . Total loss of this highly conserved and essential protein ( see below ) may be the cause of the catastrophe phenotype; however , there remains some possibility that another , tightly linked mutation may contribute to the observed phenotype . Based on significant experimentation in yeast a proposed quaternary structure for the vacuolar ATPase complex has emerged ( Figure 4F ) [14]–[18] . In this model , the two main subcomplexes , V0 and V1 have complementary functions of proton translocation and ATP hydrolysis respectively . The complexes are connected through several stalk subunits , v1d , v0d , and v1f ( not shown ) . Loss of these connecting subunits in yeast results in total loss of activity of the complex [19] . Thus in catastrophe , the loss of the v0d subunit would be predicted to result in complete loss of proton translocation throughout the embryo . If the defect in catastrophe is loss of Atp6 function the heterozygotes might be sensitive to pharmacologic inhibition of this transporter . Consistent with this , wild-type embryos placed at 24 hpf in 200 nM concanamycin A , a potent and specific inhibitor of Atp6 [20] , showed no apparent phenotype at 48 hpf ( Figure 5A ) . However , treatment of embryos heterozygous for cto resulted in punctate melanocytes and CNS degeneration , resembling the mutant ( Figure 5B ) . The mutants themselves appeared qualitatively worse , with further reductions in melanocyte pigmentation and worsening of the degenerative appearance ( Figure 5C ) . While it is apparent that loss of Atp6 results in altered cuproenzyme activity for two enzymes in the secretory pathway , it is unclear which step of global copper transport is affected in cto embryos . To address this we performed transplant experiments to determine the cell autonomy of the defect . Wild-type cells from actin::GFP transgenic zebrafish were transplanted into cto embryos and examined at 48 hpf for pigmented cells and GFP expression . Transplantation resulted in a few well-pigmented and stellate melanocytes over the head and body as well as clusters of pigmented retinal epithelial cells ( Figure 5D ) . These same embryos were mosaic for GFP expression ( Figure 5E ) . In body melanocytes the melanin obscured GFP fluorescence ( Figure 5E arrowhead ) . In contrast , the retinal pigment epithelial melanocytes display GFP fluorescence in the central area not covered by melanin ( Figure 5E arrow ) . From this we make two observations: First , the melanized melanocytes are derived from wild-type donor cells , and secondly , that nearby wild-type epidermal cells are not required for normal melanin pigmentation nor stellate appearance ( Figure 5E arrowhead ) . Thus copper metabolism must not be significantly disrupted on an organismal level , as these wild-type melanocytes in a mutant host still receive adequate copper for normal pigmentation . Also , the stellate appearance indicates that the defect that causes punctate pigment cells in cto is also cell-autonomous . The transplant experiment addresses delivery of copper to each cell , but the uptake or distribution of copper within the individual cell could also be affected in cto embryos . We hypothesized that disruption of the transporter responsible for secretory pathway acidification would result in defects in copper metabolism in this compartment . To test this we examined the sensitivity of cto embryos to partial loss of Atp7a through the use of a morpholino . Previous work from our laboratory has demonstrated that melanin synthesis following loss of Atp7a is also cell-autonomous in the melanocyte indicating that knock-down of Atp7a will allow interrogation of the pathway on a cellular rather than organismal level [6] . Injection of a splice morpholino previously shown to result in a copper deficient phenotype at a dose that does not cause pigmentation defects in wild-type or heterozygous embryos ( Figure 5F ) causes total loss of melanin pigmentation in cto embryos ( Figure 5G ) . Thus cto embryos are sensitive to loss of the secretory pathway copper transporter , Atp7a . Embryos heterozygous for the cto mutation did not show sensitivity to the Atp7a morpholino indicating that near complete loss of Atp6 activity is required to sensitize to alterations in copper metabolism . At the same time , the cytochrome oxidase activity of mitochondria derived from cto embryos is no different from wild-type indicating that copper delivery to mitochondria is normal and that the defect in copper metabolism in cto embryos is limited to the secretory compartment ( Figure 5H ) . The vacuolar ATPase has been implicated in diverse trafficking events within the cell and inhibition of this protein results in altered ion homeostasis , disrupted membrane trafficking , defective acid secretion , deficient protein degradation , and loss of protein sorting , endosomal recycling , and vesicular secretion [21]–[27] . To examine the effect of loss of this protein on cellular morphology , specifically melanocytes , we performed transmission electron microscopy focusing on the pigmented cells . Thin ( 500 nm ) plastic sections of 48 hpf embryos stained with toluidine blue did not demonstrate any further gross defects in organismal or cellular morphology beyond those observed in the pigmented cells both of the epidermis and the retinal pigment epithelium ( data not shown ) . Upon examination by electron microscopy in wild-type embryos both epidermal pigment cells as well as retinal pigment epithelial cells at 48 hpf display dark , uniformly round or ellipsoid melanosomes distributed throughout flat melanocytes ( Figure 6A–C ) . In contrast , the melanocytes of cto embryos are rounded and contain few fully melanized melanosomes , many large vacuolated structures and small vesicles surrounded by rings of melanin pigment ( Figure 6D–F ) . These latter structures have been identified as multi-vesicular bodies , the accumulation of which is reminiscent of early blocks in melanosome maturation found in the cappuccino , pallid , ruby-eye 2 , and reduced pigmentation mice which are all models of Hermansky-Pudlak syndrome and have specific early defects in melanosome biogenesis [28] . Thus among other abnormalities loss of proton transport results in early blocks in melanosome maturation . It is interesting to note that there remains active tyrosinase which produces some melanin in these aberrant structures despite the loss of the proton transporting ATPase ( Figure 6F ) .
In this work we have used the power of forward genetic screens combined with the ease of ex utero nutrient level manipulation accessible with the zebrafish to study the relationship between specific genetic alterations , the levels of the essential nutrient copper , and their combined effects on the developmental phenotype of the embryo . From these experiments we have derived a nutrient-sensitive allele of a known copper transporter that results in a juvenile skeletal phenotype . We have also implicated the vacuolar proton pump in vertebrate copper metabolism and interconnect two ion transport proteins whose individual effects on the other would not otherwise have been appreciated . The ex utero development of zebrafish provides an opportunity for manipulating the developmental levels of nutrients . Much success has been achieved in yeast using large libraries of compounds coupled with known deletion mutants to define the roles of many of the yeast proteins in cellular biology and metabolism [29] , [30] . One major advantage of yeast is the ability to absolutely control the levels of different nutrients and pharmacologic compounds and to screen large numbers at a time; however , yeast lack the complexity necessary to extend such findings to multi-cellular organisms and ultimately to understand human biology for the treatment of disease . Our work shows that the zebrafish model system can fill the niche in extending the principles of the chemical genetic screen to a vertebrate organism . Zebrafish retain the advantage of environmental exposure control while only slightly reducing the ability to screen large numbers . They also provide a system with more complex phenotypes to be examined which can then be brought back to the study of the underlying cell biology of a multi-cellular organism , particularly as the genome sequence and rapid mapping techniques improve . The first mutant which was isolated from our screen was a hypomorphic allele of atp7a . Animals bearing this allele have a normal pigmentation and notochord phenotype at 48 hpf but are sensitive to mild copper deficiency thus indicating that transporter function was impaired . This mutation reduced the protein levels to below the detection limits of our immunoblot demonstrating that only a fraction of wild-type protein expression is necessary to maintain a near-normal phenotype . This is consistent with our previous observations where very minor changes in Atp7a protein levels resulted in significant rescue of the calamity phenotype [31] . Also , the increase in severity of the calvu69 allele upon incubation with neocuproine demonstrates that even in this model of severe Menkes disease , there is still residual protein function without detectable expression [6] , [31] . Interestingly , when the gw71 protein was overexpressed in cell culture fibroblasts it was fully capable of loading copper into the secretory pathway as evidenced by the robust tyrosinase activity; yet , at the same time there was a clear sensitivity of this mutant transporter to copper levels . This mutant allele is not the first hypomorphic allele of atp7a . A less severe form of Menkes disease , Occipital Horn Syndrome , is also caused by mutations in atp7a . Children with this disease have many clinical problems similar to Menkes disease; however , as this syndrome is not fatal in early life other abnormalities can be appreciated including skeletal defects such as deforming hyperostosis and kyphoscoliosis [32] . In this context the gw71 mutant provides several important advances . First , within the screen itself it provides proof-of-concept that the screen design will result in the identification of critical proteins involved in copper transport and metabolism . Second , the gw71 allele is both viable and fertile which itself provides distinct advantages . Third , this allele demonstrates that only a fraction of wild-type levels of Atp7a protein are required for near-normal pigmentation and notochord formation , a result suggested by previous experiments [31] . Fourth , this mutant expands the hierarchy of copper metabolism previously described [6] . The differential effect on retinal pigment epithelial melanin versus the body pigmentation seen under a variety of genetic and environmental manipulations ( Compare Figures 1D , 3B , and 3D ) demonstrates an increased sensitivity of the RPE to derangements of copper metabolism . Fifth , the gw71 mutant displays an incompletely penetrant developmental hyperostosis phenotype which is easily detected . The proximal etiology of these defects is unknown . It may be related to lysyl oxidase activity which is important for zebrafish notochord development and is sensitive to nutritional copper status [12] . The increase in penetrance with copper chelation suggests that the variability may be due to nutritional differences . The lack of rescue observed with copper supplementation could be due to an inability of this ion to be translocated by the mutant Atp7a protein to the proper compartment . Alternatively , lack of rescue with copper could point to residual genetic heterogeneity leading to phenotypic differences . Whichever is the case , this aspect of the mutant phenotype may provide a model to further our understanding of this poorly understood defect . The viability of this mutant would allow a modifier screen to find mutations responsible for different aspects of the copper deficient phenotype as well as to detect any genetic variability leading to the incomplete penetrance observed in the mutant . Our second mutant contains an inactivating mutation in the vacuolar ( H+ ) ATPase subunit , Atp6v0d1 . While abolition of this protein results in loss of proton transport into the secretory pathway , the embryo is capable of developing relatively normally to about 48 hpf when defects become visibly apparent . This lag is most likely due to the persistence of maternal protein and mRNA . At this time point the changes in melanin pigmentation patterns signal the visible presence of defects in proton transport . Grossly the melanocytes become punctate which , upon ultrastructure analysis , is shown to be a loss of mature melanosomes and a rounding of the cell body with vacuolization . The observed relationship between lack of melanosome formation and cellular morphology is not understood but may suggest a toxic effect of inappropriate melanization in the multi-vesicular bodies seen with electron microscopy or may be due to a particular sensitivity of melanocytes to loss of proton transport . As it has been shown that the vacuolar ATPase is important for vesicular trafficking and endocytosis [21] , [22] , the distinct disruption of planar morphology in cto melanocytes may also be due to defects in these processes . The sensitivity to copper deficiency of the remaining melanin implicates proton transport in the homeostasis of copper metabolism . That the notochord is equally sensitive to reduced copper demonstrates that the defect is not limited to the melanocyte , but rather that there is a universal decrease in the ability of copper to adequately reach secretory cuproenzymes . Since the effect on copper metabolism in cto mutants is only revealed in the context of sub-threshold copper nutrition , without a screen of this nature , this inter-relationship of two ion transport pathways in the vertebrate organism would never have been appreciated . There are two models which could explain the defect in cuproenzyme function when proton transport is compromised . The first is that an acidic pH is important for copper incorporation into the nascent cuproproteins within the secretory pathway . The second model is that a proton gradient is required for copper transport , to balance the charge transfer across the vesicular membrane . These models are not mutually exclusive and a combination of the two could result in the final phenotype . The data presented in this paper demonstrate the power of the zebrafish model system to examine gene-nutrient interactions as well as to delineate basic cell biologic pathways . Continuing with this methodology will provide more insight into the biology of copper metabolism in a vertebrate organism . It is easy to see how screens in zebrafish similar to the one we describe have the potential to investigate the genetics of not only copper or folate metabolism , but also that this approach could be easily extended to an array of other nutrients .
Zebrafish were maintained in the Washington University Department of Pediatrics zebrafish facility according to institutional guidelines supervised by the Division of Comparative Medicine . The specific alterations of these well-characterized techniques are available in Text S1 . Mutant 48 hpf embryos were identified phenotypically . Twenty to thirty embryos were manually dechorionated and de-yolked , lysed in 75 µL RIPA buffer containing 10 µL/mL Protease Inhibitor Cocktail III ( Calbiochem ) . Unlysed material was removed by centrifugation at 1000× g for 5 minutes . For Atp7a , 50–100 µg of lysate in Laemmli buffer with 10% β-mercaptoethanol , heating for 5 min at 65°C ( not fully reducing conditions ) was loaded on a 6% SDS-polyacrylamide gel . The protein was transferred to nitrocellulose and blotted for Atp7a using a custom polyclonal antibody raised against a C-terminal peptide [31] . For Atp6v0d1 , 30–40 µg lysate in Laemmli buffer with 10% β-ME heated to 70°C for 5 minutes was loaded on a 12% SDS-polyacrylamide gel . The transferred protein was blotted for Atp6v0d1 using a mouse polyclonal raised to human recombinant protein at 1∶1000 dilution ( Abnova Corp ) . Other antibodies: Actin ( Sigma ) 1∶5000 , β-catenin ( BD Biosciences ) 1∶1000 . The Menkes patient fibroblast cell line Me344 ( gift of Mick Petris ) was maintained in 10% FBS/DMEM with Pen/Strep/Glut . Transfections were carried out on coverslips using Lipofectamine 2000 ( Invitrogen ) at a ratio of Lipo2k∶DNA of 2 . 5 for 3 hours in Optimem ( Invitrogen ) . The media was then replaced with 1% FBS/DMEM/PSG . Neocuproine was added in DMSO to the indicated concentration and the cells incubated overnight . Performed as previously described [33] . Twenty-one dpf juvenile zebrafish were fixed overnight in 4% PFA in PBS and stained as previously described [34] . Approximately 50–100 cells were extracted from wild-type ( AB ) embryos at the 1000 cell stage and placed in mutant embryos of the same age using a micromanipulator syringe and glass needle as described previously [35] . The atp7a splice morpholino e7 ( TGACAACATTAACATTCATACCCTG ) [31] was injected at a dose of 965 pg/embryo at the 1 cell stage in 10% phenol red . At 48 hpf the injected embryos were scored for pigmentation and genotyped . A crude mitochondrial fraction was prepared from groups of 45 embryos at 52 hpf by homogenizing in 250 mM sucrose , 10 mM Tris pH 7 . 4 with a loose-fitting glass-glass tissue homogenizer . The homogenate was spun at 700× g for 10 minutes . The supernatant from this spin was centrifuged at 23 , 000× g for 20 min to form a pellet containing mitochondria and large vesicles . The pellet was resuspended in 150 µL of sucrose buffer with protease inhibitors and n-dodecyl-3-D-maltoside was added to 1 mM and incubated for 10 minutes at 25°C . Cytochrome c oxidase activity was monitored by measuring the decrease in absorption of ferrocytochrome c at 550 nm using the protocol described for the Cytocox assay kit ( Sigma , USA ) . Performed as described previously [12] .
|
Copper is an essential nutrient required for multiple biologic functions . Proper uptake , transport , and excretion of copper are critical for use of this metal while reducing its inherent toxicity . While several key proteins involved in this process have been identified , there are still gaps in our understanding of copper metabolism—particularly during early development . We have used zebrafish , a genetically useful animal model system , to study genetic interactions with copper deficiency during development . We treated mutant embryonic zebrafish with a chelator that reduces the level of available copper and screened for mutants that displayed a copper deficient phenotype only in the presence of the chelator . We identified and characterized two mutants that advance our understanding of copper metabolism during the early periods of development , as well as show an interaction between copper metabolism and another fundamental pathway—that of proton transport . Our results expand our knowledge of copper metabolism and illustrate the power of this type of genetic screen in zebrafish to elucidate mechanisms of nutrient metabolism .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/animal",
"genetics",
"genetics",
"and",
"genomics/gene",
"discovery",
"nutrition",
"genetics",
"and",
"genomics/disease",
"models",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"genetics",
"and",
"genomics/gene",
"function",
"neurological",
"disorders/developmental",
"and",
"pediatric",
"neurology",
"nutrition/deficiencies",
"cell",
"biology/chemical",
"biology",
"of",
"the",
"cell"
] |
2008
|
Zebrafish Mutants calamity and catastrophe Define Critical Pathways of Gene–Nutrient Interactions in Developmental Copper Metabolism
|
Intestinal homeostasis requires precise control of intestinal stem cell ( ISC ) proliferation . In Drosophila , this control declines with age largely due to chronic activation of stress signaling and associated chronic inflammatory conditions . An important contributor to this condition is the age-associated increase in endoplasmic reticulum ( ER ) stress . Here we show that the PKR-like ER kinase ( PERK ) integrates both cell-autonomous and non-autonomous ER stress stimuli to induce ISC proliferation . In addition to responding to cell-intrinsic ER stress , PERK is also specifically activated in ISCs by JAK/Stat signaling in response to ER stress in neighboring cells . The activation of PERK is required for homeostatic regeneration , as well as for acute regenerative responses , yet the chronic engagement of this response becomes deleterious in aging flies . Accordingly , knocking down PERK in ISCs is sufficient to promote intestinal homeostasis and extend lifespan . Our studies highlight the significance of the PERK branch of the unfolded protein response of the ER ( UPRER ) in intestinal homeostasis and provide a viable strategy to improve organismal health- and lifespan .
Three highly conserved UPRER sensors coordinate the cell-autonomous response to ER stress: PERK , the transcription factor ATF6 , and the endoribonuclease IRE1 ( Fig 1B ) [4–10 , 13] . IRE1 promotes splicing of the mRNA encoding the transcription factor Xbp1 , PERK phosphorylates and inhibits the translation initiation factor 2 alpha ( eIF2α ) [11 , 14 , 15] , and ER stress-induced cleavage of ATF6 promotes its nuclear translocation and activation of stress response genes , including Xbp1 [16] . The activation of Xbp1 and ATF6 results in transcriptional induction of ER chaperones , of genes encoding ER components , and of factors required to degrade un/misfolded proteins through ER-associated degradation ( ERAD ) , thus enhancing ER folding capacity and proteostatic tolerance [17–19] . Phosphorylation of eIF2α , in turn , results in a broad , but selective decrease in protein translation , reducing the protein load in the ER , but also allowing selective translation of transcripts that contain alternative upstream open reading frames ( uORFs ) , including the transcription factor ATF4 . ATF4 target genes promote ER stress tolerance and boost antioxidant defenses [20–22] . PERK further phosphorylates and activates Nrf2 ( nuclear factor-erythroid-derived 2 ( NF-E2 ) -related factor 2 ) , a central regulator of anti-oxidant gene expression [23 , 24] . Studies in worms have shown that , in addition to these cell-autonomous responses to ER stress , local activation of the UPRER can trigger UPRER responses in distant tissues , indicating that endocrine processes exist that coordinate such stress responses across cells and tissues [25–28] . The mechanism ( s ) regulating and mediating these non-autonomous responses remain elusive . By regulating eIF2α , ATF4 and Nrf2 , PERK activation integrates the response to both protein misfolding in the ER and to misfolding-associated oxidative stress . Accumulation of un/misfolded proteins in the ER results in the production of reactive oxygen species ( ROS ) , most likely due to the generation of hydrogen peroxide as a byproduct of protein disulfide bond formation by protein disulfide isomerase ( PDI ) and ER oxidoreductin 1 ( Ero1 ) [29–31] . The coordinated control of cellular protein and redox homeostasis by the UPRER and other stress signaling pathways is likely critical to maintain SC function , as the intracellular redox state significantly impacts SC pluripotency , proliferative activity , and differentiation [32–35] . We have recently shown that this coordination is achieved in Drosophila ISCs by integration of Nrf2/CncC-mediated responses and Xbp1-mediated ER stress responses [12] . The fly orthologue of Nrf2 , CncC , counteracts intracellular oxidants and limits proliferative activity of ISCs [34] . In ISCs , CncC is inhibited in response to high ER stress ( as in Xbp1 loss-of-function conditions ) , resulting in increased oxidative stress and activation of ISC proliferation [12 , 34] . The Drosophila ISC lineage exhibits a high degree of functional and morphological similarities with the ISC lineage in the mammalian small intestine [36–38] . ISCs self-renew and give rise to transient , non-dividing progenitor cells called EnteroBlasts ( EBs ) that are lineage-restricted ( by Robo/Slit signaling and differential Notch signaling ) to differentiate into either absorptive EnteroCytes ( ECs ) or secretory EnteroEndocrine ( EEs ) cells [36 , 37 , 39] . ISCs are the only dividing cells in the posterior midgut of Drosophila and their entry into a highly proliferative state is regulated by multiple stress and mitogenic signaling pathways , including Jun-N-terminal Kinase ( JNK ) , Jak/Stat , Insulin , Wnt , and EGFR signaling [38 , 40] . During aging , flies develop epithelial dysplasia in the intestine , caused by excessive ISC proliferation and deficient differentiation of EBs [41 , 42] . This phenotype is a consequence of an inflammatory condition initiated by immune senescence and dysbiosis of the commensal bacteria , and causes metabolic decline , loss of epithelial barrier function , and increased mortality [43–45] , and is associated with a strong tissue-wide increase in ER stress [12] . Increasing ER proteostasis in ISCs ( by over-expressing Xbp1 or the ERAD-associated factor Hrd1 ) prevents the age-related over-proliferation of ISCs , suggesting that limiting ER stress-associated signaling in ISCs may be beneficial for tissue homeostasis [12] . Here , we have tested this hypothesis . We have explored the regulation of ISC proliferation by cell-autonomous and non-autonomous UPRER responses in detail , and have assessed the consequences of limiting ER stress responses in ISCs for longevity . By analyzing loss of function conditions for Ero1L we find that the induction of ISC proliferation by ER stress can be uncoupled from the production of ROS , but that ISC-specific activation of PERK is critical for the proliferative response . Interestingly , PERK activation in ISCs is triggered both by ER stress within ISCs and non-autonomously by ER stress in other cells of the intestinal epithelium , which activate PERK in ISCs through the secretion of Unpaired ligands and activation of JAK/Stat signaling in ISCs . PERK thus integrates epithelial stress responses to control ISC proliferation under challenging proteostatic conditions . Strikingly , PERK is also essential for normal cell proliferation in the ISC lineage , and excessive or chronic PERK activity in ISCs is a cause for the development of epithelial dysplasia in aging flies . Accordingly , we demonstrate that limiting PERK expression in ISCs is sufficient to extend lifespan .
In a recent study we have shown that the control of ER proteostasis in ISCs by Xbp1 and Hrd1 ( a key component of the ER-associated degradation pathway ) is both sufficient and required to limit ISC proliferation . In Xbp1 or Hrd1 loss of function conditions , ER stress is associated with increased cellular ROS , and since ISC proliferation is stimulated by ROS [12 , 34 , 46] , these results suggested that ROS production by the ER plays a critical role in the regulation of ISC proliferation during ER stress . To test this notion further , we sought to uncouple ER stress from ROS production and disrupt ER homeostasis in ISCs by means that would not result in increased ROS production . Redox homeostasis of the ER is controlled by enzymes that promote disulfide bond formation and thus act as electron acceptors ( including protein disulfide isomerase; PDI ) , and by ER oxidoreductins ( including Ero1L ) that transfer electrons from such enzymes to water , generating H2O2 [47] . Accordingly , in Ero1L loss of function conditions protein folding in the ER is perturbed , while the generation of H2O2 is reduced [47] . This provides a genetic condition in which to test whether the proliferative activity of ISCs can be influenced by ER stress in the absence of ROS production . RNAi-mediated knockdown of Ero1L in ISCs and EBs ( using the ISC/EB driver esg::Gal4 ) , or in ISCs only ( combining esg::Gal4 with EB-specific inhibition of Gal4 by Su ( H ) Gbe-mediated expression of Gal80 [12] ) , resulted in a significant increase in ISC proliferation , confirming that ER stress promotes ISC activity ( Fig 1C ) . As expected , this condition did not increase ROS levels in ISCs , as measured by Dihydroethidium ( DHE ) fluorescence ( Fig 1C and 1D ) [12 , 34] . Knocking down an unrelated ROS generating enzyme expressed in the gut , Duox , in ISCs/EBs did not induce ISC proliferation ( S1C Fig ) . Conversely , over-expressing Ero1L exclusively in ISCs was sufficient to limit ISC proliferation in animals exposed to the ER stress inducer Tunicamycin ( TM , which inhibits N-linked protein glycosylation and folding ) ( Fig 1E ) . We confirmed the effect of Ero1L on ISC proliferation by generating MARCM clones from ISCs homozygous for the Ero1L loss of function allele ero1l335qrs , or expressing Ero1LRNAi . Ero1L-deficient clones grew much faster than wild-type clones ( Fig 1F ) , but this increase was accompanied by an accumulation of small , diploid , Dl positive cells , indicating that loss of Ero1L disrupts Notch-mediated differentiation of EBs ( S1A Fig ) . Loss of ER homeostasis disrupts Notch signaling by preventing proper processing of the Notch receptor [48 , 49] , and loss of Notch in EBs results in the formation of ISC ‘tumors’ consisting of symmetrically dividing , diploid Dl+ cells [36–38] . To confirm that ER stress promotes ISC proliferation in Ero1L loss of function conditions , we over-expressed spliced Xbp1 or CncC ( both molecules improve ER homeostasis and influence ISC proliferation [12] ) . Indeed , this significantly limits over-proliferation of Ero1L-deficient ISCs ( S1B Fig ) . While these results demonstrated that ER stress can induce ISC proliferation independently of ROS production , it remained unclear whether mitotic activity of ISCs was directly stimulated by ER stress , or whether the increased number of mitotic ISCs in Ero1L loss of function conditions was a consequence of an increased rate of symmetric divisions due to the deficiency in Dl/N signaling . We therefore explored the activation of ER stress signaling pathways in Ero1L-deficient ISCs , aiming to identify potential signals that control mitotic activity . Loss of Ero1L resulted in increased eIF2α phosphorylation ( Fig 1G ) , but did not increase Xbp1 expression ( S1D Fig ) , suggesting that the PERK branch of the UPRER is selectively engaged . Knocking down PERK was sufficient to prevent eIF2α phosphorylation in Ero1L deficient ISCs ( Fig 1H ) , confirming the specific requirement for PERK for this signal . Loss of PERK also prevented ISC proliferation in Ero1L deficient ISCs ( Fig 1H and 1I ) , suggesting that PERK activity may promote ISC proliferation in response to ER protein stress independently of ROS production . To test this idea , we decided to explore the regulation of PERK activation in ISCs , and the role of PERK in the control of ISC proliferation in more detail . In a recent study we have shown that ISC-specific perturbation of ER proteostasis ( by Xbp1 or Hrd1 knock-down ) increases phosphorylation of eIF2α in ISCs [12] ( S2A Fig ) . Increased eIF2α phosphorylation is also observed in ISCs/EBs after feeding flies the ER stress inducer Tunicamycin ( TM , inhibiting N-linked protein glycosylation and folding ) ( Fig 2A ) , which also robustly induces ISC proliferation [12] . eIF2α phosphorylation in ISCs of TM treated flies is associated with increased Xbp1 splicing , a separate marker for ER stress , as determined by the expression of an Xbp1::GFP splicing reporter [50 , 51] ( Fig 2B ) . Strikingly , phospho-eIF2α ( peIF2α ) increased primarily in the progenitor cell population of flies exposed to Tunicamycin , suggesting that the PERK branch of the UPRER is activated specifically in ISCs and EBs even when ER stress is induced in a tissue-wide manner , and indicating that PERK activation has a specific role within ISCs in the regulation of the regenerative response to ER stress ( Fig 2A and 2B ) . To test this idea , we assessed ISC proliferation in conditions in which PERK was knocked down in ISCs specifically ( Fig 2C and 2D , RNAi was enhanced by co-expression of Dicer2 ) . Knockdown of PERK in ISCs and EBs , or in ISCs specifically , was sufficient to inhibit TM-induced ISC proliferation ( Fig 2C and 2D ) and prevent TM-induced phosphorylation of eIF2α in ISCs ( Figs 2C , 2D , 2E and S4A ) , confirming that PERK is required both for the phosphorylation of eIF2α and for the induction of ISC proliferation in these conditions . Knockdown of PERK in ECs , on the other hand , did not inhibit ISC proliferation , but stimulated their proliferation , similar to knockdown of ATF6 or Ire1 ( S4B Fig ) . We confirmed the role of PERK in ISC proliferation by assessing the growth of ISC lineages that were deficient in PERK , using mosaic analysis with a repressible cell marker ( MARCM ) to generate either ISCs homozygous for the PERK insertion allele PERKe01744 , or ISCs expressing dsRNA against PERK . In both conditions , clone growth was significantly delayed compared to wild-type controls , indicating that PERK activity is not only required for stress-induced ISC proliferation , but also for ISC proliferation during homeostatic regeneration ( Fig 2F ) . The impaired ISC proliferation in PERK loss of function conditions contrasts with the induction of ISC proliferation in Xbp1 or Hrd1 loss of function conditions [12] , indicating that the primary function of PERK in ISCs , rather than promoting ER proteostasis ( and thus limiting ISC proliferation ) is to serve as a sensor of ER stress and an inducer of ISC proliferation . Accordingly , knockdown of PERK was also sufficient to limit ISC proliferation in other mitogenic conditions , such as proliferation induced either by over-expressing the JNK kinase Hemipterous ( Hep ) or by knocking down Notch ( S2E and S2F Fig ) . Due to this unexpected function of PERK , and to characterize the ISC-specific PERK response , we decided to explore the transcriptome changes induced in ISCs by PERK using RNAseq . We isolated ISCs by FACS from wild-type guts expressing only YFP in ISCs ( under the control of esg::Gal4 combined with Su ( H ) Gbe::Gal80 ) or from guts expressing YFP and PERKRNAi using described protocols [52] . In both genotypes , we compared the transcriptomes of cells isolated from mock or TM treated flies ( after 24 hours of TM exposure ) using standard RNAseq procedures ( Illumina MiSeq , [45] ) ( S1 Table ) . As expected , TM treatment resulted in a significant induction of genes involved in cell cycle , mitosis and DNA replication ( as well as of genes involved in antioxidant and detoxification responses ) in wild-type ISCs . In PERK-deficient ISCs , however , the induction of the vast majority ( 89% ) of these genes was strongly reduced ( Fig 3C–3E and S1 Table ) . Our results indicated that ISCs initiate a regenerative response to both cell-autonomous as well as tissue-wide ER stress by activating PERK . We confirmed the notion of a non-autonomous control of PERK activity in ISCs by assessing the phosphorylation of eIF2α in ISCs of animals in which ER stress was induced specifically in EBs , ECs , fat body , or muscle . To perturb ER proteostasis in these cells and tissues , we knocked down Xbp1 ( using Su ( H ) Gbe::Gal4 , tub::G80ts for EBs , NP1::Gal4 , tub::Gal80ts for ECs , ppl::Gal4 , tub::G80ts for fat body , and How::Gal4 , tub::G80ts for muscle ) . Knockdown of Xbp1 in EBs or ECs results in non-autonomous activation of ISC proliferation ( Wang et al . , 2014 ) , and , consistent with an ISC-specific activation of PERK , also resulted in increased phosphorylation of eIF2α in ISCs ( Figs 4A , 4B and S2B , note eIF2α phosphorylation in ISCs neighboring GFP—expressing EBs in Fig 4A , or in Dl+ ISCs in Fig 4B ) . However , knockdown of Xbp1 in fat body or muscle did not increase ISC proliferation or stimulate eIF2α phosphorylation in ISCs ( S2C and S2D Fig ) , suggesting that the non-autonomous regulation of PERK is limited to cell/cell interactions within the intestinal epithelium . Increased ER stress in intestinal epithelial cells has been associated with intestinal inflammation in vertebrates [4 , 6 , 7 , 10] . In flies , damage or stress in ECs promotes compensatory ISC proliferation by inducing inflammatory cytokines of the IL6 family , Unpaired 1–3 ( Upd 1–3 ) . These cytokines are induced in response to JNK activation in ECs and are secreted to activate JAK/Stat signaling in ISCs and in the surrounding visceral muscle [38 , 53 , 54] . To explore whether this compensatory proliferation program is involved in the non-autonomous regulation of ISC proliferation by ER stress , we asked whether loss of Xbp1 in EBs or ECs might influence ISC proliferation and ISC-specific PERK activity by stimulating JAK/Stat signaling . Consistent with a role for stress-induced Upd expression in the non-autonomous response to ER stress , loss of Xbp1 in ECs activates the Jak/Stat signaling pathway in muscle and epithelial cells of the gut ( as determined using a reporter for Stat activity; 2XSTAT::GFP , [55] Fig 4B ) . Accordingly , knockdown of the JAK/Stat receptor Domeless , the JAK kinase Hop , or of Stat ( but not of the JAK/Stat ligands Upd , Upd2 , or Upd3 ) specifically in ISCs alleviated TM-induced ISC proliferation , and prevented phosphorylation of eIF2α in ISCs ( Figs 4C and 4D and S4C–S4G ) . The induction of ISC proliferation by loss of Xbp1 in EBs could further be inhibited by knocking down the Drosophila JNK Basket ( Bsk ) or Upd3 ( Figs 4E and S4H ) , confirming that JNK activation and Upd3 induction in EBs are required for the non-autonomous activation of ISC proliferation by ER stress in these cells . To confirm that induction of Upd/JAK/Stat signaling is sufficient to activate PERK in ISCs , we determined the phosphorylation of eIF2α in ISCs over-expressing the activated form of the JAK Kinase Hopscotch ( HopTumL ) . When HopTumL was over-expressed in ISCs only ( using esg::Gal4 combined with Su ( H ) ::Gal80 ) , eIF2α phosphorylation increased in labeled cells ( Fig 4F , note that HopTumL induces the formation of clusters of labeled cells , suggesting an accumulation of ISCs and EB-like cells under these conditions ) . eIF2α phosphorylation also increased in ISCs when HopTumL was induced in ECs ( using NP1::Gal4 , tub::Gal80ts ) , suggesting a non-autonomous response of ISCs to JAK/Stat activation in neighboring ECs ( S5B Fig ) . Similarly , over-expression of individual Upds ( Upd , Upd2 , Upd3 ) from ECs ( using NP1::Gal4 , tub::Gal80ts ) stimulated eIF2α phosphorylation specifically in ISCs ( Figs 4F and S5A ) . Taken together , these results provide a model for the non-autonomous control of ISC proliferation in response to ER stress in ECs: JNK-mediated induction of Upds from stressed ECs activates PERK via JAK/Stat signaling in ISCs , triggering ISC proliferation . The UPRER is broadly activated in the aging intestinal epithelium , and is associated with the development of age-associated dysplasia [12] . To assess the role of PERK activation in age-related ISC over-proliferation , we assessed the phosphorylation of eIF2α in the intestine of aging flies . Similar to the ISC-specific activation of PERK we observed in response to TM treatment , eIF2α phosphorylation was increased in aging intestines in an ISC-specific manner ( Fig 5A ) . This activation was due to ER stress , as promoting ER homeostasis by over-expressing Xbp1 or Hrd1 ( which maintains intestinal homeostasis by limiting age-associated ISC proliferation , [12] ) , was sufficient to limit the age-associated increase in eIF2α phosphorylation in ISCs/EBs ( Fig 5B ) . Since promoting proliferative homeostasis of the intestinal epithelium extends lifespan of flies ( Biteau et al , 2010 ) , we tested whether reducing PERK expression in ISCs was sufficient to extend lifespan . We used the RU486-inducible ISC/EB-specific 5961GS driver [44 , 56] to knock down PERK in ISCs and compare lifespan of genetically identical sibling populations . While RU486 treatment had no effect on lifespan of wild-type animals , RU486 treatment extended lifespan of animals expressing PERKRNAi under the control of 5961GS ( Fig 6A and 6B ) . Similarly , promoting ER homeostasis by over-expressing spliced Xbp1 in ISCs/EBs ( using 5961GS ) extends lifespan moderately ( S6A Fig ) . Increased lifespan of flies expressing PERKRNAi in ISCs was accompanied by improved barrier function of the intestine , as determined using a dye-penetration assay [43] ( S6B Fig ) . The age-related activation of PERK in ISCs , which is a likely consequence of both tissue-wide and cell-autonomous ER stress , thus causes intestinal dysplasia , loss of barrier function of the intestine , and increased mortality in aging flies .
Drosophila ISCs , as many other stem cell types , are controlled extensively by redox signals [12 , 34 , 46] . Our previous work , as well as the results shown here , suggests that ER-induced oxidative stress plays a central role in the control of ISC proliferation after a proteostatic challenge . Our results support the notion that ER-induced ROS is a consequence of the PDI/Ero1L system , as has been proposed in mammalian cells [20] . However , Ero1L , as a thiol oxidase , may also affect the proper folding and maturation of Notch directly ( as described previously [48] ) , inhibiting ISC differentiation , and resulting in stem cell tumors . The phenotype of Ero1L-deficient ISC lineages supports a role for Ero1L in Notch signaling ( tumors with elevated numbers of Dl+ cells ) . At the same time , our results also support a role for Ero1L in limiting ISC proliferation directly through the UPRER ( and independently of Notch signaling or oxidative signals ) , as loss of Ero1L induces PERK activity without promoting ROS production in these cells . PERK itself is required for the induction of cell cycle and DNA replication genes in ISCs responding to TM treatment , yet it also induces antioxidant genes under these conditions , suggesting complex crosstalk between PERK-mediated control of mitotic activity of ISCs and the control of redox homeostasis in these cells . The fact that loss of Ero1L activates PERK while not inducing Xbp1 in ISCs suggests selective activation mechanisms for these two branches of the UPRER . We propose that this selectivity is associated with the production of ROS and that ER protein stress activates the Xbp1 branch when associated with a ROS signal , while PERK can be activated by unfolded proteins independently of ROS production . Further studies are needed to dissect the relative contribution of ROS production , PERK activation and Notch perturbation in the control of ISC proliferation in Ero1L loss of function conditions . Our results highlight the interaction between cell-autonomous and non-autonomous events in the ER stress response of ISCs and support the notion that improving proteostasis by boosting ER folding capacity in stem cells improves long-term tissue homeostasis and can impact lifespan . The regulation of PERK activity in ISCs by the JAK/Stat signaling pathway provides a tentative mechanism for the interaction between IECs experiencing ER stress and ISCs: We propose that JNK-mediated release of JAK/Stat ligands from stressed IECs ( as described in [54] ) results in JAK/Stat mediated activation of PERK in ISCs , and that this activation is required for the proliferative response of ISCs to epithelial dysfunction . The activation of JAK/Stat signaling in the intestinal epithelium of animals in which Xbp1 is knocked down in ECs , the requirement for JNK activation and Upd expression in ECs for ISC proliferation in response to stress , and the requirement for Stat ( and Hop and Dome ) in ISCs for the activation of eIF2α phosphorylation and stress-induced ISC proliferation , support this model ( Fig 7 ) . The mechanisms by which Stat mediates activation of PERK remain unclear , and will be interesting topics of further study . Studies in worms have established the UPRER as a critical determinant of longevity , and Xbp1 extends lifespan by improving ER stress resistance [25 , 28] . Our data further support the notion that regulating ER stress response pathways is critical to increase health- and lifespan . Here , chronic PERK activation can be considered a downstream readout of the buildup of proteotoxic stress in the intestinal epithelium during aging , which then perturbs proliferative homeostasis by continuously providing pro-mitotic signals to ISCs . Knocking down PERK in ISCs limits these pro-mitotic signals , improving homeostasis and barrier function , and extending lifespan . Lifespan is generally extended when ISC proliferation is limited in older flies , but not when it is completely inhibited [34 , 44 , 45 , 57 , 58] . Accordingly , we observe lifespan extension when PERK is knocked down using an RNAi approach that does not completely ablate PERK function ( note that experiments shown in Figs 1 and 2 were performed combining PERKRNAi with Dicer2 , which experiments in Fig 6 were performed using only PERKRNAi ) . ER stress has been documented as tightly associated with intestinal inflammation and the development of IBDs in mice and humans [4 , 59 , 60] . Genetic variants in Xbp1 are associated with higher susceptibility to IBD [4] and a recent study indicates that Xbp1 can act as a tumor suppressor in the intestinal epithelium , by limiting intestinal proliferative responses and tumor development through the control of local inflammation [5] . In this context , the specific role of PERK in the control of ISC proliferation in the fly gut is consistent with the function of PERK in the intestinal epithelium of mice , where activation of PERK can promote transition of ISCs into the transient amplifying cell population [11] . While the Drosophila midgut epithelium does not contain a transit amplifying cell population , our data suggest that a role for PERK in the proliferative response of the ISC lineage to ER stress is conserved . Due to the importance of the UPRER in the maintenance of tissue homeostasis in aging organisms , therapies targeting the UPRER are promising strategies to delay the aging process . Accordingly , pharmaceuticals that can limit ER stress ( such as Tauroursodeoxycholic acid , TUDCA and 4-phenylbutyrate , PBA ) have had therapeutic success in various human disorders [61 , 62] . Interestingly , flies fed PBA show increased lifespan , yet the effects of PBA on intestinal homeostasis have not yet been explored [63] . Studies from our lab and others highlight the importance of ISC function and proliferative homeostasis in fly longevity [34 , 44 , 45 , 58] . Based on this work , it is likely that further characterization of the effects of UPRER-targeting drugs on ISC function and intestinal homeostasis will help develop clinically relevant strategies to limit human aging and extend healthspan .
The following RNAi lines were obtained from the Vienna Drosophila RNAi Center: UAS::PERKRNAi ( v16427 and v110278 ) , UAS::ATF6RNAi ( v36504 ) , UAS::IRE1RNAi ( v39561 ) , UAS::Xbp1RNAi ( v109312 , v15347 ) , UAS::Hrd1RNAi ( v6870 ) , UAS::bskRNAi , UAS::Ero1LRNAi ( v51169 ) . StatRNAi ( v106980 ) , DomelessRNAi ( v106071 ) , HopRNAi ( v40037 ) , UpdRNAi . The following RNAi lines were obtained from the Bloomington Drosophila stock center: UpdRNAi ( 33680 ) , Upd2RNAi ( 33949 ) , Upd3RNAi ( 32859 ) Fly lines w1118 , frt82B , UAS::nlsGFP , UAS::Xbp1RNAi ( TRip:HMS03015 ) were obtained from the Bloomington Drosophila stock center . The following fly lines were generously provided as indicated: y1w1; esg::Gal4/+ by Dr . S Hayashi; Su ( H ) Gbe::Gal4 by Dr . S . Bray; UAS::dEro1L and Ero1L335qrs by Dr . H . Bellen; 2xSTAT-GFP by E . A . Bach;UAS::Upd2 , UAS::HoptumL from David Bilder; UAS::Xbp1d08698 by Dr . P . Fernandez-Funez; UAS::Xbp1spliced by Dr . P . Domingos; UAS::Upd by Dr . S . X . Hou; UAS:: Upd3 by Dr . N . Buchon; PEKRRNAi combined with Dicer2 by Dr . S . Marciniak . PERKe01744 is a Piggybac insertion line obtained from the Harvard Exelixis collection . According to the annotated information on Flybase , this line has a PBac{RB} element inserted into the 1st intron of all three predicted PERK spliceforms , and exhibits recessive lethality . All flies were raised on yeast/molasses-based food at 25°C and 65% humidity on a 12 hr light/dark cycle , unless otherwise noted . For tunicamycin exposure , flies were starved in empty vials for 6–8 hrs and fed with 5% sucrose solution± 50μM tunicamycin for 24hrs followed by dissection in PBS . For TARGET experiments , flies were raised at 18°C and shifted to 29°C at certain time points after eclosion . For MARCM clone induction , adult flies were aged for 1–2 days and then heat shocked at 37°C for 45 min . All data were collected from female flies only . Guts were dissected in PBS , fixed for 45 min at room temperature in 100 mM glutamic acid , 25 mM KCl , 20 mM MgSO4 , 4 mM sodium phosphate , 1 mM MgCl2 , and 4%formaldehyde , washed for 1hr , and incubated with primary antibodies and second antibodies in washing buffer ( PBS , 0 . 5% BSA , 0 . 1% Triton X-100 ) . The following primary antibodies were used: rabbit anti-peIF2α antibody ( Cell Signaling: 3597 , 1:150 ) , rat anti-Delta ( gift from Dr . MD Rand , University of Rochester , 1:1000 ) ; rabbit anti-pH3 ( phosphorylated histone H3 , Upstate , 1:1000 ) , mouse anti-β-galactosidase ( Developmental Studies Hybridoma Bank , 1:500 ) , rabbit anti-β-galactosidase ( Cappel , 1:5000 ) , mouse anti-Armadillo ( Developmental Studies Hybridoma Bank , 1: 250 ) For Delta antibody staining , guts were fixed using a methanol-heptane method as descried ( Lin et al . , 2008 ) . Fluorescent secondary antibodies were purchased from Jackson ImmunoResearch Laboratories . DNA was stained using DAPI . Confocal imaging was performed on a Zeiss LSM700 confocal microscope and processed using ImageJ and Adobe Illustrator . ROS levels were measured as described before ( Hochmuth et al . , 2011 ) . Briefly , guts were dissected in Schneider’s medium , incubated in 30 μM ( Invitrogen ) for 5 min at room temperature in the dark , washed twice and mounted to be imaged immediately . GFP expressed under the control of esg::Gal4 or esg::Gal4 , Su ( H ) ::Gal80 was used to identify ISCs and/or EBs . 35 virgins ( 5961::GS homozygotes ) were crossed to 20 w1118; UAS::PERKRNAi ( V16427 ) , or spliced Xbp1 homozygous males . Progeny of these crosses was collected at 3 to 4 days after the first fly hatched . Flies were then separated according to sex and genotype , and females were placed into cages ( 50–80 flies/cage ) and aged at 25°C . 100 μl of 5 mg/ml solution of RU486 or vehicle ( 80% ethanol ) were added on the top of a food vial and dried overnight before fed to flies . Food was changed every other day . Demographic data were analyzed using Prism statistical software . Wild-type Flies ( esgts , Su ( H ) GbeG80> w1118 ) and flies expressing dsRNA against PERK were exposed to 50μM tunicamycin and mock for 24 hrs ( 5% sucrose solution ) , followed by YFP+ labeled ISCs FACS sorting . Total RNA was then extracted using Trizol ( Invitrogen ) and used as template to generate RNA-seq libraries for Illumina sequencing . Expression was recorded as PRKM: reads per kbp per million reads .
|
The long-term maintenance of tissue homeostasis in barrier epithelia requires precise coordination of cellular stress and inflammatory responses with regenerative processes . This coordination is lost with age , resulting in degenerative and proliferative diseases . The Unfolded Protein Response of the Endoplasmic Reticulum ( UPRER ) is emerging as a central regulator of tissue homeostasis in barrier epithelia . The UPRER adjusts the protein folding capacity of the ER in response to protein stress in stem cells and differentiated cells , and thus influences proliferative homeostasis , cell differentiation and epithelial inflammatory responses . How these responses are coordinated to maintain epithelial homeostasis in aging organisms remains unclear . In a previous study , we have found that the UPRER controls intestinal stem cell ( ISC ) proliferation in the Drosophila intestinal epithelium by influencing the intracellular redox state . How signaling through the canonical ER stress sensor PERK ( PKR-like ER kinase ) is integrated into this signaling network remained unclear . Here we show that PERK serves as a central regulator of ISC proliferation and tissue homeostasis in response ER stress . Strikingly , we find that within the intestinal epithelium , PERK is activated specifically in ISCs in response to both systemic and local ER stress , and is required for ISC proliferation under both homeostatic and stress conditions . We identify JAK/Stat signaling as an activator of PERK in ISCs in response to ER stress in neighboring cells , and find that the wide-spread age-associated increase in PERK activity in ISCs is a cause of age-related dysplasia in this tissue . Accordingly , limiting PERK activity in ISCs promotes homeostasis of the intestinal epithelium in old flies and extends lifespan .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
PERK Limits Drosophila Lifespan by Promoting Intestinal Stem Cell Proliferation in Response to ER Stress
|
Insects determine their body segments in two different ways . Short-germband insects , such as the flour beetle Tribolium castaneum , use a molecular clock to establish segments sequentially . In contrast , long-germband insects , such as the vinegar fly Drosophila melanogaster , determine all segments simultaneously through a hierarchical cascade of gene regulation . Gap genes constitute the first layer of the Drosophila segmentation gene hierarchy , downstream of maternal gradients such as that of Caudal ( Cad ) . We use data-driven mathematical modelling and phase space analysis to show that shifting gap domains in the posterior half of the Drosophila embryo are an emergent property of a robust damped oscillator mechanism , suggesting that the regulatory dynamics underlying long- and short-germband segmentation are much more similar than previously thought . In Tribolium , Cad has been proposed to modulate the frequency of the segmentation oscillator . Surprisingly , our simulations and experiments show that the shift rate of posterior gap domains is independent of maternal Cad levels in Drosophila . Our results suggest a novel evolutionary scenario for the short- to long-germband transition and help explain why this transition occurred convergently multiple times during the radiation of the holometabolan insects .
The segmented body plan of insects is established by two seemingly very different modes of development [1–4] . Long-germband insects , such as the vinegar fly D . melanogaster , determine their segments more or less simultaneously during the blastoderm stage , before the onset of gastrulation [5 , 6] . The segmental pattern is set up by subdivision of the embryo into different territories , prior to any growth or tissue rearrangements . Short-germband insects , such as the flour beetle T . castaneum , determine most of their segments after gastrulation , with segments being patterned sequentially from a posterior segment addition zone . This process involves tissue growth or rearrangements as well as dynamic travelling waves of gene expression , which result from periodic oscillations that are driven by a molecular clock mechanism [7–10] ( technical terms in bold are explained in the glossary , in S1 Text ) . The available evidence strongly suggests that the short-germband mode of segment determination is ancestral , while the long-germband mode is evolutionarily derived [1 , 2 , 11] . Although the ancestor of holometabolan ( metamorphosing ) insects may have exhibited some features of long-germband segment determination [12] , it is clear that convergent transitions between the two modes have occurred frequently during evolution [2 , 11 , 13] . Long-germband segment determination can be found scattered over all four major holometabolous insect orders ( Hymenoptera , Coleoptera , Lepidoptera , and Diptera ) . Furthermore , there has been at least one reversion from long- to short-germband segment determination in polyembryonic wasps [14] . This suggests that , despite the apparent differences between the two segmentation modes , it seems relatively easy to evolve one from the other . Why this is so , and how the transition is achieved , remains unknown . In this paper , we provide evidence suggesting that the patterning dynamics of long- and short-germband segmentation are much more similar than previously thought . Specifically , we demonstrate that shifting domains of segmentation gene expression in the posterior of the D . melanogaster embryo can be explained by a damped oscillator mechanism , dynamically very similar to the clocklike mechanism underlying periodically oscillating gene expression during short-germband segment determination . We achieve this through analysis of a quantitative , data-driven gene circuit model of the gap network in D . melanogaster . The gap gene system constitutes the topmost hierarchical layer of the segmentation gene cascade [6] . Gap genes hunchback ( hb ) , Krüppel ( Kr ) , giant ( gt ) , and knirps ( kni ) are activated through morphogen gradients formed by the products of maternal coordinate genes bicoid ( bcd ) and caudal ( cad ) . Gap genes are transiently expressed during the blastoderm stage in broad overlapping domains along the anteroposterior ( A–P ) axis of the embryo ( Fig 1A ) . They play an important role regulating spatially periodic pair-rule gene expression . Pair-rule genes , in turn , establish the precise pre-pattern of the segment-polarity genes , whose activities govern the morphological formation of body segments later in development , after gastrulation has occurred . Our aim is to go beyond the static reconstruction of network structure to explicitly understand the regulatory dynamics of the patterning process [15 , 16] . To achieve this , we use the powerful tools of dynamical systems theory—especially the geometrical analysis of phase ( or state ) space [17]—to characterize the patterning capacity of the gap gene network . We study the complex regulatory mechanisms underlying gap gene expression in terms of the number , type , and arrangement of attractors and their associated basins of attraction , which define the phase portrait . The geometry of the phase portrait in turn determines the flow of the system . This flow consists of individual trajectories that describe how the system state changes over time given some specific initial conditions . In our gap gene circuit model , initial conditions are given by the maternal Hb gradient , boundary conditions by the maternal Bcd and Cad gradients , and the state variables consists of the concentrations of regulators Hb , Kr , Kni , and Gt . Different configurations of phase space give rise to differently shaped trajectories and , thus , to different gap gene regulatory dynamics . The power of analogy between phase space and its features , and developmental mechanisms , has long been recognized and exploited . In their original "clock-and-wavefront" model , Cooke and Zeeman [18] characterize cells involved in somitogenesis in the pre-somitic mesoderm as "oscillators with respect to an unknown clock or limit cycle in the embryo . " More recently , geometrical analysis of phase space has been successfully used to study developmental processes such as vertebrate somitogenesis [19] , vulval development in nematodes [20] , A–P patterning by Hox genes [21] , and—particularly relevant in our context—the robust ( canalized ) patterning dynamics of gap genes [22–25] . To make the problem tractable , these analyses are often performed in a simplified framework . For example , in previous studies of Drosophila segmentation , models were used with a static Bcd gradient and Cad dynamics frozen after a particular time point during the late blastoderm stage [22 , 23 , 25–27] . This rendered the system autonomous , meaning that model parameters—and therefore phase space geometry—remain constant over time . However , the maternal gradients of Bcd and Cad change and decay on the same timescale as gap gene expression [28] . Taking this time dependence of maternal regulatory inputs into account leads to a nonautonomous dynamical system , in which model parameters are allowed to change over time ( see [29] and S1 Text for a detailed model comparison ) . This causes the geometry of phase space to become time-variable: the number , type , and arrangement of attractors and their basins change from one time point to the next . Bifurcations may occur over time , and trajectories may cross from one basin of attraction to another . All of this makes nonautonomous analysis highly nontrivial . We have developed a novel methodology to characterize transient dynamics in nonautonomous models [30] . It uses instantaneous phase portraits [29 , 31] to capture the time-variable geometry of phase space and its influence on system trajectories . By fitting dynamical models to quantitative spatiotemporal gap gene expression data , we have obtained a diffusion-less , fully nonautonomous gap gene circuit featuring realistic temporal dynamics of both Bcd and Cad ( Fig 1A ) [29 , 32] ( see Materials and methods and S1 Text for details ) . The model has been extensively validated against experimental data [22 , 23 , 26 , 27 , 29 , 32] and represents a regulatory network structure that is consistent with genetic and molecular evidence [6] . We have performed a detailed and systematic phase space analysis of this nonautonomous gap gene circuit along the segmented trunk region of the embryo , explicitly excluding head and terminal patterning systems [29] ( see Materials and methods for details ) . At every A–P position between 35% and 73% , we calculated the number and type of steady states in the associated phase portrait [29] . This allowed us to characterize the different dynamical regimes driving gap gene expression along the embryo trunk and to explicitly identify the time-dependent aspects of gap gene regulation [29] . In the anterior trunk region of the embryo , where boundary positions remain stationary over time , gap gene expression dynamics are governed by a multi-stable dynamical regime ( Fig 1B ) [29] . This is consistent with earlier work [23] , indicating that modelling results are robust across analyses . Here , we focus on the regulatory mechanism underlying patterning dynamics in the posterior of the embryo , which differs between autonomous and nonautonomous analyses . Posterior gap domains shift anteriorly over time [26 , 28] . Autonomous analyses suggested that these shifts are driven by a feature of phase space called an unstable manifold [23] , while our nonautonomous analysis reveals that they are governed by a mono-stable spiral sink ( Fig 1B ) . The presence of a spiral sink indicates that a damped oscillator mechanism is driving gap domain shifts in our model [17] . Here , we present a detailed mathematical and biological analysis of this damped oscillator mechanism in the posterior of the embryo , between 53% and 73% A–P position , and discuss its implications for pattern formation and the evolution of the gap gene system . Our results suggest that long-germband and short-germband modes of segmentation both use oscillatory regimes ( damped and limit cycle oscillators , respectively ) in the posterior region of the embryo to generate posterior to anterior waves of gene expression . Characterizing and understanding these unexpected similarities provides a necessary first step towards a mechanistic explanation for the surprisingly frequent occurrence of convergent transitions between the two modes of segment determination during holometabolan insect evolution .
The gap gene circuit model used for our analysis consists of a one-dimensional row of nuclei along the A–P axis [32 , 33] . Continuous dynamics during interphase alternate with discrete nuclear divisions . Our full model includes the entire segmented trunk region of the embryo between 35% and 92% A–P position . It covers the last two cleavage cycles of the blastoderm stage ( starting at the end of cleavage cycle 12 , C12 , at t = 0 , including C13 and C14A ) up to the onset of gastrulation; C14A is subdivided into 8 equally spaced time classes ( T1–T8 ) . Division occurs at the end of C13 . The state variables of the system represent the concentrations of proteins encoded by gap genes hb , Kr , gt , and kni . The concentration of protein a in nucleus i at time t is given by gia ( t ) . Change in protein concentration over time occurs according to the following system of ordinary differential equations: dgia ( t ) dt=Raϕ ( ua ) −λagia ( t ) ( 1 ) where Ra and λa are rates of protein production and decay , respectively . ϕ is a sigmoid regulation-expression function used to represent the cooperative , saturating , coarse-grained kinetics of transcriptional regulation . It incorporates nonlinearities into the model that enable it to exhibit complex behavior , such as multi-stability and damped or sustained oscillations . It is defined as ϕ ( ua ) =12 ( ua ( ua ) 2+1+1 ) ( 2 ) where ua=∑b∈GWbagib ( t ) +∑m∈MEmagim ( t ) +ha ( 3 ) The set of trunk gap genes is given by G = {hb , Kr , gt , kni} and the set of external regulatory inputs by the products of maternal coordinate and terminal gap genes M = {Bcd , Cad , Tailless ( Tll ) , Huckebein ( Hkb ) } . Concentrations of external regulators gim are interpolated from quantified spatiotemporal protein expression data [28 , 32 , 34] . Changing maternal protein concentrations means that parameter term ∑m∈MEmagim ( t ) is time dependent , which renders the model nonautonomous . Interconnectivity matrices W and E represent regulatory interactions between gap genes and from external inputs , respectively . Matrix elements wba and ema are regulatory weights . They summarize the effect of regulator b or m on target gene a and can be positive ( representing an activation ) , negative ( repression ) , or near zero ( no interaction ) . ha is a threshold parameter representing the basal activity of gene a , which includes the effects of regulatory inputs from spatially uniform regulators in the early embryo . The system of equations ( Eq 1 ) governs regulatory dynamics during interphase; Ra is set to zero during mitosis . Additional information about our model formalism can be found in S1 Text . We obtained values for parameters Ra , λa , W , E , and ha by fitting the model to data over a full spatial range covering the segmented trunk region between 35% and 92% A–P position ( see S1 Data ) [26 , 32 , 35 , 36] . Signs of parameters in the genetic interconnectivity matrices W and E were constrained during the fit to allow direct comparison with previously published models [23 , 32] . A detailed account of how we fit the model and selected solutions for analysis has been published previously [29]; we provide a summary in S1 Text . Briefly , model equations ( Eq 1 ) are solved numerically , and the resulting model output is compared to a quantitative data set of spatiotemporal gap protein profiles . The difference between model output and data is minimized using parallel Lam Simulated Annealing ( pLSA ) . Model fitting was performed on the Mare Nostrum cluster at the Barcelona Supercomputing Centre ( http://www . bsc . es ) . The best-fitting solution was selected for further analysis , as described in S1 Text ( model parameters are shown in S1 Table ) . The resulting diffusion-less , nonautonomous gene circuit has a residual error ( measured by its root mean square score ) of 14 . 53 ( see S1 Text ) . It reproduces gap gene expression with high accuracy , showing only minor defects in the shape of expression domain boundaries ( Fig 1A ) . The modelling and optimization code to reverse-engineer the gap gene network is implemented in C , using MPI for parallelization and the GNU Scientific Library ( GSL , http://www . gnu . org/software/gsl ) for data interpolation . It is available for download online at https://subversion . assembla . com/svn/flysa . Embryos derived from cad mutant germ-line clones were generated and collected as previously described [39 , 40] , and females were then mated to wild-type males . The resulting embryos all lack maternal cad activity but carry one paternal copy of the cad gene . mRNA expression patterns of the gap genes gt or kni , and the pair-rule gene even-skipped ( eve ) were visualized using an established enzymatic ( colorimetric ) in situ hybridization protocol [36] . Images were taken and processed using FlyGUI ( https://subversion . assembla . com/svn/flygui ) to extract the position of expression domain boundaries , as described in [41] . The image data and extracted boundary positions are available from figshare at https://figshare . com/s/839791c208e42b7e61fe ( DOI: 10 . 6084/m9 . figshare . 5809653 ) .
Gap domain boundaries posterior to 52% A–P position shift anteriorly over time ( Fig 1A and Fig 2A ) [26 , 28] . These domain shifts cannot be explained by nuclear movements [42] , nor do they require diffusion or transport of gap gene products between nuclei [22 , 23 , 26 , 29] ( see also S1 Text ) . Instead , gap domain shifts are kinematic , caused by an ordered temporal succession of gene expression in each nucleus , which produces apparent wavelike movements in space [23 , 26] . This is illustrated in Fig 2A for nuclei between 55% and 73% A–P position ( see Materials and methods ) . Each nucleus starts with a different initial concentration of maternal Hb , which leads to the expression of different zygotic gap genes: Kr in the central region of the embryo or kni further posterior . Nuclei then proceed through a stereotypical temporal progression , in which Kr expression is followed by kni ( e . g . , nucleus at 59% ) , kni by gt ( nucleus at 69% ) , and , finally , gt by hb ( nuclei posterior of 75%; not shown ) . No nucleus goes through the expression of all four trunk gap genes over the course of the blastoderm stage and each nucleus goes through a different partial sequence within this progression , according to its initial conditions . This coordinated dynamic behavior is what we need to explain in order to understand the regulatory mechanism underlying gap domain shifts . To do this , we carried out a systematic characterization of the dynamical regimes driving A–P gap gene patterning in a nonautonomous gap gene circuit model [29] . For every nucleus along the trunk region of the embryo , we visualized the dynamics of gap gene expression in the context of the instantaneous phase portraits that underlie them . That is , we calculated the positions and types of steady states present at every time class and plotted them ( color coded for time ) with the simulated expression dynamics for that nucleus . This yielded a full nonautonomous phase portrait associated with each nucleus . In this way , we can understand each trajectory's shape in terms of the changing geometry of the flow ( see Materials and methods for details ) . Our analysis revealed that phase portraits of nuclei between 53% and 73% A–P position are mono-stable throughout the blastoderm stage ( see , for example , Fig 2B ) . Given enough time , all trajectories would approach the only attractor present , which , at the end of the blastoderm stage ( time class T8 ) , is located close to the origin ( Fig 2B , yellow cylinder ) . Due to the nonautonomy of the system , this attractor moves across phase space over developmental time . However , this movement of the attractor is not the most important factor determining the shape of trajectories . Due to the limited duration of the blastoderm stage , the system always remains far from steady state , and posterior gap gene expression dynamics are determined by the geometry of transient trajectories relatively independently of the precise position of the attractor . Because the moving attractor positions are similar for all posterior nuclei , we were able to plot the trajectories of the different nuclei onto the same projection of phase space ( Fig 2C ) . Over time , posterior nuclei transit through buildup of Kr , then Kni , then Gt proteins . Their initial conditions are given by Hb and this determines where in the sequence they start . The plots in Fig 2B and 2C show that the ordered succession of gap gene expression is a consequence of the rotational ( spiral-shaped ) geometry of the trajectories . Eigenvalue analysis revealed that the mono-stable steady state of posterior nuclei is a special type of point attractor: a spiral sink , or focus [17 , 29] . Trajectories do not approach such a sink in a straight line but spiral inward , instead . This contributes to the curved rotational geometry of the trajectories shown in Fig 2B and 2C . From the theory of dynamical systems , we know that spiral sinks are the hallmark of damped oscillators [17] . Given that spiral sinks are the only steady states present in the mono-stable phase portraits of posterior nuclei , we concluded that , in our model , posterior gap gene expression dynamics are driven by a damped oscillator mechanism . This damped oscillator mechanism imposes the observed temporal order of gap gene expression ( Fig 2A ) . Temporal order is a natural consequence of oscillatory mechanisms , one obvious example being the stereotypical succession of cyclin gene expression driven by the cell cycle oscillator [43 , 44] . In contrast , the imposition of temporal order is not a general property of unstable manifolds ( found to drive gap domain shifts in previous autonomous analyses [23–25] ) . For this reason , our damped oscillator mechanism provides a revised understanding of gap domain shifts , which is more general and therefore constitutes an important conceptual advance over previous characterizations . Each nucleus runs through a different range of phases within a given time period ( see color wheel diagrams in Fig 2A ) , as determined by the damped oscillator . Arranged properly across space , phase-shifted partial trajectories create the observed kinematic waves of gene expression . In this sense , the dynamics of the shifting gap domains in the D . melangoaster blastoderm and those of the travelling waves of gene expression in short-germband embryos are equivalent , because they are both an emergent property of the temporal order imposed by an underlying oscillatory regulatory mechanism . In principle , domain shifts are not strictly necessary for subdividing an embryo into separate gene expression territories . Wolpert's French Flag paradigm for positional information , for example , works without any dynamic patterning downstream of the morphogen gradient [45 , 46] . This raises the question of why such shifts occur and what , if anything , they contribute to pattern formation . One suggestion is that feedback-driven shifts lead to more robust patterning than a strictly feed-forward regulatory mechanism , such as the French Flag [47 , 48] . This is supported by the fact that the unstable manifold found in autonomous analyses [23] has canalizing properties: as time progresses , it attracts trajectories coming from different initial conditions into an increasingly small and localized subvolume of phase space . This desensitizes the system to variation in maternal gradient concentrations [22] . Based on these insights , we asked whether our damped oscillator mechanism exhibits similar canalizing behavior , ensuring robust gap gene patterning . A closer examination of the spiral trajectories in Fig 2C reveals that they are largely confined to two specific sub-planes in phase space ( see S1 and S2 Movies ) . Specifically , they tend to avoid regions of simultaneously high levels of Gt and Kr , allowing us to "unfold" the three-dimensional volume of Kr-Kni-Gt space into two juxtaposed planes representing Kr-Kni and Kni-Gt concentrations ( Fig 2D ) . This projection highlights how trajectories spend variable amounts of time on the Kr-Kni plane before they transition onto the Kni-Gt plane . In order to investigate the canalizing properties of our damped oscillator mechanism , we performed a numerical experiment , shown in Fig 3A and 3B . We chose a set of regularly distributed initial conditions for our model that lies within the Kr-Gt plane ( Fig 3A ) and used this set of initial conditions to simulate the nucleus at 59% A–P position , with a fixed level of Kni ( Fig 3B ) . These simulations illustrate how system trajectories converge to the Kr-Kni or Kni-Gt plane , avoiding regions of simultaneously high Kr and Gt concentrations . Convergence occurs rapidly and is already far advanced in early cleavage cycle 14A ( Fig 3B , time class T1 ) , demonstrating that the subvolume of phase space in which trajectories are found becomes restricted long before a steady state is reached . At later stages , convergence slows down but continues confining trajectories to an increasingly restricted subvolume of phase space ( up to late cleavage cycle 14A , Fig 3B , time class T8 ) . This phenomenon can be seen as the equivalent of trajectories becoming restricted to valleys in Waddington's original landscape metaphor , which motivated the definition of the term "canalization" [49] . The canalizing behavior is robust with regard to varying levels of Kni ( S1 Fig ) . It is straightforward to interpret the exclusion of trajectories from regions of simultaneous high Kr and high Gt in terms of regulatory interactions . There is strong bidirectional repression between gt and Kr , which is crucial for the mutually exclusive expression patterns of these genes [6 , 27 , 36] . In the context of our damped oscillator mechanism , this mutual repression implies that the system must first transition from high Kr to high Kni/low Kr before it can initiate gt expression . This is exactly what we observe ( Fig 2A ) , confirming that the damped oscillator in the posterior of the D . melanogaster embryo has canalizing properties due to mutually exclusive gap genes . How do spiral trajectories switch from one plane in phase space to another ? To answer this question , we examined the flow of the system . We unfolded the Kr-Kni and Kni-Gt planes and projected trajectories and states of posterior nuclei onto this unfolded flow ( Fig 3C and S2 Fig ) . These plots reveal drastic differences in flow velocity ( magnitude ) in different regions of phase space at different points in time . At early stages , close to the origin , we observe a fast initial increase in Kr and Kni concentrations , indicated by red arrows at low Kr and Kni concentrations in Fig 3C ( C13 and T2 ) . Nuclei whose trajectories remain on the Kr-Kni plane then show a dramatic slowdown . They either continue to gradually increase levels of Kr or exhibit slow buildup of Kni , combined with consequent decrease of Kr due to repression by Kni ( Fig 3C , T4 and T6 ) . As trajectories of different nuclei approach the border between the Kr-Kni and Kni-Gt planes , the Gt component of the flow on the Kr-Kni plane becomes positive ( trajectories marked by stars in Fig 3C and S2 Fig ) . This "lifts" the trajectory out of the Kr-Kni and into the Kni-Gt plane . In the border zone between the two planes , the flow in the direction of Gt is high throughout the blastoderm stage ( Fig 3C ) , ensuring that the switch between planes occurs rapidly . Nuclei then again enter a zone of slower dynamics with a gradual buildup of Gt , combined with consequent decrease of Kni due to repression by Gt ( Fig 3C , T4 and T6 ) . Thus , the flow of our model combines relatively slow straight stretches within a plane of phase space with rapid turns at the border between planes . Similar alternating fast-slow dynamics have been observed in autonomous models [24] . These dynamics are important for gap gene patterning because they influence the width of gap domains ( through relatively stable periods of expressing a specific gap gene ) and the sharpness of domain boundaries ( through abrupt changes in gene expression at borders between planes ) . Such fast-slow dynamics are characteristic of relaxation oscillations [17] . A relaxation oscillator combines phases of gradual buildup in some of its state variables with rapid releases and changes of state , resulting from an irregularly shaped limit cycle . Although there seem to be no limit cycles present in our phase portraits , the irregular geometries of spiralling transient trajectories in our model can be understood as relaxation-like ( fast-slow ) dynamics , which , driven by a damped oscillator , govern the shape and the shift rate of posterior gap domains . In the short-germband beetle T . castaneum , an oscillator mechanism governs travelling waves of pair-rule gene expression [7 , 8] . The frequency of these repeating waves is positively correlated with the level of Cad in the posterior of the embryo: the more Cad present , the faster the oscillations [9] . In addition , a recent publication proposes that waves of gap gene expression observed in the T . castaneum blastoderm and elongating germ band may be caused by a succession of temporal gene expression switches whose rate and timing is also under control of the posterior gradient of Cad [50] . These authors speculate that Cad may control gap gene expression in D . melanogaster in an equivalent way . In D . melanogaster , changing concentrations of maternal morphogens do indeed influence posterior gap domain shifts [29 , 39] . Therefore , we asked how altered levels of Cad affect the damped oscillator mechanism regulating gap genes in D . melanogaster . We assessed the regulatory role of Cad by multiplying its concentration profile with different constant scaling factors—reducing Cad levels in space and time without affecting overall profile shape—and by measuring the dynamics and extent of gap domain shifts in the resulting simulations ( Fig 4 ) . In particular , we focus on how lowered levels of Cad affect the position of the Kr-Gt interface over time ( Fig 4A and 4B ) . Our model makes three specific predictions . First , the initial position of the Kr-Gt border interface does not change when Cad levels are decreased ( Fig 4B , C13 ) . Second , between C13 and C14A-T1 , gap domains simulated with lowered concentrations of Cad start to lag behind those simulated with wild-type levels ( Fig 4B , C13 and T1 ) . Third , from T1 onwards , shift rates become independent of Cad concentration , and boundary positions move in parallel in different simulations for the remainder of the blastoderm stage ( Fig 4B , T1–T8 ) . This last prediction is incompatible with a mechanism in which the rate of successive bifurcation-driven switches is under the direct control of Cad , which requires the shift rate to be sensitive to Cad concentration [50] . A comparison of the flow in models with reduced and wild-type levels of Cad revealed that this maternal factor affects the timing of gap domain shifts by modulating the fast-slow dynamics of the gap gene damped oscillator . While the direction of the flow remains largely constant across different concentrations of Cad , its magnitude changes significantly ( Fig 4C–4E and S3 Fig ) . The magnitude of the flow is most sensitive in the area of the Kr-Kni plane around the origin , where it is strongly reduced at early stages in simulations with lowered levels of Cad ( Fig 4C–4E , time class C12 ) . This implies a slower initial buildup of Kr and Kni protein at low Cad and hence the delayed onset of domain shifts . At later stages , when wild-type Cad levels decrease , differences in the magnitude of the flow are very subtle ( Fig 4C–4E , time class T8 , and S3 Fig , from time class C14A-T3 onwards ) . As a result of the altered early flow , the curvature of trajectories is decreased with lower Cad concentration , leading to tighter spirals . This demonstrates that the early difference in Cad levels continues to influence the behavior of the gap system into the late blastoderm stage ( S4 Fig ) . Progress along these tightened spirals is much slower than along the wider ones followed in wild type , due to the weaker flow in regions near the origin ( compare S2 Fig and S4 Fig ) . This slowed progress compensates for the tightened geometry of the spiral trajectories , preserving the rate of change in the "phase" of gap gene expression . In this way , the relative rate of the shifts remains unperturbed by changing the concentration levels of Cad , leading to the parallel trajectories after C14A-T1 depicted in Fig 4B . To experimentally test the predictions from our model , we need to carefully manipulate the levels of Cad protein in blastoderm embryos without disturbing the spatial pattern too much . This is difficult to achieve due to the lack of well-characterized hypomorphic mutants of cad in D . melanogaster and the overlapping but distinct spatiotemporal profiles of the maternal and zygotic expression contributions [51 , 80] . In the absence of more precise genetic tools , we quantified boundary shifts of Gt and Kni domains in mutant embryos derived from cad germ-line clones , which lack the maternal contribution to Cad expression . These mutants are viable as long as one paternal copy of cad is present , and exhibit reduced levels of ( zygotic ) Cad protein , with a spatial expression profile that is comparable to the wild type at the late blastoderm stage [51] . As predicted by our simulations , these mutants show delayed shifts of the posterior Gt ( Fig 5 , and S5 Fig ) and the abdominal Kni domains [39] . Here , we focus on the anterior boundary of the posterior Gt domain ( Fig 5A , arrowhead ) , which corresponds to the Kr-Gt interface measured in Fig 4 . It satisfies all three model predictions . First , its position at the onset of Gt expression in C13 is the same in mutant and wild-type embryos . This corroborates earlier analyses suggesting that maternal Hb ( and not Cad ) is the main morphogen in the posterior of the embryo [6 , 23 , 29 , 52] . Second , between C13 and C14A-T1 , it lags behind its wild-type position , exhibiting a subtle but clearly detectable posterior displacement by T1 ( Fig 5A ) . Gap domain shifts are only initiated around late C13 , when enough gap protein has accumulated to initiate cross-regulatory interactions [6 , 53] . The slower accumulation of gap protein in the posterior of the embryo therefore causes a delay in the onset of the shifts in the mutant . Third , from T1 onwards , shift rates in wild type and mutants remain more or less the same , indicating that they are robust towards changes in levels of Cad ( Fig 5 , after C14A-T1 ) . Even though the conditions of model simulations and mutants may not match perfectly , this provides clear evidence that gap domain shifts are relatively insensitive to the precise level of Cad concentration . Taken together , our experimental and modelling evidence suggest that Cad regulates the timing but not the positioning of gap gene expression in early blastoderm stage embryos of D . melanogaster . At later stages , gap domain shift rates are robust towards changes in Cad concentration . This is not entirely surprising , because the shifts result from gap–gap cross-regulatory interactions rather than depending directly on maternal input [6 , 26 , 32 , 36] . Analysis of our model shows that this robustness is entirely consistent with a damped oscillator mechanism , while a mechanism based on temporal switching under the control of Cad [50] would be much more sensitive to altered levels of the maternal gradient .
In this paper , we have shown that a damped oscillator mechanism—with relaxation-like behavior—can explain robust segmentation gene patterning of the long-germband insect D . melanogaster . Even though they may not be periodic , the kinematic shifts of gap gene expression domains in our model are an emergent property of temporally regulated gene expression driven by a damped oscillator . In this sense , they are dynamically equivalent to the travelling waves of gene expression involved in vertebrate somitogenesis [19 , 54] and short-germband arthropod segmentation [7–9 , 55 , 56] , both of which also emerge from temporal order imposed by oscillatory mechanisms . This lends support to the notion that the regulatory dynamics of segmentation gene expression in long- and short-germband insects are much more similar than is evident at first sight [57 , 58] . The mechanism described in this paper differs from an earlier proposal that gap domain shifts are driven by an unstable manifold [23] . Can these two mechanisms be distinguished experimentally ? We think they can , because the two models make different predictions for embryos misexpressing hb in the posterior region of the embryo . According to the model put forward by Manu and colleagues [23] , nuclei exposed to high maternal Hb concentrations will rapidly converge to an attractor with high zygotic Hb concentration by the end of the blastoderm stage . In contrast , our model predicts these nuclei will express high levels of Kr in addition to hb ( S6 Fig ) . Because real embryos misexpressing hb under a heat-shock promoter show high levels of Kr in the posterior embryo trunk region [59 , 60] , our model is better supported by the available experimental evidence . In addition to these empirical considerations , the proposed damped oscillator provides a more general explanation of the developmental and evolutionary dynamics of gap gene expression than the unstable manifold reported previously [23] . The spiral geometry of this manifold is contingent . It happens to traverse all the relevant expression states ( from Kr to kni to gt to hb ) , but such a succession of states is not a general characteristic of unstable manifolds . In contrast , cycling through successive states is not just typical for our proposed damped oscillator; it is the hallmark of gene expression oscillators in general . A succession of gene expression states could also be generated by a timed series of bifurcation-based switches , as suggested by Tufcea and François [61] . This relies on a precise mechanism for the temporal regulation of the switches . Zhu and colleagues [50] have recently proposed that Cad controls such a cascade of gap gene switches in both T . castaneum and D . melanogaster . The evidence presented here renders this scenario unlikely , at least in the case of D . melanogaster . One problem with the timed-switch mechanism is that it remains unclear how it could be implemented by the known interactions among gap genes [6] . Another problem is that it operates at criticality throughout the embryo—undergoing a rapid series of bifurcations . This leaves it extremely sensitive to changes in Cad concentration , unlike the robust oscillator reported here . Interestingly , there is some indication for such widespread criticality in the gap gene system from a recent study using quantitative co-expression measurements and a simplified set of gene regulatory models [62] . We could not find any evidence for this type of criticality in our model , which is based on a detailed and experimentally validated regulatory structure of the gap gene network [6 , 23 , 26 , 29 , 32] . Shifting gap domains play a central role in segmental patterning in D . melanogaster by directly regulating stripes of pair-rule gene expression . Posterior pair-rule stripes also exhibit anterior shifts in this species . They are produced by and closely reflect the expression dynamics of the gap genes [28] . In fact , dynamic shifts in gap domain positions are strictly required for the correct spatiotemporal expression of pair-rule genes in D . melanogaster [58] . In contrast , gap genes play a much less prominent role in patterning posterior segments in short-germband arthropods . Instead , periodic kinematic waves of pair-rule gene expression are thought to be generated by negative feedback between the pair-rule genes themselves ( in T . castaneum [63] ) or by an intercellular oscillator driven by Notch/Delta signalling ( in cockroaches [64] and centipedes [55 , 56] ) . The evolutionary transition from short- to long-germband segmentation has long been thought to have involved the recruitment of gap genes for pair-rule gene regulation , to replace the ancestral oscillatory mechanism [6 , 12 , 13 , 65 , 66] . The mechanistic details of how this occurred remain unclear . Gap gene–driven and segmentation clock–driven modes of patterning have been assumed to be mutually exclusive in any given region of the embryo . In contrast , our results suggest that during the replacement process , gap and pair-rule oscillators might have temporarily coexisted , which would greatly facilitate the transition . In this scenario , gap genes gradually take over pair-rule–driven oscillatory patterning in the posterior and later convert to a more switch-like static patterning mode , as observed in the anterior of the D . melanogaster embryo [23 , 27–29] . This is tentatively supported by the fact that the spatial extent of the posterior region , which is patterned by shifting gap domains , differs between dipteran species [39 , 67] . This scenario suggests that posterior gap domains shift as a result of the dynamic regulatory context into which they have been recruited during evolution . In addition , it provides an explanation for why gap domain shifts are essential for the correct placement of pair-rule stripes in D . melanogaster [58] . Seen from another angle , our results imply that equivalent regulatory dynamics ( in this case , domain shifts and travelling waves of gene expression ) can be produced by different oscillatory mechanisms . The use of divergent regulatory mechanisms to independently pattern identical expression domains appears to be very common ( see , for example , [68–71] ) . Indeed , the relative contribution of different mechanisms may evolve over time , with little effect on downstream patterning [72] . This type of compensatory evolution is called developmental system drift [73–77] . It has recently been shown to occur extensively in the evolution of the dipteran gap gene system [39 , 78] . System drift provides the necessary conditions that enable the facilitated gradual transition between the different regulatory mechanisms described above . Even though the core mechanisms that generate both behaviors differ , some aspects of segmentation gene regulation are strikingly similar between long- and short-germband insects . In different species of dipteran insects , as well as in T . castaneum , travelling kinematic waves of gene expression are involved in segment determination [9 , 26 , 39 , 50 , 67] . Cad is always involved in the initial activation of these patterns [9 , 39 , 50 , 79–82] . It also appears to control aspects of pair-rule gene regulation in centipedes [55 , 56] . From this , we conclude that the activating role of Cad in initiating these dynamics is highly conserved . In contrast , our evidence argues against a proposed universal role of Cad in regulating the rate and dynamics of travelling waves of segmentation gene expression [50] . In D . melanogaster , Cad exerts its effect primarily through regulating levels of gap gene expression; it has no direct role in the positioning of gap gene expression domains [29] . Travelling waves of gene expression that narrow and slow down over time are involved in both arthropod segmentation and vertebrate somitogenesis . It has long been recognized that these expression dynamics imply differential regulation of the rate of an oscillatory process along the A–P axis [54] . However , mechanistic explanations for this phenomenon remain elusive . A number of recent models simply assume that the concentration of some posterior morphogen determines the period of cellular oscillators , without investigating how this might arise ( see , for example , [9 , 83 , 84] ) . Experimental evidence from vertebrates suggests alteration of protein stability or translational time delays as a possible mechanism [85 , 86] . In contrast , our dynamical analysis illustrates how slowing ( damped ) oscillations can emerge directly from the intrinsic regulatory dynamics of a transcriptional network , without altering rates of protein synthesis or turnover , or even the need for external regulation by morphogens . A similar mechanism based on intrinsic oscillatory dynamics of a gene network was recently proposed for vertebrate somitogenesis [87] . It will be interesting to investigate which specific regulatory interactions mediate the effect of Cad on the T . castaneum pair-rule gene oscillator . Patterning by the gap gene system also shows interesting parallels to the developmental system governing the dorsoventral subdivision of the vertebrate neural tube . In both cases , the target domains of the respective morphogen gradients move away from their initial position over time due to downstream gene interactions , and in both cases , this involves a temporal succession of target gene expression [88] . Previous analyses suggest that this temporal succession of gene expression in the vertebrate neural tube may be caused by a succession of bistable switching events [61 , 89] . However , the possibility of damped oscillations was never explicitly investigated in any of these analyses . In light of the results presented here , it would be interesting to check for their presence in this patterning system . In summary , we argue that oscillatory mechanisms of segmentation gene regulation are not exclusive to short-germband segmentation or somitogenesis . Our analysis provides evidence that the spatial pattern of gap gene expression in the posterior region of the D . melanogaster embryo also emerges from a temporal sequence of gap gene expression driven by an oscillatory mechanism: a regulatory damped oscillator . This results in the observed anterior shifts of posterior gap domains . We suggest that the dynamic nature of posterior gap gene patterning is a consequence of the context in which it evolved and that two different oscillatory mechanisms may have coexisted during the transition from short- to long-germband segmentation . Studies using genetics and data-driven modelling in non-model organisms will reveal the regulatory circuits responsible for driving the different dynamics involved in segmentation processes , as well as the precise nature of the regulatory changes involved in transitions between them [39 , 78 , 90] . Given the insights gained through its application to gap gene patterning in D . melanogaster , phase space analysis will provide a suitable dynamic regulatory context in which to interpret and analyze these results .
|
Different insect species exhibit one of two distinct modes of determining their body segments ( known as segmentation ) during development: they either use a molecular oscillator to position segments sequentially , or they generate segments simultaneously through a hierarchical gene-regulatory cascade . The sequential mode is ancestral , while the simultaneous mode has been derived from it independently several times during evolution . In this paper , we present evidence suggesting that simultaneous segmentation also involves an oscillator in the posterior end of the embryo of the vinegar fly , Drosophila melanogaster . This surprising result indicates that both modes of segment determination are much more similar than previously thought . Such similarity provides an important step towards our understanding of the frequent evolutionary transitions observed between sequential and simultaneous segmentation .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"morphogenic",
"segmentation",
"gene",
"regulation",
"animals",
"blastulas",
"animal",
"models",
"developmental",
"biology",
"drosophila",
"melanogaster",
"model",
"organisms",
"systems",
"science",
"mathematics",
"experimental",
"organism",
"systems",
"embryos",
"morphogenesis",
"drosophila",
"research",
"and",
"analysis",
"methods",
"blastoderm",
"embryology",
"computer",
"and",
"information",
"sciences",
"gene",
"expression",
"dynamical",
"systems",
"insects",
"arthropoda",
"genetic",
"oscillators",
"eukaryota",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"organisms"
] |
2018
|
A damped oscillator imposes temporal order on posterior gap gene expression in Drosophila
|
The mortality of salmon smolts during their migration out of freshwater and into the ocean has been difficult to measure . In the Columbia River , which has an extensive network of hydroelectric dams , the decline in abundance of adult salmon returning from the ocean since the late 1970s has been ascribed in large measure to the presence of the dams , although the completion of the hydropower system occurred at the same time as large-scale shifts in ocean climate , as measured by climate indices such as the Pacific Decadal Oscillation . We measured the survival of salmon smolts during their migration to sea using elements of the large-scale acoustic telemetry system , the Pacific Ocean Shelf Tracking ( POST ) array . Survival measurements using acoustic tags were comparable to those obtained independently using the Passive Integrated Transponder ( PIT ) tag system , which is operational at Columbia and Snake River dams . Because the technology underlying the POST array works in both freshwater and the ocean , it is therefore possible to extend the measurement of survival to large rivers lacking dams , such as the Fraser , and to also extend the measurement of survival to the lower Columbia River and estuary , where there are no dams . Of particular note , survival during the downstream migration of at least some endangered Columbia and Snake River Chinook and steelhead stocks appears to be as high or higher than that of the same species migrating out of the Fraser River in Canada , which lacks dams . Equally surprising , smolt survival during migration through the hydrosystem , when scaled by either the time or distance migrated , is higher than in the lower Columbia River and estuary where dams are absent . Our results raise important questions regarding the factors that are preventing the recovery of salmon stocks in the Columbia and the future health of stocks in the Fraser River .
Many Columbia River salmon stocks are listed as threatened or endangered [1 , 2] , a result often attributed to the construction and operation of the Columbia River dams [3–5] . Here , we examine one phase of the lifecycle of Columbia River and Fraser River salmon stocks by comparing the freshwater survival of freely migrating salmon smolts down the extensively dammed Snake-Columbia River system with that of the same species migrating down the Thompson-Fraser River system , which lacks dams , using components of a large-scale acoustic telemetry system , the Pacific Ocean Shelf Tracking ( POST ) array . The Fraser and Columbia are the two largest rivers on the west coast of North America and have , or formerly had , some of the world's major salmon runs [6] . Concurrent with the start of construction of the Federal Columbia River Power System in 1938 , and especially following the completion of the last dam in the Snake River in 1975 , major declines in abundance of adult salmon returning to the Columbia have occurred [2 , 7] . Much of the salmon decline from historic abundance occurred as a result of overfishing and habitat loss before 1938 , when Bonneville , the first federal dam , became operational . However , continued sharp declines in abundance , particularly after 1977 , when the last of the Snake River dams was completed , have focused much attention on the operation of the dams [4 , 8] . A total of 13 salmon stocks in the Columbia system are now listed as threatened or endangered , with Snake River spring/summer Chinook ( Oncorhynchus tshawytscha ) and steelhead ( O . mykiss ) formally listed as threatened in 1992 and 1997 , respectively [1] . The poor adult return to the Snake River is variously ascribed to mortality on salmon smolts migrating to sea caused by the eight hydropower dams [1 , 2 , 9] , habitat disruption [2 , 10] , interactions with hatchery fish [11–13] , and changes in ocean climate affecting salmon survival after the smolts leave the river [14–16] . The Fraser River lacks dams and lies directly north of the Columbia River; within these two watersheds , the Thompson and Snake rivers form major tributaries and are located in similar climatic zones . At the end of the last ice age , salmon colonized the upper Fraser watershed ( including the Thompson River ) from the upper Columbia River , thus providing a relatively recent genetic linkage [17] . There are thus broad similarities between the two river systems , making for an interesting comparison of salmon survival during the freshwater phase of the juvenile outmigration in rivers with and without dams . Here , we compare the survival of two species of salmonid smolts in these rivers using acoustic and Passive Integrated Transponder ( PIT ) tags to measure survival from the upper reaches to the river mouth . Although identified as an important source of uncertainty [18 , 19] , an objective measure of freshwater survival has only recently become available with the construction of PIT tag detectors at dams on the Columbia and the advent of miniature radio and acoustic transmitters that can be implanted into migrating smolts . Beginning in the 1990s , the survival of migrating smolts between dams in the Snake-Columbia River watershed was measured using PIT tag technology [20–22] , a very short-range radio-frequency tag whose use is feasible because the dams channel smolts into close proximity ( ≤1 m ) to the PIT tag detectors . Prior to the development of miniature acoustic and radio tags with much greater range , it was not possible to measure the survival of migrating smolts in large rivers lacking dams , such as the Fraser , because there was no way to channel tagged fish into close proximity to receivers to detect their arrival . The POST array is a recently developed continental-scale acoustic tracking array that allows the movements and survival of individual fish to be measured directly [23–25] . Because it is based on an acoustic frequency that works in both seawater and freshwater , the technology allows tracking of fish as small as migrating salmon smolts ( ≥125 mm ) out to sea . We measured the survival of freely migrating hatchery-reared spring Chinook and wild steelhead smolts migrating out of tributaries of the undammed Thompson-Fraser River system in spring 2004–2006 by surgically implanting them with individually identifiable acoustic tags [26] and detecting the subsequent arrival of each surviving animal at the Fraser River mouth and then in the ocean ( Table 1; Video S1 ) . In all years , animals selected for tagging showed evidence of undergoing smolting , a suite of physiological changes associated with migration to sea [27]: skin color was changing to silver , and the behavior of hatchery-reared smolts showed evidence of searching for an exit from the tanks , with individuals repeatedly probing the tank walls . Median migration times after release were rapid , and smolts arrived at the Fraser River mouth , some 340 km distant , within a period of 3–17 d . Estimated survival , using the Cormack-Jolly-Seber ( CJS ) mark-recapture framework , ranged from 4%–67% [28] . The freshwater survival estimates for Thompson River smolts can be compared with two different measurements of survival of Snake River steelhead and Chinook smolts migrating down parts of the highly altered Snake-Columbia River system , which has eight major dams sited along the migration path ( Figure 1 ) . First , extensive measurements are available since 1997 of annual survival of PIT-tagged smolts migrating 516 km through the impounded section of the river from a release site in the Snake River at Lewiston , Idaho , through seven dams to the eighth and final dam at Bonneville on the Columbia River ( river Km [RKm] 223 ) [21] . Second , survival in the unimpounded lower river and estuary from Bonneville Dam to Astoria Bridge ( RKm 22 ) was measured in 2002–2004 using the same acoustic tag technology used in the Fraser River [29] , providing an estimate of survival for the final , free-flowing section of the river , and which is consistent with radio tag estimates . Radio telemetry cannot be used to measure survival in the estuary , where saltwater is present . However , our survival estimates using radio telemetry for the region above the estuary , but below Bonneville dam , for the 3-y period 2002–2004 are similar to the survival estimates reported here that were obtained using acoustic telemetry [29 , 30] . Finally , a whole-river estimate of survival was derived by multiplying the PIT tag estimates of survival in the impounded upper river by the acoustic tag estimates of survival for the lower free-flowing river , providing a combined estimate of survival covering the entire river to Astoria , Oregon , for Snake River steelhead in 2002 and 2003 and for Snake River spring Chinook in 2004 . Because the importance of the results described below partly depends on the relative performance of the PIT and acoustic tag methodologies for measuring survival , we also used the same acoustic tag technology in 2006 to provide a single measurement of survival of hatchery-reared Snake River spring Chinook salmon migrating the entire 910-km length of the Snake-Columbia River system from release at the Kooskia National Fish Hatchery to a listening line situated in the ocean across the full width of the continental shelf at Willapa Bay , 40 km north of the Columbia River mouth [24] . The 2006 experiment thus provides a directly comparable whole-river survival estimate to those made in the free-flowing Thompson-Fraser system using identical acoustic tags and surgical protocols , as well as allowing a direct comparison of the survival of acoustic-tagged smolts with independent studies of the survival of PIT-tagged Snake River Chinook in the impounded section of the river that were made in the same year ( [21]; see Video S1 for a comparison ) .
We first compared survival of PIT and acoustically tagged smolts in the impounded section of the Snake and Columbia rivers to assess survival of animals implanted with these different-sized tags in 2006 ( Figure 2; [21] ) . Survival of acoustically tagged Snake River spring Chinook smolts from the Dworshak Hatchery stock ( tagged and released at Kooskia Hatchery ) was statistically indistinguishable from the estimated survival of PIT-tagged Dworshak Hatchery Chinook in 2006 ( p > 0 . 05 ) , demonstrating that the PIT and acoustic tag methodologies provide similar survival estimates for freely migrating smolts in the impounded section of the river . Of note , the decline in smolt survival with distance , evident for the POST data , suggests that a simple model of a constant freshwater mortality rate in the Columbia may be appropriate , irrespective of location in the river . However , we can not rule out the possibility that the decline in survival with distance may be punctuated , rather than smooth , at finer spatial scales not resolved by the current POST array . Comparing survival between river systems , survival of smolts migrating the entire length of the river was either statistically indistinguishable ( spring Chinook ) between the undammed Thompson-Fraser River and the heavily impounded ( eight dam ) Snake-Columbia River system or slightly better in the Thompson-Fraser River ( steelhead; Figure 3A ) . When considered separately by river section , survival of Snake River smolts through the eight dams comprising the impounded section of the river down to Bonneville Dam was higher ( Chinook ) or statistically indistinguishable ( steelhead ) from the survival for the entire Fraser River . For both species , survival in the free-flowing lower section of the Columbia River was higher than the entire-river estimate for the Fraser River . These comparisons do not consider the distances and time that smolts must migrate to reach the location of the listening arrays in the two rivers; both values are substantially greater for Snake River smolts ( Table 1 ) . When scaled by either the migration distance or the median time for fish to reach the river mouth ( Figure 3B and 3C ) , average survival rates of spring Chinook and steelhead are significantly higher for Snake River smolts than in the undammed Thompson-Fraser river system for all comparisons ( p < 0 . 05 ) . In fact , all annual survival estimates for Snake River spring Chinook in either the dammed or undammed sections of the river exceed the average survival of spring Chinook in the Thompson-Fraser system , and all but one annual survival estimate for Snake River steelhead exceeds the average survival for the Thompson-Fraser . It is also notable that within the Columbia , in most comparisons , survival scaled by distance or time was higher for the upstream impounded section of the river than in the lower , free-flowing , section of the river . It is striking that our main finding , that survival is not worse in the Columbia despite the presence of an extensive network of dams , remains the same no matter how the data are analyzed .
Comparable survival estimates to the river mouth of the Columbia ( and higher survival rate when scaled by distance or time ) appear at odds with the conventional view that the hydropower system is one of the major current limitations to salmon recovery and that a return to more normative , pre-dam conditions will aid in recovering salmon populations [31] . The conclusion that whole-river survival in both river systems is similar partly depends on the assumption that smolts migrate similarly after tagging , that detection at the mouth of the Fraser River was not seriously underestimated , and that the tagged animals are representative of the two whole populations in each river . Several lines of evidence suggest that survival estimates in the Fraser-Thompson system are reasonable . First , migration to the river mouth of surviving smolts was rapid , with little evidence of delayed downstream movement , indicating a strongly directed migration . Second , systematic variation in the location and geometry of the lower Fraser River detection array over the 3-y period 2004–2006 always yielded detection efficiencies too high to alter the conclusion that survival rates scaled by either migration time or distance are higher on average for the Snake River populations . Because of the importance of this survival comparison to the public policy question concerning the impact of the dams on Columbia River salmon conservation , we made a particular effort in 2006 to build an extensive detection subarray consisting of six lines of paired acoustic receivers spaced a few kilometers apart within the lower Fraser River . This was done to ensure accurate detection estimates at the Fraser River mouth , as overestimation of the detection rate could potentially be incorrectly interpreted as poorer survival in the undammed river ( fewer fish reaching the mouth of the river than was actually the case ) . This alternative approach to calculating detection efficiency at the Fraser River mouth ( Table 1 ) also indicated a high detection efficiency , demonstrating that it is unlikely that survival in the Fraser was substantially underestimated . There is a paucity of data concerning interannual variation in smolt survival in either system we studied . Although we can speculate that yearly variation in survival is probably quite substantial , we do not have additional data upon which to judge variation . Third , in both rivers , the size of acoustic tags available limited our study to smolts from approximately the upper half of the size spectrum ( V7 tags: smolts ≥125 mm; V9 tags: smolts ≥140-mm fork length ) . The size of the tagged smolts whose survival we compared was therefore roughly comparable between the two rivers , but the average size of the source populations from which our tagged animals were selected was smaller in the Thompson River . As larger salmon are generally found to have better survival [32 , 33] , it seems likely that our current inability to tag the entire size range of migrating animals should be more likely to bias estimated survival upwards to a greater degree for the Thompson than the Snake River populations . Countering this empirical observation of higher survival in larger animals ( which is based on a much wider range of sizes than is representative of just-migrating salmon smolts ) , we note that our measured survival in 2006 using acoustic tags for Snake River smolts was almost identical to that obtained in the same year using PIT tags , which cover the entire size range of migrating smolts ( Figure 2 ) . Finally , a detailed examination of how Snake River survival varied in 2006 for successive 5-mm size groupings of acoustically tagged smolts [24] found no consistent relationship between smolt size and survival . Our tentative conclusion , therefore , is that survival rates of similar-sized animals are lower in the Thompson-Fraser system and that any dependence of survival on size would likely strengthen this conclusion when extended to the entire population . However , as technology develops further and reduces the size of acoustic tags , it would clearly be desirable to tag the entire range of smolt sizes and directly test this question . Evidence for stable rates of survival prior to smolt migration both before and after hydropower system completion [34] and the relatively high survival rates reported in this paper for migration down through the hydropower system and to the ocean ( 20%–30% ) sharply contrast with the very poor survival until the adults return from the ocean ( as little as 0 . 5% in some years [1 , 2 , 35] ) . The available data , therefore , indicate that although 25%–60% of smolts survive through the entire hydropower system , few return from the ocean . Thus , much of the mortality lies beyond the hydropower system , consistent with recent evidence that extensive effort put into freshwater habitat restoration may be insufficient by itself to conserve salmon populations [36] . Dam operation clearly had large impacts on the mortality of migrating salmon smolts in the 1960s and 1970s [3 , 37] , and changes to the hydropower system since then have improved survival substantially [22 , 38] . Our results suggest that survival through the hydropower system has now increased to levels similar to those experienced in both the undammed lower Columbia River and in the Fraser River , an important finding that was not technically possible before the development of the POST array . It remains unclear whether the similar rates of survival we measured result from past efforts to improve hydropower operations and reduce predators in the Columbia [39 , 40] or from unidentified problems in the Fraser River . Thompson River spring Chinook abundance has been stable or increasing since 1980 , whereas Thompson River steelhead are classified as of “Extreme Conservation Concern” [41] . However , as with many other steelhead populations located in southern British Columbia , the available evidence suggests the conservation status of Thompson River steelhead is primarily caused by poor marine survival after passage out of the river [42–44] . Poor smolt-to-adult survival is also observed for both species of Snake River smolts after migration out of the hydropower system; therefore , a common effect of ocean conditions on survival of salmon in both river systems seems likely . Modifications to dam design and operation have increased Columbia River smolt survival in the past 20 y [5 , 39 , 45 , 46] . Our initial results from the use of large-scale acoustic arrays over 5 y together with PIT tag data suggest that the overall migratory survival of salmon smolts in the Columbia and Fraser systems is now similar . This result is surprising , given that dams are often implicated as major barriers to recovery in the Columbia . However , our data do not address whether the possible delayed effects of hydropower system passage subsequently affects mortality after the fish leave the river for the ocean [9 , 48] , currently a contentious issue , nor is it clear whether survival in the Fraser River has changed during the last 100 y , as prior baseline measurements of survival are absent . There are several opposing inferences that can be made from our findings regarding the role of dams in preventing the recovery of salmon . We suggest that conservation efforts in the Columbia may be better directed towards understanding the effects of hydropower system passage on ocean survival , in addition to the extraction of small gains in survival at the dams .
Detailed surgical protocols are described elsewhere [49] . Briefly , individually identifiable Vemco ( http://www . vemco . com/products/transmitters/index_coded . php ) V9-6L acoustic tags ( 9-mm diameter , 20-mm long ) were surgically implanted into the abdominal cavities of smolts ≥140 mm in both the Thompson and Snake river stocks . Vemco V7-2L acoustic tags ( 7 mm in diameter , 22-mm long ) were surgically implanted into some groups of Thompson River smolts in the 125–140-mm size range; these groups are identified in Table 1 . Surgical procedures were annually reviewed and approved by institutional animal care committees . Elements of the POST acoustic array were used to measure the survival of the acoustically tagged smolts . POST is a large-scale passive acoustic telemetry system that sits on the sea floor and in sections of the Columbia and Fraser rivers ( http://www . postcoml . org ) . POST was designed [23] to provide a precise spatial geometry for a multitude of individually low-cost acoustic receivers that records the time of detection of individual acoustic tags; the programming of the acoustic tags was chosen to complement this geometry and to provide both high tag detection efficiencies and very long life for the tags . The full spatial scale of the array currently extends 2 , 500 km from Oregon to Alaska , and is described elsewhere [24]; data on the position of the entire POST array , including the Fraser River subarray positions , are reported in the POST database ( http://www . postcoml . org/page . php ? section=database ) , as are the detection histories of all Thompson River tagged smolts and the Snake River smolts tagged in 2006 . An animation of the movements of some of the tagged smolts on which this paper is based is shown in Video S1 . Summary data on the surgical procedures and receiver arrays used for Columbia River smolts in years prior to 2006 are similar and are described in [30] . Detection efficiencies at each line of acoustic receivers within the Fraser River were estimated for each type of tag and year , as V7 tags have a lower acoustic power output than V9 tags ( 136 versus 149 dB re: 1 μPa at 1 m ) and the geometry of the Fraser River mouth array varied between years . Aggregate detection efficiency of the ocean listening lines ( all years combined ) was estimated as 89 . 5% for V9 tags and 71 . 4% for V7 tags , thus providing a good estimate of total tagged smolts migrating out of the Fraser River . All Columbia River tagging used V9-6L tags; data and protocols for years prior to 2006 are described in [30]; in 2006 , the array was extended upstream as far as the Snake River and out into the ocean ( these data are available from the POST database ) . Dworshak Hatchery spring Chinook , a Snake River stock , were transferred to Kooskia National Fish Hatchery in the spring of 2006 and held until surgical implantation with acoustic tags and subsequent release at Kooskia . Snake River smolts were double-tagged with a PIT tag in 2006 to ensure that they were not inadvertently collected at the dams for transportation in barges and were thus forced to migrate the full length of the river . We compare their measured survival using the acoustic array with the survival of the Dworshak stock of Snake River spring Chinook smolts independently measured using the PIT tag system in the same year [21] . Survival estimates in the Columbia River measured using PIT and POST acoustic tags were regressed against distance from release site , L , after log-transformation using a fixed intercept , S ( L ) = exp ( −zL ) , yielding an estimate of the survival rate per river kilometer . PIT tag estimates of survival were measured at the dams; acoustic tag survival estimates were derived from the four in-river detection subarrays extending from the Snake River to just below Bonneville Dam plus the ocean listening line at Willapa Bay ( see Figure 1 ) . Regression coefficients of the survival rate of Dworshak Hatchery smolts were statistically indistinguishable between PIT and acoustically tagged smolts ( p > 0 . 05 ) . More extensive descriptions of the statistical measurement of survival using the acoustic array and the performance of the array is available in Text S1–S3 . The MatLAB code used to generate the Monte Carlo statistical comparisons , and the frequency histograms of the generated data are reported in Text S4 .
|
Miniature electronic technologies now allow researchers to track a salmon's migration from its birthplace in a river's headwaters in the Rocky Mountains to the North Pacific , opening a window on the mysteries of migration and survival . Surprisingly , outward migrating salmon ( smolts ) have similar survival during migration down dammed and undammed rivers , challenging widely held notions about factors affecting salmon abundance . Elements of the large-scale POST ( Pacific Ocean Shelf Tracking ) acoustic telemetry array revealed the migrations . Although salmon smolt survival to the Pacific Ocean was comparable in both the dammed Columbia and undammed Fraser rivers , it was higher in the Columbia once distance or travel time was taken into account—and higher within the hydropower system than below the dammed section . There is not yet enough evidence to determine whether ( 1 ) the Fraser has a problem that reduces salmon survival to that of a heavily dammed river or ( 2 ) factors other than dams play a larger , unsuspected role in salmon survival . Wherever future research leads on those questions , the new fish tagging technology has demonstrated itself as a useful tool for obtaining objective scientific data with important value in a number of public policy arenas .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biotechnology",
"ecology"
] |
2008
|
Survival of Migrating Salmon Smolts in Large Rivers With and Without Dams
|
Grid cells have attracted broad attention because of their highly symmetric hexagonal firing patterns . Recently , research has shifted its focus from the global symmetry of grid cell activity to local distortions both in space and time , such as drifts in orientation , local defects of the hexagonal symmetry , and the decay and reappearance of grid patterns after changes in lighting condition . Here , we introduce a method that allows to visualize and quantify such local distortions , by assigning both a local grid score and a local orientation to each individual spike of a neuronal recording . The score is inspired by a standard measure from crystallography , which has been introduced to quantify local order in crystals . By averaging over spikes recorded within arbitrary regions or time periods , we can quantify local variations in symmetry and orientation of firing patterns in both space and time .
Neurons in the entorhinal cortex are often classified based on their spatial firing patterns . Grid cells , in particular , have attracted a lot of attention , because of their highly symmetric hexagonal activity [1 , 2] . To decide whether a given cell should be classified as a grid cell or a non-grid cell [3 , 3–5] , researchers need a grid score: a number that quantifies the degree of hexagonal spatial symmetry in the firing pattern of the cell . Since the first reports of grid cells , the community has developed and refined a standard grid score [6–9] , which relies on the following procedure . Spike locations are transformed into a rate map . The peaks in the spatial autocorrelogram of this rate map are used to obtain measures of the grid spacing and the grid orientation . The autocorrelogram of the rate map is then cropped , rotated and correlated with its unrotated copy . A grid score is finally obtained from the resulting correlation-vs-angle function at selected angles ( for details , see S1 Appendix ) . This procedure assumes a consistent hexagonal pattern throughout the arena and results in a single grid score for the entire firing pattern . While this grid score has many benefits , it is inherently global and ignores local properties of firing patterns . Recent studies have shown , however , that the firing patterns of grid cells show interesting local effects , such as drifts in orientation [10] or local defects of the hexagonal symmetry [11 , 12] . These results call for new methods for quantifying the local symmetry and the local orientation of grids . Here , we introduce a method that assigns both a grid score and an orientation to each individual spike of a neuronal recording . This spike-based score is inspired by a standard measure from crystallography [13] , which has been introduced to quantify local order in crystals . By averaging over spikes recorded within arbitrary regions or time periods , we can quantify local variations in symmetry and orientation of firing patterns in both space and time . A high spatial and temporal resolution can be achieved by averaging over multiple cells . To illustrate the potential of our method , we apply it to both artificially generated data and neuronal recordings from grid cells . In the following , we first introduce the method in detail . Comparing it to the conventional correlogram-based grid score [7] , we show that both methods perform similarly when quantifying and classifying grid properties globally . We then illustrate the benefits of our spike-based score to analyze and visualize local properties of spike patterns in space and time .
The angular symmetry of an arrangement of N neighbors around a given atom in a crystal can be quantified by a complex number [13]: ψ ( M ) ≔ 1 N ∑ l = 1 N exp ( i M ϕ l ) , where i is the imaginary unit and ϕl is the angle between an arbitrary reference vector and the vector pointing from the atom in question to its l-th neighbor [13] . In the following , we will use the horizontal axis from left to right as the reference vector ( Fig 1a ) . M is a natural number that characterizes the symmetry of interest . For instance , M = 6 corresponds to a hexagonal arrangement of atoms and M = 4 to a quadratic arrangement . The computation of this score can be visualized as an addition of vectors in the complex plane ( Fig 1a ) . The absolute value |ψ ( M ) | ranges from 0 to 1 , and larger values indicate a more pronounced M-fold symmetry . If the N neighboring atoms lie on the corners of a regular M-polygon , the angle 1Marg ( ψ ( M ) ) is the orientation of this polygon ( Fig 1a ) . The function arg ( x ) maps to values between −π and π radians , so the resulting orientation is between −π/M and π/M . For hexagonal symmetry , this corresponds to values between −30 and 30 degrees . The ψ measure has been introduced to quantify symmetry in crystalline structures where neighborhood is clearly defined . For example , the nearest neighbors of a reference atom are all atoms within a distance of one lattice constant . For a given spike in a grid pattern , however , the nearest neighbors are most likely other spikes from the same grid field . To assess whether a given spike is part of a symmetric grid , we are hence interested not in neighboring spikes , but in spikes in neighboring grid fields , i . e . , spikes that are roughly one grid spacing away from the spike in question . To identify these relevant spikes , we define a neighborhood shell—a two dimensional annulus ( Fig 1b ) . We then center this neighborhood shell at the location of the spike in question and consider all spikes within the shell as its neighbors . Assuming that the grid spacing does not change drastically across the environment , we use the same shell size for all spikes . To identify a suitable size of the shell , we use the histogram of pairwise distances between all spike locations . For periodic firing patterns , this histogram has pronounced peaks ( Fig 1b ) . The first peak characterizes the size of the grid fields . The second peak corresponds to the grid spacing ℓ , and further peaks arise at multiples of the grid spacing . We therefore set the inner radius of the neighborhood shell to 5 6 ℓ and the outer radius to 7 6 ℓ ( Fig 1b ) . To reduce the noise arising from finite numbers of spikes , we determine the peaks from a smoothed histogram , obtained by applying a Gaussian filter with a standard deviation of 1% of the maximal pairwise distance . The suggested shell width of 1 3 ℓ is motivated by the typical size of grid fields , but the suggested grid score is robust to changes of the shell width . Scaling the shell width with the grid spacing is commensurate with the experimental finding that the ratio of grid spacing and grid field size is constant [15] . The detection of the correct neighborhood shell as the second peak of the histogram of pairwise distances is a crucial step in the computation of the grid score . For noisy grid patterns , the first peak of the histogram—which indicates the size of the grid fields—is sometimes hard to detect , leading to a potential misclassification of the second peak as the first , and consequently a wrong neighborhood shell . In data sets where this is the case , we use the following remedy: We define a rough cutoff distance that is larger than the grid field size and smaller than the grid spacing , and use the first peak above the cutoff to determine the shell size . For the analyzed grid patterns , taking 15% of the side length of the arena for this cutoff worked reliably . In the dataset with very noisy and elliptic grids , peak detection can fail for many cells . In this case , we exploit the fact that nearby grid cells have a similar grid spacing [2] and determine the neighborhood shell from less noisy grid cells of the same module . Whenever we use either of these remedies , we highlight it in the main text . A large | ψ k ( 6 ) | value indicates that the neighbors of the k-th spike are arranged on the corners of a hexagon , but it does not require that all corners of the hexagon are occupied . If , for example , the firing fields of a cell are arranged on a line , this leads to high |ψ ( 6 ) | values although the structure does not resemble a grid ( Fig 1c ) . This can lead to a wrongful classification of a cell as a grid cell , i . e . , to false positives . To avoid high local grid scores without an actual 6-fold symmetry , we compute | ψ k ( M ) | for a range of symmetries ( M ∈ {2 , … , 7} ) and only assign a non-zero grid score to spike k if the hexagonal symmetry is the strongest: ψ^k≔{| ψk ( 6 ) | , if| ψk ( 6 ) |>| ψk ( M ) |∀M∈{2 , 3 , 4 , 5 , 7 }0 , otherwise . We refer to ψ ^ k as the grid score of spike k . Comparing to symmetries with M > 7 did not have an effect on our results , so we restricted M to a maximum of 7 . A global grid score Ψ and the global grid orientation Θ for a given cell is calculated by averaging over the respective properties of the individual spikes . A summary of the algorithm is shown in Fig 2 . In the following , we compare the suggested local grid score to a conventional correlation-based grid score . Details on the calculation of this score , on the shuffling procedure for grid cell classification , and on the generation of artificial spike data can be found in S1 Appendix .
Grid scores are typically used for two purposes: Quantifying the hexagonality , i . e . , the gridness , of the spike map and classifying a cell as a grid cell . Given that the patterns of grid cells are rarely perfectly hexagonal , a grid score should classify a cell as a grid cell even if the hexagonal pattern is weakly distorted . For example , a grid score needs to be robust against small shifts of grid field locations or elliptic distortions like shearing . At the same time , a grid score should quantify such distortions by returning lower values for stronger distortions . So far we assumed that gridness is an arena-wide property of a firing pattern . We did not consider local defects . But recent experiments have shown that grid patterns are often locally distorted . Such distortions have been observed in both asymmetric [11 , 12] and quadratic environments [10 , 20] , and are presumably caused by boundary effects . Moreover , it has been shown that grid patterns recorded in two arenas separated by a wall do not form a coherent grid , but merge into a more coherent pattern when the wall is removed [19] . We now show that our spike-based grid score is suitable to quantify and visualize such local properties . We first illustrate this for generated spike data and then for experimental recordings . To generate an example of a grid pattern with local defects , we first generated a perfect grid and then added random vectors to the field locations ( cf . Fig 3a , see S1 Appendix for details ) only in the eastern half of the arena ( Fig 6a and 6b ) . To analyze gridness locally , we divide the arena into partitions and compute the Ψ score within each partition , i . e . , the mean over all ψ ^ k scores for all spikes k within the given partition ( Fig 6c ) . Averaging over multiple cells highlights that the gridness decays along the west-east axis but stays stable along the south-north axis , and this trend is already visible for individual cells ( Fig 6c and 6d ) . Note that due to the strong noise on the east side of the arena , we detect the neighborhood shell from the histogram of pairwise spike distances , using the first peak above a cutoff ( see Methods ) . We now apply the same analysis to recent experimental data . Wernle et al . recorded grid cells from rats that explored the northern and southern half of a 2m × 2m box [19] . The two halves were initially separated by a wall . The cells fired in grid patterns independently in both halves of the arena ( Fig 6e ) . Then the authors removed the wall and let the rat explore the entire arena . At the southern and northern boundaries , the grid fields typically stayed at the same locations as before wall removal . Close to the former location of the wall , the grid fields moved—and sometimes even merged—to form a more coherent grid pattern ( Fig 6f ) . Quantifying the local increase of coherence requires a local grid score . We therefore use our spike-based grid score to show that the rearrangement of grid fields after wall removal indeed leads to a local increase of gridness . Given that the firing patterns of the grid cells are noisy and strongly elliptic , it is difficult to determine the peaks in the histograms of pairwise spike distances ( Fig 6e and 6f ) . We thus set the central distance of the neighborhood shell manually to ℓ = 65cm—a value that we obtained by visual inspection of the histograms of pairwise spike distances of all 11 recorded cells , after wall removal . We divide the arena into three equal sized partitions . For south-north partitioning ( Fig 6g ) , the grid score is typically lowest in the partition that contains the separating wall ( Fig 6h ) . After wall removal , the grid score increases in all partitions , but the increase is most prominent in the partition that contained the separating wall ( Fig 6h ) . In contrast , a west-east partitioning does not lead to a single partition with predominantly low grid scores . Moreover , the increase after wall removal is similar for all west-east partitions ( Fig 6i and 6j ) . Our analysis thus quantifies the qualitative observation that two initially independent grids locally rearrange to form a more coherent grid . For less elliptic grid patterns , we assume our quantitative results to be even more pronounced . In summary , a spike-based grid score visually highlights local grid defects . Averaging spike scores in arbitrary partitions and potentially over many cells enables a systematic analysis of grid defects on a spatial scale smaller than the grid spacing . Early work on grid cells has shown that the firing patterns of anatomically nearby cells typically have a similar orientation [2] , which tends to align with a cardinal axis of the environment [10 , 11] . While these findings suggest a constant grid orientation , experiments from the same group have shown that the orientation can drift within the arena or exhibit local distortions [10 , 20] . To quantify these experiments , a local measure of grid orientation is needed . In the following , we study local grid orientation using the angle θk of spike k ( see the algorithm in Fig 2 for the definition ) on generated ( see S1 Appendix ) and recorded firing patterns that were kindly provided by Stensola et al . [10] . When each spike is color-coded according to its grid orientation θk , the spikes of firing fields have the same color for a perfect hexagonal grid—except the spikes at the edge of a grid field ( Fig 7a ) . In contrast , the color-coding shows a sudden switch if the orientation is changed abruptly ( Fig 7b ) or a color gradient if the grid orientation drifts continuously across the arena , both in generated data ( Fig 7c ) and in recordings ( Fig 7d ) . If the drift in orientation is along the south-north axis of the box , the gradient follows this axis ( Fig 7c and 7d ) . This can be highlighted by partitioning the arena and calculating the mean value of the orientation θk for all spikes k within each partition ( Fig 7c and 7d ) . To highlight an orientation drift recorded in multiple cells , we average over the the orientation θk for spikes from several cells . For the recordings of Stensola et al . [10] , this shows that the orientation increases monotonically from south to north , but varies less systematically from west to east , for a set of 5 cells recorded in the same rat ( Fig 7e and 7f ) . Note that changes in grid orientation distort the correlogram , which can lead to low ρ scores ( Fig 7b and 7c ) , even if the grid-like structure is clearly apparent . In contrast , the Ψ score is robust to distortions in orientation because of its locality . In summary , a spike-based orientation measure highlights local distortions in grid orientation . Averaging spike-based orientations , using arbitrary spatial partitionings , allows to further classify the distortion . Whether or not the spikes of a cell form a hexagonal spatial pattern can vary in time . For example , the activity of grid cells loses its periodicity when the light is turned off [21 , 22] . A decay of grid-like firing patterns can also be triggered by internal modifications , like the deactivation of hippocampal drive to the entorhinal cortex [23] . In the following , we show how the spike-based grid score can be used to study changes in grid cell activity with high temporal resolution . The standard procedure for computing a grid score in a given time interval is to take the spike locations from all spikes that occurred in this interval and compute the grid score for the resulting pattern . While this can be done with both the correlogram-based score and the spike-based score , it introduces constraints on the temporal resolution , because the temporal window must contain a sufficient coverage of the arena to obtain a meaningful score . This problem can be circumvented with a spike-based grid score , because for a given spike map , each spike k has a grid score ψ ^ k and a time tk . In other words , the grid score ψ ^ k at time tk quantifies the contribution of spike k to the grid pattern formed by all spikes . Just like the global grid score Ψ is calculated by averaging over the individual ψ ^ k values of all spikes , a temporally local grid score can be obtained by averaging the ψ ^ k values of the spikes in an arbitrary time interval . The resulting average quantifies how much the spikes in this interval contribute to the grid in the complete recording . To demonstrate that the spike-based grid score can characterize temporal changes in grid patterns , we consider generated data with a sudden transition from spatially homogeneous to grid cell activity ( Fig 8a ) . Plotting the individual grid scores as a function of spike time highlights the transition to the grid pattern , although the grid scores ψ ^ k of the individual spikes are computed using the locations of all spikes ( Fig 8a and 8b ) . The transition is emphasized by temporal smoothing ( Fig 8c ) . Requirements for the smoothing filter are discussed later . Note that such an analysis first of all requires that a grid pattern can be detected in the full recording , so that a meaningful neighborhood shell can be identified . To facilitate the detection of the neighborhood shell in the presence of noise , we detect the first peak above a cutoff in the histogram of pairwise spike distances ( see Methods ) . We now use this method to analyze spike patterns from cells in mice that forage in an arena in alternating trials of darkness and light [21] . For the recorded grid cells ( Fig 8d ) , the filtered time evolution of grid scores ψ ^ k is strongly correlated with darkness and light—grid scores tend to be higher in light trials ( Fig 8e , see S1 Appendix for details on the data preparation ) . The Pearson correlation coefficient between lighting condition and the smoothed grid scores ( r = 0 . 57 and r = 0 . 53 for single cells and r = 0 . 69 for the average over smoothed grid scores of two cells ) indicate that only few simultaneously recorded cells are necessary to reach confident conclusions about temporal variations in the quality of grid patterns or contextual influences on grid cell activity . But how quickly does the grid pattern decay when the light is turned off ? How rapidly does it reappear when the light is turned on again ? Previously , these questions were answered with a temporal resolution of 10 seconds using map similarity [21] ( see below for a discussion of map similarity ) . We now show that we can reach the same temporal resolution using the spike-based grid score . To this end , we create a reference spike map for each recorded grid cell in the data [21] , which comprises the locations of all spikes that were fired during light trials . Note that these subsets of spikes show a much clearer grid pattern than that of spikes during darkness ( Fig 8f ) . We then quantify how well the spikes of a given cell fit into the grid pattern of the cell’s reference spike map , by adding every spike k individually to the reference spike map and computing its ψ ^ k score as usual . We then average the ψ ^ k score of spikes that occur at similar times during light/dark trials . To this end , we divide each 2 minute trial into 12 × 10 second blocks . This division is done separately for light and dark trials . We then average the grid scores ψ ^ k over all spikes and over 73 recorded grid cells within each 10 second block . Plotting this average as a function of time shows that grids decay within 30 seconds after the light is turned off ( Fig 8g ) . Grids need a similar time to reappear after the light is turned on ( Fig 8g ) . In summary , the spike-based grid score readily contains temporal information which can be used to study distortions of the grid pattern in time . The time evolution of the grid score can be correlated with any other time varying variable to unravel its influence on the grid pattern , e . g . , light conditions , running speed or the concentration of neuromodulators .
We have shown spatially local grid scores in sub-partitions of the arena and temporally local grid scores in time intervals . The resulting values can be misleading if the spatial subinterval is too small or if the temporal subinterval is too short . For example , if we study the local grid score of a single cell in a partition that does not contain a grid field of that cell , the mean over all spike scores in that partition will be very low , suggesting a local defect . Such sampling biases could be mitigated by making the partition larger ( or the interval longer , for temporal analyses ) , at the cost of resolution . For a single cell , the resolution is thus bounded by the grid spacing . However , a reduction of sampling biases without a loss of resolution can be obtained by averaging over multiple cells . It would be beneficial if these cells have different spatial phases , because this ensures the presence of a grid field in each partition and thus reduces sampling biases . Averaging over many cells thus enables a higher spatial resolution than the grid spacing and a higher temporal resolution than the time the rat needs to sample a sufficient number of grid fields in a single cell . The correlogram-based ρ score can be made more local by computing it in spatial subregions within the arena . A window can also be used to compute the sliding average of the ρ score in the entire arena [20] , resulting in a spatial map of local grid scores . Selecting the size of the window is subject to a trade-off: while a small window is necessary for a high spatial resolution , the ρ score will fail to detect a grid if the window is too small—e . g . , smaller then the grid spacing . In contrast , the spike-based score can be computed in arbitrary spatial partitionings ( Fig 7 ) by averaging over the spikes in each partition . The side length of a partition can even be smaller than the grid spacing , in particular if multiple cells are used to compute the score in each partition ( see above ) . The flexibility to use partitionings of an arbitrary shape , e . g . , trapezoidal or triangular , can be important to study grid cells in environments with complex geometries [11 , 12] . We have shown that a spike-based grid score enables a straightforward analysis of the temporal stability of grid patterns and illustrated this on recorded data from mice foraging through an arena in light and darkness [21] . In our analysis of the decay and reappearance of grid cell activity ( Fig 8 ) , we reproduced the results shown in [21] , where the authors used map similarity for their analysis . In the map similarity measure , a rate map is correlated with a reference map . As a reference map , the authors used recordings during times of light . In each 10 second block , they collected all spike locations from that block in all trials ( separating dark and light trials ) and used these spike locations to create a rate map . If the correlation of the resulting rate map with the reference map is high , the pattern is similar , indicating that the grid has not decayed yet , and vice versa . For grid patterns without defects in field locations and with similar firing rates for different grid fields , map similarity and our measure should lead to similar results . However , our measure focuses on how much a spike is part of the grid of the reference map . In contrast , map similarity only quantifies how much spikes are part of the rate pattern of the reference map , irrespective of whether it forms a grid . If for example , the reference map has a firing field that does not belong to the hexagonal pattern , spikes that are located in this firing field will lead to low values in our measure . In contrast , map similarity assigns high values to activity within any firing field of the reference map , be it a grid field or a defect . We have shown that the spike-based Ψ score is more robust to elliptic deformations of grid patterns than the correlogram-based ρ score ( Fig 4c ) . This robustness can be crucial for the detection of grid cells , because many grid patterns are of elliptic shape . Quantifying the local improvement of a grid pattern in the scenario where two grids coalesce in a contiguous environment ( Fig 6 ) was only possible because grid patterns could be detected , even though they were strongly elliptic . A correlogram-based grid score can also be used with elliptic grid patterns , but requires an explicit compensation of the ellipticity . To this end , the ellipticity is detected and removed from the correlogram before the grid score is computed , by compressing the correlogram along the principal axis of the elliptic distortion [24] . In principle , we could apply such a transformation to all spike locations and then compute the spike-based grid scores . We have shown , however , that such an additional step is not necessary as long as the ellipticity is not too strong . If a firing pattern is grid-like , but with different grid spacings in different parts of the arena , we cannot compute the neighborhood shell from all pairwise distances . Instead , the neighborhood would need to be determined individually for each spike , using the histogram of pairwise distances only between a given spike and all other spikes . Since we are not aware of local distortions to the grid spacing , we did not consider this . For grid fields with six neighboring fields , spikes in the center of the field receive the highest grid score and grid scores decay isotropically in all direction . In contrast , in grid fields at the boundaries , spikes with high ψ ^ k form a curved stripe ( S1 Fig ) . The reason is the following: Being in the boundary grid field , moving towards the central neighbors perturbs the 6-fold symmetry stronger than moving tangentially to it ( S1 Fig ) . Such a directional dependence is not present in grid fields that have neighboring grid fields on all sides . We focused on hexagonal patterns , i . e . , grids of 6-fold symmetry . Experiments suggest that there are also band-like firing patterns , i . e . , 2-fold symmetry [3]—though their existence has been questioned [25 , 26] . Moreover , boundary effects in arenas of complex shape could cause grid deformations that result in other symmetries . The suggested spike-based grid score can be readily applied to any M-fold symmetry . For example , to study quadratic symmetry , we would replace M = 6 by M = 4 and compare to the remaining symmetries , now including 6 ( see the algorithm in S1 Appendix ) . Depending on the symmetry , the radius of the neighborhood shell needs to be chosen differently . For example , for quadratic symmetry , results are improved if the neighborhood shell is set to 2/3ℓ instead of ℓ , to only incorporate the four nearest neighboring fields ( Fig 9d ) . For band-like symmetry , the shell radius should be ℓ/2 , so that the next band is not included in the shell ( Fig 9e ) . As the experiments on spatially tuned cells become more complex , new properties will be unveiled . A spike-based grid score could be used as a single unifying method to classify grid cells and to analyze their local and temporal characteristics . The presented method could further be useful for the analysis of grid patterns in other modalities [27 , 28] or in other contexts , e . g . , in the anatomical arrangement of grid cells [29] or retinal ganglion cells [30] .
|
Grid cells are neurons in mammals whose activity depends on the location of the animal in a striking way: grid cells fire spikes at multiple locations that form a symmetric lattice of repeating hexagons . Recent experiments have shown that the symmetry of these patterns is not as stable as initially thought . For example , patterns show local distortions or drifts in orientation . Here , we propose a method to visualize and quantify such local distortions directly from the recorded spike locations . Our method is inspired by an approach used in crystallography and we show that it reliably detects distortions in grid patterns . Moreover , we demonstrate that it can be used to study the stability of grid patterns not only in space , but also in time . Our method enables researchers to analyze the spatial symmetry of neuronal activity in more detail and could thus contribute to the understanding of spatial representations in mammals .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"ecology",
"and",
"environmental",
"sciences",
"cell",
"cycle",
"and",
"cell",
"division",
"applied",
"mathematics",
"cell",
"processes",
"condensed",
"matter",
"physics",
"social",
"sciences",
"light",
"neuroscience",
"electromagnetic",
"radiation",
"simulation",
"and",
"modeling",
"algorithms",
"crystals",
"mathematics",
"materials",
"science",
"vision",
"crystallography",
"research",
"and",
"analysis",
"methods",
"solid",
"state",
"physics",
"animal",
"cells",
"soil",
"science",
"soil",
"perturbation",
"physics",
"psychology",
"cellular",
"neuroscience",
"cell",
"biology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"materials",
"sensory",
"perception"
] |
2019
|
A local measure of symmetry and orientation for individual spikes of grid cells
|
Mycobacterium tuberculosis ( Mtb ) virulence is decreased by genetic deletion of the lipoprotein LprG , but the function of LprG remains unclear . We report that LprG expressed in Mtb binds to lipoglycans , such as lipoarabinomannan ( LAM ) , that mediate Mtb immune evasion . Lipoglycan binding to LprG was dependent on both insertion of lipoglycan acyl chains into a hydrophobic pocket on LprG and a novel contribution of lipoglycan polysaccharide components outside of this pocket . An lprG null mutant ( Mtb ΔlprG ) had lower levels of surface-exposed LAM , revealing a novel role for LprG in determining the distribution of components in the Mtb cell envelope . Furthermore , this mutant failed to inhibit phagosome-lysosome fusion , an immune evasion strategy mediated by LAM . We propose that LprG binding to LAM facilitates its transfer from the plasma membrane into the cell envelope , increasing surface-exposed LAM , enhancing cell envelope integrity , allowing inhibition of phagosome-lysosome fusion and enhancing Mtb survival in macrophages .
Tuberculosis is the second leading cause of death from an infectious disease worldwide ( http://www . who . int/mediacentre/factsheets/fs104/en/index . html ) . The causative agent , Mycobacterium tuberculosis ( Mtb ) , is an intracellular bacterial pathogen that persists in phagosomes of infected macrophages . Mtb expresses a thick , waxy cell envelope of low permeability that contributes to antibiotic resistance and contains components ( e . g . glycolipids , lipoglycans and lipoproteins ) that play critical roles in regulating host responses and promoting survival of the pathogen [1]–[9] . Greater understanding of tuberculosis pathogenesis is needed as a foundation for development of drugs or vaccines to prevent or treat tuberculosis . Cell envelope components and the enzymes and transporters that control their synthesis and assembly serve as both determinants of pathogenesis and attractive targets for drug design . The cell envelope of Mtb is rich in lipids and carbohydrates [1]–[3] , [10] , including lipoglycans such as lipoarabinomannan ( LAM ) and lipomannan ( LM ) , and the glycolipid phosphatidyl-myo-inositolmannosides ( PIMs ) . The PIMs include a core acylated glycolipid structure that is incorporated into the more highly glycosylated lipoglycans , LM and LAM , which are synthesized on the on the periplasmic face of the plasma membrane from PIM precursors by the elongation of mannan and arabinan chains [11] . LAM is essential for Mtb survival; the synthesis of LAM and cell envelope arabinans is targeted by the anti-mycobacterial agent ethambutol and DprE1 inhibitors currently under development [12] . LAM inhibits Mtb phagosome-lysosome fusion , providing a mechanism for Mtb evasion of host defense [13]–[18] . This effect is specific to ManLAM , the mannose capped LAM found in slow growing strains that are pathogenic in humans or other host species ( e . g . Mtb , M . bovis BCG , M . leprae ) , and is not produced by phospho-myo-inositol-capped LAM ( PI-LAM ) found in M . smegmatis and M . fortuitum [13] . Recent studies have identified key steps in the biogenesis of LAM [1] , [19] , but the mechanisms by which LAM is assembled and organized into the cell envelope are unknown . A model for Mtb cell envelope structure is presented by Kaur et al [1] . Mtb expresses numerous , functionally diverse lipoproteins , many of which are contained in the cell envelope [20]–[22] . Mutation in lipoprotein processing by the disruption of lipoprotein signal peptidase lspA attenuates Mtb virulence [23] , causes cell envelope permeability defects and increases sensitivity to antibiotics [24] , suggesting potential roles of lipoproteins in maintaining Mtb cell envelope function . A recent proteomic study of the Mtb cell envelope identified LprG among the top 10 most abundant lipoproteins [25] . LprG ( Rv1411c ) forms an operon with Rv1410c , which encodes a putative efflux pump membrane protein , p55 [26] . Genetic deletion of this operon results in decreased virulence of Mtb in mice , and its deletion in M . smegmatis results in abnormal cell envelope morphology and permeability , suggesting its involvement in cell envelope biogenesis and/or maintenance of cell envelope integrity [26]–[30] . Despite evidence for the importance of LprG to Mtb virulence , its function remains uncertain . Mtb LprG binds to LAM , LM and PIMs when expressed in M . smegmatis , and acyl moieties of these lipoglycans and glycolipids bind in a hydrophobic binding pocket of LprG [31] . Glycolipid binding was abrogated with a mutated version of LprG ( LprG-V91W ) with altered hydrophobic binding pocket structure [31] . In the studies reported here , we expressed acylated LprG in the native species , Mtb , to determine the function of LprG in Mtb . Surface plasmon resonance ( SPR ) binding assays were used to determine the kinetics and affinities for binding of different Mtb lipoglycans and glycolipids to LprG; the results revealed novel interactions between saccharide chains of LAM and LM and a site on LprG outside of the hydrophobic pocket , as well as interactions of the lipoglycan acyl chains with the LprG hydrophobic pocket . Furthermore , an LprG null mutant Mtb ( ΔlprG ) had reduced levels of surface-exposed LAM and decreased ability to inhibit Mtb phagosome-lysosome fusion , a LAM-dependent effect that is associated with Mtb survival and virulence . We propose that LprG binding to LAM facilitates the transfer of LAM from the plasma membrane ( its site of synthesis ) to the Mtb outer membrane . This leads to LAM expression at the cell surface , contributing to the ability of Mtb to inhibit phagosome-lysosome fusion and survive in macrophages . Our results highlight the importance of LprG and cell envelope components in the survival and virulence of Mtb and reveal a potentially important target for therapeutic intervention .
These studies explored for the first time the lipoglycan/glycolipid binding function of the lipoprotein LprG expressed in Mtb . Previous studies of LprG expressed in a fast-growing saprophytic mycobacterial species , M . smegmatis , had revealed a lipoglycan/glycolipid binding function [31] , but the potential roles of LprG in lipoglycan biology in Mtb remained unexplored . The current studies provide a detailed understanding of lipoglycan/glycolipid interactions with LprG and reveal a novel role for LprG in determining the localization of LAM in the cell envelope . In order to study the association of lipoglycans and glycolipids with LprG in Mtb , the lipoprotein was expressed with a hexahistidine tag and isolated from Mtb lysate by Ni affinity and anion exchange chromatography . These studies used acylated LprG , rather than non-acylated LprG ( NA-LprG ) , which was used in prior studies , as acylation state is likely to affect subcellular localization and intersection with specific glycolipid and lipoglycan species ( acylation is linked with translocation across the plasma membrane ) . SDS-PAGE was used to analyze LprG and any co-purifying molecules . Silver stain and Western blot ( Fig . 1A ) with anti-hexahistidine antibody showed a single band at ∼25 kDa , corresponding to the approximate molecular weight of LprG . Western blot with CS-35 anti-LAM monoclonal antibody revealed a characteristic diffuse band at ∼37 kDa that was associated with acylated Mtb LprG expressed in both Mtb and M . smegmatis ( Fig . 1B ) . A polyclonal anti-Mtb antibody that detects LAM , LM and PIMs revealed bands at the expected positions for these lipoglycans ( ∼37 kDa , ∼18 kDa and ∼10 kDa; corresponding to positions of these lipoglycan/glycolipid standards on the gel ) in association with Mtb LprG expressed in Mtb and M . smegmatis but not E . coli ( Fig . 1C and 1D ) . The detection of LAM was more intense than the detection of LM and PIMs , and the association of all three lipoglycans/glycolipids was greater when LprG was expressed in Mtb relative to M . smegmatis ( Fig . 1B , C ) . These lipoglycans/glycolipids were not associated with Mtb-expressed acylated LprA , a lipoprotein that is homologous to LprG ( hexahistidine-tagged LprA was expressed and purified using the same protocol as for LprG , see Materials and Methods ) ; this control demonstrates the specificity of lipoglycan/glycolipid binding to LprG . These studies establish that acylated LprG binds LAM and LM in Mtb . While prior studies showed that binding of glycolipids to LprG involved insertion of glycolipid acyl chains into the LprG hydrophobic pocket ( blocked by mutation of the hydrophobic pocket in LprG-V91W ) [31] , we hypothesized that lipoglycans with longer saccharide chains , e . g . LAM , might have additional functional interactions involving their saccharide moieties and other sites on LAM . To obtain insight regarding structures that determine lipoglycan/glycolipid binding to LprG , which may have implications for LprG function , we used SPR to measure the affinity and kinetics of LprG-substrate interactions and the effects of alterations in glycolipid/lipoglycan structures on binding to LprG . In vitro binding assays were performed using non-acylated LprG ( NA-LprG ) or NA-LprG-V91W , since the non-acylated versions retain glycolipid/lipoglycan binding properties and present technical advantages for expression , purification and use in binding assays ( all references to LprG or LprG-V91W in SPR assays refer to the non-acylated versions ) . Expression of LprG or LprG-V91W in E . coli allowed purification of LprG without the presence of bound substrates ( Fig . 1 ) , facilitating subsequent glycolipid and lipoglycan binding assays . LprG and LprG-V91W were immobilized as ligands on sensor chips , and increasing concentrations of substrates were injected . Sensograms representing association and dissociation phases were obtained after subtracting non-specific binding to a blank sensor chip . The results provide the first determination of affinities for binding of different glycolipids and lipoglycans to LprG , allowing us to assess novel contributions of saccharide moieties to lipoglycan binding . To explore relationships between glycolipid structure and LprG-binding properties , we studied a panel of mycobacterial glycolipids and lipoglycans , including PIM2 , PIM6 , LM , ManLAM and PI-LAM , all of which share a mannosyl-phosphatidyl-myo-inositol domain with additional specific structural features ( Fig . 2A ) . All of these molecules are expressed by Mtb except for PI-LAM , which is expressed by non-pathogenic mycobacteria . PIM2 ( a precursor of PIM6 , LM and LAM ) has mannose residues at positions 2 and 6 of the myo-inositol ring of PI ( Fig . 2A ) . PIM2 bound to LprG in a dose-dependent manner ( Fig . 2B ) but with relatively low affinity ( KD = 1 . 09×10−6 M , Table 1 ) . PIM6 , LM , ManLAM and PI-LAM all bound to LprG ( Fig . 2C–F ) . Of these glycolipids , LM , which consists of a long linear mannan chain extending from position 6 of the myo-inositol ring , bound to LprG with the highest affinity ( KD = 1 . 58×10−9 M , Fig . 2D , Table 1 ) . PIM6 , which consists of a much shorter mannoside motif attached to position 6 of the myo-inositol ring , bound to LprG with the second highest affinity ( KD = 4 . 5×10−8 M , Fig . 2C , Table 1 ) . LAM is generated from LM by the addition of a branched arabinan domain with species-specific terminal caps ( ManLAM for Mtb , PI-LAM for non-pathogenic mycobacteria , Fig . 2A ) . ManLAM and PI-LAM bound to LprG with lower affinities than LM and PIM6 ( KD of 1 . 09×10−7 M for ManLAM and 1 . 6×10−6 M for PI-LAM , Fig . 2E–F , Table 1 ) . Despite sharing the common acylated core structure that was previously implicated in binding to the hydrophobic pocket of LprG , these glycolipids and lipoglycans displayed distinct kinetics for LprG binding , indicating a previously unknown contribution of polysaccharide structures to LprG . To assess the relative contributions of the hydrophobic binding pocket vs . other sites on LprG , we studied glycolipid and lipoglycan binding to LprG-V91W , which has a mutated pocket that precludes binding of the acyl chains of triacylated glycolipids [31] . PIM2 ( Fig . 3A ) and PI-LAM failed to bind to LprG-V91W , but PIM6 , LM and ManLAM bound to LprG-V91W with an affinity approximately 10- to 130-fold lower than their affinity for LprG ( Fig . 3B–D ) ( Table 1 ) . The loss of acyl-chain dependent interaction with LprG-V91W increased substrate dissociation rates ( Koff ) , indicating less stable interactions . These results indicate the presence of novel interactions of a subset of glycolipids/lipoglycans with LprG at a site outside of the hydrophobic pocket . Moreover , this suggests that these interactions include glycolipid and lipoglycan structures other than the acyl chains that bind within the hydrophobic pocket of LprG , consistent with the implication of novel contributions of saccharide structures of substrates in interactions with LprG . To further analyze the relative contributions of acyl and non-acyl structures of substrates to LprG binding , we subjected LM and ManLAM to alkaline deacylation . Deacylated LM and deacylated ManLAM both bound specifically to LprG ( Fig . 4A and B , Table 2 ) , and this binding was similar for LprG and LprG-V91W . This confirms that non-acyl structures of lipoglycans bind to a site on LprG different from the hydrophobic pocket . In support of this conclusion , loss of the acyl chains reduced the affinities of LM and ManLAM for LprG to values similar to the affinities of acylated LM and acylated ManLAM for LprG-V91W . Deacylation had little effect on the association rate for LM or ManLAM binding to LprG or LprG-V91W but resulted in increased dissociation rates with LprG . Overall , these results establish that Mtb lipoglycans and glycolipids bind to LprG via at least two different interactions at different sites on LprG , one being the acyl chain binding in the hydrophobic pocket and the other being interaction of saccharide chains ( likely mannose chains ) with LprG outside of the hydrophobic pocket . The ability of LprG to discriminate different glycolipids/lipoglycans based on differences in their saccharide structures may allow differential binding and release of different glycolipids/lipoglycans . To investigate whether mannan components of LAM and LM contribute to lipoglycan binding to LprG , we tested the ability of S . cerevisiae mannan to bind to LprG . S . cerevisiae mannan was selected since its structure is comparable to the polysaccharide components of Mtb lipoglycans , except for the absence of the phosphatidyl-myo-inositol lipid anchor present in Mtb lipoglycans . SPR assays revealed mannan binding to LprG in a dose dependent manner , although with a lower affinity compared to LAM and LM ( KD = 8 . 8×10−5 M , Fig . 5A and Table 3 , compare to Table 1 ) . Next , to test whether mannan and Mtb lipoglycans have overlapping binding site ( s ) on LprG , we assessed their ability to compete for binding to LprG in SPR assays . The potential approach to saturate immobilized LprG with mannan in a first injection and subsequently inject LAM or LM was not feasible due to the rapid dissociation rate of mannan ( 4 . 5×10−2 sec−1 ) , which allowed dissociation of pre-bound mannan before injection of LAM or LM could be completed . Accordingly , we first injected the lipoglycan ( e . g . LM , 2 . 5 µM ) , which bound to LprG ( Fig . 5B , “first injection” ) ; a subsequent injection of mannan ( 2 . 5 µM; Fig . 5B , “second injection” ) did not reveal mannan binding in contrast to the ability of mannan to bind to LprG in the absence of LM ( compare Fig . 5A and 5B ) . Similar results were obtained with LAM and mannan . These results indicate that mannan and Mtb lipoglycans compete for binding to LprG . To confirm this conclusion with a different approach , we used a solid phase binding assay that measured LprG binding to plate-immobilized ManLAM . LprG bound to plate-immobilized ManLAM in a dose-dependent fashion . When LprG was incubated with mannan , however , the binding of LprG to ManLAM was inhibited by the presence of mannan in a dose-dependent manner ( Fig . 5C ) . These data further indicate that mannan and Mtb lipoglycans compete for binding to LprG . Taken together , these results support our other evidence that mannose residues contribute to the binding of Mtb lipoglycans to LprG , in addition to contributions of acyl chain interactions . We hypothesized that LprG may serve a function in the synthesis of LAM ( and possibly LM and PIM ) or their assembly into the cell envelope ( possibly mediating their transfer from anchorage in the plasma membrane to allow localization at sites more peripheral in the cell envelope ) . This hypothesis suggests that deletion of LprG might alter the expression or localization of these molecules and influence the cell envelope properties of Mtb . Accordingly , we studied the effects of LprG expression on Mtb cell envelope properties using Mtb H37Ra ΔlprG and Mtb H37Rv ΔlprG strains with deletion of the lprG gene , and H37Rv ΔlprG::lprG-Rv1410c , a complemented version of H37Rv ΔlprG expressing the operon encoding LprG and p55 . To assess whether deletion of lprG affects the expression of the glycolipids to which LprG binds , e . g . LAM , we used Western blotting to assess the total bacterial expression of LAM and flow cytometry to detect the expression of LAM at the bacterial cell surface . Whole cell lysates of Mtb H37Ra and H37Ra ΔlprG were analyzed by SDS-PAGE and Western blotting with anti-LprG , which confirmed the absence of LprG in Mtb H37Ra ΔlprG ( Fig . S1A ) . Western blotting with anti-LAM monoclonal antibody CS-35 or anti-Mtb polyclonal antibody ( Fig . S1B , C ) and detection of glycolipids and lipoglycans with carbohydrate staining ( Fig . S1D ) showed that Mtb H37Ra and H37Ra ΔlprG had similar total cellular expression of LAM and LM . Assessment of lipoglycan and glycolipid expression in Mtb H37Rv , H37Rv ΔlprG and H37Rv ΔlprG::lprG-Rv1410c strains by SDS-PAGE analysis , thin layer chromatography and periodic acid/Schiff staining revealed comparable amounts of LAM , LM and PIMs in all three strains . Since deletion of lprG did not affect overall expression of LAM in Mtb , we assessed the hypothesis that LprG may affect the distribution of LAM in Mtb . Specifically , we considered that LprG may affect LAM assembly into the cell envelope , possibly by mediating removal of LAM from anchorage in the plasma membrane to allow its localization to sites more peripheral in the cell envelope . To determine whether the deletion of lprG reduces the amount of LAM detected at the bacterial cell surface , Mtb H37Rv , the knockout strain H37Rv ΔlprG , and the complemented control strain H37Rv ΔlprG::lprG-Rv1410c were stained with rabbit anti-ManLAM anti-serum or control normal rabbit serum and analyzed by flow cytometry . Specific LAM staining was >2-fold higher on Mtb H37Rv relative to Mtb H37Rv ΔlprG , and 1 . 8-fold higher on Mtb H37Rv ΔlprG::lprG-Rv1410c relative to Mtb H37Rv ΔlprG ( Fig . 6 ) . Furthermore , similar results were obtained with Mtb H37Ra ( specific LAM staining was 7-fold higher on Mtb H37Ra relative to Mtb H37Ra ΔlprG ) . In summary , deletion of lprG did not affect total bacterial expression of LAM but did alter LAM distribution to substantially decrease its cell surface exposure . Bacterial phagosomes fuse with lysosomes to produce phagolysosomes , which can mediate killing of some bacteria . Inhibition of phagosome-lysosome fusion is one means by which Mtb evades host defenses , and LAM is known to inhibit phagosome-lysosome fusion ( i . e . inhibit Mtb phagosome maturation ) [13]–[15] . We hypothesized that decreased surface expression of LAM in Mtb ΔlprG would decrease the ability of this Mtb strain to inhibit phagosome maturation . To test this hypothesis , murine bone marrow-derived macrophages were incubated with LysoTracker Red to label lysosomes and then infected with FITC-labeled Mtb H37Rv , H37Rv ΔlprG or H37Rv ΔlprG::lprG-Rv1410c . Mtb phagosome-lysosome fusion was assessed by co-localization of LysoTracker Red with intracellular FITC-Mtb ( within DAPI-stained cells ) as assessed by fluorescence microscopy ( Fig . 7A ) . After 1 h of infection , Mtb H37Rv demonstrated 50% co-localization with LysoTracker Red , whereas Mtb H37Rv ΔlprG showed 82% co-localization ( p = 0 . 0002 ) . Mtb H37Rv ΔlprG::lprG-Rv1410c showed reversion to the wild-type phenotype with 55% lysosomal co-localization ( not significantly different from Mtb H37Rv , p = 0 . 6181 ) ( Fig . 7B ) . As a positive control for uninhibited phagosome-lysosome fusion , heat-killed Mtb of all three strains proceeded to near complete phagosome-lysosome fusion ( 96–97% co-localization of Mtb and lysosomal markers ) , consistent with prior observations that heat-killed Mtb does not inhibit phagosome maturation [16] , [17] . In conclusion , deletion of lprG reduces the availability of LAM for interactions with the host cell machinery that produce inhibition of phagosome-lysosome fusion . These results and the observation that deletion of lprG reduced cell-surface expression of LAM by Mtb both indicate that LprG controls the localization of LAM and suggest that LprG has a role in the insertion or assembly of lipoglycans into the Mtb cell envelope , perhaps by mediating its removal from the plasma membrane to allow localization to more peripheral sites in the cell envelope . The ability of LprG to affect the expression of LAM on the Mtb cell surface has important implications , as this mechanism determines the access of LAM to interaction with host cells .
The studies reported here explore the importance of determinants of cell envelope architecture in host-pathogen interactions and pathogenesis of bacterial infections . In particular , we shed new insight into the role of LprG expression in the pathogenic species , Mtb . Our data reveal novel structure-function relationships that determine glycolipid binding by LprG , including aspects of LAM binding that are specific to the LAM species expressed by Mtb ( ManLAM , as opposed to PI-LAM expressed by non-pathogenic mycobacteria ) . Specifically , we demonstrate that LprG expressed in Mtb binds to ManLAM , a major immunomodulatory molecule of the Mtb cell envelope . Differences in the kinetics and affinities of LprG binding by ManLAM , PI-LAM , LM , PIM2 , PIM6 and deacylated versions of LM and LAM reveal novel contributions of the saccharide moieties of these substrates to LprG binding . ManLAM , PI-LAM , LM , PIM2 and PIM6 all share the core acylated PIM structure , allowing their acyl chains to bind in the LprG hydrophobic pocket , as indicated by reductions in binding of all of these species to pocket-mutated LprG V91W relative to LprG ( Table 1 ) . The V91W mutation reduced the binding affinity by ∼10–100-fold for ManLAM , LM , and PIM6 , and reduced binding to undetectable levels for PIM2 or PI-LAM . However , variation in the saccharide moieties of PIM , LM and LAM also influenced binding to LprG , revealing novel binding interactions of saccharide components at an additional site outside of the hydrophobic pocket ( which persist with the LprG V91W pocket mutant ) . Deacylated ManLAM and deacylated LM were found to bind LprG , albeit with lower affinity than the acylated versions , and their binding was not substantially affected by the V91W pocket mutation . This further establishes a contribution of the substrate saccharide moieties to binding at a site outside of the LprG hydrophobic pocket . PIM2 lacks the longer mannan chain extensions from the PIM core structure that are seen in the other glycolipids , and its lower affinity for LprG binding may reflect lack of mannan interactions with LprG . Variation of the terminal capping of LAM may also affect LprG binding , given the reduced affinity of PI-LAM ( which lacks terminal mannose residues ) relative to ManLAM . In summary , these results indicate that binding interactions between Mtb lipoglycans and LprG extend beyond the lipoglycan acyl chains and implicate a role for saccharide chains of Mtb lipoglycans in binding to LprG . First , deacylated Mtb lipoglycans bind to LprG ( albeit at lower affinity than acylated lipoglycans ) . Second , acylated lipoglycans bind to LprG even when the hydrophobic pocket where the acyl chains bind is mutated to prevent their interaction ( albeit with lower affinity than to wild-type LprG ) . Notably , the affinity of binding of the deacylated lipoglycans was not affected by mutation of the hydrophobic pocket ( Table 2 ) . Interestingly , in these two situations ( acylated lipoglycan binding to pocket-mutated LprG vs . deacylated lipoglycan binding to LprG ) , whereby different approaches the contribution of acyl chains to LprG binding has been removed and the remaining interactions are implicated to be with saccharide chains , the binding affinities are very similar , consistent with our model ( Tables 1 and 2 , compare KD for binding of acylated vs . deacylated LAM to LprG vs . LprG-V91W , or compare KD for binding of acylated vs . deacylated LM to LprG vs . LprG-V91W ) . The discovery of contributions of lipoglycan polysaccharide moieties to LprG binding suggested that these interactions involve lipoglycan mannan residues , since this property is shared by lipoglycans with mannose saccharides ( e . g . LAM ) and that lack other saccharide components ( e . g . arabinan ) found in LAM . Furthermore , we directly tested the contribution of mannan structures to LprG binding by use of S . cerevisiae mannan , which has structural similarity to the mannan chains of Mtb LM and LAM and also binds to LprG , although with a lower affinity . In competition binding assays , LM blocked mannan binding to LprG ( Fig . 5B ) , and mannan inhibited LAM binding to LprG ( Fig . 5C ) . These two findings further support the hypothesis that mannan and the Mtb lipoglycans ( LAM , LM ) compete for a binding site on LprG and further support the hypothesis that mannan chains of these Mtb lipoglycans contribute to their binding to LprG . In summary , although saccharide interactions are not essential for low affinity glycolipid binding to LprG , mannan chains interact with LprG to enhance substrate-LprG binding . We propose a model for lipoglycan binding to LprG that involves both acyl chain binding in the LprG hydrophobic pocket and mannan chain interactions with LprG outside of the pocket . In addition to determining the structural basis for LprG-glycolipid interactions , we assessed the functional contributions of LprG using genetic deletion models . Although lprG deletion did not alter the total cellular expression of LAM , LM and PIM by Mtb , our results reveal the striking discovery that LprG controls expression of LAM at the cell surface of Mtb , presumably via its lipoglycan binding function . Thus , LprG influences the spatial organization of LAM within the Mtb cell envelope , perhaps via a role in LAM insertion or assembly into the cell envelope . One possibility is that LprG mediates the removal of lipoglycans from acyl-chain anchorage in the plasma membrane to allow their localization to more peripheral sites in the cell envelope . Altered localization of LAM in the cell envelope may have significant implications for cell envelope functions and host-pathogen interactions . Deletion of LprG affects general measures of cell envelope integrity ( e . g . malachite green decolorization , Congo red binding ) , consistent with other studies that associated altered cell envelope permeability with deletion of the lprG-Rv1410c operon [28] , [30] , [32] . Since LAM is implicated in inhibiting the fusion of Mtb phagosomes with lysosomes [13]–[15] , we tested the hypothesis that reduced expression of surface-exposed LAM in the Mtb cell envelope would decrease LAM interactions with host cells and limit inhibition of phagosome maturation in Mtb H37Rv ΔlprG . Consistent with our hypothesis , we observed decreased inhibition of Mtb phagosome-lysosome fusion with Mtb H37Rv ΔlprG relative to wild-type Mtb H37Rv . Thus , deletion of lprG reduced the ability of Mtb to inhibit phagosome maturation , consistent with the known role for LAM in inhibiting phagosome maturation and our discovery that deletion of lprG diminishes expression of LAM at the bacterial cell surface , which we surmise diminishes its availability to interact with host cells . In summary , our results reveal novel structural determinants of the binding of PIMs , LM and LAM to LprG and a novel role for LprG in determining LAM distribution in the cell envelope . While LprG is not necessary for LAM biosynthesis , it controls the expression of LAM at the bacterial cell surface , perhaps via a role in transferring LAM from the plasma membrane for localization in the cell envelope . Thus , LprG determines the accessibility of LAM for interactions with host cells to regulate phagosome-lysosome fusion and other host-pathogen interactions . These findings have important implications for fundamental determinants of host-pathogen interactions during infection with Mtb . LprG clearly influences Mtb phagosome maturation; it is required for optimum inhibition of phagosome maturation by Mtb , and it may thereby contribute to immune evasion by this mechanism . The influence of LprG on LAM localization in the Mtb cell envelope may affect the integrity of the cell envelope and its effectiveness as a permeability barrier , which is critical to Mtb survival within the host . Findings in this study provide novel insight into mechanisms of pathogenesis and host resistance to Mtb infection , and they suggest the possibility that pharmacologic disruption of LprG expression or function might result in a leaky cell envelope ( making Mtb less resistant to drugs and antibiotics ) and reduced ability of Mtb to evade host defense mechanisms by inhibiting phagosome-lysosome fusion . In conclusion , LprG is a critical determinant of Mtb virulence and host-pathogen interactions during Mtb infection and may be used as a potentially important target for therapeutic intervention .
The Institutional Animal Care and Use Committee of Case Western Reserve University approved all animal studies ( protocol 2012–0007 ) . Studies were performed in accordance with recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . Mtb strains H37Rv and H37Ra were obtained from American Type Culture Collection ( Manassas , VA ) . LprG null strains were generated from Mtb H37Rv ( H37Rv ΔlprG; N . Banaei ) and H37Ra ( H37Ra ΔlprG; E . T . Richardson ) using a specialized transducing phage targeting 571 bp within the lprG gene locus for homologous recombination with a hygromycin resistance cassette . Mtb H37Rv ΔlprG::lprG-Rv1410c was then generated by complementing Mtb H37Rv ΔlprG with the native Rv1411c/1410c operon expressed off the kanamycin-selective integrating plasmid pMV306 . Bacteria were cultured in Middlebrook 7H9 broth ( Difco , Lawrence , KS ) supplemented with 10% albumin/dextrose/catalase ( BD , Franklin Lakes , NJ ) , 0 . 05% Tween 80 and 0 . 2% glycerol plus 100 µg/ml hygromycin B ( Invivogen , San Diego , CA ) for H37Ra ΔlprG and H37Rv ΔlprG or 50 µg/ml kanamycin ( Sigma , St . Louis , MO ) for H37Rv ΔlprG::lprG-Rv1410c . To assess bacterial expression of lipoglycans , lysates were prepared from Mtb H37Ra and H37Ra ΔlprG . Bacteria were grown to mid log phase in 50 ml of Sauton's medium , pelleted , suspended in PBS , and sonicated in an ice bath for 40 min in 10-second cycles at 50 W in a Misonix sonicator ( Misonix Inc . , Farmingdale , NY ) . Insoluble debris was pelleted at 2000× g for 5 min , and supernatants were stored at −80°C . Samples were analyzed by SDS-PAGE as described below . Constructs for expression of recombinant lipoproteins LprG , LprA , non-acylated LprG ( NA-LprG ) and non-acylated LprG-V91W ( NA-LprG-V91W ) were cloned previously [6] , [31] . LprG , LprA , NA-LprG and NA-LprG-V91W were expressed in wild-type Mtb H37Ra , M . smegmatis or E . coli and purified as described [6] , [31] . For expression in Mtb H37Ra , constructs were digested with NdeI and HindIII ( New England Biolabs , Ipswich , MA ) and ligated into the shuttle vector pVV16 ( provided by J . Belisle , Colorado State University , Fort Collins , CO ) under control of the constitutively active hsp60 promoter and in-frame with a C-terminal hexahistidine tag . Mtb was transformed by electroporation with a Gene Pulser ( Bio-Rad , Hercules , CA ) set at 2 . 5 kV , 25 µF , and 800 ohms . Mtb was grown with kanamycin selection ( 50 µg/ml ) to late log phase , isolated by centrifugation at 6000 g for 20 min at 4°C , and suspended for 15 min at 37°C in lysis buffer . Bacteria were disrupted by four passages through a French press ( 2000 psi ) . Insoluble material was removed by ultracentrifugation at 100 , 000 g for 1 h at 4°C . The supernatant was incubated for 2–4 h at 4°C with Ni beads ( Qiagen , Valencia , CA ) , which were then washed 3 times with 25 volumes of wash buffer ( 50 mM sodium phosphate , 1 M NaCl , 20 mM imidazole , 10% glycerol , pH 8 . 0 ) . Bound protein was dissociated with elution buffer consisting of 50 mM sodium phosphate , 300 mM NaCl , 450 mM imidazole , pH 8 . 0 . Samples were exchanged into 20 mM Tris , pH 8 . 0 using PD-10 columns ( GE Healthcare , Uppsala , Sweden ) and further purified by anion-exchange chromatography . Recombinant LprG was bound to HiTrap Q FF columns ( GE Healthcare ) and eluted by stepwise addition of 50 , 150 , 200 and 1000 mM NaCl . Recombinant LprG eluted with 50–200 mM NaCl was collected and concentrated using 10-kDa cutoff Centricon units ( Amicon , Billerica , MA ) . Protein yields were determined by BCA protein assay ( Pierce , Rockford , IL ) . Protein preparations or bacterial lysates were electrophoresed on 12% or 4–20% Mini-PROTEAN TBX Precast SDS-PAGE gels ( BioRad , Hercules , CA ) . Protein purity was analyzed with Silver Stain Plus ( BioRad ) . For Western blot analysis , materials were transferred to polyvinylidene difluoride membranes ( Millipore ) , which were blocked with 5% milk in PBS supplemented with 0 . 1% Tween-20 ( PBST ) for 1 h at room temperature . Membranes were incubated overnight at 4°C with mouse monoclonal anti-hexahistidine antibody ( Santa Cruz Biotechnology , Santa Cruz , CA ) , rabbit polyclonal anti-Mtb antibody ( GENway , Hayward , CA ) , mouse monoclonal anti-LAM antibody ( CS-35 ) and mouse monoclonal anti-LprG antibody ( clone α–Rv1411c , NR-13806 , NIH Biodefense and Emerging Infections Research Resources Repository; BEI , Manassas , VA ) . Blots were washed three times with PBST , incubated for 1–2 h at room temperature with secondary antibodies conjugated to horseradish peroxidase , washed three times with PBST and visualized using a chemiluminescence kit ( Pierce ) . Alternatively , membranes were stained to detect carbohydrate with the Glycoprotein Staining Kit ( Pierce , Rockford , IL ) . Purified ManLAM and LM from Mtb were deacylated by treatment of 1 mg ManLAM or LM with 0 . 1 N NaOH for 2 h at 37°C . The reaction mixture was neutralized with 10% acetic acid to pH ∼8 . 0 , and dialyzed against distilled water for one day at room temperature . The sample was dried , loaded onto a Bio-gel P100 column and eluted with 0 . 1 N acetic acid . Fractions were collected and analyzed by SDS-PAGE to confirm loss of migration due to deacylation . Fractions containing deacylated ManLAM or LM were pooled , dialyzed against distilled water and dried . Sample purity was assessed by gas chromatography-mass spectrometry . SPR binding experiments were performed on a BIAcore 3000 instrument using CM5 ( carboxymethylated ) sensor chips and HBSN running buffer ( 10 mM HEPES and 150 mM NaCl , pH 7 . 4 ) ( GE Healthcare , Uppsala , Sweden ) . Non-acylated ( NA ) -LprG was used for SPR studies and purified from E . coli to prevent prior occupancy of LprG binding site ( s ) by Mtb glycolipids and lipoglycans that might hinder the binding of exogenous purified LprG ligands used in SPR analyses . Our previous studies [31] reveal that NA-LprG retains its ability to bind to Mtb glycolipids and lipoglycans , and a similar spectrum of LprG-associated molecules was found associated with acylated LprG expressed in Mtb ( Fig . 1 ) . In addition , it is unlikely that the acyl chains of LprG would interact with residues at or near to the lipoglycan binding pocket , as the N-terminal site of acylation of LprG is ∼50 Å away from the lipoglycan pocket ( discussed by Drage et al [31] ) . Acylated LprG ( Fig . 1 ) and NA-LprG ( Figs . 2–5 ) both bind LAM , LM and PIMs . NA-LprG and NA-LprG-V91W ligands were immobilized on CM5 sensor chips by amine coupling [33] . A 1∶1 mixture of 0 . 1 M n-hydroxysuccinimide ( NHS ) and 0 . 1 M ethylene diamine ( GE Healthcare ) was injected to activate carboxyl groups on the CM5 matrix , and 30 ml of a 40 mg/ml solution of ligand in 10 mM acetate buffer pH 4 . 5 was then injected . Residual NHS-esters were deactivated with 1 M ethanolamine ( GE Healthcare ) . Control flow cells were activated and deactivated in the same manner but without protein ligand . Analytes were ManLAM , LM and PIMs purified from Mtb H37Rv ( BEI ) , PI-LAM purified from M . smegmatis ( BEI ) , and mannan from Saccharomyces cerevisiae ( Sigma ) . The acylation state of PIMs was assessed by thin layer chromatography and mass spectrometry , revealing that the glycolipids were tri-acylated and tetraacylated ( other acylation states were not significantly detected ) . Analytes were diluted in HBSN running buffer and injected at increasing concentrations ( 0 . 312 µM to 2 . 5 µM ) at a flow rate of 30 µl/min . At the end of the injection , complexes were allowed to dissociate for 10 min . Chip surfaces were regenerated by 2–3 injections of 20 ml 50–90% ethylene glycol . Injections were separated by an equilibration delay of 30 min with HBSN at a flow rate of 5 ml/min . The final amount of bound analyte , expressed in resonance units ( RU ) , was calculated by subtracting the RU of the control flow cell from the RU of the ligand-conjugated cell . Sensograms were analyzed using BIA evaluation 3 . 1 software ( GE Healthcare ) . To assess binding competition by SPR , a first injection of LM ( 2 . 5 µM for 3 min ) was followed by a buffer wash ( ∼10 min , reflecting the time necessary for the machine to switch to another injection ) and then a second injection of mannan from Saccharomyces cerevisiae ( 2 . 5 µM for 3 min ) . This injection order was chosen due to the fast dissociation rate for mannan . Overlaid sensograms of the serial injections ( with the signal at the start of each injection normalized ) reveal the effect of prior LM-LprG binding on mannan-LprG interactions . To assess the ability of mannan to compete with LAM for binding to LprG , solid phase binding assays were performed as described [34] in 96-well plates ( Corning Incorporation , NY , USA ) coated with ManLAM . ManLAM was added ( 100 ng in 50 µl of 50% ethanol per well ) , and plates were air-dried for 1 h at 37°C . Subsequent steps were at room temperature . Plates were washed in wash buffer ( 10 mM Tris-HCl , pH 7 . 4 , 140 mM NaCl , 1 mM CaCl2 , 0 . 005% Tween 20 ) to remove unbound ManLAM , blocked with 5% milk in wash buffer for 2 h , washed twice in wash buffer , and incubated overnight with 100 µl of NA-LprG ( 1 µM ) that had been pre-incubated for 30 min with or without mannan ( 10–500 µM ) in wash buffer minus Tween . The plates were washed extensively , incubated with monoclonal anti-LprG antibody ( clone α-Rv1411c , NR13806 , BEI ) for 2 h , washed three times , incubated for 1 h with goat anti-mouse IgG conjugated to horseradish peroxidase ( Cell Signaling Technology , Danvers , MA ) , washed three times and incubated with o-Phenylenediamine dihydrochloride ( Sigma ) in 0 . 05 M phosphate-citrate buffer , pH 5 . 0 ( 50 µl/well ) in the dark for 30 min . To stop the color development reaction , 50 µl of 2N H2SO4 was added . OD450 was determined on a Bio-Rad plate reader . Mtb cultures ( Mtb strains H37Rv , H37Rv ΔlprG , H37Rv ΔlprG::lprG-Rv1410c , H37Ra and H37Ra ΔlprG ) were seeded to OD600 of 0 . 05 and harvested after one week . Bacteria were declumped by 10 passages through a 23-gauge needle , incubated in blocking buffer ( 1% heat inactivated normal rabbit serum , 0 . 01% Tween 80 , PBS ) at room temperature for 1 h , and washed twice with FACS buffer ( 1% BSA , 0 . 01% Tween 80 , PBS ) . Bacteria were incubated for 1 h with 500 µl of FACS buffer containing rabbit anti-ManLAM antiserum ( NR-13821 , BEI , Manassas , VA ) or control normal rabbit serum ( Life Technologies , Grand Island , NY ) , washed twice with FACS buffer , stained for 1 h with Alexa Fluor 647-anti-rabbit IgG ( Life Technologies , Grand Island , NY ) , washed twice with FACS buffer , fixed with 2% paraformaldehyde for 1 h at room temperature , and suspended in PBS containing 0 . 01% Tween 80 . Anti-LAM labeling of bacteria was assessed with an LSR II flow cytometer ( BD Bioscience , San Jose , CA ) . Data were analyzed using FlowJo software version X ( TreeStar , Ashland , OR ) . Female 8–12 week old C57BL/6J mice were obtained from the Jackson Laboratory ( Bar Harbor , ME ) and housed under specific pathogen-free conditions . Bone marrow-derived macrophages were cultured as described previously [6] . Briefly , bone marrow was flushed from isolated bones in DMEM , and cell suspensions were homogenized and filtered through a 70 µm screen . Bone marrow cells were pelleted and subjected to red blood cell lysis in ACK lysing buffer ( Lonza , Walkersville , MD ) , pelleted again , and cultured in D10F consisting of DMEM ( HyClone , Logan , UT ) supplemented with 10% heat-inactivated fetal bovine serum ( Gibco , Carlsbad , CA ) , 50 µM 2-mercaptoethanol ( Bio-Rad , Hercules , CA ) , 1 mM sodium pyruvate ( HyClone ) , 10 mM HEPES ( HyClone ) , 100 units/ml penicillin , 100 µg/ml streptomycin ( HyClone ) and 25% LADMAC cell conditioned medium as a source of M-CSF . Cultures were incubated at 37°C in a humidified , 5% CO2 atmosphere . Medium was changed once on day 5 of culture , and differentiated bone marrow-derived macrophages were used on day 7 . Macrophages were harvested using 0 . 25% trypsin/0 . 02% EDTA ( HyClone ) and 250 , 000 cells per well were plated in D10F without penicillin/streptomycin onto autoclaved glass cover slips ( Fisher Scientific , Pittsburgh , PA ) for in vitro infections and microscopy . Mtb strains ( H37Rv , H37Rv ΔlprG , H37Rv ΔlprG::lprG-Rv1410c ) were grown to mid-log phase ( OD600 = 0 . 5 to 0 . 6 ) , pelleted , suspended in PBS with 0 . 05% Tween 80 , and declumped by vigorous vortexing . Remaining clumps were removed by centrifugation for 15 min at 100 g . Some aliquots of Mtb were heat-killed by incubation at 80°C for 20 min . For fluorescent labeling of bacteria , FITC ( Sigma ) was added from a stock solution in DMSO to a concentration of 0 . 2 mg/ml FITC in 1% DMSO for 30 min at 37°C with gentle rotation . Bacteria were washed three times in PBS with 0 . 05% Tween 80 and suspended in D10F without antibiotics . Bone marrow-derived macrophages ( grown on coverslips ) were incubated for 30 min with 100 nM LysoTracker Red ( Molecular Probes , Carlsbad , CA ) in D10F , and FITC-labeled bacteria were added to a multiplicity of infection ( MOI ) of 10 for 30 min at 37°C . Macrophages were washed and incubated in D10F containing 100 nM LysoTracker Red for 30 min , washed with PBS , and fixed for 1 h in 2% paraformaldehyde . Coverslips were washed three times in PBS and inverted onto ProLong Gold mounting media with DAPI ( Molecular Probes ) . Slides were cured overnight and imaged using a Deltavision RT epifluorescence microscope . Image files were deconvolved using the Softworx software package ( Applied Precision , Issaquah , WA ) . Co-localization was determined by identifying intracellular bacteria ( FITC-labeled organisms within DAPI-labeled cells ) with overlapping lysosomal marker ( LysoTracker Red ) . Colocalization data were obtained by counting at least six independent fields from each sample , and the results are representative of two independent experiments . Control samples consisting of uninfected/LysoTracker Red-treated or infected/LysoTracker Red-untreated bone marrow-derived macrophages were analyzed to set the lower threshold for FITC and LysoTracker Red positivity , and all infected and LysoTracker-labeled macrophages were analyzed with identical intensity cutoffs . Statistical analyses were performed using GraphPad Prism 5 . 01 software ( La Jolla , CA ) . Fisher's exact test or Students two-tailed t-test was used to analyze the statistical significance of differences between groups .
|
The causative agent of tuberculosis , Mycobacterium tuberculosis ( Mtb ) , persists in phagosomes inside infected macrophages . Mtb expresses lipoarabinomannan ( LAM ) , which inhibits fusion of phagosomes with lysosomes as a means for Mtb to evade host defense . LAM is present in the cell envelope , which surrounds Mtb and interfaces with the host , but its localization remains unclear . We show that LprG , an Mtb lipoprotein , binds LAM and controls its distribution in the cell envelope . A mutant strain of Mtb that lacks LprG has less LAM at the surface of the cell envelope . This decreases LAM-mediated inhibition of phagosome-lysosome fusion , thereby impairing an immune evasion mechanism . We propose that LprG facilitates transfer of LAM from the plasma membrane into the cell envelope , enhancing its interaction with the host and ability to regulate host defense . Our results reveal mechanisms that determine bacterial cell envelope function and influence host-pathogen interactions and pathogen evasion of host defense .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biomacromolecule-ligand",
"interactions",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"immunology",
"microbiology",
"bacterial",
"diseases",
"immunomodulation",
"infectious",
"diseases",
"proteins",
"medical",
"microbiology",
"tuberculosis",
"microbial",
"pathogens",
"pathogenesis",
"immune",
"response",
"biochemistry",
"immunity",
"host-pathogen",
"interactions",
"biology",
"and",
"life",
"sciences",
"glycobiology"
] |
2014
|
Mycobacterium tuberculosis Lipoprotein LprG Binds Lipoarabinomannan and Determines Its Cell Envelope Localization to Control Phagolysosomal Fusion
|
In many bacteria the ParA-ParB protein system is responsible for actively segregating DNA during replication . ParB proteins move by interacting with DNA bound ParA-ATP , stimulating their unbinding by catalyzing hydrolysis , that leads to rectified motion due to the creation of a wake of depleted ParA . Recent in vitro experiments have shown that a ParB covered magnetic bead can move with constant speed over a DNA covered substrate that is bound by ParA . It has been suggested that the formation of a gradient in ParA leads to diffusion-ratchet like motion of the ParB bead but how it forms and generates a force is still a matter of exploration . Here we develop a deterministic model for the in vitro ParA-ParB system and show that a ParA gradient can spontaneously form due to any amount of initial spatial noise in bound ParA . The speed of the bead is independent of this noise but depends on the ratio of the range of ParA-ParB force on the bead to that of removal of surface bound ParA by ParB . We find that at a particular ratio the speed attains a maximal value . We also consider ParA rebinding ( including cooperativity ) and ParA surface diffusion independently as mechanisms for ParA recovery on the surface . Depending on whether the DNA covered surface is undersaturated or saturated with ParA , we find that the bead can accelerate persistently or potentially stall . Our model highlights key requirements of the ParA-ParB driving force that are necessary for directed motion in the in vitro system that may provide insight into the in vivo dynamics of the ParA-ParB system .
A variety of mechanisms exist within bacteria to spatially localize proteins within the small confines of a bacterial cell , from reaction diffusion processes that set up waves [1] to spatial occlusion due to the highly crowded environment [2–5] . One such protein system that has been observed to display highly dynamic spatial localization within bacteria is the ParA-ParB system . These two proteins are responsible for actively transporting DNA , whether it be the replicating chromosome [6 , 7] or much smaller plasmids [8] , within the cell . With respect to the chromosome , it has been observed that ParB bound to sites near the origin of replication processively moves from one pole to the other via a gradient in ParA that is bound on the nucleoid surface [9] . When the system is used to segregate plasmids , ParB bound plasmids are seen to move along the ParA bound nucleoid , eventually settling into positions that are equally spaced along the cell length [10] . The exact mechanism by which the ParA-ParB system generates directed transport is not entirely resolved . Biochemical experiments have shown that ParA dimerizes in the presence of ATP and is able to to bind to DNA as ParA-ATP [11] . ParB is able to bind a specific DNA sequence known as parS and aids the autohydrolysis of ParA-ATP , causing it to unbind from the DNA [12 , 13] . Recent studies have also shown that the structure of the bacterial nucleoid plays a role as well . Specifically , it has been shown that conformational changes in the nucleoid can disrupt plasmid positioning [14] and a model suggests that chromosomal elasticity could provide the required translocation force that transports partition complexes [15] . Based on the experimental evidence , several models have emerged to explain the operation of the ParA-ParB-parS system and rule out mechanisms that generate predictions which are inconsistent with observations . One model assumes that ParA-ATP binds into longitudinal filaments over the nucleoid , and that ParB initiates depolymerisation and is then carried along by a depolymerisation force [16 , 17] . The filamentous structure provides directionality to the segregation of DNA . Another model considers that ParA dimers bind uniformly over the nucleoid surface and that ParB linked DNA moves via a diffusion-ratchet mechanism as it creates a wake of ParA [18] . A chemotactic force is hypothesized to exist between the two proteins that biases the otherwise free diffusion of ParB bound DNA [19] . Recent work argues that the precise nature of how ParA binds to the nucleoid surface is inconsequential for the equi-positioning of plasmids in-vivo [14] . It showed that the plasmid’s ability to move along a gradient in ParA concentration was sufficient to explain their resulting positioning . Additionally a diffusion/immobilization mechanism where freely diffusing ParB complexes are immobilized through interactions with ParA was also ruled out . Another model has ruled out freely diffusing ParB complex biased by ParA concentrations by computing that such a mechanism was insufficient to provide directionality to the motion of the partition complex [15] . Recent experimental work has managed to amazingly reconstitute the ParA-ParB system in vitro which provides a new context in which to explore the mechanism of how this two protein system produces active transport . Specifically , it was found that a ParB coated magnetic bead can undergo directed motion when confined to move along the plane ( without rolling ) of a DNA coated flow cell that was bound with ParA-ATP protein [20 , 21] . Multiple beads which were identically prepared were observed to show either directed motion with differing speeds or underwent free diffusion . The beads with directed motion had diffusion constants that were around 3 times lower than that of the freely diffusing beads and had paths with persistence lengths which were many times ( ∼20 ) the radius of the micron sized bead . ParA in the vicinity of the bead was hydrolyzed and released from the surface as the bead moved and was later recovered when the bead had moved away . After an initial lag time , the bead would begin to move across the surface , creating a wake in ParA . Thus from the experimental observations , persistent motion seems to result from the creation of a ParA gradient that can provide a chemotactic force which can drive the ParB covered bead [19] . However , how does this gradient form ? And how does the motion of the bead depend on the various system parameters ? Here we develop a completely deterministic model for the operation of the in vitro ParA-ParB system that complements prior modelling work on the in vivo system . We consider the ParB decorated bead to be an over-damped particle under the influence of attractive forces from ParA proteins on the surface ( see Fig 1 ) . As shown experimentally , beads experiencing directed motion had little diffusion [21] , and so we consider a bead’s motion to be completely deterministic and proportional to this chemical force . ParA kinetics on the surface is also completely deterministic and is driven by the presence of the bead which removes ParA within its vicinity . The only noise we consider is in the initial spatial distribution of ParA and we find that this is sufficient to generate a spontaneous ParA gradient which can drive the motion of the bead . Under most conditions , we find that the bead moves with constant speed , and that this depends on the ratio of the range of the force to that of ParA removal by ParB . Interestingly , we find that the bead can attain a maximum speed which depends on this ratio . We also consider ParA rebinding to the surface since in the in vitro experiment , depleted ParA regions would recover once the bead had moved away [21] . ParA surface diffusion may also serve as a possible mechanism for recovery of ParA . There are two regimes for ParA recovery on the surface: undersaturated , where there are excess binding sites on the DNA for ParA or saturated , where there is more free ParA than there are binding sites on the surface . We find that in the undersaturated regime persistent acceleration of the bead is possible . In the saturated regime , depending on the rate of rebinding and the degree of ParA cooperativity , the bead can be made to stall . From our modeling we find that distinguishing cooperative binding or ParA surface diffusion from that of non-cooperative rebinding using the current in vitro assay would be challenging as their effects on bead motion are all qualitatively similar . Nevertheless , the model does make predictions that could be readily testable using the in vitro system and suggests ways to tune the operation of the ParA-ParB system .
In Fig 1 we show a schematic of our minimal model for the in vitro ParA-ParB system . In the experiment by Vecchiarelli et al . [21] , a ParB decorated bead was put into contact with a surface that was covered with strands of DNA on which ParA-ATP was bound . The observed motion of the bead was predominantly unidirectional , and so we begin by considering a 1D model for the system . We represent the surface bound ParA-ATP with a concentration , a ( x , τ ) . It is initialized with a mean concentration that fluctuates uniformly from position to position with a magnitude δa . The bead of radius R is located at a position xp , and we assume there is a central force that acts on it due to the interactions between ParA-ATP and ParB . ParB on the bead also stimulates the removal of ParA-ATP in the vicinity of the bead . We assume that bead motion is in the overdamped regime so that the drag force balances the net force due to ParA-ParB interaction . Both the chemotactic force and the rate of removal decay with distance from the center of the bead . In the absence of surface diffusion or ParA rebinding the system dynamics for our minimal model are given by the following deterministic equations ( for further details , see Methods ) : ∂ a ( x , τ ) ∂ τ = - e - ( x - x p ) 2 / 2 c 2 a ( x , τ ) , ( 1 ) v = d x p d τ = A 0 ∫ d x e - ( x - x p ) 2 / 2 x - x p 1 + ( x - x p ) 2 a ( x , τ ) . ( 2 ) Here the parameter A0 combines several system parameters: the initial mean ParA concentration , the amount of ParB on a bead , the ParA-ParB interaction force , and the drag on the bead ( see Methods ) . The parameter c is the ratio of the lengthscale over which the bead removes nearby ParA-ATP to the lengthscale over which it experiences a force due to the surface bound ParA . We assume that both the rate of removal and force decay as Gaussian functions and c is the ratio of their standard deviations σr and σf respectively . Physically , c should be less than 1 , since ParA can not be removed over distances greater than that from which it exerts an attractive force on the bead . An estimate for these two lengthscales can be inferred from several experimental observations . First , it was found from in vivo measurements that active motion of plasmids require a ratio of ParB to ParA of around 5 to 1 ( there are about 580 ParB bound proteins to 120 ParA in a cell [15] ) . We assume that the in vitro system requires a similar ratio to generate motion of the ParB decorated beads . The experiment conducted by Vecchiarelli et . al [21] estimates the number of ParB molecules on a bead that can interact with ParA on the surface to be 4800 . They also found the number of ParA molecules present per square micron on the surface near the bead to be 400 . Given the required ratio of ParA to ParB molecules , a bead with 4800 ParB molecules would need to interact with ∼ 1000 ParA molecules to generate motion . Given the ParA molecule surface density , 1000 ParA molecules would cover 1000/400 ∼ 2 . 5 μm2 of the surface leading to an effective force range , σf ∼ 0 . 9 μm . From the experiment , it was also observed that the radius over which the bead removed ParA from the surface ( σr in our model ) was on average 225 nm . Thus given these experimental observations , we estimate that c = σr/σf ∼ 0 . 2−0 . 4 given the lower and upper bounds on the measured values . As we will show below , this value for c is sufficient for generating directed motion of a bead . The only source of noise in our system is in the initial conditions that describe the ParA concentration at every point . Starting with the bead at rest , integrating the above deterministic equations show that after a short time lag the bead begins to move ( Fig 2A ) . This lag period has also been observed in the in vitro experiments [21] . The cause of this movement is the non-zero net force that builds along a particular direction due to the noise in the initial bound ParA distribution which breaks the symmetry around the bead . As the bead travels it removes ParA-ATP leaving a wake behind itself and creates a wavefront of bound ParA in front ( Fig 2B ) . Since the bead is over-damped it has no inertia from its previous step and movement along the chosen direction is sustained because the backward pulling force due to the ParA behind the bead is less than that due to the ParA in front . In two dimensions , a similar lag period is observed during which the bead reduces the ParA concentration ( Fig 3B ) . After symmetry breaking the bead maintains motion along a particular direction as the attractive forces along the sides of the bead roughly cancel out whereas the forward motion is sustained due to the depleted concentration behind the moving bead . For the model given by Eqs 1 and 2 , following the symmetry breaking the bead attains a uniform speed as can be seen by the linear displacement of the bead with time ( Fig 2A ) . The average speed is independent of the magnitude of the noise in the initial ParA distribution ( Fig 2C ) , though we find that the standard deviation of the speed does increase with the noise level . We also find by increasing A0 ( for example by increasing the initial mean ParA concentration ) , the bead’s speed increased , keeping c fixed . The constant of proportionality between speed and A0 is found to be dependent on c . Given that the bead attains a uniform speed we analytically solved Eqs 1 and 2 in this limit ( see S1 Text ) . The steady state ParA distribution for a bead moving with constant speed was found that was then used to solve for the steady state net force acting on the bead ( see S1 Fig ) . This force balances the drag force on the bead moving at constant speed , v , that leads to a non-linear equation that can be solved for the speed in terms of the two free parameters , A0 and c . In Fig 3A , we show that the analytical results for the predicted speed , v , as a function of c , match well with the simulated results found from integrating Eqs 1 and 2 . We observe that the speed of the bead is maximized at a particular value of c for a fixed value of A0 and this value of c at which the speed is maximized changes with changing A0 . The presence of a maximum speed was not unexpected , since in the limit c → 0 , no ParA is removed and hence there is no motion and when c → 1 , too much ParA is removed and the gradient is weakened , lessening the speed . The two dimensional simulation shows a typical trajectory on a 2d substrate that has a noisy initial distribution of ParA ( see Fig 2B ) . Now the bead can spontaneously move in any direction and moves roughly in a directed fashion over the surface . Similar dependences on A0 and c were also found . The analytical solution was extended to 2d by considering the entire speed of the bead to be along an axis ( see S1 Text for details ) . It predicted similar features of maximum speed at a particular value of c , which was confirmed by simulation . Interestingly , for values of c less than ∼ 0 . 3 the bead is faster when placed in a 1d system . But for values of c greater than ∼ 0 . 3 the bead would attain higher speed on a 2d surface ( Fig 3C ) . This can be intuitively understood as follows: when c < 0 . 3 on a 2d surface the wake of reduced ParA behind the bead is thinner than the bead’s effective radius that is attracted by the remaining ParA . This implies that there is a backward pulling force component due to unremoved ParA behind the bead . This force component decreases as c > 0 . 3 and the wake width nears the effective bead radius . This leads to lower speeds for c < 0 . 3 and higher speeds for c > 0 . 3 in 2d . Next we investigated the inclusion of ParA rebinding to a one-dimensional substrate . We assumed that at each position x along the surface there was a certain concentration of binding sites for ParA on the DNA substrate , d ( x ) , whose average is D0 , and fluctuates an amount δd ( see Fig 4 ) . We now also consider that there is ParA in the buffer , given by the quantity ab ( τ ) . Diffusion is quick in the buffer and so ParA in the buffer does not depend on position . The amount of ParA in the system is limited by the initial amount in the buffer which is set to ab ( 0 ) = As . For non-cooperative binding of ParA to the substrate , rebinding depends on the amount of unbound sites available at a given location , namely ( d ( x ) − a ( x , τ ) ) and the rebinding rate , kr . We included cooperative rebinding with a term that depends on the amount of ParA on the surface , and is governed by a cooperative rate , kc . Including this favors rebinding to sites that possess larger amounts of bound ParA . Including rebinding changes Eq 1 to the following equation , ∂ a ( x , τ ) ∂ τ = - e - ( x - x p ) 2 / 2 c 2 a ( x , τ ) + a b ( τ ) [ d ( x ) - a ( x , τ ) ] [ k r + k c a ( x , τ ) ] . ( 3 ) While the equation for the ParA per binding site available in the buffer , ab ( τ ) is given by: d a b ( τ ) d τ = ∫ d x e - ( x - x p ) 2 / 2 c 2 a ( x , τ ) - a b ( τ ) [ d ( x ) - a ( x , τ ) ] [ k r + k c a ( x , τ ) ] / L ( 4 ) Since the ParA released due to the bead diffuses rapidly in the buffer it adds to the free concentration ab uniformly . As the ParA released by the bead is distributed over the entire system , the system size ( L ) and rate of rebinding ( kr ) are coupled in our model . Increasing the length of the system would effectively decrease the rate of rebinding as the ParA released would be diluted over a larger area . The finite size effects of our model also become evident when the bead nears a boundary of the surface . In the 1d model , on reaching an edge the bead stopped there until the ParA concentration was recovered enough in its wake and then it would begin to move back to the other end . In 2d , our model found that a bead would change directions when it encountered a boundary but would never come to a halt , unlike in 1D ( see S2 Fig ) . Depending on the initial amount of ParA in the buffer , ab ( τ = 0 ) = As , there are two possible regimes defined by the quantity ϕ = As/D0: ϕ > 1 is the saturated regime where there is an excess of ParA and ϕ < 1 in which the system is undersaturated , and there is always an excess of binding sites for ParA . When the system was initialized with limited ParA such that ϕ < 1 , there was still the possibility of further rebinding at every point . On introducing the bead into the system , a spontaneous gradient forms and the bead starts traveling in a particular direction as in the previous sections . As the sites in front of the bead have the capability to bind more ParA , the wavefront attracting the bead increases in magnitude and the bead gains further speed ( Fig 4B ) . Thus in the undersaturated regime , we predict that it may be possible to observe persistent acceleration of the bead ( Fig 5A ) . In the regime ϕ > 1 , there is excess ParA in the buffer and during the equilibration phase ( before the ParB decorated bead is introduced ) , the bound ParA is nearly equal to the saturating limit d ( x ) while free ParA in the buffer still remains . Bead motion in this regime is similar to that described in the first section . A wavefront forms due to hydrolyzation and as all sites ahead of the bead are saturated the leading edge can not grow ( Fig 4C ) . Indeed , we find that the bead attains a uniform speed , experiencing only an initial short burst of acceleration . Within these two regimes a variety of bead behaviours can be observed on varying ϕ . In the undersaturated regime , when ϕ ≪ 1 the bead accelerates persistently , never reaching a saturated speed within the surface length L . We carried out a simple analytical calculation ( see S3 Text ) in this limit to determine the dependence of the bead’s speed on time which matched our simulation results ( see S3 Fig ) . For values of ϕ ≲ 1 it is possible for the bead’s speed to increase and saturate to a constant value . This happens when the bound ParA wavefront in front of the bead rises to saturate all the binding sites . In the oversaturated regime , when ϕ ⩾ 1 there is a possibility of the buffer ParA filling the wake region completely . If this occurs rapidly enough , we expect that there should be potential to stall the bead . For a fixed value of kr , we observed that increasing the ϕ led to reduced bead speeds until a value ϕstop was reached , at which the bead did not commence motion ( Fig 5C ) . Besides depending on ϕ , the dynamics of the bead also depend on the rate of rebinding , kr . We find that in the undersaturated regime when ϕ ≲ 1 the final value to which the bead speed tends , decreases with increasing kr to a saturated value ( Fig 5B inset ) . Hence , it is not possible to stall the bead no matter how high the rate of the rebinding is . In the saturated regime however the bead can be made to stop by increasing ϕ and this ϕstop depends on kr . Some analytics that show this dependence are given in the S3 Text section of this paper and agree well with our simulated results ( S3 Fig ) . We then considered the case of cooperative rebinding , where we assumed that rebinding depended not only on the amount of free binding sites ( d ( x ) − a ( x , τ ) ) but also on the amount of ParA bound at a given location . This cooperative rebinding was governed by a rate constant , kc while a non-zero kr gave the rate of non-cooperative rebinding from the buffer ( Eq 3 ) . In the regime where ϕ < 1 the bead behaviour was similar for simulations with and without any cooperativity in rebinding . As the amount of ParA in the buffer was scarce , it immediately redistributed to all unsaturated sites . In the ϕ < 1 regime , introducing cooperative rebinding increases the acceleration of the bead as lesser ParA rebinds to the wake and more ParA rebinds to the wavefront ahead . In the ϕ > 1 regime the role of the rates of rebinding become more prominent as there is ParA available for rebinding , but only in the regions from where the ParA is hydrolyzed . Increasing the cooperative rebinding rate led to the depletion zone filling in faster , making it possible for the wake to recover and stall the bead at lower values of ϕ ( Fig 5D ) . Hence , introducing cooperativity in ParA rebinding led to the bead stopping at lower values of ϕ than when there was no cooperative binding . Lastly , we considered the effect of surface diffusion of ParA in the absence of rebinding to determine if there were any significant differences in bead behaviour from that of rebinding from a well mixed buffer alone . We again included a saturating limit for the amount of ParA concentration that could exist at every point x . This coupled the amount of available ParA in the system to the binding site distribution , d ( x ) as ParA could diffuse to a point only if it had the capacity to bind more ParA . We assumed that ParA surface diffusion was governed with a diffusion coefficient , κ . The equation describing the bead dynamics remains the same while the equation describing a ( x , τ ) in the presence of surface diffusion and no rebinding is given by ( see S2 Text for details ) : ∂ a ( x , τ ) ∂ τ = - e - ( x - x p ) 2 / 2 c 2 a ( x , τ ) + κ [ d ( x ) ∂ 2 a ( x , τ ) ∂ x 2 - a ( x , τ ) d 2 d ( x ) d x 2 ] . ( 5 ) As the concentration of binding sites is spatially noisy this coupling maintains the spatial noise necessary for the formation of the spontaneous gradient in ParA that initiates bead motion . The system was initiated in a state such that all sites on the surface have bound ParA equal to the binding distribution . When ParA protein’s surface diffusion is low , the wake can not fill in rapidly enough to stall the bead and so it attains a steady state speed . By increasing the diffusion constant , just as was the case for rebinding rates , the bead could be made to stall ( Fig 6 ) . Thus qualitatively , the behaviour is nearly identical to the situation where only rebinding occurred from the buffer .
In this paper we have presented a minimal model for the operation of the in vitro ParA-ParB system . The model involved a tug-of-war between attractive forces exerted on a ParB decorated bead by surface bound ParA in front of and behind the bead . ParB on the bead would remove ParA from the surface , tilting the balance in favor of one side , leading to directed motion . For a range of parameter values , we found that spatial noise in the initial ParA distribution was sufficient to break spatial symmetry , causing the spontaneous formation of a ParA gradient between the front and back of the bead , leading to motion . It was found experimentally that identical beads could display different speeds when placed on the same DNA substrate . Our model provides some insight into parameters that influence a bead’s speed . The first such parameter is A0 which depends on both the initial ParA concentration as well as the amount of ParB on the bead . A simple explanation for the observed differences in speed , would be that the amount of ParB may not be the same on each bead , and hence , even though the surface bound ParA concentration that each bead sees is the same , the effective A0 would be different . Another contributing factor that could influence the amount of ParB on a bead that can interact with the surface , is bead bound ParA . Experimentally it was found that the ParA content on the beads undergoing directed motion was 25 ± 5% less than on beads which diffused freely [21] . This suggests that increased ParA on the bead lowers its ability to interact with surface bound ParA and transforms the bead into a freely diffusing particle . In our model , ParA bound to beads would affect A0 through the effective change in the amount of ParB that can interact with the surface . The speed of a bead also depends on the parameter c which is the ratio of the lengthscale of the ParA removal kinetics to the lengthscale of the ParA-ParB attraction force . We speculate that a change in this ratio could be experimentally achieved by changing the size of the DNA linker that binds ParB to the micron-sized bead . The changes in bead speed due to different c could be characterized by looking at the shape and size of their wakes . The model displayed a rich variety of behaviour when rebinding of ParA to the surface was considered . We showed that when the surface was saturated and there is always free ParA in the buffer that it should be possible to stall the bead . For the situation where the surface is unsaturated and free ParA can always find free sites to bind , we predict that persistent acceleration of the bead results , with the counter intuitive result that lesser total amount of ParA can actually lead to higher speeds . In order to potentially see acceleration , one would likely have to study beads on much narrower tracks so that the released ParA could have an appreciable effect on the wavefront when it rebinds . These predictions that only depend on the amount of ParA in relation to binding sites should be readily testable experimentally . A topic of some debate about the operation of the ParA-ParB system is the role of cooperative binding for ParA and whether the formation of filaments or ParA clusters is essential . We included cooperative rebinding of ParA in our model and found that it was qualitatively indistinguishable from non-cooperative rebinding in regards to the bead dynamics . More complex dynamics , that include having multiple beads ( the in vitro version of multiple plasmids in a cell ) may be able to disentangle whether cooperative rebinding has any detectable effect . This will be a topic for further exploration . We also allowed for ParA surface diffusion and found that it too led to dynamical behaviours that would be hard to distinguish it from non-cooperative rebinding . Although our model was developed to capture essential features of the in vitro ParA-ParB system , we feel that it may serve as a useful coarse grained model for studying the in vivo system . Future work towards this end would allow for ParB to diffuse and include discretizing the system to consider stochastic kinetic effects . Neither of these is currently in the continuum model presented here , but likely play a role in vivo . In summary , the model presented here makes several non-trivial predictions that should further aid the dissection of the operational mechanisms of this active transport system .
To derive the dimensionless Eqs 1 and 2 we start with a 1d version of our model in real space-time coordinates X and t . The concentration of ParA at every point on the surface is given by Am ( X , t ) which can only be removed by the ParB decorated bead . The rate of removal depends on the bead’s position , Xp and decays with distance from the bead . We assume that the rate has a Gaussian form , centered on the bead with a characteristic range of removal given by the parameter , σr . The bead has attractive forces acting on it due to its interactions with the ParA on the surface . The force between the bead and ParA on the surface decays with distance from the bead , and is directed along the surface in proportion to the X component of the vector connecting the bead to the surface location . Similar to the rate of removal , we consider the magnitude of the force to have a Gaussian form with a characteristic range given by σf . The total force is found by integrating over the entire surface . We consider that the dynamics of the bead is in the over damped regime and so the net force , F is proportional to the bead’s speed , dXp/dt . Putting all these assumptions together we start with the following dimensionful equations for the ParA-ParB system: ∂ A m ( X , t ) ∂ t = - γ 0 e - ( X - X p ) 2 / 2 σ r 2 A m ( X , t ) ( 6 ) β d X p d t = F = ∫ d X F 0 e - ( X - X p ) 2 / 2 σ f 2 X - X p R 2 + ( X - X p ) 2 A m ( X , t ) . ( 7 ) Here γ0 gives the rate of ParA removal , F0 is a multiplicative constant that scales the force per unit ParA and ParB concentration exerted on the bead and β is the drag coefficient of the bead given by β = 6πηR . The above equations are reducible to a dimensionless version by making the following transformations: X → R x , X p → R x p , t → τ / γ 0 ( 8 ) A m ( X , t ) → a 0 a ( x , τ ) ( 9 ) Where x , xp and τ are dimensionless variables and a0 is a multiplicative constant to scale the ParA in the system . Under these transformations and introducing the ratio c = σr/σf the dimensionless equations governing the dynamics become: ∂ a ( x , τ ) ∂ τ = - e - ( x - x p ) 2 / 2 c 2 a ( x , τ ) ( 10 ) v = d x p d τ = A 0 ∫ d x e - ( x - x p ) 2 / 2 x - x p 1 + ( x - x p ) 2 a ( x , τ ) ( 11 ) Here A0 = F0 a0/βR and c are the primary parameters of the system on which v depends . It should be noted that varying the constant A0 might imply varying the magnitude of initial ParA concentration a0 or magnitude of the force of attraction exerted by the ParA per unit ParB present on the bead , F0 or inversely varying the radius of the bead R . All these dependencies have been suitably combined into the single dimensionless parameter A0 , which when varied reflects changes in concentration of initial ParA , since both the strength of the attractive force and radius of the bead are assumed fixed and not readily changeable . Apart from A0 the only other parameter affecting bead speed is the ratio , c . We numerically integrate the above equations for a 1d system with surface length L = 70 and spacing dx = 0 . 02 using the Euler step method in steps of dτ = 0 . 01 . For each simulation we fixed an average initial ParA concentration and obtained the ParA concentration for every site by adding uniform noise with a magnitude δa . The resultant profile was a spatially noisy distribution about the mean ParA concentration . The bead was placed in the center of this surface to replicate the in vitro experimental process such that the bead is surrounded by ParA in all directions . The forces attracting the bead from the ParA are integrated using Simpson’s rule to obtain the total vector force on the bead and the change in its position can be calculated using Eq 2 . Bead speeds were obtained by doing a linear fit to the position vs . time graphs , ignoring the initial lag period . Similar transformations can be extended to obtain the following set of equations for a 2d system: ∂ a ( x , y , τ ) ∂ τ = - e - ( r - r p ) 2 / 2 c 2 a ( x , y , τ ) ( 12 ) ( r 2 = x 2 + y 2 , r p 2 = x p 2 + y p 2 ) d x p d τ = A 0 ∫ ∫ d x d y e - ( r - r p ) 2 / 2 x - x p 1 + ( r - r p ) 2 a ( x , y , τ ) ( 13 ) d y p d τ = A 0 ∫ ∫ d x d y e - ( r - r p ) 2 / 2 y - y p 1 + ( r - r p ) 2 a ( x , y , τ ) ( 14 )
|
Segregating genetic material is essential for cell survival over multiple generations . The process underlying the required spatio-temporal organization of DNA is mediated by the ParA-ParB-parS system . Recently , experiments have shown that directed motion can be reconstituted in vitro . In these experiments , a magnetic bead was covered with the protein ParB and was able to move ballistically over a surface of DNA that was bound by the protein ParA . How does this active transport spontaneously emerge ? In this paper we present a deterministic model for the dynamics of ParA-ParB proteins . We show how spatial noise in surface bound ParA is sufficient for the creation of a gradient in ParA that can drive motion of ParB in vitro . The model explains certain key aspects of the in vitro ParA-ParB system and leads to testable predictions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
Operational Principles for the Dynamics of the In Vitro ParA-ParB System
|
Schistosomiasis is a chronic parasitic trematode disease that affects over 240 million people worldwide . The Schistosoma lifecycle is complex , involving transmission via specific intermediate-host freshwater snails . Predictive mathematical models of Schistosoma transmission have often chosen to simplify or ignore the details of environmental human-snail interaction in their analyses . Schistosome transmission models now aim to provide better precision for policy planning of elimination of transmission . This heightens the importance of including the environmental complexity of vector-pathogen interaction in order to make more accurate projections . We propose a nonlinear snail force of infection ( FOI ) that takes into account an intermediate larval stage ( miracidium ) and snail biology . We focused , in particular , on the effects of snail force of infection ( FOI ) on the impact of mass drug administration ( MDA ) in human communities . The proposed ( modified ) model was compared to a conventional model in terms of their predictions . A longitudinal dataset generated in Kenya field studies was used for model calibration and validation . For each sample community , we calibrated modified and conventional model systems , then used them to model outcomes for a range of MDA regimens . In most cases , the modified model predicted more vigorous post-MDA rebound , with faster relapse to baseline levels of infection . The effect was pronounced in higher risk communities . When compared to observed data , only the modified system was able to successfully predict persistent rebound of Schistosoma infection . The observed impact of varying location-specific snail inputs sheds light on the diverse MDA response patterns noted in operational research on schistosomiasis control , such as the recent SCORE project . Efficiency of human-to-snail transmission is likely to be much higher than predicted by standard models , which , in practice , will make local elimination by implementation of MDA alone highly unlikely , even over a multi-decade period .
Schistosomiasis is a neglected tropical disease ( NTD ) having an estimated global prevalence of 240 million infected persons , many of whom experience significant morbidity within the infected communities of Africa , the Mideast , South America , Asia , and the Philippines [1] . For global control of the disease schistosomiasis , the World Health Organization ( WHO ) recommends delivery of the anti-helminthic drug , praziquantel , via mass drug administration ( MDA ) , with attempts at local elimination , where possible [1 , 2] . Unlike the very effective MDA experience obtained for other helminthic NTDs such as onchocerciasis and lymphatic filariasis [3 , 4] , there remain significant concerns about the feasibility of schistosomiasis elimination using MDA alone [5] . This is in part due to that fact that MDA has been unable to interrupt schistosomiasis transmission in many endemic areas , even after a decade or more of repeated MDA [6 , 7] . This failure to interrupt transmission has often been marked by a significant rebound of infection prevalence following termination of MDA [8–10] , or of concurrent mollusciciding interventions [11] . The highly uneven landscape distribution of suitable intermediate host snail habitat , combined with weather- and climate-related seasonal differences in snail abundance , mean that there is often a quite varied patchwork of transmission zones within any given region slated for parasite control [12–16] . Understanding the mechanisms that drive infection rebound is crucial for the development and implementation of more efficient control strategies [1] . Conventional predictive models of transmission suggest that where rebound is slow , there can be progressive reduction of parasite burden after each MDA cycle , so we may expect to bring Schistosoma burden under control and achieve elimination of transmission [17] . On the other hand , rapid rebound of parasite burden following treatment serves to impede long term progress towards elimination goals , and necessitates additional MDA effort and/or introduction of complementary environmental control measures to achieve parasite elimination [5 , 18 , 19] . In developing transmission models , important but often overlooked determinants of schistosome transmission are the ecology and population biology of the intermediate snail host and accurate assessment of the human-to-snail force of infection ( FOI ) . As part of the transmission cycle , Schistosoma must infect very specific intermediate host snail species , then undergo a process of extensive asexual multiplication within the snail’s body in order to create the free-swimming cercariae that will infect the next round of human hosts [20 , 21] . For Schistosoma parasites of humans , local presence of freshwater snail species of genera Bulinus , Biomphalaria , Oncomelania , or Neotricula , is essential to the transmission of Schistosoma haematobium , S . mansoni , S . japonicum , and S . mekongi , respectively [22] . Because snail infection is an obligate stage for parasite transmission , ecological factors that favor the presence and abundance of these ‘vector’ snails also foster local risk for these Schistosoma spp . infections and for their related human disease states , either urogenital or intestinal schistosomiasis [15 , 20 , 23] . Conventional transmission models assume the snails’ FOI is a linear function of human infectivity ( see , e . g . [24–26] ) . Under this assumption , any drop in human infectivity ( e . g . via MDA-related reduction in local egg excretion ) , will proportionately reduce the rate of local snail infections , which in turn will slow reinfection of human hosts . However , the empiric field data from recent large-scale , cluster- randomized operational research trials of anti-schistosomal MDA [27] have demonstrated a broad range of community-level parasitological responses , ranging from highly effective reductions in prevalence and intensity at some locations , to the existence of highly resistant “hotspots” ( Fig 1 ) , where infection levels persist at or near baseline levels despite effective implementation of MDA [28 , 29] . While the current simplified deterministic models mimic the average effects of MDA across all communities , the failure to account for broad village-by-village variability is a challenge to the general utility of transmission model-based predictions . Prior modeling studies of other micro-and macro-parasite systems have established that the form assumed for the transmission coefficient ( beta ) can have a significant impact on the projected outcomes for disease ecology models [30–32] . As a simplifying initiative , most modelling approaches typically assume ‘density-dependent’ kinetics , in which two well-mixed populations provide a constant per capita rate of exposure [33] . However , some models have elected to employ ‘frequency-dependent’ kinetics , which can exhibit saturation of transmission at higher host densities or in the presence of non-random mixing [30 , 34 , 35] . In such models , landscape patchiness , associative movement networks , time-dependence , and heterogeneity in host susceptibility can explain the failure of standard ‘mass action’ transmission coefficients to accurately capture the trajectory of disease transmission in real-world settings [30 , 33 , 36–38] . These transmission features are common among vector-borne macroparasites such as the Schistosoma species studied here . Where such features exist , it is apparent that fine-scale transmission events in linked territories can serve to drive larger meta-population patterns of infection prevalence [30] . To explain observed heterogeneities in Schistosoma transmission , we undertook a closer examination of the intermediate snail host and its infection by humans . Schistosoma transmission and parasite development have multiple time scales , ranging from “fast” larval dynamics ( hours , days ) , to “slow” ( month , years ) host-parasite-snail dynamics . In the current study , we focused on these slower dynamics , so larval stages did not enter our model formulation explicitly . However , we saw the need to have an accurate account of their effect on human and snail infection . Conventional modeling approaches assume each FOI to be proportional to its source infectivity and population size [33] . We reexamined the conventional model assumptions , and derived a newer formulation of human-to-snail FOI that combines human host infectivity , demographics and snail population inputs . Among other salient features of the proposed FOI is its nonlinear dependence on human egg output . This functional form could be linked to the magnitude of the post-MDA prevalence rebound and to the consequent success or failure of long-term control . We explored the effect of modifying snail FOI in simulating MDA responses for typical endemic communities , comparing “nonlinear” vs . “linear” models . The two models produced markedly different outcomes , particularly in high-intensity transmission settings .
The schistosome parasite maintains a complex life cycle , transiting between human and snail hosts , with the transition mediated by two larval stages , the egg-derived miracidium ( for human-to snail movement ) , and the snail-derived cercaria ( for snail-to-human movement ) [21] . For this study , we applied a previously developed dynamic model that describes this biological process . We denote the corresponding forces of infection λ ( for snail-to-human ) , and Λ ( for human-to-snail ) . The former ( λ ) represents the mean rates of worm accumulation by human hosts , the latter ( Λ ) , the mean rate of snail invasion by miracidia . Each force depends on its host carrier’s infectivity , population abundances , and the frequency and pattern of their contact ( human water exposure and water contamination rates ) . In our setup , human FOI is proportional to infected snail prevalence ( 0 < y < 1 ) , λ = A y , with transmission coefficient A . Snail FOI is a function of human infectivity , E ( mean egg release ) , but its derivation requires careful analysis . Most conventional models employ linear Λ = B E with transmission coefficient B [25 , 26 , 39] . Here , instead , we propose a nonlinear ( saturable ) form of snail FOI , Λ=Λ0 ( 1−e−bE ) ( 1 ) The derivation of ( 1 ) is outlined in the supporting information S1 File . It employs some natural assumptions on miracidial dynamics from human egg release , its diffusive spread , and the process of snail invasion . We assumed miracidia randomly cluster about snail host , with a Poisson distributed “miracidia/snail” ratio . The resulting saturable ( exponential ) function ( 1 ) is the probability of successful invasion ( see S1 File ) . To study the effect of nonlinear FOI , we programmed two coupled human-snail model systems , termed M1 ( having a linear snail FOI factor ) , and M2 ( having a nonlinear FOI given by Eq ( 1 ) ) . Nonlinear Λ had two coefficients ( Λ0 , b ) that , through local model calibration , reflected important local environmental , biological , and behavioral inputs . Λ0 can be viewed as maximal rate of miracidial invasion in a given environment . It depends on local snail density ( which determines “mean travel time” to reach target ) , and on search strategies employed by miracidia ( see [40] for a general discussion of encounter rates ) . Coefficient b is related to the mean miracidia production by human hosts and the probability of snail invasion by miracidia . Additional factors that enter b include mean population density ( host/snail ) , and human-snail contact ( exposure/contamination ) rates . Different types of human and snail models can be coupled via FOI terms λ , Λ . Here we adopted a stratified worm burden ( SWB ) approach ( for the human part ) , developed in earlier works [41–44] , but one can also use a simpler MacDonald-type mean worm burden ( MWB ) system [24 , 45] . The basic differences between models M1 and M2 , and their projected control outcomes , are due primarily to the Λ -function , whereas a specific formulation for the human side of the coupled model proved less influential . Importantly , there can be a convergence between linear and nonlinear FOI systems: Function ( 1 ) can be approximated by a linear function Λ ( E ) ≈Λ0bE ( 2 ) at small contagion levels ( i . e . , b E ≪ 1 ) . So our nonlinear Λ ( 1 ) can be viewed as an extension of linear form ( 2 ) to reflect larger values of human infectivity . Specifically , the M1 and M2 FOIs depart significantly as E or b grow large; the latter , in particular , embodies higher human-to-snail ratios or higher contact rates . Notably , the two FOI systems can also give markedly different values of transmission coefficients , even when calibrated against the same datasets . For snail infection modeling , we used a standard simple S-I transmission system ( x–susceptible ( S ) , y–infected ( I ) ) with stationary population density ( x + y = 1 ) . The prevalence variable 0 < y ( t ) < 1 , solves differential equation dydt=Λ ( 1−y ) −νy , ( 3 ) with snail FOI , Λ . For the present analysis , a human SWB model was used , consisting of variables h→ ( t ) ={hm ( t ) }- ( worm burden strata ) that undergo dynamic changes due to worm accumulation and loss processes . The detailed exposition of SWB approach has been described in detail in previous publications [39 , 41 , 42] , and it is briefly summarized in S1 File . A conventional MWB setup [39 , 41 , 42] can also be used if desired . It has a single dynamic variable , MWB w ( t ) , that obeys differential equation dwdt=λ− ( γ+μ ) w ( 4 ) with human FOI ( λ = Ay ) depending on snail prevalence ( 3 ) , and loss term ( γ + μ ) which combines worm mortality , γ , and host turnover , μ . The two models , MWB and SWB , share common input parameters ( λ , γ , μ ) . In fact , MWB Eq ( 4 ) follows from the SWB if one takes the first moment ( mean ) of the {hm}-distribution , w ( t ) =∑m>0mhm ( t ) . The main difference between the SWB and MWB approaches lies in their assumptions on within-humans worm distribution patterns , {hm} , and the resulting human infectivity E ( see , e . g . [41] ) . The SWB imposes no constraints on variables {hm} , whereas MWB uses a priori assumptions to express E as a function of w ( t ) . Typically in the MWB model , {hm}are assumed to follow a negative binomial ( NB ) with prescribed aggregation constant , k . In both systems , human infectivity is a product of mean mated worm count ( MMC ) Φ , and worm fecundity ρ , with E = ρΦ . The MWB gives MMC as a function of variable w , Φ ( w , k ) , while SWB function Φ ( h→ ) depends on worm burden strata h→={hm} . Two different FOI {λ ( y ) , Λ ( E ) } couple transmission dynamics between human and snail hosts , and give rise to a coupled SWB-snail model . The setup can be can be extended to demographically-structured populations made of multiple risk/age groups , each carrying specific burden distributions . In our analysis we employed structured host communities made of child ( C ) and adult ( A ) age groups , with age-specific FOI and transmission coefficients , λC = ACy , λa = Aay . The combined infectivity of such system depends on MMC Φi of each group , their age-specific worm fecundities ρi , population fractions ( Hc + Ha = 1 ) , and contact ( exposure/ contamination ) rates ωi . Their combination gives the following dimensionless form E=ρ ( HcΦc+ωHaΦa ) ( 5 ) Factor ρC is the mean worm fecundity of the child group , while weight ω is the product of relative ( child-to-adult ) fecundity and exposure factors ( S1 File ) . The child age-group worm fecundity is subsumed as a factor in the transmission coefficient , b , so doesn’t enter the model explicitly . Calibration of the coupled systems proceeded in two steps: ( i ) human egg-count ( diagnostic test ) data were employed to estimate snail-to-human FOI and worm fecundity ( λi , ρi ) for each human subgroup . The outcome was a best-fit posterior distribution of the model parameter space; ( ii ) next , the calibrated human parameters were combined with additional environmental/behavioral ( snail ) data to estimate transmission coefficients Ai ( snail-to-human ) , and either {B , ω} ( for linear FOI ) , or triplet {Λ0 , b , ω} ( for nonlinear , Λ; see S1 File , Part B for details ) . In our predictions , we used similar snail inputs ( baseline prevalence , y* ) and the relative adult/child exposure factor , ω , in both model systems M1 and M2 . However , nonlinear FOI ( M2 ) had an additional parameter , b , which encoded the relative human/snail population factor ( H/N ) . In our sensitivity analysis , we varied b to simulate a broad range of environments and explore its effect on MDA outcomes . Drug treatment with praziquantel kills a large fraction of adult Schistosoma worms , and its clearing efficacy is estimated at 80–95% [5] . In our simulations , we have set this value at 85% ( using a surviving worm fraction , ε = . 15 ) . The key inputs for MDA program simulation consisted of target group sizes ( children , adults ) , their coverage levels ( e . g . 0 < fc < 1 , 0 < fA < 1 ) , and the timing or frequency of MDA delivery ( annual , biennial , etc . ) . In our numeric simulations , MDA was implemented as an instantaneous event , whereby worm burden of each group is reduced depending on its coverage and drug efficacy , so the dynamical transmission system was reinitialized at time td after each control event . For the SWB system , an MDA event results in reshuffling of burden strata , so that each higher-burden stratum shifts to lower-burden strata hm → hεm ( see [42 , 43] ) . For a corresponding MacDonald-like MWB system , each MDA event with coverage f , and efficacy ε , would reduce MWB w ( td ) by a factor ε f + ( 1 − f ) . For analysis and MDA simulations we modeled three communities from past Kenyan control-surveillance studies 1983–92 [46] , and 2000–2009 [12 , 13] , having heavy ( H ) , moderate ( M ) or light ( L ) infection levels ( see Table 1 ) . This dataset was extensively used in our previous SWB work [10] , and in more recent papers [41–43 , 47] . The latter have employed refined SWB methodology to account for in-host biology ( worm mating , aggregation , random egg release ) , and have introduced more advanced calibration methodologies . The modeled high-intensity community ( H ) was subject to longitudinal study spanning nine years , with two MDA sessions ( in 2001 and 2003 ) , and three population-wide surveillance screenings ( in 2001 , 2003 , and 2009 ) . For the purpose of the current comparative modeling analysis , we divided the village population into child ( 0–20 year old ) and adult ( 20+ years ) age groups ( Table 2 ) based on Kenyan demographics . Additional model parameters included in the simulations were worm mortality and snail survival as described in S1 File , Table A1 . The two study models ( M1 , M2 ) were calibrated for each of our high- , moderate- and low-intensity sample communities following [42] . The calibration procedure involved two-steps: ( i ) individual egg-count test data at baseline ( Year 2001 ) were employed to define a posterior distribution of likely parameter choices ( λ , ρ , k ) for age-groups C and A . The calibration results ( marginal distributions of human parameters and their statistics ) are described in S1 File , part B . The next step used the estimated human parameters ( from our first-stage calibration’s posterior distribution ) to estimate transmission coefficients . Snail-to-human transmission coefficients Ai ( i = C , A ) were identical for M1 and M2 . The human-to-snail components were different: {B , ω} for the linear-FOI model M1 , and {Λ0 , b , ω} for the nonlinear M2 . All depended on infected snail prevalence ( both prepatent and patent ( i . e . , cercaria-shedding ) ) , which was fixed at value y* = 0 . 3 , consistent with PCR-based snail surveillance findings in the Kenyan environment [48] . Patent snail density , which is responsible for transmission , was assumed to be proportional to infected snail prevalence , y ( t ) . There were two additional inputs ( y* , ω ) for M1 , and 3 additional inputs ( y* , ω , b ) for M2 . The relative adult/child exposure ratio , ω , was set at 1 . 5 , and b combined a transmission coefficient ( miracidia contagion release the by the child age group ) times relative host population abundance ( human/snail ) ( see S1 File , part B ) . Because these values have been less well studied , in sensitivity analysis we allowed broad range of uncertainties: 0 . 5 < ω < 5; 0 . 5 < b < 5 , for both of these transmission variables .
The calibrated model community , using a consistent choice of transmission uncertainties ( y , ω , b ) , was subjected to a series of control experiments to explore the effect of snail FOI assumptions ( model M1 vs . M2 ) and the role of ( y , ω , b ) on long term MDA outcome patterns in different environmental settings . A typical 10-year history for a high-risk community is shown in Fig 2 . The model parameters used in this simulation are listed in Table 3 . For this analysis , annual community MDA was used , with an estimated 75% annual coverage for children , and 35% biennial coverage was used for adults . The simulation results show large differences between M1 and M2 projections , with the M1 system rapidly approaching elimination , whereas M2 becomes locked in a limit-cycle pattern and does not approach elimination ( Fig 2 ) . This qualitative distinction between the models—mainly that M2 model was considerably less likely to achieve MDA-mediated elimination—persisted for a range of parameter choices and MDA coverage . In sensitivity analysis of our prediction by random sampling of model parameters ( human and environmental ) over a broad range of values with identical M1 and M2 communities subjected to the same MDA regimen , significant differences remained in projected outcomes . History envelopes ( Fig 3 ) show ensemble mean and 95% CI for the multiple simulated 10-year MDA programs . The M1 histories consistently go to elimination , while the M2 histories settle into recurrent limit cycles that fail to achieve elimination . To help validate our approach , we used an observed longitudinal dataset collected over 9-year period for the base case high-risk community , Milalani , in Kwale County , Kenya [46 , 49] . The community was screened in ( 2001 , 2003 , 2009 ) , with two MDA sessions run in 2001 ( community-wide coverage 79% ) , and in 2003 ( community-wide coverage 41% ) . The results of study are summarized in Table 4 . To assess prediction potential of linear and nonlinear models , both systems were fitted to the baseline infection dataset ( 2001 ) . As explained in Methods , this yields a posterior ensemble of best-fit calibrated human parameters ( λ , k , ρ ) . We then sampled random choices from this posterior ensemble , along with three additional environmental inputs , ( ω , b , y* ) , to get the estimated transmission parameters for M1 and M2 ( see Table 3 ) . Each virtual community ( parameter choice ) was simulated over a 9-year period subject to two MDAs . Typical model outcomes are shown in Fig 4 , with comparison to observed field data . On both follow-up years ( 2003 , 2009 ) , we observed significant relapse toward pre-control ( endemic ) levels of infection . Of the two calibrated models , the nonlinear M2 was able to reproduce this pattern for child and adult groups . However , the M1 model did not capture post-treatment prevalence values with its slower intrinsic relapse rate . We again tested parameter sensitivity for robustness of our predictions . This test was run separately for three environmental inputs: i ) the relative exposure factor was varied in the range 0 . 5 < ω < 5 , ii ) the child transmission rate was varied in the range 0 . 5 < b < 5 ( for M2 ) , and iii ) a random variation of best-fit panel parameter inputs ( λi , ρi , ki ) of the calibrated community was used in each replicate simulation . In each case , an ensemble of 9-year histories was simulated . Solution envelopes of these ensembles along with their mean path are plotted in Fig 5 ( panels a , b , and c ) . The envelopes are less sensitive to relative exposure factor ω , but child transmission b had more pronounced effect . The uncertainties due to human inputs { ( λi , ρi , ki ) } , come from the baseline posterior calibration , as shown in panel ( c ) of Fig 5 . In all cases , observed data points lie within prediction envelopes . As discussed earlier in Methods , nonlinear snail FOI becomes approximately linear at low levels of human infectivity . To explore the effect of a reduced transmission environment on long term MDA , we subjected three sample communities with heavy ( H ) , moderate ( M ) or light ( L ) transmission intensity , respectively , to the same 10-year control regimen , and compared projected prevalence outcomes for M1 vs . M2 simulations ( infection prevalence ) . Fig 6 shows the comparative results . The difference in simulations is unambiguous for the high risk community , where M1 predicts gradual decline towards elimination , whereas M2 shows strong rebound to moderately high prevalence levels ( 15–30% for children ) each year . For moderate risk areas , the two curves for M1 and M2 are closer , although M2 still predicts a persistent cycle of reinfection . For the low risk community ( L ) the discrepancy between models appears marginal , with M1 and M2 closely following each other .
In this modeling study , we systematically compared two model structures for Schistosoma transmission to better understand the importance of non-linear snail vector dynamics for model prediction of long-term intervention outcomes . We calibrated two transmission models with identical human host inputs but different human-to-snail transmission coupling—a conventional model with linear FOI assumption ( M1 ) and a more complex model assuming a nonlinear saturable FOI for snails ( M2 ) –using longitudinal data collected in coastal Kenya [10 , 42] . We subjected both models to a series of numeric experiments simulating different MDA regimens , and found marked differences in long-term epidemiologic predictions . The conventional M1 model predicted efficient control ( reaching targeted reductions , then elimination ) after relatively few rounds of MDA , even in the face of low or moderate treatment coverage levels . The proposed M2 model , however , found many settings to be highly refractory to MDA treatment impact , with persistent Schistosoma re-infection even with high treatment coverage levels . In the model validation , we found the M2 model with its non-linear snail FOI formulation to be more reflective of empirically observed data [10 , 29] . Going forward , these findings have clear implications for program monitoring and evaluation and future control implementation for schistosomiasis control , suggesting that a non-linear FOI function should be incorporated for more realistic projections in future Schistosoma transmission modeling . Empirical evidence from other host-pathogen systems [33 , 38 , 50–53] suggest that there is likely to be a continuum in transmission kinetics that must be considered when modeling the observed transmission patterns found in settings where host numbers and distribution are varied . Although they are more complex and require more data , more nuanced modelling systems are expected to yield better understanding of parasite dynamics and the impact of control interventions [33] . Previous modeling work on S . japonicum transmission by Liang and colleagues [54] has incorporated multiple human risk groups identified by location and occupation , as well as seasonal aspects of snail reproduction and development . When calibrated against field data , this model more accurately projected the re-emergence of infection in high-risk communities when MDA and other interventions were stopped . Prediction of ‘bounce-back’ risk will be essential in determining the design of follow-up surveillance programs as local elimination is attempted . As noted above , the accurate calibration of such models requires more information about the control areas . However , the greater precision of model projections should improve the efficiency of program interventions [54] . In the presence of nonlinear FOI , a relatively small infective human host pool can exert a disproportionate , leveraged effect on snail infection . Hence , even a steep drop of human infectivity post MDA may result in only a marginal drop of snail infections , and this phenomenon , in turn , may result in a vigorous rebound or human infection to pre-treatment levels as noted in the SCORE project persistent hotspots [28 , 29] . In our analysis , we have used independent longitudinal data from communities in rural Kenya to formally compare the proposed non-linear snail FOI models with more conventional models to understand the impact of this effect on long-term model prediction . In our study system , the concept of a nonlinear , saturable pattern for snail FOI in Schistosoma transmission environments ( as proposed in the M2 model ) has biological plausibility: i ) Water contamination occurs in pulses , as infected humans only intermittently contaminate their environment with urine or feces [55] . Human treatment coverage is non-random , with people who are non-adherent to MDA perhaps the most likely ones to contaminate the snail environment ( i . e . , as effective superspreaders ) ; ii ) the miracidia that hatch from contaminating eggs selectively home onto local vector snails in order to infect them [56] , iii ) because of substantial asexual reproduction of the sporocyst , each infected intermediate host snail has the potential to release thousands of infective cercariae [57 , 58] , and iv ) cercariae sense human skin lipids , and preferentially swim toward any persons coming into contact with affected water bodies [59] . These nonlinear features all bias the transmission process in favor of higher levels of human infection and post-MDA reinfection . Specifically , this means that the extra-human phase of Schistosoma transmission is not a random , mass action process , although , for simplicity’s sake , many current models of transmission have assumed that it is . The coupled human-snail transmission dynamics in a model of schistosomiasis transmission are driven by two FOI: human-to-snail ( Λ ) , and snail-to-human ( λ ) . Each FOI is dependent on its source population size and infectivity , and given the predictive limitations of conventional models , our findings suggest that future models should include an updated accounting of these parasite invasion processes . The two obligate trematode hosts ( human and snail ) are treated differently in mathematical models of schistosome transmission but their FOIs are often assumed to be linear functions of the combined host infectivity . While such an assumption appears justified for human FOI , λ , snail FOI Λ requires more careful elaboration . In our current analysis , we derived a nonlinear saturable snail FOI function , which embodied several essential environmental ( e . g . type of water source , sanitation ) , demographic ( e . g . age distribution ) , and behavioral inputs ( e . g . contact with water , defecation practices ) , including human/snail population densities ( H , N ) and their contact/exposure rates . Given the difficulty of empirically measuring many of these aspects , we calibrated a composite estimate of FOI that reflected many complex and often heterogeneous factors . The conventional linear and proposed nonlinear functions were approximately equal at low levels of human contagion , where the linear FOI could be viewed as an adequate approximation of what is actually a nonlinear Λ . However , the two FOI versions diverged at higher levels of contagion , and so yielded very different transmission parameter estimates when fitted to the same human-snail infection data . The M1 and M2 models , based on the two different systems , also responded differently to strong perturbations , as occurs with MDA interventions; the M2 models predicting substantially faster post-MDA rebound as compared to M1 models . The human part of our present coupled system analysis employed SWB methodology [39 , 41–43] , but the qualitative conclusions of the M1-M2 comparison would remain true for other transmission models , including MacDonald-type MWB models [24 , 45] . Only the nonlinear ( M2 ) was able to accurately reproduce the strong rebound of infection seen in the dataset in years 3 and 9 of the Kenya project . This would predict that such communities will be resilient to any attempts at targeted elimination of transmission . In many cases the temporal differences between the two model systems ( M1 & M2 ) were large , in that M1 community model projections typically achieved control targets over a short time-span with moderate effort , compared to M2 models , where infection was projected to persist much longer and to require extended treatment intervention . In a separate project , we have explored , in greater depth , possible elimination strategies using combined MDA and environmental snail control , and we predict that the only way to achieve target reduction in high transmission communities would be via implementation of additional environmental interventions , e . g . combining MDA with molluscicide-based snail control [11 , 44] . For the nonlinear M2 system , three factors contribute independently to snail FOI estimation , accounting for a variety of MDA responses ranging from near-linear , efficient reduction /elimination in lower prevalence communities , to a highly resilient “locked” pattern of reinfection , whereby each MDA-mediated drop in prevalence is matched by post-treatment rebound . This latter feature could provide a key to the hotspot phenomenon observed in many control programs ( see , e . g . [28 , 29] ) . Indeed , it can explain why adjacent communities with near identical baseline human infection can produce divergent MDA responses based on variations in their local snail environment and in human behavior [60] . Importantly , while the proposed non-linear model demonstrates improved predictive value , this benefit should be balanced with the need for additional community data and more complex parameter estimation . The principle finding of this study is that a relatively simple non-linear function , on average , outperforms a linear function even when considering parameter uncertainty . Our analysis suggests a defining role of transmission environment ( and its resultant snail FOI ) for predicting MDA control outcomes . The heterogeneity and connectedness across Schistosoma transmission landscapes [16 , 45 , 61] , along with substantial parasite replication in the snail host , appear to make Schistosoma infection control much more challenging than for the filarial parasites that are transmitted by insect vectors [3 , 4] . In particular , MDA-based ‘transmission control’ for schistosomes will be particularly fragile in the face of persistent non-adherence to treatment ( or sanitation ) by a small group of infected residents or migrants [44 , 45 , 62] . In summary , there are substantial complexities in the human and snail factors that can affect Schistosoma transmission dynamics and related predictions of MDA-based schistosomiasis control outcomes . This study finds that nonlinear human-snail coupling ( FOI ) can improve model prediction . Although other model structures could also provide broad agreement with the data , nonlinear snail FOI could provide a plausible explanation of strong MDA resilience ( hotspots ) observed in the SCORE studies and the observed heterogeneous community responses reported elsewhere [28 , 29] . The present work will motivate future studies to apply these ideas to connected human-snail environments ( see [14] , [63] ) , and to the analysis of recent control datasets to develop tools to more accurately predict hotspots and explore strategies for their efficient control .
|
Infection with blood fluke Schistosoma parasites is a major cause of disease burden around the world . Control of schistosomiasis , which is transmitted through intermediate host freshwater snails , is a priority for national and global health programs working in at-risk regions of Africa , the Mideast , Asia , the Philippines , and South America . Program planning often relies on mathematical models to project the impact of different schedules of mass drug administration ( MDA ) of the anti-schistosomal drug , praziquantel , in these areas . In practice , though , recent projections of standard models have failed to capture the variability of MDA program impact on community levels of infection , especially in high-risk zones . In the present study , we developed a modification of the conventional modeling approach that takes more detailed account of human-to-snail transmission . Inclusion of a revised , nonlinear form for the model’s snail infection function had profound effects on long term predictions of the impact of MDA programs for Schistosoma control . In specific , our proposed snail parameters helped to explain the persistent rebound of Schistosoma prevalence in certain high risk communities . The efficiency of human-to-snail transmission is likely to be much higher than predicted in standard models , which makes local elimination by implementation of MDA , alone , highly unlikely .
|
[
"Abstract",
"Introduction",
"Methods",
"and",
"models",
"Results",
"Discussion"
] |
[
"schistosoma",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"helminths",
"tropical",
"diseases",
"parasitic",
"diseases",
"animals",
"systems",
"science",
"gastropods",
"mathematics",
"neglected",
"tropical",
"diseases",
"population",
"biology",
"infectious",
"disease",
"control",
"snails",
"fecundity",
"infectious",
"diseases",
"computer",
"and",
"information",
"sciences",
"nonlinear",
"systems",
"schistosoma",
"haematobium",
"molluscs",
"population",
"metrics",
"helminth",
"infections",
"schistosomiasis",
"eukaryota",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"organisms"
] |
2018
|
The human-snail transmission environment shapes long term schistosomiasis control outcomes: Implications for improving the accuracy of predictive modeling
|
In phenotypically heterogeneous microbial populations , the decision to adopt one or another phenotype is often stochastically regulated . However , how this stochasticity affects interactions between competing microbes in mixed communities is difficult to assess . One example of such an interaction system is the competition of an Escherichia coli strain C , which performs division of labor between reproducers and self-sacrificing toxin producers , with a toxin-sensitive strain S . The decision between reproduction or toxin production within a single C cell is inherently stochastic . Here , combining experimental and theoretical approaches , we demonstrate that this stochasticity in the initial phase of colony formation is the crucial determinant for the competition outcome . In the initial phase ( t < 12h ) , stochasticity influences the formation of viable C clusters at the colony edge . In the subsequent phase , the effective fitness differences ( set primarily by the degree of division of labor in the C strain population ) dictate the deterministic population dynamics and consequently competition outcome . In particular , we observe that competitive success of the C strain is only found if ( i ) a C edge cluster has formed at the end of the initial competition phase and ( ii ) the beneficial and detrimental effects of toxin production are balanced , which is the case at intermediate toxin producer fractions . Our findings highlight the importance of stochastic processes during the initial phase of colony formation , which might be highly relevant for other microbial community interactions in which the random choice between phenotypes can have long-lasting consequences for community fate .
Interactions like cooperation and competition between different organisms govern ecosystem dynamics , influencing ecosystem composition [1] , maintenance of biodiversity [1–4] , and the microbiota–host relationship [5–9] . Despite the detailed knowledge of individual interaction mechanisms on the one hand [10 , 11] and large-scale sequencing based microbiome studies on the other hand [7–9 , 12 , 13] , the fundamental problem in microbial ecology is the need for predictive model systems that combine experiments with theoretical modelling to explain how ecosystem dynamics emerge from interactions between single cells [14 , 15] . Simple bacterial model systems allow the elimination of distorting factors and enable the investigation of fundamental processes governing these dynamics , such as stochasticity , under well-defined experimental conditions [2–4 , 16 , 17] . In principle , cooperative [18 , 19] and competitive [10 , 20] interactions can occur between members of the same or of different species [21] and are mediated by various mechanisms [11 , 14 , 21–24] . Cooperative interactions are often realized by phenotypic heterogeneity , which is a division of labor within isogenic populations [24] , in which only a subpopulation produces a public good . Competitive interactions are achieved indirectly by competition for resources such as nutrients and space [22] but can also act directly by production of bacteriocins [7]—protein-based toxins produced by many microbes [10] . One class of bacteriocins , the colicins , are synthesized only by a subpopulation of the producer strain [25–27] . The ColicinE2 is one of these heterogeneously produced colicins by E . coli . Here , similar to many other colicins [25] , phenotypic heterogeneity is a consequence of the regulation via the noisy SOS response [26 , 28 , 29] , which can be triggered by the inducing agent mitomycin C ( MitC ) [25] , with higher MitC levels increasing the fraction of toxin producers [30] . The dynamics of ColicinE2 production have been investigated in detail [30] . Individual bacteria switch into the toxin-producing state stochastically [31] . Once a cell has switched into the producing state , it produces the toxin for approximately 60 min and subsequently releases the produced toxin into the environment upon cell lysis [30 , 32] . Cells that do not produce the toxin continue to reproduce . Hence , two different types of cells ( toxin-producing and reproducing cells ) are present , and a division of labor is established . Especially at small cell numbers , the stochastic switching dynamics can have important consequences . However , for large cell numbers , the stochastic nature of toxin production is less important , and the system approaches a steady state with constant producer fraction . The combined action of the two cell types ( toxin-producing and reproducing cells ) can be seen as a specific type of indirect cooperation that does not rely on communication but in which the tuning of producer fractions is determined by the architecture of the gene regulatory network and its response to an external signal ( MitC concentration ) [30] . Hence , the production of ColicinE2 by E . coli incorporates both direct competition ( by production of toxins ) and cooperation ( by means of division of labor/phenotypic heterogeneity ) . The competition between colicinogenic and susceptible strains of bacteria has been studied as a model system for allelopathy [33–35] . Theoretical studies highlighted the importance of colicin production cost , toxin effectiveness , and initial strain ratios as determinants of the outcome of competition [33] . Furthermore , long-term coexistence of both strains was found to emerge only in structured habitats [33] . More recent studies included experiments and focused on coexistence and biodiversity of toxin producer , toxin-sensitive , and/or resistant strains [1 , 3–5 , 36] . However , these studies mainly neglected the cooperative aspect of toxin production by the ColicinE2 producer strain ( C strain ) . In particular , to our knowledge , the influence of heterogeneity and stochasticity in toxin production on competition outcome and C strain success is largely unexplored , and quantitative experimental validation of theoretical predictions remains lacking . To address this problem , we employed an experimental approach using a stereoscopic microscope with zoom functionality , which enables investigations over multiple length-scales . In particular , this setup allowed us to correlate well-defined initial conditions with the macroscopic competition outcome of bacterial range expansions , hence , studying the impact of stochastic effects during initial colony formation on competition outcome . Computational modelling of the interaction by means of stochastic , spatially extended , lattice-based simulations supplemented the experiments . Using these two tools , we investigated the interaction of a C strain with a strain sensitive to the toxin ( S strain ) ( Fig 1A ) . This system exhibits both indirect intra-strain cooperation between reproducers and toxin producers within the C strain population and inter-strain competition between the C and S strain , mediated via toxin action and denial of access to resources by spatial exclusion . Our data revealed that the competition dynamics can be divided into two phases . In the initial phase of competition , at small cell numbers , stochastic effects originating from the stochastic toxin production dynamics and random initial positioning influenced the number of viable C clusters at the colony edge . In the second , deterministic phase , the degree of division of labor and the number of viable C clusters at the colony edge determined the final outcome of competitions .
Mixed bacterial communities of fluorescently labelled C and S strains were prepared on solid growth media with an initial C:S ratio of 1:100 . This ratio boosted the competitiveness of the S strain by facilitating spatial exclusion . At the same time , this ratio allowed us to investigate if and how the C strain is able to outcompete the S strain , in case the S strain is initially dominant in cell number . We imaged the communities using a stereoscopic microscope with a zoom function , which enabled us to acquire time-lapse recordings of the competition from the near single-cell level up to mature , macroscopic colonies . We then extracted colony area and colony composition using customized image and data analysis software ( S1 Fig , Methods ) . To assess the influence of cooperation within the C strain described above , as well as stochasticity in toxin production on competition outcome and C strain success , we performed range expansion experiments at four concentrations of the inducing agent MitC ( 0 . 0 , 0 . 005 , 0 . 01 , 0 . 1 μg/ml ) . After 48 h of competition , we observed four qualitatively distinct outcomes based on the relative area occupied by the particular strain: domination by C or S , coexistence , and extinction of both strains ( Fig 1B , S2 Fig and S1–S4 Videos ) . Domination signifies that one strain occupies over 90% of the colony area , coexistence denotes occupancies of between 10% and 90% , and occupation of a total area of less than 106 μm2 constitutes extinction . This was in contrast to competitions performed under well-mixed conditions ( S3 Fig ) , in which toxin release by the C strain lead to a growth arrest of the S strain in all cases . In our experiments , we observed a strong dependence of competition outcome distribution on the inducer concentration ( Fig 1C ) . Without MitC , either S won , or the competition resulted in coexistence . At low inducer concentrations , C won in the majority of cases , while at high inducer concentrations C succumbed , and one observed either domination by S or extinction of both strains ( Fig 1C ) . Hence , domination by C was only observed at intermediate inducer concentrations . Although we could identify clear differences between the outcome distributions , we unexpectedly observed multistability—the presence of multiple competition outcomes under similar initial conditions ( the same inducer concentration ) . To gain a mechanistic understanding of the interaction , we characterized the dynamics of the competition quantitatively . First , we analyzed the impact of stochastic effects in both toxin production and initial positioning on competition outcome ( Fig 2 ) . In a second step , we used experimental time-lapse data ( Fig 3 ) as well as computational modelling ( Fig 4 ) to investigate competition parameters that dictate the deterministic competition dynamics , such as growth rate , division of labor ( toxin producer fraction ) within the C strain , and toxin sensitivity . This enabled us to disentangle the influence of deterministic and stochastic effects and to identify the origin of multistability . Stochastic effects that are mainly important in this C–S competition are the random initial spatial positioning of the C strain within the C–S colony and the stochastic switching from the replicating to toxin-producing phenotype of C cells at low cell numbers ( Fig 2A ) . Our method enabled us to precisely determine the position and number of C cells in initial colonies and subsequently analyze their influence on competition outcome . Surprisingly , we found that the fraction of C cells in the colony after 48 h FC is only weakly correlated to various initial spatial measures , such as initial C cell number NC , 0 , individual C cell distance to the colony center |x→C , 0 , i| , average distance of C cells from the center RC , 0 , and the average distance of C cells from the center of all C cells DC , 0 ( Fig 2A dotted arrow , S4 Fig , Methods , and see S1 Data for correlation table ) . In contrast , the second source of stochasticity , the stochastic toxin dynamics , was observed to strongly influence the final C fraction FC ( Fig 2B ) . For small cell numbers , stochastic switching has a big influence on the toxin production dynamics of individual C cells and can decide the fate of the single S–C communities . Qualitatively , this can be understood when considering different scenarios ( S2 Fig ) : Early toxin production prevented C’s success in case there were too few C cells left to proliferate . Consequently , the S strain dominated or the whole S–C colony went extinct . In case of delayed toxin production , a considerable C strain population had formed , and enough reproducing C cells remained after toxin release to maintain the C strain population . This resulted in C domination or coexistence outcomes . These observations inspired us to take a closer look at the initial phase of colony formation ( phase one: t < = 12 h ) where stochastic effects play an important role due to small cell numbers . We found that we could quantify our qualitative observations by NC , Edge , the number of viable C cell clusters at the colony edge after 12 h . Linear regression showed that the buildup of viable C clusters at the colony edge was significantly influenced by each of the spatial variables NC , 0 , RC , 0 , DC , 0 , and MitC as a proxy for the switching rate into the toxin-producing state ( Fig 2D–2F , Methods , and S1 Data ) . We found that variations in NC , Edge together with the influence of the inducer concentration MitC , could explain the observed multistability ( Fig 2C ) . As we will describe in the subsequent paragraphs , the deterministic dynamics were strongly influenced by the inducer concentration that determined the interaction regime . However , due to the stochastic nature of single-cell dynamics and initial spatial positioning , the formation of viable C cell clusters was subject to noise . Given a favorable deterministic regime ( 0 . 005 , 0 . 01 μg/ml MitC ) , we can distinguish three cases: ( 1 ) NC , Edge ≥ 1 lead to success of the C strain or transient coexistence , ( 2 ) NC , Edge = 0 and there was no viable C cluster remaining lead to domination of the S strain , ( 3 ) NC , Edge = 0 and there was no C cell cluster remaining and the random spatial distribution lead to unlikely but possible complete extinction . In cases of transient coexistence , growing C clusters were not large enough to cover over 90% of the area within the time-frame of our experiments but will take over in the long run . In the uninduced deterministic regime ( 0 . 0 μg/ml MitC ) , C success was prevented due to limited toxin action , and we only observed S success and coexistence with a linear dependence of FC on NC , Edge ( Fig 2C ) . In the highly induced case ( 0 . 1 μg/ml MitC ) , buildup of viable C cell clusters was suppressed ( Fig 2F ) , and depending on the spatial initial position of the S strain , S could survive , or we observed extinction . To formally analyze the effects of the different variables , we employed a linear statistical model incorporating the variables MitC , NC , Edge , NC , 0 , RC , 0 and DC , 0 and their interactions to explain FC ( Methods , S1 Table ) . We found that significant effects stem from MitC , NC , Edge , their interaction , and the interaction term of NC , Edge and NC , 0 . As expected from the previous results , large and medium effect sizes ( based on η2 statistic ) were only attributable to MitC and NC , Edge , respectively . In order to verify our findings with higher significance , we performed simulations ( see below for details ) with 16 random initial conditions IC , repeated 30 times each , over a range of 17 different switching rates sC . Intriguingly , we could observe different outcomes of the competitions under the very same initial conditions ( sC = 0 . 02 , example in S4C Fig ) , underlining the result that the spatial initial conditions do not fully determine the competition outcome . Analogously to our experimental analysis , NC , Edge was determined , and the response variable FC was modelled in a linear statistical model ( Methods , S1 Table ) . Here , all variables and their interaction terms are statistically significant , but only sC and NC , Edge have large η2 effect sizes ( see S1 Table ) . This supports the experimental result that spatial initial conditions act only indirectly via NC , Edge on the final C fraction FC . In order to probe the relation between stochastic and deterministic phases , we performed experiments ( at 0 . 005 μg/ml MitC ) and simulations with increased initial density , as inoculation density is known to influence colony pattern formation [37] . Intuitively , one expects a decrease in multistability , with competition outcome distributions becoming increasingly deterministic . However , our observations indicate a counteracting mechanism for intermediate densities ( S5 Fig ) . For 2- and 4-fold initial density , S success was decreased , but we also observed an increased fraction of coexistence outcome in experiments . These cases of coexistence were believed to be transient , and are most likely explained by an increased S area that established before the toxin stops S’s growth . We speculate that during the 48 h of competition , the C strain did not have enough time to catch up , and the final fraction was below 90% . If we increased the density further , S’s head start might not have been enough to act against the increased probability to have high NC , Edge values , and we observed mostly C success with few coexistence events . In contrast to these ambiguous results for the outcome distributions , we found a decrease in variance of the C fraction dynamics averaged over the time-course of the experiment , with increasing density ( OD 0 . 1: 0 . 2146 , OD 0 . 2: 0 . 1630 , OD 0 . 4: 0 . 1231 , OD 0 . 8: 0 . 1013 ) indicating more reproducible dynamics with increasing initial cell density . Note that the simulation results differed slightly from the experimental observations . First , with increasing initial cell density , spotting of the C-S mixture resulted in an increased coffee stain effect in experiments . Hence , the density is not homogeneously increased as is the case in the simulations . Secondly , in simulations in which viable C clusters were surrounded by dead S cells , preventing spatial expansion and dominance of the C strain , the total colony size remained very small , and the outcome was therefore classified as extinct . Up to now , we have learned that stochastic effects in initial positioning and toxin production dynamics influence the number of viable C cell clusters at the colony edge after 12 h ( NC , Edge ) . Furthermore , we found that NC , Edge together with the inducer concentration MitC determining the interaction dynamics of the second phase strongly predicts the outcome of competitions . However , an exact analysis of the influence of MitC on the deterministic dynamics in the second phase was missing . Therefore , we investigated competition parameters that dictate the deterministic competition using experimental time-lapse data and computational modelling . In the first step , we evaluated the growth rates ( area expansion rates ) for both strains in the absence of the competing strain ( Fig 3B , Methods ) . We found that the growth rates for S and the uninduced C strain were similar to those seen in previous studies [3] , with C growing slightly faster on average . At high inducer levels , C’s growth rate falls considerably , owing to increased toxin release with concomitant lysis [30] . We speculate that the observed increased growth rate of the S strain can be attributed to selection for fast growing S cells at high MitC concentrations . We furthermore determined the growth rates of the entire colonies in competition experiments ( Fig 3C ) . We found that in cases of final S strain dominance growth of the entire mixed colony is enhanced compared to the single S colony growth upon induction with MitC . Hence , competition may promote growth of S due to the selective effect of sublethal colicin concentrations ( S3D Fig ) . In competition experiments that lead to final C strain dominance , the growth rate of the entire mixed colony is reduced compared to the single C strain colony growth rate . Here , although the C strain is dominating in the end , its growth is reduced by initial spatial exclusion by S , which restricts the territory available to the former . The reduced effective growth of C in competition experiments emphasized that growth rate alone cannot explain its dominance at low inducer concentrations . Hence , we assessed the dependence of the second parameter , producer fraction within the C strain population , on MitC concentration using high-resolution microscopy ( Fig 3D , Methods and S6 Fig ) . The dynamics of toxin production have been studied in detail for liquid environments using single-cell fluorescence time-lapse microscopy [30] . Here , the strongest population response is observed 75 min after induction , and it was found that the fraction of cells producing the toxin increases with the external stress level ( MitC ) . While in the absence of external stress , only a basal producer fraction was found with few cells producing the toxin; low external stress resulted in a significant fraction of cells producing and releasing the toxin . In contrast , high MitC concentrations lead to a synchronized response of all cells [30] . We verified , that this correlation between the external stress level and the fraction of toxin-producing cells within the C strain population was also present for the experimental conditions used in this study ( Fig 3D ) . In the absence of MitC , an intrinsically low producer fraction ( 7 . 0 ± 1 . 5% ) was detected . At low MitC concentrations , the mean producer fraction increased to 14 . 5 ± 1 . 9% ( 0 . 005 and 0 . 01 μg/ml MitC combined ) , while the highest inducer concentration leads to a synchronized response ( 57 . 8 ± 3 . 2% ) with a subsequent collective decrease in producer fraction due to cell lysis , in accordance with previous studies [30] . This suggested that the observed shift in competition outcome distribution with varying inducer concentration can be attributed to the change in toxin-producer fraction . To verify that , indeed , increased toxin production was determining the interaction dynamics of the second phase , we repeated our competition experiments with a strain SYFP that is genetically identical to the C strain but lacks the original colicin-producing pColE2-P9 plasmid and hence is not able to produce the toxin ( Methods ) . Lysis in this strain is achieved through the additional lysis gene expression from a reporter plasmid encoding a yellow fluorescent protein ( YFP ) [30] . For this SYFP–S competition , we found mostly dominance of the S strain and coexistence of both strains at all inducer concentrations ( S7A Fig ) . Similar to this all-OFF scenario ( no cell able to produce the toxin ) , also a scenario with all C cells producing the toxin ( all-ON ) reduced C strain success ( S7B Fig ) . At a very high MitC concentration of 0 . 4 μg/ml , all cells produced the toxin simultaneously [30] which lead to the total extinction of both the C and S strain in nearly all competition experiments . This showed that , indeed , the change in producer fraction , and consequently the toxin amount released , was causing the shift in competition outcome distributions , and demonstrated the importance of division of labor for C strain success . To characterize individual interaction dynamics , we extracted the transition time from the relative occupied area curves ( Fig 3E ) . The transition time was defined as the time it takes for C to capture more than half of the colonized area . We observed that average transition times were shorter for the higher inducer concentration and that long-term C dominance was possible only if the transition occurred early . This again emphasized that population fate was mostly determined in the initial phases of colony formation . To corroborate our experimental finding of toxin-producer fraction as the key parameter of the second phase’s interaction dynamics determining C strain success , we developed a theoretical model of the competition and simulated the system with parameters we obtained in our experiments . We used a stochastic lattice-based model that enabled us to explicitly incorporate the stochastic positioning and phenotypic heterogeneity of the C strain ( Methods and Fig 4 ) . Phenotypic heterogeneity in the model is a result of stochastic switching [31] from the reproducing ( normal ) state C to the toxin-producing state Con . Because of cell lysis accompanying toxin release , producing cells can only decay and cannot switch back in the model . Our numerical simulations reproduced the four different outcomes observed in competition experiments ( Fig 4A , S5–S8 Videos ) and demonstrated that dominance of the C strain could be achieved by only varying the switching rate sC , i . e . , the propensity of a reproducer cell to switch to the producing state ( Fig 4C ) . In particular , upon altering sC , which corresponds to varying the inducer concentration , we observed the same qualitative changes in the final C strain fraction and competition outcome as were seen in experiments ( Fig 4D ) for a constant toxin effectivity sS . With low and high producer fractions , we observed no domination of the C strain , whereas the C strain took over for intermediate inducer concentrations . More explicitly , at ~50% toxin-producing C cells , we found mostly dominance of the C strain ( Fig 4C ) but also dominance of the S strain and coexistence of both strains . The question arose if one observed the same competition outcome if the C strain was able to release the toxin while growing , which is the case for other toxin-producing species [25] . In such a scenario , we found that the C strain dominates in nearly all cases ( S7C Fig ) . Hence , the C strain is similarly successful if the toxin can be released without cell lysis by all toxin producers , even if the growth rate and the toxin sensitivity is reduced . This indicates that cell lysis bears a high cost for the toxin producer . The third important interaction parameter , the toxin effectivity sS , represents the total toxin impact and incorporates toxin amount released by the C strain and toxin sensitivity of the susceptible strain S . Because of its complexity , it is hard to exactly determine this parameter experimentally . However , its effect on the dynamics can be explored in the model . We simulated the system for a range of sS and sC values , yielding phase diagrams for each of the four competition outcomes ( Fig 4C , S8A Fig ) . In agreement with our experimental findings , dominance of C was indeed prominent in regions of intermediate producer fractions . Furthermore , this held for a broad range of toxin effectivities sS . Conversely , S dominance was most likely where C’s toxin production strategy failed . Coexistence was mostly found at low toxin producer fractions or low toxin effectivity . Extinction events occurred all over regions of the phase diagram where C prevailed and became more prominent for higher sC values . However , the simple computational model could not reproduce the high incidence of extinction events seen in experiments at high inducer levels . A more detailed model incorporating the synchronicity of the toxin-production response to external stress yielded higher numbers of extinction events ( S8B Fig and S9 and S10 Videos ) , highlighting the fact that population extinction at high producer fractions requires a synchronous response . Taken together , the theoretical model clearly showed that only varying the switching rate and thereby the toxin producer fraction within the C strain population is sufficient to explain why the C strain can be dominant at intermediate inducer concentrations ( = intermediate toxin producer fractions ) . To explicitly test the influence of relative strain growth rate and toxin sensitivity , we repeated the competition of C with three other strains experimentally ( at 0 . 005 μg/ml MitC ) and varied the growth rate and toxin sensitivity parameters in simulations ( Fig 5 and S9C Fig ) . Experimentally , this included altering the growth rate of the competitor and its sensitivity to the toxin ( Methods ) . Boosting the growth rate of S ( SNFP ) increased the effect of its competition strategy , spatial exclusion , and improved the S strain’s competitiveness considerably . Resistance to the toxin ( RRFP ) neutralized the deleterious effect of C on its competitor and prevented dominance of C . Furthermore , increasing the resistant strain’s growth rate ( RNFP ) enabled it to consistently prevail over C . These results ( Fig 5B ) were confirmed by our theoretical simulations , using the appropriate growth rates and toxin sensitivities for the competitor mutants obtained in experiments ( S9A Fig and S2 Table ) . This underlined the ability of our model to accurately predict competition outcome distributions . Exploring the dependence of outcome probability on relative growth rate and toxin sensitivity further ( S9C Fig ) , we found a trade-off between both parameters . The phase diagrams show a dividing line of coexistence along a diagonal axis that represents the growth rate and sensitivity trade-off . For instance , if a strain was sensitive and slow , it succumbed to the toxin producer . Conversely , a strain that was less sensitive did not need a high relative growth rate to thrive in competition . These insights can be used to predict the influence of growth rate changing fluorophore expression ( S9A Fig ) on the deterministic competition dynamics . The costly expression of red fluorescent protein ( RFP ) compared to no fluorescing protein ( NFP ) ( SRFP compared to SNFP ) [38] influenced competition outcome considerably ( Fig 5 , S9A Fig and S2 Table ) in accordance to our theoretical results . Likewise , we expect the relatively small decrease in growth rate of the C strain compared to the corresponding wild-type strain due to fluorophore expression ( Methods ) to only weakly influence the interaction . Taken together , the results show that stochastic and deterministic effects can be disentangled . In the first phase ( t < 12 h ) , at low densities , stochastic effects of toxin production and spatial distribution influence the number of viable C cell clusters at the colony edge after 12 h ( NC , Edge ) . Together with a measure for the division of labor ( MitC in experiments , sC in simulations ) the deterministic dynamics of the second phase can be well predicted . In this second competition phase , the strains simply expand in range afterwards , following deterministic dynamics governed by the effective fitness differences between the two strains [17] .
Our results clearly showed , how stochastic processes influenced the initial phase of colony formation and how effective fitness differences stemming from division of labor within the toxin-producing population dictated the deterministic interaction dynamics of the second phase . We demonstrated that the outcome of competitions fought under the same condition is not fixed [33–35] . Instead , we could observe four different competition outcomes in our experiments . This multistability was a result of the stochastic toxin-production dynamics . This can be understood if we disentangle the dynamics into two distinct phases . In phase one , the stochastic toxin-production dynamics of individual cells at early time points determine the fate of a given S–C colony by affecting the number of viable C cell clusters at the colony edge NC , Edge ( phase one , Fig 2 ) . In the second phase , the stochastic toxin production manifests itself in a constant toxin-producer fraction constituting the division of labor/phenotypic heterogeneity . The exact degree of heterogeneity sets effective fitness differences between both strains and , together with the number of C clusters at the colony edge , it dictates the deterministic macro dynamics and thereby competition outcome distributions ( phase two , Fig 2 ) . As a result , in case of sufficient toxin action ( 0 . 005 and 0 . 01 μg/ml MitC ) , the effective fitness differences were so large that once a viable toxin-producer cluster had formed at the colony edge ( NC , Edge > 0 ) , it took over the colony in the long run . Hence , the coexistence observed at 48 h was only transient , S never won as we observed NC , Edge > 0 , and the dynamics were fully determined by the initial phase . However , in the case of limited toxin action ( 0 . 0 μg/ml MitC ) , the effective fitness difference was not as clear . The observed coexistence was persistent as the effective fitness differences were not large enough to let the C strain take over during the time-course of our experiments ( S2 Video ) and after prolonged competition ( >48 h ) . Furthermore , in the absence of MitC , we observed few cases in which the S strain won , although we had observed NC , Edge > 0 . These cases were in accordance to previous studies that showed how inevitable random fluctuations due to genetic drift can cut off access to the exterior for similarly expanding colony domains [39] . Therefore , we conclude that for the uninduced case , the initial phase does not fully determine the outcome but predicts the final fraction of the C strain well ( see also Fig 2C ) . By taking advantage of the multiscale functionality of the presented experimental method , we could investigate both the parameters for the deterministic dynamics as well as the influence of stochasticity in toxin production and initial conditions . In addition , we are confident that this method can be used to investigate a variety of interaction types in detail , thereby advancing our fundamental understanding of bacterial interactions . The computational model used in this study nicely complemented our experiments . When used with experimentally observed parameters , it was able to predict competition outcome distributions in good accordance with the experimental observations . Variations of toxin-producer fraction , growth rate , and the switch from sensitive to resistant strains showed outcome distributions similar to experiments . Furthermore , the model could be exploited to investigate conditions that we could not create experimentally , such as varying toxin effectivity sS , replication of simulations with the very same initial conditions , and simulations of competitions with a non-lysing C strain . Our conclusions about stochasticity as the origin of multistability are robust with respect to a wide range of different parameters influencing the deterministic interaction dynamics , such as toxin effectivity , growth rate , or exact switching rate . Obviously , the initial influence of stochasticity decreases with increasing density . The observed trade-off between the beneficial effect of antibiotic production and the effective fitness disadvantage the producer strain exerts on itself is in accordance with Geradin et al . [36] . In this study , the fitness disadvantage was further increased by the presence of resistant bacteria . In summary , our results show that stochasticity in the initial phase of colony formation can be the crucial factor that determines the ensuing population dynamics and consequently controls the final composition of the community . This is in accordance with a recent study of the effects of stochastic assembly on host-associated microbial communities [40] and underlines how random events at the single-cell level can influence the fate of microbial communities in the long run . Besides toxin production , there are many phenotypically heterogeneous traits in which decisions between one or another phenotype happen stochastically . Therefore , our findings are relevant for a broad range of microbial communities , in which the random choice between phenotypes during the initial phases of colony formation effectively fixes the community's composition and ultimate fate .
The E . coli colicin-producing strain EMO3-C [30] ( just called C in the main text ) is a derivative of E2C-BZB1011 ( COriginal ) [1] and carries the pMO3 plasmid encoding the Yellow Fluorescent Protein ( YFP ) [30] . On this plasmid , the yfp gene is under the control of the same promoter as the ColicinE2 operon . Hence , YFP expression correlates directly with ColicinE2 expression and serves as a visual marker for toxin-producing cells . Moreover , in Mader et al . [30] , we showed that every cell that produces YFP also lyses . As cell lysis is coupled to toxin expression in this strain , YFP expression can be used as a proxy to assess the number of cells that produce the colicin . The competitor strains SRFP , SNFP , RRFP , and RNFP [3] are derivatives of BZB1011 ( S ) and E2R-BZB1011 ( R ) , respectively [1] , and carry plasmids on which the RFP mCherry or NFP is encoded . RFP expression is under the control of the pBAD promoter , which provides for continuous expression in the presence of arabinose . In contrast , YFP in EMO3-C is only expressed when a cell produces colicin . The SRFP and SNFP strains are sensitive to the toxin , due to the lack of the original colicinE2-producing pColE2-P9 plasmid that also encodes an immunity protein protecting the otherwise genetically identical C strain from ColicinE2-induced DNA damage . The strains RRFP , and RNFP are resistant towards the toxin due to impaired colicin uptake [41] . Please note that the expression of the RFP mCherry impairs an additional metabolic cost and hence reduces growth rates relative to the NFP strains [3] ( S7A Fig ) . In contrast , the expression of YFP ( in EMO3-C ) does not considerably affect the growth rate of this strain ( growth rate in liquid media [see below for calculation] 0 . 735 ± 0 . 47 1/h for E2C-BZB1011 ( COriginal ) and 0 . 728 ± 0 . 026 1/h for EMO3-C ) . The respective fluorescence proteins expressed in the S strain ( SRFP ) and the C strain ( YFP ) were chosen to ensure a similar growth rate of these strains in the absence of an external stressor ( Fig 3B ) . Strain SYFP [30] is a derivative of BZB1011 ( S ) [1] . It lacks the pColE2-P9 plasmid but carries the plasmid pMO2 [30] . Similar to pMO3 , this plasmid encodes YFP , with the yfp gene being under the control of the same promoter as the ColicinE2 operon . Hence , YFP expression in SYFP correlates directly with ColicinE2 expression , although this strain is unable to produce the toxin due to a lack of original pColE2-P9 plasmid . In contrast to pMO3 , pMO2 additionally encodes the gene required for toxin release through cell lysis that is again under the control of the ColicinE2 operon promoter [30] . Hence , this strain is able to lyse upon induction of the ColicinE2 operon promoter , ensuring similar lysis gene expression dynamics as present in EMO3-C . As this strain is not expressing the colicin and has therefore a reduced metabolic cost , it grew slightly faster than EMO3-C , with a growth rate in liquid media of 0 . 972 ± 0 . 029 1/h . Experiments were performed on solid M63 growth medium [42] ( 1 . 5% agar ) containing 100 μg/ml ampicillin and 0 . 2% arabinose . Furthermore , the medium was supplemented with the SOS agent MitC at different concentrations ( 0 . 0 , 0 . 005 , 0 . 01 , 0 . 1 μg/ml ) . The growth medium was prepared in rectangular 127 . 8 x 85 . 5 mm Greiner OneWell plates ( Catalogue Number 670190 ) . Prior to experiments , bacterial cultures were grown separately in liquid M63 medium , with 100 μg/ml ampicillin and 0 . 2% arabinose shaken at 300 rpm at 37°C . Overnight cultures were diluted to OD600 = 0 . 1 and grown to OD600 = 0 . 2 . The EMO3-C culture was then centrifuged through a 100-kD filter to remove colicin molecules ( 62 kD ) already present in the starter culture . Prior to transfer onto agar , starter cultures were diluted to OD600 = 0 . 1 . For competition experiments , EMO3-C cells were mixed with a competitor strain X ( SRFP , SNFP , RRFP , or RNFP ) at a C:X ratio of 1:100 shortly before transfer . This C:X ratio was chosen on the basis of earlier results demonstrating that strain coexistence was only possible at reduced C strain ratios [3] . For control experiments , monoclonal cultures were transferred without mixing . For experiments with varying initial density , starter cultures were concentrated via centrifugation , diluted to chosen OD600 ( 0 . 2 , 0 . 4 , 0 . 8 ) and processed further as described above . To verify known heterogeneities with regard to toxin production by the C strain in liquid environments [30] for the conditions used in our range-expansion experiments , we determined the YFP fluorescence intensity of the C strain on agar plates using a high-resolution setup consisting of a Nikon Eclipse 90i with a Nikon Intensilight light source and a 50× objective operated by Nikon NIS Elements Advanced Research software ( version 3 . 22 . 01 ) . The system was maintained at 37°C by a custom-built heat box . Starter cultures of the C strain were grown from OD600 = 0 . 05 to 0 . 2 , diluted to OD600 = 0 . 01 , and 1-μl droplets were pipetted onto agar plates prepared as described above . The plates were then incubated at 37°C , and images were recorded every 60 min for 5 h in bright-field and fluorescence channels . Applying a fluorescence threshold of 200 fluorescence units , cells were classified into ON and OFF phenotypes ( S6 Fig and Fig 3D ) . Note that these experiments tend to underestimate the producer fraction because already lysed cells were difficult to distinguish from living cells in the OFF state . A minimum of 169 cells ( for each data point of Fig 3D ) for each inducer concentration was investigated . Bacterial population growth , the influence of colicinE2 on growth of S , as well as the C–S interaction under well-mixed conditions were analyzed with a plate reader ( POLARstar OPTIMA , BMG Labtech ) . Overnight cultures were generated as described above , diluted to OD600 = 0 . 1 and grown to OD600 = 0 . 2 , and for measurements , the cultures were diluted again to OD600 = 0 . 1 . Bacterial population growth and interaction was followed for 18 h , while the cultures were maintained at constant shaking at 37°C . Optical density ( 600 nm ) and respective fluorescence channels ( YFP and RFP ) were measured every 15 min . After blank correction , the population growth rate ( GR ) was obtained as follows: Population growth curves as represented by OD600 were fitted by using the linear fit function ( fL = a +b * x , with a the y-intercept and b the slope of the function ) of the IGOR PRO 6 . 36 software to fit the natural logarithm of the population growth curves in the exponential growth phase . The population growth rate ( GR ) was then calculated as: GR=bln ( 2 ) for uninduced and 0 . 005 μg/ml MitC induced samples and averaged . To test the influence of colicinE2 on the growth of S , the colicin was extracted from the supernatant of a C strain culture induced with 0 . 7 μg/ml MitC , using 10 K centrifugal filter units ( Amicon ) . The obtained total protein concentration was 1 . 14 mg/ml . This protein solution was then added to S strain cultures at the concentrations/dilutions indicated in S3D Fig For C–S interaction experiments under well-mixed conditions , the C and S strain were applied at the 1:100 ratio in accordance to C–S interactions on the solid agar surface . Competition experiments were performed over a period of 48 h using the following multiscale set-up , which allows us to monitor up to 77 competition experiments in parallel . A Nikon SMZ 25 stereoscopic microscope with a custom-built mount was assembled on a Newport Isostation table . Image acquisition was performed by a Nikon Qi1 CCD camera controlled by a Nikon DS-U3 camera control unit . A Märzhäuser SCAN 130 x 85 scanning stage controlled via a Märzhäuser TANGO 2 facilitated parallel investigation of multiple communities on the experimental plate . Bright-field illumination and fluorescence excitation were provided by a Lumencor Sola SE II LED lamp . Nikon P2-EFL GFP-B and P2-EFL RFP-L filter blocks were used for fluorescence excitation and emission filtering , and a customized OG-570 long pass filter reduced phototoxicity from bright-field illumination . Microscope , camera , scanning stage , and LED lamp were operated via Nikon NIS-Elements AR 4 . 30 . 01 64-bit software with the requisite plug-ins . A gas incubation and heating system for multi-well plates ( Ibidi ) ensured constant environmental conditions ( 37°C and 80% humidity ) . Aliquots of the inoculum culture were deposited on the experimental plate by a Labcyte Echo 550 Liquid Handler using acoustic droplet ejection [43] . Minimal transferrable volumes of 2 . 5 nl result in initial bacterial communities of approximately 150 single bacterial cells that are distributed within circular areas of approximately 450 μm in diameter ( S1 Fig ) . ColicinE2 can easily diffuse through the agar plate and an effective toxin range of 100–400 μm [27 , 44] was described . To prevent touching of neighboring S–C colonies during competition ( maximum competition colony size is 4 . 9 mm [diameter] ) and to reduce colicin interactions in between spots to a minimum , the distance between the center of two spots was chosen to be 9 mm . Hence , a minimum of 4 mm unoccupied territory is present at the end of a competition experiment in between two neighboring colonies . Experiments have been repeated 2–3 times . Only communities containing C cells in the initial colonies were analyzed , resulting in a total of 75 , 87 , 85 , and 87 ( 0 . 0 , 0 . 005 , 0 . 01 , 0 . 1 μg/ml MitC ) competitions for S–C ( Fig 1 ) , 87 , 108 , 108 , 128 competitions for S–X ( Fig 5 ) , and 106 , 83 , 107 , 108 competitions for S–SYFP ( S9 Fig ) . See S1 Data , sheet “Number of Observations” for details . The total time course of each experiment ( 48 h ) was divided into four distinct zoom levels to accommodate bacterial population growth . Settings for lamp intensity , detection gain , or exposure time must be adjusted for each zoom level and are listed in S3 Table . Before the start of the actual experiment , images were taken at every zoom level for background correction purposes at every spot . Acquired images had a resolution of 1280 x 1024 pixels and 12-bit depth . Please note that the fluorescence data obtained correspond to cells present in the top layer of the colony with contributions of the bacterial layers ( up to 200 ) beneath the top layer . Hence , the fluorescence signal represents an average value over the entire colony depth and is only used to monitor the dominant strain in a particular colony section . We developed our image processing routines with Mathworks MATLAB software ( version 2013b ) . Image processing included image segmentation to isolate the bacterial population from the background and classification to distinguish between different strains based on labelling via fluorescent protein expression ( S1C Fig ) . The use of four different zoom levels made it necessary to use four distinct parameter sets for image segmentation , in order to account for varying object dimensions and image characteristics . Additional complexity arose from the different fluorescence intensities of the fluorescent proteins expressed by the strains used . Consequently , the requirements on the fluorescence analysis varied depending on the strains present . Therefore , parametric fine-tuning was performed to match the outcome of the experiment . Details of the image processing routine can be found in the S1 Text and in S3 Table . Recorded data curves resulting from image analysis were screened for obvious mis-segmentations and mis-classifications . Identified points were replaced by NaN values to correct the curves . Examples for replaced data points are extinction scenarios , in which the image recognition algorithm could not handle empty images . Finalized data curves are exported and further processed with the statistical programming language R ( Version 3 . 2 . 3 ) , different R packages , and Matlab ( Version 2013b ) software . Growth rates of the entire colonies ( single strain and mixed S–C colonies ) were obtained in the linear area growth regime by linear fitting ( S9B Fig ) . All competition experiments were screened for presence of C cells in the initial colony . Colonies without C cells were not considered for further analysis . Classification of C cells happened via the presence of a YFP signal , absence of a RFP signal in a cell , and/or were deduced from the effect of secreted toxin at later time points . Then , the positions of each C cell and the colony center were determined . To quantify the initial spatial distribution of C cells , three variables were used . The number of initial C cells NC , 0 represents the sum of classified C cells at the beginning of the experiment . From the position of initial C cells x→C , 0 , i with respect to the colony center , the average radial distance of these cells from the center was calculated by the center of mass formula RC , 0=|x→C , 0|=|1NC , 0∑i=1NC , 0x→C , 0 , i| . The average distance of C cells from their center of mass was calculated by DC , 0=|1NC , 0∑i=1NC , 0 ( x→C , 0 , i−x→C , 0 ) | as a measure for the spread of C cells over the colony . In a second step , it was checked whether initial C cells grow into viable clusters that reach the colony edge during the first phase of competition or die early due to toxin production and concomitant lysis . The number of these viable C cell clusters at the colony edge after 12 h was aggregated in the variable NC , Edge . Similarly , simulations with 16 different fixed initial conditions IC were repeated 30 times each , for the whole range of sC values ( sC = 0 . 02 example in S8C Fig ) . Again , the number of viable C clusters at the colony boundary at the end of the initial phase NC , Edge was determined . For the experimental data , the influence of NC , 0 , RC , 0 , DC , 0 , and MitC on NC , Edge was modelled after standardization using the following linear model . For both experiment and simulation , the influence of the different factors MitC , NC , Edge , NC , 0 , DC , 0 , and RC , 0 on final C fraction FC was analyzed using a linear model that also incorporated interactions between these factors . For the experimental analysis , MitC was treated as a categorical variable to account for the qualitative difference between different induction regimes . The other variables were treated as continuous variables . The model had the explicit form: FC=β0+MitC ( β1+β6NC , Edge+β7NC , 0+β8DC , 0+β9RC , 0 ) +NC , Edge ( β2+β10NC , 0+β11DC , 0+β12RC , 0 ) +NC , 0 ( β3+β13DC , 0+β14RC , 0 ) +DC , 0 ( β4+β15RC , 0 ) +β5RC , 0 Analysis of the simulation was performed analogously using the three variables sC , NC , Edge , and IC . Here all three variables were treated as categorical variables . The linear model had the explicit form: FC=β0+sC ( β1+β4NC , Edge+β5IC ) +NC , Edge ( β2+β6IC ) +β3IC Significance and effect sizes of the factors were analyzed using the Anova and EtaSq functions of the statistical programming language R ( Version 3 . 2 . 3 ) that rely on the type I ( sequential ) sum of squares , where the order of terms in the model matters . Terms in the model are ordered according to the experimental design: MitC/sC is externally tuned , NC , Edge is found to be very important , exact spatial details/IC are of minor importance . A stochastic lattice-based computational model was used to simulate competition between the C strain and a competitor strain X ( X being SRFP , SNFP , RRFP , or RNFP ) . The model incorporated spatial coarse-graining steps during the simulation in analogy to the zooming used in the experimental procedure . Once the simulated expanding colony reached the boundaries of the 250 x 250 pixel lattice , the current colony was coarse-grained by a factor of five , resulting in a coarser 50 x 50 lattice that was positioned in the middle of a new 250 x 250 lattice . Reaction rates were adjusted accordingly . Zooming did not only accommodate the microscopic details of the initial conditions but also permitted simulations to be completed in feasible computational times . Initial communities used for simulations were created in accordance with experimental conditions , having random spatial distributions of C and X cells in an approximate 1:100 ( C:X ) ratio within a circular field approximately 450 μm in diameter , with at least one initial C cell . Initial colony density was chosen in accordance with experimental conditions ( S4 Table ) . Five different species of agents were used in this model: viable C and X cells , colicin-producing Con cells , growth-inhibited Xstop cells , and unoccupied agar sites A . Reactions were modelled using a Moore neighborhood ( eight nearest neighbors ) , in which the rates for diagonal growth were scaled by a factor of 1/2 . Possible reactions comprised reproduction of viable C and X cells , C cells switching to a producing Con state , subsequent lysis of the Con cell with concomitant colicin release , and transition to a growth-inhibited Xstop state by S cells in response to the action of colicin ( Fig 4B ) . As soon as colicin was released by a lysing Con cell , an exponential colicin profile was assumed to originate from this position , as performed in previous studies [3] . Rates for these reactions are given in S2 Table . Growth rates , r , were obtained by fitting simulated single-strain curves to experimental control data for the S strain without induction by MitC and assuming a simple linear relationship between area growth and microscopic growth rates [45] ( S2 Table , S9B Fig ) . The lysis rate of Con cells adopted here was suggested by earlier experimental findings [30] . The C strain’s switching rate to the producing phenotype sC was chosen such that it covered a broad range of producer fraction values ( S4 Table ) . Assuming a steady state for the toxin producer dynamics ∂tCon=sC⋅C−dCon⋅Con , the relationship between switching rate sC , the lysis rate of toxin producers dCon , and the number of reproducing C and toxin-producing Con cells can be approximated by ConC=sCdCon . The remaining free parameter , toxin effectivity sX = σx ⋅ ntox , is composed of two terms , toxin sensitivity of the competitor strain σx and toxin amount factor ntox representing the amount of toxin produced by Con . Because the toxin effectivity parameter was hard to determine experimentally , the system was simulated for a range of values , testing for robustness and dependence on this parameter ( Fig 4C , S8A Fig , S4 Table ) . A toxin sensitivity of σx = 1 , 500 and toxin amount factor ntox = 1 . 0 were chosen as “standard conditions . ” For simulations of the impact of growth rate variation , a fixed switching rate ( sC = 0 . 015 ) and fixed C strain growth rate were used in combination with a range of toxin sensitivities and relative competitor growth rates and a fixed toxin amount factor ntox = 1 . 0 ( S4 Table ) . For every sX , sC or rX , sC pair simulations were repeated 48 times . Besides the purely stochastic simulations , we performed simulations in which synchronous toxin production and release was explicitly implemented . In these simulations , the C strain grew without switching to the producing state for 50 min and did not replicate for a further 50 min , at which point all cells produced and released their toxin simultaneously ( synchronicity scenario I , S8B Fig ) . Additionally , a second synchronicity scenario was implemented where C cells reproduced for 100 min without switching to the producing state , prior to producing and releasing the toxin collectively ( synchronicity scenario II , S8B Fig ) .
|
Competition is the dominant interaction type between species of bacteria . Bacterial toxin-mediated competition is often accompanied by a division of labor between toxin-producing cells and reproducers within a species . In populations with large cell numbers , the stochastic dynamics that determine if an individual cell switches into the toxin-producing state are often unnoticed . Consequently , we know little about how stochastic effects influence bacterial competition . Here , combining experimental and theoretical efforts , we study the competition of a toxin-producing strain with a toxin-sensitive strain . By correlating the initial conditions—at near single-cell level—to the macroscopic competition outcome , we investigate both the importance of the division of labor as well as the influence of the stochastic toxin production dynamics on competition outcome . Our results highlight the impact of the initial phase of competition as a major determinant for the success of the toxin-producing population .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"toxins",
"pathology",
"and",
"laboratory",
"medicine",
"toxicology",
"toxic",
"agents",
"simulation",
"and",
"modeling",
"luminescent",
"proteins",
"phase",
"diagrams",
"yellow",
"fluorescent",
"protein",
"bioassays",
"and",
"physiological",
"analysis",
"color",
"codes",
"crystallographic",
"techniques",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"proteins",
"biochemistry",
"predictive",
"toxicology",
"data",
"visualization",
"biology",
"and",
"life",
"sciences",
"phase",
"determination",
"fluorescence",
"competition"
] |
2017
|
Effects of stochasticity and division of labor in toxin production on two-strain bacterial competition in Escherichia coli
|
Obesity is a growing epidemic characterized by excess fat storage in adipocytes . Although lipoprotein receptors play important roles in lipid uptake , their role in controlling food intake and obesity is not known . Here we show that the lipoprotein receptor LRP1 regulates leptin signaling and energy homeostasis . Conditional deletion of the Lrp1 gene in the brain resulted in an obese phenotype characterized by increased food intake , decreased energy consumption , and decreased leptin signaling . LRP1 directly binds to leptin and the leptin receptor complex and is required for leptin receptor phosphorylation and Stat3 activation . We further showed that deletion of the Lrp1 gene specifically in the hypothalamus by Cre lentivirus injection is sufficient to trigger accelerated weight gain . Together , our results demonstrate that the lipoprotein receptor LRP1 , which is critical in lipid metabolism , also regulates food intake and energy homeostasis in the adult central nervous system .
The low-density lipoprotein receptor-related protein 1 ( LRP1 ) is a large cell surface receptor ubiquitously expressed in a variety of organs including adipose tissue , liver , and brain [1] . Previous tissue-specific knockout studies showed that hepatic LRP1 mediates the metabolism of apolipoprotein E ( apoE ) -rich chylomicron remnants [2] and that adipocyte LRP1 modulates postprandial lipid transport and glucose homeostasis [3] . Furthermore , LRP1 has a pivotal role in preventing atherosclerosis by restricting smooth muscle cell proliferation and protecting vascular wall integrity [4] . In the central nervous system ( CNS ) , LRP1 is highly expressed in neurons and plays critical roles in lipoprotein metabolism , neurotransmission , synaptic plasticity , cell survival , and clearance of the amyloid-β ( Aβ ) peptide , critical in the pathogenesis of Alzheimer's disease ( AD ) [5]–[7] .
To investigate the roles of LRP1 in the adult CNS , we generated conditional Lrp1 forebrain knockout mice ( LRP1-KO ) by crossing Lrp1 floxP mice [2] with αCamKII-Cre mice [8] . Because αCamKII-Cre is only expressed in neurons of the adult brain [8] , the essential function of LRP1 during embryonic development [9] is preserved . Lrp1 deletion in different brain regions was assessed by comparing LRP1 protein expression levels between LRP1-KO ( Lrp1flox+/+/Cre+/− , LRP1 knockout ) and WT ( Lrp1flox+/+/Cre−/− , Lrp1 floxP littermate control ) mice at 3 , 6 , 9 , and 12 mo of age . LRP1 expression was not significantly decreased in the LRP1-KO mice at 3 mo of age . However , from 6 to 12 mo of age , LRP1 expression was decreased by ∼75% in the cortex , hypothalamus , and hippocampus of LRP1-KO mice ( Figure S1A–S1C ) . The residual LRP1 observed in these regions likely represents LRP1 expressed in glial cells [10] . Inactivation of the Lrp1 gene was specific to the forebrain , as no changes in LRP1 expression were detected in the cerebellum or in peripheral tissues ( Figure S1D and S1E ) . LRP1-KO mice were apparently indistinguishable from control littermates during the first 6 mo of life but showed significantly accelerated body weight gain starting at 7 mo of age ( Figure 1A and S2A ) . This increased weight gain in LRP1-KO mice correlated closely with the observed decrease in LRP1 expression in the CNS . On a normal chow diet , LRP1-KO mice became obese at about 12 mo of age ( Figure 1B ) and had approximately 2-fold increased body fat content compared to their littermate controls ( Figure 1C ) . No changes in snout-anus length between LRP1-KO mice and WT mice were observed ( Figure S2B ) . LRP1-KO mice ate significantly more than controls ( Figure 1D ) and had significantly decreased energy expenditure , as revealed by decreased O2 consumption and CO2 production ( Figure 1E and 1F ) . Together , these results indicated that LRP1 expression in the brain controls body weight and adiposity by regulating food intake and energy expenditure . Obesity in LRP1-KO mice was associated with hyperlipidemia and insulin resistance . LRP1-KO mice showed a 3-fold increase in circulating free fatty acids ( FFA ) and ∼50% increase in circulating triglycerides ( TG ) ( Figure 1G and 1H ) . Plasma cholesterol levels were not significantly altered in LRP1-KO mice ( Figure 1I ) . At about 12 mo of age , LRP1-KO mice showed approximately 2-fold increase in plasma insulin levels ( Figure 1J ) . While fasting blood glucose levels appeared to be normal ( unpublished data ) , LRP1-KO mice had significantly decreased tolerance to exogenous glucose when assessed by intraperitoneal glucose tolerance test ( GTT ) . A marked increase in both magnitude and duration of blood glucose in response to glucose injection was observed in LRP1-KO mice ( Figure 1K ) . LRP1-KO mice also failed to reduce blood glucose levels after insulin injection ( insulin tolerance test , ITT ) ( Figure 1L ) . Thus , the obese phenotype in LRP1-KO mice was characterized by hyperlipidemia and insulin resistance , resembling the metabolic syndrome pathology in humans . Increase in fat storage observed in LRP1-KO mice was accompanied by approximately 5-fold increase in plasma leptin concentrations ( Figure 1M ) . Because leptin secreted by adipose tissue plays a major role in body weight regulation [11]–[13] , we hypothesized an existing leptin-resistant condition in LRP1-KO mice . To directly assess a role for LRP1 in leptin signaling we measured the levels of phospho-Stat3 ( P-Stat3 ) [14] in the hypothalamus of LRP1-KO and WT mice . A significant decrease in the levels of P-Stat3 was observed in LRP1-KO mice ( Figure 2A , 2B and S2C , S2D ) . In addition , deletion of LRP1 in the adult brain led to leptin insensitivity as shown by impaired leptin-stimulated phosphorylation of hypothalamic Stat3 ( Figure 2C and 2D ) . Central leptin sensitivity was also evaluated by introcerebroventricular ( ICV ) leptin administration . Chronic ICV infusion of leptin caused a rapid reduction in body weight ( Figure 2E ) and food intake ( Figure 2F ) in WT control mice but had minimal effects on LRP1-KO mice ( Figure 2E and 2F ) , suggesting that LRP1 deletion led to leptin resistance . In the arcuate nucleus of the hypothalamus , leptin regulates energy imbalance by inhibiting expression of orexigenic neuropeptide Y ( NPY ) and agouti-related protein ( AgRP ) [13]–[17] . We found that deletion of LRP1 markedly increased NPY and AgRP mRNA levels in the hypothalamus ( Figure 2G ) . Further , using immunofluorescence staining , we demonstrated that LRP1 colocalized with AgRP but not with POMC ( Figure S3 ) , suggesting that LRP1 may specifically regulate leptin signaling in AgRP neurons . These results are consistent with compromised leptin signaling in the hypothalamus of LRP1-KO mice and further support leptin signaling as the potential mechanism by which LRP1 regulates food intake and energy expenditure in the adult mice . To further elucidate the molecular and cellular mechanisms underlying LRP1 regulation of leptin signaling we used GT1-7 cells , a neuronal cell line derived from the mouse hypothalamus [18] . LRP1 knockdown in GT1-7 cells by LRP1-specific siRNA significantly decreased the phosphorylation of both Stat3 and leptin receptor ObR ( Figure 2H , 2I and S4A ) . While treatment of control GT1-7 cells with leptin dramatically increased the phosphorylation of both Stat3 and ObR , it only had lesser effects on LRP1 knockdown cells ( Figure 2H , 2I and S5 ) . In addition , we found that LRP1 knockdown significantly reduced the phosphorylation of the extracellular signal-regulated kinase ( ERK ) , which is also downstream to the leptin receptor signaling ( Figure S4B and S4C ) [19] . Previous studies have shown that the leptin receptor mediates Stat3 phosphorylation via janus kinase 2 ( JAK2 ) activation [13]–[15] . Using co-immunoprecipitation , we confirmed that LRP1 knockdown in GT1-7 cells decreased the interaction between ObR and JAK2 ( Figure 2J , S6A , S6B , S6D , and S7A ) . We also found that LRP1 knockdown decreased leptin-mediated phosphorylation of JAK2 ( Figure S7B and S7C ) . Because LRP1 had been previously shown to serve as co-receptor to PDGF signaling via the PDGF receptor [4] , we next examined the potential for a direct association between LRP1 and the leptin receptor . Using GT1-7 cellular extracts , we found that the LRP1 antibody , but not a control antibody , co-immunoprecipitated the leptin receptor ObR in the presence of leptin but not in the absence of leptin ( Figure 2K , S6C , and S6E ) . To further confirm direct binding of LRP1 to the leptin/leptin receptor complex , we followed binding of 125I-leptin to GT1-7 cells with chemical cross-linking [20] . Under these conditions , radiolabeled leptin migrated as a high molecular weight complex on SDS-PAGE that was immunoreactive with leptin , leptin receptor , and LRP1 antibodies , but not with a control antibody or an antibody to the LDL receptor ( LDLR ) ( Figure 2L , S6F , and S8 ) . The specificity of each antibody in these co-immunoprecipitation experiments was confirmed by antigen blocking . Together , these results suggest that LRP1 regulates leptin signaling by forming a complex with leptin and the leptin receptor . The hypothalamus is central to the control of food intake and energy expenditure; therefore LRP1 deletion in the hypothalamus of LRP1-KO mice is likely the major contributor to the obese phenotype . To further test this hypothesis , we deleted the Lrp1 gene specifically in the hypothalamus using lentivirus delivery technology ( Figure 3A ) . Direct injection of Cre lentivirus into the arcuate nucleus of the hypothalamus of Lrp1 floxp mice resulted in significant neuronal LRP1 deletion , as measured by immunofluorescence and Western blotting ( Figure 3B–3E and S9 ) and reduction of leptin signaling , as measured by reduced levels of P-Stat3 ( Figure 3D and 3E ) . Strikingly , Cre lentivirus-injected Lrp1 floxp mice also showed significantly greater body weight gain , increased food intake , and higher fat content compared to control GFP lentivirus-injected mice ( Figure 3F–3H ) . Plasma levels of leptin ( Figure 3I ) were significantly increased in Cre lentivirus-injected mice . In addition , Cre lentivirus-injected mice had markedly increased expression of NPY and AgRP ( Figure 3J ) . Cre lentivirus-injected mice also showed leptin insensitivity as detected by impaired leptin-stimulated phosphorylation of hypothalamic Stat3 ( Figure S10A and S10B ) . As controls , Cre lentivirus injection into either the cortical region of the Lrp1 floxp mice or the arcuate nucleus of the hypothalamus in wild-type mice ( C57BL/6 ) did not result in accelerated weight gain ( Figure 3K and 3L ) . Having demonstrated that deletion of LRP1 led to obesity and decreased leptin signaling , we were prompted to evaluate whether overexpression of LRP1 could rescue the obese phenotype . Co-expression of mLRP2 , a functional LRP1 minireceptor [21] , reduced the body weight gain and ameliorated the hyperphagia associated with hypothalamic cre overexpression in Lrp1 floxp mice ( Figure 4A and 4B ) . mLRP2 also partially restored leptin signaling ( Figure 4C and 4D ) . Finally , we examine if LRP1 was also involved in the obese phenotype of ob/ob mice . LRP1 protein expression levels were significantly decreased in the hypothalamus of ob/ob mice when compared to wild type mice ( Figure S11A and S11B ) . These results strongly support a critical role for LRP1 in leptin signaling in the hypothalamus . Leptin binding to the long form of the leptin receptor ( ObRb ) stimulates the tyrosine kinase JAK2 to phosphorylate Stat3 at tyrosine residues . P-Stat3 dimers subsequently enter the nucleus and regulate transcription of target genes , such as AgRP and POMC [13]–[15] . We found that LRP1 directly binds to leptin and the leptin receptor complex and is required for leptin receptor phosphorylation and Stat3 activation . Further , LRP1 knockdown in GT1-7 cells decreased the interaction between leptin receptor and JAK2 , which reduced the leptin-mediated phosphorylation of JAK2 . Interestingly , LRP1 is expressed in AgRP but not in POMC neurons , and disruption of LRP1 significantly increases the expression of AgRP and NPY , indicating that LRP1 deletion impairs lepitn signaling in AgRP neurons . Taken together , our results suggest that LRP1 modulates leptin signaling likely via regulation of JAK2 activation in AgRP neurons . It has been reported that leptin receptors are also expressed in various cortical regions and hippocampus [22]–[23] , which are associated with learning and memory . In leptin receptor-deficient mice , hippocampal CA1 region exhibits impaired long-term potentiation ( LTP ) and long-term depression ( LTD ) [24] , which are widely considered major cellular mechanisms underlying learning and memory . Further , leptin enhances NMDA receptor function and modulates hippocampal synaptic plasticity [25] . LRP1 is also widely expressed in the cortex and hippocampus and play essential roles in neurotransmission , and synaptic plasticity [5]–[7] . LRP1 deletion in the cortex and hippocampus region may also decrease leptin signaling and further affect synaptic function , learning , and memory . Neuronal inactivation of LRP1 achieved by expression of the Cre transgene under control of the synapsin I promoter [26] has been previously reported [27] . Unlike the CamKII-Cre model , in which Lrp1 disruption was restricted to the forebrain and hypothalamic areas , LRP1 deletion in the Syn-Cre model was detectable much earlier and throughout the entire CNS , first manifesting by impairment of motor function and systemic tremors about 3 wk after birth . SynI-Cre/Lrp1 mutant mice were also hyperphagic , but in contrast to the CamKII-Cre/Lrp1 model , they were hyperkatabolic , and the majority of Syn-Cre/Lrp1 mice died prematurely between 6 and 9 mo of age . This severe and earlier phenotype emphasized a critical role for CNS LRP1 in motor circuitry and muscle control but effectively occluded another pivotal role for LRP1 in the CNS , i . e . , the regulation of leptin signaling and metabolic homeostasis described here . In summary , we have uncovered a new molecular pathway that regulates leptin signaling and energy homeostasis . Our results demonstrate that LRP1 , a lipoprotein receptor that plays important roles in lipid metabolism , also regulates food intake and body energy homeostasis in the CNS . Because ICV administration of apoE was recently shown to suppress food intake in the rat [28] , it is plausible to speculate that brain lipoproteins may also play a role in regulating food intake , likely via LRP1 . Interestingly , we found that LRP1 knockout resulted in increased apoE levels in the hypothalamus ( Figure S12 ) , consistent with the fact that apoE metabolism is impaired in the LRP1 deficient mice . Therefore , LRP1 provides a critical link between peripheral and central energy metabolism and could serve as a novel therapeutic target to lessen obesity in human patients .
All tissue culture media and serum were from Sigma . Anti-leptin receptor antibody was purchased from Santa Cruz , anti-JAK2 antibody from Invitrogen , anti-phosphotyrosine antibody from Upstate , anti-P-Stat3 and anti-Stat3 antibodies from Cell Signaling , and anti-actin antibody from Sigma . In-house anti-LRP1 antibody has been described previously [29]–[30] . Peroxidase-labeled anti-mouse antibody and ECL system were from GE Healthcare . Carrier-free Na125I was purchased from Perkin Elmer Lifescience . Recombinant human insulin was from Eli Lilly and recombinant mouse leptin was from R&D Systems . For feeding studies , animals were singly housed for 2 wk . Food intake was measured daily for consecutive 7-d or 14-d periods . Metabolic rates were measured by indirect calorimetry ( Windows Oxymax Equal Flow system , Columbus Instruments ) at the Washington University Diabetes Research and Training Center ( DRTC ) . Mice were housed individually in air-tight respiratory cages through which room air was passed at a flow rate of 0 . 51/min . O2 and CO2 contents of exhausted air were determined by comparison with O2 and CO2 contents of standardized sample air . VO2 and VCO2 were normalized to lean body mass [31] . For GTT studies , mice were fasted overnight and D-glucose at 2 g/kg of body weight was injected intraperitoneally . Blood glucose was monitored at 0 , 15 , 30 , 60 , 90 , and 120 min after glucose injection . Results were expressed as mean blood glucose concentration from at least six animals per genotype . For ITT studies , mice were fasted for 6 h and human insulin at 1 IU/kg of body weight was injected intraperitoneally . Blood glucose levels were monitored at 0 , 15 , 30 , 60 , and 120 min after insulin injection . Results were expressed as mean percent of basal blood glucose concentration from at least seven animals per genotype . For leptin sensitivity , 12-mo-old mice were fasted for 24 h and injected intraperitoneally with leptin ( 1 mg/kg body weight ) or PBS as control . Hypothalamic extracts were prepared 45 min after injection and immunoblotted with P-Stat3 and Stat3 antibody . For chronic intracerebroventricular ( ICV ) infusion of leptin , stereotaxic implantation of an intraventricular cannula to the third ventricle of 12-mo-old mice was performed by the Washington University Hope Center In Vivo Animal Models Core . After 1 wk , an osmotic mini-pump ( Alzet minipump ) was attached to the ICV catheter . The mini-pump delivered a constant infusion of leptin ( 50 ng/h ) or artificial cerebrospinal fluid ( aCSF ) for 14 d . Food intake and body weight were monitored . MRI , a whole-body magnetic resonance analyzer for mice ( Echo Medical Systems ) , was used to perform quantitative magnetic resonance analysis of fat-free mass and fat mass . Fat content was calculated as a percent of fat mass over the total body mass . RT-PCR analysis was performed as described previously [14] . Details can be found in Text S1 . ON-TARGET plus™ siRNA for LRP1 and control siRNA were purchased from Dharmacon Research . Cells were transiently transfected with 50 nM of the siRNA duplex using Lipofectamine 2000 ( Invitrogen ) according to the manufacturer's instructions and harvested for processing 48 h post-transfection . Experiments were performed with 125I-leptin cross-linked to unlabeled GT1-7 cells as described before [20] . Details can be found in Text S1 . Western blot analysis was performed as described previously [20] . Details can be found in Text S1 . Venous blood was taken from 12-mo-old mice and centrifuged at 7 , 000 rpm at 4°C for 5 min to separate plasma . Plasma FFAs , TG , and cholesterol analyses were performed by the Washington University DRTC . Serum leptin and insulin levels were also measured by the DRTC using ELISA methods . The lentivirus plasmids pHR'EF-Cre-WPRE-SIN and pHR-EF-GFP-WPRE-SIN have been described previously [32] . Cre lentivirus and GFP lentivirus were produced by the Washington University Hope Center Viral Vectors Core . Eight to 10-wk-old Rosa-26 reporter mice [33] or Lrp1 floxp mice were stereotaxically injected with lentivirus into the arcuate nucleus of the hypothalamus ( 4 µl , 4 . 6×108 TU/ml ) with an air pressure injector system . The injection was performed by the Washington University Hope Center In Vivo Animal Models Core . Staining was performed as described previously [32] . Details can be found in Text S1 . Staining was performed as described previously [14] . Details can be found in Text S1 . All data represent the average of at least triplicate samples . Error bars represent standard error of the mean . Statistical significance was determined by Student's t test and p<0 . 05 was considered significant .
|
The World Health Organization estimates that at least 1 in 10 adults worldwide are obese , and in some western countries , a far greater percentage ( 25% or more ) is affected . Obesity is a serious concern because it increases the risk of cardiovascular disease , type 2 diabetes , and some cancers , among other health problems . Despite recent advances in understanding the disease mechanism , effective treatments are still lacking . Lipoprotein receptors play critical roles in lipid metabolism , but their potential roles in controlling food intake and obesity in the central nervous system have not been examined . Here we show that deletion of LRP1 , a member of the LDL ( low density lipoprotein ) receptor family , in the adult mouse brain results in obese phenotype characterized by increased food intake , decreased energy consumption and decreased leptin signaling . We further show that deletion of the Lrp1 gene specifically in the hypothalamus ( a region of the brain ) by using Cre lentivirus injection is sufficient to trigger accelerated weight gain . Together , our results present a novel function of LRP1: the direct regulation of leptin signaling and energy balance in the adult central nervous system . Hence , LRP1 represents a very promising new therapeutic target for the design of innovative and more effective therapies for obesity .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"diabetes",
"and",
"endocrinology/obesity",
"neuroscience/neuronal",
"signaling",
"mechanisms",
"neurological",
"disorders/neuroendocrinology",
"and",
"pituitary"
] |
2011
|
Lipoprotein Receptor LRP1 Regulates Leptin Signaling and Energy Homeostasis in the Adult Central Nervous System
|
Antitumor cyclopalladated complexes with low toxicity to laboratory animals have shown leishmanicidal effect . These findings stimulated us to test the leishmanicidal property of one palladacycle compound called DPPE 1 . 2 on Leishmania ( Leishmania ) amazonensis , an agent of simple and diffuse forms of cutaneous leishmaniasis in the Amazon region , Brazil . Promastigotes of L . ( L . ) amazonensis and infected bone marrow-derived macrophages were treated with different concentrations of DPPE 1 . 2 . In in vivo assays foot lesions of L . ( L . ) amazonensis-infected BALB/c mice were injected subcutaneously with DPPE 1 . 2 and control animals received either Glucantime or PBS . The effect of DPPE 1 . 2 on cathepsin B activity of L . ( L . ) amazonensis amastigotes was assayed spectrofluorometrically by use of fluorogenic substrates . The main findings were: 1 ) axenic L . ( L . ) amazonensis promastigotes were destroyed by nanomolar concentrations of DPPE 1 . 2 ( IC50 = 2 . 13 nM ) ; 2 ) intracellular parasites were killed by DPPE 1 . 2 ( IC50 = 128 . 35 nM ) , and the drug displayed 10-fold less toxicity to macrophages ( CC50 = 1 , 267 nM ) ; 3 ) one month after intralesional injection of DPPE 1 . 2 infected BALB/c mice showed a significant decrease of foot lesion size and a reduction of 97% of parasite burdens when compared to controls that received PBS; 4 ) DPPE 1 . 2 inhibited the cysteine protease activity of L . ( L . ) amazonensis amastigotes and more significantly the cathepsin B activity . The present results demonstrated that DPPE 1 . 2 can destroy L . ( L . ) amazonensis in vitro and in vivo at concentrations that are non toxic to the host . We believe these findings support the potential use of DPPE 1 . 2 as an alternative choice for the chemotherapy of leishmaniasis .
Protozoan parasites of the Leishmania genus induce cutaneous , mucocutaneous and visceral diseases in man and animals . According to the World Health Organization , about 1 . 5 million of new human cases of cutaneous leishmaniasis and 500 , 000 of visceral leishmaniasis are registered annually [1] . Leishmania ( Leishmania ) amazonensis , one of the causative agents of human cutaneous leishmaniasis in the Amazon region , Brazil , is associated with both the simple and diffuse forms of the disease [2] . The first-line drugs used for treatment of leishmaniasis are pentavalent antimonial compounds , while amphotericin B and pentamidine are used as the second-line chemotherapy . However , the use of these compounds is limited by toxicity to the host and the development of resistance by the parasites [3] , [4] . Thus , the development of new leishmanicidal drugs is an important goal and several compounds including synthetic , natural products extracted from plants and marine sources have shown different degrees of efficacy in the treatment of experimental leishmaniasis [5]–[7] . The in vitro and in vivo demonstration that the viability of the Leishmania parasites is reduced by inhibitors of cysteine proteases [8]–[10] encouraged the use of virtual screening to identify additional inhibitors [11] , [12] . The demonstration that antitumor drugs may also display antileishmanial activity has also stimulated the screening of these compounds in vitro and in clinical trials [13] . Cyclopalladated complexes have shown in vitro and in vivo antitumor activity and low toxicity in animals [14]–[16] and more recently one of them exhibited lethal effects on human leukaemia cells while was ineffective against normal human lymphocytes [17] . The leishmanicidal and tripanocidal activity of cyclopalladated complexes has also been demonstrated [18]–[20] . Furthermore , there is evidence that palladacycle complexes may destroy tumoral cells by inhibition of cathepsin B activity and their inhibitory effect on Leishmania cysteine proteases in vitro was also demonstrated [18] , [21] . The present study describes the effect of one palladacycle compound called DPPE 1 . 2 on promastigotes , intracellular amastigotes and cutaneous lesions in mice infected with L . ( L . ) amazonensis .
Eight-week-old female Golden hamsters were obtained from breeding stocks maintained at the Universidade de Campinas ( São Paulo , Brazil ) and female BALB/c mice 6 to 8 weeks old were acquired from Universidade Federal de São Paulo ( São Paulo , Brazil ) . This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the Brazilian National Council of Animal Experimentation ( http://www . cobea . org . br ) . The protocol was approved by the Committee on the Ethics of Animal Experiments of the Institutional Animal Care and Use Committee at the Federal University of São Paulo ( Id # CEP 1844/08 ) . The L . ( L . ) amazonensis strain used ( MHOM/BR/1973/M2269 ) was kindly provided by Dr . Jeffrey J . Shaw , Instituto Evandro Chagas , Belém , Pará , Brazil and maintained as amastigotes by inoculation into footpads of Golden hamsters every 4 to 6 weeks . Amastigote suspensions were prepared by homogenization of excised lesions , disruption by four passages through 22-gauge needles , and centrifugation at 250×g for 10 min; the resulting supernatant was centrifuged at 1 , 400×g for 10 min , and the pellet was resuspended in RPMI 1640 . The suspension was kept under agitation for 4 h at room temperature and centrifuged at 250×g for 10 min . The final pellet contained purified amastigotes which were essentially free of contamination by other cells [22] . L . ( L . ) amazonensis promastigotes were grown at 26°C in 199 medium ( Gibco ) supplemented with 4 . 2 mM sodium bicarbonate , 4 . 2 mM HEPES , 1 mM adenine , 5 µg/ml hemin ( bovine type I ) ( Sigma , St Louis , MO , USA ) and 10% fetal calf serum ( FCS ) ( Cultilab , SP , Brazil ) . The palladacycle compound DPPE 1 . 2 ( Figure 1 ) was obtained from N , N-dimethyl-1-phenethylamine ( DMPA ) , complexed to 1 , 2-ethane-bis ( diphenylphosphine ) ( DPPE ) ligand and synthesized as previously described [16] . Stock solutions at 1 . 45 mM were prepared in dimethylsulfoxide ( DMSO ) ; for in vitro use , the drug was diluted to the appropriate concentration in cell culture medium , and for in vivo injections the stock was diluted in PBS . The promastigote cultures at 1×106 parasites/ml were kept in 199 culture medium as described above containing between 1 . 25 nM and 150 nM of DPPE 1 . 2 . Parasites were counted daily in a Neubauer chamber for three days . The leishmanicidal effect of DPPE 1 . 2 on intracellular amastigotes was evaluated in mouse bone marrow derived macrophages infected with L . ( L . ) amazonensis . Bone marrow-derived macrophages were generated from bone marrow stem cells isolated from BALB/c mice [23] . Cells were counted , added ( 8×105 ) and cultured on glass coverslips inserted in 24-well tissue culture plates containing RPMI 1640 medium buffered with 15 mM of HEPES , 20 mM of sodium bicarbonate and supplemented with 1 mM L-glutamine , 20% of fetal calf serum ( FCS ) and 30% L929 cell conditioned medium ( LCCM ) . Cultures were kept at 37°C in an atmosphere of air/CO2 ( 95/5% ) . After 5 days , the medium was changed for RPMI containing 10% of FCS and macrophages were infected at a multiplicity of 2 amastigotes per macrophage . After 24 h , infected cultures were treated with different drug concentrations ( 150 to 500 nM ) for 3 days . The coverslips were fixed with methanol , stained with hematoxylin-eosin ( HE ) and intracellular amastigotes were counted . Results are expressed by the infection index , obtained by multiplying the percentage of infected macrophages by the average number of amastigotes per macrophage . At least 200 macrophages were scored in each 3 coverslips . Amphotericin B ( Sigma-Aldrich , St Louis , MO , USA ) and Glucantime ( Sanofi-Aventis , Brazil , 300 mg/ml , 81 mg/ml SbV ) were used as standard drugs for treatment of L . ( L . ) amazonensis promastigotes and intracellular amastigotes , respectively . DPPE 1 . 2 cytotoxicity to macrophages was tested by a MTT micromethod described previously [24] after incubation of bone marrow derived macrophages with 150 to 2 , 000 nM of DPPE 1 . 2 for 3 days . Macrophages were also incubated with the highest concentration of DMSO used for DPPE 1 . 2 solubilization ( 0 . 04% ) . The formation of formazan was measured by adding 3- ( 4 , 5-dimethylthiazol-2-yl ) -2 , 5-diphenyltetrazolium bromide ( MTT; Molecular Probes , Eugene , OR , USA ) 0 . 5 mg/ml and incubation of the cultures at 37°C in the dark . After 4 h the medium was removed , 200 µl of DMSO was added per well and the absorbance was measured using an ELISA reader at 540 nm ( Labsystems Multiskan ) . For evaluation of in vivo leishmanicidal activity of DPPE 1 . 2 female BALB/c mice 6 to 8 weeks-old were infected subcutaneously at the right hind-foot with 1×105 L . ( L . ) amazonensis amastigotes . Fifteen days after infection , the animals were randomly separated in 3 groups of 12 mice each . Treated animals received in the foot lesions every other day doses of 60 mg/kg/day ( 16 . 8 mg [Sbv]/kg/day ) of Glucantime for 1 month ( total of 900 mg/kg–252 mg [Sbv]/kg/day ) or doses of 320 µg/kg/day of DPPE 1 . 2 ( total of 4 . 8 mg/kg ) . Stock solutions of DPPE 1 . 2 were prepared daily in PBS after solubilization in DMSO ( final concentration of 0 . 1% ) . Control group received the same number of injections of PBS . Infection was monitored once a week by measuring the diameter of foot lesions with a dial caliper ( Mitutoyo Corp . , Japan ) . Parasite burden from infected feet was determined by a limiting dilution method , as previously described [25] . Serum concentrations of urea , creatinine , bilirubin and transaminases were determined in BALB/c mice at the end of treatment , using sets of commercial reagents ( Doles Reagentes e Equipamentos para Laboratórios , Ltda , Brazil ) . Proteolytic activity of L . ( L . ) amazonensis promastigotes and amastigotes was determined by zymography employing electrophoretic separation of parasite lysates under unheated and nonreduced conditions resolved on 10% acrylamide gels containing 0 . 1% copolymerized gelatin ( Gibco-BRL ) by low-voltage ( 50 V ) electrophoresis [26] . Proteolytic activity in the gels was detected after 1 h of incubation , under agitation , in 0 . 1 M sodium acetate buffer , pH 5 . 0 , containing 2 . 5% Triton X-100 , followed by 2 h of incubation in the acetate buffer in the absence of Triton X-100 and Coomassie blue staining . Some gel strips after electrophoresis were incubated in buffer solutions in the presence of either protease inhibitor E-64 ( trans-epox-isuccinil-L-leucinamide- ( 4-guanide-butane ) or orthophenanthroline or DPPE 1 . 2 . Molecular weight markers ( Pharmacia LKB ) were visible on the background of stained gelatin when used in a 5-fold excess . Cathepsin activities were monitored with the fluorogenic substrates Z-Phe-Arg-AMC ( for all cathepsins ) , Z-Arg-Arg-AMC ( for cathepsin B ) , and Z-Leu-Arg-AMC ( for cathepsins K , V , and S ) ( commercially obtained from Sigma , St . Louis , MO , USA ) using 1 µl of L . ( L . ) amazonensis amastigote cell lysate ( 1×109 amastigotes disrupted in 200 µl PBS ) , 2 mM DTT ( dithiothreitol ) , 1 ml of four-component buffer comprised of 25 mM acetic acid , 25 mM Mes ( 4-Morpholineethanesulfonic acid ) , 75 mM Tris , and 25 mM glycine , pH 5 . 0 , 10 µM of each fluorogenic substrate and 50 µM of DPPE 1 . 2 . The effect of DPPE 1 . 2 on the parasite enzyme activity was tested by incubation of the L . ( L . ) amazonensis lysate with DPPE 1 . 2 for 2 minutes in the buffer solution a 37°C; the fluorogenic substrate was then added and the fluorescence of the released fluorophore , 7-amino-4-methylcoumarine ( AMC ) , was measured over time . The remaining enzyme activities were determined and expressed as a percentage of the activity of the control experiment . Parasite lysate was also incubated with 10 µM of the fluorogenic substrate Abz-Gly-Ile-Val-Arg-Ala-Lys ( Dnp ) -OH ( Sigma , St . Louis , MO , USA ) , specific for cathepsin B [27] , in the presence of either increasing concentrations of DPPE 1 . 2 or CA074 , a specific inhibitor of cathepsin B . The cathepsin activity was monitored spectrofluorometrically using the fluorogenic substrates on a Hitachi F-2000 spectrofluorometer equipped with a thermostated cell holder . The fluorescence excitation ( λEx ) and emission ( λEm ) wavelengths , for the fluorescence of AMC , were set at 380 nm and 460 nm , respectively , while the parameters for the fluorescence of Abz-peptide fragments resulting from the Abz-Gly-Ile-Val-Arg-Ala-Lys ( Dnp ) -OH hydrolysis were set at λEx = 320 and λEm = 420 nm . To determine the statistical differences between groups ANOVA and Student's t test were used and P values<0 . 05 or lower were considered statistically significant . IC50 and CC50 values were determined by GraphPad Prism , version 5 . 0 .
Axenic cultures of L . ( L . ) amazonensis promastigotes were grown in the presence of 1 . 25 to 150 nM of DPPE 1 . 2 . Significant inhibition of parasite growth was detected after 2 and 3 days of treatment with 1 . 25 to 25 nM of DPPE 1 . 2 . At 75 and 150 nM the drug inhibited 84% and 96% , respectively , of parasite growth 1 day after treatment and nearly 100% of promastigotes were killed after 2 and 3 days in the presence of these concentrations of DPPE 1 . 2 . A growth curve similar to control was observed when L . ( L . ) amazonensis promastigotes were cultured in the presence of the highest concentration of DMSO used for DPPE 1 . 2 solubilization ( 0 . 04% ) . As a control drug parasites were grown in the presence of amphotericin B . After 72 of incubation , the IC50 values for both drugs were determined ( Table 1 ) . Treatment with DPPE 1 . 2 resulted in a significant , dose-dependent decrease in infection index of L . ( L . ) amazonensis-infected macrophages with an inhibition of 92% for 500 nM of DPPE 1 . 2 ( IC50 of 128 . 35 nM; 95% confidence limits , 111 . 2–164 . 2 nM ) ( Figure 2A ) . Infected cultures were also treated with the highest concentration of DMSO used for DPPE 1 . 2 solubilization ( 0 . 04% ) ; these concentrations did not reduce the viability or the infection of macrophages ( data not shown ) . The cytotoxicity of DPPE 1 . 2 on macrophages was evaluated by the MTT method and the CC50 was determined ( 1 , 267 nM; 95% confidence limits , 1 , 15–1 , 52 nM ) . Figure 2B ) . The IC50 value expressed as µg/ml of pentavalent antimony [Sbv] was 178 . 5 µg/ml ( 95% confidence limits , 108 . 1–294 . 6 µg/ml ) and treatment with the drug at this concentration resulted in 40% of macrophage toxicity ( data not shown ) . The calculated CC50 for Glucantime expressed as µg/ml of pentavalent antimony [Sbv] was 266 . 3 µg/ml ( 95% confidence limits , 252 . 2–290 . 3 µg/ml ) . BALB/c mice infected with L . ( L . ) amazonensis were treated every other day with 320 µg/kg/day of DPPE 1 . 2 for 1 month injected in foot lesions . As can be observed in Figure 3 , starting from 24 days of treatment the animals which received DPPE 1 . 2 showed a significant decrease of foot lesion size compared to controls . Starting from 16 days of treatment , the animals that received Glucantime also exhibited significantly smaller foot lesions compared to untreated control , as well as to animals treated with DPPE 1 . 2 . Parasite load was also evaluated by limiting dilution in foot lesions of BALB/c mice one month after end of the treatment with either DPPE 1 . 2 or Glucantime . Figure 4 shows that BALB/c mice treated with either DPPE 1 . 2 or Glucantime displayed a reduction of parasite load of 97% and 99% , respectively , compared to untreated animals . To evaluate hepato and nephrotoxicity of DPPE 1 . 2 serum levels of transaminases , urea and creatinine were determined . No statistically significant alterations were detected between groups ( data not shown ) . Parasite proteolytic activity was determined by zymography after electrophoresis of L . ( L . ) amazonensis extracts in SDS-PAGE with gelatin coupled gels . Figure 5A shows that most of the proteolytic activity of L . ( L . ) amazonensis promastigotes that migrates as a 60 kDa band was abolished in the presence of orthophenanthroline , while DPPE 1 . 2 did not show any effect on this activity . On the other hand , amastigotes displayed a strong activity at a molecular mass of 30–35 kDa that was totally inhibited by either DPPE 1 . 2 or E-64 , indicating that DPPE 1 . 2 inhibits cysteine protease activity of L . ( L . ) amazonensis amastigotes . A spectrofluorometric assay using specific substrates for cathepsins in the presence of DPPE 1 . 2 was also carried out . Figure 5B shows that the L . ( L . ) amazonensis amastigote extract exhibited high hydrolytic activity on all substrates tested . Although DPPE 1 . 2 inhibited the enzymatic activity on all substrates , a significantly higher reduction on cathepsin B activity could be observed in the presence of this palladacycle complex ( 75% ) . Figure 5C shows that the activity of L . ( L . ) amazonensis extract on a most specific substrate for cathepsin B was significantly inhibited either by DPPE 1 . 2 or CA074 ( data not shown ) . The calculated IC50 values for DPPE 1 . 2 and CA074 were not significantly different ( 2 . 25±0 . 11 µM and 0 . 7±0 . 08 µM , respectively ) , strongly suggesting that DPPE 1 . 2 inhibits L . ( L . ) amazonensis cathepsin B .
The present results document the leishmanicidal effect of the palladacycle complex DPPE 1 . 2 on L . ( L . ) amazonensis . This compound destroyed L . ( L . ) amazonensis promastigotes at very low concentrations . Extension of this study to L . ( L . ) amazonensis-infected macrophages also showed an effective leishmanicidal activity of DPPE 1 . 2 against amastigotes , whereas the drug displayed 10-fold less toxicity to macrophages . Although similar leishmanicidal effect was observed with Glucantime , significantly higher concentrations of this antimonial were necessary to destroy L . ( L . ) amazonensis amastigotes . The leishmanicidal activity of DPPE 1 . 2 is comparable to that obtained with several compounds tested against L . ( L . ) amazonensis like mesoionic salt derivatives , primary S-nitrosothiols , aureobasidin A , julocrotine , tamoxifen , elatol [28]–[33] . However , higher concentrations of these compounds were used to destroy L . ( L . ) amazonensis , whereas an effective leishmanicidal effect was observed with DPPE 1 . 2 at nanomolar range . The leishmanicidal activity of metal complexes of gold , platinum , iridium , rhodium and osmium has also been investigated . However , most of them destroyed only L . ( L . ) donovani promastigotes , while few reduced the parasitism in infected animals [34]–[36] , impairing the comparison with DPPE 1 . 2 data . Among other palladium complexes previously tested against Leishmania only one was an effective inhibitor of promastigote growth , while none of them reduced the intracellular amastigote burden [18] . The leishmanicidal effect of DPPE 1 . 2 was also demonstrated in vivo . Although the reduction of parasite load in foot lesions of L . ( L . ) amazonensis-infected mice treated with DPPE 1 . 2 was similar to that obtained with Glucantime , this antimonial compound was used in 200 times higher concentration . Treatment with DPPE 1 . 2 led to a significant reduction of parasite load in foot lesions ( 97% ) , but did not result in sterile cure in infected mice . However , it is important to emphasize that the BALB/c strain is highly susceptible to L . ( L . ) amazonensis infection . These mice develop a gradual increase of foot lesions characterized by a large infiltrate of macrophages harboring a high number of amastigotes , thus mimicking the anergic form of diffuse cutaneous leishmaniasis caused by L . ( L . ) amazonensis [2] . The apparent lack of toxicity of DPPE 1 . 2 to BALB/c mice was demonstrated by hepatic and renal function assays after treatment with the drug , corroborating data that showed the low toxicity of palladacycle complexes in the treatment of mice against tumor cells [16] . More recently , the high selectivity index of one cyclopalladacycle complex with trypanocidal activity was also demonstrated , suggesting the use of this compound for treatment of Chagas' disease [19] . As reported in the literature , the antitumor property of palladacycle complexes can be attributed , at least in part , to their inhibitory activity on the cysteine protease cathepsin B [21] . This information led us to test for the possible effect of DPPE 1 . 2 on L . ( L . ) amazonensis protease activity . We showed that DPPE 1 . 2 did not inhibit the activity of the metalloprotease gp63 , the major surface protein of Leishmania promastigotes [37] . On the other hand , the high cysteine protease activity expressed in L . ( L . ) amazonensis amastigotes was inhibited by DPPE 1 . 2 and the most significant inhibition was observed on the cathepsin B activity . However , the drug did not affect the cysteine proteinase activity of mouse macrophages ( data not shown ) . Several studies demonstrated the involvement of cathepsin L-like ( cpL ) and cathepsin B-like ( cpB ) in Leishmania growth and virulence in vitro and in vivo [38]–[40] . Furthermore , cysteine proteinase inhibitors have been reported to kill Leishmania in vitro and in vivo [8]–[10] . In the present study we show that in vitro DPPE 1 . 2 inhibited L . ( L . ) amazonensis cathepsin B at higher concentrations than those necessary to kill L . ( L . ) amazonensis . These findings argue against a relationship between the leishmanicidal effect of DPPE 1 . 2 and the inhibition of L . ( L . ) amazonensis cathepsin B and suggest that other relevant targets may account for the leishmanicidal effect of the drug . Palladacycle complexes have been associated with organelle-specific effects in tumor cells such as the lysosomal and mitochondrial permeabilization that can trigger apoptosis [41] , [42] . The induction of L . ( L . ) amazonensis apoptosis by DPPE 1 . 2 has not been investigated . Leishmania killing is associated to macrophage activation by IFN-γ and TNF-α and the production of nitric oxide [43] , whereas TGF-β is an immunosupressor cytokine known to exacerbate visceral and cutaneous leishmaniasis [44]–[47] . Interestingly , Leishmania cathepsin B is involved in the conversion of latent TGF-β to its biologically active form [48] . Since we have shown here that DPPE 1 . 2 inhibits parasite cathepsin B , killing of L . ( L . ) amazonensis in infected mice treated with the drug may be associated to protective responses arising from lower expression of the active form of TGF-β . This possibility is now under investigation . In conclusion , the effectiveness of DPPE 1 . 2 in destroying L . ( L . ) amazonensis in vitro and by intralesional administration in vivo at concentrations non toxic to the host support further studies of the leishmanicidal activity of the palladacycle as an additional choice to available chemotherapies .
|
Leishmaniasis is an important public health problem with an estimated annual incidence of 1 . 5 million of new human cases of cutaneous leishmaniasis and 500 , 000 of visceral leishmaniasis . Treatment of the diseases is limited by toxicity and parasite resistance to the drugs currently in use , validating the need to develop new leishmanicidal compounds . We evaluated the killing by the palladacycle complex DPPE 1 . 2 of Leishmania ( Leishmania ) amazonensis , an agent of human cutaneous leishmaniasis in the Amazon region , Brazil . DPPE 1 . 2 destroyed promastigotes of L . ( L . ) amazonensis in vitro at nanomolar concentrations , whereas intracellular amastigotes were killed at drug concentrations 10-fold less toxic than those displayed to macrophages . L . ( L . ) amazonensis-infected BALB/c mice treated by intralesional injection of DPPE 1 . 2 exhibited a significant decrease of foot lesion sizes and a 97% reduction of parasite burdens when compared to untreated controls . Additional experiments indicated the inhibition of the cathepsin B activity of L . ( L . ) amazonensis amastigotes by DPPE 1 . 2 . Further studies are needed to explore the potential of DPPE 1 . 2 as an additional option for the chemotherapy of leishmaniasis .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"biology",
"microbiology",
"parasitology"
] |
2012
|
In Vitro and In Vivo Activity of a Palladacycle Complex on Leishmania (Leishmania) amazonensis
|
T . pallidum subsp . endemicum ( TEN ) is the causative agent of bejel ( also known as endemic syphilis ) . Clinical symptoms of syphilis and bejel are overlapping and the epidemiological context is important for correct diagnosis of both diseases . In contrast to syphilis , caused by T . pallidum subsp . pallidum ( TPA ) , TEN infections are usually spread by direct contact or contaminated utensils rather than by sexual contact . Bejel is most often seen in western Africa and in the Middle East . The strain Bosnia A was isolated in 1950 in Bosnia , southern Europe . The complete genome of the Bosnia A strain was amplified and sequenced using the pooled segment genome sequencing ( PSGS ) method and a combination of three next-generation sequencing techniques ( SOLiD , Roche 454 , and Illumina ) . Using this approach , a total combined average genome coverage of 513× was achieved . The size of the Bosnia A genome was found to be 1 , 137 , 653 bp , i . e . 1 . 6–2 . 8 kbp shorter than any previously published genomes of uncultivable pathogenic treponemes . Conserved gene synteny was found in the Bosnia A genome compared to other sequenced syphilis and yaws treponemes . The TEN Bosnia A genome was distinct but very similar to the genome of yaws-causing T . pallidum subsp . pertenue ( TPE ) strains . Interestingly , the TEN Bosnia A genome was found to contain several sequences , which so far , have been uniquely identified only in syphilis treponemes . The genome of TEN Bosnia A contains several sequences thought to be unique to TPA strains; these sequences very likely represent remnants of recombination events during the evolution of TEN treponemes . This finding emphasizes a possible role of repeated horizontal gene transfer between treponemal subspecies in shaping the Bosnia A genome .
Uncultivable human pathogenic treponemes include T . pallidum subsp . pallidum ( TPA ) , causing syphilis , T . pallidum subsp . pertenue ( TPE ) , causing yaws , and T . pallidum subsp . endemicum ( TEN ) , causing bejel , which is also known as endemic or nonvenereal syphilis . Infections caused by TPE and TEN are commonly denoted as endemic treponematoses . While yaws is found in warm , moist climates , bejel is found in drier climates . In both cases , infection is spread by direct contact ( e . g . skin-to-skin or skin-to-mucosa ) . In addition , bejel can also be transmitted by contact with contaminated utensils [1] , [2] . The current , and widespread , belief that yaws and bejel are non-sexually transmitted may simply reflect that these diseases mostly affect children that have not reached sexual maturity [3] , [4] . Diagnosis of endemic treponematoses comprises clinical symptoms , epidemiological data , and serology . Since there is significant clinical similarity between the symptoms of syphilis and endemic treponematoses , and serology cannot discriminate between infection with TPA , TPE , and TEN strains , the epidemiology plays a major role in establishing a diagnosis . While yaws remains endemic in poor communities in Africa , Southeast Asia , and the western Pacific , bejel is predominant in western Africa and in the Middle East ( reviewed in [2] , [4] ) . Imported cases of yaws and bejel have been documented in children in Europe and Canada [5] , [6] . With the accumulation of genetic data , molecular targets that can be used to differentiate treponemal subspecies , at the molecular level , have become available [2] . Endemic syphilis has been described almost everywhere in Europe since the 16th century ( for review see [7] ) and often was described under different names , e . g . the disease that appeared in Brno , CZ in 1575 was called morbus Brunogallicus , although it is not clear whether this infection was not perhaps caused by the syphilis treponeme [8] . The Bosnia A strain was isolated in 1950 in Bosnia , a country in southern Europe , from a 35-year old male with mucous patches under the tongue and on the tonsils; additionally , the patient showed secondary lesions ( papules ) on the face , trunk and extremities . Material for experimental inoculation of laboratory animals was taken from an ulcer on the shaft of the penis [9] . Although several other isolates were collected from bejel patients , only one additional strain of T . pallidum subsp . endemicum ( Iraq B ) is currently propagated in laboratory settings . In this study , the complete genome sequence of the T . pallidum subsp . endemicum Bosnia A strain was obtained using a combination of next-generation sequencing approaches and compared to the genomes of the four TPE strains ( Samoa D , CDC-2 , Gauthier , Fribourg-Blanc isolate ) and five TPA strains ( Nichols , DAL-1 , Chicago , SS14 , Mexico A ) , all of which have been determined in recent years [10]–[15] .
Bosnia A DNA was provided by Dr . Sylvia M . Bruisten from the Public Health Service , GGD Amsterdam , Amsterdam , The Netherlands . Bosnia A genomic DNA was amplified using the pooled segment genome sequencing ( PSGS ) method as described previously [11] , [15] . Briefly , Bosnia A DNA was amplified with 214 pairs of specific primers to obtain overlapping PCR products ( Table S1 ) . To facilitate sequencing of paralogous genes containing repetitive sequences , PCR products were mixed in equimolar amounts into four distinct pools . Prior to next-generation sequencing ( 454-pyrosequencing , Illumina and SOLiD ) , the PCR products constituting each pool were labeled with multiplex identifier ( MID ) adapters and sequenced as four different samples . Two genomic regions were not amplified during PSGS and therefore were not used for sequencing the whole genome ( gaps between coordinates 332290–335395 and 1123251–1123648 according to the Nichols sequence , AE000520 . 1 [16]; see Table S1 ) . Sequences in these regions were Sanger sequenced at the University of Washington in Seattle ( WA ) , USA . Whole genome DNA sequencing was done using the Applied Biosystems/SOLiD 3 System platform ( Life Technologies Corporation , Carlsbad , CA , USA ) combined with the Roche/Genome Sequencer FLX Titanium platform ( 454 Life Sciences , Branford , CT , USA ) and with the Illumina/Solexa HiSeq 2000 approach ( Illumina , San Diego , CA , USA ) . SOLiD sequencing was performed at SeqOmics Ltd ( Mórahalom , Hungary ) , 454-pyrosequencing and Illumina sequencing were performed at The Genome Institute , Washington University School of Medicine ( St . Louis , MO , USA ) . SOLiD , 454 , and Illumina sequencing resulted in average read lengths of 40 bp , 504 bp and 100 bp and the total average depth coverage of 234× , 138× and 141× , respectively . 454 and Illumina sequencing reads were obtained from 4 distinct pools ( sequenced as 4 different samples – see Table S1 ) and were separately assembled de novo using a Newbler assembler ( 454 Life Sciences , Branford , CT , USA ) or TIGRA [17] , respectively . The resulting 454 and Illumina contigs obtained for each pool were then aligned to the corresponding sequences ( representing each pool sequence ) of the reference CDC-2 genome ( CP002375 . 1 [11] ) using Lasergene software ( DNASTAR , Madison , WI , USA ) . All gaps and discrepancies between these platforms within each pool were resolved using Sanger sequencing . Altogether , 20 genomic regions of the Bosnia A genome were amplified and Sanger sequenced . The final overlapping pool sequences were joined to obtain complete genome sequence of the Bosnia A strain . The SOLiD sequencing results were mapped to the reference Samoa D genome ( CP002374 . 1 [11] ) using the CLC Genomics Workbench ( CLC bio , Cambridge , MA , USA ) and were processed as mentioned above . The genome sequence obtained from SOLiD was then compared with the consensus genome sequence obtained from 454 and Illumina . All discrepancies were resolved using Sanger sequencing . Two TPE genomes ( CDC-2 or Samoa D ) were used as reference genomes for contig alignments since only few minor genetic differences have been found to be specific within individual TPE strains [11] . Due to low coverage , one genomic region ( Treponema pallidum interval; TPI ) , was amplified with specific primers using a GeneAmp XL PCR Kit ( Applied Biosystems , Foster City , CA , USA ) [18] , [19] . This TPI-48 interval contained paralogous genes tprI and tprJ . The PCR product was purified using a QIAquick PCR Purification Kit ( QIAGEN , Valencia , CA , USA ) according to the manufacturer's instructions and Sanger sequenced using internal primers . The tprK ( TENDBA_0897 ) , arp ( TENDBA_0433 ) , and TENDBA_0470 genes were amplified and cloned into the pCR 2 . 1-TOPO cloning vector ( Invitrogen , Carlsbad , CA , USA ) . Nine independent clones for the tprK and arp genes and seven clones for TENDBA_0470 were sequenced as previously described [11] . A total of 7 genomic regions ( in genes TENDBA_0040 , TENDBA_0348 , TENDBA_0461 , TENDBA_0697 , TENDBA_0859 , TENDBA_0865 and TENDBA_0966 ) revealed intra-strain variability in the length of homopolymeric ( G- or C- ) stretches . The prevailing length of these regions was determined by TOPO TA-cloning and Sanger sequencing . At least five independent clones were sequenced as previously described [15] . The final whole genome sequence of the Bosnia A strain was assembled from SOLiD , 454 and Illumina contigs . In addition , Sanger sequencing was used for finishing the complete genome sequence and for additional sequencing including paralogous , repetitive and intra-strain variable chromosomal regions . Geneious software v5 . 6 . 5 [20] was used for gene annotation based on the annotation of the TPE CDC-2 genome [11] . Genes were tagged with TENDBA_ prefix . The original locus tag numbering corresponds to the tag numbering of orthologous genes annotated in the TPE CDC-2 genome [11] . The TENDBA_0897 gene , coding for TprK , showed intra-strain variable nucleotides and therefore nucleotides in variable regions were denoted with Ns in the complete Bosnia A genome . For proteins with unpredicted functions , a gene size limit of 150 bp was applied . Protein domains and functional annotation of analyzed genes were characterized using Pfam [21] , CDD [22] and KEGG [23] databases . Whole genome nucleotide alignments of five TPA strains , four TPE strains and the Bosnia A strain were used for determination of genetic relatedness using several approaches including calculation of nucleotide diversity ( π ) and construction of a phylogenetic tree . All positions containing indels in at least one genome sequence were omitted from the analysis . There were a total of 1 , 128 , 391 nucleotide positions aligned in the final dataset . TPA strains comprised Nichols ( re-sequenced genome CP004010 . 2 [14] ) , DAL-1 ( CP03115 . 1 [13] ) , SS14 ( re-sequenced genome CP004011 . 1 [14] ) , Chicago ( CP001752 . 1 [10] ) , and Mexico A ( CP003064 . 1 [12] ) genomes , while TPE strains included Samoa D ( CP002374 . 1 [11] ) , CDC-2 ( CP002375 . 1 [11] ) , Gauthier ( CP002376 . 1 [11] ) and Fribourg-Blanc ( CP003902 . 1 [15] ) . Whole genome alignments were constructed using Geneious software [20] and SeqMan software ( DNASTAR , Madison , WI , USA ) . Nucleotide differences among studied whole genome alignments were analyzed using DnaSP software , version 5 . 10 [24] . An unrooted phylogenetic tree was constructed from the whole genome sequence alignment using the Maximum Parsimony method and MEGA5 software [25] . To test , whether the mosaic character of identified loci were a result of intra-strain recombination , potential donor sites were screened from the entire Bosnia A genome using several computer programs and algorithms including RDP3 [26] , EditSeq software ( DNASTAR , Madison , WI , USA ) , BLAST ( http://blast . ncbi . nlm . nih . gov ) , and Crossmatch ( http://www . phrap . org/phredphrapconsed . html ) . We failed to find any potential donor sites in the Bosnia A genome . We also failed to find any TPA- or TPE-specific NGS reads in the regions having a mosaic character . The complete genome sequence of the Bosnia A strain was deposited in the GenBank under accession number CP007548 .
Sequencing of the TEN Bosnia A strain genome using three independent next-generation sequencing platforms yielded a total combined average coverage of 513× . The summarized genomic features of the Bosnia A strain in comparison to previously sequenced TPA and TPE strain genomes are shown in Table 1 . The size of the Bosnia A genome ( 1 , 137 , 653 bp ) was 1 , 628–2 , 828 bp shorter than the sizes of previously published genomes for TPA and TPE strains [10]–[15] . The overall gene order in the Bosnia A genome was identical to other TPE and TPA strains . Altogether , 1125 genes were annotated in the Bosnia A genome including 54 untranslated genes encoding rRNAs , tRNAs and other ncRNAs ( short bacterial RNA molecules that are not translated into proteins ) . A total of 640 genes ( 56 . 9% ) encoded proteins with predicted function , 137 genes encoded treponemal conserved hypothetical proteins ( TCHP , 12 . 2% ) , 141 genes encoded conserved hypothetical proteins ( CHP , 12 . 5% ) , 145 genes encoded hypothetical proteins ( HP , 12 . 9% ) and 8 genes ( TENDBA_0082a , TENDBA_0146 , TENDBA_0316 , TENDBA_0370 , TENDBA_0520 , TENDBA_0532 , TENDBA_0812 and TENDBA_1029; 0 . 7% ) were annotated as pseudogenes . The average and median gene lengths of the Bosnia A genome were calculated to 979 . 2 bp and 831 bp , respectively . The intergenic regions covered 52 . 6 kbp and represented 4 . 63% of the total Bosnia A genome length . In general , other calculated genomic parameters were similar to other TPE strains . When compared to TPA strains , the Bosnia A genome contained a 635 bp long insertion in the tprF locus . In this respect , the Bosnia A genome was similar to TPE strains . When compared to both TPA and TPE genomes , the Bosnia A genome contained a 2300 bp long deletion involving the tprF and G loci ( TPANIC_0316 and TPANIC_0317 in the Nichols genome CP004010 . 2 [14] ) . Moreover , the predicted TENDBA_0316 gene ( 1860 bp in length ) was a chimera encompassing the tprG 5′-region , tprI-like sequence and the tprF 3′-region , and was hence designated as tprGI as previously described by Centurion-Lara et al . [27] ( Table 2 ) . Two insertions of 65 bp and 52 bp , respectively , resulted in the prediction of two hypothetical genes , TENDBA_ 0126b and TENDBA_548a . The same orthologs were also predicted in TPE but not in TPA strains ( Table 2 ) . Besides the annotated pseudogenes in the Bosnia A genome ( see above ) , 8 additional genes ( orthologous to TP0129 , TP0132 , TP0135 , TP0266 , TP0318 , TP0370 , TP0671 and TP1030 ) were considered pseudogenes . The same genes were also considered pseudogenes in TPE strains [11] , [15] ( Table 1 ) . Sequence relatedness of the Bosnia A genome to other Treponema pallidum genomes is shown in Fig . 1 . This unrooted tree was constructed using several available whole genome sequences of uncultivable pathogenic treponemes . The image clearly showed clustering of the Bosnia A strain with the TPE strains . The Bosnia A genome was found to be 99 . 91–99 . 94% and 99 . 79–99 . 82% identical to the TPE and TPA genomes , respectively ( Table 3 ) . The nucleotide diversity between TPE strains and the Bosnia A strain ( 0 . 00063±0 . 00032 to 0 . 00086±0 . 00043 ) was about three times lower than the nucleotide diversity between TPA strains and the Bosnia A strain ( 0 . 00181±0 . 00090 to 0 . 00212±0 . 00106 ) . For comparison , calculated π values between the Bosnia A strain and individual TPA strains were of the same order of magnitude as π values between TPA and TPE strains ( Table 4 ) . To identify Bosnia A-specific differences , the Bosnia A genome was compared to the available genomes of TPE strains [11] , [15] and TPA strains [10] , [12]–[14] . The Bosnia A strain-specific sequences were defined as those not present in both TPA and TPE strains and altogether comprised 406 differences ( indels and substitutions with a total length of 2772 bp ) equally distributed along the Bosnia A genome ( Fig . 2 ) . Differences in coding regions included 9 deletions , 5 insertions and 360 nucleotide substitutions for a total of 2728 bp ( Table 5 ) . Those 360 substitutions resulted in 197 Bosnia A-specific amino acid differences in the putative proteome . Most of the nucleotide substitutions were found in the TENDBA_0136 , TENDBA_0548 , TENDBA_0856 , TENDBA_0859 and TENDBA_0865 genes ( Table 5 ) . Bosnia A-specific frameshift mutations ( caused by three deletions and one insertion ) resulted in significant gene truncation ( TENDBA_0082a , TENDBA_0316 and TENDBA_1029 ) or elongation ( TENDBA_0126b ) ( Table 2 ) . Other detected indels resulted in 6 protein shortenings ( TENDBA_0067 , TENDBA_0136 , TENDBA_0225 , TENDBA_0548 , TENDBA_0859 , and TENDBA_0865 ) and 4 protein elongations ( TENDBA_0856 , TENDBA_0859 , TENDBA_0897 , and TENDBA_0898 ) ( Table 5 ) . All affected genes code for hypothetical proteins of unknown function except for TENDBA_0898 coding for RecB ( exodeoxyribonuclease V beta subunit; EC3 . 1 . 11 . 5 ) . TENDBA_0136 and TENDBA_0865 have been predicted to be putative outer membrane proteins . In addition , TPA and TPE orthologs to TENDBA_0136 have been experimentally shown to bind human fibronectin [28] . TENDBA_0856 has been predicted to be putative lipoprotein . No putative conserved domains have been detected in hypothetical proteins except for TENDBA_0067 , TENDBA_0225 and TENDBA_1029 containing TPR ( tetratricopeptide ) domain , LRR_5 ( leucine rich repeat ) domain and DbpA ( RNA binding ) domain , respectively ( Table 5 ) . All nonsynonymous substitutions have been identified outside the predicted domains . Genome sequences differentiating the Bosnia A strain from the TPA but not TPE strains are shown in Fig . 2 . These sequences were found to be regularly distributed along the Bosnia A genome and altogether comprised 1422 differences ( indels and substitutions of total length of 2335 bp ) . In the coding regions , 2128 bp including 13 deletions , 9 insertions and 1296 substitutions differentiated genomes of TPA strains from Bosnia A and other TPE strains ( Table 6 ) . A set of 1296 substitutions resulted in 631 amino acid differences in the encoded proteins . Most of the differences were found in genes TENDBA_0117 ( tprC ) , TENDBA_0131 ( tprD ) , TENDBA_0133 , TENDBA_0134 , TENDBA_0136 , TENDBA_0304 , TENDBA_0314 , TENDBA_0462 , TENDBA_0619 , TENDBA_0620 ( tprI ) , and TENDBA_0621 ( tprJ ) ( Table 6 ) . Except for TENDBA_0103 coding for RecQ ( ATP-dependent DNA helicase; EC3 . 6 . 4 . 12 ) and TENDBA_0027 coding for HlyC ( putative hemolysin ) , all other affected genes code for hypothetical proteins of unknown function . TENDBA_0134 has been predicted to be putative outer membrane protein . TENDBA_0462 and TENDBA_0858 have been predicted to be putative lipoproteins . No putative conserved domains have been detected in hypothetical proteins except for TENDBA_0067 and TENDBA_0304 conatining TPR ( tetratricopeptide ) domain and peptidase_MA_2 domain , respectively ( Table 6 ) . All nonsynonymous substitutions have been identified outside the predicted domains . Genome sequences differentiating the Bosnia A strain from TPE but not TPA strains are shown in Fig . 2 . These sequences were also found to be regularly distributed along the Bosnia A genome and , altogether , comprised 197 differences in genome positions ( containing indels and substitutions encompassing a total of 635 bp ) . Three deletions , three insertions and 174 substitutions ( Table 7 ) were found within the Bosnia A coding regions , encompassing a total of 612 bp . The 174 substitutions resulted in 101 amino acid differences in the putative encoded proteins . Most of the substitution differences were found in genes TENDBA_0136 , TENDBA_0488 , TENDBA_0577 , TENDBA_0856a/TENDBA_0858 , TENDBA_0859 , TENDBA_0865 and TENDBA_0968 ( Table 7 ) . An insertion of 378 bp in TENDBA_1031 ( tprL ) resulted in a gene elongation ( Table 2 ) . TENDBA_0488 codes for Mcp ( methyl-accepting chemotaxis ) protein . All other genes code for hypothetical proteins of unknown function . Two genes have been predicted to encode putative outer membrane proteins ( TENDBA_0136 and TENDBA_0865 ) and one gene has been predicted to encode putative lipoprotein ( TENDBA_0858 ) . No putative conserved domains have been detected in hypothetical proteins ( Table 7 ) . Despite the overall sequence similarity of the Bosnia A genome to TPE strains , several chromosomal sequences were found to be almost identical to sequences in TPA strains . The Bosnia A sequence in the TENDBA_0577 locus was identical to four out of 5 orthologous sequences of completely sequenced TPA strains ( Fig . 3 ) . In the TENDBA_0968 locus , stretches of TPA- and TPE-like sequences were found ( Fig . 3 ) and a similar pattern was also found in TENDBA_0858 ( not shown ) . In addition , TENDBA_0326 ( tp92 , bamA ) was identical to the orthologous sequence of TPA SS14 ( coordinates 1593–1649 , Fig . 3 ) and to all TPA strains ( with the exception of the TPA Mexico A strain ) between coordinates 2127–2494 . The TPA Mexico A strain is , in this region , similar to TPE strains [12] , [29] . While the latter TPA-like sequences in TENDBA_0326 were almost 0 . 4 kbp long , other TPA-like sequences were usually relatively short , ranging from about 50–70 bp . However , TPA-like sequences of the Bosnia A strain were clearly different from Bosnia A-specific sequences with sporadic nucleotide positions identical to TPA sequences ( TENDBA_0856; Fig . 3 ) . The previously reported 378 bp insertion almost identical to TPA strains ( differing only in one nucleotide position [27] ) was confirmed in TENDBA_1031 as well as the nucleotide mosaic in the TP0488 ( mcp2-1 ) locus; revealing a sequence identical to TPA Mexico A ( with the exception of 2 single nucleotide substitutions [12] ) . Altogether , at least seven TPA-like sequences having 5 or more nucleotide positions identical to TPA sequences and not interrupted by TPE-like nucleotide positions were found in the Bosnia A genome .
The first complete genome sequence of the bejel-causing agent , T . pallidum subsp . endemicum ( TEN ) strain Bosnia A , was determined using three independent next-generation sequencing techniques . Because the total combined coverage was >500× and all sequencing ambiguities were resolved with Sanger sequencing , the quality of this new genome is very high . This allowed us to carry out a comparative analysis of the Bosnia A genome with the already available treponemal genomes [10]–[15] , [30] with a high degree of confidence that our results would not be affected by sequencing errors . In several of the previously published genomes , the whole genome sequence was compared to whole genome fingerprinting data to assess the quality of the genome sequence . In each of the previously tested genomes , the sequencing error rate was less than 10−4 [11] , [12] , [15] , [30] . The genome length of strain Bosnia A ( 1 , 137 , 653 bp ) is about 2 kbp shorter than the length of TPE or TPA genomes . This is caused by a 2300 bp deletion in the tprF and tprG loci . This deletion was also confirmed in the TEN Iraq B sequence [27] suggesting that this is a common feature of bejel strains . An identical deletion was also found in the T . paraluisleporidarum ec . Cuniculus genome ( formerly denoted T . paraluiscuniculi Cuniculi A [30] , [31] ) . Moreover , this type of deletion was observed during PCR amplification of the tprF and tprG loci in other treponemal genomes ( M . Strouhal , D . Šmajs; unpublished data ) . This fact , together with the presence of repeats in the flanking regions suggests that this 2300 bp deletion is a result of polymerase slippage and that this deletion could have happened several times independently during evolution . In fact , no other similarities between the Bosnia A and T . p . ec . Cuniculus genome were found with respect to other identified indels in the T . p . ec . Cuniculus genome . The overall genetic similarity of Bosnia A to the sequenced TPE strains is 99 . 91–99 . 94% , at the DNA level . For comparison , the sequence similarity between TPA and TPE strains is greater than 99 . 8% [11] , [15] . This enormous sequence similarity among TPA , TPE and TEN strains is the molecular basis for the long established fact that individual etiological agents of syphilis and endemic treponematoses ( yaws and bejel ) cannot be distinguished by their morphology or serology . Although syphilis , yaws , and bejel show differences in their geographical distribution , mode of transmission , invasiveness and pathogenicity , it is known that the clinical symptoms of these diseases overlap and one disease can mimic the others . Interestingly , in very dry areas , yaws symptoms are almost the same as bejel symptoms [32]; which again reflects the extremely high sequence similarity between TPE and TEN strains . In many or perhaps most cases , the final diagnosis is therefore often based on the epidemiological context of the infection . However , at the same time , even small genomic differences ( although not known at present ) have the potential to influence the phenotypic differences between the clinical manifestations of syphilis , yaws and bejel . Additional whole genome sequences of TPA , TPE and TEN strains will help to identify a set of invariant differences between the etiological agents of these diseases , which could help answer this question . At the same time , the TEN Bosnia A strain is clearly distant from the cluster of TPE strains . However , additional TEN whole genome sequences will be needed to assess the variability within TEN strains . To our knowledge , there is only one additional laboratory stock of TEN , i . e . strain Iraq B . Previous studies on the Iraq B isolate revealed a high degree of similarity to Bosnia A [27] , [29] , [33]–[36] suggesting that this strain is more related to Bosnia A than to TPE strains . Most prominent genetic changes between Bosnia A and TPE and/or TPA genomes resulting in protein truncations or elongations were located in just 14 genes . These genes encoded TprA , F , G , and L proteins , RecQ protein , ethanolamine phosphotransferase , and treponemal conserved hypothetical proteins ( 3 ) or hypothetical proteins ( 5 ) . Both Tpr and RecQ proteins were found to also be affected in the T . p . ec . Cuniculus genome [30] . While the tprA gene was functional in Bosnia A and TPE strains but not among TPA strains ( except for strain Sea 81-4; see [37] ) , tprF and tprG were partially deleted ( similarly to T . p . ec . Cuniculus genome ) and the tprL gene was elongated in a way that was similar to that seen in TPA strains . These changes were already described in detail by Centurion-Lara et al . [27] . Tpr proteins likely play an important role in treponemal infectivity , pathogenicity , immune evasion and host specificity . Tpr proteins induce an antibody response during infection and exhibit heterogeneity both within and among T . pallidum subspecies and strains [38]–[40] . In the T . p . ec . Cuniculus genome , a mutation in recQ resulted in a predicted RecQ protein without a C-terminal or DNA-binding domain [41]; on the other hand in Bosnia A the frameshift reversion led to a functional recQ gene ( similar to that seen in TPE genomes [11] ) . Other prominent changes seen in the Bosnia A strain include a different number of tandem repeat units in TENDBA_0433 ( encoding Arp ) and TENDBA_0470 genes ( encoding conserved hypothetical protein ) compared to orthologous genes in individual TPE and TPA strains . The same number of 60-bp tandem repeat units ( all of Type II ) within the arp gene was found in the Bosnia A genome as previously described [42] . Variable numbers of tandem repeat units in genes orthologous to TENDBA_0470 have already been described in TPE and TPA strains [11] , [15] , [19] . The genome of Bosnia A showed several genetic loci with sequences identical to TPA sequences ( Fig . 3 ) . The TENDBA_0577 gene encoded treponemal conserved hypothetical protein of unknown function with predicted cytoplasmic membrane localization . This gene was completely identical to TPA orthologs and differed from TPE orthologs by deletion of 12 nucleotides and substitution of 5 nucleotides . Recent studies of σ factor RpoE ( TP0092 ) binding sites identified gene TP0577 ( orthologous to TENDBA_0577 ) as one of 22 putative TP0092-controlled ORFs [43] . The TENDBA_0577 thus could possibly code for a protein integrated in the stress response pathway during the first days post infection . Similarly , the 378 bp insertion in TENDBA_1031 is with exception of a 1 nucleotide insertion almost identical to orthologs of the TPA strain ( but not to TPE strains ) . In other genes ( TENDBA_0968 , TENDBA_0858 ) , 50–70 bp long sequences identical to one or several TPA strains were found indicating that the genome of Bosnia A incorporated sequences identical to TPA strains . Most of the above mentioned genes were found to evolve under positive selection in TPA-TPE comparisons [11] . In fact , previous papers found this type of mixed TPA and TPE sequences in TPA Mexico A and South Africa strains [12] , [29] . Moreover , previous reports have shown that TEN strain Bosnia A contains the same nucleotide mosaic at the TP0488 ( mcp2-1 ) locus as TPA Mexico A ( with the exception of 2 single nucleotide substitutions ) . Despite the numerous efforts to identify potential donor sites within TPA Mexico A that could explain the existence of these sequences by intra-strain recombination [12] , no such sites have been identified in the Mexico A genome . Similarly , no donor sites have been identified in the Bosnia A genome either . It is likely that these sequences identical to TPA in the Bosnia A genome could result from inter-strain recombination event between TPA and TEN strains during a simultaneous infection of multiple hosts during the TEN evolution . Although the overall genome sequence of Bosnia A is related to TPE strains , horizontal gene transfer appears to be the mechanism that introduced at least seven chromosomal sequences related to TPA SS14 , TPA Mexico A , and other TPA strains . In fact , both the TPA SS14 and Mexico A sequences are required and sufficient to provide sequences to Bosnia A genome . Moreover , at least two subsequent transfers had to occur to introduce both SS14- and Mexico A-specific sequences . Experimental infection with either TPA , TPE or TEN strains did not result in complete cross-protection [9] . In addition , recombination mechanisms are more active during treponemal infection and represent important genetic mechanisms for avoiding the host immune response [40] . Moreover , the absence of modification and restriction systems and the presence of genes for homologous recombination in pathogenic treponemes [16] appear to allow incorporation of foreign DNA molecules with subsequent integration into chromosomal DNA . Therefore , uptake of TPA DNA by a TEN strain during a simultaneous infection of multiple hosts appears to be a possible explanation . It is clear that TPA strains can be classified as SS14-like ( SS14 , Mexico A ) and Nichols-like strains ( Nichols , DAL-1 , Chicago ) [14] , [44] and that most of the TPA strains causing infections throughout the world are in fact SS14-like strains [36] . However , it is not clear if the SS14 and Mexico A sequences in the Bosnia A genome reflect a greater prevalence of SS14-like strains in the human population or an accidental coincidence of transfers from SS14-like strains . Moreover , there are several loci in the Bosnia A genome similar to the TENDBA_0856 locus ( TENDBA_0483 , TENDBA_0858 , TENDBA_0865 ) that represent regions of Bosnia A-specific sequences with only sporadic nucleotide positions that are identical to TPA sequences . These sequences may be identical to other , yet unidentified , TPA strains or isolates . If such TPA isolates are identified in the future , they may help to unravel the evolution of TPA and TEN treponemes .
|
Uncultivable treponemes represent bacterial species and subspecies that are obligate pathogens of humans and animals causing diseases with distinct clinical manifestations . Treponema pallidum subsp . pallidum causes sexually transmitted syphilis , a multistage disease characterized in humans by localized , disseminated , and chronic forms of infection , whereas Treponema pallidum subsp . pertenue ( agent of yaws ) and Treponema pallidum subsp . endemicum ( agent of bejel ) cause milder , non-venereally transmitted diseases affecting skin , bones and joints . The genetic basis of the pathogenesis and evolution of these microorganisms are still unknown . In this study , a high quality whole genome sequence of the T . pallidum subsp . endemicum Bosnia A strain was obtained using a combination of next-generation sequencing approaches and compared to the genomes of available uncultivable pathogenic treponemes . Relative to all known genomes of Treponema pallidum subspecies , no major genome rearrangements were found in the Bosnia A . The Bosnia A strain clustered with other yaws-causing strains , while syphilis-causing strains clustered separately . In general , the Bosnia A genome showed similar genetic characteristics to yaws treponemes but also contained several sequences thought to be unique to syphilis-causing strains . This finding suggests a possible role of repeated horizontal gene transfer between treponemal subspecies in shaping the Bosnia A genome .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"genetics",
"biology",
"and",
"life",
"sciences",
"microbiology",
"molecular",
"biology"
] |
2014
|
Whole Genome Sequence of the Treponema pallidum subsp. endemicum Strain Bosnia A: The Genome Is Related to Yaws Treponemes but Contains Few Loci Similar to Syphilis Treponemes
|
Although mosquitoes serve as vectors of many pathogens of public health importance , their response to viral infection is poorly understood . It also remains to be investigated whether viruses deploy some mechanism to be able to overcome this immune response . Here , we have used an RNA-Seq approach to identify differentially regulated genes in Culex quinquefasciatus cells following West Nile virus ( WNV ) infection , identifying 265 transcripts from various cellular pathways that were either upregulated or downregulated . Ubiquitin-proteasomal pathway genes , comprising 12% of total differentially regulated genes , were selected for further validation by real time RT-qPCR and functional analysis . It was found that treatment of infected cells with proteasomal inhibitor , MG-132 , decreased WNV titers , indicating importance of this pathway during infection process . In infection models , the Culex ortholog of mammalian Cul4A/B ( cullin RING ubiquitin ligase ) was found to be upregulated in vitro as well as in vivo , especially in midguts of mosquitoes . Gene knockdown using dsRNA and overexpression studies indicated that Culex Cul4 acts as a pro-viral protein by degradation of CxSTAT via ubiquitin-proteasomal pathway . We also show that gene knockdown of Culex Cul4 leads to activation of the Jak-STAT pathway in mosquitoes leading to decrease viral replication in the body as well as saliva . Our results suggest a novel mechanism adopted by WNV to overcome mosquito immune response and increase viral replication .
Flaviviruses , such as West Nile virus ( WNV ) and dengue virus ( DENV ) , pose a huge burden on public healthcare system worldwide . With more than half of world’s population at risk of infection , the geographic distribution of these mosquito-borne flaviviruses is expanding due to increased travel , trade and climate change [1] . First isolated in Uganda in 1937 , WNV is now endemic in parts of Africa , Europe , the Middle East , Asia , Australia and the Americas [2] . Transmitted by Culex mosquitoes and causing an acute febrile illness that can lead to severe neurological disease , there is currently no specific vaccine or anti-viral for WNV approved for use in humans [3] . The mammalian response to flavivirus infection has been well studied . Mosquito immune pathways are less well understood but some recent studies have shown that they may play an important role during infection in the vector [4 , 5] . Although lacking essential components of the mammalian innate and adaptive immune systems , such as interferons , antibodies , B cells , T cells and MHC antigens , mosquitoes have been shown to respond to viral infection by a range of mechanisms including RNA interference ( RNAi ) and by activation of several evolutionarily conserved signal transduction pathways , include the Toll , Imd/JNK and Jak-STAT [4–7] . Transcriptome analysis using genome-wide microarrays [8–11] have also revealed complex dynamics of mosquito transcripts during infection and identified changes in expression of genes from diverse cellular processes , including ion binding , transport , metabolic processes and peptidase activity . Gene expression is also tissue-specific , with differences reported between midgut and salivary glands [10] . The ubiquitin-proteasomal system is one of the major protein degradation pathways in cells and has been shown to be important during flaviviral infection in mammalian cells [12] . Using a complex set of processes , it affects a myriad of cellular pathways [13] . Ubiquitin itself is a highly conserved 76-amino-acid protein that is highly conserved in sequence from yeast to human [14] . Ubiquitylation is brought about by a cascade of enzymes . E1 , ubiquitin activating enzyme , transfers activated ubiquitin to E2 , the ubiquitin conjugating enzyme . E3 , ubiquitin protein ligase , binds ubiquitin-charged E2 and the substrate , facilitating ligation of ubiquitin to the internal lysine residue on the substrate [15 , 16] . Substrate specificity is largely determined by the E3 ligase . The E3 family is characterised by the presence of the HECT ( Homologous to E6-AP Carboxyl Terminus ) [17] , RING ( Really Interesting New Gene ) finger [18] , U-box [19] and PHD ( Plant Homeo-Domain ) [20] or LAP ( Leukemia-Associated Protein ) finger domains [21] . E3 ligases can work as single or multi-subunit complex . The multi-subunit complexes include a RING-finger subunit and a member of cullin family that binds the RING-finger protein [22] . They also include structural adaptor proteins that link cullin to substrate recognition elements [23] . The ubiquitin-proteasomal pathway has been shown to be important in all stages of viral infection in mammalian cells , including entry , transcription , replication , assembly and exit of virus particles from the cell . A number of ubiquitin enzymes have been found to be involved in the flavivirus infection process . UBE1 has been shown to be important for dengue virus infection in primary human endothelial cells [24] . E3 ubiquitin ligase , CBLL1 ( HAKAI ) has been found to be important during WNV endocytosis [25] , but not in dengue entry [26] . Interestingly , there are also examples of viruses encoding ubiquitin ligases , which lead to evasion of host immunity , by degradation of immune proteins . For example , Kaposi’s sarcoma-associated herpes virus ( KSHV ) immediate-early transcription factor RTA encodes ubiquitin E3 ligase activity that targets IRF7 , a key mediator of type I interferon induction , for proteasome-mediated degradation [27] . As an alternate strategy , viruses also encode adaptors that recruit and redirect host E3 ligases to ubiquitylate host proteins , leading to degradation . Adenoviruses express proteins which recruit a cullin-based E3 ligase to target p53 for degradation [28] . Several paramyxoviruses limit the activity of interferons by targeting STATs for ubiquitylation and degradation via interaction of their highly conserved V proteins and host Cul4A RING E3 ligase [29 , 30] . Here we use an unbiased approach of transcriptome analysis by deep sequencing of the Culex cell transcriptome to identify genes that are differentially regulated following WNV infection . The results show multiple cellular pathways are involved during infection . Using mosquito infection studies in vitro and in vivo , we determine that the ubiquitin-proteasomal system plays a major role during WNV infection in mosquitoes . We also show that Culex cullin ( orthologous to mammalian Cul4A/B ) is induced by WNV infection and blocks Jak-STAT signalling , increasing WNV replication .
Hsu cells ( Culex quinquefasciatus cell line ) were infected with WNV ( NY99-4132 strain ) at multiplicity of infection ( MOI ) of 10 and total RNA was collected at 48 h post-infection ( hpi ) for high-throughput transcriptome sequencing as given in Methods . The experiment was conducted in duplicate and since Pearson correlation coefficients of 0 . 994 ( control ) and 0 . 991 ( WNV ) indicated close agreement between the biological replicates ( S1 Table ) , data from the replicates was combined for further analysis . The raw sequencing and downstream files were deposited at NCBI with accession code GSE60229 . A total of 265 unique transcripts were identified as differentially accumulated ( >2-fold ) between control and WNV-infected cells , with 130 transcripts up-regulated and 135 transcripts down-regulated after infection ( S2 Table ) . Ingenuity Pathway Analysis ( Ingenuity Systems , www . ingenuity . com ) performed on all differentially regulated genes indicated that various pathways and cellular processes were involved during infection , including the immune signalling , the cell-cycle/apoptosis and proteasomal pathways , metabolism genes and the cell transport machinery ( Fig 1A ) . Transcripts , which could not be functionally annotated were grouped as other or unknown . Genes in the ubiquitin-proteasomal pathway comprised 12% of differentially regulated transcripts and were selected for further analysis . To validate the RNA-Seq analysis , the relative abundance of seven differentially up-regulated transcripts in the ubiquitin-proteasomal pathway was determined by real-time RT-qPCR using target-specific primers . The results confirmed that all of the transcripts were up-regulated during infection but the fold-increase was consistently overestimated by RNA-Seq analysis ( Fig 1B ) . Previous studies have shown that ubiquitin-proteasomal pathway plays an important role in mammalian cells during flavivirus infection [25 , 31] . Studies have also shown that ubiquitin-related genes are upregulated following bacterial infection in mosquito cells [32] . Experiments were performed to determine the significance of the proteasomal pathway in WNV infected Culex cells . Hsu cells were pretreated with 0 . 1 , 1 or 10 μM MG132 ( specific proteasomal inhibitor ) from 1 h prior to WNV infection ( MOI 10 ) , and supernatant media and total RNA from cells were collected at 48 hpi . Real-time RT-qPCR using primers specific for WNV NS1 gene showed a dose-dependent decrease in viral replication , with the NS1 RNA level reducing by more than 80% following treatment with 10 μM MG132 ( Fig 2A ) . Cell viability assay showed no significant toxicity for MG132 at concentrations used in the assay ( S1 Fig ) . Plaque assays performed on the supernatant media showed a similar dose-dependent decrease in viral titers , with a 20-fold reduction in cells treated with 10 μM MG132 ( 30 pfu/ml ) compared to untreated controls ( 600 pfu/ml ) ( Fig 2B ) . These results suggest the proteasomal pathway is required for efficient WNV replication in Culex cells . A previous report suggests the proteasomal pathway has a role during entry of WNV into mammalian cells [25] . To test this in the mosquito system , Hsu cells were pre-treated with MG132 at various concentrations ( 0 . 1 , 1 and 10 μM ) , infected with WNV ( MOI 10 ) at 4°C for 30 min and then incubated at 30°C . Real-time RT-qPCR using NS1 primers conducted on total RNA from cells collected at 6 hpi showed no significant effect of MG132 treatment , indicating that the proteasome does not play a role during WNV entry to Culex cells ( Fig 2C ) . To determine whether the proteasomal pathway plays significant role post-entry , cells were infected with WNV and treated with 10 μM MG132 from 2 , 4 , or 6 hpi . Real-time RT-qPCR using NS1 primers conducted on total RNA from cells collected at 48 hpi showed ~80% reduction in viral RNA in all treated samples ( Fig 2D ) . Experiments were also performed to rule out significant contribution by extracellular viral RNA in real time RT-qPCR results ( S2 Fig ) . These results suggest involvement of the proteasomal pathway post-viral entry , possibly during viral transcription . Amongst the differentially expressed genes in the ubiquitin-proteasomal pathway , CPIJ010574 was highly up-regulated ( >20-fold in RNA-Seq and >12-fold in real-time RT-qPCR ) following WNV infection in Culex cells ( Fig 1B and S2 Table ) . CPIJ010574 is orthologous to mammalian Cullin4A and Cullin 4B ( ~70% amino acid identity with both ) ( S3 Fig ) and is therefore referred to here as CxCul4 . Cullin-RING ligases , such as Cullin4A , have been found to be upregulated during WNV infection of mammalian cells and implicated in various cellular processes such cell cycle regulation , signal transduction , DNA replication as well as viral replication [33] . Initially , in vivo validation of RNA-Seq analysis was performed using a mosquito infection model . Female Culex annulirostris mosquitoes were infected with WNV ( NY99-4132 strain ) by blood-feeding . Total RNA was collected from whole carcasses and dissected midguts at 24 hpi . Real-time RT-qPCR performed using CxCul4-specific primers showed approximately 4- and 12-fold increases in mRNA in carcass and midgut , respectively , compared with control mosquitoes fed on uninfected blood ( Fig 3A ) . Gene knock-down experiments were conducted to determine the significance of CxCul4 during WNV infection . Hsu cells were transfected with long dsRNA against CxCul4 , infected with WNV at 24 h post-transfection , and total RNA and supernatant media were collected at 48 hpi . As a control , cells were either left untransfected ( No dsRNA ) or were transfected with dsRNA against GFP ( GFP dsRNA ) . Real-time RT-qPCR showed a significant decrease ( >80% ) in CxCul4 mRNA , indicating efficient knock-down of the gene ( S4 Fig ) . There was also a significant decrease ( >75% ) in WNV NS1 RNA in cells treated with CxCul4 dsRNA , indicating a decrease in viral replication , compared to control ( No dsRNA and GFP dsRNA ) cells ( Fig 3B ) . Plaque assays conducted on supernatant media showed greater than 10-fold decrease in viral titers following CxCul4 knock-down ( 75 pfu/ml ) compared with No dsRNA ( 900 pfu/ml ) and GFP dsRNA ( 800 pfu/ml ) ( Fig 3C ) . In a parallel experiment , western blots performed using anti-NS5 antibody showed a significantly lower level of WNV NS5 in cells treated with CxCul4 dsRNA compared with control cells ( GFP dsRNA ) ( Fig 3D ) . To further investigate the role of CxCul4 , Hsu cells were transfected with a plasmid containing CxCul4 cloned under insect promoter ( OpIE2 ) , infected with WNV at 24 h post-transfection , and total RNA and supernatant media were collected at 48 hpi . Real-time RT-qPCR showed increased WNV NS1 levels ( >4-fold ) in cells transfected with CxCul4 compared with empty vector ( Control ) , indicating increased viral replication ( Fig 3E ) . Plaque assays conducted on supernatant media also showed significantly increased viral titer ( > 10-fold ) in cells over-expressing CxCul4 compared with control cells ( Fig 3F ) . In a parallel experiment , western blots using V5 antibody was performed on total cell lysates collected from Hsu cells transfected with CxCul4 to confirm exogenous expression level of CxCul4 ( Fig 3G ) . Combined , the above data suggest that CxCul4 plays a significant pro-viral role during WNV infection . Previous reports have shown that mammalian Cul4A plays a role in degradation of STAT2 by an ubiquitin-proteasomal-dependent pathway . To determine whether CxCul4 plays a similar role during WNV infection , Hsu cells were transfected with plasmid expressing CxCul4 and infected with WNV 24 h post-transfection . Total RNA was collected 48 hpi and real-time RT-qPCR was performed using primers for Vir1 , a reporter gene regulated by Jak-STAT pathway [34] , and CxVago , a reporter gene regulated by the TRAF-Rel2 pathway [35] . As expected , expression of both CxVir1 and CxVago was upregulated following WNV infection . However , whilst there was no significant difference in CxVago expression following CxCul4 overexpression , CxVir1 upregulation was suppressed significantly in CxCul4-overexpressing cells compared with empty vector transfected cells ( Fig 4A ) . Previous studies have shown that , once expressed , CxVago is secreted and activates the Jak-STAT pathway in Culex cells [7] . Therefore , Hsu cells were transfected with CxVago and the supernatant medium was collected 72 h post-transfection . Fresh Hsu cells were then transfected with the CxCul4 overexpression plasmid and , at 24 h post-transfection , the medium was replaced with medium containing CxVago to activate the Jak-STAT pathway . Real-time RT-qPCR conducted on total RNA collected 48 h after media replacement showed expression of CxVir1 was upregulated in control ( empty vector transfected ) cells but upregulation of CxVir1 was suppressed significantly in cells overexpressing CxCul4 ( Fig 4B ) . These results indicate that CxCul4 functions to inhibit the Jak-STAT pathway in Culex cells . To further confirm that CxCul4 is a negative regulator of Jak-STAT pathway , Hsu cells were transfected with a plasmid ( p6x2DRAF-Luc ) containing a Firefly luciferase reporter gene under Drosophila STAT-responsive elements , along with dsRNA against CxCul4 . After 24 h , cells were infected with WNV and luciferase activity was measured 16 hpi . The results showed that CxCul4 knockdown resulted in increased luciferase activity in WNV-infected cells ( Fig 4C ) . As a positive control , cells were stimulated with heat-inactivated E . coli for 1 h , which also showed increased luciferase activity . To determine whether Cul4 acts via the proteasomal pathway , Hsu cells were transfected with the CxCul4 overexpression plasmid and , at 24 h post-transfection , infected with WNV and treated simultaneously with either MG132 ( 10 μM ) or the ubiquitin activating enzyme E1 inhibitor PYR-41 ( 10 μM ) . Real-time RT-qPCR conducted on total RNA collected 48 hpi showed decreased CxVir1 expression in cells overexpressing CxCul4 compared with control ( empty vector ) cells in WNV infected cells . However , treatment of cells overexpressing CxCul4 with MG132 or PYR-41 resulted in significantly higher levels of CxVir1 expression compared with untreated cells . Real-time RT-qPCR for WNV NS1 showed there was a decrease in viral replication in cells transfected with CxCul4 and treated with MG132 or PYR-41 compared with the control ( CxCul4 transfection alone ) ( Fig 4D ) . To further determine significance of CxCul4 in STAT signaling , Hsu cells were transfected with dsRNA against CxCul4 . As a control , cells were either not transfected ( No dsRNA ) or transfected with dsRNA against GFP ( GFP dsRNA ) . Cells were infected with WNV 24 hours post-transfection and total RNA was collected 48 hpi . Real time RT-qPCR using CxVir1 primers showed increased expression after WNV infection which was further increased in cells transfected with CxCul4 dsRNA ( Fig 4E ) . CxVago on the other hand showed increased expression after WNV infection with no further increase in cells with CxCul4 dsRNA . To determine that action of CxCUl4 is via CxSTAT , cells were transfected with CxSTAT dsRNA along with dsRNA against CxCul4 . Cells were infected with WNV 24 hours post-transfection . Total RNA was collected 48 hpi and real time RT-qPCR using WNV NS1 primers showed significant increase in cells with CxSTAT dsRNA and decrease in cells with CxCul4 dsRNA . Cells containing both CxCul4 and CxSTAT dsRNA showed no decrease in WNV NS1 levels indicating that CxCul4 action required CxSTAT ( Fig 4F ) . These results confirm that pro-viral effect of CxCul4 is via ubiquitin-proteasomal pathway . To determine which WNV protein expression leads to overexpression of CxCul4 , Hsu cells were transfected with select individual WNV genes cloned in insect vector ( pIZ-V5/His ) . Total RNA was collected 48 hours post-transfection and real time RT-qPCR was performed using CxCul4 primers . As a control , real time RT-qPCR was performed using CxRpL32 primers . Western blot was performed on duplicate cells to confirm transfection and protein expression using anti-V5 antibody ( S6 Fig ) . The results showed up-regulation of CxCul4 expression in cells transfected with WNV-NS1 ( 2 fold ) and WNV-NS5 ( 3 fold ) genes ( Fig 5A ) . This indicates NS1 and NS5 may be involved in upregulation of CxCul4 after WNV infection in Culex cells . As a control , expression of CxCullin3 ( CPIJ011310 ) was measured after WNV infection or overexpression of WNV NS1 and NS5 . Results showed no significant change in expression level ( Fig 5B ) , indicating effect of WNV is specific to CxCul4 . Experiments were performed to determine whether WNV infection leads to STAT degradation via CxCul4 . Hsu cells were transfected with dsRNA against CxCul4 and infected with WNV at 24 h post-transfection . Total cell lysates were collected at 24 hours post-infection and Western blot was performed using anti-CxSTAT antibody . As a control , cells were either transfected with GFP dsRNA ( Control ) . The results ( Fig 5C ) showed cells infected with WNV showed decreased CxSTAT levels , which returned to baseline in cells transfected with dsRNA against CxCul4 . Although , we cannot rule out the possibility that reduced degradation of CxSTAT upon CxCul4 knockdown is due to lower virus replication , these results combined with other data indicate that WNV infection leads to degradation of CxSTAT via CxCul4 . To further establish this , Hsu cells were transfected with WNV-NS1 , WNV-NS5 or WNV-NS3 genes with or without transfection with dsRNA against CxCul4 . As a control cells were transfected with dsRNA against GFP . Cell lysates were collected 48 hours post-transfection . Western blot was performed using anti-CxSTAT antibody . The results showed decreased STAT levels in cells transfected with WNV-NS1 or WNV-NS5 ( Fig 5D ) . The levels returned to baseline in cells also transfected with dsRNA against CxCul4 . In WNV-NS3 transfected cells , there was no significant change in STAT level . These results indicate that WNV infection of Culex cells leads to degradation of CxSTAT via CxCul4 , activated by WNV NS1 and NS5 . CxCul4 , ortholog of mammalian Cullin4A/B , has previously been shown to be responsible for degradation of STAT via ubiquitin-proteasomal pathway . Cullin-RING ubiquitin ligases are responsible for ubiquitylation of target proteins . To determine whether degradation of CxSTAT via CxCul4 occurs by ubiquitylation , Culex cells were transfected with dsRNA against CxCul4 ( or GFP dsRNA as control ) . The cells were infected with WNV and cell lysates collected 48 hours post-infection . Immunoprecipitation was performed using anti-ubiquitin antibody , followed by Western blot using anti-ubiquitin and anti-CxSTAT antibodies . The results ( Fig 5E ) showed that STAT was co-immunoprecipitated with ubiquitin after WNV infection in control ( GFP dsRNA ) cells , indicating ubiquitylation of CxSTAT . CxSTAT was not co-immunoprecipitated with ubiquitin in cells transfected with dsRNA against CxCul4 . The results suggest that CxCul4 plays a major role in ubiquitylation of CxSTAT after WNV infection . To validate the significance of CxCul4 during viral infection of mosquitoes , female Culex annulirostris mosquitoes were microinjected with dsRNA against CxCul4 . As a control , mosquitoes were microinjected with dsRNA against GFP . The mosquitoes were infected with WNV ( NY99 strain ) by blood-feeding 24 h post-microinjection . At 10 days post-infection , saliva from individual mosquitoes was collected in a capillary tube ( see methods ) and total RNA was collected from whole mosquito carcass and midgut . As a parallel experiment , mosquitoes were collected at day 2 and day 10 post-infection . Western blot was performed on whole mosquito using anti-CxSTAT and anti-actin antibodies . Results showed increased levels of CxSTAT by day 2 post-infection . The levels decreased to baseline by day 10 post-infection in control mosquitoes ( GFP dsRNA ) ; however in mosquitoes with CxCul4 dsRNA , STAT levels remained high ( Fig 5E ) . Real-time RT-qPCR results showed efficient knock-down ( >60% ) of CxCul4 mRNA in carcass after Cul4 dsRNA microinjections ( Fig 6A ) . The results also showed a decrease ( >50% ) in viral RNA ( NS1 ) in whole mosquito carcass ( Fig 6B ) and midgut ( Fig 6D ) in CxCul4 dsRNA-injected mosquitoes compared with the control , confirming pro-viral effect of CxCul4 . Interestingly , real-time RT-qPCR results also showed a significant increase in CxVir1 mRNA expression in carcass ( 3-fold ) ( Fig 6C ) and midgut ( 5-fold ) ( Fig 6E ) in CxCul4-knockdown mosquitoes , indicating increased Jak-STAT signaling . Plaque assays performed on mosquito saliva showed a significantly lower virus titer in mosquito saliva microinjected with CxCul4 dsRNA ( 88 pfu/mosquito ) compared with control ( GFP dsRNA ) ( 429 pfu/mosquito ) ( Fig 6F ) , indicating a lower likelihood of virus transmission by CxCul4 silenced mosquitoes . It should be noted that saliva from 3 mosquitoes with CxCul4 dsRNA injection showed higher viral titers ( similar to controls ) . Further analysis showed that CxCul4 was not efficiently knocked down in these three mosquitoes ( Fig 6A , inset ) . These results further validate that CxCul4 has a pro-viral effect by inhibiting Jak-STAT pathway during WNV infection in mosquitoes . It should also be noted that ~20–30% of mosquitoes from each group did not show any virus in the saliva , possibly indicative of vector competence of these mosquito species .
A number of studies have also been performed to identify differentially regulated genes during flavivirus infection in mammalian cells [31 , 36–39] . Transcriptome studies using genome-wide microarrays or next-generation deep sequencing of the transcriptome , as well as genome-wide siRNA studies , have revealed a number of pathways to be important during infection process [25 , 31] . These include immune pathways , apoptotic pathways , cellular transport proteins , proteasomal pathways among others . Previous studies using microarray analyses to characterize the transcriptional response of mosquitoes to flavivirus infection have shown that a multitude of pathways appear to be involved , including the immune , cell-cycle/apoptosis , metabolic and proteasomal pathways as well as other cellular processes such as the transport machinery [10 , 11] . Here we adopted an unbiased approach using high-throughput deep sequencing of the mosquito cell transcriptome during a synchronized ( high multiplicity ) infection with WNV . Our data also identified differentially regulated genes from a number of cellular pathways of which we selected ubiquitin-proteasomal pathway for further functional analysis . Proteasomal inhibition has been shown previously to affect replication of a wide range of viruses in mammalian cells , including herpesviruses [40] , poxviruses [41] , hepadnaviruses [42] , adenoviruses [43] , influenza viruses [44] , retroviruses [45] , coronaviruses [46] , paramyxoviruses [47] and rotaviruses [48] . Microarray studies have also shown that several ubiquitin-related genes are induced following infection in mosquito cells [32] . The ubiquitin-proteasomal pathway is one of the major pathways involved in post-translational modifications , leading to degradation of proteins . It has also been implicated in various stages of viral infection in mammalian cells including virus entry , replication and exit [49] . The largest family of ubiquitin ligases ( E3 ) is a group of proteins that contain a RING domain , a structural motif comprising eight cysteine and histidine residues that form the interface with E2 ( ubiquitin conjugating ) enzyme [18] . Substrate specificity is determined by variant RING ubiquitin ligases and also by multiprotein complexes that contain a conserved RING protein . Cullin RING ubiquitin ligases are one such family of proteins , containing eight mammalian members , each with different substrate specificity [22] . These proteins have been implicated in many cellular processes including cell-cycle regulation and cell signaling . Because of this , cullin RING ubiquitin ligases are frequently targeted by viruses to evade host immunity [33] . Viruses redirect these ligases to select specific host proteins for degradation in order to prevent the host response and promote viral replication and dissemination . For example , viruses have been shown to use cullins to prevent cellular apoptosis by degrading p53 [50] or inhibit the innate antiviral response by blocking interferon signaling [30] . In mammals , interferon is a key component of a major innate defensive response against invading viruses by activating Jak-STAT pathway and the expression of antiviral genes in neighbouring cells [51] . Some members of the Paramyxoviridae have been shown to hijack cullin4A ubiquitin ligase complexes to overcome the interferon response , by promoting degradation of STAT proteins . Specifically , the V protein of simian virus 5 causes degradation of STAT1 [29] , while the parainfluenza virus V protein causes degradation of STAT2 via the Cul4A substrate adaptor DDB1 [52] . Recently , Culex Vago , which is induced in mosquitoes in response to WNV infection , was found to be functionally similar to mammalian interferon in that it is activated via the TRAF-NF-κB pathway and , upon secretion , activates the Jak-STAT pathway and an antiviral response in neighbouring cells [7 , 35] . Our results here suggest that Culex STAT is regulated by Culex Cul4 ( mammalian Cul4A/B ortholog ) , which has proviral activity during WNV infection , and that WNV may target STAT for degradation by inducing Culex Cul4 via NS1 and NS5 proteins . Suppression of the mammalian interferon-mediated Jak-STAT pathway has been shown to be a common property of vector-borne flaviviruses . Dengue NS5 protein has been shown to target STAT2 for degradation [53] and WNV NS5 suppressed phosphorylation of STAT1 [54] , leading to decreased signaling in mammalian system . Interestingly , our results suggest that the mosquito Jak-STAT pathway is also targeted by WNV via cullin RING ubiquitin ligase . Although , we have identified WNV NS1 and NS5 proteins to be responsible for upregulation of Cul4 , the exact mechanism of this activation remains unknown . WNV NS1 and NS5 proteins have significantly different protein sequences , however both are involved in formation of viral replication complex . Studies are currently underway to determine whether this plays any part in the described mechanism . Our in vivo results also suggest that CxCul4 proviral activity may be mainly localized in the mosquito midgut and knockdown of CxCul4 increases Jak-STAT signaling in the midgut , thus decreasing the overall viral load in the body and , in turn , the saliva of these mosquitoes . Our unbiased transcriptome analysis also indicated that other genes involved in the ubiquitin-proteasomal pathway , including E1 , UBC and other E3 ligases , as well as ubiquitin-specific proteases are differentially expressed during WNV infection , suggesting that this pathway has a critical role in the invertebrate response to infection and/or the viral host evasion strategy . Our results suggest that proteasomal pathway ( and CxCullin4 ) does not play any role in WNV entry in mosquito cells ( S5 Fig ) . This is different than previously reported mechanism in mammalian cells [25] . This may be due to different mechanisms of viral entry in mammalian cells versus mosquito cells . It is also interesting that a number of reads from our data did not map to the published Culex quinquefasciatus transcriptome . Many of these transcripts may be expressed from unannotated genes which contribute to novel host-pathogen interactions in invertebrates . Here we show that cullin RING ubiquitin ligase ( CxCul4 ) plays an important role during WNV infection of Culex mosquitoes and dengue infection in Aedes albopictus cells ( S7 Fig ) . In vitro and in vivo experiments show that Culex Cul4 ( mammalian Cul4A/B ortholog ) is induced by WNV infection and acts as a proviral factor by ubiquitylation of STAT in Culex mosquitoes , thus inhibiting the antiviral response . It remains to be seen whether mammalian Cul4A/B act in a similar manner during viral infection . This study opens up a new avenue of research in viral evasion of the mosquito immune system .
Hsu ( Culex quinquefasciatus ) and RML12 ( Aedes albopictus ) cells were maintained at 28°C in Leibovitz's L-15 medium ( Gibco #11415 ) containing 10% tryptose phosphate broth solution , 15% heat-inactivated fetal bovine serum , and 1% penicillin-streptomycin solution . West Nile virus ( NY99–4132 strain ) was used for the study . C6/36 ( Aedes albopictus ) cells were maintained in RPMI medium at 28°C and were used to propagate the virus . Vero cells maintained in EMEM at 37°C were used for plaque assays . Total RNA was collected from Hsu cells ( control and WNV infected ) using Qiagen RNeasy kit following the manufacturer’s instructions . RNA was quantified and checked for quality using a bioanalyser . The experiment was conducted in duplicate with two RNA-Seq libraries generated for WNV-infected cells and two for uninfected control cells . Sequencing was conducted in a single lane of an Illumina HiSeq2000 ( Micromon Facility , Monash University , Australia ) , generating more than 80 million reads per sample ( S1 Table ) . Sequence reads ( paired , 100 bp ) were mapped using Tophat v2 . 0 . 8 ( http://ccb . jhu . edu/software/tophat/index . shtml ) [55] to Culex quinquefasciatus ( Johannesburg strain , CpipJ2 . 1 ) transcripts available from VectorBase ( http://www . vectorbase . org ) [56] . Differential isoform expression analysis was conducted using Cufflinks v1 . 3 . 0 ( http://cole-trapnell-lab . github . io/cufflinks/ ) [57] . Analysis of RNA sequencing quality was performed with FastQC version 0 . 10 . 1 ( http://www . bioinformatics . babraham . ac . uk/projects/fastqc/ ) using the default options . Pathway analysis of differentially regulated transcripts was performed using Ingenuity Pathway Analysis v9 . 0 ( Ingenuity Systems , www . ingenuity . com ) . Total RNA was extracted from cells using the Qiagen RNA extraction kit according to the manufacturer's protocol . Reverse transcription was performed with random hexamer primers using the First Strand Synthesis kit ( Invitrogen ) . Real-time RT-qPCR was performed using gene-specific primers . As an internal control , real-time RT-qPCR was also performed using the housekeeping gene , RpL32 . The control was set arbitrarily at 1 and fold-increase over control was calculated by the ΔΔCt method . The experiments were conducted at least three times , each in triplicates . The results were plotted in graph format as mean ± SD . MG132 and PYR-41 ( Sigma-Aldrich ) were used at the described concentrations dissolved in DMSO . For controls , cells were treated with DMSO . For pre-treatment , cells were initially treated with MG132 , followed by WNV infection for one hour . This was followed by replacing the medium with medium containing MG132 at appropriate concentrations . Cells were lysed in RIPA lysis buffer ( 25 mM Tri-HCl , 150 mM NaCl , 1% NP-40 , 0 . 1% SDS ) containing protease inhibitor ( Halt protease inhibitor cocktail , Pierce ) . The cell lysates were collected by centrifugation at 16 , 000 × g for 10 min at 4°C . Protein samples ( 10 μg ) were loaded onto polyacrylamide gradient gels ( 4–12% ) . After electrophoresis and transfer to nitrocellulose membranes , proteins were blotted using anti-Culex STAT ( rabbit polyclonal against peptide VVIVHGNQEPQSWATITWDNAFADINRV PFHVPDKVSWNLLAEALNTKYRASTGRSMTQENMHFLC ) , anti-WNV NS1 , anti-WNV NS5 [58] or anti-V5 ( Invitrogen ) antibody followed by anti-rabbit or anti-mouse secondary antibodies . After adding substrate , the membrane was exposed to film to detect protein levels . Anti-β-actin antibody ( Abcam ) was used in immunoblots as a loading control . Plaque assays were performed as previously described [7] . In brief , supernatant media from cells infected with WNV ( 10-fold dilutions ) were added onto confluent Vero cell monolayers in 6-well plates . After 1 h incubation at 37°C , the cells were overlaid with medium containing agar . Plaques formed within 72 hpi were counted and the results were plotted graphically . The experiments were conducted at least twice , each with duplicates . Gene-specific dsRNA ( ~400 nt ) were prepared using the MEGAscript RNAi kit according to the manufacturer's protocol . dsRNAs were transfected into Hsu cells using Cellfectin according to a previously described protocol [7] . dsRNA against green fluorescent protein ( GFP ) was used as a knock-down specificity control . STAT activation experiments were performed as previously described [59] . The STAT reporter plasmids were kindly provided by Prof . Martin Zeidler ( University of Sheffield ) [60] . In brief , Hsu cells were transfected with STAT reporter plasmid p6x2DRAF-Luc ( multimerised Drosophila STAT-responsive element with a Firefly luciferase reporter ) and a control plasmid , pAct-Renilla ( Renilla luciferase gene under control of the Drosophila actin 5C promoter for constitutive expression ) . Cells were also transfected with dsRNA against Cul4 or GFP . Cells were stimulated with heat inactivated E . coli , WNV ( MOI = 1 ) or PBS ( Control ) for 1h . Luciferase activity was measured at 16 h post-stimulation . Heat-inactivated bacteria were prepared by incubating 1μl of E . coli DH5α in 5 ml LB medium at 37°C for 16 h . Cells were harvested by centrifugation and resuspended in 0 . 5 ml of PBS . Heat inactivation was achieved by incubating the cells for 10 min at 90°C . 1 μl of the suspension was used per well . Hsu cells under various conditions were treated with MG132 ( 1 μM ) for 6 hours to prevent degradation of ubiquitylated proteins . Immunoprecipitation was performed using Pierce Crosslink IP kit ( Thermo Scientific ) following manufacturer’s protocol . Briefly , cells were lysed using lysis buffer and after preclearing were incubated with anti-ubiquitin antibody ( Abcam ) crosslinked to beads in column for 18 hours . Flow-through was collected and immunoprecipitate was eluted using elution buffer . The samples were separated on polyacrylamide gel and Western blot was performed using anti-ubiquitin and anti-CxSTAT antibodies . Culex annulirostris mosquitoes were maintained in a diurnal cycle ( 12h/12h ) with temperatures alternating for 23 and26°C and 65% humidity . Three to five day-old female mosquitoes ( n = 40 ) were blood-fed on chicken skin membranes with WNV ( 1 . 13×10∧6 pfu/ml ) or 199 medium ( as control ) and the mosquitoes were incubated at 25°C and 65% humidity in an environmental cabinet ( Thermoline Scientific , Smithfield , Australia ) with a wet cotton pad ( 10% sucrose solution ) provided daily as a food source . At 24 hpi , surviving females were collected for analysis . For this , mosquito midguts were dissected and homogenized using a bead-beater . RNA extracted using the RNeasy kit ( Qiagen ) by pooling four samples and was used for real-time RT-qPCR as described above . Three to five day-old female mosquitoes ( n = 40 ) were microinjected with dsRNA against CxCul4 or GFP ( 200 ng/mosquito ) . Mosquitoes were blood-fed with WNV ( 1 . 13×10∧6 pfu/ml ) one day later and were maintained in a diurnal cycle ( 23/26°C ) in an environmental cabinet . At 10–12 days post-infection , saliva was collected in capillary tubes containing 5 μl FCS for 10 min using a protocol described previously [61] . Midguts were dissected and homogenized using a bead-beater . RNA was extracted using the RNeasy kit from the midgut and carcass of individual mosquitoes and was used for real-time RTqPCR as described above . Saliva from each mosquito was diluted in 300 μl of L-15 medium and was used to determine viral titer by plaque assay as described above . Standard error of the mean ( SEM ) was calculated and data analyzed using the non-paired Student's t-test for single mean comparisons .
|
Mosquitoes are responsible for transmitting a large number of human and livestock viruses , like West Nile , dengue and Japanese encephalitis viruses . Infection of female mosquitoes with these viruses during blood feeding elicits an immune response . It is not known how the viruses manage to replicate in spite of this antiviral response . We used an unbiased transcriptome sequencing approach to identify genes differentially regulated after WNV infection resulting in 265 transcripts from various cellular pathways . Ubiquitin-proteasomal pathway , responsible for protein degradation , was found to be important during viral infection in mosquito cells . Using in vitro and in vivo infection models , we identified Culex Cul4 to be acting as pro-viral protein , increasing viral titers . Knockdown of Cul4 in Culex mosquitoes decreased viral titers in mosquito saliva . Identification of this novel immune evasion mechanism adopted by WNV provides new insights into transmission of arbovirus and interaction of WNV with its mosquito vector .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Cullin4 Is Pro-Viral during West Nile Virus Infection of Culex Mosquitoes
|
We used the bioorthogonal protein precursor , homopropargylglycine ( HPG ) and chemical ligation to fluorescent capture agents , to define spatiotemporal regulation of global translation during herpes simplex virus ( HSV ) cell-to-cell spread at single cell resolution . Translational activity was spatially stratified during advancing infection , with distal uninfected cells showing normal levels of translation , surrounding zones at the earliest stages of infection with profound global shutoff . These cells further surround previously infected cells with restored translation close to levels in uninfected cells , reflecting a very early biphasic switch in translational control . While this process was dependent on the virion host shutoff ( vhs ) function , in certain cell types we also observed temporally altered efficiency of shutoff whereby during early transmission , naïve cells initially exhibited resistance to shutoff but as infection advanced , naïve target cells succumbed to more extensive translational suppression . This may reflect spatiotemporal variation in the balance of oscillating suppression-recovery phases . Our results also strongly indicate that a single particle of HSV-2 , can promote pronounced global shutoff . We also demonstrate that the vhs interacting factor , eIF4H , an RNA helicase accessory factor , switches from cytoplasmic to nuclear localisation precisely correlating with the initial shutdown of translation . However translational recovery occurs despite sustained eIF4H nuclear accumulation , indicating a qualitative change in the translational apparatus before and after suppression . Modelling simulations of high multiplicity infection reveal limitations in assessing translational activity due to sampling frequency in population studies and how analysis at the single cell level overcomes such limitations . The work reveals new insight and a revised model of translational manipulation during advancing infection which has important implications both mechanistically and with regards to the physiological role of translational control during virus propagation . The work also demonstrates the potential of bioorthogonal chemistry for single cell analysis of cellular metabolic processes during advancing infections in other virus systems .
Much of our understanding of the molecular mechanisms operating during virus infection comes from population studies . The classic single-step virus growth cycle , the identification and characterisation of virus encoded transcripts and proteins , and the associated mechanisms governing temporal regulation of their production and turnover have been founded on population studies of infected cells in culture systems [1] . However , it is becoming clear in many fields that while analysis of the average behaviour in total infected cell populations is vital , information at the individual cell level is also critical for a true understanding of the processes governing the outcomes of infection . Such analyses may support and refine conclusions from population studies , but can also yield results which are not accounted for in population studies and provide conceptually new mechanistic insight [2–5] . In this regard , while much effort has focussed on analysis of levels and variations in transcription patterns at the single cell level , we know much less with regard to protein synthesis . All viruses manipulate the host cell translational apparatus to promote the synthesis of their proteins and to supress cellular antiviral responses . At the same time , cells modulate both their qualitative translational output and their translational apparatus in the attempt to suppress virus replication [6–15] . Thus , overall infected cell protein synthesis results from a complex and temporally regulated interplay of multiple distinct translational objectives for the host and virus , in addition to selective controls on the abundance and localisation of individual protein species . However , global protein synthesis has been almost universally studied by population methods such as gel electrophoresis and autoradiography , Western blotting or mass spectrometry , potentially masking dynamic and diverse individual cell behaviour [16–21] . A complete understanding of infected cell protein metabolism requires a parallel approach to spatial aspects of protein synthesis and temporal alterations in these processes at the single cell level during the progression of infection . Traditional steady-state analysis using antibodies , or gene fusion to fluorescent proteins , provide powerful tools for the investigation of individual proteins [22–24] . However , global spatial analysis requires a different approach . Recent advances in bioorthogonal chemistry [25] have facilitated the development of new techniques based on the in vivo incorporation of metabolic precursors containing designed chemical end-groups . Subsequent highly specific covalent bond-forming reactions , commonly termed “click chemistry” , then link the macromolecular products incorporating these precursors to capture reagents via a dedicated , paired end-group [26–29] . The chemical pairs most routinely used are the azide- and alkyne moieties which are small , inert and can be introduced to a variety of precursors [28 , 30–33] . Thus , for protein synthesis it is possible to label newly synthesised proteins over a specified timeframe using the methionine analogues homopropargylglycine ( HPG ) or azidohomoalanine ( AHA ) , and then covalently couple those de novo synthesised proteins to fluorescent capture reagents . This enables analysis either biochemically e . g . by SDS-PAGE and in-gel fluorescence , or spatially to simultaneously visualise the overall levels and localisation of the “translatome” by microscopy [32 , 33] . We recently used these techniques to provide the first spatial and kinetic analysis of bulk newly synthesised proteins during a single-step replication cycle of herpes simplex virus ( HSV ) , providing new insight into protein synthesis and trafficking , including the formation of novel nuclear depots into which newly synthesised host and viral proteins trafficked [34] . Here we visualise global protein translation during HSV cell-to-cell transmission . Current models indicate that HSV progressively suppresses infected cell translation , by multiple processes but particularly involving the structural component vhs , an RNase that is the product of the UL41 gene [14 , 35–39] . Our results agree with the considerable data from several laboratories that vhs is a key determinant of translational suppression during HSV infection . However from spatial analysis at the single cell level during virus spread we now demonstrate a very early biphasic switch , combining efficient suppression of translation with subsequent recovery to normal levels and show that such oscillations would be averaged out to the unimodal kinetic that is currently proposed from population studies . We also show a highly correlated relocalisation to the nucleus of eIF4H , a factor known to interact with vhs [40 , 41] in cells exhibiting translational suppression . Moreover , we show that a single particle of HSV-2 was sufficient to promote translational shutdown , a result not approachable by current methods . Together with other results , this work has significant implications for our understanding of the mechanism ( s ) and role of translational control , leading to a new interpretation of that would not be gained from population studies . The work also demonstrates the broad potential of chemical biology for spatial and biochemical studies of virus infection , applicable to other viral and indeed bacterial systems .
A schematic indicating the principle of HPG incorporation into proteins and then ligation with azide-linked fluorophores is summarised in S1a Fig . Much accumulated data has demonstrated that HPG has no effect on global rates of protein synthesis nor protein degradation [42–48] . In examining parameters for virus transmission studies , we first pulsed cells with HPG for 30 min at 1 hr after high multiplicity infection with HSV ( MOI 10 ) or mock infection and analysed protein synthesis by SDS-PAGE and in-gel fluorescence ( S1b Fig ) . As a control we also included cycloheximide ( CHX , 100 μg/ml ) to block de novo protein synthesis . The results demonstrate efficient labeling of uninfected cell proteins ( lane 7 ) and almost complete inhibition by CHX ( lane 5 ) , demonstrating as expected that HPG incorporation reflects de novo protein translation . In population experiments such as this ( which analyse the total cells in the sample ) , we observed very little change at this early time in overall levels of translation in HSV infected cells compared to uninfected cells ( c . f . lanes 7 and 8 ) . Using a 30 min labeling interval as a benchmark , we examined different labeling intervals to optimise spatial analysis of translation in individual cells ( S1c Fig ) . Although shorter pulse times ( 5–10 min ) gave a detectable signal , we selected 30 min as the standard interval for all subsequent analyses since it gave a good signal and dynamic range . Typical results showing spatial analysis of a field of uninfected cells pulse-labeled with HPG for 30 min is shown in Fig 1a , with an individual cell shown at higher magnification in the right-hand panel . Pronounced localisation of newly translated proteins to the nucleus and nucleolus together with distribution throughout cytoplasmic organelles reflects previous observations by ourselves and others [34 , 42–44 , 47] . Overall incorporation levels within individual cells were relatively homogeneous . The standard protocol for analysis of translation during HSV cell-to-cell spread is shown in Fig 1b . Confluent cell monolayers were infected at extremely low multiplicity ( approximately 1 in 4000 cells infected ) , and infection allowed to proceed after addition of neutralising antibody to prevent secondary infection from free virus . After approximately 20–24 hr , cultures were pulse-labeled with HPG for 30 min and analysed by simultaneous click chemistry and immunofluorescence . Multiple early plaques were then inspected with each showing identical features regarding the main outcomes described below . Advancing infections were imaged with x63 or x40 objectives encompassing approximately 40 or 120 cells respectively on each field and at least 10 fields were evaluated . Each panel of a figure is representative of these fields . Typical images , analysing active protein synthesis in relation to virus spread ( in this case with HSV-2 ) are shown in Fig 1c . The extent of infection is marked in this case by antibody to the late capsid protein VP5 ( red channel ) . Total cell nuclei ( DAPI staining ) are shown in the blue channel and active protein synthesis during the labeling interval is shown in the green channel . The merged channels are shown in the right hand panel . Various zones discussed in the text are delineated by white dotted lines . A subset of individual cells are labeled for spatial reference . The extent of the focus of infection can therefore be seen , with typical abundant nuclear accumulation of VP5 together with efficient protein synthesis in those cells ( HPG , green channel ) . Strikingly , immediately surrounding this area of infection there is an extensive area of cells , almost completely surrounding the central focus , where protein synthesis has been suppressed to virtually background levels . Cells in this zone appear morphologically normal ( phase channel ) and as indicated are not yet synthesising detectable levels of virus protein , at least VP5 in this case . Cells more distant to this zone ( i . e . external to reference cells 1–4 ) exhibit normal levels and distribution of protein synthesis that would be expected of uninfected cells . VP5 , while an abundant protein and a sensitive marker for infection , is nevertheless expressed from early to late times in infection . We repeated these experiments using the immediate-early ( IE ) protein ICP4 as a marker for very early infection and expression of viral proteins , also examining HSV-1 in this case ( Fig 1d ) . Here infection is spreading from the top of the panel downward . Again a broad zone , representing almost the entire front of the advancing infection , exhibited pronounced suppression of protein synthesis ( green channel , representative cells marked 1–4 within the dotted lines ) . Cells in the lower region of the panel , more distant from the advancing front , again exhibited normal levels of protein synthesis . Delineating cells on the basis of protein synthesis i . e . within the dotted lines , it can be seen that cells in the shutoff zone express either undetectable or very low levels of ICP4 , while cells at the top region exhibit much more abundant ICP4 and importantly overall protein synthesis levels have recovered to virtually normal . A final example of this translational shutoff at the extreme front of an advancing infection , together with recovery in the immediately adjacent cells in shown in Fig 1e , in this case with infection marked by a glycoprotein gB . While the conclusions were clear from visual inspection , to spatially delineate the translational shutoff area in a more quantitative manner , we used the DAPI signal to create a mask for individual cell nuclei and then quantitated the protein synthesis signal within individual cells of the entire field , normalised for nuclear area . We set a threshold for significant shutoff at 30% or below the maximum for a field ( i . e . an approximately 3-fold reduction , colour coded yellow ) . All other cells were coded pink . This method is somewhat conservative since translation levels in uninfected cells ( i . e . in mock-infected monolayers ) were relatively consistent , rarely exhibiting levels below 50–60% of the maximum in the field . Setting the threshold at 30% may miss cells that were in partial shutoff but this does not materially alter our conclusions . Thus , individual cells outside the shutoff zone ( right hand side of the field ) exhibited relatively limited variation . Cells within the shutoff could be clearly observed ( coded yellow ) with cells interior to this , and now gB positive , exhibiting restored levels of protein synthesis at approximately similar levels to those on the other side of the shutoff zone at the right hand side of the field . Cells at the extreme front of a spreading plaque must be either uninfected or at the very earliest stages of infection . Cells that are progressively located inward from the boundary are generally at a later stage of infection . The simplest interpretation of these results therefore is that there is a biphasic suppression and recovery of protein synthesis . It is not likely that cells in the shutoff zone are uninfected . Rather cells at the leading edge , exhibiting efficient suppression of global translation over a broad spatial front , are likely at the very earliest stages of infection where ICP4 expression is not yet detectable , or is at very low levels . Importantly , cells immediately adjacent on the inward side have recovered from this shutoff to restore comparatively normal levels of total protein synthesis and progressively increasing accumulation of steady state ICP4 , while cells immediately adjacent on the outward side represent cells which are not yet infected . Although it is presently impossible to precisely analyse it , considering the extent of infection and spatial distribution of normal levels of synthesis immediately adjacent to cells with almost total suppression , the biphasic suppression and recovery is likely to occur over a very short period of time ( see below ) . To examine the degree of translational shut off at a single cell level in a more quantitative manner , we repeated our standard analysis , additionally directly comparing HSV-1 and HSV-2 ( Fig 2 ) . The results are quantitated for individual cells within zones delineated on the basis of a ) translational levels , b ) VP5 levels and c ) spatial relationship . The selective regional suppression was again extremely efficient ( Fig 2a ) . Almost every cell in the shutoff zone , zone 2 ( Z2 , in merged panel ) , directly adjacent to the advancing infected zone ( Z1 ) , exhibited very substantial shutoff of translation . Translation within cells in the advancing focus itself ( Z1 ) , even those cells immediately adjacent to the shutoff zone , showed virtually normal levels of overall protein synthesis , comparable to uninfected distal cells in Z3 , i . e . those exterior to Z2 . The quantitative analysis showed that while there was some cell-to-cell variability in the uninfected zone , this was modest and few if any cells were more than 1 . 5–2 fold different from the mean . By contrast translation in cells in the shutoff zone for HSV-1 were reduced to 5–15% of the mean of those in the uninfected zone , and there was a more uniform and almost complete shutoff for HSV-2 . Importantly cells within the infected focus Z1 , had returned to absolute levels only marginally lower than those in the uninfected zone . Generally large areas of an advancing infection exhibited very efficient regional shutoff although this did not always result in a complete annulus of shutoff around the infection ( likely due to inherent asynchrony in phasing , see below and Discussion ) , Nevertheless such a spatial distribution with an extended ring of translational shutoff could frequently be observed ( see e . g . Fig 1c , S2a Fig ) We examined spatial features of translational shutoff in several other cell types with generally similar results but noted one distinctive feature ( S2 Fig ) . As indicated above , in Vero cells at 24 hr post infection ( p . i . ) , translational shutoff at the advancing edge of virus spread was efficient , in this case forming a distinct and almost complete annulus around the advancing focus ( S2a Fig Vero , asterisked cells ) . By contrast in parallel in human skin keratinocytes ( HaCaT cells ) , while plaque formation was clearly advancing , shut off was quite difficult to discern , with no clear distinction at the periphery of the focus versus more distant uninfected cells at the perimeter of the field ( S2 Fig HaCaT ) . However , when the pulse was delayed until later in the advancing keratinocyte infection , although general incorporation in uninfected distal cells remained similar , pronounced shutoff was now clearly discernible in cells immediately surrounding the now expanded focus of infection ( S2b Fig , HaCaT 50 hr p . i . ) . While there are several possible explanations for these observations , including paracrine effects progressively influencing distant uninfected cells ( see Discussion ) , these results may reflect distinct phases of temporal modulation of translational control during infected cell to uninfected cell transmission , a process that would not be possible to observe with single-step population analysis . Overall translation declines progressively during HSV infection , with several mechanisms contributing to such decline including the vhs function ( UL41 ) [35 , 37 , 49–57] . The kinetics and the extent depends on several factors including virus strain , with HSV-2 promoting a more rapid decline , consistent with our results above . Nevertheless , this translational decline is generally explained as a continuum within any one cell , of progressive unimodal early repression dependent upon vhs ( among other factors ) and continued later repression by independent mechanisms , summed across the population . Our conclusion from single cell analysis of a very rapid early but biphasic programming of suppression and recovery of overall translation is not accounted for in current models . In population based single-step experiments ( S1 Fig , high MOI infection and SDS-PAGE ) , we did not observe significant early shutoff of translation . We repeated these experiments using high MOI infection and analysing protein translation in the same cultures , either by imaging analysis at the single cell level or by SDS-PAGE and in-gel fluorescence . The protocol ( Fig 3a ) was adopted after evaluating procedures to obtain as synchronised an infection as possible ( MOI 10 with pre-incubation at 4°C ) , at the earliest time frame and with efficient labeling sensitivity as possible ( see experimental procedures ) . Cells were processed and colour coded as for Fig 1 . In mock-infected cells we observed the typical intracellular distribution of newly synthesised proteins and minor cell to cell variability ( Fig 3b , mock , HPG and right hand panel ) . In contrast , in HSV infected cells , we observed distinct variation in translational levels at 1 hr after infection , with three classes of spatially interspersed cells ( Fig 3b , HSV-1 1 hr ) . In class A cells ( labeled A ) , there was little alteration in levels of protein synthesis ( HPG and right hand panel ) from those seen in mock-infected cells and either no , or barely discernible , ICP4 expression . In class B cells , there was a pronounced translational shutoff with such cells expressing low but clearly detectable levels of ICP4 . In other cells ( class C ) , overall protein synthesis levels were at least as high as those in mock-uninfected cells . In these cells , the qualitative patterns of localisation of newly synthesised proteins exhibited a combination of those seen in uninfected cells ( including abundant nuclear and nucleolar accumulation ) together with additional distinct features termed NPDs , which are hallmarks of advancing infection [34] . These cells showed increased levels of ICP4 accumulation compared to class B cells . Taking into account the results during HSV cell to cell spread after low MOI infection , we interpret these data as follows . Class A cells represent cells that are either uninfected , or formally at the extremely early stages of infection , even prior to translational shutoff and prior to detectable ICP4 levels . Although at MOI 10 essentially all cells ( 99 . 995% ) will be infected at some particle level , infection will show some degree of asynchrony and it is difficult to discriminate between these two possibilities . However , we think it is more likely that class A cells are infected but not yet exhibiting significant translational shutoff or ICP4 synthesis . Class B cells ( approximately 25% ) represent cells which are at some stage within the suppression/recovery cycle in protein synthesis , exhibiting much reduced levels of protein synthesis together with limited levels of ICP4 . Class C cells are then further advanced , exhibiting increased levels of ICP4 , but also increased levels of total protein synthesis compared to class B cells . Consistent with this , pulsing later in infection ( 4 hr ) the vast majority of cells now exhibited totals levels of protein synthesis that were similar to uninfected cells , together with increased and more uniform levels of ICP4 ( Fig 3b , 4 hr ) . As expected , these cells also showed the advancing features of NPD formation ( white arrowheads ) combined with abundant nuclear import [34] and early stages of replication compartment formation . However in absolute terms , the majority of total protein synthesis at this stage still represented host protein synthesis , as seen from the parallel analysis of the same cells at the same time point by SDS-PAGE and in-gel fluorescence ( Fig 3c ) . Thus while at a single cell level , a significant fraction of the cells exhibited very substantially reduced levels of total protein synthesis at 1–1 . 5 hr , this was not evident from SDS-PAGE analysis of active protein synthesis and the considerable bulk of synthesis represented host proteins ( exemplified by representative host species ( Fig 3c , cf . lanes 1–3 , host species labeled H1-4 ) . Analysis of advancing cell-cell transmission therefore reveals a process , whereby efficient shutoff and restoration of translation takes place in a rapid manner that is not as readily experimentally assessed during spatial analysis of high MOI infection and or by biochemical analysis at the population level . We constructed a modelling simulation which accounts for this contrast ( Fig 3d ) . The simulation , while necessarily oversimplified , allows variable inputs for several parameters including; overall time for all cells to become infected ( e . g . , 45 min in Fig 3d ) , rate of infection ( degree of asynchrony , e . g . 5% of cells every 2 . 25 min ) , the overall translational cycle duration and the restoration level as a percentage of that in uninfected cells ( see Fig 3d legend ) . This simulation which is within physiological expectation for a high MOI infection , shows that every cell in a population could be completely supressed to zero translational activity ( black lines , each representing 5% of the population ) and with a biphasic oscillation return to a definable high level in a certain time frame , yet this would not be discernible from the averaged population output at any time ( red broad line ) . This represents a type of technical sampling limit well known in analysis of oscillating outputs , combined with averaged population analysis inherent in current biochemical techniques versus information gained from single cell analysis of metabolic processes ( see Discussion ) . Cell-to-cell transmission during plaque spread involves high numbers of particles transmitting to the surrounding uninfected cells . To examine whether a single particle would be sufficient to induce efficient global translational shutoff , we infected cells with progressively decreasing amounts of virus , down to a MOI 0 . 005 where 1 cell in 200 would be infected with an infectious particle and statistically the probability of infection by more than one particle is extremely low ( <0 . 005% ) . We then pulse-labeled cultures with HPG 1 hr after infection . Under these conditions , detection of individual particles themselves was difficult but antibody to gB gave brighter signals than several other antibodies evaluated including ones to VP5 . The results ( Fig 4 ) showed that the numbers of cells showing translational shutoff titrated downwards in proportion to increasing virus dilution ( Fig 4a ) . At MOI 0 . 01 isolated individual cells showing pronounced shutoff of translation were observed , a feature never seen in mock infected cultures ( Fig 4a ) . Moreover such cells contained detectable fluorescent punctae , stained by the antibody to gB . At MOI 0 . 005 , isolated individual cells could still be observed , with virtually complete translational shutoff , examples of which are shown ( marked by asterisks , Fig 4b and inset ) . In these cells , defined by the virtual absence of detectable translation , corresponding single particles could be observed ( marked by arrows , panels i inset ) . We note that although there have been few previous reports on the fate of membrane proteins very early after infection , translational shutoff indicates that these cells were infected , and thus the signal for gB , indicates that the membrane , or at least a significant population of this component remains in a tight localised focus . ( As explained below , translational shutoff requiring simply virus attachment rather than infection would be not be consistent with any known mechanism nor with further results from this work ) . These results make no assumption about non-infectious particles and indeed it could be that certain particles we detect did not go on to make an infectious pfu , but were sufficient to promote shutoff . However even accounting for particle/pfu ratios , ( in our stocks approximately 50 ) , the vast majority of cells at MOI 0 . 005 would be infected by a single particle . While it may be , and indeed is likely , that particles go undetected in this experimental set up , nevertheless taken together our results strongly support the conclusion that at most a few particles and very likely a single particle of HSV-2 can be sufficient to induce a very profound translational shutoff . We note however that while HSV-1 exhibited prominent regional shutoff during cell-cell spread ( Figs 1–3 ) , we could not observe shutoff under conditions of single particle HSV-1 infection . As indicated above , overall protein translation declines progressively during HSV infection , contributed to by multiple mechanisms including the HSV vhs function [14 , 35–39] . To examine the influence of vhs on regional translational suppression at the single cell level , we examined a series of deletion or catalytic vhs mutants in HSV-1 and HSV-2 , with representative results shown in Fig 5 . While HSV-1 [17] exhibited efficient regional shutoff ( Fig 5a , HPG asterisked cells surrounding VP5+ve cells ) , for HSV-1[17] . Δvhs we never observed significantly declined translational shutoff at the plaque periphery . This lack of regional shutoff for a vhs mutant was even clearer with HSV-2 , given the profound suppression for the parental virus . We first tested a complete deletion mutant HSV-2 . ΔUL41 ( ΔUL41 ) , versus its repaired counterpart , ΔUL41-Rep . For the repaired virus , there was virtually complete cessation of translation in numerous cells immediately surrounding the advancing focus of infection ( Fig 5b , ΔUL41-Rep , asterisks ) . For the mutant , no zone could be identified and there was little discernible difference in cells immediately adjacent to the infected focus versus more distant cells . To complete these analysis , we examined a HSV-2 vhs mutant containing a single amino acid substation , D215N , which inactivates the RNase function of the protein together with its repaired version [58] . The repaired virus again showed very pronounced regional shutoff , with virtually every cell surrounding the focus of infection being suppressed while cells within the focus showed overall levels of translation similar to those more distant peripheral cells ( Fig 5b , D125N-Rep ) . By contrast , again we never observed significantly altered levels of protein synthesis in cells surrounding the focus of D215N mutant infection ( Fig 5b , D215N ) . Since in the D215N mutant , vhs is packaged into the virion in similar copy numbers to the repaired mutant [58] , these results indicate that the translational shutoff is tightly linked to the RNase activity of vhs and that infection and tegument deposition is necessary for shutoff . The results also imply that , while other mechanisms play a role in host translational shutoff , in this regional shutoff at the earliest stages of infection , vhs plays a critical role . These data link the regional oscillation in translational levels to vhs function . We next wished to examine any corresponding regional alteration of factors within the translational apparatus related to vhs function . In vivo , vhs is thought to promote the degradation of mRNAs that are in the process of translational initiation by interaction with candidate cell translation factors , notably eIF4A , an RNA helicase and eIF4H , an accessory factor that stimulates eIF4A [38 , 41 , 59 , 60] . We therefore simultaneously analysed protein synthesis by HPG incorporation and steady state levels and localisation of a series of translation factors during HSV cell-cell spread . S3 Fig shows representative images for eIF2α , eIF4AII , eIF4B , and eIF4G at the leading edge of virus spread , containing examples of individual cells with significant shutoff ( cells marked 1 ) adjacent to external cells with normal levels of translation ( cells marked 2 ) . We could discern no difference in levels or localisation of any of these components in cells where translation was suppressed compared to adjacent cells showing normal translation ( S3 Fig ) . However we observed a very striking and clear cut specific relationship between translation and localisation for another factor , eIF4H ( Fig 6 ) . In this case , virus spread is advancing from the bottom left with cells at the top being uninfected . A profound shutoff zone is seen across virtually the entire advancing face ( Fig 6a , HPG and DAPI , cells marked with asterisks ) . While eIF4H is localised predominantly within the cytoplasm in cells in the unaffected top zone , there was a pronounced and highly correlated relocalisation of eIF4H from cytoplasm to the nucleus in cells precisely within the shutoff zone ( Fig 6a , HPG versus eIF4H , asterisks and arrows mark the same cells for each read out ) . Interestingly , eIF4H localisation in cells within the interior infected zone also appeared more prominently nuclear . To extend this , we also examined eIF4H localisation early after infection at high MOI at a time ( 2 hr ) when all cells would be expected to be infected ( but the asynchronous pattern of shutoff and recovery would be expected within individual cells across the population , as described in Fig 3 ) . The results showed a clear and pronounced relocation from the mainly cytosolic localisation in mock-infected cells to a distinctly nuclear localisation in many HSV-infected cells ( Fig 6b , arrows indicate eIF4H localisation in representative cells for mock and infected ) . An important feature of these results in shown in Fig 6c . This shows the typical pattern of protein synthesis levels and localisation in mock-infected cells , alongside the mainly cytoplasmic eIF4H distribution . In HSV infected cultures we detected two patterns . A representative cell showing almost complete shutoff is shown together with pronounced eIF4H translocation into the nucleus . However infected cells exhibiting translational recovery , combined with progressive NPD formation , clearly show the retention of eIF4H in the nucleus notwithstanding translational recovery . The significance of these results for shutoff and translational mechanism are discussed further below . Finally considering the link with vhs function , we wished to determine whether eIF4H relocation was dependent upon vhs activity . We therefore compared regional translational activity and eIF4H localisation for the vhs deletion ( ΔUL41 ) and repaired virus ( ΔUL41-Rep ) . Consistent with the data above , for ΔUL41-Rep , we observed a pronounced zone of efficient translational suppression ( Fig 6d , zone marked by asterisks , HPG and DAPI ) surrounding the advancing infected focus ( identified by phase microscopy and indicated by top semicircle ) , with a clear correlation between translational shutoff and eIF4H relocalisation to the nucleus ( Fig 6d , HPG and eIF4H , asterisks ) . Both the regional shutoff and eIF4H relocalisation were abolished in ΔUL41 . We never observed either process for the mutant with a typical field shown in Fig 6e .
From many previous studies , it is generally understood that HSV infection leads to increased , temporally regulated virus protein synthesis combined with a suppression of cellular protein synthesis , with overall translational rates declining from those seen in uninfected cells [35–37 , 39 , 52 , 53 , 57 , 61–64] . This general decline in global cellular translation is proposed to occur in distinct phases reflecting firstly a vhs-dependent early shut-down ( associated with the vhs RNase activity and potentially other vhs activities ) , transitioning into a general vhs-independent decline in translation , through multiple specific and non-specific pleotropic effects of infection ( for reviews see [52 , 65] ) . Our understanding of these processes has been gained from studies of cell populations ( or biochemical analysis of population extracts ) during single-step replication . The principle conclusions when applied to a multistep system , encompassing infected to uninfected cell transmission that occurs in a natural infection , would result in a spatial representation of translational activity as summarised in the schematic ( Fig 7a ) . Thus infection initiating at the focal centre of infection ( red circle ) radiates outward ( arrow ) , with naïve uninfected cells at the extreme perimeter . Immediately adjacent internal cells represent the earliest/most recently infected cells and progressively more central cells represent the progressively later/“older” infected cells . Absolute translation levels are indicated in green with declining levels indicated by increasing grey . A cross sectional slice ( arrowed line ) indicating relative overall levels from peripheral to central cells , is summarised in the graph below , where the x-axis indicates space from external to internal and therefore time from new to old . Though idealised , this represents a first approximation of what would be anticipated from current models but this is not what is observed . Instead we observed a distinct regionalisation of overall translational activity , with early pronounced suppression at the periphery together with restoration of translational levels in the more interior cells ( Fig 7b and lower graphic ) . While overall translation is certainly reduced in the inner-most/older cells , this is not as extensive as anticipated . This advancing zone of translational oscillation has important implications . While other factors and viral genes could play some role ( e . g . activation and counter-activation of PKR mediated translational repression or stress mediated ATF6/PERK signalling ) , [15 , 66–71] , it is clear that the main mechanism of initial suppression ( though not necessarily translational recovery ) is viral mediated rather than cell mediated . This conclusion stems from the observation that vhs is critical and without vhs activity relatively little suppression of translation was observed , even at the single cell level . Because host-promoted suppression of translation can be counteracted by virus factors [11–15] , it could be anticipated that in the absence of vhs , these measures and countermeasures could balance out . However the relative contributions of each of these processes have not been definitively dissected . Moreover , a main point of this current study is that such dissection by population-based studies would in any case be extremely difficult . As described above , population based averaged output may not have the resolution to discriminate distinct processes and factors influencing translation activity in space and time ( e . g . Fig 3d and accompanying text ) . In future studies , it will be interesting to examine translation at the single cell level , ( in both single-step and cell-cell transmission models ) , using single or double viral mutants where vhs is inactivated with or without other viral genes such as ICP34 . 5 or US11 , thus blocking virus-induced translational suppression and virus-encoded counter-suppression that might have different contributions at different times . In this regard it is interesting to consider what factors would potentially extend the regional shutoff zone , with the result simulated in Fig 7c . By this , we mean not simply more pronounced translational suppression in a given target cell population , but increasing the numbers of cells in a broader regionally defined zone , shifting the balance of the efficiency , or duration of shutoff versus the recovery in translation . These outcomes would be dictated by multiple factors , potentially with changing relative contributions at different stages of infection . Moreover , although vhs plays a key role in the regional shutoff at the earliest stages of infection , other factors or gene products which affect translational activity generally ( or in a gene specific manner ) , could impact the overall spatial stratification . Examples include , increased half-life of input vhs , increased specific activity of vhs ( as has been speculated upon previously [72] , advanced de novo expression of new vhs activity , decreased activity or expression of virus factors that negatively modulate input vhs , such as VP16 and VP22 [72–76] , and decreased capacity of the host to restore translation levels . Linked to this therefore is the question of the mechanism ( s ) involved in translational recovery , both with regard to host and viral functions . At its simplest , translational recovery could be explained by progressive loss of input vhs protein and relatively unaltered cellular translation machinery , without a requirement for specific vhs dampening and/or a translational restoration processes . However clearly other factors are likely to play a role . With regard to the role of transcription in translational recovery , while it could be proposed that that abundant de novo host transcription was required , it could also be that translational suppression involved only a specific subpopulation of mRNAs ( e . g those at some stage of the active translation cycle ) , and that translational recovery could ensue with mRNAs already made but in transport or not yet in active translation ( see also below ) . In addressing the question of whether any de novo virus gene expression is required , or at least promotes translational recovery , clearly this cannot be approached using cycloheximide or other translational inhibitors ( to block de novo virus protein synthesis ) , since this would inhibit the activity to be measured . Blocking transcription , at a delayed time point after allowing some virus spread , and then pulsing with HPG to examine regional shutoff could in principle help address the question . The logic would be that , if the timing was appropriate , blocking de novo transcription in a cell which had received vhs and exhibited translational shutoff , ( but had not yet restored normal levels ) , might result in prolonged shutoff within cells , thus extending the zone and numbers of cells suppressed . However such attempts did not yield interpretable results ( in both single step and multi-step replication models ) . Although after high multiplicity infections vhs can clearly degrade host mRNA in the presence of actinomycin D , in attempts to examine overall spatial translational activity , we found that protein synthesis in control cultures was affected by actinomycin D even early after application and the dynamic range of the assays was severely reduced . A more fruitful approach for future studies will be to examine the extent of regional translational activity at the single cell level during infection with mutants , especially in components such as ICP34 . 5 , VP16 , or VP22 which have been reported to directly interact with vhs and to suppress its activity later in infection . One additional aspect of note stems from the comparative analysis of human keratinocytes , i . e . a cell type that is perhaps more physiologically relevant for HSV infection . Thus , at times when in Vero cells broad regional shutoff in the advancing zone was readily discerned , this was not apparent in the keratinocyte model , even for infection with HSV-2 . Overall protein synthesis measures , i . e . levels and general localisation , were approximately equivalent in HaCaT and Vero . However , as infection progressed in keratinocytes , distinct shutoff zones in advance of the antigen positive cells became much more apparent . This result is illustrated schematically in Fig 7d and warrants speculation on possible explanations . It could be that in keratinocytes , the numbers of virus particles transmitting from cell to cell early in infection is insufficient to induced translational shutoff . Then as infection progressed , whether through increased virus yields per cell or increased transmission rates to susceptible cells , shutoff was more efficiently induced because more particles infected the cells . This is possible but considering that a single particle of vhs , at least of HSV-2 , could elicit shutoff in Vero cells , this explanation would require also that HaCaT cells were intrinsically more resistant to vhs-mediated shutoff , ( although later non-specific translational suppression might still occur ) , a proposal that will be assayed in the future analysis using the single particle assay . Cell-type differences in the robustness or resilience of the translational apparatus ( for example the availability of translation factors ) has been previously alluded to , including the prospect that such intrinsic differences could influence the activity or requirement for vhs [77] . Such cell-type differences could help explain the observations reported here . However it could also be , at least in certain cell types , that quite distinct processes influence the outcome of infection when uninfected , naïve cells are in contact with or exposed to infected cells . For example , it could be in certain cells that paracrine mediators from infected cells signal to uninfected cells , promoting their relative resistance to translational shutoff and that as infection progressed these paracrine processes waned , or were counteracted by the virus , resulting in more efficient shutdown later in the progressing infection . Whatever the precise explanation , these results would inherently not be obtained in a model system of population analysis of single–step infection , and reinforce the utility of single cell analysis in transmission models in revealing additional complexity in virus-host interactions . Our results strongly indicate that a single particle of HSV can be sufficient to promote translational shutoff . This stems from the considerations; 1 ) at the extremely low input virus used , the statistical probability that more than one particle was infecting the cells is extremely small , 2 ) we could observe the single particles responsible and 3 ) under the same conditions HSV-1 was unable to promote shutoff . Although in these experiments it is possible that some particles may have escaped detection and formally that some cells may be infected by more than one particle , taken together our results strongly indicate that a single particle of HSV-2 can suffice to promote transient , global shutdown . Unfortunately no useful antibodies are available for immunofluorescence analysis of virion associated input vhs , ( or even de novo synthesised vhs ) to directly assess vhs distribution in relation to shutoff and analysis of input vhs by immunofluorescence has never to our knowledge been reported . It is estimated that vhs is a relatively minor virion component with approximately 200 molecules incorporated per particle on average [51 , 62 , 78 , 79] and our results indicate that the catalytic activity of these molecules is sufficient to promote pronounced , reversible translation shutdown . It will be interesting in future work to combine the HPG pulse labeling approach with single molecule RNA in situ hybridisation to host mRNAs , to examine questions on the abundance , location , selectivity and possibility temporal reversibility of RNA molecule abundance in relation to spatial outputs of active translation . We are currently attempting to combine bioorthogonal labeling and click chemistry with simultaneous RNA FISH to address such questions . In principle , the activity at the single particle level of HSV-2 versus HSV-1 could be due to a number of factors , including increased vhs copy number in HSV-2 , increased specific activity ( or half-life ) of HSV-2 vhs , or less efficient/slower reversal of vhs activity , for example by association with de novo expressed VP16 or VP22 . While formally each of these could contribute , it has been shown both in the context of virus infection and in reconstruction assays measuring activity of the individual protein , that HSV-2 vhs is more potent than HSV-1 [80–82] . The simplest explanation therefore to account for the observation that a single particle ( or low numbers ) of HSV-2 but not HSV-1 is sufficient for shutoff is the comparative potency at the level of specific activity of HSV-2 vhs . In addition to considerations of what governs the activity of input vhs early after infection , the result raises the possibility that other types of particles from an HSV-2 infection notably L-particles could , even via a single particle , contribute to transient shutoff of translation , in advance of or associated with infection by a complete infectious particle . Indeed it has previously been shown that vhs is incorporated into L-particles and that L-particles may exhibit vhs activity [78 , 79] . It is interesting to compare conclusions from distinct techniques . In previous studies [78] of high multiplicity infections , using population analysis of protein synthesis by 35S-methinone radiolabeling/SDS-PAGE analysis , the results indicated that approximately 400 virion particles were required to induce at least some translational suppression ( measured at 5 hr p . i . ) . In contrast our data indicate that a single particle , at least of HSV-2 , can induce pronounced early shutoff which is then reversed . In this regard , it is conceivable that the formation of a localised zone of translational shutdown , contributed to by single infectious virions or L-particles , is important for the strong physiological role of vhs in pathogenesis [72 , 83–89] . Our results also reveal a pronounced relocation of a cellular translation factor , eIF4H , from its mainly cytosolic localisation to a mainly nuclear localisation . This relocation was tightly linked to translational suppression . Not only was this relationship highly statistically significant , in that shutoff at the advancing front and eIF4H relocation were almost completely congruent , but eIF4H is also one of the cellular translation factors previously shown to directly interact with vhs in biochemical studies [41] . Classical biochemical studies have shown that vhs binds directly both to eIF4H and to eIF4A , an RNA helicase with which eIF4H itself interacts [40 , 41 , 90 , 91] . These interactions are thought to underpin the recruitment of vhs to the larger eIF4F cap binding complex , with such recruitment helping explain the specificity of vhs RNase activity in the degradation of mRNAs versus non-mRNAs in vivo [40 , 92–95] . However , the precise mechanism of vhs activity , target selectivity and associating factors remain to be fully identified [52 , 65] . Of the translation factors analysed here by immunofluorescence , there was clear selectivity in the relocalisation of eIF4H . While we could not observe any distinct alteration in eIF4AII , it remains possible that e . g . subpopulations of eIF4AII or indeed other translation factors are also relocalised in a manner correlating with translational suppression . While these results do not imply a causal association between eIF4H relocalisation and vhs-dependent translational suppression , nevertheless it is reasonable to integrate our results with the biochemical analyses cited above . Such analyses , further comparing w/t and selected vhs mutants , suggest that eIF4F association correlates with vhs binding to eIF4A , rather than eIF4H . At the same time however , recruitment of active vhs to eIF4F is not of itself sufficient to selectively degrade mRNAs in vivo and binding to eIF4H is thought to be required [38 , 40 , 41 , 90] . Moreover , siRNA mediated knockdown of eIF4H result in a decline in the activity of vhs in degrading host mRNA early after infection , independent evidence that eIF4H is specifically involved in vhs mRNA degradation and thus translational suppression [96] . We first considered that eIF4H relocation could be a general effect of infection , independent of vhs activity . However this was not the case , and relocation clearly required catalytically active vhs . Thus several not mutually exclusive explanations could account for the combined observations of on the one hand , a required interaction between vhs and eIF4H for mRNA degradation and on the other , eIF4H nuclear relocation tightly correlating with vhs-dependent translational suppression . Perhaps most plausible is the proposal that relocation is not causal , but is a downstream consequence of vhs activity . Once targeted by vhs for its recruitment to mRNA and after mRNA cleavage , eIF4H could be released from the RNase complex and then be somehow altered or lack a recycling/targeting mechanism for cytoplasmic retention . It could then ( eIF4H is 248 residues long with a mol . wt of ca . 27kD ) passively relocate to the nucleus and/or be retained there . An alternative explanation , but one also invoking relocation as a consequence of vhs function , would be some form of active shunting of the eIF4H cytoplasmic pool to the nucleus . While requiring vhs function , this would not exclude participation of other virus factors in relocation . Moreover , it appears that the majority of eIF4H is retained in the nucleus during translational recovery . This result may reflect a qualitative change in the nature of the translational apparatus during translational recovery , for example with eIF4H playing a different or dispensable role in the restoring infected cell . It could also be that this relocalisation reflects a broader change in the nature of the translational apparatus and the qualitative processes operating occurring during translational recovery compared to prior to shutoff . The precise explanation for the relocation of eIF4H and the consequences for later infected cell translation requires further investigation . Finally , it is interesting to speculate on the implications of vhs mediated biphasic translational oscillation on its role ( s ) in vivo . Although vhs deletion does have an effect on replication , vhs is dispensable in tissue culture and deletion mutants replicate comparatively well [62 , 83] . However vhs plays very significant roles in virulence and pathogenesis in vivo in various animal models [72 , 83–85 , 88 , 97 , 98] . It is becoming increasing clear that the loss of virulence and pathogenic outcomes of vhs mutants is due to an important role in immune evasion [reviewed in 52] . Proposals in this regard include vhs involvement in down regulation of MHC class I and II , suppression of the production of proinflammatory cytokines and blocking dendritic cell activation [99–102] . One attractive possibility therefore is that the oscillating nature of translational activity mediated by vhs is an integrated function , wherein the rapid suppression of translation is sufficient to compromise host cell functions , ( e . g . protein delivery , immune cell activation ) immediately after infection but where sustained translational suppression would impact on virus progression . Thus translation needs to be restored to provide necessary cellular de novo synthesised proteins , as well as viral proteins for progressive replication . In conclusion , while considerable insight into metabolic processes during virus infection has been gained from classical molecular and biochemical analysis at the population level , information at the individual cell level is also critical for a true understanding of the processes governing the outcomes of infection . Such analyses cannot answer all questions of detailed mechanisms and cannot replace molecular and biochemical investigation of these processes . However , these techniques of bioorthogonal chemistry in combination with immunofluorescence , complement biochemical and population approaches and can reveal new insight not appreciated by classic techniques . Our results support previous data from biochemical analysis but also reveal new understanding from which we propose a revised model of translational control during advancing infection which has important implications both mechanistically and with regards to the physiological role of translational control during infection in vivo . This work opens avenues for future investigation of infected cell translation , the mechanisms involved in suppression , the function of vhs and the role ( s ) of other viral and host factors . It is also highly applicable to other virus systems .
African Green Monkey kidney fibroblast ( Vero ) cells , obtained from the European Cell Culture Collection , Porton Down , UK ) and human epithelial keratinocytes ( HaCaT ) , obtained from Professor Gill Elliott , University of Surrey , UK , were grown in Dulbecco’s Modified Eagle’s Medium ( DMEM; Gibco ) supplemented with 10% Fetal Bovine Serum ( FBS; Gibco ) , and penicillin/streptomycin ( Gibco ) . The viruses used in this study were parental strains HSV-1[17] , HSV-1[KOS] and the vhs mutant HSV-1[17] . Δvhs [82] and for HSV-2 , HSV-2[186] and mutants ΔUL41 , ΔUL41-REP , D215N , and D215N-REP [58] . For studies during cell-cell transmission , single particle infections were routinely performed by infecting a confluent monolayer of cells ( 2x105 cells ) with 50 PFU with neutralising human serum added 1 hr post infection . HPG-pulse labeling intervals were initiated approximately 24 hr later . Studies during single step replication were performed at a multiplicity of infection ( MOI ) of 10 . Particle/pfu ratios of selected virus stocks ( routinely a ratio of approximately 50 ) were determined as previously described [103] . To analyse translation at the earliest times possible and in as synchronised a manner as possible ( see e . g . , Fig 3 ) , cells were infected at +4°C for 1 hr , shifted to 37°C for 45 min to allow infection to proceed and 45 min later methionine depleted and pulse-labeled with HPG for 30 min . Omitting a methionine depletion stage or shortening the labeling interval in the attempt to analyse translation at even earlier times resulted in a vastly reduced signal . The protocol adopted was the optimal compromise for sensitive labeling at the earliest time possible . For analysis of the earliest stages of single particle infection , infections were performed at MOIs of 0 . 1 , 0 . 01 and 0 . 005 in methionine depleted medium , and the cultures then incubated in labeling medium for 30 min as described . The following antibodies were used: mouse anti-ICP4 MAb ( Virusys , 1:500 ) ; mouse anti-VP5 MAb ( Virusys , 1:200 ) ; mouse anti-gB MAb ( Sigma-Aldrich , 1:100 ) ; mouse anti-protein disulfide isomerase ( PDI ) Mab ( Abcam , 1:50 ) ; mouse anti-eIF4H Mab ( Santa Cruz , 1:50 ) ; mouse anti-eIF2α Mab ( Santa Cruz , 1:50 ) ; mouse anti-eIF4AII Mab ( Santa Cruz , 1:50 ) ; mouse anti-eIF4B Mab ( Santa Cruz , 1:50 ) ; mouse anti-eIF4G Mab ( Santa Cruz , 1:50 ) ; and Alexa Fluor 594 goat anti-mouse IgG ( Molecular Probes , 1:500 ) . For immunofluorescence analysis , cells on glass coverslips were fixed at times indicated in 4% paraformaldehyde ( Pierce ) for 10 min , permeabilised with 0 . 5% Triton X-100 ( Sigma ) for 5 min , and blocked with phosphate-buffered saline ( PBS; Sigma ) containing 10% FBS for 30 min at room temperature ( RT ) . Cells were immunolabeled for 1 hr at RT with primary antibodies and 45 min with secondary antibodies , followed by click chemistry and mounting in ProLong Gold Antifade Mountant ( Molecular Probes ) . Images were with a Zeiss Axiovert 135 TV microscope system using Zeiss 63x ( Plan-APOCHROMAT , 1 . 4 numerical aperture ) , 40x ( Plan-Neofluar , 0 . 75 numerical aperture ) , or 20x ( Plan-Neofluar , 0 . 5 numerical aperture ) objectives and a Retiga 2000R camera with Image Pro Plus 7 . 0 software . Alternatively , images were acquired with a Zeiss Laser Scanning Confocal Microscope system using 488 nm and 543 nm lasers with Zeiss LSM 5 software . Each channel was collected separately , with images at 1024 x 1024 pixels , with 4x averaging and without or with a variable zoom factor . Single confocal sections were acquired or multiple z-sections at 1 μm intervals which were then compiled for maximum projection display . For all low moi early infected cell foci or plaques were randomly inspected for the presence of zones exhibiting regional shutoff at the periphery of the plaque . When comparing viruses , e . g . w/t versus a vhs mutant all plaques within a virus type showed an identical phenotype with regard to the presence or absence of such zones . For qualitatively or quantitatively assessing the extent of shutoff , as explained in the text ( pg 11 ) , inherent asynchrony in infection and other variables , means that there was not a continuous zone of shutoff around the developing plaque . Rather there are regional zones with defined features of pronounced translational suppression within and between areas of normal translation , where the inner part of the zone was at the edge of a distinct infected area , ( defined by virus antigen or morphology ) Selected representative fields exhibiting such zonal shutoff were examined for translation ( by click chemistry , see below ) , antigen presence ( by immunofluorescence ) and cell numbers ( by DAPI and phase ) . Usually at least 10 fields were specifically evaluated and imaged using x40 or x63 lenses , ( approximately 120 cells and 40 cells on average per field respectively ) . Thus for representative images , each panel in a figure is a representative image of a zone from about 50 plaques , 10 fields , and 400–500 cells . From systematic analysis of the HPG concentration and duration of pulse , we optimised protocols for HPG incorporation , click chemistry and fluorescence detection as follows . Cells on coverslips were mock-infected or infected with HSV by standard procedures . At times indicated , medium was removed and replaced with L-methionine-free DMEM ( Sigma-Aldrich ) containing 2% FBS for 45 min to deplete methionine prior to the addition of HPG ( Molecular Probes ) at a final concentration of 0 . 5 mM for an optimised labeling time of 30 min in L-methionine-free DMEM . When the pulse-labeled cells were to be analysed in parallel for localisation of specific antigens , immunofluorescence with primary and secondary antibodies was carried out as standard ( see above ) . Samples were then subjected to click reaction in a buffer freshly prepared in each case ( premixed for 2 min ) and containing 10 μM Alexa Fluor 488-azide ( Sigma ) ; 1 mM CuSO4; 10 mM sodium ascorbate; 10 mM amino-guanidine and 1 mM Tris ( 3-hydroxypropyltriazolylmethyl ) -amine ( THPTA; Sigma-Aldrich ) in PBS pH 7 . 4 . The reaction was allowed to proceed by incubation for 2 hr at RT in the dark . After removal of the reaction cocktail , cells were washed with PBS and mounted on slides in ProLong Gold Antifade Mountant . Images were acquired as described above . For quantitation of the global protein synthesis at the single cell level , we used the thresholding and object quantification modules of Image Pro Plus software ( Media Cybernetics ) . We evaluated several routes for individual cell quantification . We found that the cytoplasm of cells frequently overlapped between adjacent cells ( observed better by HPG incorporation than by phase microscopy ) . Moreover , the nuclear signal represented a considerable fraction of the total . Considering these features , we therefore used the nuclear signal as a measurement of relative levels of protein synthesis between cells . While not measuring total level of protein synthesis , this parameter give a more accurate measurement of between-cell variations within populations than attempting to delineate entire cell boundaries . HPG-labeled cells were co-stained with DAPI to allow outlining of the entire nucleus and creation of a mask for each field . The masks were applied onto the HPG-green channel in which the HPG intensity of each individual cell ( tagged as objects ) was quantitated and normalised to nuclear area . To delineate the cells within different zones ( distal uninfected cells , shutoff zone and interior infected focus zone ) , we set a threshold for significant suppression to be at 30% or below of the maximum intensity for the field and coded this zone as the shutoff zone ( yellow ) . Cells having intensities above this threshold were coded pink . This method is somewhat conservative since translation levels in uninfected cells ( i . e . in mock-infected monolayers ) were relatively consistent , rarely exhibiting levels below 50–60% of the maximum in the field . Setting the threshold at 30% may therefore miss cells that were in partial shutoff but this does not materially alter our conclusions . Cells were mock-infected or infected with HSV by standard procedures . In control experiments , cells were incubated with HPG in the absence or presence of 100 μg/ml cycloheximide ( CHX ) , added 1 hr before methionine depletion ( S1b Fig ) . Cells were lysed in PBS containing 2% SDS and diluted to 1% SDS before the click reaction . 100 μg of protein samples were subjected to the click reaction as follows . Click reaction buffer consisting of the capture reagent ( 0 . 1 mM IRDye 800CW Azide Infrared Dye from LI-COR ) ; 1 mM CuSO4 , 2 mM Tris- ( 2-Carboxyethyl ) phosphine ( TCEP; Sigma-Aldrich ) , 0 . 2 mM Tris ( benzyltriazolylmethyl ) amine ( TBTA; Sigma-Aldrich ) was freshly prepared . Following the addition of the click mixture , samples were placed on a rotating mixer for 1 . 5 hr at RT , and the reaction was stopped by addition of EDTA to a final concentration of 10 mM . Subsequently , proteins were precipitated ( chloroform/methanol , 0 . 25:1 , relative to the sample volume ) . The precipitated proteins were pelleted by centrifugation at 14 , 000 rpm for 5 min , washed with methanol and air dried for 10 min . The pellets were then resuspended in 1x SDS sample buffer , boiled for 5 min , and 20 μg of proteins were loaded on 12% SDS-PAGE gels . Following electrophoresis , gels were washed with water , fixed in solution containing 40% methanol , 10% acetic acid , 50% water for 5 min and washed with water . In-gel fluorescence detection of translated proteins was performed using a LI-COR Odyssey infrared imaging system , and the protein loading was assessed by Coomassie blue staining .
|
All viruses reprogram protein synthesis within infected cells for the production of their own proteins and for suppression of host antiviral responses . On the other hand , cells also modulate translation to suppress virus replication . Thus , global protein synthesis in infected cells represents the temporally regulated interplay of multiple translational processes . While protein synthesis has been generally studied by methods that investigate the average behaviour in cell populations , information at the individual cell level is also critical for a true understanding of the processes governing infection . We used novel techniques in chemical biology that enables spatial analysis of translation levels at the single cell level , examining protein synthesis during cell-to-cell spread of herpes simplex virus . This work leads to a new interpretation of previous models of translational suppression that would not be gained from population studies and demonstrates the broad potential of chemical biology for spatial and biochemical studies of virus infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"herpes",
"simplex",
"virus",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"chemical",
"compounds",
"pathogens",
"messenger",
"rna",
"microbiology",
"organic",
"compounds",
"geoinformatics",
"biochemical",
"analysis",
"viruses",
"protein",
"synthesis",
"dna",
"viruses",
"amino",
"acids",
"bioassays",
"and",
"physiological",
"analysis",
"immunologic",
"techniques",
"herpesviruses",
"spatial",
"analysis",
"chemical",
"synthesis",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"geography",
"proteins",
"medical",
"microbiology",
"gene",
"expression",
"immunoassays",
"microbial",
"pathogens",
"chemistry",
"methionine",
"immunofluorescence",
"biosynthetic",
"techniques",
"sulfur",
"containing",
"amino",
"acids",
"biochemistry",
"rna",
"organic",
"chemistry",
"nucleic",
"acids",
"protein",
"translation",
"viral",
"pathogens",
"earth",
"sciences",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"organisms"
] |
2018
|
A bimodal switch in global protein translation coupled to eIF4H relocalisation during advancing cell-cell transmission of herpes simplex virus
|
Retroviruses affect a large number of species , from fish and birds to mammals and humans , with global socioeconomic negative impacts . Here the authors report and experimentally validate a novel approach for the analysis of the molecular networks that are involved in the recognition of substrates by retroviral proteases . Using multivariate analysis of the sequence-based physiochemical descriptions of 61 retroviral proteases comprising wild-type proteases , natural mutants , and drug-resistant forms of proteases from nine different viral species in relation to their ability to cleave 299 substrates , the authors mapped the physicochemical properties and cross-dependencies of the amino acids of the proteases and their substrates , which revealed a complex molecular interaction network of substrate recognition and cleavage . The approach allowed a detailed analysis of the molecular–chemical mechanisms involved in substrate cleavage by retroviral proteases .
Retroviruses are associated with a broad range of diseases that include tumors , immunodeficiency syndromes , and neurological disorders [1] . They affect a large number of species , from fish and birds to mammals and humans , with global socioeconomic negative impacts [1] . Each year the HIV pandemic causes more than 3 million deaths despite advances in the development of anti-HIV therapies [2] . The seemingly endless capability of retroviruses to escape antiviral drugs undermines treatment strategies and prompts the need for new broad-spectrum therapeutic agents [3] . Retroviral proteases process viral precursor polyproteins into structurally and functionally mature proteins that combine into infectious viral forms . As such , these proteins are key targets for the design of therapeutic inhibitors [4 , 5] . To date , the majority of protease inhibitors for treatment of HIV have been peptide mimetics , and most of them were specifically designed against only one of the HIV-1 proteases , namely the HXB2 ( “wild-type” ) HIV-1 protease [6 , 7] . Unfortunately , this strategy has led to failures to retard the replication of strains bearing drug-resistant protease mutations [3 , 8] . Although efficiently hydrolysable protease substrates have served as excellent templates for peptide-mimetic inhibitor design , it is difficult to predict which combination of amino acids will make the best substrate over multiple proteases [6] . Analysis of protease mutations associated with drug resistance is also confounded by the existence of many viral subtypes carrying naturally occurring polymorphisms [9] . The genomic differences among HIV-1 proteases can be as high as 30% and range from 10%–70% within the retroviral protease class [3] . Mutations contributing to viral resistance to antiviral drugs in one particular HIV subtype are found frequently in equivalent positions in the genes of other HIV subtypes or other retroviral proteases [9–14] . Still , the roles of specific mutations are only partly understood [5] . Here we report the development and experimental validation of a novel strategy for the molecular analysis of retroviral proteases . We hypothesized that merging essentially all available knowledge of retroviral proteases and their interactions with their substrates into a unified model would provide broad insight into the function of these enzymes and facilitate the analysis of retroviral drug resistance mechanisms . The modeling that we here report is based on the multivariate analysis of sequence position–physicochemical properties of the amino acids of 61 retroviral proteases from nine viral species and reveals a complex network of physicochemical interactions involved in protease recognition and cleavage of substrates . The approach provides novel insights into the molecular mechanisms involved in substrate cleavage by retroviral proteases in general as well as in relation to drug resistance .
The model was based on an extensive survey of publicly available data from multiple retroviral proteases and their substrates from 16 years of retrovirus research during 1990 to 2005 , combined into a single dataset ( Table S1 ) . Because retroviral proteases are inherently dynamic structures that undergo significant structural changes with binding , we described each structurally aligned amino acid of the 61 retroviral proteases by their principal physicochemical properties ( i . e . , their z-scales z1–z5 ) , rather than using the proteins' static 3-D structures ( see Materials and Methods , Figure S1 , and Table S2 ) [15 , 16] . Similarly , we described the retroviral protease substrates by considering the same principal physicochemical properties of every single amino acid of the octapeptide sequence spanning the P4 to P4′ position ( see Materials and Methods for details ) . Protease cleavage rates are dependent on the constituents of the experimental assay ( e . g . , pH and salt concentrations ) [17] . To account for differences in the assays , additional assay descriptors were introduced ( Table S3 ) . Substrate recognition and cleavage involve many dynamic noncovalent and covalent interactions between the substrate and the enzyme . Such complex processes can be accounted for by introducing “cross-terms” into the multivariate modeling . Cross-terms are formed as a product of multiplication of any two of the descriptors and reflect the simultaneous influence of two particular physicochemical properties on activity . Cross-terms can be viewed as description of specific interactions , which do not necessarily need to occur by physical contact of amino acids with each other . In a mathematical sense , cross-terms represent approximate nonlinear contributions of combination effects , regardless of whether these occur due to close contact or not [18 , 19] . The descriptors of the retroviral proteases , substrates , assays , and cross-terms were correlated to the experimentally determined substrate cleavage rates ( kcat/Km ) using partial least squares ( PLS ) regression modeling ( Table 1 ) . Our results show that it is possible to obtain acceptable models only after inclusion of cross-terms between the descriptors of amino acids in the substrates and proteases , and between amino acids at different positions in the substrates ( Table 1 ) . Moreover , including assay descriptors in the modeling further increased the validity of the model ( Table 1 ) . The performance of the cleavage rate model ( CRM ) is summarized in Table 2 and shown graphically in Figure 1A . To validate the model further we examined its capacity to predict the activity of naturally occurring and artificially mutated retroviral proteases externally . This was afforded by excluding all data for eight retroviral strains one at a time in their entirety , and then predicting the excluded data using models constructed from the remaining data ( see Materials and Methods for details ) . This analysis showed that the models could accurately predict the activities of the excluded retroviral proteases , most notably for the HIV-2 protease with an accuracy of 93% ( root mean square error of prediction [RMSEP] = 0 . 52 ) , and for the HIV-1 protease mutants 86% ( RMSEP = 0 . 65; Figure 2 ) . By contrast , state-of-the-art model building using only the HXB2 HIV-1 protease data failed to give acceptable models ( Table 1; see Materials and Methods for further details ) . Our approach thus resulted in a statistically well-validated model for the rate of cleavage of peptide substrates by retroviral proteases . However , although the model was capable of predicting kcat/Km values , it could not distinguish cleavable from noncleavable sequences . To allow such predictions , we constructed a cleavability model ( CLM ) by correlating substrate and protease descriptors and their cross-terms to a vector representing cleavability or noncleavability ( Table 2; see Materials and Methods for details ) . The CLM , which was based on the data for all 61 proteases with all cleavable peptides used for the construction of the CRM as well as an additional large set of noncleavable peptides , almost perfectly classified cleavable and noncleavable substrates ( 97% ) and performed excellently in external predictions of cleavability of new sequences ( 90 . 1% ± 1 . 2%; see Table 2 and Materials and Methods ) . Encouraged by these results , we confirmed the predictive power of the models by independent experimental validation . We constructed a virtual peptide library and applied in silico screening to it , first by using the CLM , then followed by the CRM . This process resulted in an unbiased set of 30 novel peptides , selected according to diversity criteria , of which 15 were predicted as cleavable and 15 as noncleavable; the predicted cleavage rates for the cleavable ones ranging over almost three orders of magnitude ( see Materials and Methods for details ) . The peptides were subsequently synthesized and assayed for their cleavability by the HXB2 HIV-1 protease and three HIV-1 proteases harboring mutations associated with drug resistance . The analysis showed that the CLM could correctly recognize all cleavable substrates as cleavable , and all noncleavable substrates as essentially noncleavable ( 100% accuracy; Table 3 ) . Moreover , the experimentally determined cleavage rates of the cleavable peptides agreed well with the CRM predicted rates on HXB2 and mutated HIV-1 proteases ( 68% accuracy; RMSEP = 1 . 01; Figure 1B ) . Addition of all experimental data to the CRM further increased CRM predictability according to cross-validation; the Q2 increased from 0 . 52 to 0 . 54 . Analysis of the regression coefficients of the CRM allows analysis of the physicochemical properties of the amino acid residues of both the substrates and the proteases required for catalytic activity . The model verifies the well-known hydrophobic requirement for the P3–P3′ residues of the substrates by the retroviral protease cleavage sites ( i . e . , the regression coefficients for the z1 terms of the P3–P3′ residues of the substrates are negative , as can be seen from the physicochemical property map derived from the model; Figure 3 ) [20] . In fact , the map shows that P3 and P3′ can accommodate various amino acids with a preference for hydrophobic residues , which is in perfect agreement with previous findings [13] . Moreover , it is well-known that β-branched or polar amino acids are not tolerated in the P1 position of retroviral substrates [13] . In the model , this limitation is reflected in that the z1 , z2 , and z4 terms for the P1 position are most favorable for aromatic amino acids and methionine , while polar or β-branched residues are disfavored . Specificity studies for retroviral proteases have found that highly complicated and not easily interpretable interior interactions take place in the substrates [21] . Such interactions become easy to interpret in the model by the cross-terms , however . For example , the large regression coefficients from particular physicochemical properties of the P1′–P2 , P1′–P1 , P1′–P3′ , and P1–P3 residue pairs indicate the presence of interactions for these pairs , while no such interactions take place for the P3–P2 pair according to the model , in accordance with experimental results [22] . Cooperativity between P1–P2 , P1–P3 , and P1–P4 residue pairs , indicated by the model , has also been shown to be important for specificity features [21] . In earlier specificity studies of retroviral proteases , many series of substrates were used , each of which often differ only by one or two residues , prohibiting a complete analysis of all residues at every subsite [21] . Merging all the available data thus provides a comprehensive picture that reveals important cross-dependencies between several different residue positions in the substrates ( i . e . , the regression coefficients where particular substrate–substrate cross-terms are significantly large ) , and demonstrates that a complex interaction network between residues in the substrates is involved in their cleavage ( Figure 3 ) . In a similar way as above for analysis of substrates , the regression coefficients of the descriptor terms for protease amino acid residues can identify the physicochemical properties of the nonconserved amino acids in the proteases that determine substrate cleavage ( e . g . , the model reveals that hydrophobic amino acids are preferred at the position corresponding to position 82 in HIV-1 protease to afford a high catalytic activity of the proteases ) . This is due to the fact that the regression coefficient z1 for position 82 is the largest one and negative . Another example is that hydrophobic , small-size amino acid residues ( e . g . , Ile or Pro ) are preferred at position 81 , since both regression coefficients for hydrophobicity ( z1 ) and size ( z2 ) are among the largest and also negative at this position for the model . To assess a cumulative importance of all physicochemical properties for each protease residue relatively to other residues , we computed and compared the absolute value sums of z1–z5 regression coefficients for each individual position , which is thus a measure representing the overall importance of an amino acid in eliciting chemical effects in the protease when compared to the same measure of other amino acids in the model . Our results from this analysis reveal the most important nonconserved positions involved in catalytic activity of the retroviral proteases ( Figure 4A; see Materials and Methods for details ) . One of the most important amino acid residues shown by the CRM was the threonine of the aspartate proteases' catalytic triad , Asp-Thr ( Ser ) -Gly ( i . e . , the T26 residue in HIV-1 protease that is substituted to serine in Rous sarcoma and avian myeloblastosis virus retroviral proteases; Figure S1 ) [23] . Six further amino acid positions ( corresponding to R8 , D30 , V32 , V82 , I84 , and L90 residues in the HXB2 HIV-1 protease ) were also identified as important . These positions agree well with the amino acids known to be associated with high resistance to protease inhibitors [24 , 25] . The model also identified P81 and N83 amino acid positions , which are known to play a key role in regulation of retroviral protease function [26] . The role of the I64 residue , also indicated by the model , appears to be indirect , as it is located farther way from the substrate cleavage cleft ( Figure 4A ) . The substrate-protease cross-terms of the CRM were then in a similar fashion used to identify the major cross-dependencies of the protease and substrate amino acids for cleavage activity , which thus reveal the major interaction effects that determine substrate specificity ( Figure 4B–4D; see Materials and Methods for details ) . We then found that P3′ substrate residues form the strongest cross-dependencies with retroviral protease amino acids corresponding to L24 , D29 , I84 , and L97 residues in the HIV-1 protease ( Figure 4D ) . Notwithstanding that D29 directly contributes to the S3′ subsite , the effect of residues L24 , I84 , and L97 distal to the S3′ subsite is indirect [13] . Further analysis indicated the importance of direct interactions between the P1 residue and the P81 and V82 protease amino acids ( Figure 4C ) , which form a part of the S1 subsite [13] . The P1′ residue , on the other hand , shows a major indirect interaction with the L90 amino acid ( Figure 4C ) . This is a position for a distantly acting , commonly appearing drug-resistant mutation , L90M in the HIV-1 protease , which has been observed to increase the cleavage activity of HIV-1 protease for natural substrates mutated in the P1′ position [27] . Although the P3 amino acid may interact directly with various amino acids in the S3 pocket , our results suggest that the P3 amino acid specificity is determined indirectly by effects arising from the I13 and E34 residues ( Figure 4C ) . This result is in alignment with other reports , where the polymorphic mutation I13V was linked with the mutation of Thr to Ala at the P3 position of the natural cleavage site p24/p2 [28] . Moreover , mutations at the E34 position have been seen in clinical HIV samples after protease inhibitor treatment [29] . The analysis further demonstrated that the P4 and P4′ residues form a large number of important cross-terms with protease amino acids ( Figure 4B–4D ) . The P4 and P4′ positions can broadly tolerate a variety of amino acids ( Figure 3 ) . However , mutations could occur in the P4 and P4′ positions of natural cleavage sites under antiviral drug pressure , which compensate a decreased catalytic activity of drug-resistant retroviral multi-mutants with the mutations depicted in Figure 4B–4D . Indeed , resistance mutations are known at such positions . For example , the Cbz group of the retroviral protease inhibitor TL-3 occupies the S4 subsite and interacts with the F53 residue , where mutation to a smaller Leu causes a decreased susceptibility of TL-3 by an order of magnitude [30] . Moreover , a substantial overlap exists between retroviral protease residues associated with specificity ( Figure 4B–4D ) and residues involved in resistance development ( I13 , I50 , F53 , V82 , I84 , N88 , L90 , and I93 ) [24] .
A goal of any successful antiretroviral therapy must be to ensure complete inhibition , or at least a fair retardation , of the replication of all the multiple viral strains that constantly emerge in an infected organism . The present approach allows concomitant analysis of many mutated target proteases in silico , and is useful to aid the analysis of the roles of such mutations in drug resistance . The Stanford HIV drug resistance database contains more than 24 , 000 HIV protease polymorphisms and resistance mutation sequences [31] . Performing high-throughput screenings and ligand optimizations to search for a drug suited to such a multitude of targets is an insurmountable task . Traditional structure-based drug design is built on the “lock-and-key principle , ” in which a drug is designed to be a snug fit with its target protein [3] . It is not well-suited to concomitant design of multiple targets that undergo conformational changes and show dynamically regulated differences in the shape of their active sites . Our results show that combining multiple proteases from many retroviral strains encompasses the mutational space information of retroviral proteases better than using the protease from a single strain . Thereby , it becomes possible to obtain models that allow interpretations of the molecular mechanisms involved in retroviral protease cleavage site processing . The multiple-protease–based models thus allow localization of physicochemical effects that rule substrate cleavage and predict multiple positions where compensational mutations could occur that restore substrate cleavage following the appearance of protease inhibitor resistance mutations . The validity of the models are proven not only by applying state-of-the-art statistical validation methods , but also by their ability to a priori accurately predict the cleavage rate of entirely novel peptides and proteases . Interestingly , the model also reveals several amino acids outside of the enzymes' binding pocket , such as I13 , L24 , E34 , I64 , I84 , L90 , and L97 , as being important for catalytic activity . It is well-known that retroviral proteases are flexible proteins , and it is likely that these positions contribute with long-range conformational effects that indirectly affect protein function and mobility [18] . The regression coefficients of terms and cross-terms of the model contain a large amount of chemical information that would be of direct value in designing a substrate that is efficiently cleaved over a group of protease mutants . Another option for such design would be to apply virtual screening of peptide libraries using the model . In addition , we show that inclusion of new experimental data leads to a model with improved predictability . Iterating the process should thus give models that afford increasingly accurate predictions of peptides with particular properties ( e . g . , having broad specificity over multiple resistance mutations ) . Analyzing such new entities experimentally and including the new data into new models would lead to further improved models and would refine the understanding of how retroviral proteases overcome drug resistance .
Data for substrate cleavage by 61 retroviral proteases were collected in an extensive survey that included publicly available data for retroviral proteases during 1990–2005 [32–64] . The survey included proteases from the following viruses: HIV-1 , HIV-2 , AMV ( avian myeloblastosis virus ) , RSV ( Rous sarcoma virus ) , HTLV-1 ( human T cell leukemia virus type 1 ) , BLV ( bovine leukemia virus ) , Mo-MuLV ( Moloney murine leukemia virus ) , EIAV ( equine infectious anemia virus ) , and FIV ( feline immunodeficiency virus ) ; its outcome is summarized in Tables S1 and S2 . In some cases fully denaturated proteins had been exposed to HIV-1 or HIV-2 proteases [61 , 62 , 64] . Noncleavable octapeptides were in these cases extracted from the noncleavable fragments located between the observed cleavage sites by using an eight-residue-long sliding window . Some of the data were generated in-house as described below ( see Materials and Methods further below ) . Description of proteases . The 61 retroviral protease sequences included in the study ( Table S2 ) were aligned using the template shown in Figure S1 . A total of 94 amino acids could be fully aligned over all the proteases , but only the positions lacking gaps in all proteases , as well as those being nonconserved , were considered . These then amounted to 85 amino acids , which were described by their five principal physicochemical properties , or “z-scales” [16] . These z-scales roughly represent hydrophobicity ( z1 ) , steric properties ( z2 ) , polarizability ( z3 ) , and polarity and electronic effects of amino acids ( z4 , z5 ) ( z-scales are the principal components of 26 physicochemical properties of amino acids , which include: molecular weight , van der Waals volume , heat of formation , energy of highest occupied molecular orbital , energy of lowest unoccupied molecular orbital , log P , α-polarizability , absolute electro-negativity , absolute hardness , total molecular surface area , polar molecular surface area , nonpolar molecular surface area , number of hydrogen bond donors , number of hydrogen bond acceptors , indicator of positive charge in the side chain , indicator of negative charge in the side chain , NMR α-proton shifts at pD = 2 . 7 and 12 . 5 , and seven descriptors representing thin-layer chromatographic mobilities using different stationary and mobile phases [16] ) . Thus , every protease was described by 85 × 5 = 425 descriptors , which comprised the physicochemical property space information of the series of proteases used herein . It shall be noted that amino acids entirely conserved in a library do not yield any additional information and their importance can therefore not be assessed unless the library is extended by further mutations of these positions . Description of substrates . We restricted the length of the substrates to octapeptides ( P4-P3-P2-P1-P1′-P2′-P3′-P4′ , where P4 represents substrate N-terminus amino acid and P4′ represents C-terminus substrate amino acid ) , since generally only eight amino acid residues are involved in the interaction process with eight subsites ( S4-S3-S2-S1-S1′-S2′-S3′-S4′ ) of a retroviral protease , with the cleavage site being between the P1 and P1′ amino acids . Each one of the eight amino acids of the substrates were described by the same five z-scales as above , yielding 8 × 5 = 40 total descriptors for each substrate . This comprised the physicochemical space information of the series of substrates used herein . Description of assay conditions . Descriptors for eight constituents of the experimental assays according to the published data used [32–64] were included in the modeling in order to normalize for the differences in assay conditions . The descriptors used are given in Table S3 and accounted for variations in pH , sodium chloride , 2-mercaptoethanol , EDTA , DMSO , dithiothreitol , nonidet-P40 , and glycerol concentrations . Description of cross-dependencies of proteases , substrates , and assays . The mutual dependencies of protease , substrate , and assay properties were described by cross-terms . These cross-terms were formed by multiplication of any two of the above-described descriptors of proteases , substrates , and assays . To simplify the discussion in the following , the above blocks of descriptors for assays , proteases , and substrates will be referred to as A , B , and C descriptor blocks , respectively . The cross-terms were then formed by multiplications yielding A × A , A × B , A × C , B × C , and C × C cross-term blocks . Each one of these blocks were in the subsequent modeling used in various combinations , together with the A , B , and C blocks , to demonstrate their respective importance and to find the most suited combination for creation of optimal models . All ordinary protease , substrate , and assay descriptors were mean-centered and scaled to unit variance prior to computation of cross-terms . In addition , we applied block-scaling for each type of descriptors to account for their differences in number and mutual correlation [65] . ( Block-scaling gives each block a variance square root of Nb , where N is the number of descriptors in block b . Block-scaling thus gives each variable the variance 1/ ( Nb ) 1/2 . The procedure avoids a situation where large blocks of descriptors mask small ones . ) ( The B × B cross-terms block was not formed due to its huge number of descriptors [i . e . , 90 , 100 descriptors] ) . Two types of models were created . One aimed to delineate whether or not a peptide is cleavable by retroviral proteases . This model , called CLM , was trained against a vector formed by assigning +1 to a hydrolysable substrate and −1 to a nonhydrolysable . The other model aimed to model the cleavage rate of cleavable substrate , and was called CRM . In the latter case , the model was trained against the vector formed from the logarithm of the experimentally determined kcat/Km values ( mM−1h−1 units ) , log ( kcat/Km ) . CRM . All experiments listed in Table S1 , where substrate cleavage rates had been determined , were used for the construction of CRM , and comprised 760 observations . Protease , substrate , and assay descriptors , and cross-terms thereof , were used as detailed in Table 2 . The preprocessed descriptors ( see below ) were correlated to measured cleavage rates log ( kcat/Km ) units by PLS regression modeling using Simca-P+ 10 . 0 software ( Umetrics AB , http://www . umetric . com ) . In the model building , inclusion of various descriptor blocks were attempted and the data were subjected to PLS regression modeling ( see models 1–5 and the single target model [STM] in Table 1 for details ) [65] . While models 1–5 utilized all the 760 log ( kcat/Km ) values obtained from Table S1 , the STM comprised only 212 experiments for the HXB2 HIV-1 protease of Table S1 . Models were subjected to validation ( see below ) , and model 4 and the CRM were the only ones considered acceptable ( R2 > 0 . 7 and Q2 > 0 . 4 ) [66] . As CRM also outperformed model 4 , it was the one used herein . For the CRM containing descriptors of substrates , proteases , assays , and their cross-terms as shown in Table 2 , the regression equation can be expressed as follows: where A , B , and C represent the number of descriptors in assay , substrate , and protease blocks respectively , a , b , and c correspond to assay , substrate , and protease descriptors respectively , and coeff denotes a coefficient for a corresponding descriptor or a cross-term . CLM . All data listed in Table S1 were considered for the CLM . Assay descriptors were not included . This was because the assay conditions used have only minor effects on substrate cleavability . In some cases , the assay conditions also had not been specified . All in all , the dataset comprised 2 , 163 peptide–protease combinations . However , 13 experiments of these differed only by assay descriptors and were therefore excluded . This resulted in a final dataset with a total of 2 , 150 observations , which was used for the model creation . Proteases , substrates descriptors , and cross-terms were used for the CLM construction as denoted in Table 2 . The descriptors , preprocessed as described below , were correlated to the peptide cleavability ( +1/−1 ) by PLS regression modeling using Simca-P+ [65] . For the CLM containing descriptors of substrates , proteases , and their cross-terms as shown in Table 2 , the regression equation can be expressed as follows: where B and C represent the number of descriptors in substrate and protease blocks , respectively , b and c correspond to substrate and protease descriptors , respectively , and coeff denotes a regression coefficient for a corresponding descriptor or a cross-term . Validation of models . The goodness-of-model fits were quantified by R2 . This unitless fraction indicates the portion of the total variation of the response that is explained by the model and shows how well a model fits the data [65 , 66] . We also computed the root mean square error of estimation ( RMSEE ) to determine the internal calculation error within the model: where yi and y icalculated denote the observed and calculated rates by the CRM [65] . N denotes the number of calculated observations . Cross-validation is a method of estimating the accuracy of a regression model . In cross-validation the dataset is divided into several parts ( seven were used herein ) , with each part used to test a model fitted to the remaining parts , resulting in the cross-validated regression coefficient Q2 [67 , 68] , where a higher Q2 denotes a better predictability [66] . In bootstrap validation the dataset is repeatedly and randomly permutated , yielding new dataset samples with replacements from the original dataset [69 , 70] . New models are then built on permutated data , and R2 , Q2 , and correlation coefficients between original and permutated response values are estimated . Intercept values for R2 ( iR2 ) and Q2 ( iQ2 ) reflecting R2 and Q2 of random response data were computed from repeated random permutations of the data ( 100 repeats were done herein ) [70] . Negative iQ2 indicates that it is impossible to get predictive models based on random data . External validation for the CLM was performed by randomly dividing the dataset into two parts ( 30% and 70% ) . The smaller part was excluded and predicted based on a model created from the remaining 70% of the data . This procedure was repeated ten times . For each external validation round we calculated the prediction accuracy ( i . e . , the fraction of correctly classified substrates to cleavable or noncleavable versus all observations included in the test set ) . External validation of the CRM was performed by excluding all data for eight retroviral strains one at a time in their entirety , and then predicting the excluded data using models constructed from the remaining data . ( In the case of HIV-1 proteases , the HXB2 HIV-1 protease and HIV-1 proteases with five artificial stabilizing mutations , Q7K + L33I + L63I + C67A + C95A , were kept in the model , and the external predictions were performed for the remaining 23 drug-resistant HIV-1 mutants . ) The prediction accuracy for each model was estimated as the fraction of protease–substrate pairs with prediction error < 1 . 0 log ( kcat/Km ) to all protease–substrate pairs used for the respective external prediction . This critical threshold was set based on 2-fold RMSEE for the CRM ( 0 . 49 log ( kcat/Km ) ; Table 1 ) . We also used RMSEP to evaluate model predictive ability for external datasets [65] . RMSEP can be compared with the root mean square error of internal cross-validation ( RMSECV ) , which illustrates the error of predictions within the model [65] . RMSEP was computed as follows: where yi denotes the observed rate and y ipredicted the externally predicted rate by the CRM . RMSECV was calculated in an identical fashion , using for y ipredicted the predicted rates obtained during internal cross-validation of the CRM [65] . N denotes the number of predicted observations . The correlation coefficient , r , for the experimentally observed versus predicted cleavage rates by the CRM ( Figure 1B and Figure 2A–2H ) was determined , and the statistical significance , p , of the correlation was assessed . The p-value obtained is the probability that a correlation this great or greater ( in the positive direction only ) would be seen if there was no linear relationship between observed and predicted cleavage rates . An in-house add-in to Excel ( Microsoft , http://www . microsoft . com ) was used for the test of correlation . All significance tests were one-sided . Analysis of CRM . All descriptors used for the CRM construction were mean-centered and scaled to unit variance , as described above . This transformation unified the different ranges of descriptor values allowing a comparative analysis of their coefficients . The larger an absolute value is of a descriptor's coefficient , the larger its impact is on the model's outcome . To construct the retroviral protease substrate physicochemical fingerprint map shown in Figure 3 , we analyzed CRM substrate and substrate–substrate cross-term descriptor coefficients . First , we compared the absolute values of the coefficients to find the largest ones . A total of 17 z-scales of the substrates' amino acid residues were then identified to be highly important and are shown in Figure 3 as a red sphere if its regression coefficient had a positive value , and as a blue sphere if it was negative . In a similar way we identified 20 highly important substrate–substrate cross-terms . These are represented in Figure 3 as red lines if the corresponding cross-term coefficient had a positive value , and as blue lines if it was negative . To determine the most important protease amino acids shown in Figure 4A , we compared the sum of the absolute values of the five z-scale descriptor coefficients for each of the 85 aligned amino acid positions of the proteases . Summation of the coefficients allowed us to simultaneously capture all the physicochemical property effects caused by each of the amino acids considered . The ten amino acids with the largest sums of their coefficients and consequently the largest contribution on cleavage rate according to the model were the ones depicted in Figure 4A . Figure 4A was produced using the Visual Molecular Dynamics ( VMD ) program , version 1 . 8 . 3 [71] . The model was further analyzed by considering protease–substrate amino acid interactions as described by cross-terms . Every protease–substrate amino acid pair yields 25 cross-terms , as five z-scales of each amino acid multiplied makes 25 cross-terms . To capture the most important protease–substrate interactions , we calculated the sum of the absolute values of 25 cross-term coefficients for each substrate–protease amino acid pair . We then compared all obtained sums and identified the protease–substrate interactions with the largest influence on the model's outcome . In total , the 20 most important protease–substrate amino acid pairs are presented in Figure 4B–4D . Figure 4B–4D was also produced using VMD [71] . It may be noted that whereas the regression coefficients arising from the substrates only , or the proteases only , relate to the overall activity of all the substrates and all the proteases , respectively , the coefficients of the substrate–protease cross-terms relate to specificity ( i . e . , the ability of a particular substrate to prefer a particular protease ) . In silico substrate screening . The active site of HIV-1 protease accommodates a sequence of eight amino acid residues ( P4–P4′ ) of a substrate , and cleaves it between the P1 and P1′ residues . The potential number of substrates consisting of natural amino acids is therefore 208 , but this large number was not computationally feasible to assess . We therefore constructed a smaller library of octapeptide sequences by considering only the natural amino acids that can frequently be found in retroviral protease substrates as follows: for the P4 position , amino acids were R , S , K , P , G , or A; for P3 , Q , A , R , K , G , or E; for P2 , N , E , A , G , T , I , L , or V; for P1 , F , Y , W , M , or L; for P1′ , P , F , A , L , W , or V; for P2′ , L , Q , V , A , T , or I; for P3′ , D , S , Q , T , M , V , R , or I; and for P4′ , T , M , Q , V , P , G , R , or S . This resulted in a virtual library of 6 × 6 × 8 × 5 × 6 × 6 × 8 × 8 = 3 , 317 , 760 entries . The library was first screened using the CLM to filter out all noncleavable substrates for the HXB2 HIV-1 protease . We considered a substrate noncleavable if its predicted cleavability parameter was less than −0 . 3 . This resulted in 2 , 463 , 379 cleavable sequences ( ∼74% of the initial library ) . We then used the CRM to predict the actual rate of cleavage for the cleavable octapeptides . From these we chose 15 substrates; seven with a predicted cleavage rate log ( kcat/Km ) of more than 4 . 2 U , and eight with a rate less than 4 . 2 U . To ensure that peptides were dissimilar , we first randomly selected substrates with predicted log ( kcat/Km ) > 4 . 2 U , allowing at most four amino acids to be identical with the corresponding positions of the substrates in the dataset and in the already chosen substrates in the test set . If none of the remaining substrates met the requirements , five-amino-acid similarity was allowed . The same procedure was applied for the eight substrates with the predicted log ( kcat/Km ) < 4 . 2 U ( Table 3 , numbers 4–18 ) . Next , we consecutively chose 15 substrates predicted to be noncleavable by the CLM by HXB2 HIV-1 protease , allowing at most four amino acids to be identical at any same positions among all the substrates already selected ( including the cleavable substrates already chosen above ) . If none of the remaining substrates met the requirements , a five-amino-acid similarity was allowed ( Table 3 , numbers 19–33 ) . We then used the CLM to predict cleavability of the chosen 30 substrates by mutant HIV-1 proteases I84V , L90M , and I84V + L90M . ( The outcome for the predicted cleavability of the 30 chosen substrates was essentially the same for the mutant HIV-1 proteases as for the HXB2 HIV-1 protease . ) Following this , we then again applied the CRM and predicted the cleavage rate of the 15 cleavable substrates chosen for the three mutant HIV-1 proteases , I84V , L90M , and I84V + L90M . A total of 33 octapeptide sequences were engineered into internally quenched fluorogenic substrates . Fluorogenic substrates were synthesized by solid-phase peptide synthesis using an automated multiple peptide synthesizer ( MultiPep; Intavis AG Bioanalytical Instruments , http://www . intavis . com; Table 3 ) . Reagents were purchased from Fluka ( http://www . fluka . org ) , Applied Biosystems ( http://www . appliedbiosystems . com ) , Bachem ( http://www . bachem . com ) , or Novabiochem ( http://www . emdbiosciences . com/html/NBC/home . html ) . The following amino acid derivatives were used in the synthesis: Fmoc-L-Ala-OH , Fmoc-L-Arg ( Pbf ) -OH , Fmoc-L-Asn ( Trt ) -OH , Fmoc-L-Asp ( Ot-Bu ) -OH , Fmoc-L-Gln ( Trt ) -OH , Fmoc-L-Glu ( Ot-Bu ) -OH , Fmoc-L-Glu ( EDANS ) -OH , Fmoc-Gly-OH , Fmoc-L-Ile-OH , Fmoc-L-Leu-OH , Fmoc-L-Lys ( Boc ) -OH , Fmoc-L-Lys ( DABCYL ) -OH , Fmoc-L-Met-OH , Fmoc-L-Phe-OH , Fmoc-L-Pro-OH , Fmoc-L-Ser ( t-Bu ) -OH , Fmoc-L-Thr ( t-Bu ) -OH , Fmoc-L-Trp ( Boc ) -OH , Fmoc-L-Tyr ( t-Bu ) -OH; and Fmoc-L-Val-OH . PyBOP was used as an activating reagent and Tenta Gel R PHB-Arg ( Pbf ) -Fmoc resin ( capacity 0 . 16 mmol/g ) as a polymeric support . The peptides were synthesized at a 5-μmol scale using the automated standard protocol optimized for Fmoc chemistry provided with the MultiPep synthesizer . Each cycle included deprotection of the Fmoc group by 20% piperidine in DMF and washing of the support with DMF; coupling ( i . e . , the N-deblocked peptidyl-resin was treated with a solution of the appropriate Fmoc amino acid derivative , PyBOP , and NMM in DMF for 25 min ) and washing of the support with DMF , capping ( i . e . , treatment of the polymer with a 2% solution of acetic anhydride in DMF for 5 min ) , and washing of the support with DMF . The final synthetic step on MultiPep included deprotection with 20% piperidine in DMF , washing of the support with DMF and CH2Cl2 , and drying . The peptide was deprotected and cleaved from the resin with deprotection mixture ( TFA–triisopropylsilane–1 , 2-ethanedithiol–water , 92 . 5:2 . 5:2 . 5:2 . 5 ) for 3 h at room temperature , triturated with tert-butyl-methyl ether , taken up in MeCN/water , lyophilized , and purified by high-performance liquid chromatography ( HPLC ) ; their structures were confirmed by mass spectrometry . Analytical HPLC was performed on a Waters ( http://www . waters . com ) system ( Millenium32 workstation , 2690 separation module , 996 photodiode array detector ) equipped with Vydac RP C18 90 Å reversed-phase column ( 2 . 1 × 250 mm; http://www . vydac . com ) . Small-scale preparative HPLC was carried out on a system consisting of a 2150 HPLC pump , 2152 LC controller , and 2151 variable wavelength monitor ( LKB , Sweden ) and Vydac RP C18 column ( 10 mm × 250 mm , 90 Å , 201HS1010 ) , with the eluent , an appropriate concentration of MeCN in water + 0 . 1% TFA , a flow rate of 5 mL/min , and detection at 280 nm . Freeze-drying was carried out at 0 . 01 bar on a Lyovac GT2 Freeze-Dryer ( Steris Finn-Aqua , http://www . steris . com ) equipped with a Trivac D4B ( Leybold Vacuum , http://www . oerlikon . com ) vacuum pump and a liquid nitrogen trap . Peptides were checked by liquid chromatography/mass spectrometry using a Perkin Elmer PE SCIEX API 150EX instrument equipped with a turboionspray ion source ( PerkinElmer Life and Analytical Sciences , http://las . perkinelmer . com ) and a Dr . Maisch Reprosil-Pur C18-AQ ( http://www . dr-maisch . com ) , 5 μm , 150 mm × 3 mm HPLC column , using a gradient formed from water and acetonitrile with 5 mM ammonium acetate additive . When not otherwise specified , chemicals were reagent grade from Sigma ( http://www . sigmaaldrich . com ) . Wild-type HIV-1 protease ( HXB2 clone ) and its three mutants , the I84V , L90M , and I84V + L90M genes , were a kind gift of Professor Helena Danielson , Uppsala University . Protease expression , isolation , refolding , and analysis were performed as previously described [72] . Rates of cleavage of the synthesized internally quenched fluorogenic substrates by the HXB2 and mutant HIV-1 proteases were assayed fluorimetrically ( ex , 355 nm; em 490–10 nm ) in black 96-well plates ( Nunc , http://www . nuncbrand . com ) under the conditions detailed in Table S3 , assay 17 , using a PolarSTAR OPTIMA microplate reader ( BMG LABTECH , http://www . bmglabtech . com ) . Substrate stock solutions were 1 mM dissolved in DMSO ( Table 3 , numbers 1–33 ) . A typical reaction mixture ( total volume 100 μL ) contained variable concentrations of peptide substrates in 0 . 1 M acetic acid and 1 . 1 M sodium chloride ( pH 5 . 0 was achieved with sodium hydroxide solution ) and 35 ng of enzyme . Reaction was conducted at 37 °C for 60 min ( cycle time , 60 s , with 5 s shaking after each cycle ) . Each experiment was repeated at least three times , and the average value was taken as a final result ( Table 3 ) . The kinetic data was analyzed by nonlinear fit using the GraFit program and the basic equation for Michaelis–Menten kinetics [73] . The obtained kcat/Km constants were converted into mM−1h−1 units for before further use .
The Protein Data Bank ( http://www . pdb . org ) accession numbers for the proteins discussed in this paper are HIV-1 protease ( 1aid ) , HIV-2 protease ( 1ida ) , HTLV-1 protease ( 2b7f ) , FIV protease ( 4fiv ) , RSV protease ( 1bai ) , AMV protease ( 1mvp ) , and EIAV protease ( 1fmb ) . The Swiss-Prot ( http://www . expasy . org/sprot ) accession numbers for the proteases discussed in this paper are Mo-MuLV protease ( P03355 ) and BLV protease ( P10270 ) .
|
Retroviruses are associated with a broad range of diseases that include tumor formation , neurological disorders , and immunodeficiency syndromes , including those of HIV . The extraordinary mutational plasticity of HIV-1 causes the rapid appearance of highly diverse quasi-species in a very short time , leading to severe problems with drug resistance . We here present and validate experimentally a novel approach for the analysis of the molecular interaction networks involved in the recognition process of substrates by natural and drug-resistant retroviral proteases . By combining a large number of wild-type and mutant retroviral proteases from nine different viral species , and their interactions with a large number of substrates , we have created a unified model incorporating all the proteases' mutational space . Our results reveal that a complex physicochemical interaction network is involved in substrate recognition and cleavage by aspartate proteases and unravel detailed molecular mechanisms involved in drug resistance . These findings provide novel implications for understanding important features of HIV resistance and raise the possibility of developing completely novel strategies for the design of protease inhibitors that will remain effective over time despite rapid viral evolution .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"biochemistry",
"infectious",
"diseases",
"hiv",
"drug",
"resistance",
"multivariate",
"modelling",
"proteases",
"computational",
"biology",
"retroviruses"
] |
2007
|
A Look Inside HIV Resistance through Retroviral Protease Interaction Maps
|
Recent in-vitro studies have suggested that mast cells are involved in Dengue virus infection . To clarify the role of mast cells in the development of clinical Dengue fever , we compared the plasma levels of several mast cell-derived mediators ( vascular endothelial cell growth factor [VEGF] , soluble VEGF receptors [sVEGFRs] , tryptase , and chymase ) and -related cytokines ( IL-4 , -9 , and -17 ) between patients with differing severity of Dengue fever and healthy controls . The study was performed at Children's Hospital No . 2 , Ho Chi Minh City , and Vinh Long Province Hospital , Vietnam from 2002 to 2005 . Study patients included 103 with Dengue fever ( DF ) , Dengue hemorrhagic fever ( DHF ) , and Dengue shock syndrome ( DSS ) , as diagnosed by the World Health Organization criteria . There were 189 healthy subjects , and 19 febrile illness patients of the same Kinh ethnicity . The levels of mast cell-derived mediators and -related cytokines in plasma were measured by ELISA . VEGF and sVEGFR-1 levels were significantly increased in DHF and DSS compared with those of DF and controls , whereas sVEGFR-2 levels were significantly decreased in DHF and DSS . Significant increases in tryptase and chymase levels , which were accompanied by high IL-9 and -17 concentrations , were detected in DHF and DSS patients . By day 4 of admission , VEGF , sVEGFRs , and proteases levels had returned to similar levels as DF and controls . In-vitro VEGF production by mast cells was examined in KU812 and HMC-1 cells , and was found to be highest when the cells were inoculated with Dengue virus and human Dengue virus-immune serum in the presence of IL-9 . As mast cells are an important source of VEGF , tryptase , and chymase , our findings suggest that mast cell activation and mast cell-derived mediators participate in the development of DHF . The two proteases , particularly chymase , might serve as good predictive markers of Dengue disease severity .
Dengue virus infection is associated with disease , ranging from Dengue fever ( DF ) to Dengue hemorrhagic fever ( DHF ) and/or Dengue shock syndrome ( DSS ) . As severe diseases typically develop in individuals suffering secondary Dengue virus infection , host immunological factors appear to play a role in DHF and DSS [1] . DHF and DSS are characterized by increased vascular permeability and hemorrhagic manifestations [2] , with the former phenotype recognized as the hallmark of these severe forms of Dengue . However , the cellular factors and immune molecules underling the development of DHF and DSS are not well understood . Recent studies on Dengue virus infection have demonstrated that the serum levels of vascular endothelial cell growth factor ( VEGF ) -A ( formerly VEGF ) are elevated in DHF patients [3] . VEGF/vascular permeability factor ( VPF ) was first identified and characterized as a potent stimulator of endothelial permeability [4] , and was shown to increase vascular permeability 50 , 000 fold more efficiently than histamine [5] . VEGF was subsequently reported to promote the proliferation , migration , and survival of endothelial cells [6] . VEGF is a member of a growing family of related proteins that includes VEGF-B , -C , -D , and placental growth factor [7] . A potential candidate for the VEGF-binding molecule is the soluble form of its receptor . At least two types of VEGF receptors are expressed on endothelial cells; both are transmembrane receptor tyrosine kinases , namely , VEGFR-1 or Fms-like tyrosine kinase 1 ( Flt-1 ) , and VEGFR-2 or kinase insert domain receptor ( KDR ) [8] . VEGFR-1 is expressed on monocyte-macrophage lineages other than endothelial cells , whereas VEGFR-2 is expressed primarily on endothelial cells and their progenitors [9] , [10] . In addition to its role in promoting endothelial permeability and proliferation , VEGF may contribute to inflammation and coagulation . For example , under in-vitro conditions , VEGF induces the expression of several types of cell adhesion molecules , including E-selectin , intercellular adhesion molecule 1 ( ICAM-1 ) , and vascular cell adhesion molecule 1 ( VCAM-1 ) , in endothelial cells and promotes the adhesion of leukocytes [11] , [12] . Moreover , VEGF signaling up-regulates tissue factor mRNA expression , and protein and procoagulant activities [13] . The proinflammatory/procoagulant effects of VEGF are mediated , at least in part , by the activation of the transcription factors NF-κB , Egr-1 , and NFAT . VEGF has been implicated as a pathophysiological mediator in several human disease states , including rheumatoid arthritis , cancer , and inflammatory bowel disease [14]–[16] . Dengue patients typically exhibit increased levels of urinary histamine , which is a major granule product of mast cells and whose levels correlate with disease severity [17] . A large autopsy study of 100 DHF cases from Thailand found that mast cells in connective tissue around the thymus exhibited swelling , cytoplasmic vacuolation , and loss of granule integrity , which are suggestive of mast cell activation [18] . Although recent in-vitro studies have also reported the involvement of mast cells in Dengue virus infection [19] , [20] , the potential role of mast cells in severe Dengue disease has not yet been explored . The activation of mast cells , which reside mainly in tissues and are associated closely with blood vessels and nerves [21] , [22] , is tightly linked with local increases in vascular permeability in allergic disease . Mast cells are key effector cells in IgE-dependent immune responses , such as those involved in the pathogenesis of allergic disorders or in certain instances of immunity to parasites [23] . Recent works have revealed another aspect of mast cell effector function , and mast cells play important roles in inflammation and host defenses against foreign pathogens [24] , [25] . Mast cells synthesize and release a range of biologically active substances , including proteases , biogenic amines , cytokines , chemokines , and lipid mediators [26] . Mast cell proteases are key protein components of secretory mast cell granules and are essential for innate antimicrobial inflammatory responses [27]–[29] . It is estimated that mast cell proteases account for >25% of total mast cell protein [30] and that human skin mast cells contain a total of ∼16 µg tryptase and chymase per 106 cells [31] . Mast cell proteases , tryptase , and chymase are serine proteases with trypsin- or chymotrypsin-like substrate specificities , and are the major proteins stored and secreted by mast cells . Measurement of the serum ( or plasma ) levels of these proteases are recommended in the diagnostic evaluation of systemic anaphylaxis and mastocytosis , with total tryptase levels generally reflecting either the increased burden of mast cells in patients with all forms of systemic mastocytosis , or the decreased burden of mast cells associated with cytoreductive therapies in these disorders . Tryptase and chymase levels generally reflect the magnitude of mast cell activation and are typically elevated during systemic anaphylaxis . Secreted tryptase and chymase promote inflammation , matrix destruction , and tissue remodeling by several mechanisms , including the destruction of procoagulant , matrix , growth , and differentiation factors , and the activation of proteinase-activated receptors , urokinase , metalloproteinases , and angiotensin . In addition , these two serine proteases also modulate immune responses by hydrolyzing chemokines and cytokines , and can also suppress inflammation by inactivating allergens and neuropeptides responsible for inflammation and bronchoconstriction . Thus , similar to mast cells themselves , mast cell serine proteases play multiple roles in host defenses , which may be either beneficial or harmful depending on the specific conditions . As substantial levels of tryptase and chymase are only found in mast cells , these proteases are considered to be selective markers of mast cell activation [26] . The importance of cytokines and chemokines together with mast cells in the pathogenesis of Dengue virus infection has been demonstrated [19] , [20] , however , the roles of the mast cell-specific proteases , tryptase and chymase , remain unclear . Here , to determine the roles of mast cells and mast cell-derived mediators in DHF and DSS , we first measured the levels of VEGF , soluble forms of VEGFR-1 and -2 , tryptase , and chymase in the plasma of Dengue patients and healthy control subjects . Moreover , because IL-9 has been reported as a T cell-derived growth factor of mast cells [32]–[34] and more recently has been implicated as a Th17-derived cytokine that contributes to inflammatory diseases , the involvement of IL-9 and IL-17 in Dengue infection was also investigated .
The study was performed at two hospitals , Children's Hospital No . 2 in Ho Chi Minh City ( HCMC ) and the Center for Preventive Medicine in Vinh Long Province ( VL ) , Vietnam . The enrolment was a consecutive sequence of hospitalized children at each hospital . The inclusion criteria on admission to the hospital were age ( 6 months to 15 years old ) and ethnicity ( Kinh race ) . A total of 103 subjects from HCMC and VL were enrolled in this study during 2002–2005 ( Table 1 ) . The patients were suspected to have Dengue virus infection based on clinical symptoms at admission . After hospitalization , the patients were diagnosed using standardized serology techniques , as described below , and the WHO ( 1997 ) classification criteria for Dengue virus infection [35] . It was reported that the sensitivity of WHO criteria for DSS in Vietnam was only 82% , mainly due to the lack of evidence for thrombocytopenia [36] . Therefore , we basically followed the WHO criteria , but included patients lacking a significant reduction of platelet count , which accounted for no more than 11% of all DHF/DSS cases . Our classification scheme met the requirements of the simplified Integrated Management of Childhood Illness ( IMCI ) classification system , which is based on plasma leakage as a hallmark of severe dengue disease ( DHF/DSS ) [37] . Plasma samples were obtained from the 19 DF , 43 DHF , and 41 DSS patients on the day of admission , and an additional 189 plasma samples from healthy , unrelated school children living in HCMC and VL who had no symptoms of Dengue virus infection were collected as control samples . Eighteen ( male: 12 , female: 6 ) plasma samples were also collected from school children with a febrile illness ( 38 . 9±0 . 9°C: mean±SD ) without an obvious source of infection , including Dengue virus . This study was approved by the institutional ethical review committees of the Institute of Tropical Medicine , Nagasaki University , Jikei University School of Medicine in Tokyo , and the Pasteur Institute in Ho Chi Minh City . Written informed consent was obtained from the parents or legal guardians of the subjects upon enrollment . The sample collection and serological diagnosis performed in this cohort study were identical to those reported in our previous study [38] . Blood samples were collected from patients with suspected Dengue infection at the time of admission ( day 0 ) and twice during the following four days ( days 2 and 4 ) . Plasma samples were used for the titration of anti-Dengue virus IgM and IgG antibodies , virus isolation , and RT-PCR for the determination of viral serotype . Dengue virus infection was determined by previously established serologic criteria for IgM/IgG ELISAs to Dengue virus ( DEN 1–4 ) and Japanese encephalitis virus in paired plasma , collected with at least three-day intervals [39] . IgM and IgG ELISAs were performed using kits obtained from the Pasteur Institute , HCMC and were considered positive if the ratio of optical density ( OD ) of test sera to the OD of negative control plasma was ≥2 . 3 [35] . The cases were diagnosed as secondary infection when the DV IgM-to-IgG ratio was <1 . 8 [40] . Dengue virus serotyping was performed as previously reported [38] . Briefly , acute plasma samples were used to inoculate C6/36 ( Aedes albopictus ) cells , virus was obtained , and the Dengue virus serotype was then identified using either a direct or indirect fluorescent antibody technique with monoclonal antibodies supplied by the Centers for Disease Control and Prevention ( Fort Collins , CO , USA ) [41] . Viral RNA was also extracted from the acute plasma samples with the QIAamp Viral RNA Mini Kit ( Qiagen , Hilden , Germany ) for the molecular detection of Dengue virus and confirmation of its serotype , as previously described [41] . Briefly , cDNA from the Dengue virus genome RNA was amplified with the Ready-to-go RT-PCR test kit ( Amersham , MA , USA ) using a consensus primer set ( D1 and D2 ) [31] . The serotype was then determined by semi-nested PCR using specific primer sets ( TS1 , TS2 , TS3 , and TS4 ) to amplify serotype-specific fragments from the regions encoding the capsid and membrane proteins of Dengue virus [39] . The plasma levels of VEGF ( VEGF-A ) , sVEGFR-1 , sVEGFR-2 , IL-9 , and IL-17 in samples from Dengue patient ( DF , DHF , and DSS ) and control groups ( febrile illness and healthy subjects ) were measured by ELISA kits ( R&D Systems , Minneapolis , MN , or Peprotech Inc . , Rocky Hill , NJ ) . The levels of tryptase or chymase in plasma from the Dengue patients and control groups , and the culture supernatants of mast cells were examined by ELISA kits ( CSB , Newark , ED or Otsuka Pharmaceutical Co . , Tokushima , Japan ) . The human mast cell/basophil line KU812 [42] and human mast cell line HMC-1 [43] were maintained in RPMI 1640 medium or IMDM ( Invitrogen , Grand Island , NY ) . In the infection experiments , Dengue virus 2 ( DV16681 strain ) was propagated in the C6/36 cell line , and virus titers were then determined by plaque assay using BHK-21 cells [44] . In control experiments , the virus was rendered nonreplicative by placing a sample aliquot under a germicidal lamp ( 125 mJ/10 min , UV irradiation at 254 nm ) at a distance of 5–6 cm , followed by a 30-min incubation on ice [44] . For infection , HMC-1 or KU812 cell pellets were adsorbed at 4°C for 90 min with aliquots of Dengue virus or UV-inactivated virus , Dengue virus in combination with human dengue virus immune serum ( 1∶1 , 000 or 1∶10 , 000 final dilution ) , UV-inactivated virus in combination with human dengue virus immune serum ( 1∶1 , 000 or 1∶10 , 000 final dilution ) , or Dengue virus in combination with normal human serum ( 1∶1 , 000 final dilution ) ( premixed at 4°C for 90 min ) . Dengue virus 2 convalescent-phase sera were used in the antibody-dependent enhancement of Dengue virus infection . Mast cells were infected at a multiplicity of infection ( MOI ) of 3 plaque forming units ( pfu ) /cell . Following adsorption , cells were washed and plated in 96-well plates ( 0 . 25 mL/well ) at 1×106 cells/mL and then incubated at 37°C in 5% CO2 for 24 h . To examine the effects of IL-9 on the production of VEGF from mast cells , mast cells treated with Dengue virus and antibody were incubated with or without recombinant human IL-9 ( 200 ng/mL , Peprotech , Rocky Hill , NJ ) . Activation of mast cells with Compound ( C ) 48/80 ( 300 µg/ml , Sigma-Aldrich , St . Louis , MO ) was used as a positive control , and the culture supernatant of C6/36 cells was used as a negative control . For the measurement of VEGF levels in culture supernatants , culture supernatants were collected from each well after incubation and then stored at −80°C until being subjected to ELISA . KU812 cells were inoculated with Dengue virus ( MOI , 3 ) or Dengue virus-antiserum combinations . After incubation for 24 h , cells were fixed with 4% paraformaldehyde , washed , and then permeabilized with 0 . 1% saponin for 1 h at room temperature . Samples were then washed and incubated with mouse anti-Dengue virus monoclonal antibody 1B7 [45] and isotype-matched mouse IgG2a antibody ( negative control , R&D Systems ) on ice for 1 h , which were employed as primary antibodies . Subsequently , samples were washed and incubated with FITC-labeled anti-mouse IgG antibody ( R&D Systems ) for 1 h on ice . Cytospins were made for each sample , and positive cells were observed by fluorescence microscopy . Plasma VEGF , sVEGFRs , IL-9 , IL-17 , tryptase , and chymase levels were compared between the Dengue ( DF , DHF , or DSS ) and control groups ( febrile illness and healthy subjects ) using the unpaired Student's t test . VEGF levels in the in-vitro experiments were also compared between the Dengue virus infection and control ( UV-inactivated Dengue virus and Medium alone ) samples using the unpaired Student's t test . A value of p<0 . 05 was considered statistically significant .
As mast cells are an important source of VEGF [46] , [47] , we first measured VEGF levels in plasma samples from the DF ( n = 19 ) , DHF ( n = 43 ) , and DSS ( n = 41 ) patient groups , and the control group , which consisted of febrile illness and healthy subjects . On day 0 ( admission ) , the VEGF plasma levels were significantly higher in DHF and DSS than those in DF , and febrile illness and healthy subjects ( Fig . 1A ) . The sVEGFR-1 levels in plasma were higher in DSS than those in DF , DHF , febrile illness and healthy subjects ( Fig . 1B ) . In contrast with sVEGR-1 , the levels of sVEGFR-2 were dramatically decreased in DHF and DSS compared with DF or febrile illness and healthy subjects ( Fig . 1C ) . We next examined the levels of VEGF and sVEGFRs in DHF ( n = 21 ) and DSS ( n = 27 ) patients during the admission period ( Fig . 2 ) . The VEGF levels in DHF and DSS , and sVEGFR-1 levels in DSS were significantly higher than those of DF or healthy controls on the day of admission ( day 0 ) ; however , 2–4 days later ( convalescence ) , their levels had gradually declined to comparable levels with DF , febrile illness , and healthy subjects by the convalescent phase ( day 4; VEGF DF: 0 . 61±+0 . 24 ng/ml , febrile illness: 0 . 57±0 . 11 ng/ml , and healthy subjects: 0 . 52±0 . 17 ng/ml; sVEGFR-1 DF: 180 . 9+55 . 3 pg/ml , DHF: 223 . 6±136 pg/ml , febrile illness: 201 . 3±167 . 1 pg/ml , and healthy subjects: 195 . 1±59 . 1 pg/ml ) . The plasma levels of sVEGFR-2 in DHF and DSS patients were significantly lower compared to those of DF , febrile illness , and healthy subjects; however , the levels were comparable between these groups by day 4 ( Fig . 2 ) . Taken together , these findings suggested the possibility that VEGF and sVEGFRs participated in severe Dengue virus infection . We also measured the tryptase and chymase levels in plasma collected from the Dengue patients ( day 0 ) and controls by ELISA . Plasma tryptase levels increased significantly in DHF and DSS compared with DF , febrile illness , and healthy subjects ( Fig . 3 ) . In contrast , the chymase levels were increased significantly in DSS compared with DF , DHF , febrile illness , and healthy subjects ( Fig . 3 ) . We next measured the plasma levels of tryptase and chymase in DHF ( n = 21 ) and DSS ( n = 27 ) patients during the admission period and found that the protease levels had gradually declined by days 2 and 4 to a comparable level with those of DF , febrile illness , and healthy subjects ( chymase DF: 4 . 8±2 ng/ml , DHF: 6 . 7±2 . 4 ng/ml , febrile illness: 6 . 2±2 ng/ml , healthy subjects: 4 . 5±1 . 3 ng/ml , tryptase DF: 7 . 4±4 . 3 ng/ml , febrile illness: 6 . 7±2 . 6 ng/ml , healthy subjects: 7 . 7±3 . 4 ng/ml ) ( Fig . 2 ) . These results suggested that mast cells and mast cell-derived proteases participated in the severe form of Dengue virus infection . As IL-9 has been reported as a T cell-derived mast cell growth factor [32]–[34] and more recently , is implicated as a Th17-derived cytokine that can contribute to inflammatory diseases , we investigated the involvement of IL-9 and IL-17 in Dengue virus infection . The levels of IL-9 and IL-17 in Dengue patients on day 0 , and those in blood samples collected from febrile illness and healthy subjects were measured by ELISA . The analysis showed that IL-9 and IL-17 levels were significantly increased in DHF and DSS compared with those in DF , febrile illness , and healthy subjects ( Figs . 4A and B ) . Although these results suggested that IL-9 and IL-17 participate in Dengue virus infection , IL-9 may act additively or synergistically with other factors , such as other Th2 cytokines , to induce optimal mast cell responses . To examine the possibility that Th2 cytokines affect mast cell responses in Dengue virus infection , IL-4 levels were also examined in plasma from Dengue patients and control groups ( Fig . 4C ) . We found comparable levels of IL-4 between Dengue patients and control groups , suggesting the involvement of IL-9 and -17 in Dengue virus infection . To investigate if Dengue virus induces VEGF production from mast cells , the in-vitro production of VEGF in the human mast cell lines KU812 and HMC-1 was examined . KU812 and HMC-1 cells were inoculated with Dengue virus in the presence of either human Dengue virus-immune or normal human serum , and VEGF levels in the culture medium were assessed 24 h after viral inoculation . As the antibody-dependent enhancement of infection in KU812 and HMC-1 cells was observed at 1∶1 , 000 and 1∶10 , 000 dilutions of human Dengue virus-immune serum in preliminary experiments ( data not shown ) , a 1∶1 , 000 dilution was used in the in-vitro experiments in this study . The production of VEGF was observed in both KU812 and HMC-1 cells after exposure to Dengue virus in the presence human Dengue virus-immune serum , however , VEGF levels were higher in KU812 cells ( Table 2 ) . No significant increase of VEGF level was observed when Dengue virus was inoculated with normal human serum ( 1∶1 , 000 final dilution ) or when UV-inactivated Dengue virus was inoculated with human Dengue virus immune or normal human serum . In addition , no VEGF production by KU812 and HMC-1 cells was observed after mock-infection with human Dengue immune or normal human serum . These results suggested the importance of antibody to Dengue virus for mast cell secretion of VEGF in vitro . As it is known that KU812 and HMC-1 cells are permissive to Dengue virus infection when the virus is inoculated together with human Dengue immune serum [20] , the antibody-dependent infection of KU812 cells with Dengue virus was examined by immunofluorescence analysis in the presence and absence of human Dengue immune serum 24 h after the inoculation . Positive immunofluorescence was only observed in cells infected in the presence of human Dengue virus-immune serum , suggesting the occurrence of permissive infection of Dengue virus ( Fig . 5 ) . To determine the role of IL-9 in VEGF production by mast cells , KU-812 and HMC-1 cells were inoculated with Dengue virus and human Dengue virus-immune serum ( 1∶1 , 000 final dilution ) in the presence and absence of IL-9 . Although a low level of VEGF production by KU-812 and HMC-1 cells was observed without IL-9 , VEGF levels were significantly increased in the presence of IL-9 ( Table 3 ) . The effect of IL-9 on VEGF production by KU812 and HMC-1 cells was not observed in the presence of normal human serum ( data not shown ) . Taken together , these findings suggested the possibility that Dengue virus induces VEGF secretion from human mast cells during infection , and that IL-9 supports the production of VEGF in mast cells .
Recently , Srikiatkhachorn et al . [48] compared the plasma levels of VEGF-A and sVEFGR-1 and -2 between DHF and DF patients , and found a rise of VEGF-A and decline of sVEGFR-2 levels in DHF patients , with the severity of plasma leakage inversely correlating with sVEGFR-2 levels . These findings seemed to be consistent with our present results that VEGF and sVEGFR-2 were significantly increased and reduced , respectively , in DHF and DSS patients . Although the reason why sVEGFR-2 levels are decreased in DHF and DSS patients is not clear , as VEGF binding to VEGFR-2 on endothelial cells results in receptor phosphorylation , changes in endothelial cell morphology and proliferation , and maintenance of physiological condition of blood vessels , decreased sVEGFR-2 levels in severe Dengue patients might represent the dysfunction of homeostasis in vascular endothelial cells and correlate with increased plasma leakage [49] . We additionally observed a significant increase of sVEGFR-1 levels in DSS patients , which suggests that activation of monocytes/macrophages by Dengue virus leads to increased expression of soluble and surface VEGFR-1 on cells during severe Dengue infection , as was previously reported [49] . Regarding the relationship between VEGF level and severity of Dengue virus infection , Tseng et al . [3] observed the elevation of circulating VEGF levels in adult DHF patients during the early phases of Dengue infection , as compared to DF patients and healthy controls . In a study of a pediatric population with DHF , Srikiatkhachorn et al . [6] also detected a rise in circulating VEGF in the early febrile and defervescent stages of Dengue infection , but not during the later convalescent stage . However , subsequent studies reported contradictory findings , as increased circulating VEGF concentrations were not observed during the early febrile and toxic stages in DHF , but lower VEGF concentrations were detected in patients with more severe Dengue infection [50]–[52] . Several underlying reasons may explain these differences , such as poor study design , small sample size , and the lack of a standardized collection methodology and storage of blood samples used for the measurement of VEGF . In addition , VEGF is also expressed at low levels in a wide variety of normal adult human and animal tissues , and at higher levels in a few selected sites , namely , podocytes of the renal glomerulus , cardiac myocytes , prostatic epithelium and semen , and certain epithelial cells of the adrenal cortex and lung [53] . Dovrak et al . [54] reported that VEGF is substantially overexpressed at both the mRNA and protein levels in a high percentage of malignant animal and human tumors , as well as in many transformed cell lines . Thus , studies of VEGF production by mast cells during Dengue virus infection are complicated by these alternate sources of VEGF in human and animals , and may affect circulating VEGF levels . Incubation of KU812 and HMC-1 cells with Dengue virus in the presence of human Dengue virus-immune serum resulted in enhanced VEGF production , which was not observed when UV-inactivated Dengue virus was incubated with human Dengue virus-immune serum or when Dengue virus was used alone to infect KU812 cells ( Table 2 ) . As the permissive infection of Dengue virus was observed in KU812 cells ( Fig . 5 ) , these findings suggest the critical importance of antibodies to Dengue virus for VEGF production by highly infected mast cells and indicate that infected mast cells can secrete VEGF without stimulation through FcεRI . Our results appear consistent with the findings that Dengue virus infection induces the production of chemokines by human mast cells without stimulation of FcεRI in the presence of human Dengue immune serum [22] . Brown et al . [55] reported that FcγRII plays a dominant role in antibody-enhanced Dengue virus infection of the human mast cell lines HMC-1 and KU812 , and in the associated CCL5 release . In studies of DHF epidemics , Halstead et al . [56] and Guzman et al . [57] demonstrated that secondary infection is the most important host risk factor for DHF . Boesiger et al . and Grützkau et al . [46] , [47] reported that mouse and human mast cells produce and secrete VPF/VEGF , and release VEGF upon stimulation through FcεRI or after challenge with chemical mast cell activators . Notably , the FcεRI-dependent secretion of VEGF by either mouse or human mast cells is significantly increased in cells that have undergone upregulation of FcεRI surface expression by preincubation with IgE . As Koraka [58] reported that Dengue virus-specific IgE levels were significantly higher in DHF and DSS patients compared to those in DF and non-Dengue patients , FcεRI may be important for mast cell activation via IgE antibody in Dengue virus infection . However , we did not determine whether the patient sera collected in the present contained IgE antibody against Dengue virus . To clarify the importance of FcεRI for VEGF production by mast cells in Dengue virus infection , further studies are needed . High levels of VEGF in culture supernatants were detected when KU812 and HMC-1 cells were cultured in the presence of IL-9 ( Table 3 ) . As IL-9 enhances the survival of mast cells and induces their production of proinflammatory cytokines , including Th1 and Th2 cytokines [59] , it is possible that IL-9 primes HMC-1 and KU812 cells in vitro to respond to Dengue virus infection by promoting VEGF production . To evaluate the contribution of IL-9 and IL-17 to Dengue virus infection , we measured the plasma levels of these two cytokines in Dengue patients and found that both IL-9 and IL-17 were significantly increased in DHF and DSS compared with DF , febrile illness and healthy subjects . These findings suggest that Th9 and Th17 cells contribute to the inflammatory response to severe Dengue virus infection . It is possible that IL-9 may act additively or synergistically with other factors , such as additional Th2 cytokines , to induce the mast cell response observed in this study . However , as the level of IL-4 was not increased in the plasma of Dengue patients , our findings suggest the independent involvement of IL-9 secreted by Th2 cells in Dengue virus infection . Recently , IL-9-producing cells have been described as a new subset of the T helper cell population separate from Th2 that produces IL-9 in large quantities and contributes uniquely to immune responses [60] , [61] . This cell population has been named ‘Th9’ , and IL-9 secreted by T cells , particularly Th9 cells , may regulate chronic allergic inflammation [62] . Moreover , IL-9 has been recently proposed to function as a Th17-derived cytokine that contributes to inflammatory diseases [36] . Tryptase and chymase levels were significantly increased in DHF and DSS , and DSS , respectively , on admission compared with DF , febrile illness , and healthy subjects ( Fig . 3 ) . However , 2–4 days after admission , the levels of these proteases had returned to similar levels with the other patient groups ( Fig . 2 ) . These findings support the concept that mast cells and mast cell degranulation play important roles in the pathogenesis of DHF/DSS and might be suitable targets for new therapies and prevention of Dengue infection . However , it is presently unclear whether Dengue virus infection in mast cells directly induces chymase and tryptase production and secretion . Recently , Kitamura-Inenaga et al . [63] reported that encephalomyocarditis virus infection results in mast cell chymase and tryptase production in vivo , and additionally , viral infections have been shown to cause the accumulation of mast cells in the nasal mucosa during the first days of a symptomatic , naturally acquired respiratory infection [64] . However , the relevance and underlying mechanisms of mast cell infection and activation in the setting of viral infections remain to be characterized in detail . Immunocytohistochemical studies in human tissues have identified two mast cell phenotypes distinguishable by their neutral protease content , namely the ‘mast cell-tryptase’ ( MCT ) phenotype and the ‘mast cell-tryptase-chymase’ ( MCTC ) phenotype [65] . MCT appears to be associated with immune system-related mast cells that play a primary role in host defenses and are preferentially located at mucosal surfaces . MCT mast cells are increased in number in areas of T lymphocyte infiltration and in allergic disease , and are reduced in number in acquired and chronic immunodeficiency syndromes [65] . In contrast , the MCTC phenotype appears to be associated with non-immune system-related mast cells that primarily function in angiogenesis and tissue remodeling , rather than immunologic protection , and are found predominantly in submucosal and connective tissues . In addition , MCTC mast cells are not increased in numbers in areas of heavy lymphocytic infiltration and are not decreased in number in immunodeficiency syndromes [65] . In the present study , a significant increase of chymase was observed in the plasma of DSS patients as compared with those of DF , DHF , and the control group , suggesting the possibility that MCTC mast cells contribute to the pathogenesis of severe forms of Dengue virus infection . However , further study is needed to clarify the roles of tryptase and chymase in severe Dengue virus infection . Concerning the ability of mediators produced by mast cells other than VEGF to activate endothelial cells , King et al . [22] reported that Dengue virus plus Dengue virus-specific antibody treatment results in selective production of the T-cell chemoattractants RANTES , MIP-1α , and MIP-1β by KU812 and HMC-1 human mast cell-basophil lines . In addition , Brown et al . [66] demonstrated that antibody-enhanced Dengue virus infection of primary human cord blood-derived mast cells ( CBMCs ) and HMC-1 cells results in the release of ICAM-1 and VCAM-1 , which subsequently activate human endothelial cells . St . John et al . [67] reported that the response to mast cell activation involves the de novo transcription of cytokines , including TNF-α and IFN-α , and chemokines , such as CCL5 , CXCL12 , and CX3CL1 , which are well characterized to recruit immune effector cells , including cytotoxic lymphocytes , to sites of peripheral inflammation . In conclusion , we found that mast cells and mast cell-derived mediators , namely VEGF , and the mast cell-specific proteases tryptase and chymase participate in the development of severe forms of Dengue virus infection , which is accompanied by elevated circulating levels of IL-9 and -17 . As tryptase and chymase are known as selective markers of non-immune system-related activation of mast cells in submucosal and connective tissues , these two proteases , particularly chymase , might serve as good predictive markers of Dengue disease severity .
|
To clarify the involvement of mast cells in the development of severe Dengue diseases , plasma levels of mast cell-derived mediators , namely vascular endothelial cell growth factor ( VEGF ) , tryptase , and chymase , were estimated in Dengue patients and control subjects in Vietnam . The levels of the mediators were significantly increased in Dengue hemorrhagic fever ( DHF ) and Dengue shock syndrome ( DSS ) patients compared with those of Dengue fever ( DF ) and control ( febrile illness and healthy subjects ) patients , and the soluble form of VEGF receptors ( sVEGFR ) -1 and -2 levels were significantly changed in the patients with severe disease . After 2–4 days of admission , the mediator levels had returned to similar levels as those of DF and control subjects . Furthermore , the levels of the Th17 cell-derived mast-cell activators IL-9 and -17 were increased in DHF and DSS . In-vitro production of VEGF in human mast cells was significantly enhanced in the presence of IL-9 when these cells were inoculated with Dengue virus in the presence of human Dengue virus-immune serum . As mast cells are an important source of VEGF , and tryptase and chymase are considered to be specific markers for mast cell activation , mast cells and mast cell-derived mediators might participate in the development of DHF/DSS .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"neglected",
"tropical",
"diseases"
] |
2012
|
Association of Mast Cell-Derived VEGF and Proteases in Dengue Shock Syndrome
|
Mycobacterium africanum , made up of lineages 5 and 6 within the Mycobacterium tuberculosis complex ( MTC ) , causes up to half of all tuberculosis cases in West Africa , but is rarely found outside of this region . The reasons for this geographical restriction remain unknown . Possible reasons include a geographically restricted animal reservoir , a unique preference for hosts of West African ethnicity , and an inability to compete with other lineages outside of West Africa . These latter two hypotheses could be caused by loss of fitness or altered interactions with the host immune system . We sequenced 92 MTC clinical isolates from Mali , including two lineage 5 and 24 lineage 6 strains . Our genome sequencing assembly , alignment , phylogeny and average nucleotide identity analyses enabled us to identify features that typify lineages 5 and 6 and made clear that these lineages do not constitute a distinct species within the MTC . We found that in Mali , lineage 6 and lineage 4 strains have similar levels of diversity and evolve drug resistance through similar mechanisms . In the process , we identified a putative novel streptomycin resistance mutation . In addition , we found evidence of person-to-person transmission of lineage 6 isolates and showed that lineage 6 is not enriched for mutations in virulence-associated genes . This is the largest collection of lineage 5 and 6 whole genome sequences to date , and our assembly and alignment data provide valuable insights into what distinguishes these lineages from other MTC lineages . Lineages 5 and 6 do not appear to be geographically restricted due to an inability to transmit between West African hosts or to an elevated number of mutations in virulence-associated genes . However , lineage-specific mutations , such as mutations in cell wall structure , secretion systems and cofactor biosynthesis , provide alternative mechanisms that may lead to host specificity .
Mycobacterium africanum is a member of the Mycobacterium tuberculosis complex ( MTC ) that causes up to half of all tuberculosis cases in West Africa [1] . First identified by Castets in 1968 , it was originally characterized as having biochemical characteristics intermediate between Mycobacterium tuberculosis , which consists of lineages 1 , 2 , 3 , 4 , and 7 and is the main cause of human tuberculosis , and Mycobacterium bovis , an animal-adapted lineage that causes bovine tuberculosis [2] . Later work divided M . africanum into two lineages , M . africanum West African type I and M . africanum West African type II , which became known as lineages 5 and 6 , respectively , within the MTC [3 , 4] . Lineages 5 and 6 cause a disease similar to classically defined M . tuberculosis , although it has been suggested that human disease caused by these lineages may differ compared to that caused by lineages 1–4 . For example , patients with lineage 6 disease have been reported to show attenuated ESAT-6 responses compared to patients with classical M . tuberculosis lineage disease [5 , 6] . In addition , in liquid culture systems it has been reported that M . africanum has a slower growth rate with a larger bacillary size than M . tuberculosis [7 , 8] . While some studies have found that M . africanum is less virulent than M . tuberculosis , both in animal models and human patients [7 , 9–11] , others show that there is no difference [12] . Though these contradicting results may be due to differences in the study populations , they underscore how little is known about lineages 5 and 6 . Contributing to this lack of knowledge , while lineages 1–4 are widely distributed around the globe , lineages 5–7 are limited to certain regions of Africa [13] . Lineage 7 has only been found in Ethiopia [14] , and lineages 5 and 6 are found almost exclusively in patients living in West Africa , with very few cases occurring outside of this region , mostly involving recent immigrants from West Africa [1] . The reason for the apparent geographic restriction of lineages 5 and 6 is unknown . One hypothesis is the presence of an undiscovered animal reservoir endemic to West Africa , which is supported by the close relationship between lineages 5 and 6 and the animal-adapted lineages of the MTC [15 , 16] . Another hypothesis is that lineages 5 and 6 have a unique predilection for humans with genetic backgrounds common in West Africa . In fact , using a retrospective epidemiological study of the MTC in Ghana , Asante-Poku et al . showed that lineage 5 is associated with the Ewe ethnic group [17] . A third hypothesis is that lineages 5 and 6 are unable to compete with other lineages outside of West Africa , either due to loss of fitness or decreased transmissibility , thus explaining their limited global distribution [7] . Historically , mycobacterial subspecies were defined by biochemical assays , but , as genetic tools became more readily available , it is now possible to identify genomic regions that define MTC lineages [18] . The publication of the whole genome sequence of M . africanum GM041182 , a single lineage 6 strain , provided valuable insights into the genetics of this lineage [19] . For instance , the authors identified lineage 6-specific pseudogenes , a novel region not present in M . tuberculosis , and single nucleotide polymorphisms ( SNPs ) in key genes , all of which may play a role in the geographic restriction of lineage 6 . A later study sequenced four additional lineage 6 isolates and was able to confirm many of these findings , but also showed that not all mutations identified in M . africanum GM041182 are shared by other members of this lineage [8] . To our knowledge , no study has closely analyzed the genetics of lineage 5 . From these studies , it is clear that more sequenced isolates are needed to fully characterize the genetics of lineages 5 and 6 and to illuminate mechanisms that may explain its geographic isolation . Toward this end , we sequenced 92 clinical MTC isolates from Mali , a country in West Africa in which 26 . 2% and 1 . 6% of tuberculosis cases are caused by lineage 6 and lineage 5 , respectively [20] [1] . Using these and previously published data , we performed both alignment- and assembly-based comparative analyses to further refine our understanding of lineage-specific genomic features that might explain the geographic distribution of lineages 5 and 6 . To our knowledge , this is the largest collection of lineage 6 strains sequenced to date , and the first in depth whole genomic characterization of lineage 5 .
101 strains were selected from clinical isolates collected in Bamako , Mali [20] , and included all strains identified by spoligotyping as M . africanum , M . tuberculosis T1 , or M . bovis . Of these strains , 92 were still viable and were submitted for whole genome sequencing . These 92 strains will be referred to as the “Mali Collection” ( S1A Table ) . In addition , to improve MTC lineage representation , we selected additional whole genome assemblies that matched the quality of our assemblies . These included four finished M . bovis genomes available from GenBank ( M . bovis AF2122/97 [21] , M . bovis BCG Mexico [22] , M . bovis BCG Pasteur 1173P2 [23] , and M . bovis BCG Tokyo 172 [24] ) , a set of 40 M . tuberculosis strains ( 9 lineage 1 strains , 12 lineage 2 strains , 7 lineage 3 strains , and 12 lineage 4 strains ) from South Africa [25] , the finished M . africanum genome from Genbank ( M . africanum GM041182 ) [19] , and our outgroup , M . canettii CIPT 140010059 [26] . Combined with the Mali Collection , these 137 strains will be referred to as the “Assembly Collection” ( S1B Table ) . Finally , all 161 strains ( 122 lineage 2 , two lineage 3 , and 37 lineage 4 ) from a study in China were included in the variant analysis to improve geographical and lineage representation [27] . The samples from the China study ( S1C Table ) combined with the samples from South Africa and Mali ( for a total of 289 strains ) will be referred to as the “Alignment Collection” . The study protocols for the Mali samples were approved by the Ethics Committee of the University of Bamako and the Institutional Review Board of the National Institute of Allergy and Infectious Diseases , National Institutes of Health ( NIAID/NIH ) , Bethesda , MD , USA . For all samples , written informed consent was obtained from study participants prior to cohort enrollment [20] . For the South African samples , Biomedical Research Ethics Council ( BREC ) approval from the University of KwaZulu-Natal was granted for collection of sputum specimens from study participants and for whole genome sequencing of clinical strains . Written informed consent was obtained from study participants prior to cohort enrollment , or waived by BREC [25] . Drug resistance to isoniazid , rifampicin , ethambutol and streptomycin was tested for all Mali strains as previously described [20] . We confirmed those results by submitting 17 strains to National Jewish Health in Colorado for agar proportion testing of isoniazid , rifampicin , ethambutol , ofloxacin , niacin , kanamycin , ethionamide , capreomycin , amikacin , cycloserine and para-aminosalicylic acid , as well as radiometric testing of ciprofloxacin and pyrazinamide . The agar proportion results confirmed the mycobacterial growth indicator tube ( MGIT ) tests performed in Mali . Genotypic drug resistance was determined for rifampicin , isoniazid , ethambutol , streptomycin , ofloxacin , kanamycin and ethionamide using genetic markers from line-probe assays ( S2 Table ) . Extraction of genomic DNA was performed on 10 mL cultures grown in 7H9 broth using the CTAB-lysozyme method as previously described [28] . Library preparation and whole genome sequencing ( WGS ) were performed as previously described [29–31] . GenBank accessions for all strains used in this analysis can be found in S1B Table , along with assembly statistics for the new sequences generated at the Broad Institute ( 92 sequences from Mali generated for this study , and 40 sequences from South Africa ) . All genomes in our Assembly Collection were uniformly annotated by transferring annotations from M . tuberculosis H37Rv . The reference M . tuberculosis H37Rv genome ( accession CP003248 . 2 ) was aligned to draft assemblies using Nucmer [32] . This alignment was used to map reference genes over to the target genomes . Using this methodology , annotations were successfully transferred onto all 137 strains for 3466 of the M . tuberculosis H37Rv genes; the rest of the M . tuberculosis H37Rv genes transferred to a subset of the genomes . For those genes not cleanly mapping to M . tuberculosis H37Rv , the protein-coding genes were predicted with the software tool Prodigal [33] . tRNAs were identified by tRNAscan-SE [34] and rRNA genes were predicted using RNAmmer [35] . Gene product names were assigned based on top blast hits against the SwissProt protein database ( > = 70% identity and > = 70% query coverage ) , and protein family profile search against the TIGRfam hmmer equivalogs . Additional annotation analyses performed include Pfam [36] , TIGRfam [37] , Kyoto Encyclopedia of Genes and Genomes ( KEGG ) [38] , clusters of orthologous groups ( COG ) [39] , Gene Ontology ( GO ) [40] , enzyme commission ( EC ) [41] , SignalP [42] , and Transmembrane Helices; Hidden Markov Model ( TMHMM ) [43] . Reads from each isolate were aligned against the 43 spacer sequences traditionally used in wet lab spoligtyping [28 , 44] . From these alignments , the number of matching reads was used to determine if the spacer was present . The spacer was considered absent if the read count total was in the lowest quartile of counts . Spacers were defined as present by using a Bonferroni corrected p-value based on an exponential distribution of the average absent spacer counts . If the p-value was <0 . 01 the spacer was considered to be present . The spacer pattern was matched to the SITVITWEB database to generate a named spoligytpe for each isolate and to determine the spoligotype international type ( SIT ) [45] . SYNERGY2 [46–48] , available at http://sourceforge . net/projects/synergytwo/ , was used to identify cluster-based orthogroups across our Assembly Collection of 137 genomes , which we will refer to as “SYNERGY orthogroups” . In addition , for each M . tuberculosis H37Rv gene , we defined a second set of annotation transfer-based ortholog groups as the set of genes for which annotations were transferred from this M . tuberculosis H37Rv gene in our annotation protocol , which we will refer to as “M . tuberculosis H37Rv-based orthologs” . Genes without M . tuberculosis H37Rv orthologs were manually examined in the context of their SYNERGY orthogroups to identify lineage-specific novel genes . Phylogenetic trees were generated by applying RAxML [49] to a concatenated alignment of 3343 single-copy core SYNERGY orthogroups ( excluding orthogroups with paralogs ) across all 137 organisms . Bootstrapping was performed using RAxML’s rapid bootstrapping algorithm ( 1000 iterations ) . Calculations of ANI were done as previously described [50 , 51] using the SYNERGY orthogroups calculated from the Assembly Collection . PAUP [52] was used to reconstruct gain and loss of M . tuberculosis H37Rv-based orthologs at ancestral nodes of the Assembly Collection phylogenetic tree using parsimony . In order to analyze changes in gene content , we used a cost matrix with values of 10 for a gene gain , 5 for a gene loss , and 0 . 2 for an increase or decrease in copy number . We looked for orthologs found within all members of one clade , and absent in other clades . As a further filter , we also required that orthogroups be found in >80% of the clade of interest , and <20% of other strains . We performed this analysis for four key clades: lineage 5 , lineage 6 , the clade including M . bovis and lineage 6 , and the clade including lineages 5 , 6 and M . bovis . In addition , we selected the Pfam gene categories most expanded or reduced in each clade of interest . We determined significance using Fisher’s test ( Q<0 . 05 ) . For each of the clades described above , we compared the strains below this node versus all other strains in our analysis . For our Alignment Collection , reads were mapped onto a reference strain of M . tuberculosis H37Rv ( GenBank accession number CP003248 . 2 ) using BWA version 0 . 5 . 9 . 9 [53] . In cases where read coverage of the reference was greater than 200x , reads were down-sampled using Picard [54] prior to mapping . Variants , including both single nucleotide polymorphisms ( SNPs ) and multi-nucleotide polymorphisms , were identified using Pilon version 1 . 5 as described [29] and were used to construct phylogenetic trees using FastTree [55] . We defined lineage-specific variants for lineage X , as those occurring in at least 95% of the strains of lineage X ( true positive rate >95% ) , missing in less than 5% of strains of lineage X ( positive predictive value >95% ) , occurring in less than 5% of the strains that do not belong to lineage X ( true negative rate >95% ) and not occurring in at least 95% of strains not belonging to lineage X ( negative predictive value >95% ) . The absolute number of true positives must exceed seven . Formulas are schematically presented in S3 Table . Mutations were considered M . africanum-specific ( lineage 5 and 6-specific , identified as LIN-Maf in S4 and S5 Tables ) if they met these cutoffs for lineage 5 and 6 combined , and were present in both lineage 5 strains . Similarly , mutations were considered M . tuberculosis-specific if they met these cutoffs for lineages 1–4 combined . No M . tuberculosis-specific mutations were identified . Due to inclusion of only two lineage 5 strains in our dataset , no lineage-specific variants were identified in lineage 5 . Thus , for this lineage only , we used a less stringent requirement to define lineage-specific variants: we required that variants be present in both lineage 5 strains and in <5% of the strains in each other lineage . We classified each gene containing a lineage-specific variant into functional group categories , including GO [40] , KEGG [38] , Pfam [36] , and COG [39] . We then evaluated enrichment using Fisher's Exact test and corrected for multiple comparisons using the Storey method for functional group categories [56] . A pseudogene was defined as any gene that had a loss of function mutation anywhere within the coding sequence . Loss of function mutations were defined as nonsense mutations , or insertions or deletions with lengths that were not multiples of 3 base pairs or were greater than 30 base pairs . Lineage-specific pseudogenes were determined using the same definitions as for variants on a per gene basis ( positive predictive value > 95% , negative predictive value >95% , true positive rate >95% , true negative rate >95% , number of true positives >7 , with the exception of lineage 5 , which used the SNP cutoffs of pseudogene in both lineage 5 strains and in <5% in each other lineage ) . The effect of select non-synonymous mutations on protein function was assessed using the online version of SIFT at default settings [57] , unless there was low confidence in the prediction , in which case SIFT was run for each of the four available databases ( UniRef90 from April 2011 [default] , UniProt-SwissProt 57 . 15 from April 2011 , UniProt-TrEMBL from March 2009 and NCBI nonredundant from March 2011 ) . Peptide binding was predicted using the NetMHCII online tool with default settings [58] .
Our collection of 92 clinical MTC strains was isolated from patients presenting with pulmonary tuberculosis at Point G , Bamako , Mali between 2006 and 2010 as part of a cross-sectional study to analyze the diversity of the MTC in Mali [20] . All patients were Mali natives , with the exception of one patient born in central Africa ( S1A Table ) . We sequenced this collection using the Illumina sequencing platform , and the resulting reads were both assembled into contigs and aligned against the M . tuberculosis H37Rv reference genome . Based on our phylogenetic reconstructions , our collection included one lineage 1 , two lineage 2 , zero lineage 3 , 63 lineage 4 , two lineage 5 and twenty-four lineage 6 strains ( Fig 1 ) . The spoligotype distribution of our collection is representative of what has previously been observed in West Africa , except that we had a higher proportion of SIT53 ( T1 ) strains and a lower proportion of SIT181 ( AFRI_1 ) strains ( Figs 1 and S1 ) [45] . In order to perform statistical comparisons of the M . tuberculosis , M . africanum and M . bovis lineages , our newly sequenced dataset ( the “Mali Collection” ) was combined with data from additional strains from GenBank and South Africa ( “Assembly Collection” , Fig 2 ) , as well as data from China ( “Alignment Collection”; see Materials and Methods and S1 Table ) . These additional comparator genomes enabled us to examine in detail the distinguishing characteristics of lineages 5 and 6 that might explain their geographic restriction . Since this represents the largest collection of whole genome sequences of lineage 5 and 6 strains to date , we used our Assembly Collection to conduct a detailed examination of their phylogeny and characteristics in relation to other members of the MTC , including M . bovis and M . tuberculosis . M . bovis is considered an animal strain that mainly infects cattle and rarely humans , while M . tuberculosis is human adapted , and lineages 5 and 6 are thought to be intermediate between the two [1 , 15] . Using our Assembly Collection , we constructed a high-resolution phylogenetic tree using 3 , 343 single-copy core orthogroups ( sets of orthologs ) conserved across all 137 strains ( Materials and Methods ) . This tree was rooted using the outgroup M . canettii and agreed with phylogenies observed in other studies , including the fact that each of the lineages was clearly separated from the other , with lineage 5 being more closely related to human-adapted strains and lineage 6 being more closely related to M . bovis ( Fig 2 ) [13 , 15] . It has been previously shown , using average nucleotide identity ( ANI ) analysis , that separate bacterial species share <65–90% of genes and have no more than 94–95% ANI among shared genes [50 , 51] . Using gene content and nucleotide variation among shared genes , we examined the genetic distances between strains within the Assembly Collection to understand how mycobacterial species fit within this framework . In agreement with previous studies showing the close relationship between MTC subspecies , including M . africanum , we observed that there was little diversity between the lineages analyzed [61] . Strikingly , values from inter-lineage comparisons of M . tuberculosis , M . bovis , and M . africanum strains overlapped those from intra-lineage comparisons , showing very little separation , with >99% ANI and >94% fraction of shared genes ( Fig 3 ) . These results are in agreement with previous observations that these different organisms should not , in fact , be named different species [61] . In contrast , MTC pairwise comparisons with M . canettii revealed a clear separation between the two groups , suggesting that they occupy distinct niches ( Fig 3 ) . M . canettii is a smooth tubercle bacilli that causes human tuberculosis in East Africa and is considered an emerging pathogen in some parts of the world , but its natural host ( s ) and reservoirs remain unknown [62] . Thus , it might be argued , based on these data and the traditional cutoffs set by ANI analysis , that all MTC members should be named the same species , and that even M . canettii should be included since pairwise identities with MTC exceeded these thresholds ( Fig 3 ) . However , as Smith et al . have previously discussed [61] changes in nomenclature can cause confusion in the literature , and so we will continue to refer to M . africanum-associated lineages as either lineage 5 or 6 within the MTC . Despite the fact that lineages 5 and 6 are so closely related to lineages 1–4 , as demonstrated by Fig 3 , they are still unique in being geographically restricted compared to these other lineages . One hypothesis for this restriction is that they are less fit , unable to compete with other lineages within the MTC . To examine this possibility , we looked within the Mali Collection for clues that lineage 6 strains were not as diverse as strains from lineage 4 , the other predominant lineage within the region . We analyzed the breadth of pairwise diversity within lineage 6 using the ANI output and compared this diversity to that of lineage 4 strains isolated within Mali . ANI diversity was not statistically different when comparing these two groups of strains ( Fig 4 ) . Although this result does not eliminate the possibility of differing ecologies , such as an animal reservoir for lineage 6 , as has previously been hypothesized [16] , it does suggest that lineage 6 has not undergone a recent selective sweep or population bottleneck that would make lineage 6 populations circulating within Mali less diverse than lineage 4 populations [63] . In addition to being diverse , we also observed highly similar lineage 6 strains among this collection . Three pairs of lineage 6 strains were separated by less than 10 SNPs relative to M . tuberculosis H37Rv ( Fig 1B; see Materials and Methods ) , including isolates from both HIV-positive and immunocompetent patients . There were also six such clusters within lineage 4 . A cutoff of 12 SNPs has previously been used to determine recent transmission [64] . Thus , strains separated by less than 10 SNPs provide evidence of transmission , suggesting that 6 of 24 ( 25% ) of our lineage 6 strains and 13 of 63 ( 21% ) of our lineage 4 strains were involved in recent transmission events , confirming previous observations based on alternative genotyping approaches that there is robust ongoing transmission of lineage 6 within this region [9] . Given the reports of lineages 5 and 6 strains having decreased virulence [7 , 9–11] , we hypothesized that altered virulence may contribute to geographical restriction , either due to changes in host requirements or to a reduction in fitness . To test this hypothesis , we examined lineage-specific pseudogenes ( truncated genes ) and non-synonymous SNPs in known essential genes , slow growth genes , and genes required for virulence in mice and growth in macrophages to determine whether lineages 5 and 6 had an enrichment of defects in these genes that might contribute to overall altered virulence [65–67] . Although both lineages 5 and 6 had lineage-specific mutations in these gene categories , so did other lineages ( S4A and S5 Tables ) , and the proportion of mutated genes in lineage 6 was not significantly different from that of the other MTC lineages [8] ( Fig 5 ) . Lineage 4 was not included on this graph because it only had one lineage-specific mutation in an intergenic region when aligned to M . tuberculosis H37Rv , which is a member of lineage 4 , and lineage 5 was excluded due to low sample size . We performed a similar analysis on the full length of genes encoding known T cell antigens as defined by Comas et al . [4] to explore whether alterations in these genes might be restricting host specificity , but again we observed no significant difference in the proportion of lineage 6-specific mutations that fell within these genes as compared to lineages 1 , 2 and 3 ( Fig 5 ) . Similarly , we looked for enrichment of lineage-specific mutations in COG , GO , KEGG , Pfam and TIGRfam gene categories , but found no enrichment in any of these categories , either for pseudogenes or non-synonymous SNPs ( Q > 0 . 05 ) . These results corroborate our observations from ANI that the lineages of the MTC are very similar in their overall genetic composition and suggest that lineage 6 is not enriched for mutations in virulence genes relative to other lineages . However , while the overall number of mutations in virulence genes was not enriched , we identified mutations in these genes that might have an impact on virulence that will be discussed below . Studies have shown that lineages 5 and 6 evolve drug resistance less often compared to other MTC lineages , including the study from which these sequenced strains were obtained [20 , 68] . Thus , one hypothesis for the limited geographic range of lineages 5 and 6 could be decreased fitness relative to strains better able to evolve antibiotic resistance . In this case , we might expect that mutations driving drug resistance in these two lineages would be different from those evolving in more successful lineages . Thus , we analyzed our newly sequenced strains from Mali for the presence of mutations known to confer drug resistance and used in common nucleic acid-based commercial tests [59 , 60] for the detection of drug resistance [69–75] ( S2A Table ) . Forty ( 60% ) strains in lineages 1–4 and only four ( 15% ) of the lineage 5 and 6 strains were phenotypically resistant to at least one of the four tested drugs . We observed that mutations used in commercial tests were sensitive in detecting phenotypic resistance to rifampicin , isoniazid and ethambutol ( S2B Table and Figs 1 and S1 ) . Streptomycin resistance mutations were not included among our list of known resistance mutations ( S2A Table ) . Therefore , we searched for potential resistance mutations in a set of genes previously known to affect streptomycin resistance , including rrs , rpsL , and gidB [76–78] . We identified a point mutation in gidB that caused a non-synonymous change ( leucine to serine at residue 79 ) that is predicted to affect protein function [57] ( S1D Fig; see S1 Text for more details ) . This mutation was found in 23 streptomycin resistant strains and no streptomycin susceptible strains in our dataset and likely represents a previously uncharacterized mutation that confers resistance to this drug . Previous studies have identified loss of function mutations in gidB affecting streptomycin resistance [77] , as well as point mutations in the region of gidB close to residue 79 , including at residues 75 and 82 [78] . In addition , we identified known mutations in genes associated with resistance to drugs that were not phenotypically assessed , including ofloxacin , kanamycin , and ethionamide . Using the list of mutations in S2A Table , we found that 25 ( 38% ) of the Mali strains belonging to lineages 1 , 2 or 4 could be classified as MDR ( multi-drug resistant; resistant to isoniazid and rifampicin ) , and two ( 3% ) could be classified as pre-XDR ( pre-extensively drug resistant; resistant to isoniazid , rifampicin , plus either ofloxacin or kanamycin ) . In contrast , three ( 11% ) of the lineage 5 and 6 strains could be classified as MDR , and one ( 4% ) could be classified as pre-XDR . The presence of these pre-XDR strains is of particular concern , as XDR has not been reported in Mali , and testing is not currently performed routinely for second line antibiotics [79 , 80] . Similar resistance-conferring mutations were found among the lineages ( S2 Fig ) . Although we cannot eliminate the possibility of cross resistance and other alternate genetic mechanisms of lineage 5 and 6 drug resistance , or of differences in drug tolerance or rates of persister cells , it appears that the mechanism of genetic drug resistance was similar between lineages 2 , 4 , 5 and 6 . Thus , although the sample size was small , our results suggest that drug resistance , while less frequent in lineage 6 , evolves through acquisition of similar mutations as observed in lineages 2 and 4 in Mali , including combinations of mutations leading to pre-XDR , and that this resistance could be detected using current molecular diagnostic approaches . Previous analyses pinpointing lineage 6-specific genomic features have compared limited numbers of strains , which might have caused these studies to miss important features or to identify features that are not actually found in a broader set of lineage 6 strains [8 , 19] . Also , these studies have not examined genomes of lineage 5 in detail . Using both our Alignment and Assembly Collections , containing representatives from lineages 1 through 6 and M . bovis , we sought to robustly identify distinguishing features of lineages 5 and 6 , focusing on traits that could have caused geographic restriction . Using our Assembly Collection , at each node labeled A-D in Fig 2 ( representing genetic diversification events that may correlate with ecological specialization ) , we identified gene gains and losses ( Table 1; Materials and Methods ) . Many of our findings agreed with previous observations describing regions of difference , as determined through genomic hybridizations [3 , 18] . However , we also identified a small number of genes that were not previously identified as being part of these known regions of difference ( see S1 Text for full details ) , including a gain of genes encoding a PE-PGRS and hypothetical protein at the last common ancestor of lineage 6 and M . bovis , the loss of Rv1523 ( a methyltransferase ) and Rv3514 ( PE-PGRS57 ) in lineage 5 , and the loss of a gene encoding a TetR family regulator and the gain of one PPE protein-encoding gene at the last common ancestor of lineages 5 , 6 and M . bovis . In addition , using our alignments to M . tuberculosis H37Rv , we identified a number of lineage-specific mutations , including pseudogenes that affect protein function ( Tables 2 , S4 , S5 and S6 ) . From these data , we identified 681 lineage 6-specific mutations shared across all lineage 6 strains , including eight truncated pseudogenes . These data also provided the first in-depth analysis of lineage 5 assemblies , which revealed 952 lineage-specific mutations and 43 pseudogenes as shared by our two lineage 5 strains ( see S1 Text ) . The larger number for lineage 5 compared to other lineages likely results from our small sample size . Key categories of lineage-specific mutations and pseudogenes that might contribute to the geographic restriction of lineages 5 and 6 are discussed below , and in more detail in the S1 Text . One distinguishing clinical characteristic of lineage 6 is an attenuated T cell response to ESAT-6 , one of the proteins secreted through the ESX secretion system , as compared to patients infected with lineages 1–4 [5] . This altered immune response supports the hypothesis that lineage 5 and 6 have specificity for a particular host immunogenic background . Although our data cannot address whether ESAT-6 production has been affected , we observed non-synonymous polymorphisms , including indels , in genes encoding ESX secretion systems that could contribute to the different immune responses of lineage 6-infected patients ( Table 3 ) . Furthermore , we observed lineage-specific mutations in ESX-encoding genes in all lineages , suggesting that each lineage may have unique interactions with the host ( Table 3; S1 Text ) . Lineages 5 and 6 had lineage-specific mutations , including pseudogenes , in genes encoding multiple components of cofactor biosynthetic pathways , including molybdenum , vitamin B12 , and vitamin B3 ( S1 Text and Tables 3 , S4 and S5 ) . Molybdenum cofactors are key catalysts for redox reactions , and are hypothesized to have played an important role in the evolution of pathogenic mycobacteria [81] . In addition , mycobacteria are one of the few bacterial pathogens with the ability to synthesize vitamin B12 [82] . Thus , both of these cofactors have specifically evolved in mycobacteria and loss of these cofactor biosynthetic pathways could affect the function of proteins that use these cofactors , which include proteins that are important for many cellular functions . These mutations may affect the host range of lineages 5 and 6 , supporting the hypothesis of a unique host preference . It has been shown previously that M . africanum GM041182 has a distinct physiology as compared to that of M . tuberculosis H37Rv , including a larger cell size and slower growth rate [7] . Possibly explaining these differences , we identified lineage 6-specific non-synonymous SNPs in genes encoding the L , D transpeptidases , ldtA and ldtB ( Rv0166c and Rv2518c ) , previously shown to form cross-linkages within peptidoglycan ( Tables 3 and S4A ) [83] and to be key drivers of cell shape , size , surface morphology , growth and virulence [84] . Lineage 5 also contained a non-synonymous SNP predicted to affect LdtA protein function ( Tables 3 and S4A ) . No other lineages had a lineage-specific mutation in an L , D-transpeptidase . We observed that lineages 5 and 6 had lineage-specific mutations in genes encoding adenylate cyclases , the enzymes that synthesize cyclic AMP ( cAMP ) , an important cell signaling molecule . Although the affected genes were different between the two lineages , no other lineage had lineage-specific mutations predicted to affect adenylate cyclase function . Deletion of one of the 17 adenylate cyclases in M . tuberculosis , Rv0386 , has been shown to reduce virulence and alter the immune response [85] . Bentley et al . also previously found that this gene was a pseudogene in M . africanum GM041182 , although here we find that pseudogenization of Rv0386 was not lineage specific ( S6A Table ) . Nevertheless , given the number of affected adenylate cyclases , there may be differences in cAMP signaling within lineages 5 and 6 , leading to altered pathogenicity . In order to shed light on the reported lower rates of drug resistance in lineages 5 and 6 , we screened our lineage-specific mutations to investigate if there were any changes in known drug resistance genes that were not on the list of mutations used before and that might affect the development of antibiotic resistance [20 , 68] . In lineage 6 , we observed two lineage-specific non-synonymous mutations in rpoB , and one lineage-specific non-synonymous mutation in embC ( S1A and S1C Fig and Tables 3 and S4A ) not previously implicated in antibiotic resistance . Lineage 5 strains had non-synonymous mutations in genes encoding AtpH ( Rv1307 ) and AtpG ( Rv1309 ) , both of which are subunits of ATP synthase [86] ( S4A Table ) , and a target of bedaquiline , a new antibiotic reserved for the treatment of drug resistant tuberculosis [87] . Both of these mutations were predicted to affect protein function by SIFT [57] , and may affect bedaquiline efficacy in countries with a high proportion of patients infected with lineage 5 . Thus , both of these lineages have non-resistance conferring mutations in genes associated with drug resistance that might influence the frequency at which drug resistance develops in these lineages . M . tuberculosis H37Rv contains four mammalian cell entry ( MCE ) operons , which play an important role in mycobacterial virulence [88] . In addition to confirming earlier reports that lineage 6 strains lacked one of these four operons ( operon 3; Table 1 ) [18] , we observed lineage-specific mutations in several of the other MCE operons ( lineage 6 had mutations in operons 1 and 2; lineage 5 had mutations in operons 1 and 3 ) . We also observed a non-synonymous mutation in mce1B that was shared by lineages 5 and 6 strains and was predicted by SIFT to affect Mce1B protein function [57] . In comparison , the other lineages had nearly identical MCE operons as compared to M . tuberculosis H37Rv ( Tables 3 and S4A ) .
Our study describes the largest collection of sequenced lineage 6 isolates to date , and , to our knowledge , the first in-depth analysis of the genetics of lineage 5 . Through our work , we have characterized the genetic basis of antibiotic resistance in lineage 6 strains from Mali , shown that M . africanum and M . tuberculosis are part of the same species , and better defined the mutations and changes in gene content that typify these lineages . Collectively , this work provides insights into these understudied lineages and provides testable hypotheses as to why they are geographically restricted . We evaluated 92 Mali MTC isolates using both assembly and alignment-based approaches . Our assemblies revealed several new regions of difference and our alignments identified smaller lineage-specific changes . In addition to our conclusion that M . africanum is not a separate species , we observed that some M . africanum-M . tuberculosis pairs of strains have greater average nucleotide identity than some pairs of strains from within the same lineage . Furthermore , our ANI data demonstrated that there is comparable diversity in lineages 4 and 6 , suggesting that lineage 6 has not undergone a recent population bottleneck . This emphasizes the extremely close relationship between all MTC lineages , highlighting the role that small changes within the MTC have played in geographical restriction and altering host preferences . Since our assemblies were of very high quality , we were able to observe changes in genes that previous studies could not , thus providing a prioritized list of genes for investigating lineage 5 and 6 characteristics . One hypothesis for the geographical restriction of lineages 5 and 6 is the presence of an unknown non-human reservoir . M . africanum has been found in animals , including monkeys , cows , pigs and hyrax [89–94] . Unfortunately , given genomic data from human clinical isolates alone , we cannot address this hypothesis directly . However , given the similar level of diversity between lineage 4 and 6 in Mali and the evidence of person-to-person transmission , a non-human reservoir seems unlikely to explain the geographic restriction , as lineage 6 appears well adapted to spread in humans living in this geographic setting , unlike M . bovis in this and other settings [95–97] . Another hypothesis for why lineages 5 and 6 occur almost exclusively in West Africa is a preference for hosts of West African ethnicity , supported by previous evidence , including a study linking lineage 5 to the Ewe ethnic group [17] . We identified lineage-specific mutations in ESX genes in every lineage , indicating that each lineage may interact uniquely with the host immune system . Mycobacteria have five ESX secretion systems , also known as type VII secretion systems , which secrete small proteins across the bacterial cell envelope and are important to mycobacterial virulence [98 , 99] . For example , ESX-1 secretion is lost as part of RD1 in M . bovis BCG vaccine strains , resulting in loss of ESAT-6 and CFP-10 secretion , and thus attenuation of the bacterium [100 , 101] . The lineage-specific mutations in ESX genes could lead to alterations in the pathogen-host immune interaction , resulting in a requirement in lineages 5 and 6 for the West African immune system . In fact , an altered response to ESAT-6 in patients infected by lineage 6 has previously been reported [5] . Thus , the specific ESX mutations in lineages 5 and 6 could represent adaptations to the niche of the West African host . Lineage-specific mutations in cobalamin biosynthesis could also contribute to adaptation of these lineages to the specific ecological niche of the West African host . The hypothesis of adaptation to a different host cofactor environment for lineages 5 and 6 is supported by several studies that have found increased levels of vitamin B12 plasma concentrations in West Africans compared to Europeans and Mexicans [102 , 103] . One unique characteristic of mycobacteria compared to many other bacterial genera is that they are capable of synthesizing vitamin B12 . Furthermore , vitamin B12 may play a crucial role in M . tuberculosis infection [82 , 104 , 105] . Thus , the lineage-specific mutations in cobalamin pathways in lineages 5 and 6 may alter these strains’ ability to synthesize vitamin B12 , which may be tolerated in West African hosts with higher levels of plasma B12 . Adaptation to this B12-rich West African niche might prevent these lineages from infecting other ethnic groups with lower B12 bioavailability; however , further studies would be required to confirm this hypothesis . A third hypothesis for the geographic restriction of lineages 5 and 6 is that they are less fit , either for transmission or in-host virulence , resulting in a decreased ability to survive outside of West Africa . Several papers have shown no difference in transmission rates between M . tuberculosis-associated strains and M . africanum-associated strains [5 , 9 , 106 , 107] . In agreement with previous findings , our Mali Collection revealed three pairs of lineage 6 strains separated by 10 or fewer SNPs when aligned to M . tuberculosis H37Rv , suggesting recent transmission of strains between patients within the ethnic backgrounds prevalent in Mali [64] . That these transmission events were not exclusive to HIV positive patients suggests that a compromised immune system is not required for a transmission event . These results indicate that lineages 5 and 6 do not have a reduced ability to transmit . A decrease in fitness could also be reflected in a decrease in virulence . It has been hypothesized that M . africanum is less virulent within humans , mice and guinea pigs than is M . tuberculosis [7 , 9–11] . However , lineage 6 was not enriched for mutations in virulence and growth-related genes compared to lineages 1 , 2 and 3 , suggesting that lineage 6 does not contain an overall numerical loss of virulence or growth-associated genes . Despite this , individual mutations can still greatly affect disease outcome , and analysis of lineage-specific mutations identified several potential mechanisms that could lead to changes in how lineages 5 and 6 proliferate and cause disease . The lineage-specific mutations discussed above that could relate to a niche adaptation in hosts of West African ethnicity , including the lineage-specific mutations in ESX genes and cofactor biosynthesis genes , are also involved in virulence . Another key set of virulence genes with lineage 5 and 6-specific mutations are the MCE operons . The MCE operons play an important role in the virulence of M . tuberculosis , particularly in mycobacterial growth in macrophages [67] . Antibodies to MCE1 proteins have been identified in patients [108] , and operons 1–3 are required for virulence in mice [88] . Despite this apparent role in virulence , lineage 6 contains mutations that affect protein function in operons 1–3 , while lineages 1–3 have nearly identical MCE operons to M . tuberculosis H37Rv , suggesting one potential mechanism of decreased virulence . Another set of virulence-related genes with lineage-specific mutations are adenylate cyclases , which synthesize cAMP , an important second messenger [109] . M . tuberculosis encodes 17 adenylate cyclases , and deletion of one of them ( Rv0386 ) has been shown to affect virulence and host response [85] , highlighting the importance of this set of genes to pathogenicity . Both lineages 5 and 6 contained lineage-specific mutations predicted to affect the protein function of several adenylate cyclases , suggesting altered cAMP signaling in these strains , and a potential effect on the virulence of lineages 5 and 6 . Another pathway that affects bacterial growth and host response is the synthesis of the cell wall . Both lineage 5 and 6 contained lineage-specific mutations in L , D-transpeptidases . L , D-transpeptidases are critical to the structure of mycobacterial peptidoglycan and are involved in bacterial structure and growth [84] , providing a possible explanation for the reported changes in cell size and doubling time in M . africanum GM041182 compared to M . tuberculosis H37Rv [7] . An altered cell wall could support either the hypothesis of decreased virulence , or suggest the need for a specific host immune system . In addition , we saw high variability in PE , PPE and PE-PGRS genes , including changes in gene content . These repetitive regions are difficult to sequence and are often ignored , but may play a crucial role in antigenicity and the host-pathogen interaction [110 , 111] . Using our high quality assemblies and alignments , we were able to identify lineage-specific mutations in these genes , as well as altered gene content . These mutations highlight the possibility of a critical role for these proteins in host-pathogen interactions and emphasize the need for a more detailed analysis of these regions . Furthermore , there were also a number of mutated hypothetical proteins and proteins of unknown function , all of which may play a critical as yet undiscovered role . In addition to exploring mechanisms of geographic restriction , we also identified mutations that may have clinical implications for the region . We found that in Mali , M . africanum-associated and M . tuberculosis-associated strains evolved antibiotic resistance through similar mutations , and thus standard line-probe assays can still be utilized in West Africa . However , we also found a gidB polymorphism not previously described which might account for much of the streptomycin resistance in Mali . Also of concern , we identified several cases of pre-XDR in Mali , suggesting that Mali may need to begin testing for XDR cases . Furthermore , we identified lineage 5 or 6-specific mutations that may affect the evolution of drug resistance , particularly bedaquiline . Thus , whole genome sequencing surveys like this one are useful in revealing new mechanisms for drug resistance , informing development of molecular diagnostics . One weakness of our study was that we were limited in our sample size for lineage 5 and M . bovis strains . Our collection was not representative of M . bovis genomic diversity , as three of the four M . bovis strains in our analysis were attenuated M . bovis BCG vaccine strains . However , we only used the M . bovis strains in our ANI and gene content analysis , and required that any observations be consistent with wild-type M . bovis sequence , AF2122/97 , and our results corroborated all previous findings of M . bovis regions of difference . Another weakness of our study was that our observations may be specific to Mali , since all lineage 5 and 6 isolates sequenced for our study were isolated in Mali , although these lineages are found throughout West Africa . However , our lineage 6 isolates were genetically diverse , and represented multiple spoligotypes , and our isolates from other lineages did not cluster separately on the phylogenetic tree from strains isolated from South African patients . Thus , our collection reflected substantial diversity and did not originate from a clonal outbreak . In fact , the study from which we selected our samples found a wide diversity of strains in Mali , which covered 55% of all known spoligotyped strains [20] . Furthermore , based on spoligotyping , many similar strains can be found in neighboring countries [54 , 68 , 112–115] . And , finally , studies that employ genomic data alone are insufficient to address causality . However , we believe that this in-depth genomics analysis of the neglected pathogen , “M . africanum” , provides a strong foundation from which causal relationships between lineage-specific variation and geographic restriction can be made . This collection provides valuable insights into the distinguishing genomic features of M . africanum . Here , we have analyzed in detail the genomes of lineage 5 and 6 isolates from Mali and identified several potential genetic reasons for the geographical restriction of lineages 5 and 6 , such as alterations in vitamin B12 pathways and genes associated with virulence , which provide a guide to future studies focusing on the effects of specific genes . Although we cannot specifically point to a single reason why these lineages are geographically restricted , we have found mutations that support the hypothesis of a unique requirement for a host of West African ethnicity and for the hypothesis of loss of bacterial fitness . These hypotheses are not mutually exclusive , and we anticipate that these observations will be able to inform and fast-track experiments on mycobacterial pathogenicity and virulence , particularly with regard to this unique member of the MTC .
|
Mycobacterium africanum consists of two lineages , lineages 5 and 6 , within the Mycobacterium tuberculosis complex ( MTC ) that cause human tuberculosis in West Africa , but are found rarely outside of this region . Our analysis of the whole genome sequences of 26 lineage 5 and 6 isolates , and 66 isolates from other lineages within the MTC , reveal that M . africanum does not meet modern criteria to be considered an independent species . We analyzed drug resistance-associated genes and found that drug resistance evolves within these lineages through similar mechanisms as observed for the rest of the MTC in Mali . Though we did not see an elevated number of mutations in virulence-associated genes in these two lineages , we identified a number of lineage-specific mutations , pseudogenes and changes in gene content that may impact virulence and host specificity , and improve , overall , our understanding of what make these lineages unique .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2016
|
Whole Genome Sequencing of Mycobacterium africanum Strains from Mali Provides Insights into the Mechanisms of Geographic Restriction
|
Q fever is a widespread zoonosis that is caused by Coxiella burnetii ( C . burnetii ) , and ruminants are identified as the main sources of human infections . Some human cases have been described , but very limited information was available about Q fever in ruminants on Reunion Island , a tropical island in the Indian Ocean . A cross-sectional study was undertaken from March 2011 to August 2012 to assess the Q fever prevalence and to identify the major risk factors of C . burnetii infection in ruminants . A total of 516 ruminants ( 245 cattle , 137 sheep and 134 goats ) belonging to 71 farms and localized in different ecosystems of the island were randomly selected . Samples of blood , vaginal mucus and milk were concomitantly collected from females , and a questionnaire was submitted to the farmers . Ticks from positively detected farms were also collected . The overall seropositivity was 11 . 8% in cattle , 1 . 4% in sheep and 13 . 4% in goats . C . burnetii DNA was detected by PCR in 0 . 81% , 4 . 4% and 20 . 1% in cow , sheep and goat vaginal swabs , respectively . C . burnetii shedding in milk was observed in 1% of cows , 0% in sheep and 4 . 7% in goats . None of the ticks were detected to be positive for C . burnetii . C . burnetii infection increased when the farm was exposed to prevailing winds and when there were no specific precautions for a visitor before entering the farm , and they decreased when a proper quarantine was set up for any introduction of a new ruminant and when the animals returned to the farm at night . MLVA genotyping confirmed the role of these risk factors in infection .
Q fever is a widespread zoonosis that is caused by Coxiella burnetii ( C . burnetii ) , an obligate intracellular bacterium [1]–[4] . The reservoir includes mammals , birds and arthropods , mainly ticks [5] . Ruminants ( sheep , goats and cattle ) are identified as the main sources of human infections [6] , [7] . Humans are infected mainly by inhalation of an aerosol contaminated with parturient products from the urines or feces of infected animals [8] . The risk of transmission of C . burnetii is dependent on the prevalence of shedder ruminants and on the level of shedding . C . burnetii is shed by ruminants mainly by birth products , but it may be shed via the vaginal mucus , milk , feces , urine and semen [9] . To control the spread of C . burnetii among animals as well as from animals to humans , the detection of shedders of C . burnetii and the knowledge of the prevalence of the infection are imperative . The risk of zoonosis also depends on the level of C . burnetii in the products of the infected animals . Serological tests ( complement fixation , indirect immunofluorescence and enzyme-linked immunosorbent assays ( ELISA ) ) are classically used in epidemiological studies to detect carriers of antibodies against C . burnetii . Serological tests indicate previous exposure [10] to C . burnetii and are not appropriate for the identification of shedder ruminants , especially because seronegatives are present among them [11] , [12] . This lack of sensibility in this technique is lower using ELISA [13] . Isolation of C . burnetii is not performed for epidemiological investigation because it is difficult , time consuming and requires confined level L3 laboratories . Conventional polymerase chain reaction presents a very useful method for the detection of C . burnetii DNA [9] , [14] . The real-time PCR assays are now recognized as the most convenient tools because these tests have excellent sensitivity , specificity and permits investigators to obtain quantifiable information . Real-time PCR is adapted to large scale studies because this technique can be semi-automated , thus reducing the risk of sample contamination and permitting gained time . Reunion Island is a French overseas department that has a population of approximately 800 , 000 inhabitants . Reunion Island is a hotspot in the Earth's crust located in the Indian Ocean , east of Madagascar , approximately 200 km south-west of Mauritius , the nearest island . The island is 63 km long and 45 km wide and covers an area of 2 , 512 km2 . Cities are concentrated on the surrounding coastal lowlands . The climate is tropical and humid , with two main seasons: a hot rainy season from December to March , and a dry and cold season from April to November . The eastern coast ( the “windward” coast ) experiences rainfall of approximately 2 , 000 mm per year , whereas the western coast ( the “leeward” coast ) has an annual rainfall of less than 2 , 000 mm . The domesticated animal populations on the island comprise approximately 40 , 000 cattle , 30 , 000 goats and 2 , 000 sheep . To date , no information was available about Q fever in humans and animals . The present study aimed to provide epidemiological information about Q fever in the animal population of Reunion Island using available diagnostic tools and appropriate samples . The data will be used to appreciate the prevalence of C . burnetii infection in the three main domestic ruminant species: cattle , sheep and goats at both the animal and herd levels , as well as to identify the major risk factors of infection .
The research protocol was implemented with the approval of the Direction of Agriculture , Food and Forestry ( DAAF ) from the French Ministry of Agriculture , under the European animal welfare regulation ( project license number 102498 ) . No endangered or protected species were involved in the survey . All the farmers gave their permission to be included in the study and for the samples . The animals were sampled without suffering . The animals were considered positive when at least one sample ( blood , swab or milk sample ) tested positive by either serology or PCR . The serological and PCR data were analyzed using a generalized linear mixed model ( glmmML library , R software ) , where the individual health status was the binomial response , and the variables from the questionnaire were the explicative factors . All of the explicative variables were categorical . The number of categories per variable was limited , such that frequencies of categories were only >10% . These variables were selected from a preliminary step aimed at lowering the chance of obtaining results affected by multicollinearity in the dataset [18] . All bilateral relationships between these variables were evaluated ( χ2 ) . A two-stage procedure was used to assess the relationship between explanatory variables and the health status of the animals . Logistic regression was used according to the method described by Hosmer and Lemeshow [19] . In the first stage , a univariate analysis was performed to relate Q fever positivity to each explanatory variable . Only factors associated ( Pearson χ2-test , P<0 . 25 ) with Q fever positivity were offered to a full model for multivariable analysis [20] . The second stage involved a logistic multiple-regression model . The contribution of each factor to the model was tested with a likelihood-ratio χ2 through a stepwise procedure ( backward and forward ) . At the same time , the simpler models were compared to the full model by the Akaike information criterion [21] . This process was continued automatically until a model was obtained with all factors significant at P<0 . 05 ( two-sided ) . Goodness-of-fit of the final model was assessed using Pearson χ2 , Deviance and the Hosmer–Lemeshow tests [19] .
The evaluation of the specificity of the real time PCR assays was reported by Klee et al . ( 2006 ) [22] . In the present study , all negative samples from INRA were negative , confirming the specificity of the assays . All of the samples that were found positive by INRA were confirmed positive . The detection threshold determined from dilution series of synthetic DNA showed that this PCR allowed the detection of 24 samples of 500 copies of the genome/mL and one sample of 250 C . burnetii particles/ml sample , the number of IS1111 elements in the genome being determined to be close to 20 for the Nine Mile strains [22] . Coefficients and Ct averages of intra- and inter-assays were 0 . 46% and 1 . 4% , 26 . 54 and 26 . 63 , respectively . Bacterial load in vaginal samples by ml of transport medium ranged from 50 , 600 to 255 , 000 for cattle , from 82 , 400 to 314 , 000 for sheep and from 112 , 000 to 385 , 000 for goats . The typability of the two loci was 85 . 1% ( 40 of 47 positive PCR for goats ) . We obtained nine genotypes among the 40 amplified DNA samples ( Table 1 ) . The overall seropositivity was 11 . 8% ( 95% CI 7 . 8 – 15 . 9 ) in cattle , 1 . 4% ( 95% CI 0 – 3 . 5 ) in sheep and 13 . 4% ( 95% CI 8 . 2–25 . 6 ) in goats . C . burnetii DNA was detected by PCR in 0 . 81% ( 95% CI 0–1 . 9 ) of cow vaginal swabs , 4 . 4% ( 95% CI 0 . 9 – 7 . 8 ) of ewe vaginal swabs and 20 . 1% ( 95% CI 13 . 3 – 26 . 9 ) of goat vaginal swabs . C . burnetii shedding in milk was observed in 1% ( 95% CI 0 . 2 –1 . 8 ) of cows , 0% in sheep and 4 . 7% ( 95% CI 0 – 11 . 2 ) in goats . Twenty-one out of 46 ( 95% CI 32 – 60 ) cattle farms were found to be positive either in serology or PCR , 50% ( 95% CI 33 – 67 ) of sheep farms and 41% ( 95% CI 18 – 64 ) of goat farms . All of these farms were spread throughout the island ( figure 1 ) . The within-herd prevalence in the positive farms ranged from 20% to 40% in cattle farms and from 30% to 90% in small ruminant farms . None of the ticks collected were detected to be positive for C . burnetii . After variable selection ( Table 2 ) , the logistic multiple-regression model indicated that the risk of C . burnetii infection was increased when the farm was exposed to prevailing winds ( OR = 2 , 11; 95% CI [1 , 13; 3 , 99] ) and when there were no specific precautions for a visitor before entering the farm ( OR = 3 , 13; 95%CI [1 , 57; 6 , 70] ) , and decreased when a proper quarantine was set up for any introduction of new ruminant ( OR = 0 , 06; 95%CI [0 , 01; 0 , 17] ) and when the animals went back to the farm at night ( 0 , 53; 95%CI [0 , 42; 0 , 64] ) ( Table 3 ) .
To the best of our knowledge , this is the only documented epidemiological study on Q fever in ruminants in Reunion Island , highlighting that Coxiella burnetii is endemic in this territory . For our epidemiological survey , we used both serological and PCR techniques to better understand the characteristics of Q fever . Complement fixation technique remains widely used by laboratories in many countries to assess the seroprevalence of C . burnetii infection . This method yields good results for routine diagnosis at the herd level , but multiple studies have concluded ( World Organisation for Animal Health 2010 ) that CFT is less sensitive than ELISA testing . Following international suggestions , ELISA results are deemed reliable for the screening of seroprevalence [23] , [24] . However , serological tests ( complement fixation or ELISA ) only detect antibody-carriers against C . burnetii , demonstrating the previous exposure to the pathogen but not the current shedding of the pathogen [10] . Because we aimed to assess the overall pattern and characteristics of Q fever in Reunion island , the detection of shedders of C . burnetii was important because they are one of the critical points for the control of spreading of the bacteria among animals and from animals to humans [3] . Polymerase chain reaction ( PCR ) has been used to detect C . burnetii DNA in biological samples . Additionally , we employed a real-time PCR technique that is currently being developed with the aim of providing quantifiable information . The technique allows a priori scaling in the importance of sources of bacterium with regards to the risk of transmission of C . burnetii among animals and from animals to humans . Finally , on the contrary to conventional PCR , real-time PCR can be automated , leading to both a lower risk of sample contamination and a more time-efficient method of detection [25] . Because we found a bacterial load 40 to 310 times higher than the detection threshold , the probability for false negatives remained low . Finally , we observed nine different MLVA genotypes with very good typability compared to that obtained from Roest et al . ( 53% ) [26] , possibly due to the number of cycles ( 60 ) . In our study , the overall seropositivity was 11 . 8% in cattle , 1 . 4% in sheep and 13 . 4% in goats . These results are much lower than those observed in Europe; for example , ELISA testing showed 38 . 0% in cattle and 6 . 0% in sheep for individual seropositivity in Hungary [27] . In Northern Spain , ELISA anti-C . burnetii antibody prevalence was slightly higher in sheep ( 11 . 8±2 . 0% ) than in goats ( 8 . 7±5 . 9% ) and beef cattle ( 6 . 7±2 . 0% ) [28] . Our seroprevalence rates were also lower compared to the results from other tropical countries . The seroprevalence in cattle was estimated to be between 40% and 59 . 8% in Nigeria , Sudan and Zimbabwe , and only 4% Chad [29] . The seroprevalence in sheep has been reported to vary between countries: 62 . 5% in Sudan , 22 . 5% in Egypt and 11% in Chad . Additionally , differences were also observed for seroprevalence in goats: 53% in Sudan , 16 . 3% in Egypt and 10% in Zimbabwe [29] . Our PCR results were quite surprising , with a low prevalence of C . burnetii in cow and ewe vaginal swabs ( 0 . 81% and 4 . 4% , respectively ) , but very high prevalence of 21% among goats . Generally , such high rates are observed after a Q fever-related termination of pregnancy as described by Cantas et al . ( 2011 ) [30] and Berri et al . ( 2005 ) [7] . Indeed , in our study , six small ruminant farms have indicated terminations among pregnant ruminants and our samples were collected within one month after these events . Shedding of C . burnetii in vaginal mucus lasts for one to five weeks [31] . In addition , it has been shown that most of the goats that had aborted or delivered normally in naturally infected herds shed the bacteria [32] , [33] . Our findings confirmed these previous results because the within-herd bacterial prevalence in the farms that reported pregnancy terminations was estimated to be between 70% and 90% . This study demonstrates that the risk of Q fever infection of ruminants increased when farms or grazing pastures are in the way of prevailing winds , confirming the airborne route of transmission for C . burnetii . In contrast to other studies [34] , ticks , which were all detected to be negative for C . burnetii , appeared to not be involved in the contamination process . However , the systematic use of deltamethrin may have reduced the tick population and altered their ability to carry C . burnetii . Infection of animals or humans and contamination of the environment with C . burnetii requires transport through the atmosphere . It is assumed that C . burnetii is absorbed or fixed at the aerosol surface and becomes airborne . C . burnetii is resistant to heat and dryness and can survive for more than 150 days in the environment . Most ruminants , especially sheep and goats , spend their days grazing outside in the production areas of the highlands or on the eastern coast , where the highest density of ruminants and farms is met and where the winds are blowing most of the year . Additionally , manure is often used as a fertilizer in market gardening in these areas , potentially contributing to the spread of C . burnetii [35] . It is notable that contaminated aerosols are a major mechanism whereby C . burnetii is transmitted to humans [36] . MLVA genotyping results were in agreement with this risk factor because genotype 5 was observed in five farms located in a 3 km radius of the same area as the eastern windy part of the island [37] . Even if no correlation between pastures exposed to prevailing winds and animals kept at night in their barn was observed , these two variables support the assumption that C . burnetii may be transmitted via airborne route . Indeed , our study also showed that the risk of infection for ruminants was lower when the animals were kept in the barn at night . Generally , older cows that stayed in the cow barn for longer periods of time than young animals are more frequently infected . Hence , the probability of being exposed to the bacterium increases with exposure time [38] . However , in Reunion Island , most of the barns , particularly for sheep and goats , are open spaces sheltered from the wind . In these cases , the probability of infection by droplets and aerosols transmitted by wind is lower . A lack of precautionary measures for visitors ( such as washing hands and changing clothes and boots ) before entering the farm was also associated with a higher risk for infection of ruminants with C . burnetii . The visitors , including veterinarians , food factory staff and professional hoof trimmers , may act as mechanical carriers and transfer the pathogen from infected to non-infected herds . This route of transmission has already been highlighted in previously reported articles [39] , [40] , suggesting that farm personnel often act as mechanical transmitters of contaminated fomites from an infected herd to uninfected ones . Conversely , the risk of infection for ruminants was decreased when a proper quarantine was set up before any introduction of new animals to the farm . New ruminants are introduced after purchasing or , in the case of goats , when a male is borrowed from another farm to improve the reproductive performances . We should mention that young goats are often used in religious celebrations on Reunion Island . Again , MLVA genotyping results stressed the risk of infection when no quarantine is set up because , in our study , genotype 7 was detected in only in the two farms that purchased live animals from farm A [40] . A recent study reported that purchase of animals increased the risk of introducing C . burnetii infection into cattle herds [23] . This assumption stresses the risk of introduction of the bacteria both biologically and mechanically . Animals that live in close contact can become infected with C . burnetii because bacteria are shed from infected animals by vaginal secretions , placenta , urine or feces . A previous study described the occurrence of pregnancy terminations in goat herds that were exposed to three goats from another herd that reportedly kidded prematurely during a fair [41] . Moreover , when cows were imported into an area of endemic infection , 40% of uninfected cows became C . burnetii-infected within six months [42] . Viable bacteria have been isolated from sperm of seropositive bulls [43] . Our results demonstrate that even with a relative low seroprevalence in ruminants , C . burnetii is circulating consistently in the island . This was particularly evident in goats , where 21% animals were PCR positive . Questions emerged regarding the potential impact of C . burnetii on the general population as well as persons at risk , such as pregnant women . Thus , we have begun another study to assess the consequences of this bacterium on human health .
|
Q fever is a disease that could be transmitted from animals ( cattle , sheep and goats ) to humans and caused by a bacterium called Coxiella burnetii ( C . burnetii ) . Some human cases exhibiting characteristic clinical signs of that disease have been detected on Reunion Island , a tropical island in the Indian Ocean , but to date , we did not know if these animals could be seen as potential sources of the disease . Thus , a study was undertaken from March 2011 to August 2012 to detect the presence of that bacterium in these animals and to understand how they could get infected themselves . A total of 516 ruminants ( 245 cattle , 137 sheep and 134 goats ) belonging to 71 farms and localized in different environments of the island were selected . Samples of blood , vaginal mucus and milk were concomitantly collected from females , and a questionnaire was submitted to the farmers . Ticks from positively detected farms were also collected . We observed 11 . 8% of cattle , 1 . 4% of sheep and 13 . 4% of goats had already been in contact with the bacterium . Coxiella burnetii was also directly detected in some vaginal and milk samples . None of the ticks were detected to be positive for C . burnetii . We found that the ruminants could be infected when their farm was exposed to prevailing winds because the bacterium can be transported by the wind , and when there were no specific precautions for visitors before entering the farm , because they could act as mechanical carriers of Coxiella . Conversely , keeping new animals under surveillance for some days to detect any signs of the disease before they enter the farm or keeping the animals in the barn at night limit the risk of infection .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"bacterial",
"diseases",
"infectious",
"diseases",
"veterinary",
"diseases",
"zoonoses",
"veterinary",
"epidemiology",
"medicine",
"and",
"health",
"sciences",
"veterinary",
"microbiology",
"biology",
"and",
"life",
"sciences",
"q",
"fever",
"veterinary",
"science"
] |
2014
|
Emergence of Coxiella burnetii in Ruminants on Reunion Island? Prevalence and Risk Factors
|
Rates of spontaneous mutation determine the ability of viruses to evolve , infect new hosts , evade immunity and undergo drug resistance . Contrarily to RNA viruses , few mutation rate estimates have been obtained for DNA viruses , because their high replication fidelity implies that new mutations typically fall below the detection limits of Sanger and standard next-generation sequencing . Here , we have used a recently developed high-fidelity deep sequencing technique ( Duplex Sequencing ) to score spontaneous mutations in human adenovirus 5 under conditions of minimal selection . Based on >200 single-base spontaneous mutations detected throughout the entire viral genome , we infer an average mutation rate of 1 . 3 × 10−7 per base per cell infection cycle . This value is similar to those of other , large double-stranded DNA viruses , but an order of magnitude lower than those of single-stranded DNA viruses , consistent with the possible action of post-replicative repair . Although the mutation rate did not vary strongly along the adenovirus genome , we found several sources of mutation rate heterogeneity . First , two regions mapping to transcription units L3 and E1B-IVa2 were significantly depleted for mutations . Second , several point insertions/deletions located within low-complexity sequence contexts appeared recurrently , suggesting mutational hotspots . Third , mutation probability increased at GpC dinucleotides . Our findings suggest that host factors may influence the distribution of spontaneous mutations in human adenoviruses and potentially other nuclear DNA viruses .
DNA viruses have been traditionally viewed as slowly-evolving entities , but this notion has been challenged in the last decade after the discovery of several highly diverse and fast-evolving DNA viruses [1–6] . The pace of evolution should be dependent on the rate at which new spontaneous mutations are produced , yet it is currently accepted that DNA virus mutation rates are typically much lower than those of RNA viruses [7] . However , as opposed to RNA viruses , few mutation rate estimates have been obtained for DNA viruses , which include four bacteriophages ( φX174 , m13 , λ , and T4 ) , herpes simplex virus , and human cytomegalovirus [7–11] . Moreover , these estimates were derived from indirect , phenotype-based methods of mutation detection , used very small portions of the viral genome , or suffered from bias due to selection acting on population mutation frequencies . Therefore , we currently lack an unbiased , genome-wide view of how spontaneous mutations are produced in DNA viruses . Although next-generation sequencing ( NGS ) has made it possible to analyze genetic variation in full-length DNA virus genomes with unprecedented detail , its relatively low per-read accuracy has prevented detection of rare variants , including new spontaneous mutations . This problem has been solved in recently-developed methods that increase the accuracy of NGS by orders of magnitude [12 , 13] , now permitting an in-depth characterization of DNA virus spontaneous mutation rates . Adenoviruses are non-enveloped icosahedral , viruses with double-stranded linear DNA genomes of 26–45 kbp . They infect a broad range of vertebrates , and over 60 serotypes of human adenoviruses have been identified and grouped into seven species ( A-G ) and serotypes [14] . Human adenoviruses can cause a wide variety of diseases including eye , gut and respiratory infections that are typically not clinically relevant for healthy individuals , yet potentially life-threating for immuno-compromised patients [15] . The adenovirus genome is flanked by inverted terminal repeats ( ITR ) containing the origins of replication and encodes early-expressed proteins required for DNA replication ( transcription units E1A , E1B , E2 , E3 and E4 ) , late structural proteins ( transcription units L1 , L2 , L3 , L4 and L5 ) , and intermediate-expressed proteins ( IX and IVa2 ) . Transcription units can be oriented in either direction and undergo extensive alternative splicing to yield diverse mRNA products , referred to as genes . Adenoviruses constitute an excellent model for studying the evolution of DNA viruses due to their high prevalence , broad tropism , tractability , and relatively simple genomes . Sequencing of isolates from patients has revealed peaks of genetic diversity at some regions of the hexon , fiber , penton base genes , and E3 , which encode exposed domains of the capsid required for host cell binding and entry , or proteins involved in virus-host interactions and immune evasion [16 , 17] . Some of these hypervariable regions are being currently used for genotyping adenovirus isolates by PCR and Sanger sequencing , as well as by NGS [18] . Here , we have used high-fidelity NGS to analyze the genome-wide rate of spontaneous mutation of human adenovirus C5 ( HAdv5 ) under controlled laboratory conditions . After an endpoint dilution step to remove pre-existing diversity and short-term culturing to minimize the effects of natural selection , we sequenced >98% of the viral genome with >1000-fold coverage and extremely high accuracy . This allowed us to identify >200 spontaneous mutations produced at different positions of the viral genome . The estimated rate of spontaneous mutation was 1 . 3 × 10−7 per site per cell infection cycle , and the mutational spectrum was dominated by G-to-A and C-to-T base transitions . We found that different transcriptions units mutated at roughly similar rates , indicating that hypervariable regions originate mainly by selection and not by mutational hotspots . However , we found several sources of mutation rate heterogeneity . First , GpC dinucleotides showed increased mutation probability . Second , we identified two 5 kpb regions approximately mapping to E1B , IX , IVa2 and L3 which showed a significant reduction in the number of accumulated mutations . Third , several individual genome sites located within low-complexity sequence contexts exhibited the same mutations recurrently in independently replicating viruses , suggesting the presence of mutational hotspots . Some of these mutation rate heterogeneities correlated with changes in diversity in publicly available patient-derived sequences .
HAdv5 was subjected to three serial end-point dilution steps in HeLa cells and then re-amplified by two serial transfers in liquid culture at high multiplicity of infection ( MOI ) to obtain sufficient viral genome copies to carry out DNA extraction and NGS without PCR amplification , such that we could avoid PCR-driven sequencing errors ( Fig 1 ) . The endpoint dilution steps ensured that viral growth was initiated from a single infectious unit . Since , after three serial endpoint dilutions , all pre-existing genetic diversity should have been removed , variants observed in the sequenced populations should correspond to newly-produced , spontaneous mutations . HAdv5 DNA was purified from the cytoplasm of infected cells to avoid carrying over large amounts of nuclear cellular DNA , and directly subjected to Duplex Sequencing ( DS ) using the Illumina platform . DS relies on template tagging and strand complementarity to increase base call accuracy by orders of magnitude compared to conventional NGS [13 , 19] . For each of three independent biological replicates , 99% of the HAdv5 genome was sequenced with an average coverage >2500 in each replicate ( Fig 2A ) . We found 68 , 78 , and 62 different single-base substitutions in 93 . 2 , 115 . 7 , and 123 . 7 Mb sequenced , respectively , yielding an average per-base mutation frequency of ( 6 . 4 ± 0 . 7 ) × 10−7 ( Table 1 ) . These mutations were present at frequencies below 1% in sequence reads ( S1 Table ) . Hence , it is highly unlikely that these were pre-existing polymorphisms carried forward from the initial population , because any variant that may have survived the three endpoint dilutions should have reached very high population frequencies after such drastic bottlenecks . To test whether the observed variants were sequencing errors , we performed DS of a purified E . coli plasmid pUC18 , which should exhibit low diversity given the high replication fidelity of bacteria . This yielded three single-nucleotide substitutions in 21 . 0 Mbp sequenced , which implies a maximal per-base sequencing error rate of 1 . 4 × 10−7 assuming that the plasmid contained no diversity . Thus , at least 80% of the observed variants should correspond to real mutations . Based on the estimated sequencing error , the net mutation frequency was ( 6 . 4–1 . 4 ) × 10−7 = 5 . 0 × 10−7 . In addition to single-nucleotide substitutions , the HAdv5 sequences contained 70 point insertions and deletions ( S1 Table ) , but these types of mutations were present in the control plasmid at similar frequencies , indicating that most should be sequencing errors . Therefore , we did not use insertions and deletions for mutation rate estimation . To further check the nucleotide substitutions detected by DS , we set out qPCR assays aimed at selectively amplifying wild-type and mutant alleles in four genome sites . For mutations detected by DS , the Ct values obtained in qPCRs designed to amplify the mutant allele were delayed by approximately 6 to 12 cycles compared to those in which the wild-type sequence was amplified . In contrast , for mutations not detected by DS , this difference increased to 12–20 cycles ( S1 Fig ) . This further supports the conclusion that most DS mutations were real . G→A transitions and the reverse complementary C→T transitions were the most abundant type of mutations detected by DS ( 42% ) , followed by A→G/T→C ( 24% ) , whereas transversions were 2 . 0 times less abundant than transitions ( Table 1 ) . We found that the nearly twofold excess of G→A and C→T changes over A→G/T→C was accounted for by a higher mutation probability at GpC motifs . Specifically , of the 3167 such dinucleotides in the HAdv5 genome , 50 contained G→A or C→T mutations , whereas the 13 , 380 G and C bases that were not part of GpC motifs showed only 39 mutations , revealing a 2 . 7-fold excess probability of G→A /C→T mutation at GpC dinucleotides ( Fisher test: P < 0 . 001 ) . In contrast , CpG motifs only showed a 1 . 3-fold excess C→T/G→A mutation probability compared to other G and C bases . The observed mutational pattern at GpC motifs is unlikely to be explained by selection , because HAdv5 was passaged only twice after the bottleneck , minimizing the ability of selection to purge spontaneous mutations . Furthermore , these passages were done at high MOI , which reduces the efficacy of selection [20–23] . The lack of significant selection was further supported by analysis of synonymous and non-synonymous variation . Based on the observed mutational spectrum , we expected 71 . 5% of base substitutions at coding regions to be non-synonymous under a neutral model . The observed fraction was 74 . 5% , a value that did not deviate significantly from this expectation ( dN/dS = 1 . 16 , chi-square test: P = 0 . 327 ) . In the absence of selection , the observed mutation frequency should equal the mutation rate per cell infection cycle times the number of infection cycles elapsed [7] . Each of the two high-MOI transfers allowed for one infection cycle and , using our estimate of the HAdv5 burst size ( ca . 104 infectious units per cell ) , approximately two additional cycles were required for growth of the initial infectious unit isolated by endpoint dilution ( see Methods ) . Hence , the calculated point mutation rate per cell infection cycle was 5 . 0 × 10−7 / 4 = 1 . 3 × 10−7 . Genetic diversity is not uniformly distributed in adenovirus genomes [16 , 17] , raising the question whether mutation rates vary accordingly . Although mutations were generally well-scattered along the HAdv5 genome , we found two regions of approximately 5 kpb showing significantly fewer mutations than expected under the assumption of a constant rate ( Fig 2B ) . Specifically , the region encompassing genome sites 18 , 500 to 23 , 500 showed only seven total base substitutions in the three biological replicates , whereas the expected number was 29 . 17 ( Binomial test: P < 0 . 001 ) , the reduction being statistically significant in each of the three replicates ( P ≤ 0 . 014 ) . This region approximately maps to genes located in transcription unit L3 , which encodes a protease and two capsid proteins ( VI and hexon protein ) . Another low-mutation region encompassed genome sites 500 to 5500 , which includes transcription units E1B , IX , and IVa2 ( Binomial test: P = 0 . 001 ) , although in this case the reduction was significant in only two of the three replicates . One possible factor driving mutation depletion may be purifying selection acting specifically at these regions . However , the percentage of non-synonymous mutations in L3/E1B-IVa2 was similar to the rest of the genome ( 76 . 9% versus 74 . 5% , respectively ) , suggesting no differential selection pressure at the protein-coding level . We also found no major differences in the frequency of GpC dinucleotides between the low-mutation regions and the rest of the genome , and the reduction in mutation frequency in L3 was still significant after removing all G→A /C→T changes from the analysis ( Binomial test: P < 0 . 001 ) . Hence , the mechanisms driving mutation rate variation along the viral genome remain unclear . Aside from the reduced mutation frequency in these two regions , the observed number of mutations per transcription unit correlated well with the expected number assuming a constant mutation rate ( Fig 2C; Chi-square test: P = 0 . 672 ) . Finally , to further test for mutation clustering , we obtained the empirical distribution of the distance between consecutive mutations . Assuming a constant rate , this distance should follow a geometric distribution with parameter equal to the per-site mutation probability . The data were in broad agreement with this null model , although there was a slight increase in the proportion of mutations showing small distances , suggesting some level of clustering . Interestingly , we identified 11 individual genome sites that exhibited the exact same mutation in at least two of the three biological replicates , suggesting the presence of mutational hotspots ( Table 2 ) . For instance , genome site 9417 , which maps to transcription unit E2B ( terminal protein precursor gene ) showed a T-to-G substitution in 0 . 28% , 0 . 54% and 0 . 77% of the reads of each replicate , a frequency that exceeds the genome-wide mutation frequency by orders of magnitude . This site was flanked by a G-rich motif ( GGTGGGG ) , such that the mutation produced a G heptamer . Most other mutations were point insertions . A common feature of these recurrently appearing mutations was a low-complexity sequence context with frequent homopolymeric runs . This type of sequence context reduces replication fidelity by inducing frequent polymerase slippage , but also elevates the sequencing error rates . Based on this , we cannot rule out the possibility that these mutations were sequencing artefacts . In genome site 14 , 073 , which maps to the end of L1 , we found a frequent insertion that was further accompanied by a marked decrease in sequencing coverage , from >1000 to approximately 200 ( Fig 2A ) . In contrast , other similar motifs such as a poly-T 11-mer at genome positions 34 , 337–34 , 347 and a poly-T 10-mer at genome positions 1161–1170 did not show a similarly marked decrease in sequencing coverage . Under a null model in which the mutation rate shows no appreciable variation along the HAdv5 genome , mutated sites should represent a random sample of genome sites , hence , should not show particularly elevated diversity or evolvability . To test this , we downloaded from GenBank 35–52 adenovirus C sequences for each gene ( except the L3 hexon gene , for which only 15 well-aligning sequences could be retrieved; S1 Dataset ) and calculated Li and Nei´s nucleotide diversity per site , defined as the probability that pairs of sequences differ at that particular site ( Fig 3A ) . Of the 29 , 858 genome sites examined , 2463 ( 8 . 2% ) were polymorphic . Overall , the distribution of diversity values was similar for genome sites showing mutations in our experimental system and for those showing no mutations ( Fig 3B ) , supporting no major effects of mutation rate variation on in vivo diversity of HAdv5 . However , when we focused on the low-mutating regions L3 and E1B-IVa2 , the fraction of polymorphic sites in database sequences dropped to 5 . 2% , versus 9 . 1% outside these regions ( Fisher test: P < 0 . 001; Fig 3C ) , suggesting that reduced mutation rate limits in vivo diversity in these regions . We also tested whether the 11 sites showing recurrent mutations in our experimental design exhibited the same mutations in database sequences . If these sites were true mutational hotspots , database sequences should tend to show variation at these sites too . In contrast , if these were DS artefacts , most should not be polymorphic or , alternatively , they may systematically show sequence changes similar to those detected by DS if low-complexity sequence contexts also led to errors in database sequences . Interestingly , recurrently mutated sites that mapped to non-coding regions showed variation in GenBank sequences , whereas those mapping to coding regions did not ( Table 2; Fig 3D; S2 Fig ) . This strongly suggests that most of the proposed mutational hotspots are real and that , in patient-derived sequences , purifying selection has removed these mutations from coding regions , but not from non-coding regions .
Per-base mutation rates correlate negatively with genome sizes over a broad range of DNA microorganisms including viruses , bacteria , and unicellular eukaryotes [7 , 24 , 25] . As a result , the genomic mutation rate varies weakly , and a quasi-constant rate of approximately 0 . 003 mutations per genome per round of copying was suggested [24] . For HAdv5 , our calculated mutation rate per cell infection cycle is 1 . 3 × 10−7 or , equivalently , 0 . 0046 per 35 . 9 kbp genome , in good agreement with the suggested rule . In recent work with human cytomegalovirus , de novo mutations were identified in longitudinal patient samples and , using the estimated duration of the cell infection cycle for this virus in vivo , the calculated mutation rate was 2 . 0 × 10−7 [11] . Previous work with murine cytomegalovirus gave a very similar estimate of 1 . 4 × 10−7 , although this mutation rate was measured per day instead of per cell infection cycle [8] . In herpes simplex virus , the mutation rate was estimated by scoring null mutations in the tk gene using ganciclovir [10] . This yielded an estimated rate of 5 . 9 × 10−8 per cell infection cycle [7] . Therefore , mutation rates for different human DNA viruses measured by widely different methods vary within approximately twofold around 10−7 mutations per base per cell infection cycle . Genome sizes range from 35 . 9 kbp for HAdV to 150 kpb for herpes simplex virus and 230–236 kbp for cytomegaloviruses . As a result , genomic mutation rates vary by approximately an order of magnitude , and are substantially lower than the 0 . 003 expected value for the largest DNA viruses ( Fig 4 ) . This cast some doubts on the proposed relationship between genome sizes and mutation rates , at least for viruses . However , estimates for cytomegaloviruses and herpes virus suffer from limitations imposed by scoring only a few , potentially unrepresentative , genome sites , and from biases associated to selection , and hence should be taken with caution , making it premature to draw conclusions . We suggest , though , that factors other than genome size per se determine DNA virus mutation rates , such as whether the virus encodes its own polymerase or whether the viral genomic DNA is single-stranded or double-stranded . For instance , viruses encoding their own polymerases should have greater capacity to optimize mutation rates in response to selective pressures such as mutational load , the costs of replication fidelity , or adaptation to novel environments , compared to viruses that use host-encoded polymerases . Also , single-stranded DNA should be more prone to spontaneous damage and host-mediated editing than double-stranded DNA . Future work should help clarify the molecular mechanisms and evolutionary processes underlying mutation rate variation across viruses . Whereas from an evolutionary standpoint mutation rates per cell infection cycle are meaningful because the cell infection cycle is the equivalent of a viral generation , from a biochemical perspective use of per round of copying estimates better reflects the actual fidelity of replication . If approximately 104 HAdv5 infectious units are produced per cell , there should be at least log2 ( 104 ) = 13 . 3 rounds of semi-conservative replication per cell , giving a rate of 1 . 3 × 10−7/ 13 . 3 = 9 . 8 × 10−9 per-round-of copying . Notice that this estimate is robust to gross uncertainties in burst size measurements . If , for instance , we assume only 103 HAdv5 infectious units per cell , the estimated mutation rate per round of copying would change only weakly , i . e . 1 . 3 × 10−7/ log2 ( 103 ) = 13 . 0 × 10−9 . The mutation rate of normal human cells is on the order of 10−9 per round of copying [25] . In contrast , typical mutation rates for tumoral cells are at least 100-fold higher [27] , with recent estimations for various types of cancers ranging from 10−7 to 10−6 [28] . Here , we used human cervix tumor HeLa cells for HAdv5 growth , suggesting that the HAdv5 mutation rate per round of copying was similar or even lower than that of the host cell . Whether adenoviruses use cellular post-replicative repair is unclear and the fact that tumoral cells as those used here typically show aberrant repair pathways precludes us from addressing this question here . Use of normal cells for HAdv5 growth may help clarify this point in future studies . It is well-established that most DNA viruses cause genome instability and interact with repair and DNA damage response ( DDR ) pathways [29 , 30] . For instance , the adenovirus E4orf6 protein recruits an ubiquitin ligase and promotes the proteasomal degradation of TOPBP1 , an activator of DDR via ATR [31] , and defects in the adenoviral E4 gene lead to the formation of concatemers of viral genomes with heterogeneous junctions [32] . However , it remains unclear whether activation or suppression of the DDR determines DNA virus mutation rates . Although two 5 kpb regions with an over two-fold reduction in mutation rate were identified , we found no dramatic differences in mutation rate across adenovirus transcription units . This contrasts with the ability of some organisms to critically increase the mutation rate of some loci . Targeted hyper-mutation has been described in the immunoglobulin genes of B lymphocytes [33] , contingency loci encoding surface proteins in some bacteria [34] , and genes encoding tail fiber proteins in some DNA bacteriophages [35] . The involved mechanisms include DNA editing , polymerase slippage in DNA tandem repeats , and error-prone reverse transcription . Adenovirus genetic diversity shows ample across-gene variation and is highest in genes involved in virus-host interactions [16 , 17] . In principle , adenoviruses should also benefit from targeting mutations to these specific loci , but we found no such targeting in HAdv5 . Potential hotspots were restricted to low-complexity sequence contexts and did not span entire genes but only specific nucleotide sites . Interestingly , one such possible hotspot , the poly-A motif in the 3´end of gene L1 , has also been identified as a recombination hotspot and is located between a relatively conserved genome region ( 3´of the hotspot ) and a more variable region ( 5´of the hotspot ) [17] . It can be speculated that frequent replication errors in this low-complexity region may recruit proteins of the post-replicative system , which would increase the likelihood of recombination events . Finally , we also found that GpC dinucleotides were 2 . 7 times more prone to mutation than G or C bases alone . In vertebrates , cytosines in CpGs are frequently methylated and their spontaneous deamination produces thymidines , leading to high rates of C→T substitutions [36] but , in contrast , there is little evidence for methylation or elevated mutation rates at GpC motifs [37] . Furthermore , adenovirus DNA is poorly methylated [38 , 39] and hence this mutational pathway should be infrequent . As such , the potential mechanisms leading to increased mutation rate at GpC motifs in HAdv5 remain to be investigated .
HAdv5 was a generous gift from Dr . Ramón Alemany ( Bellvitge Biomedical Research Institute ) and was propagated in HeLa cells from ATCC . Cells were free of mycoplasma as determined by a PCR test . Cells were cultured at 37°C and 5% CO2 in Dulbeco’s Eagle’s medium ( DMEM ) supplemented with 10% FBS and antibiotics . HeLa H1 cells at 70–80% of confluency were used for plaque assays in 6-well plates . After viral adsorption , cells were washed with PBS and incubated for 4–5 days at 37°C with 5% CO2 in a semi-solid medium containing DMEM supplemented with 1% FBS , 1% penicillin-streptomycin , and 0 . 8% noble agar overlaid with a nutrient medium layer of DMEM supplemented with 1% FBS , 1% penicillin-streptomycin , glucose and GOP supplement . Cell monolayers were fixed with 4% formaldehyde , and stained with 2% crystal violet . Viral titers were expressed as plaque forming units ( PFU ) /mL . Viral infections were done in DMEM supplemented with 5–10% FBS and 1% penicillin-streptomycin until cytopathic effect was observed . The virus was first subjected to three transfers at endpoint dilutions in 96-well plates ( 10 days per transfer ) . The last transfer was then further amplified in a 10-cm plate ( 3 days ) and the supernatant was used to inoculate five 10-cm plates of confluent HeLa cells at an MOI of 5 PFU/cell ( 3 days ) . Cells were harvested by centrifugation and washed with PBS . A suspension of approximately 107 cells was incubated 30 min on ice with 1 mL cell lysis buffer ( 50 mM Tris-HCl pH 8 . 5 , 150 mM NaCl and 1% Triton X-100 ) . Cellular debris were removed by centrifugation at 14 , 000 rpm for 10 min at 4°C , and the supernatant was treated with 10 mg/mL RNAse A for 1 h at 37°C . Then , viral lysis buffer ( 10% SDS , 0 . 5M EDTA and 10 mg/mL proteinase K ) was added and samples were incubated for 1 h at 56°C . Viral DNA was extracted with phenol/ chloroform and resuspended in ddH20 for direct DS sequencing . As a sequencing control , pUC18 DNA was purified using the NucleoSpin Plasmid Kit ( Macherey-Nagel , Germany ) . All DNA samples were quantitated using the Qubit dsDNA BR Assay kit ( Life Technologies , USA ) prior to sequencing . The increased accuracy of DS is based on ligation of sequencing adapters containing random yet complementary double-stranded nucleotide sequences [19] . These molecular tags allow tracing each strand of the original double-stranded template and removal of mutational artefacts that appear in only one of the two strands . DS adapters were constructed by annealing of two oligonucleotides , one of which contained a 12-nt single-stranded randomized sequence tag . Annealed primers were extended using the Klenow fragment , digested to obtain cohesive ends , and used as the final DS adapters for library preparation as previously described [13] . The purified HAdv5 DNA was fragmented using a Covaris sonicator and size selection was performed with Ampure X beads . Subsequent steps including sequencing library construction were exactly performed as detailed in previous work [13] . Samples were run on a NextSeq machine ( Illumina ) with 150 bp reads . FastQ files were processed with the DS software pipeline ( https://github . com/loeblab/Duplex-Sequencing ) using BWA 0 . 6 . 2 , Samtools 0 . 1 . 19 , Picard-tools 1 . 130 and GATK 3 . 3–0 , and GenBank accession AY601635 as reference sequence . The computational workflow relies into three major steps: tag parsing and initial alignment , single stranded consensus sequence ( SSCS ) assembly , and duplex consensus sequence ( DCS ) assembly . Finally , the processed DCS data were realigned to the reference genome to analyze each genomic position and count mutations . Default parameters were used except for family size , which was reduced from 3 to 2 to increase the number of reads . Analysis of HAdv5 and puC18 sequencing outputs at family size 2 and 3 yielded nearly identical mutation frequencies , indicating that the reduced family size did not increase sequencing error appreciably ( S2 Table . Accession AY601635 was used to extract HAdv5 gene coordinates , and deduced changes in protein sequences were obtained with a homemade R script . The actual sequence of the virus used in the experiments , as obtained from the DS consensus , was identical to AY601635 except for two single-base substitutions: a T→C change at position 18 , 813 , and a G→C change at position 30 , 598 . The DCS-final . bam . pileup text file was used to visualize insertions/deletions . The DS output is available from the NCBI SRA database ( www . ncbi . nlm . nih . gov/sra; accession SRP091328 ) . We set out to check mutations at the following genome sites: 191 ( G→A ) , 7295 ( A→G ) , 9417 ( T→G ) , and 33 , 565 ( A→G ) . For each , we designed three primers for SYBR green-based qPCR: one located approximately 200 bp upstream of the mutated site and two for which the 3´base matched exactly the mutated site , one containing the mutant base and another containing the wild-type base ( WT , see S3 Table for primer information ) . We performed two separate qPCR reactions under the same conditions , using the mutant or wild-type primer , and a modified Taq DNA pol as provided in Agilent´s Brilliant III Ultra-Fast SYBR Green qPCR Master Mix . Since pairing of the 3´base is critical for DNA extension and Taq DNA pol has no 3´exonuclease activity , each of these two alternate qPCRs should amplify selectively the mutant or wild-type DNA . All reactions were carried using 0 . 3 ng of the purified HAdv5 DNA and 200 nM of each primer . Preliminary gradient PCRs were carried out using WT primers to set the annealing temperature as high as possible ( 67°C , 69°C , 78°C , and 68 . 5°C for each of the sites , respectively ) . The thermal qPCR profile was as follows: an initial denaturation step ( 95°C 3 min ) , 40 cycles of amplification ( 95°C 5 s , annealing temperature 10 s , 72°C 10 s ) , and a final melting cycle ( 95°C 30 s , 65°C 30 s , 95°C 30 s ) . Data analysis and Ct estimation were carried out using the AriaMx software provided by the manufacturer . Each qPCR was performed in triplicate . To estimate the number of PFUs produced per infected cell ( burst size ) , HeLa cells were infected with HAdv5 at an MOI of 10 PFU/cell in a 24-well plate . The supernatant was collected at 3 and 7 days post inoculation and titrated by the plaque assay . The burst size was calculated as the number of PFUs produced per culture at day 3 divided by the number of cells , giving 9 . 86 × 103 PFU/cell . Titers at day 7 were slightly higher , yielding a burst size of 2 . 60 × 104 PFU/cell . Since each well of a 96-well plate contained approximately 3 × 104 cells , we estimate that roughly two infectious cycles were required to fully infect a well initially containing a single PFU . To infer Nei and Li´s nucleotide diversity , sequence AY601635 was used as query to carry out Blastn analyses on a per-gene basis , and all hits corresponding to human adenovirus C sequences were retrieved and aligned using the Muscle algorithm implemented in Mega 7 ( www . megasoftware . net ) . The analysis was done gene by gene to facilitate alignment . For E3 , since there are abundant alternative coding regions , we arbitrarily defined our query as a region spanning AY601635 genome positions 27 , 500 to 31 , 000 . Similarly , our query for the E4 regions encompassed AY601635 genome positions 32 , 500 to 35 , 119 . A , T , G , C base frequencies ( f ) at each site were obtained using a custom script and Nei and Li´s diversity was calculated as Ⅱ = 1 –fA2 –fT2 –fG2 –fC2 . To analyze sequence polymorphism in sites showing recurrent mutations in our system , an approximately 300 pb region flanking the site of interest was used as query for a Blastn . Human adenovirus C sequences were retrieved and visually inspected to assess polymorphism .
|
Next-generation sequencing has provided a powerful tool for studying microbial genetic diversity but suffers from relatively low per-base accuracy , limiting our ability to detect low-frequency polymorphisms and spontaneous mutations . However , this limitation has been solved recently by the development of high-fidelity deep sequencing techniques . Taking advantage of these advancements , here we provide the first unbiased genome-wide characterization of the rate of spontaneous mutation of a human DNA virus ( adenovirus 5 ) under controlled laboratory conditions . The adenovirus genome shows a relatively low mutation rate , consistent with high replication fidelity and the action of post-replicative repair . We also found evidence for mutation rate heterogeneities and regions of genetic instability in the viral genome . Together with previous reports , our findings indicate that DNA viruses with large double-stranded genomes mutate significantly slower than those with small single-stranded genomes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"sequencing",
"techniques",
"microbial",
"mutation",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"insertion",
"mutation",
"pathogens",
"microbiology",
"viruses",
"mutation",
"substitution",
"mutation",
"dna",
"viruses",
"sequence",
"motif",
"analysis",
"molecular",
"biology",
"techniques",
"microbial",
"genomics",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"viral",
"genomics",
"medical",
"microbiology",
"microbial",
"pathogens",
"molecular",
"biology",
"point",
"mutation",
"virology",
"viral",
"pathogens",
"gene",
"identification",
"and",
"analysis",
"genetics",
"adenoviruses",
"biology",
"and",
"life",
"sciences",
"mutation",
"detection",
"genomics",
"organisms"
] |
2016
|
Genome-Wide Estimation of the Spontaneous Mutation Rate of Human Adenovirus 5 by High-Fidelity Deep Sequencing
|
To survive , organisms must extract information from the past that is relevant for their future . How this process is expressed at the neural level remains unclear . We address this problem by developing a novel approach from first principles . We show here how to generate low-complexity representations of the past that produce optimal predictions of future events . We then illustrate this framework by studying the coding of ‘oddball’ sequences in auditory cortex . We find that for many neurons in primary auditory cortex , trial-by-trial fluctuations of neuronal responses correlate with the theoretical prediction error calculated from the short-term past of the stimulation sequence , under constraints on the complexity of the representation of this past sequence . In some neurons , the effect of prediction error accounted for more than 50% of response variability . Reliable predictions often depended on a representation of the sequence of the last ten or more stimuli , although the representation kept only few details of that sequence .
Organisms often operate in unknown and uncertain environments . Therefore , extracting aspects of past observations , which are maximally predictive of the relevant future , is essential for survival . It has been suggested that the sensory cortex evolved to extract the statistical regularities of the world [1] . Adaptation of the nervous system to the statistical structure of the input is reflected in studies of neuronal responses to natural stimuli . For example , in the auditory system , auditory nerve fibers–part of the auditory periphery–have been shown to achieve high coding efficiency by implementing a “tuned” nonlinear filter that selectively amplifies the anticipated signal [2] . Similarly , in the visual system , Laughlin [3] showed that the contrast-response function of interneurons in the fly's compound eye approximates the cumulative probability distribution of contrast levels in natural scenes . The central auditory system shows sensitivity to stimulus statistics as well . Event-related potentials recorded in humans show sensitivity to deviant stimuli . This sensitivity may occur rather early , at the mid-latency potentials range [4] , and has been intensively studied in the context of the mismatch negativity ( MMN ) , peaking about 150 ms after the point of deviance [5] . Similar sensitivity occurs also in the responses of single neurons in auditory cortex: using oddball sequences composed of two frequencies with different probabilities , Ulanovsky et al . [6] found that neurons in cat auditory cortex responded more strongly to a given tone frequency when it was rare than when it was common . This sensitivity , named stimulus-specific adaptation ( SSA ) by Ulanovsky et al . [6] , has been by now shown in multiple mammalian species and even in birds ( See [7] for review ) . SSA may be linked to deviance detection in the time frame of the mid-latency potentials rather than to MMN ( Grimm , Escera and Nelken 2015 ) . Thus , we hypothesize that neurons in the auditory system encode some notion of a prediction error . The coding of prediction error is considered central to learning [8] , memory formation [9] , and decision-making [10] . Prediction errors are also known to be related to efficient coding where only the unexpected at one stage of processing should be transmitted to the next stage [11] . This paper has two goals . The first is to present a theory of prediction error from first principles . For an organism that operates in a statistically stationary world ( as is generally the case in laboratory experiments ) , prediction quality is limited by the statistical structure of the data that the organism collects from the past . The theory depends only on this statistical structure , and not on any assumptions about specific brain mechanisms . We assume that the brain forms a reduced representation of the past , which serves to generate predictions of future events . We use information theory to quantify both the complexity of these reduced representations and the predictive information they carry with respect to future events [12] . The term complexity refers here to the rate of information ( specifically , the mutual information in bits/s or , equivalently but more conveniently here , in bits/stimulus ) that the representation carries about the past . Predictive information refers to the rate of information ( again , mutual information in bits/s or bits/stimulus ) that the reduced representation carries about future events . Both terms are defined precisely below . Extraction of the predictive aspects of the past stimuli can be formalized as an optimization problem: minimize the complexity of the reduced representation of the past while preserving a predefined level of predictive information . A reduced representation provides a predictive probability for every event–this is the probability assigned to the event just before it actually occurred , given the reduced representation of the past . Our theory uses these predictive probabilities to calculate prediction errors–events with low predictive probability generate large prediction error , while events with a high predictive probability generate small prediction errors . The constrained optimization problem we use to calculate the reduced representations is a special instance of the Information Bottleneck ( IB ) principle [13] . The IB principle applies to any two random variables X and Y with a known joint distribution . Like the special case described here in details , the IB principle provides a way of finding reduced representations ( as defined later in the paper ) of X that are maximally informative about Y . Here we apply it to the past of the stimulation sequence ( X ) and to its future ( Y ) . Much of the detailed discussion below can be considered as a primer for the use of the IB principle . The second goal of the paper is to demonstrate the usefulness of the theory by applying it to the neuronal responses evoked by random tone sequences consisting of two frequencies with varying probabilities ( ‘oddball’ sequences ) . We show that prediction errors calculated by the theory correlate well with neuronal responses . Most importantly , we use the theory to extract parameters of the reduced representations that underlie these responses: we show that these reduced representations have a long duration ( typically N ≥ 10 preceding stimuli ) but keep only coarse details about the sequence that was presented ( i . e . , the reduced representations have low complexity ) . These results show how to establish properties of the neural code rigorously from first principles .
In order to demonstrate the relevance of our approach to neuronal processing , we studied responses of single neurons in primary auditory cortex ( A1 ) to oddball sequences . Oddball sequences are generated by selecting two stimuli ( two pure tones in our case ) and then forming a sequence composed of these two stimuli in which one of the two is common and the other is rare ( Fig 3A ) . Usually , the overall number of times each of the stimuli occur in the sequence is fixed . We will however test here a slightly different statistical model , in which the probability of each of the tones is fixed . We will furthermore assume that successive tones are selected independently of each other with a given probability ( ‘Bernoulli sequences’ ) . This assumption , which is a very good approximation to the experimental setting when the number of tones in the sequence is large , makes the application of the theory particularly transparent . Neurons in auditory cortex are sensitive to stimulus probability in such sequences [6 , 19–23] . The majority of neurons in cat auditory cortex produced a larger average response to the same tone frequency when it was rare than when it was common ( Fig 3B ) . These responses often had substantial sustained components that could outlast stimulus duration ( e . g . [24] ) , resulting in large spike counts within the counting window we used ( 0–330 ms after stimulus onset , for stimuli whose total duration was 230 ms , same as in [6 , 20] ) . We interpret these responses as encoding a prediction error: in the oddball sequences , the probability of the next stimulus to be either of the two frequencies is roughly given by the probability with which it occurred in the past . The prediction error , −log2 ( p ) , for the tone when rare is therefore larger than for the same tone when common . Thus , responses of neurons in auditory cortex seem to fit qualitatively the notion of prediction error as defined here . Using the theory presented above , we aimed to quantify this intuition rigorously . We start by describing in substantial detail the family of optimal reduced representations in the case of the oddball sequences , and then we show how we used these representations to extract information about the memory duration and complexity of the reduced representation underlying the dependence of neuronal responses in auditory cortex on tone probability . In the experiments , sequences with a number of different probability values have been used: each stimulus was used both as common ( with probabilities of 90% and 70% ) and as rare ( with probabilities of 10% and 30% ) , as well as in equiprobable sequences . Formally , the stimulus sequence can be modeled as a Bernoulli process , where one stimulus is drawn with probability p ( 0<p<1 ) and the other with probability 1-p . We assume that the parameter p is drawn from a uniform distribution prior to generating each stimulus block . This assumption is not crucial–it changes the exact numerical values , but not the trends that will be discussed below , as long as all the tone probability conditions that occurred in the experiment are allowed under the prior . These assumptions determine the joint probability PN ( past , future ) characterizing the statistical structure of the stimulus sequence , where past denotes the sequence of the last N stimuli and future denotes the next stimulus in the sequence ( see Methods ) . The memory duration , N , is a parameter of the model and will be discussed in more details below . As we discussed above , the most detailed ( and complex ) representation in this case would consist of a perfect encoding of the past sequence by identifying each of the 2N possible configurations of the stimulation sequence with a unique state m ( Fig 2A; case ( i ) is this representation for N = 4 ) . This is , however , an unnecessarily detailed representation since the number of occurrences of the B tone among the last N is a sufficient statistic . The IB method identifies the relevant information for predicting the next stimulus: in our case , it assigns a single and unique state to all sequences that share the same number of occurrences of A tones and B tones ( Fig 2A; case ( ii ) ) . This reduced representation filters out non-relevant details of the past ( i . e . the exact order of the tones ) : it maintains the maximal predictive information possible ( for a given N ) , but at a lower complexity ( 2 . 3 bits instead of 3 . 6 bits for N = 4 ) . Fig 2B ( gray line ) displays the maximal predictive power at each complexity of the reduced representation for N = 4 , and cases ( i ) and ( ii ) can be directly compared to each other . Using the IB method enabled us to reduce the complexity of the representation even further . While representations that are simpler than the minimal sufficient statistic ( case ii ) do reduce the quality of the prediction of the next stimulus , the loss in predictive power may be relatively small . For example , for a memory of N = 4 tones , constraining the reduced representation complexity to 1 bit results in a ‘noisy’ representation with 3 unique states ( Fig 2A and 2B , case ( iii ) ) . In this case , given the reduced representation alone , it is impossible to recover the exact number of B tones that occurred in the past ( their exact order was already lost in the reduction to the minimal sufficient statistic ) . Nevertheless , this solution is optimal in the sense that its predictive power is maximal among all possible reduced representations with a complexity of 1 bit ( for N = 4 ) . For longer past durations and large complexities , the low sensitivity of predictive power to complexity is even more prominent . For a memory of N = 10 tones , the complexity reduction when using the sufficient statistic is larger than for N = 4 , and further reducing the complexity of the representation by another 57% ( from 3 . 46 to 1 . 49 bits ) results in a loss of merely 12% in its predictive power ( from 0 . 223 to 0 . 196 bits; Fig 2B , black ) . We examined the tradeoff between the complexity and the predictive power of the reduced representation for memory durations from N = 1 to N = 50 ( Fig 4A ) by applying the IB method to the probability distributions of the oddball paradigm PN ( past , future ) . Note that these probability distributions depend on N , the duration of the past . We used the IB algorithm to find , for each N , 200 reduced representations whose complexity spanned the full allowed range ( see below ) , and whose predictive power was maximal for each value of the complexity . Each curve in Fig 4A is produced by linearly interpolating between these 200 points . The 200 points , however , sample the relevant ranges so densely that the linear interpolation is imperceptible . Thus , each point along the curves shown in Fig 4A corresponds to a unique optimal reduced representation: as described above , such reduced representation is given by the two conditional distributions P ( m|past ) and P ( future|m ) . The curves in Fig 4A plot the predictive power as a function of the complexity for these 200 reduced representations ( the N = 4 and N = 10 cases are the same lines plotted in Fig 2B ) . These convex curves describe quantitatively the tradeoff between complexity and predictive power . Their convexity is a general property of the tradeoff postulated by the IB principle ( see [13] ) . For each N , complexity ( abscissa of Fig 4A ) spans the range between 0 and the complexity of the sufficient statistic . The predictive power ( ordinate in Fig 4A ) spans the range between 0 and the mutual information between past and future , which is equal to the mutual information between the sufficient statistic and the future . Thus , the rightmost point of each line shows the complexity of the sufficient statistic ( abscissa ) and the predictive power of the sufficient statistic ( ordinate ) . More complex representations do exist ( e . g . the full past ) , but do not result in an increased predictive power . As complexity is lowered below that of the sufficient statistic ( moving leftward along a curve ) , predictive power is lost , but at least initially the loss of predictive power is rather minor . The curves separate achievable and non-achievable combinations of complexity and predictive power . Any point on or below the curve is achievable , in the sense that there is ( at least one ) pair of conditional distributions P ( m|past ) and P ( future|m ) with the corresponding complexity and predictive power . Any point above the curve is non-achievable in this sense–there is no way to process the information from the past , keeping that level of complexity , and still get better predictions than those specified by the curve . The dependence of optimal predictive power on N , the duration of the past , is illustrated in Fig 4B for different constraints imposed on the complexity ( 0 . 5 , 1 , 2 bits and with no constraints ) . Two effects are readily apparent . First , for each level of complexity , increasing past duration much above N = 10 does not result in a major increase in predictive power . Second , as hinted above , although the complexity of the sufficient statistic for long memory duration may be as high as 4–5 bits ( Fig 4A ) , there is a strong effect of ‘diminishing returns’–complexity above 2 bits does not add substantial amount of predictive power to the reduced representation , even for a past duration of N = 50 stimuli ( Fig 4B ) . One way of understanding these effects is by noting that the predictive power reflects the precision of the probability estimates of the B tone . While this precision increases with both N and complexity , beyond a certain point increasing the precision at which the probability of the B tone is known does not improve much the predictions anymore . To test the hypothesis that neurons in the auditory cortex represent the prediction error derived from a reduced representation of the recent past , we correlated the prediction errors , derived from the reduced representations we computed , with the neuronal responses to oddball sequences ( Fig 5 ) [6 , 20] . We calculated separately the prediction errors for each one of the reduced representations we computed above ( defined by memory duration N = 1 to N = 50 and 200 complexity values ranging from 0 to the maximum possible for each N , for a total of 200*50 = 10 , 000 reduced representations ) . For each reduced representation , for each neuron in the dataset , and for each of the two frequencies with which the neurons were tested , we used the actual sequence of tone presentations used in the experiment to calculate a corresponding sequence of prediction errors . We followed the prescription described above . For each tone presentation along the sequence , we used the previous N tone presentations as the past . Each past leads to a state m of the reduced representation , with an associated estimate P ( future|m ) for the probability of the upcoming stimulus . The prediction error associated with that stimulus was −log2 P ( future|m ) . As explained above , the prediction error itself is a random variable . The transformation from past to reduced representation , P ( m|past ) , provided a set of probabilities to be in each of the states of the reduced representation , and through them to each of the possible prediction errors ( see Methods ) . Therefore , instead of a single prediction error for each stimulus , we calculated a set of prediction errors , one for each possible state of the reduced representation m , together with their probabilities . To compare the prediction errors with the neuronal responses , we used linear regression of the neuronal responses against the prediction errors , weighted by the corresponding probabilities ( see Methods for details ) . For illustration purposes , in Fig 5 we plot neuronal responses ( spike counts evoked by individual tone presentations , ordinate ) against the average prediction error ( −log2 P ( future|m ) averaged over all possible values for the state m of the reduced presentation , each with its probability P ( m|past ) ) for that same stimulus presentation . In the experiments , neurons were tested using blocks composed of two tones whose frequency separation was about half octave ( fhigh/flow = 1 . 44; this is the Δf = 0 . 37 case of [6 , 20] ) . The probabilities of the two tones in each block were fixed , but their order was random . For the main analysis ( Figs 5 and 6 ) , we used a conservative counting window ( 0–330 ms after stimulus onset; tone duration was 230 ms and onset-to-onset interval was 730 ms ) , and we used the responses of all 68 neurons tested with 3 blocks in which the tone probabilities were 90%/10% , 50%/50% , and 10%/90% ( for the low frequency and high frequency tones respectively ) . The use of responses from these 3 blocks ensured sampling of the full range of relevant values of prediction errors . The analysis was performed separately for each frequency , resulting in 136 combinations of neuron and frequency , with 117/136 combinations showing responses to the corresponding tone that were significantly larger than the spontaneous rate . Prediction error was significantly correlated with neuronal responses in many cases . Fig 5A compares responses of one neuron ( spike counts evoked by single presentations of one of the two frequencies with which it was tested ) with the expected prediction errors calculated for each tone presentation using three different optimal reduced representations . These representations had a complexity of 1 bit and past durations of N = 5 , N = 10 and N = 15 stimuli . Each block had 400 stimuli , of which 90% ( 360 ) , 50% ( 200 ) or 10% ( 40 ) consisted of the relevant frequency . To avoid using stimuli whose past sequence was not well defined , the first 50 stimuli of each block were removed from the analysis . Due to the randomized nature of the sequences , for each combination of neuron and test frequency we ended with 311–322 stimuli in the p = 90% condition ( blue points ) , 168–182 stimuli in the p = 50% condition ( black points ) , and 28–39 stimuli in the p = 10% condition . As the tones presented in the 10% condition were , by definition , rare , they were associated with large prediction errors , and therefore the red points are mostly concentrated at the right of the scatter plots . Similarly , the tones presented in the 90% condition were , by definition , common , and were associated with small prediction errors . Therefore the blue points are mostly concentrated at the left of the scatter plots . On the other hand , for the 50% condition , prediction errors varied quite substantially , leading to a larger dispersion along the abscissa between ( and partially overlapping ) the two extremes . While the correlation with the prediction errors corresponding to N = 10 was slightly larger than for shorter or longer memory durations , this example mainly illustrates the finding that the correlation between the prediction error and neuronal responses was sometimes only weakly dependent on past duration , a finding we will return to later . To illustrate the range of goodness of fit that could be achieved , Fig 5B compares single-trial responses of three neurons ( columns; the rows show the two frequencies used for testing each neuron ) versus their expected prediction errors , calculated using reduced representations with complexity of 2 bits and past duration of N = 10 stimuli . Neurons 1 and 2 had strong dependence of their responses on prediction error , while Neuron 3 was typical for the data . To quantify these observations for the entire data set , we performed weighted linear regression analysis , separately for each combination of a neuron and frequency that evoked significant responses ( n = 117 combinations ) , using the entire set of pre-calculated optimal solutions . The distribution of the resulting best goodness-of-fit scores ( rmax2 ) is displayed in Fig 6A . We tested the significance of the fit separately for each combination of neuron and frequency . The goodness of fit scores were calculated as the maximum r2 over all possible reduced representations , so that standard significance tests could not be applied due to the inherent multiple comparisons involved . We therefore tested the significance of a fit using a permutation test . We repeated the same analysis for the measured spike counts but replacing the stimuli by 20 randomly permuted sequences with the same tone probabilities , thus breaking the short-term relationships between responses and associated prediction errors . The effect of prediction error on the neuronal responses was considered to be significant ( p<0 . 05 ) if rmax2 for the actual stimulation sequence was larger than rmax2 calculated using 20 different random permutations . The prediction error significantly influenced the neuronal responses in about 2/3 of cases ( 78/117 combinations , 67%; see Fig 6A , red bars ) , supporting our hypothesis that auditory cortical neurons represent this formal notion of prediction error in their responses . Most neurons contributed to Fig 6A twice , so the figure could overstate the degree to which prediction error and neuronal responses are related to each other . Fig 6B is a scatter plot of the goodness of fit values for the two frequencies , with the corresponding histograms for the two frequencies separately . There was a mild correlation between the goodness of fit values for the two frequencies ( r = 0 . 34 , n = 68 , p<0 . 05 , using the normal approximation to the Fisher z-transformed correlation ) , suggesting that the representation of prediction error in the neuronal responses was at least to some degree a property of the neuron itself . Considering each frequency by itself , 41/62 neurons that responded significantly to the low frequency and 37/55 of those that responded significantly to the high frequency showed a significant dependence of neuronal responses on prediction error . More conservatively , 52 neurons had significant responses to both frequencies , and 27 of these neurons showed significant dependence of the responses to both frequencies on prediction error . Thus , depending on the criterion , between 1/2 and 2/3 of the neurons showed significant dependence ( p<0 . 05 ) of the neuronal responses on prediction error . In order to characterize the past duration and the complexity of the representations that best fitted the neuronal responses , we analyzed the parameters of these representations that best explained the data . For this analysis we used only combinations of neuron and test frequency with explanatory power of rmax2≥0 . 1 ( adjusted correlation coefficient > 0 . 33 , n = 34/117 , see Methods ) . We found that the dependence of the goodness of fit scores on the parameters ( memory duration and complexity ) could be rather weak once either past duration or complexity were large enough , so that many different models achieved approximately the same goodness of fit . Therefore , we defined , for each combination of neuron and test frequency , a set of good representations consisting of those that achieved goodness-of-fit scores higher than an ( admittedly somewhat arbitrary ) threshold: r2≥0 . 9∙rmax2 . The parameters of these ‘good representations’ ( past duration , complexity and predictive power ) are summarized in two-dimensional , color-coded histograms ( Figs 6C–6E ) . For each reduced representation , the histograms show the fraction of cases for which that specific representation belonged to the set of good representations . The reduced representations are index in three different ways ( by complexity and predictive power , Fig 6C; complexity and duration , Fig 6D; predictive power and duration , Fig 6E ) . Thus , Fig 6D shows that reduced representations with duration N < 10 stimuli were outside the set of good representations for most cases . Thus , for a reduced representation to be good for a large fraction of cases , it had to have a relatively long past duration , N ≥ 10 stimuli ( corresponding to 7 . 3 seconds or longer ) . Similarly , Fig 6D shows that for a reduced representation to be good for a large fraction of cases , it could have a complexity as low as about 2–3 bits ( white line in Fig 6D ) –reduced representation with lower complexities tended to be outside the set of good representations for most cases . Moreover , although the resulting reduced representations were relatively coarse ( low complexity ) their predictive power almost matched the predictive power attainable with models that have maximal complexity ( Fig 6C and 6E; compare with Fig 4A and 4B ) . Thus , the good reduced representation with shortest memory tended to have a long memory ( N>10 ) , coarse ( complexity of 2–3 bits relative to maximal complexity of ~5 bits ) , but keep a high degree of predictive power .
We present here an information theoretic formulation for the problem of sensory perception and prediction in the brain . We suggest that the representation of past stimuli reflects a tradeoff between the complexity of the reduced representation and its predictive power . The prediction errors , calculated from the statistics of the stimulation sequence using first principles , were significantly associated with the measured neuronal responses . Importantly , these results suggest that the relevant time scale for the calculation of the prediction error is very long relative to the time scale usually considered in sensory coding–on the order of several seconds or longer . Our results therefore suggest that neurons in auditory cortex rely on surprisingly long time scales to calculate prediction errors , although the representations of the past , which underlie the computation of the prediction errors over these time scales , were found to be rather coarse . Attneave [28] and Barlow [29] already suggested that neural information processing might follow principles of information theory . Information theory provides universal bounds on the minimal expected prediction error that can be achieved , independent of other assumptions on the implementation of a predictive process by the brain . In this sense the theory is normative: it specifies absolute bounds that cannot be improved in any way , and that are achieved by specific reduced representations of the past sequence . These bounds are governed by only two parameters , the complexity of the reduced representation ( or , alternatively , its predictive power ) and the duration of the past memory used for perception . The notion of predictive coding is not new . The importance of the interplay between the representation complexity and the accuracy of predictions has been noticed before ( e . g . [12] ) . In recent years , the tradeoff has been used extensively by Friston [30] . Friston used an approximation to the exact Bayesian inference problem , which is very difficult . Our approach circumvent this difficulty: crucially , it differs from previous attempts to formally use the interplay between complexity and prediction quality in that the prediction error in our formulation is derived from a reduced representation rather than directly from the explicit past stimulus sequence . Most importantly , one unique aspect of our approach is that we use a model-independent ( information theoretic ) bound . In spite of the generality of this approach , we could apply it to real experimental situations: we show how to use it for studying rigorously the neuronal code in one simple case . It turns out that the neural responses reflect the prediction error derived from efficient reduced representations , allowing us to extract bounds on the duration of the memory that underlie the observed neuronal responses as well as on its complexity: memory duration is remarkably long ( longer than 10 stimuli back ) but rather coarse ( with a complexity less than 2 bits ) . The theory presented here can be considered as part of a more general theory of neural function . Since the reduced representation is assumed to be internal to the organism , its complexity may be related to the metabolic cost required to maintain and update it . As we illustrated in the case of the oddball sequences , less complex representations in our sense are also simpler to implement biologically–they require a smaller number of states , and incorporate noisy assignments of past sequences to states of the representation . On the other hand , constraints on expected prediction error may stand for general constraints on future value . Thus , our theory can be seen as an application of the general principle that organisms attempt to minimize metabolic or other costs subject to future value constraints . This principle unifies most known learning and control theoretic models , potentially linking information theoretic measures with general biological constraints directly [31] . Our approach provides a principled way to study long-term dependencies in neuronal responses , which are inaccessible to many models of auditory responses [32–34] . These models relate the neuronal responses with the preceding content of the stimulus within a short time window ( typically 50–100 ms ) . Attempts to include context in such models end up with analyzing relatively short-term contextual effects , from a few ms [16] to a few tens of ms [35] and up to a few hundreds of ms [36 , 37] . Only a few attempts to quantify longer-range dependencies have been published . For example , Ulanovsky et al . [20] demonstrated ( in a previous analysis of the data presented here ) that the larger responses to rare tones could not be accounted for by taking into account only a recent past ( N ≤ 4 preceding stimuli , 3 seconds ) . They concluded that the larger responses to rare stimuli depended on longer segments of the stimulation sequence , although the effective memory duration and its content were left unspecified in these studies . Similarly , studies that fitted more mechanistic models of synaptic depression to such data [19 , 23] concluded that such models cannot fully account for the data , but such conclusions were limited by the restricted range of models that have been considered . Our main experimental observation in this paper is that the dependence of neuronal responses on prediction error calculated over long time scales can be made precise . This observation supports the notion that auditory cortex neurons carry a prediction error signal . Furthermore , we specify for the first time potential properties of the reduced representation underlying auditory cortical responses to oddball sequences at such long time constants . In particular , we demonstrate that the neuronal responses are compatible with reduced representations that have long memory duration but low complexity . Importantly , the descriptive power of our theory is not restricted to statistically simple sequences such as the oddball paradigm presented here . The full power of the theory resides in more complex stimulation sequences that may lack sufficient statistics ( e . g . [21] ) . Of particular interest are the statistical regularities related to the syntax and semantics of language and music , which span multiple temporal scales . The tools developed here allow for a formal examination of the sensitivity of neurons to the complex statistical regularities of real-world soundscapes and therefore present a broad framework for characterizing sensory perception both qualitatively and quantitatively .
The neuronal responses are those described in [6 , 20] , which also contain the detailed experimental methods . In brief , extracellular recordings were made in primary auditory cortex ( A1 ) of halothane-anesthetized cats . Anesthesia was induced by Ketamine and Xylazine and maintained with halothane using standard protocols authorized by the committee for animal care and ethics of the Hebrew University—Hadassah Medical School . Single units were spike sorted on-line using template-based sorting , and in most cases they were also sorted off-line , to improve unit isolation . Stimuli were presented to the animal through sealed , calibrated earphones . For the oddball paradigm , two frequencies were selected close to the best frequency of the neuron , with a frequency ratio fhigh/flow = 1 . 44 . Ulanovsky et al . ( 2003; 2004 ) defined the frequency difference slightly differently , and this is their Δf = 0 . 37 condition . The two frequencies were presented in a number of blocks . Each block contained 400 pure tone stimuli of identical duration ( 230 ms ) , inter-stimulus interval ( 736 ms onset to onset ) and tone level ( approximately 40 dB above the neuron’s minimal threshold ) . The blocks differed by the number of times each frequency was presented . For example , in the 90%/10% block , 360 of the stimuli had frequency flow and 40 had frequency fhigh , presented using a random permutation . We also used blocks with probabilities 70%/30% , 50%/50% , 30%/70% , and 10%/90% . The dataset was composed of 99 neurons tested in the 90%/10% and the 10%/90% cases . Of these , 68 were also tested in the 50%/50% ( these are the neurons used in the main analysis ) ; and of those 68 neurons , 29 neurons were additionally tested in the 70%/30% and 30%/70% conditions . For a detailed presentation of the IB principle & algorithm see Tishby et al . [13] . In short , given a joint distribution P ( x , y ) , the IB finds a compressed representation of x denoted by m that is most informative on the target variable y . The compression of the representation is quantified by the mutual information between x and m , given by I ( x;m ) , and the information that m carries on the target variable y is quantified by I ( m;y ) . Using the IB algorithm we effectively pass the information that x provides about y through a ‘bottleneck’ formed by the reduced representation m , defined by P ( m|x ) . In practice , the reduced representation is determined by minimization of the Lagrangian , L[P ( m|x ) ]=I ( x;m ) −βI ( m;y ) with respect to P ( m|x ) . The positive Lagrange multiplier β , associated with the constraint on I ( m;y ) , controls smoothly the tradeoff between preserving relevant information and the compactness of the representation . The optimal assignment that minimizes the IB Lagrangian satisfies the following equations: P ( m|x ) =P ( m ) Z ( x , β ) exp ( −βDKL[P ( y|x ) ||P ( y|m ) ] ) P ( m ) =∑xP ( x ) P ( m|x ) P ( y|m ) =∑xP ( y|x ) P ( x|m ) where Z ( x , β ) is a normalization ( partition ) function . We used here an iterative algorithm ( over the set of self consistent equations ) to find the optimal solution P ( m|x ) for a given P ( x , y ) and β . In our case , x and y are the past and future of the stimulus , respectively , and m is the reduced representation of the past . We computed reduced representations for the oddball paradigm , for a given past duration of N stimuli . First , we calculated the joint distribution PN ( past , future ) corresponding to the oddball sequences . For any given value of the parameter p ( the probability of the ‘high tone’ in the block ) the events are independent by construction . Thus , for a uniform prior over p , the posterior probability can be calculated explicitly , PN ( past , future=high ) =∫01 ( Nk ) pk+1 ( 1−p ) N−kdp=k+1 ( N+1 ) ( N+2 ) where k = 0 , 1 , . . , N indicates the number ‘high tones’ in past . Then we applied the IB method over PN ( past , future ) to find reduced representations of the past that are predictive with respect to the future ( i . e . , the next stimulus ) . We repeated this procedure with 200 different values of β to span the entire range of complexity and predictive power levels , and with 50 different values of N spanning past durations from N = 1 to N = 50 . We ended up with a set of 200 × 50 reduced representations , where each one was defined through the conditional probability distributions: P ( m|past ) and P ( future|m ) . We used the conditional probability distributions P ( m|past ) and P ( future|m ) to calculate prediction errors along the stimulation sequence , as follows . For each stimulus ( future ) we considered its previous N stimuli ( past ) to determine the probability of entering each state m of the reduced representation using P ( m|past ) . The probability with which the future stimulus is then expected P ( future|m ) was used to calculate the prediction error −log2 P ( future|m ) associated with that stimulus . Since this quantity depends on the unknown state m , we used the expected value of the prediction error ( with respect to m ) to generate Fig 5: E ( future|past ) =−∑mP ( m|past ) log2P ( future|m ) Note that the expected prediction error with respect to the stimulus distribution P ( past , future ) is negatively related to the predictive power , I ( m;future ) : ⟨E ( future|past ) ⟩P ( past , future ) =−∑past , m , futureP ( past , future ) P ( m|past ) log2P ( future|m ) =−∑past , m , futureP ( past , m , future ) log2P ( future|m ) =−∑m , futureP ( m , future ) log2P ( future|m ) =H ( future|m ) =H ( future ) −I ( m;future ) where H ( Future ) is the entropy of the future stimulus and does not depend on the reduced representation . It follows that maximizing the predictive power is equivalent to minimizing the expected prediction error . For the main analysis , spike counts were measured in a window of 330 ms , starting at stimulus onset and ending 100 ms after stimulus offset . For the analysis displayed in Fig 7 , shorter windows were used as well . We collected the neuronal responses ( represented as spike counts ) for each neuron and for each test frequency ( ‘low’ or ‘high’ ) across the oddball stimuli blocks . For each combination of neuron , test frequency , and reduced representation ( 200 × 50 pre-calculated reduced representations ) we calculated the prediction error at each stimulus along the actual stimulation sequence used in the experiments . Since the prediction error is calculated with respect to a past window of up to N = 50 stimuli , the first 50 stimuli in each block were omitted from the analysis , resulting in ( 400 − 50 ) × 3 = 1050 stimuli . Dividing the stimuli further into ‘low’ and ‘high’ tone-frequencies , resulted in about 525 stimuli for the analysis of each combination of neuron and test frequency . The reduced representation determines the prediction error associated with each stimulus for each state −log2 P ( future|m ) . For each stimulus we considered the previous N stimuli , using them to estimate the probability of entering each state m by P ( m|past ) . These probabilities served as weights in calculating the regression between spike counts and prediction error values . Using the method of weighted linear regression allowed us to take the uncertainty in the unknown state m into account . Finally , the fraction of explained variance based on this weighted linear regression ( weighted r2 ) was used to measure the goodness-of-fit associated with each one of the reduced representations . For each combination of neuron and test frequency we denoted the highest fraction of explained variance over the entire set of reduced representations by rmax2 . We considered reduced representations that achieved a score of at least 90% of rmax2 as ‘good representations’ . For each combination of neuron and test frequency used in the main analysis , we constructed two-dimensional binary maps indicating the set of ‘good representations’ over the parameter space ( complexity , predictive power and past duration: see Fig 6C–6E ) . For Fig 6 , We calculated the averages of these binary maps over a subset of the population , which had high explanatory power ( rmax2≥10%; n = 34/117 ) .
|
A crucial aspect of all life is the ability to use past events in order to guide future behavior . To do that , creatures need the ability to predict future events . Indeed , predictability has been shown to affect neuronal responses in many animals and under many conditions . Clearly , the quality of predictions should depend on the amount and detail of the past information used to generate them . Here , by using a basic principle from information theory , we show how to derive explicitly the tradeoff between quality of prediction and complexity of the representation of past information . We then apply these ideas to a concrete case–neuronal responses recorded in auditory cortex during the presentation of oddball sequences , consisting of two tones with varying probabilities . We show that the neuronal responses fit quantitatively the prediction errors of optimal predictors derived from our theory , and use that result in order to deduce the properties of the representations of the past in the auditory system . We conclude that these memory representations have surprisingly long duration ( 10 stimuli back or more ) , but keep relatively little detail about this past . Our theory can be applied widely to other sensory systems .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"sequencing",
"techniques",
"auditory",
"cortex",
"medicine",
"and",
"health",
"sciences",
"computational",
"biology",
"brain",
"random",
"variables",
"neuroscience",
"learning",
"and",
"memory",
"probability",
"distribution",
"mathematics",
"forecasting",
"statistics",
"(mathematics)",
"cognition",
"computational",
"neuroscience",
"memory",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"sensory",
"physiology",
"animal",
"cells",
"mathematical",
"and",
"statistical",
"techniques",
"molecular",
"biology",
"probability",
"theory",
"auditory",
"system",
"cellular",
"neuroscience",
"cell",
"biology",
"anatomy",
"physiology",
"neurons",
"single",
"neuron",
"function",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"sensory",
"systems",
"cognitive",
"science",
"statistical",
"methods"
] |
2016
|
The Representation of Prediction Error in Auditory Cortex
|
The factors influencing variation in the clinical forms of Chagas disease have not been elucidated; however , it is likely that the genetics of both the host and the parasite are involved . Several studies have attempted to correlate the T . cruzi strains involved in infection with the clinical forms of the disease by using hemoculture and/or PCR-based genotyping of parasites from infected human tissues . However , both techniques have limitations that hamper the analysis of large numbers of samples . The goal of this work was to identify conserved and polymorphic linear B-cell epitopes of T . cruzi that could be used for serodiagnosis and serotyping of Chagas disease using ELISA . By performing B-cell epitope prediction on proteins derived from pair of alleles of the hybrid CL Brener genome , we have identified conserved and polymorphic epitopes in the two CL Brener haplotypes . The rationale underlying this strategy is that , because CL Brener is a recent hybrid between the TcII and TcIII DTUs ( discrete typing units ) , it is likely that polymorphic epitopes in pairs of alleles could also be polymorphic in the parental genotypes . We excluded sequences that are also present in the Leishmania major , L . infantum , L . braziliensis and T . brucei genomes to minimize the chance of cross-reactivity . A peptide array containing 150 peptides was covalently linked to a cellulose membrane , and the reactivity of the peptides was tested using sera from C57BL/6 mice chronically infected with the Colombiana ( TcI ) and CL Brener ( TcVI ) clones and Y ( TcII ) strain . A total of 36 peptides were considered reactive , and the cross-reactivity among the strains is in agreement with the evolutionary origin of the different T . cruzi DTUs . Four peptides were tested against a panel of chagasic patients using ELISA . A conserved peptide showed 95 . 8% sensitivity , 88 . 5% specificity , and 92 . 7% accuracy for the identification of T . cruzi in patients infected with different strains of the parasite . Therefore , this peptide , in association with other T . cruzi antigens , may improve Chagas disease serodiagnosis . Together , three polymorphic epitopes were able to discriminate between the three parasite strains used in this study and are thus potential targets for Chagas disease serotyping .
Chagas disease , a zoonosis caused by the protozoan parasite Trypanosoma cruzi , affects approximately 10 million people in the Americas . Approximately 14 , 000 deaths occur annually , and 50 , 000–200 , 000 new cases are diagnosed each year [1] . During the acute phase of infection , diagnosis is based on parasitological methods [2]; however , in the chronic phase , such parasitological approaches have a low sensitivity , between 50–65% , because of low levels of parasitemia [3] , [4] . The chronic phase is also characterized by a strong and persistent humoral immune response , thus the measurement of IgG antibodies specific for parasite antigens should be performed for diagnosis [5] . However , serological methods from different laboratories have been observed to be inconclusive or contradictory [6]–[8] . These discrepancies are mainly related to technical errors and antigen composition because crude or semi-purified protein extracts of epimastigotes , a parasite stage not found in the mammalian host , are generally used [6] , [9] . Moreover , false-positive results are frequently observed because of the cross-reactivity of crude preparations of T . cruzi antigens with sera from individuals infected with Leishmania sp . and T . rangeli [10]–[12] . The use of recombinant antigens and synthetic peptides as a substitute for parasite lysates has increased reproducibility and , in addition , does not require the maintenance and processing of live parasites [13] , [14] . Despite recent advances in Chagas disease diagnostics , the methods available still have limitations related to low specificity and sensitivity [15] , [16] . Among the factors that compromise the performance of diagnostic tests , the genetic variability of the parasite is known to contribute to false-negative results in Chagas disease serodiagnosis [17] . Epidemiological , biochemical , and molecular studies have demonstrated that the T . cruzi taxon is extremely polymorphic [18]–[21] . Recently , T . cruzi strains were reclassified into six DTUs ( discrete typing units ) called TcI to TcVI [22] , and there is much speculation regarding whether this parasite variability could be associated with different disease prognoses . Although T . cruzi infection results in a broad spectrum of clinical forms as indeterminate , cardiac , and digestive forms , the determinant factors involved in the development of each clinical form have not been elucidated , though it is likely that genetic factors of the host and parasite are involved [23] . However , no study to date has found an unequivocal association between the infecting parasite DTU and the clinical forms of the disease . Nevertheless , this hypothesis has not been discarded because correlations between the geographic distribution profiles of different T . cruzi DTUs and a higher frequency of specific clinical forms have been reported [21] . Indeed , digestive manifestations are more common in the central region of Brazil and the southern part of South America , where infection by TcII , TcV , and TcVI predominates; in contrast , such manifestations are rare in the northern part of South America and in Central America , where infection caused by TcI is more common [24] . Correlation studies between the parasite DTU and clinical forms of Chagas disease are challenging because most of the techniques require parasite isolation from patient blood or parasite genotyping directly from infected tissues . Because many T . cruzi populations are polyclonal , hemoculture may select sub-populations of parasites more adapted to in vitro growth conditions [25] . Moreover , because of different tissue tropisms of some T . cruzi strains [26] , in infections caused by polyclonal populations and/or co-infections , the clones circulating in the patient blood may not be the same as those found in tissue lesions . The current methodologies to genotype the parasite from tissue biopsies are laborious and expensive , thus limiting the number of samples that can be analyzed . Within this context , a parasite typing method based on the detection of strain-specific antibodies from patient sera could resolve many of these problems . Thus far , there is only one study that proposes the use of an antigen to discriminate among T . cruzi DTUs [17] . This study is based on an antigen named TSSA ( trypomastigote small surface antigen ) , belonging to the TcMUC III protein family , which can differentiate between humans infected with TcI , TcIII , and TcIV and those infected with TcII , TcV , and TcVI . In the present study , we performed a genomic screen to identify polymorphic and conserved linear B-cell epitopes in the predicted proteome of the CL Brener T . cruzi strain in an attempt to identify targets for the serotyping and serodiagnosis , respectively , of T . cruzi-infected patients . The results were validated using sera from experimentally infected mice and chagasic patients .
The design and methodology of all experiments involving mice were in accordance with the guidelines of COBEA ( Brazilian College of Animal Experimentation ) , strictly followed the Brazilian law for “Procedures for the Scientific Use of Animals” ( 11 . 794/2008 ) , and were approved by the animal-care ethics committee of the Federal University of Minas Gerais ( protocol number 143/2009 ) . The study protocol involving human samples from Bolivia was approved by the ethics committees of the study hospital , A . B . PRISMA , Johns Hopkins University and the U . S . Centers for Disease Control and Prevention . All subjects provided written informed consent before blood was collected . As for the Brazilian patients , written informed consent was obtained from the participants and was approved by the Ethics Committee of the Federal University of Minas Gerais ( UFMG ) , under protocol number No . 312/06 . Each experimental group was composed of six 2–4-week-old C57BL/6 male mice . The mice were infected with 50 Colombiana or 500 Y trypomastigotes . For the CL Brener clone , we used three mouse groups infected with 50 , 100 , or 500 trypomastigotes . Infection was confirmed by the observation of trypomastigote forms in blood collected from the tail at seven days after intraperitoneal inoculation . One additional group was infected with 1×105 T . rangeli trypomastigotes , and the infection was confirmed by PCR [27] . Six un-infected mice were used as the control group . The chronic phase of infection was confirmed after approximately 3 months by negative parasitemia and the presence of anti-parasite IgG ( as tested against T . cruzi and T . rangeli crude antigens ) by ELISA [28] . Mouse blood was then obtained by cardiac puncture; coagulation was performed at room temperature for 30 minutes , and the serum was obtained after centrifugation at 4000×g for 15 minutes . Blood samples from chagasic patients from Bolivia were collected in a public hospital in Santa Cruz de la Sierra . DNA was extracted from patient blood samples and parasite genotyping was performed as previously described [29] . Infection by TcI parasite lineage was confirmed for six samples ( Supplementary Figure S1 ) . Samples from 10 chagasic patients previously characterized to be infected with TcII [30] and 56 samples from chagasic patients infected with untyped parasites collected from Rio Grande do Norte State , Brazil , were also used . Samples from 14 patients infected with L . braziliensis and 14 patients with visceral leishmaniasis both known to be un-infected with T . cruzi and the sera from 24 un-infected humans were used as specificity and negative controls , respectively . Epimastigotes of the Colombiana and CL Brener clones , and Y strain of T . cruzi and T . rangeli SC-58 were maintained in a logarithmic growth phase at 28°C in liver infusion tryptose ( LIT ) medium supplemented with 10% fetal bovine serum , 100 µg/mL streptomycin , and 100 units/mL penicillin [31] . A total of 1×106 T . cruzi epimastigotes/mL were incubated in triatomine artificial urine ( TAU ) medium for 2 hours at 28°C . L-proline ( 10 mM ) was added to the medium , and the metacyclic forms were obtained after 72 hours at 28°C [32] . Trypomastigotes and amastigotes were obtained from rhesus-monkey epithelial LLC-MK2 cells infected with metacyclic forms cultured in RPMI medium supplemented with 2% fetal bovine serum at 37°C and 5% CO2 [31] . Differentiation of T . rangeli epimastigotes to trypomastigotes was induced with 106 parasites/mL in DMEM medium ( pH 8 ) for 6 days at 28°C [33] . Linear B-cell epitopes were predicted for all the proteins of the CL Brener genome release 4 . 1 [34] using the Bepipred 1 . 0 program with a cutoff of 1 . 3 [35] . The BepiPred program assigns a score to each individual amino acid in a sequence , therefore only amino acids with prediction Bepipred score ≥1 . 3 were considered for the downstream analysis . Proteins encoded by the pair of Esmo and Non-Esmo alleles were aligned using the CLUSTALW program [36] , and each pair of amino acids aligned received a polymorphism score according to the following scale: 0 for identical amino acids; 1 for different amino acids with similar physical-chemical properties; 2 for a mismatch involving amino acids with dissimilar physical-chemical properties; and 3 for a gap position . A perl script based on a sliding window approach that uses a fixed window size of 15 amino acids and an increment of one amino acid identified all 15-mer subsequences in which each individual amino acid has a bepipred score ≥1 . 3 . Those peptides with a polymorphism score above 6 ( sum of the individual amino acid polymorphism scores ) and a mean BepiPred score ≥1 . 3 were classified as polymorphic epitopes; those peptides identical between the Esmo and Non-Esmo haplotypes and with a mean BepiPred prediction score ≥1 . 3 were classified as conserved epitopes . The selected peptides were compared with the predicted proteins from the genomes of L . infantum , L . major , L . braziliensis , and T . brucei ( release 4 . 1 ) [37] using the BLASTp algorithm [38] . Peptides with at least 70% similarity along 70% of the length were discarded . After elimination of peptides with potential cross-reactivity with Leishmania and T . brucei , 50 Esmo-like peptides , 50 Non-esmo-like peptides and 50 peptides conserved with the highest mean Bepipred score were selected . Peptides were synthesized on pre-activated cellulose membranes according to the SPOT synthesis technique [39] . Briefly , Fmoc-amino acids were activated with 0 . 05 mM HOBt and 0 . 1 mM DIC and automatically spotted onto pre-activated cellulose membranes using the MultiPep SPOT synthesizer ( Intavis AG ) . The non-binding sites of the membrane were blocked with 10% acetic anhydride , and the Fmoc groups were removed with 25% 4-methyl piperidine . These processes were repeated until peptide chain formation was complete . After synthesis , side-chain deprotection was performed by adding a 25∶25∶1 . 5∶1 solution of trifluoroacetic acid , dichloromethane , triisopropylsilane , and water . The amino acid coupling and side-chain deprotection were monitored by staining the membrane with 2% bromophenol blue . The immunoblotting methodologies followed a previously described protocol [39] . First , the membrane containing peptides was blocked with 5% BSA and 4% sucrose in PBS overnight and incubated with infected and control mouse sera diluted 1∶5 , 000 in blocking solution for 1 hour . After washing three times with PBS-T ( PBS; 0 . 1% Tween 20 ) , the membrane was incubated with the secondary HRP-conjugated anti-mouse IgG antibody ( Sigma-Aldrich ) diluted 1∶10 , 000 in blocking solution for 1 hour . After a third wash , detection was performed using ECL Plus Western blotting ( GE Healthcare ) , following the manufacturer's instructions , with the Gel Logic 1500 Imaging System ( Kodak ) . The densitometry measurements and analysis of each peptide were performed using Image Master Platinum ( GE ) , and the relative intensity ratio ( RI ) cutoff for positivity was determined at 2 . 0 . The soluble peptides were synthesized in solid phase on a 30-µmol scale using N-9-fluorenylmethoxycarbonyl [40] with PSSM8 equipment ( Shimadzu ) . Briefly , Fmoc-amino acids were activated with a 1∶2 solution of HOBt and DIC . The active amino acids were incorporated into Rink amide resin with a substitution degree of 0 . 61 . Fmoc deprotection was then performed using 25% 4-methylpiperidine . These steps were repeated until the synthesis of each peptide was complete . The peptides were deprotected and released form the resin by treatment with a solution of 9 . 4% trifluoroacetic acid , 2 . 4% water , and 0 . 1% triisopropylsilane . The peptides were precipitated with cold diisopropyl ether and purified by high-performance liquid chromatography ( HPLC ) on a C18 reverse-phase column using a gradient program of 0 to 25% acetonitrile . The peptides were obtained with 90% purity , as confirmed by mass spectrometry using Autoflex Speed MALDI/TOF equipment . Each well of flexible ELISA polyvinylchloride plates ( BD Falcon ) was coated with 2 µg of soluble peptide . After blocking with 5% BSA in PBS for 1 hour at 37°C , followed by three washing steps with PBS containing 0 . 05% Tween 20 ( PBS-T ) , the plates were incubated with human or mouse serum ( dilution 1∶100 ) . The plates were washed three times with PBS-T , and secondary HRP-conjugated anti-human or anti-mouse IgG antibody was added for 1 hour at 37°C , followed by four washes . A solution containing 0 . 1 M citric acid , 0 . 2 M Na2PO4 , 0 . 05% OPD , and 0 . 1% H2O2 at pH 5 . 0 was used for detection; the reaction was stopped with 4 N H2SO4 , and the absorbance was measured at 492 nm . The mean optical density value at 492 nm plus three times the standard deviation of the negative serum was used as the cutoff value . For affinity ELISA , 6 M urea was added for 5 min at 37°C after incubation with the primary antibodies; the remainder of the protocol was the same [41] . The results are shown as an affinity index ( AI ) determined as the ratio between the absorbance values of the samples treated and not treated with urea . An AI value lower than 40% represented low-affinity antibodies , between 41 and 70% was classified as intermediate affinity and higher than 70% as high affinity . Genomic DNA extraction was performed using the GFXTM Genomic Blood DNA Purification kit ( GE Healthcare ) following the manufacturer's instructions . The DNA samples were quantified using a NanoDrop Spectrophotometer ND-1000 ( Thermo Scientific ) . The PCR products amplified with the primers listed in Supplementary Table S1 were subjected to sequencing at both ends using the ABI Prism 3730×l DNA Analyzer ( Applied Biosystems ) by Macrogen Inc ( Korea ) . Total RNA was isolated from 108 epimastigotes , 106 trypomastigotes , and 108 LLC-MK2 cells infected with approximately 105 intracellular amastigotes of the Colombiana , Y , and CL Brener strains using the NucleoSpin RNA II RNA extraction kit ( Macherey-Nagel ) following the procedures described by the manufacturer . RNA from 108 LLC-MK2 cells was also extracted and used as a negative control . The concentration and purity of the RNA samples were measured with a NanoDrop Spectrophotometer ND-1000 ( Thermo Scientific ) . cDNA was synthesized using 10 ng of total RNA and the High Capacity cDNA Reverse Transcription Kit ( Applied Biosystems , Foster City , CA ) using random hexamer primers according to the manufacturer's instructions . Specific primers for each Esmo and Non-Esmo allele were designed , and the primer specificity was verified by electronic PCR using the entire parasite genome as a template . The primers used are listed in the Supplementary Table S2 . Real-time PCR reactions were performed in an ABI 7500 sequence detection system ( Applied Biosystems ) . The reactions were prepared in triplicate and contained 1 mM forward and reverse primers , SYBR Green Master Mix ( Applied Biosystems ) , and 20 ng of cDNA . Standard curves were prepared for each experiment for each pair of primers using serially diluted T . cruzi CL Brener genomic DNA to calculate the relative quantity ( Rq ) values for each sample . qRT-PCRs for the constitutively expressed GAPDH gene were performed to normalize the expression of the specific alleles . All statistical analyses were performed using Graph Prism 5 . 0 software . First , the normal distribution of data was evaluated by the Kolmogorov-Smirnov test; because all they showed a Gaussian profile , an unpaired t test was used for the comparative analysis between the two sets of data , and an ANOVA was used for three or more experimental groups . P-values lower than 0 . 05 were considered statistically significant . The sensitivity , specificity , and accuracy of the peptides were also calculated for the human samples . The sensitivity is represented by Se = TP/ ( TP+FN ) , where TP ( true positive ) is the number of sera from individuals infected with T . cruzi above the cutoff value and FN ( false negative ) is the number of sera from infected individuals below the cutoff for the conserved peptide . For the polymorphic peptides , TP was defined as the number of sera from individuals infected with a specific strain above the cutoff value , and FN is the number of these sera below the cutoff for polymorphic peptides . The specificity is represented by Sp = TN/ ( TN+FP ) , where TN ( true negative ) is the number of sera from individuals infected with L . braziliensis or un-infected individuals below the cutoff and FP ( false positive ) is the number of sera from these samples with reactivity the for conserved peptide . For the polymorphic peptides , TN was defined as the number of sera from individuals infected with a non-specific strain or L . braziliensis and uninfected individuals below the cutoff , and FP is the number of sera from these samples with reactivity . The accuracy is calculated as Ac = ( TP+TN ) / ( TP+TN+FP+FN ) .
We performed B-cell epitope prediction for 3 , 983 proteins derived from pair of alleles of CL Brener genome . We decided to restrict our analysis to this dataset because the CL Brener clone is a recent hybrid between the TcII and TcIII DTUs and evidence suggests that the latter is an ancient hybrid between TcI and TcII [42] . Therefore , it is likely that polymorphic epitopes in the pairs of alleles of CL Brener could also be polymorphic for its parental genotypes and other T . cruzi strains . In the CL Brener hybrid diploid genome , it is possible to identify two haplotypes: “Esmo” , which is more similar to TcII; and “Non-Esmo” , which is more similar to TcIII [34] . A total of 1 , 488 predicted epitopes were classified as conserved between the two haplotypes , and 428 were classified as polymorphic . We next excluded epitopes also present in Leishmania major , L . infantum , L . braziliensis and T . brucei to minimize the chance of cross-reactivity , because these parasites share many antigens with T . cruzi [10] , [12] , [16] , thus reducing the number of conserved and polymorphic epitopes to 1 , 086 and 242 , respectively . A total of 50 conserved , 50 polymorphic Esmo-specific , and 50 polymorphic Non-Esmo-specific peptides with high epitope prediction scores were selected for the construction of peptide arrays . The reactivity of the peptides was tested using a pool of sera from six C57BL/6 mice chronically infected with Colombiana ( TcI ) , Y ( TcII ) , or CL Brener ( TcVI ) strains and un-infected mice as the control group ( Figure 1 ) . The quantification of the reactivity was performed by densitometric analysis ( Supplementary Table S3 ) . A peptide was considered reactive and antigenically conserved if its intensity signal with all T . cruzi strains was two times higher than its signal with the sera from un-infected mice . A peptide was considered reactive and antigenically polymorphic if its intensity signal with a specific strain was two times higher than the values with the other two strains and the un-infected mice . A total of 36 peptides were considered reactive with at least one strain ( Figure 1E ) . A conserved peptide with the highest reactivity with all T . cruzi strains ( C6_30_cons ) and three polymorphic peptides specific for Colombiana ( A6_30_col ) , Y ( B2_30_y ) , and CL Brener ( B9_30_cl ) were selected for soluble synthesis , and their reactivity was validated by ELISA . Because immunoblotting assays are semi-quantitative techniques , we validated the results with quantitative ELISA and affinity ELISA assays using individual sera from six C57BL/6 mice chronically infected with the Colombiana ( TcI ) , Y ( TcII ) , or CL Brener ( TcVI ) strains and the un-infected mice as a control group . For the conserved peptide C6_30_cons , no significant difference in the reactivity among the sera from animals infected with different T . cruzi strains was observed ( Figure 2A ) . The sera from mice infected with the Colombiana strain had a higher antibody titer against the A6_30_col peptide compared to the sera from mice infected with the Y strain . More importantly , the affinity antibodies discriminated Colombiana infection from those caused by the other two strains ( Figure 2B ) . An expected recognition profile was also observed for the peptide B2_30_y ( Figure 2C ) : sera from mice infected with the Y strain had a significantly higher antibody titer than those from mice infected with Colombiana , and the highest affinity antibodies generated by the Y strain discriminated its infection from those caused by Colombiana and CL Brener . With regard to peptide B9_30_cl , conventional ELISA was able to discriminate CL Brener infection from that caused by Y and Colombiana ( Figure 2D ) . Because the infection caused by different T . cruzi strains has specific evolution and mortality rates in a mouse model [43] , we infected mice with a distinct parasite inoculum for each strain to reach the chronic phase when the sera were collected . Thus , to evaluate whether the differences in reactivity observed in the ELISA experiments were dependent on the inoculum , we tested the reactivity of sera from mice infected with 50 , 100 , or 500 CL Brener trypomastigotes ( Supplementary Figure S2 ) . There was no significant variation among the different CL Brener inocula , suggesting that the antigenic variability among the parasite strains is the main factor responsible for the distinct recognition profile of the peptides tested in the ELISA experiments . The evaluation of cross-reactivity with sera from mice infected with T . rangeli and from Leishmaniasis patients demonstrated that the peptides are T . cruzi specific ( Supplementary Figures S3 and S4 ) . We next analyzed the polymorphisms of the epitopes identified in this study and predicted their reactivity with sera from individuals infected with T . cruzi strains representative of each DTU ( TcI to TcVI ) . To this end , we first subjected the peptide sequences to an AlaScan analysis [44] to identify the amino acid residues critical to antibody binding . We found that the pattern GXXXXMRQNE in the carboxy-terminal region of conserved peptide C6_30_cons is important for the interaction with the antibodies generated in infection caused by the three T . cruzi strains ( Figures 3A , B , and C ) . As for the polymorphic epitopes , the patterns PPXDXSLXXP in peptide A6_30_col ( Figure 3D ) , QPQPXPQXXXQP in B2_30_y ( Figure 3E ) , and DEXXXXG in B9_30_cl ( Figure 3F ) are critical for binding with the antibodies generated by Colombiana , Y , and CL Brener infections , respectively . We then sequenced the genomic DNA encoding these four epitopes in strains representative of the six T . cruzi DTUs to predict whether the peptides would be recognized in infections caused by different parasite DTUs . It is expected that a peptide would be recognized in infections caused by a specific strain if the amino acid residues critical for antibody recognition are encoded by its genome . Based on this criterion , we predicted that conserved peptide C6_30 would be able to identify infection caused by four of the six T . cruzi DTUs ( Figure 4A ) , whereas peptides A6_30_col and B2_30_y are expected to identify infections caused only by TcI and TcVI ( Figure 4B ) and TcII and TcVI ( Figure 4C ) , respectively . Interestingly , the A6_30_col and B2_30_y epitopes are identical to the Non-Esmo- and Esmo-like CL Brener haplotypes , respectively , reinforcing the hypothesis that the nature of the CL Brener hybrid may have contributions of both the TcI and TcII genomes . B9_30_cl is predicted to identify patients infected with TcIII or TcVI ( Figure 4D ) . Although all peptides are derived from the CL Brener genome , the sera from mice infected with this strain had lower antibody affinities for the A6_30_col and B2_30_y peptides than did the sera from mice infected with the Colombiana or Y strain ( Figure 2 ) . Because CL Brener is a hybrid strain [34] , [42] , [45] , the polymorphic epitopes encoded by its pairs of alleles may have distinct expression levels that could explain the differences in their reactivity . To investigate this further , we designed allele-specific primers for the genes that encode the epitopes to evaluate their expression levels in the trypomastigote and amastigote forms , the parasite stages found in mammalian hosts ( Figure 5 ) . As expected based on the T . cruzi phylogeny [46] , Y expressed only the Esmo-like variants , and Colombiana expressed only the Non-Esmo variants , except for the B9_30 transcript . CL Brener expressed both alleles of all genes , except for the B9_30 transcript . The conserved peptide was expressed by both the Esmo and Non-Esmo haplotypes of CL Brener ( Figure 5A ) . The polymorphic Non-Esmo peptide A6_30_col was expressed by the Colombiana and CL Brener strains ( Figure 5B ) , and CL Brener also expressed the Esmo-like allele for this peptide . The opposite profile was observed for the polymorphic Esmo B2_30_y peptide , whereby only the Y and CL Brener strains expressed the Esmo-like allele and CL Brener also expressed the Non-Esmo allele of this peptide ( Figure 5C ) . CL Brener only expressed the Esmo-like variant of the B9_30_CL epitope , and its level of expression was approximately 5 times higher than in the Y strain ( Figure 5D ) . All previous results were based on a mouse model because the amount of the inoculum , infective strain , and time of infection can be adequately controlled . To test the potential application of these peptides for serodiagnosis and serotyping of human infection , we performed ELISA experiments with sera from chagasic patients with parasites genotyped as TcI or TcII , and healthy individuals . The conserved peptide C6_30_cons showed 95 . 8% sensitivity , 88 . 5% specificity , and 92 . 7% accuracy for the identification of chagasic patients , and no significant differences in the reactivity of sera from patients infected with TcI or TcII was observed ( Figure 6 ) . As expected , peptide A6_30_col showed much higher reactivity with the sera from patients infected with TcI ( Figure 7A ) , with 100% sensitivity , 91 . 9% specificity , and 92 . 6% accuracy; peptide B2_30_y identified most of the individuals infected with TcII ( Figure 7B ) , with 80% sensitivity , 94 . 8% specificity , and 92 . 6% accuracy . Additionally , none of the sera from patients infected with TcI recognized the B2_30_y peptide , and peptide B9_30_cl showed a low reactivity with both TcI and TcII ( Figure 7C ) . All peptides were also T . cruzi specific because the majority of the sera from the patients infected with L . braziliensis were non-reactive ( Supplementary Figure S4 ) .
Despite efforts to identify new targets for the immunodiagnosis of Chagas disease , the impressive genetic variability of T . cruzi strains has imposed serious limitations on the development of high-sensitivity methods [47]–[50] . Additionally , serological cross-reactivity with Leishmania and T . rangeli infections [11] , [12] , [16] compromises the specificity of Chagas disease diagnosis . Therefore , the identification of new T . cruzi-specific antigens conserved among the parasite strains has been recognized as an important research area for Chagas disease diagnosis and control [47] . The polymorphic nature of T . cruzi isolates , on the other hand , opens new avenues for the development of serotyping methodologies to identify the parasite DTU causing infection based on a serological survey . For instance , this would allow large-scale epidemiological studies aimed at correlating the strain causing the infection with the clinical forms of Chagas disease , an open question that has been hampered by the limited number of samples that can be analyzed by the current genotyping methodologies [48] . To the best of our knowledge , only one study has identified a polymorphic epitope among the T . cruzi DTUs [49] . This marker is an immunodominant B-cell epitope of TSSA ( trypomastigote small surface antigen ) , a representative of the TcMUC III gene family . The TSSA-I and TSSA-II isoforms serologically discriminate between animals infected with T . cruzi I from those infected with T . cruzi II , according to the previous DTU classification ( TcII-VI in the current classification ) , respectively . In a serological survey of chagasic patients from Argentina , Brazil , and Chile , anti-TSSA antibodies recognized only the TSSA-II isoform , suggesting that the TcII-VI DTUs are the cause of Chagas disease in those regions . In a more recent study , however , this same research group analyzed the diversity of the TSSA gene in several representatives of each of the six T . cruzi DTUs and found a complex pattern of sequence polymorphism . Based on their analysis , the epitope considered to be specific for TcII-VI was shown to identify the TcII , V , and VI DTUs . In addition , the peptide previously described as TcI specific shares key features with TcIII and IV . Therefore , there is no T . cruzi DTU-specific serological marker identified thus far . The goal of this work was to identify conserved and polymorphic linear B-cell epitopes of T . cruzi for Chagas disease serodiagnosis and serotyping using ELISA . This technique was selected because it is a quantitative assay and easily automated , thus allowing the analysis of a large number of samples . In recent years , synthetic peptides used as antigens have shown high sensitivity and specificity in diagnostic tests [50] . Peptides have several advantages over chemically purified or recombinant antigens because their production does not involve the manipulation of living organisms and can be obtained with a high level of purity [51] . Recently , the use of peptide arrays has allowed the immunoscreening of a large number of epitope candidates [39] . Thus , an approach based on a synthetic peptide array was chosen to screen of a large number of potential antigens by immunoblotting , followed by ELISA validation . Initially , we screened the CL Brener genome to predict epitopes that are polymorphic and conserved between the Esmo and Non-Esmo haplotypes . The rationale underlying this strategy is that , because the CL Brener strain is a recent hybrid between the TcII and TcIII DTUs and there is evidence suggesting that the latter is an ancient hybrid between TcI and TcII [46] , it is likely that the polymorphic epitopes between the CL Brener alleles would also be polymorphic among distinct T . cruzi strains . The Colombiana ( TcI ) and CL Brener ( TcVI ) clones and Y ( TcII ) strain were selected for this study to evaluate the degree of polymorphism of epitopes in TcII , a direct representative of one CL Brener parental DTU , and TcI , a more distant DTU of CL Brener , along with CL Brener . The immunoscreening of 150 high-scoring peptides resulted in the identification of 36 novel epitopes , indicating that our computational approach for the prediction and prioritization of epitope candidates was successful . Our rate of success ( 24% ) was slightly higher than previously described ( 19 . 5% ) for T . cruzi using a similar validation approach [50] . We found that only 11% ( 4/36 ) of the reactive peptides are shared among the three parasite strains ( Figure 1E ) , highlighting the problem with identifying high-sensitivity antigens for the serodiagnosis of Chagas disease due to the high degree of T . cruzi polymorphism . One of the conserved epitopes identified in this study , peptide C6_30_cons , has proven to be a new conserved T . cruzi antigen with a potential application in Chagas disease serodiagnosis ( Figures 2A , 6A , S2 , and S3 ) . Together , the three polymorphic epitopes were able to discriminate among infections caused by the three different T . cruzi strains included in this study and , thus , have the potential to be used for the serotyping of infections caused by this parasite . ELISA experiments using human sera confirmed the predictive reactivity of A6_30_col and B2_30_y ( Figure 6 ) . A6_30_col was able to identify 100% of the patients infected with TcI . As expected , the serum samples obtained from Brazilian patients known to be infected with TcII were reactive only with the C6_30 conserved and B2_30_y peptides . These results confirm the potential use of this peptide set for Chagas disease serotyping . The peptide A6_30_col and B9_30_cl are derived from RNA binding proteins and RNA polymerase III , respectively ( Supplementary Table S3 ) . Both are predicted to have an intracellular localization . Indeed , humoral response against intracellular antigens is quite common in trypanosomatids as shown by the work described by da Rocha et al . , 2002 [52] that performed immunoscreening of an amastigote cDNA library using sera from chagasic patients . About 70% of the amastigote antigens identified in this study is derived from intracellular parasite proteins . Similar to Leishmania infection , it is postulated that during T . cruzi infection a proportion of trypomastigotes/amastigotes cells are destroyed , thus releasing substantial amounts of multicomponent complexes containing intracellular antigens [53] . This reactivity could be the result of high abundance of these antigens as circulating complexes during the parasite infection due to high and constant expression of nuclear and house-keeping genes; higher stability due to formation of nucleoprotein particles more resistant to degradation; and their increased capacity to be processed by antigen-presenting cells because multicomponent particles are taken into the cell more efficiently than soluble antigens [54] . It is worth noting that the conserved and polymorphic epitopes identified in this study encompass repetitive regions . Interestingly , two of these peptides have proline-rich regions ( Figure 4 ) that may be involved in protein-protein interactions in prokaryotes [55] and eukaryotes [56] . It has been demonstrated that the overall immunogenicity of proteins harboring tandem repeats is increased , as is the antigenicity of epitopes contained within repetitive units [52] , [57] . Therefore , one expects that repeats receive a high B-cell epitope prediction score . Furthermore , the polymorphic epitopes containing repeats were top ranked for an additional reason: our polymorphic scale applied to the CL Brener pair of alleles attributes the highest score to a gap position in the alignment , a situation always present when the contraction or expansion of a repetitive region occurs in one sequence but not in another . This criterion was used because it is well known that repetitive sequences evolve faster than other regions of the genome [58] , hence it is expect that they display a high level of polymorphism among distinct parasite strains . Additionally , because it is known that the number of repetitive antigenic motifs may affect antibody binding affinity [59] , we hypothesized that polymorphic repetitive epitopes would be differentially recognized by the sera of animals and human infected with distinct parasite DTUs , an assumption that was reinforced by our results . The cross-reactivity of the epitopes among the sera from mice infected with distinct parasite strains is in agreement with an origin hypothesis of the different T . cruzi DTUs . In two-way comparisons , the CL Brener and Y strains , the two more phylogenetically related strains , shared a higher number of epitopes ( 4 ) compared to Y and Colombiana ( 1 ) , whereas CL Brener and Colombiana did not share any epitope ( Figure 1E ) . Interestingly , despite the fact that all of the peptides are derived from the CL Brener genome , a smaller number of epitopes were identified in this strain compared with Y and Colombiana ( Figure 1E ) . We speculate that the co-expression of alleles that encode the polymorphic epitopes in CL Brener may affect the titer of the antibody and/or its affinity for the variant epitopes . For example , the pattern of expression and the reactivity of the polymorphic peptides A6_30_col and B2_30_y ( Figures 4B , 4C , 5B , and 5C ) suggest that the co-expression of polymorphic epitopes in CL Brener could induce low-affinity antibodies . A similar phenomenon has been described for the polymorphic T . cruzi trans-sialidase ( TS ) multigene family , whereby TS displays a network of B-cell cross-reactive and polymorphic epitopes that delays the generation of high-affinity neutralizing antibodies and hamper an effective elicitation of a humoral response against these proteins [60] . Whether this is a more general adaptive immune evasion strategy that affects antibody affinity maturation , particularly in the case of hybrid strains , remains to be investigated . Altogether , the results demonstrated that peptide C6_30_cons is a new T . cruzi antigen conserved in the majority of DTUs of this parasite . Using this peptide , in association with other T . cruzi antigens , may improve the serodiagnosis of Chagas disease . The three polymorphic epitopes identified were able to discriminate among infections caused by the three different T . cruzi strains included in this study and , thus , have the potential to be used for the serotyping of infections caused by this parasite . This is the first study on the genomic scale to identify DTU-specific antigens . The genome sequencing of other T . cruzi strains will help identify new strain-specific and conserved epitopes and increase the number of antigen candidates for Chagas disease serodiagnosis and serotyping . The development of a robust panel of strain-specific epitopes may allow large-scale epidemiological studies aimed at correlating the infective strain with the variability in clinical outcomes observed in chagasic patients .
|
Serological tests are preferentially used for the diagnosis of Chagas disease during the chronic phase because of the low parasitemia and high anti-T . cruzi antibody titers . However , contradictory or inconclusive results , mainly related to the characteristics of the antigens used , are often observed . Additionally , the factors influencing variation in the clinical forms of Chagas disease have not been elucidated , although it is likely that host and parasite genetics are involved . Several studies attempting to correlate the parasite strain with the clinical forms have used hemoculture and/or PCR-based genotyping . However , both techniques have limitations . Hemoculture requires the isolation of parasites from patient blood and the growth of these parasites in animals or in vitro culture , thereby possibly selecting certain subpopulations . Moreover , the level of parasitemia in the chronic phase is very low , hindering the detection of parasites . Additionally , direct genotyping of parasites from infected tissues is an invasive procedure that requires medical care and hinders studies with a large number of samples . The goal of this work was to identify conserved and polymorphic linear B-cell epitopes of T . cruzi on a genome-wide scale for use in the serodiagnosis and serotyping of Chagas disease using ELISA . Development of a serotyping method based on the detection of strain-specific antibodies may help to understand the relationship between the infecting strain and disease evolution .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2013
|
Identification of Strain-Specific B-cell Epitopes in Trypanosoma cruzi Using Genome-Scale Epitope Prediction and High-Throughput Immunoscreening with Peptide Arrays
|
The gastrointestinal microbiota influences immune function throughout the body . The gut-lung axis refers to the concept that alterations of gut commensal microorganisms can have a distant effect on immune function in the lung . Overgrowth of intestinal Candida albicans has been previously observed to exacerbate allergic airways disease in mice , but whether subtler changes in intestinal fungal microbiota can affect allergic airways disease is less clear . In this study we have investigated the effects of the population expansion of commensal fungus Wallemia mellicola without overgrowth of the total fungal community . Wallemia spp . are commonly found as a minor component of the commensal gastrointestinal mycobiota in both humans and mice . Mice with an unaltered gut microbiota community resist population expansion when gavaged with W . mellicola; however , transient antibiotic depletion of gut microbiota creates a window of opportunity for expansion of W . mellicola following delivery of live spores to the gastrointestinal tract . This phenomenon is not universal as other commensal fungi ( Aspergillus amstelodami , Epicoccum nigrum ) do not expand when delivered to mice with antibiotic-depleted microbiota . Mice with Wallemia-expanded gut mycobiota experienced altered pulmonary immune responses to inhaled aeroallergens . Specifically , after induction of allergic airways disease with intratracheal house dust mite ( HDM ) antigen , mice demonstrated enhanced eosinophilic airway infiltration , airway hyperresponsiveness ( AHR ) to methacholine challenge , goblet cell hyperplasia , elevated bronchoalveolar lavage IL-5 , and enhanced serum HDM IgG1 . This phenomenon occurred with no detectable Wallemia in the lung . Targeted amplicon sequencing analysis of the gastrointestinal mycobiota revealed that expansion of W . mellicola in the gut was associated with additional alterations of bacterial and fungal commensal communities . We therefore colonized fungus-free Altered Schaedler Flora ( ASF ) mice with W . mellicola . ASF mice colonized with W . mellicola experienced enhanced severity of allergic airways disease compared to fungus-free control ASF mice without changes in bacterial community composition .
The gut microbiome is a dynamic ecosystem that profoundly influences immune function throughout the body [1–3] . Commensal microorganisms are recognized by the host immune system and can alter systemic immune response or produce bioactive metabolites which are absorbed into the bloodstream and have pharmacological effect on distant organ systems [4–7] . The commensal microbial composition of the gut can therefore have a distant effect on immune function in the lung and other organ systems; this is the concept of the gut-lung axis . In addition to bacteria , the healthy gastrointestinal system contains a community of commensal fungi that live in the gut and continually interact with the gastrointestinal mucosa [8–10] . As many commensal fungi require fastidious culture conditions , non-culture-based assays are essential to comprehensively profile intestinal fungal communities . Like 16S sequencing of bacterial DNA , targeted amplicon sequencing of the first internally transcribed spacer region ( ITS1 ) of the fungal ribosomal RNA genome can be used to generate a profile of the fungal organisms in a sample and their relative abundance [11] . Although there are several sample processing and analysis considerations distinct from 16S analysis , ITS1 sequencing has been used successfully to profile gastrointestinal fungal communities in a both mice and humans [9 , 11] . ITS1 sequences have high variability across the phylogenetic tree and can generally be used to classify organisms to the genus and often species level . Studies using ITS1 sequencing to profile the human gastrointestinal mycobiota generally report a diverse community containing more than 50 unique genera with substantial variation across individuals [8] . Fungi are particularly important in asthma . Inhaled fungi are a well-described trigger for asthma , and patients with fungal sensitization have increased asthma incidence and more severe and fatal disease compared to patients sensitized to other allergens [12 , 13] . Interestingly , many commensal fungal species commonly found in the human gastrointestinal tract are triggers of allergic respiratory illness when inhaled such as Aspergillus , Cladosporium , Penicillium , and others [14] . We know that avoidance of airborne fungi can improve control of asthma in some patients , but it is uncertain whether variation in the commensal fungal species which reside in the gut and continually interact with the host immune system may also alter the severity of allergic airways disease . This is an emerging area of investigation , but tantalizingly , one recent study of human infants suggested that gut fungal dysbiosis may be more strongly associated with the development of allergic wheeze than bacterial dysbiosis [15] . Prior studies of the intestinal mycobiota have largely focused on Candida albicans and other Candida species . These are opportunistic pathogens that are recognized to be subject to intestinal overgrowth after exposure to oral antibacterials . In mouse models , intestinal overgrowth of Candida can be induced after treatment with the antibiotic cefoperazone and exposure to a bolus of live yeasts [16] . Mice with Candida overgrowth have exacerbated allergic airways disease , which has been suggested to occur via fungal secretion of prostaglandins that are absorbed into systemic circulation [4] . Building on these studies , we previously investigated whether suppressing natural commensal fungal populations with anti-fungal drugs could suppress allergic airway disease . We found , instead , that antifungal treatment surprisingly exacerbated allergic airways disease [17] . Although anti-fungal therapy depleted native Candida populations , ITS1 sequencing analysis of their gastrointestinal mycobiota suggested that other , relatively drug-resistant fungal organisms increased in abundance; specifically , Aspergillus amsteoldami , Wallemia mellicola , and Epicoccum nigrum . Further , when administered together as a cocktail by oral gavage , these three fungi exacerbated allergic airways disease . These studies have led us now to investigate the idea that relatively modest changes in naturally occurring non-Candida intestinal fungi may influence pulmonary immune responses . In this manuscript , our experimental results have led us to focus on Wallemia mellicola , a common environmental fungus . W . mellicola is a ubiquitous spore forming filamentous basidiomycete that is a common component of house dust and an agent of food spoilage [18–21] ( S1A and S1B Fig ) . W . mellicola is slow growing and may therefore be less commonly detected by culture than other commensal fungi such as Candida and Aspergillus . Wallemia spp . are highly xerotolerant and have been reported to secrete several toxins when grown in culture such as walleminol , walleminone , and wallimidione [22–24] . The metabolic products generated by Wallemia residing in the mammalian intestine are unknown . W . mellicola does not generally act as a human pathogen , but there are some reported associations between Wallemia exposure and lung disease . Specifically , asthmatic patients have a high incidence of immune sensitization to Wallemia , and Wallemia was identified as one of a handful of environmental fungi associated with increased risk of asthma in individuals living in water damaged homes [25 , 26] .
We have previously observed that a fluconazole-induced intestinal fungal dysbiosis state enhances the severity of allergic airways disease in mice [17] . Fluconazole therapy has complex and surprising effects on commensal gut fungal communities . Although fluconazole moderately depletes the overall burden of commensal fungi including Candida spp . , fungi that are relatively resistant to fluconazole expand in population ( Fig 1A ) . Notably , this is not simply an increased relative abundance due to mismatched decline in individual species abundance compared to total fungal burden . Rather , the absolute quantity of selected gut fungal species can increase during fluconazole therapy . Our prior observation that fluconazole depletion of gut mycobiota exacerbates the severity of allergic airways disease suggests that previously-characterized effects of intestinal Candida overgrowth on allergic airways disease may not be restricted to Candida [17] . To investigate further , we first sought to define conditions that generate a fungal dysbiosis in the gut by means other than inducing Candida overgrowth or repeated oral gavage with fungi . We examined conditions required for enhancing intestinal colonization with several non-Candida commensal fungal species that are natively found in the intestinal microbiota of specific pathogen free mice ( SPF ) at our animal facility: Wallemia mellicola , Aspergillus amstelodami , and Epicoccum nigrum . These were selected because we have previously observed that these species expand in relative population abundance in our fluconazole treated mice who develop exacerbated sensitivity to the house dust mite model of asthma [17] . We initially attempted simply a gavage of high dose of cultured live fungi from each species into the mouse gastrointestinal tract . This did not result in sustained expansion of the population of any fungus . We hypothesized that resistance to expansion of the fungal populations was due to competition from other commensal gastrointestinal microbes . Previous authors have shown that mice with an intact commensal bacterial community resist colonization with Candida albicans , but that antibiotic depletion of mouse commensal bacteria renders them vulnerable to C . albicans overgrowth after exposure [27] . We therefore performed 7 days of treatment of mice with cefoperazone followed by one-time gavage of each fungus ( Fig 1B ) . Cefoperazone treatment had a devastating effect on bacterial communities , depleting total commensal bacterial burden of the gut by nearly 10 , 000x fold , but gut bacterial abundance recovered to previous levels within 8 days of cessation of antibiotics ( Fig 1C ) . A single gavage of W . mellicola into a mouse with antibiotic depleted bacterial microbiota resulted in sustained and substantial increase in the population of Wallemia above baseline ( Fig 1D ) . The Wallemia-expanded mycobiota persisted after discontinuation of antibacterials and without any further W . mellicola gavage to support the population . W . mellicola is present as a minor component of the commensal mycobiota in our mice at baseline , but mice treated with cefoperazone and gavaged with sterile water ( Fig 1D ) did not experience expansion of this fungus suggesting that both antibacterial depletion and exposure to a bolus of W . mellicola are necessary to generate a Wallemia-expanded dysbiosis state . To be certain that W . mellicola can survive and grow in the mouse intestines , we further colonized germ-free mice with W . mellicola and observed that we could culture organisms from the stool after 10 days ( S1C Fig ) . Antibiotic treatment was not similarly sufficient to allow for enhanced colonization with Epicoccum nigrum or Aspergillus amstelodami , suggesting that antibiotic depletion of gut bacteria is not universally able to facilitate expansion of a commensal fungal species ( Fig 1D ) . These organisms may not directly compete with bacteria for their gut ecological niche , or they may compete with bacteria that are not affected by cefoperazone therapy . We next explored the characteristics of the Wallemia-expanded mycobiota . W . mellicola colonization was predominantly in the cecum and colon and restricted to the gastrointestinal tract with no W . mellicola detected in extra-gastrointestinal organ systems ( Fig 2A ) . Notably , we did not detect W . mellicola by rtPCR or culture in mouse lungs . Total gastrointestinal fungal abundance in mice with Wallemia-expanded microbiota remained similar to untreated mice , but the population of fungi had shifted such that W . mellicola had markedly increased in relative abundance ( Fig 2B and 2C ) . This contrasts with the Candida albicans overgrowth state where the total fungal burden in the gastrointestinal tract increases by orders of magnitude after mice are subjected to a similar protocol of cefoperazone depletion of gut bacteria followed by C . albicans gavage ( Fig 2C ) . Finally , there was no evidence to suggest that W . mellicola was behaving as an infectious pathogen in these mice . Specifically , mice with Wallemia-expanded mycobiota showed no weight loss , behavioral changes , or stool changes throughout the experiments , and there was no histological evidence of colonic inflammation ( Fig 2D and 2E ) . Together , these data suggest that stable W . mellicola colonization is best thought of as a model of altered intestinal fungal community rather than gastrointestinal fungal overgrowth . We next examined whether W . mellicola and the other commensal fungi studied in these experiments are found in the human gastrointestinal tract . Prior studies using ITS1 sequencing have detected sequences from Wallemia spp . , Aspergillus spp . , and Epicoccum spp . in human gastrointestinal samples , but species-specific PCR based assay of human gastrointestinal samples for the three fungal species discussed in this manuscript has not previously been described [28–30] . We extracted DNA from stool specimens from 9 healthy human subjects and performed rtPCR using species-specific primers to directly assess for DNA from each of the three relevant fungal species . We detected Wallemia mellicola in 3 of 9 human samples and Aspergillus amstelodami in 7 of 9 human samples ( Fig 3A ) . We did not detect Epicoccum nigrum DNA in any human samples in our cohort , as no amplification was observed in any sample tested by rtPCR with Epicoccum nigrum specific primers . These results are consistent with prior studies showing that the composition of the human gut mycobiota varies across individuals and suggest that Wallemia mellicola and Aspergillus amstelodami may be capable of residing in the human gastrointestinal tract . We next examined the three human subjects who had W . mellicola DNA detectable in their stool . The total amount of W . mellicola DNA per stool weight was similar between Wallemia-colonized humans and SPF mice in our facility who are natively colonized with W . mellicola but less than mice who underwent the Wallemia-expanded colonization protocol ( Fig 3B ) . This is not surprising because none of these healthy individuals carried a diagnosis of asthma or had recent antibiotic use , and we observed that antibiotic therapy was necessary to generate the Wallemia expansion dysbiosis in our mouse experiments . Human subjects found to be colonized with Wallemia do not have enhanced total fungal burden compared to non-Wallemia-colonized subjects ( Fig 3C ) . Having established a mouse model of sustained W . mellicola colonization of the gut , we next sought to determine whether mice with Wallemia-expanded intestinal dysbiosis have an altered immune response to inhaled aeroallergens . We generated mice with a Wallemia-expanded mycobiota as described above by gavage of W . mellicola conidia into antibiotic-treated animals . Control mice were housed and treated identically to the Wallemia-expanded mice including the antibiotic treatment , but they received sterile water gavage rather than W . mellicola conidia gavage . We induced allergic airways disease in both groups by weekly intratracheal house dust mite ( HDM ) sensitization . Mice with expanded intestinal population of W . mellicola had increased severity of allergic airways disease compared to control by multiple measures . Wallemia-expanded mice demonstrated markedly greater bronchoalveolar lavage ( BAL ) cellularity driven primarily by increased alveolar eosinophils ( Fig 4A and 4B ) along with enhanced airways hyperresponsiveness ( AHR ) to methacholine challenge ( Fig 4C ) . Histological analysis of the lungs showed enhanced goblet cell hyperplasia in Wallemia-expanded mice compared to controls ( Fig 4D and 4E ) , and these mice also had higher BAL levels of IL-5 and serum IgG1 to HDM detectable at the end of the experiment compared to controls by ELISA ( Fig 4F and 4G ) . To determine whether this effect on allergic airways disease might possibly be due to the initial bolus of Wallemia conidia , rather than the sustained colonization , we compared live W . mellicola gavage to gavage with heat-killed organisms . Heat-killed organisms did not influence allergic airways disease ( S2A Fig ) . Similarly , we observed that an initial gavage with live W . mellicola did not influence disease if no pretreatment with antibiotics was provided to allow for sustained colonization ( S2B Fig ) . Finally , to be certain that established colonization was essential , we delayed beginning the HDM sensitization for a week after the live W . mellicola gavage and found that the presence of Wallemia still exacerbated disease ( S2C Fig ) . Together the data support the conclusion that enhanced colonization with live W . mellicola in the intestines exacerbates susceptibility to HDM allergic airways disease . To begin to understand the mechanism by which expansion of gastrointestinal W . mellicola may have altered pulmonary immune response to HDM , we extracted the mediastinal lymph node from both groups and cultured lymphocytes in vitro . Five days after in vitro restimulation with HDM , lymphocytes extracted from mice with a Wallemia-expanded mycobiota had increased percentage of CD4+ T-cells positive for Th2 cytokine IL-13 by intracellular staining and increased supernatant IL-13 concentration ( Fig 4H–4J ) . We did not observe any differences in IFNγ or IL-17 between the two groups by either intracellular staining or supernatant cytokine levels , suggesting that the Wallemia-expanded mycobiota has little effect on Th1 and Th17 response in this setting ( S3 Fig ) . Finally , to determine whether this augmented pulmonary immune response could be due to stimulation by direct migration of W . mellicola to the lungs , we examined for the presence of W . mellicola in the lungs by multiple methods . We plated homogenized lung from Wallemia-expanded mice on antibacterial treated SDB agar plates and observed no growth of W . mellicola . We also performed PCR of whole lung homogenate and no amplification was observed in any sample tested by rtPCR using W . mellicola specific primers . Together , these results suggest that an intestinal dysbiosis state characterized by enhanced presence of W . mellicola has a distant effect on pulmonary immune response characterized by increased eosinophilic airway inflammation , goblet cell hyperplasia , and enhanced secretion of IL-13 by mediastinal lymphocytes in response to HDM . We have shown population expansion of intestinal W . mellicola enhances the severity of allergic airways disease in response to HDM allergen challenge . However , these experiments are insufficient to establish that intestinal Wallemia itself alters pulmonary immune response . Rather than having a direct effect , it is possible that Wallemia population expansion may alter or suppress commensal bacteria or other fungi which are themselves responsible for altering asthma severity . Notably , Wallemia species have been reported to secrete antibacterial compounds when grown in culture , so we hypothesized that expanded growth of Wallemia might alter bacterial community composition [31] . To determine whether expanded W . mellicola colonization altered bacterial and fungal communities , we analyzed fecal bacterial and fungal communities in Wallemia-expanded mice via 16S and ITS1 rDNA sequencing respectively . To control for potential cage-effects [32] , each group ( n = 8 ) was spread across 4 independent cages . Principal coordinates analysis ( PCoA ) and pair-wise differential abundance analysis with LEfSe [33] suggested that W . mellicola expansion altered bacterial communities ( Fig 5A , 5B , S4A and S4B Fig ) . Further , we observed changes to fungal communities in Wallemia-expanded mice in addition to the expected expanded W . mellicola population ( Fig 5C , 5D , S4C and S4D Fig ) . However , enhanced colonization with W . mellicola did not affect levels of E . nigrum or A . amstelodami ( Fig 5E ) . Together , the data suggest that enhanced growth of Wallemia mellicola in the gut alters intestinal bacterial and fungal populations , and this makes it difficult to conclude whether Wallemia mellicola itself is sufficient to alter allergic immune responses in the lung . To investigate whether the presence of W . mellicola in the gastrointestinal microbiota itself is sufficient to alter airway response to allergens , we employed a gnotobiotic mouse model that offers more precise control of intestinal microflora . Altered Schaedler Flora ( ASF ) mice are gnotobiotic animals with a stable microbiota consisting of eight defined bacterial species . Importantly for this study , they are fungus-free . Being colonized with bacteria , ASF mice are healthier than germ-free mice and have more mature immune systems [34–36] . While ASF mice and germ-free mice are initially “fungal-free” , gastric gavage with live W . mellicola is sufficient to establish fungal colonization ( Fig 6A and 6B ) . W . mellicola grows similarly in both types of animals , although the total fungal burden remains substantially lower than SPF mice from our facility . Interestingly , W . mellicola colonization of ASF animals did not require prior antibiotic-mediated depletion of bacteria suggesting that the ASF bacteria do not substantially compete for the niche required by this fungus . We further observed that colonization of ASF animals with W . mellicola did not alter ASF bacterial microbiota . ASF mice have fecal bacterial levels similar to SPF mice , and colonization of ASF mice with W . mellicola did not grossly alter the total bacterial burden ( Fig 6C ) . Upon measuring fecal levels of each of the 8 constituent ASF bacteria , we did not observe any Wallemia-induced quantitative changes in the population ( Fig 6D , S5A and S5B Fig ) , although these results do not exclude the possibility that W . mellicola may alter metabolic products produced by the ASF bacterial community . Together the data suggest that these animals are a strong model for evaluating the specific ability of intestinal W . mellicola to influence allergic airways disease . We next performed intratracheal HDM sensitization on these mice to induce allergic airways disease . We compared ASF mice to ASF mice colonized with W . mellicola . Like conventional mice with an expanded population of W . mellicola , we observed that colonizing ASF with W . mellicola increased the severity of HDM-induced allergic airways disease . Interestingly ASF mice demonstrated a mildly blunted response to HDM at baseline compared to SPF mice , but the pattern of disease was like that observed in SPF mice with antibiotic-associated expansion of Wallemia , including increased eosinophilic airway infiltration , increased histological inflammation with hyperplasia of mucous producing goblet cells , elevated BAL IL-5 , and increased serum IgG1 to HDM ( Fig 7 ) . The data suggest that intestinal W . mellicola is sufficient to exacerbate allergic airways disease since the effect occurs without substantial alteration to other commensal bacteria .
We have shown that altered composition of the gastrointestinal mycobiota enhances the severity of allergic airways disease with enhanced eosinophilic airway inflammation and increased IL-13 production by mediastinal lymphocytes in response to HDM allergen stimulation . Interestingly , these effects are not due to a fungal overgrowth state where bloom of a single organism results in exponential population expansion of the total gastrointestinal fungal burden . Rather , the W . mellicola dysbiosis described herein is a shift in the composition of the commensal fungal community that occurs without substantial increase in the total fungal burden yet still produces a significant change in pulmonary immune response to inhaled allergens . The term “dysbiosis” is generally used to describe altered gut microbial ecosystem that results in negative host effects but is not an infectious state . We believe that the dysbiosis state described in this manuscript is not unique to Wallemia mellicola . Rather Wallemia mellicola dysbiosis may just be one representative example of a gut microbial pattern that alters pulmonary and systemic immune response . Other alterations of the gastrointestinal mycobiota community characterized by expansion of different fungal species may have distinctive beneficial or harmful effects on respiratory immune function . The mechanism by which gastrointestinal W . mellicola population expansion alters pulmonary immune function is unknown , but there are several possibilities . W . mellicola may produce a toxin or metabolite that is absorbed , circulates systemically , then acts like a drug to modify the pulmonary immune response . This phenomenon has been observed with several other gut commensal microorganisms such as Clostridium orbiscindens which produces a small molecule metabolite ( desaminotyrosine ) that is systemically absorbed and alters pulmonary interferon response to influenza infection and prostaglandin E2 produced by intestinal Candida as discussed earlier [4 , 5] . Interestingly , Wallemia species have been described to secrete a variety of mycotoxins such as walleminol , walleminone , and wallimidione when cultured in vitro [22–24] . Mycotoxins produced by other similar environmental fungal species have been shown to promote inflammatory cytokine secretion by lung alveolar macrophages [37] . Fungal mycotoxin production is generally influenced by the environmental conditions , and further study is needed to determine whether Wallemia produces toxins or other bioactive metabolites during growth in the mammalian intestine , whether these are absorbed into systemic circulation , and whether they affect host immune cells . It is also possible that Wallemia indirectly triggers production of different metabolites by the bacterial microbiota that then affects immune responses to challenge . Future studies will need to be designed to address this possibility . Alternatively , W . mellicola may be recognized by the gastrointestinal host immune system and result in differential immune cell trafficking to the lungs . Gut commensal fungi have previously been shown to promote trafficking of different immune cell populations to non-GI organ systems . For example , migration of RALDH+ dendritic cells to peripheral lymph nodes in young mice is enhanced specifically by the presence of certain commensal gut fungal species [38] . W . mellicola is not generally considered to be a human pathogen , so there has been little prior study of immune response to this organism . Wallemia is known to elicit serum IgE and IgG responses [25 , 39] , but other innate and adaptive immune responses are not well characterized . Further study is needed to understand how W . mellicola is recognized by intestinal immune and epithelial cells and the consequences . An important observation in this study is that oral antibiotic therapy places mice at risk for expansion of the dysbiosis-associated fungus Wallemia mellicola , but that mice with intact microbiota resist W . mellicola expansion after exposure . Our W . mellicola dysbiosis state in mice was established under conditions like those that a human asthma patient might experience . Wallemia are common environmental and food spoilage fungi , and therefore it is plausible that individuals may be exposed to live Wallemia conidia in their food or environment during a course of antibiotic therapy . Cefoperazone is a third-generation cephalosporin , and antibacterial medications in this class are widely used in patients with asthma and other respiratory diseases . Although antibacterials are not indicated for an uncomplicated asthma exacerbation , patients with severe asthma nonetheless receive frequent courses of broad spectrum antibacterial medications for a variety of indications throughout their lifetime [40 , 41] . Furthermore , frequent courses of antibacterials , particularly early in life , have been associated with an increased incidence asthma [42 , 43] . The concept that an intact microbiota resists colonization by pathogens such as Clostridium difficile has been previously established [44] , and we now describe that a non-pathogenic commensal fungus can expand in the face of an antibiotic depleted microbiota and enhance the severity of allergic airways disease in mice . Intestinal fungal dysbiosis might therefore be an unrecognized but potentially important risk of each course of antibiotic therapy in patients with asthma and other respiratory disease . Further studies are needed to determine whether a phenomenon like the Wallemia dysbiosis state described in this manuscript can occur in humans during routine broad spectrum antibacterial therapy and more generally to what extent that gut mycobiota changes in humans alters pulmonary and systemic immune function .
All experiments involving research animals were performed in accordance with the recommendations outlined in the Guide for the Care and Use of Laboratory Animals . All research animal protocols were approved by the institutional animal use and care committee at Cedars-Sinai Medical Center ( IACUC #6670 and #5160 , PHS assurance number A3714-01 ) . All studies involving humans and human samples were approved by the Cedars-Sinai Medical Center Institutional Review Board ( IRB #0003358 , Federalwide Assurance number 00000468 ) . All human subjects were adults age >18 who provided written informed consent . All specimens were assigned an anonymized sample ID with no connection to patient identifiable information . Specimens were collected by the Cedars-Sinai MIRIAD IBD Biobank . 7-8-week-old C57BL/6 female mice were purchased from Jackson Laboratory and housed in specific pathogen free conditions at the Cedars-Sinai animal facility . Germ-free mice C57BL/6 mice were obtained from Taconic Farms then housed and bred in microbially sterile flexible film isolators . A separate colony of altered Schaedler flora ( ASF ) mice was generated by gavage of live cultures of the 8 ASF bacterial species ( Taconic Farms ) into germ-free mice [35] . These mice were subsequently housed , bred , and raised in a separate flexible film isolator that was designated exclusively for maintenance of the ASF mouse colony . PCR based assays were performed to confirm stable presence of all 8 ASF bacterial species in subsequent generations of mice . ASF experiments were performed in age matched cohorts of 7-8-week-old male mice who had been born in a flexible film isolator to mothers colonized with ASF bacteria . Quality control testing including microbial culture and PCR was regularly performed on both the germ-free and ASF colonies to ensure that no contaminating microorganisms were introduced into either colony . Notably , ASF animals were verified to be fungus-free by FungiQuant rtPCR assays of both stool and environmental samples [45] . All animal experimental protocols were approved by the Institutional Animal Use and Care Committee ( IACUC ) at the Cedars-Sinai Medical Center ( IACUC #5160 and #6670 ) . Fungal cultures of Wallemia mellicola ( ATCC 42694 ) , Aspergillus amstelodami ( ATCC 46362 ) , Epicoccum nigrum ( ATCC 42773 ) were obtained from American Type Culture Collection ( ATCC ) and grown on Sabouraud dextrose agar for 7–14 days at 23°C as previously described [17] . Note that due to a recent taxonomic revision , the Wallemia strain examined in this study ( ATCC 42694 ) was previously identified as Wallemia sebi but has been reclassified as Wallemia mellicola [19] . As documented in this manuscript , Wallemia mellicola is a naturally-occurring commensal microbe in mice and humans . Fungal conidia suspensions were generated by flooding a Sabouraud dextrose agar plate with mature fungal culture growth with 10 mL sterile water , gently washing by pipet to dislodge conidia , then passing the spore suspension through a 40 μm filter to exclude hyphal fragments . Conidia were centrifuged at 1600 rpm x4 minutes , resuspended in 1 mL sterile water , counted using hemocytometer , and the volume of water was adjusted to produce a 5 x 107 conidia /mL suspension for gavage . For experiments involving heat-killed Wallemia , conidia were prepared as described and exposed to 95° C for 10 minutes then plated on Sabouraud dextrose agar to confirm non-viability . Medicated drinking water was administered in selected experiments as follows: Fluconazole powder ( Sigma-Aldrich , PHR1160 ) was dissolved in deionized water at a concentration of 0 . 5 mg/mL or cefoperazone sodium salt ( Alfa Aesar , J65185 ) was dissolved in deionized water at a concentration of 0 . 5 mg/mL . The medicated water was provided as the exclusive source of drinking water for mice over the duration of antibiotic therapy , and mice were permitted to drink the medicated water ad libitum . The water was protected from light exposure by foil , and the medicated water was exchanged with freshly mixed solution every 3–5 days . The fungal expanded colonization protocol described herein was adapted from experiments originally described by Noverr and Huffnagle [16] . SPF mice were treated with cefoperazone drinking water for 7 days to deplete intestinal bacteria . Mice subsequently received gavage of 5 x 106 live conidia from a single species ( Wallemia mellicola , Aspergillus amstelodami , or Epicoccum nigrum ) or 5 x 106 live yeast ( C . albicans ) suspended in 100 μL deionized water . Antibiotic drinking water was discontinued 4 hours after gavage . Stool collection was performed at several timepoints ( Days #1 , 7 , 14 , 21 ) and fungal DNA was extracted from stool as later described . Stool pellets were examined for consistency and for the presence of blood . Mice ( n = 5 per group ) first underwent the colonization protocol consisting of 7 days of cefoperazone 0 . 5 mg/mL followed by a single gastric gavage of either 5 x 106 Wallemia spores suspended in 100 μL water ( Wallemia colonization expansion group ) or 100 μL sterile water ( control group ) . Four hours after gavage , the cefoperazone water was discontinued in both groups and all mice were supplied with non-medicated drinking water for remainder of the experiment . Mice subsequently underwent house dust mite sensitization once every 7 days for a total of three treatments . Dermatophagoides pteronyssinus house dust mite suspension was purchased from Greer Labs and all mice in each experiment received HDM from the same lot . HDM sensitization was performed by first anesthetizing mice with isoflurane and then direct intratracheal administration of 100 μg HDM suspended in 50 μL of sterile saline . Mice were euthanized 48 hours after the final HDM treatment . Bronchoalveolar lavage ( BAL ) was performed immediately after euthanizing by cannulating the trachea with a 20 gauge catheter , instilling 1 mL sterile saline into the trachea with visual confirmation of inflation of the bilateral lungs , and aspirating the return lavage fluid . This was repeated for a total of 4 x 1 mL lavage for each mouse with >80% return ( 3 . 2 mL ) defined as an adequate BAL . Cell suspension was counted using hemocytometer , then BAL solution was centrifuged at 1600 rpm x 4 minutes with supernatant removed for ELISA analysis and cell pellet resuspended in 1 mL FACS buffer for flow cytometry staining . Blood was aspirated by direct cardiac puncture , allowed to clot , then centrifuged at 6000 rpm x6 minutes to extract serum . A single lung lobe was fixed in formalin for H&E and PAS staining . To assess for Wallemia presence in the lungs , in separate experiments otherwise performed identically but with no BAL and n = 3 mice per group , the entire bilateral lung was resected en-bloc , the lung was suspended in 500 μL sterile PBS , morsellized with sharp scissors using sterile technique , and split to be plated in Sabouraud Dextrose Agar ( SBD ) with ampicillin 50 μg/mL for culture or homogenized with 0 . 5 mm bead and then processed for DNA extraction as described below for Wallemia rtPCR analysis . To perform the in-vitro restimulation experiments , the mediastinal lymph node was harvested after the house dust mite sensitization experiments described above . A single cell suspension was generated by digestion with 200 u/mg type 4 collagenase ( Worthington ) and DNase 1 ( Sigma-Aldrich ) for 30 minutes at 37°C followed by passage through a 70 μm filter . Cells were resuspended in RPMI-1640 medium ( Cellgro ) supplemented with 2% FBS , beta-mercaptoethanol , and penicillin/streptomycin at a concentration of 2 . 5 x 106 cells/mL . Cells were distributed into 24 well tissue culture plates , stimulated with 25 μg HDM , then incubated x5 days at 37°C in a cell culture incubator with 5% carbon dioxide . After incubation , supernatants were collected for ELISA and cell surface staining for flow cytometry was performed as described below . Cells were then stimulated for 4 hours with 20 nM PMA ( phorbol 12-myristate 13-acetate , Sigma-Aldrich ) and 1 . 3 μM ionomycin ( Sigma-Aldrich ) in the presence of GolgiStop ( BD Biosciences ) , permeabilized , and then intracellular staining was performed as subsequently described . Airway hyperresponsiveness to methacholine was measured 48 hours after the final HDM challenge . To perform this measurement , Wallemia-expanded and antibiotic only control mice were assessed using a Buxco FinePointe resistance and compliance system ( Harvard Bioscience ) . Sodium pentobarbital ( 40 mg/kg i . p . ; Butler Co . ) was used to anesthetize mice before tracheostomy tube placement for mechanical ventilation . Within the sealed plethysmograph mouse chamber , transpulmonary pressure ( i . e . , Δ tracheal pressure -Δ mouse chamber pressure ) and inspiratory volume were continuously monitored online by an adjacent computer . Airway resistance index was measured and calculated via the manufacturer supplied software ( Buxco ) . After a baseline period in the Buxco apparatus , anesthetized and intubated mice received escalating doses ( 0 mg/mL , 2 . 5 mg/mL , 5 mg/mL , 10 mg/mL ) of methacholine suspended in PBS by nebulizer ( 20 μL volume ) over 30 seconds , and airway responsiveness to this bronchoconstrictor was measured over 3 minutes with an additional 1 minute for return to baseline before administration of next methacholine dose . At the end of the assessment of airway responsiveness each mouse was euthanized by exsanguination . Fluorophore-conjugated antibodies directed against each of the following molecules were used to perform cell surface staining of BAL samples: Siglec-F ( E50-2440 , PE , BD Bioscience ) , CD11c ( N418 , violetFluor 450 , Tonbo Bioscience ) , CD11b ( M1/70 , Alexa Fluor 700 , eBioscience ) , Ly-6G ( RB6-8C5 , PE-Cy7 , Tonbo Bioscience ) , CD3e ( 145-2C11 , APC , eBioscience ) CD19 ( APC , eBioscience ) . BAL CD3-CD19-Siglec-F+CD11c- cells were classified as eosinophils and CD3-CD19-Siglec-F+CD11c+ cells were classified as alveolar macrophages [46] . Antibodies directed against the following were used to stain mediastinal lymphocytes CD3e ( 145-2C11 , PerCP-Cy5 . 5 , eBioscience ) , and CD4 ( GK1 . 5 , FITC , BioLegend ) , and intracellular staining was performed after cell membrane permeabilization with saponins ( PermWash Buffer , BioLegend ) per manufacturer instructions using fluorophore-conjugated antibodies directed against IL-13 ( eBio13A , PE-Cy7 , eBioscience ) , IFNγ ( XMG1 . 2 , APC , eBioscience ) , and IL-17 ( eBio17B7 , PE , eBioscience ) . Sample data was acquired with a LSRII ( BD Biosciences ) and data was analyzed with FlowJo version 10 . 4 . 2 ( BD Biosciences ) . Supernatant and BAL cytokine measurements were performed using commercial IL-13 ( eBioscience ) , IL-17 ( BioLegend ) , IFNγ ( BioLegend ) , and IL-5 ( BioLegend ) ELISA kits . All commercial ELISA assays were performed according to the manufacturer’s supplied instructions . To measure serum HDM-specific IgG1 , a 96-well plate ( Corning Costar 3361 ) was first coated overnight at 4°C with 25 mg/ml HDM in PBS . After plate washing , serum samples and controls were added and incubated at 23°C for 2 hours followed by washing and incubation with a detection biotinylated anti-mouse IgG1 ( Biolegend , RMG1-1 ) at 1:750 dilution for 60 minutes . After washing , streptavidin-HRP ( Biolegend ) was added for 30 minutes . The plate was washed again , and the ELISA was developed with BD OptEIA TMB substrate set ( BD Biosciences ) . OD measurements were recorded at 450–570 nm using a CLARIOstar ( BMG Labtech ) plate reader . A reference standard for HDM-specific IgG1 ELISAs was generated by pooling serum samples from HDM-sensitized mice , and varied dilutions of this standard were used to generate a standard curve . ELISA results were scaled such that this reference sample has a relative abundance of HDM-specific IgG1 of 1 , and the same pooled reference standard was used for all HDM-specific IgG1 ELISAs reported in this manuscript . Whole mouse lung was perfused with phosphate buffered saline , resected en-bloc , and then fixed in 4% paraformaldehyde . Samples were embedded in paraffin , sectioned at 4 μm , and stained with hematoxylin and eosin or periodic acid-Schiff ( PAS ) . Histological scoring was performed as previously described [47 , 48] . Briefly , five airways were randomly selected on each PAS stained specimen slide: one primary airway , two separate secondary conducting airways , and two tertiary conducting airways . In each airway , one hundred sequential airway epithelial cells were identified and the number of PAS+ goblet cells in this segment were counted and divided by the total number of epithelial cells to generate the airway goblet cell index . The goblet cell index score for a specimen was the mean goblet cell index of the five examined airways . Prior to slide review and assignment of scores , all histology slides were assigned a random specimen ID such that the reviewer scoring each slide was blinded to the experimental group and to all specimen-associated experimental data . The blinding was broken only after the histological scoring was completed for all slides . Human subjects were recruited after approval by Cedars-Sinai Medical Center Institutional Review Board ( IRB #0003358 ) . All volunteers self-reported as healthy with no chronic gastrointestinal or respiratory conditions and no acute illness at the time of study participation . A stool sample was collected from each volunteer and immediately frozen at −80°C . Samples were deidentified for this study by assigning a unique specimen ID number that was not connected to any subject demographic data . An approximately 50 mg portion of each frozen sample was processed for DNA extraction and rtPCR analysis using the same protocols as described for mouse stool specimens . DNA for rtPCR or fungal and bacterial sequencing was isolated from 1–2 fecal pellets following lyticase treatment , bead beating , and processing using QIAmp DNA mini kit ( Qiagen ) as previously described using a protocol optimized to lyse fungal cell walls for recovery of fungal DNA [11] . DNA extraction from organ tissue was performed by first homogenization tissue with a 0 . 5 mm steel bead in TissueLyser II ( Qiagen ) at frequency 30s−1 x4 . 5 minutes then proceeding per DNA extraction protocol used for stool samples . A primer-probe based detection system with iTaq Universal Supermix ( BioRad ) was used for FungiQuant and all species specific fungal rtPCR assays while the 16S , ASF , and pan-Candida rtPCR was performed using a SYBR Green supermix ( BioRad ) based assay . TaqMan rtPCR assays were run on an Eppendorf Mastercycler realplex2 machine using the following parameters: initial denaturation step of 95°C for 2 minutes followed by 45 cycles of 95°C x15s then 60°C x1 minutes . The SYBR green based rtPCR assay was run on an Eppendorf Mastercycler realplex2 machine using the following parameters: initial denaturation step of 95°C for 10 minutes followed by 40 cycles of 95°C x15s then 60°C x30s then 72°C x32s . The following primer pair and probe sequences were used for rtPCR assays: Wallemia mellicola F: GGC TTA GTG AAT CCT TCG GAG , W . mellicola R: GTT TAC CCA ACT TTG CAG TCC A , W . mellicola probe: 5’- ( FAM ) TGT GCC GTT ( ZEN/ ) GCC GGC TCA AAT AG ( 3lABkFQ ) -3’ [17] . Aspergillus amstelodami F: GTG GCG GCA CCA TGT CT , A . amstelodami R: CTG GTT AAA AAG ATT GGT TGC GA , A . amstelodami probe: 5’ ( FAM ) CAG CTG GAC ( ZEN ) CTA CGG GAG CGG G ( 3lABkFQ ) -3’ [17] . Epicoccum nigrum F: TTG TAG ACT TCG GTC TGC TAC CTC TT , E . nigrum R: TGC AAC TGC AAA GGG TTT GAA T , E . nigrum probe: 5’- ( FAM ) CAT GTC TTT ( ZEN ) TGA GTA CCT TCG TTT CCT CGG C ( 3lABkFQ ) -3’ [17] . FungiQuant F: GGR AAA CTC ACC AGG TCC AG FungiQuant R: GSW CTA TCC CCA KCA CGA FungiQuant probe: 5’- ( FAM ) TGG TGC ATG GCC GTT ( 3lABkFQ ) -3’ [45] . 16S F: ACT CCT ACG GGA GGC AGC AGT , 16S R: ATT ACC GCG GCT GCT GGC . Pan-Candida F: GCA AGT CAT CAG CTT GCG TT Pan-Candida R: TGC GTT CTT CAT CGA TGC GA [49] . Detection of the ASF bacteria was performed using eight previously published organism-specific primer pairs [50] . A standard curve using dilutions of defined content of a pure sample of single species fungal DNA was created for PCR reactions to quantify their fungal DNA content . Fungal ITS1 amplicons were generated in 20 μL PCR reactions using 3 μL of each sample with 35 cycles using Invitrogen Platinum SuperFi DNA Polymerase at an annealing temperature of 48°C using the primers ITS1F ( CTTGGTCATTTAGAGGAAGTAA ) and ITS2 ( GCTGCGTTCTTCATCGATGC ) . Amplicons were then used in the second PCR reaction , using Illumina Nextera XT v2 barcoded primers to uniquely index each sample and 2x300 paired end sequencing was performed on the Illumina MiSeq , according to manufacturer’s instructions . Raw data processing and run de-multiplexing was performed using on-instrument analytics as per manufacture recommendations . Sequence reads from this study are available from the Sequence Read Archive under the project ID "PRJNA451226" . 16S sequence data were processed and analyzed as previously described including OTU assignment by alignment with the GreenGenes reference database ( release of May 2013 ) at 97% identity [51] . For analysis of ITS1 sequence data , raw FASTQ data were filtered to enrich for high quality reads including removing the adapter sequence by cutadapt v1 . 4 . 1 [52] , truncating reads not having an average quality score of 20 over a 3 base pair sliding window , removing any reads that do not contain the proximal primer sequence or any reads containing a single N ( unknown base ) by a custom script . Filtered pair-end reads were then merged with overlap into single reads using SeqPrep v1 . 1 wrapped by QIIME v1 . 9 . 1 [53] . The processed high-quality reads were firstly aligned to previously observed host sequences ( including rRNA , olfactory receptor and uncharacterized genes in human and mouse ) to deplete potential contamination , then operational taxonomic unit ( OTU ) were picked by aligning filtered reads to the Targeted Host Fungi ( THF ) custom fungal ITS database ( version 1 . 6 ) [11] , using BLAST v2 . 2 . 22 in the QIIME v1 . 9 . 1 wrapper with an identity percentage ≥97% . 16S OTUs with average relative abundance >0 . 0001 and ITS1 OTUs with average relative abundance >0 . 001 were considered to be present for downstream analysis . PCoA of Bray–Curtis dissimilarities were evaluated and plotted using R package phyloseq v1 . 13 . 3 [54] . The pair-wise differential abundance analysis was conducted by using linear discriminant analysis with LEfSe v1 . 0 . 7 at default settings [33] . Microbiota profiling analysis is described in detail above . Additional statistical analysis was performed using GraphPad Prism 7 ( GraphPad Software ) and JMP Pro 13 . 0 . 0 ( SAS Institute Inc ) . Bar and line graphs report the mean value with error bars depicting the SEM unless otherwise specified .
|
The microbiome is the ecosystem of bacteria , fungi , viruses , and other microorganisms that live in and on us . The intestines are where most of these organisms reside , and the composition of the intestinal microbiota has a profound effect on immune system function throughout the body . In this manuscript , we observe that expansion of a certain species of house dust fungus ( Wallemia mellicola ) can occur in the intestines of mice after they are treated with antibiotics and exposed to the fungus , but that mice with an intact and healthy intestinal microbiota resist this expansion . After expansion of this fungal population , the mice are more prone to develop asthma-like inflammation in their lungs when exposed to allergens . Although it is not known whether the same phenomenon can occur in people with asthma , we have also identified this fungus as a component of the microbiota of some healthy humans . Elimination of this organism from the intestines is not as simple as taking an antifungal medication because antifungal medications can also disrupt the balance of the intestinal microbial community and may similarly worsen allergic airways disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"microbiome",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"drugs",
"microbiology",
"animal",
"models",
"fungi",
"model",
"organisms",
"antibiotics",
"experimental",
"organism",
"systems",
"pharmacology",
"fungal",
"diseases",
"fungal",
"pathogens",
"bacteria",
"microbial",
"genomics",
"digestive",
"system",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"mycology",
"medical",
"microbiology",
"microbial",
"pathogens",
"mouse",
"models",
"gastrointestinal",
"tract",
"eukaryota",
"anatomy",
"genetics",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"genomics",
"organisms"
] |
2018
|
Expansion of commensal fungus Wallemia mellicola in the gastrointestinal mycobiota enhances the severity of allergic airway disease in mice
|
Currently , there is no effective vaccine to halt HIV transmission . However , pre-exposure prophylaxis ( PrEP ) with the drug combination Truvada can substantially decrease HIV transmission in individuals at risk . Despite its benefits , Truvada-based PrEP is expensive and needs to be taken once-daily , which often leads to inadequate adherence and incomplete protection . These deficits may be overcome by next-generation PrEP regimen , including currently investigated long-acting formulations , or patent-expired drugs . However , poor translatability of animal- and ex vivo/in vitro experiments , and the necessity to conduct long-term ( several years ) human trials involving considerable sample sizes ( N>1000 individuals ) are major obstacles to rationalize drug-candidate selection . We developed a prophylaxis modelling tool that mechanistically considers the mode-of-action of all available drugs . We used the tool to screen antivirals for their prophylactic utility and identify lower bound effective concentrations that can guide dose selection in PrEP trials . While in vitro measurable drug potency usually guides PrEP trial design , we found that it may over-predict PrEP potency for all drug classes except reverse transcriptase inhibitors . While most drugs displayed graded concentration-prophylaxis profiles , protease inhibitors tended to switch between none- and complete protection . While several treatment-approved drugs could be ruled out as PrEP candidates based on lack-of-prophylactic efficacy , darunavir , efavirenz , nevirapine , etravirine and rilpivirine could more potently prevent infection than existing PrEP regimen ( Truvada ) . Notably , some drugs from this candidate set are patent-expired and currently neglected for PrEP repurposing . A next step is to further trim this candidate set by ruling out compounds with ominous safety profiles , to assess different administration schemes in silico and to test the remaining candidates in human trials .
Pre-exposure prophylaxis ( PrEP ) to prevent HIV infection ( using drugs which are licensed for its treatment ) has been assessed in people at high risk of sexual transmission . Of the available agents , once-daily tenofovir and emtricitabine ( Truvada ) have been extensively studied , and demonstrate protective efficacy ( 59–100% [1 , 2] ) in individuals who are adherent to the medication; conversely poor medication adherence explains the lack of protection observed in some trials [3] . However , major shortcomings of Truvada-based PrEP are its costs [4] , a residual infection risk and the necessity for daily drug intake ( which often leads to inadequate adherence ) . These deficits may be overcome by next-generation PrEP regimen , including patent-expired antivirals and long-acting formulations . Studies assessing next-generation PrEP regimen are underway [5] , but rational selection of which agents to advance into PrEP trials based on their intrinsic pharmacology and mode of action has not been comprehensively or systematically undertaken . Moreover , studies have focussed on patent-protected compounds [6] , which are likely unaffordable in resource-constrained settings [4] hit hardest by the epidemic . The considerable sample sizes ( N > 1000 individuals ) and clinical trial duration required ( years ) to test any new candidate against tenofovir-emtricitabine , and the need to assess regimens with forgiveness for missed dosing or episodic , event-driven PrEP make the current strategy of empirical drug selection costly and prone to failure . We chose to explore an alternative strategy by developing a mathematical modelling tool to assess the per-contact efficacy of anti-HIV drugs . This approach allows prediction of prophylactic utility by integrating drug specific factors ( pharmacokinetic/pharmacodynamic ( PK/PD ) attributes ) and attributes of the targeted risk group in order to probe and discard candidates , accelerate drug development and markedly reduce costs . In this work , we are particularly interested in agents where existing patents had already , or are about to expire , in order to maximise the potential impact for low and middle income countries . Various epidemiological modelling approaches have been used to predict the public health benefits of PrEP [7] and the risk of emergent drug resistance [8–10] . These approaches are highly dependent on ad hoc parameter assumptions [11] ( specifically the per-contact PrEP efficacy ) , which may explain the different and contradictory predictions which have emerged . Knowledge of the per-contact PrEP efficacy , ideally concentration-prophylaxis relationships , are currently lacking and parameters derived from animal models poorly translate into human efficacy . Concentration-prophylaxis relationships are particularly critical to define lower concentrations in human trials that can attain e . g . > 90% protection: I . e . , ideally a PrEP candidate should be dosed such that the concentrations stay above this target ( e . g . 90% protection ) and at the same time avoid adverse effects in all individuals . For prophylaxis , there is a general void of information regarding drug-specific and drug-class specific concentration-prophylaxis relationships . While the potency of drugs to inhibit HIV replication can readily be measured in vitro , researchers are often unaware that this measure of drug potency may not coincide with the potency to prevent HIV infection ( prophylactic potency ) and consequently PrEP trial design may be flawed , incurring costs and putting individuals at risk . In a top-down approach , Hendrix et al . [12] analyzed available clinical data for Truvada to define concentration-prophylaxis relationships . However , this approach is naturally limited to PrEP candidates where sufficient clinical data already exists and is not able to disentangle the potency of the administered drugs from confounding factors . More mechanistic , bottom-up approaches integrate various host- and viral factors [13–18] to predict the probability of viral extinction . Despite their advantages , these approaches conventionally do not establish concentration-prophylaxis relations , or they are specific to particular drugs [17] or drug classes [18] . In this work , we will first analyze the drug-class specific relation between in vitro potency and PrEP efficacy and its dependency on the amount- and type of transmitted virus . Utilizing pharmacokinetic and pharmacodynamic data for all treatment-approved drugs , and simulating typical viral exposures during sexual contact , we will then screen all treatment-approved drugs for their PrEP utility and assess the sensitivity of the prophylactic endpoint with regard to uncertainties in viral dynamics parameters and with regard to variabilities in drug concentration , which can typically result from inter-individual metabolic differences or differences in medication adherence . Our central aim is to provide a tool to screen out drug candidates with a lack of- or uncertain prophylactic efficacy .
The infection probability PI ( Y0 ) for some initial state Y0 is the complement of the extinction probability PE ( Y0 ) P I ( Y 0 ) = 1 - P E ( Y 0 ) , ( 1 ) where Y0 denotes the initial viral population in a replication enabling ( target-cell ) environment . Throughout the article we will use Y = [V , T1 , T2]T , i . e . the state of the viral dynamics is defined by infectious viruses , early- and productively infected cells as outlined below . The extinction probability is defined by P E ( Y 0 ) ≔ P ( Y t = [ 0 0 0 ] | Y 0 = [ V T 1 T 2 ] ) ( 2 ) for t → ∞ . In words , the probability that all viral compartments will eventually go extinct . The prophylactic efficacy φ then denotes the reduction in infection probability per contact , φ = 1 - P I ( Y 0 | D ) P I ( Y 0 | ⌀ ) ( prophylactic efficacy ) , ( 3 ) where PI ( Y0|D ) and PI ( Y0|⌀ ) denote the infection probabilities in the presence- and absence of prophylactic drugs D respectively . The term PI ( Y0|D ) was computed using a mathematical model of the viral dynamics ( below ) and by mechanistically considering the direct effects of the distinct antivirals on viral replication whereas PI ( Y0|⌀ ) is computed analogously , assuming the absence of drug D = 0 . For the ease of notation we introduce the unit vectors V ^ , T ^ 1 and T ^ 2 which represent the states where only one infected compartment is present ( either virus , early- or late infected cells ) V ^ = [ 1 0 0 ] , T ^ 1 = [ 0 1 0 ] , T ^ 2 = [ 0 0 1 ] ( 17 ) While free virus is typically transmitted , our framework also allows to study prophylactic efficacy for arbitrary initial states . Using the notation above , any state of the system can be expressed as a linear combination of the unit vectors above . For example , 5 ⋅ V ^ ⊕ 3 ⋅ T ^ 1 ⊕ 12 ⋅ T ^ 2 denotes the state where we have 5 viruses , 3 early infected cells and 12 late infected cells . In S1 Text we provide a detailed derivation of infection/extinction probabilities after viral exposure . Herein , we will provide a sketch of the central idea . Starting from a single virus Y 0 = V ^ , we can write the Chapman-Kolmogorov equation: P E ( Y 0 = V ^ ) = ∑ n = 0 ∞ P ( Y r = n · V ^ | Y 0 = V ^ ) · P E ( Y r = V ^ ) n . ( 18 ) In words , the extinction probability P E ( Y 0 = V ^ ) is given by the probability that n viruses are produced in a single replication cycle r , ℙ ( Y r = n ⋅ V ^ | Y 0 = V ^ ) , and that all of these viruses eventually go extinct , considering all possible values of n . Herein we assumed statistical independence , i . e . P E ( Y r = n ⋅ V ^ ) = P E ( Y r = V ^ ) n . Furthermore , the extinction probabilities for parent- and progeny virus are identical when the inhibitor efficacy is constant , i . e . P E ( Y 0 = V ^ ) = P E ( Y r = V ^ ) . Next , we construct the embedded Markov chain [26] corresponding from the continuous-time Markov jump model depicted in Fig 1 with parameters in Table 1 ( details in S1 Text ) . This allows to derive algebraic formulas for ℙ ( Y r = n ⋅ V ^ | Y 0 = V ^ ) , n = 0 … ∞ . Substituting these into Eq ( 18 ) , rearranging and solving for P E ( Y 0 = V ^ ) yields a quadratic formula . Solving the quadratic formula , and using PE ( ⋅ ) = 1 − PI ( ⋅ ) we derive analytical solutions for the infection probabilities after exposure to a single virus V ^ , early- T ^ 1 and late infected cell T ^ 2: P I ( Y 0 = V ^ ) =max ( 0 , a 4 ( D ) a 1 ( D ) + a 4 ( D ) · a 5 ( D ) a 2 + a 5 ( D ) ( 1 - 1 R 0 ( D ) ) ) ( 19 ) P I ( Y 0 = T ^ 1 ) =max ( 0 , a 5 ( D ) a 2 + a 5 ( D ) · ( 1 - 1 R 0 ( D ) ) ) ( 20 ) P I ( Y 0 = T ^ 2 ) =max ( 0 , 1 - 1 R 0 ( D ) ) . ( 21 ) where R0 ( D ) denotes the basic reproductive number , i . e . the average number of viruses produced from a single founder virus [41] in a single replication cycle under the action of drug D . Using our model we have R 0 ( D ) = a 4 ( D ) a 1 ( D ) + a 4 ( D ) ⋅ a 5 ( D ) a 2 + a 5 ( D ) ⋅ a 6 ( D ) a 3 . The first solution PI ( ⋅ ) = 0 of Eqs ( 19 ) – ( 21 ) are valid in the regimen where R0 ( D ) ≤ 1 , i . e . in the regimen where extinction is certain . The second solution describes the case where infection may occur , i . e . R0 ( D ) > 1 . The pre-terms in the second solution of Eqs ( 19 ) and ( 20 ) denote the bottlenecking probabilities that a late-infected , virus producing cell is reached , starting from a free virus ( Eq ( 19 ) ) or starting from an early infected cell ( Eq ( 20 ) ) respectively . We can assume statistical independence during the onset of infection ( i . e . competition for target cells is negligible ) as noted before . Hence , for any given combination of free virus , early-stage infected cell and late-stage infected cell the extinction probability is given by P E ( Y 0 = [ V T 1 T 2 ] ) = ( P E ( Y 0 = V ^ ) ) V · ( P E ( Y 0 = T ^ 1 ) ) T 1 · ( P E ( Y 0 = T ^ 2 ) ) T 2 , ( 22 ) where the exponents V , T1 and T2 denote the number of free virus , early- and late-stage infected cells present and where we notice that PE ( ⋅ ) = 1 − PI ( ⋅ ) . Initial viral exposure after sexual intercourse occurs at tissue sites typically not receptive for establishing and shedding HIV infection ( e . g . mucosal tissues ) . Hence , the virus needs to pass several bottlenecks and physiological barriers to reach a replication enabling ( target-cell ) environment where infection can be established and from where it can shed systemically [42] . To determine realistic inoculum sizes after sexual exposure to HIV , we previously developed a data-driven statistical model linking plasma viremia in a transmitter to the initial viral population Y0 in a replication-enabling environment [18] ( Supplementary Note 4 therein for details ) . Herein , we used the ‘exposure model’ to compute drug efficacy estimates after homosexual exposure presented in section Prophylactic efficacy of treatment-approved antivirals . In brief , this ‘exposure model’ was developed to capture key clinical observations: ( i ) the average HIV transmission probabilities per exposure as reported in [20 , 21 , 43] . ( ii ) the fact that viral loads in the untreated transmitter population are approximately log-normal distributed [18 , 44–46] ( μ = 4 . 51 , σ = 0 . 98 ) and ( iii ) the observation that the plasma viremia in the transmitter is the most dominant factor determining HIV transmission [44 , 47–49] . More specifically , it was reported that each 10-fold increase in the transmitter’s viral load increases the transmission probability per coitus by approximately 2 . 45-fold [47] ( similar values confirmed in [49] ) . The aforementioned clinical observations can be summarised in the formula below: P ¯ trans = ∫ ν = 0 ∞ P ( VL = ν ) · ( ∑ n = 0 ∞ P ( Y 0 = n · V ^ | VL = ν ) · P I ( Y 0 = n · V ^ ) ) ( 23 ) where P ‾ t r a n s is the average transmission probability per exposure/coitus ( given in ( i ) ) , P ( VL = ν ) is the probability density of viral load in the donor ( log-normal distributed , given in ( ii ) ) , P I ( Y 0 = n ⋅ V ^ ) is the infection probability when n viruses reach a replication enabling site ( computed from the virus dynamics model above with P I ( Y 0 = V ^ ) ≈ 0 . 0996 ) and P ( Y 0 = n ⋅ V ^ | V L = ν ) denotes the ‘exposure model’ ( the probability that n viruses reach a replication-enabling compartment after viral exposure from a transmitter with virus load ν ) . For the ‘exposure model’ , we assumed a binomial distribution P ( Y 0 = n · V ^ | VL = ν ) = ( ⌈ ν m ⌉ n ) · r n · ( 1 - r ) ⌈ ν m ⌉ - n ) ( 24 ) where m = log10 ( 2 . 45 ) is given by ( iii ) [47] and the success probability r was estimated in a previous work [18] ( Supplementary Note 4 therein ) , e . g . rhomo = 3 . 71 ⋅ 10−3 for homosexual exposure . However , the model can be adapted to the different exposure types ( e . g . heterosexual , needle-stick , etc … ) . In this model , the success probability r summarises both the extent of local exposure , as well as the probability of passing all bottlenecking physiological barriers and reaching a replication enabling target cell compartment . Lastly , in line with Keele et al . [22] , we observed that if infection occurs in our model it is established by a very low number of viruses after homosexual contact and usually by a single founder virus after heterosexual contact .
Drug-specific inhibition of viral replication can be studied in vitro , for example in single-round turnover experiments [40] or even more mechanistically using enzymatic assays in conjunction with appropriate mathematical models [50] . Since the infection risk per exposure is already low in untreated individuals [20 , 21] , exploring the prophylactic efficacy ( reduction in infection risk ) in the clinic is difficult , requiring very long ( several years ) clinical trials with many individuals ( N > 1000 ) to achieve statistically evaluable results . Systematic evaluation of concentration-effect relations is not feasible in this context , notwithstanding ethical concerns . We wanted to gain a deeper insight how in vitro measurable direct drug efficacy η translates into prophylactic efficacy φ ( reduction in infection probability per exposure ) in a drug-class specific manner . Particularly , since different antiviral drug classes inhibit distinct stages in the HIV replication cycle , we wanted to elucidate how these different mechanisms of action affect prophylaxis . We combined Eqs ( 11 ) – ( 16 ) with Eqs ( 19 ) – ( 21 ) into Eq ( 3 ) to predict prophylactic efficacy . When relating direct drug effects η to prophylactic efficacy φ we observed striking drug-class specific differences as illustrated in Fig 2 . Using parameters from Table 1 we found that the prophylactic efficacy φ may be less than predicted by in vitro measurable direct drug effects η . The sole exception are reverse transcriptase inhibitors ( RTI ) in case of exposure to a single virus particle Y 0 = V ^ where the two measures of drug efficacy coincide . While the prophylactic efficacy after exposure to a single virus are moderately less than the direct effects of co-receptor antagonists CRA and integrase inhibitors InI respectively ( Fig 2A ) , there is a profound difference for protease inhibitors , which do not seem to reduce HIV transmission unless their direct efficacy η exceeds ≈ 95% . Interestingly , a similar observation using a different mathematical model and only distinguishing RTIs and PIs has been made by Conway et al . [13] . While HIV-transmission typically occurs after exposure to free virus , it is still useful to study the prophylactic efficacy of distinct drug classes in the hypothetical case when infected cells were present in the exposed individual . A realistic example for this scenario is post-exposure prophylaxis ( PEP ) : During PEP , drugs are taken shortly after virus exposure and initial viral replication steps may have taken place generating early- or late infected cells . As can be seen in Fig 2B and 2C , the prophylactic efficacy of all drugs profoundly deteriorates compared to their direct effects , i . e . only very effective ( in terms of η ) drugs may prevent systemic infection once cells become infected in the exposed individual . An exception are integrase inhibitors: their prophylactic efficacy φ is moderately less than their direct effect η ( panel B ) if only early infected cells T1 ( before proviral integration ) were present . Thus , while the prophylactic efficacy of all other drug classes is profoundly less than their direct effects once infected cells emerged , integrase inhibitors may still potently prevent infection . An intuitive explanation for the deterioriation of prophylactic efficacy can be made in terms of changes in drug-target stoichiometry: For example , after exposure to a single virus V ^ , drugs from the classes of CRAs , RTIs and InIs need to block a single reaction to foster viral extinction . For PIs however , the same is only achieved if maturation of the entire viral progeny is inhibited ( possibly hundreds of particles ) . Similarly , when considering a single early infected cell T ^ 1 , CRAs and RTIs can only prevent further viral expansion after viral progeny has emerged . Subsequently , for each viral particle ( possibly hundreds ) the respective target processes ( receptor binding , reverse transcription ) need to be blocked by the inhibitors . Along the same lines of argumentation it is also evident that prophylactic efficacy is generally more favourable in the case of PrEP , compared to post-exposure prophylaxis ( PEP ) , where initial viral replication may have occurred . In vitro measured drug potency IC50 , IC90 usually guides the design of PrEP trials [51] . In particular , dosing regimen are designed so that the majority of individuals achieve drug levels just above the 90% inhibitory concentrations IC90 . However , it has never been rigorously investigated whether these ‘target concentrations’ are sufficient to provide 90% protection against HIV infection . Integrating Eq ( 10 ) into Eqs ( 11 ) – ( 16 ) , ( 19 ) and ( 3 ) allows to predict the concentration-prophylaxis profile for different HIV-1 inhibitor classes . Rearranging this composite equation reveals how in vitro measured drug potency IC50 , IC90 can be translated into prophylactic potency ( 50% and 90% reduction in infection risk , EC50 and EC90 , respectively ) , guiding clinical trial design . The derived analytical expressions for the prophylactic efficacy ( reduction in infection risk ) indicate that the shape of the concentration-prophylaxis profile varies considerably for different HIV-1 inhibitor classes with important consequences for their prophylactic endpoints ( % reduction in HIV transmissibility ) . After exposure to a single virion Y 0 = V ^ , the overall shape of the concentration-prophylaxis profile for co-receptor antagonists ( CRAs ) , reverse transcriptase inhibitors ( RTIs ) and integrase inhibitors ( InIs ) is a classical Emax equation ( the equation of choice for evaluating concentration-effect relations ) , see S1 Text for derivation . φ ( V ^ ) =R 0 ( ⌀ ) R 0 ( ⌀ ) - 1 · D m IC 50 m ( 1 υ ) + D m≈ R 0 ( ⌀ ) ≫ 1 D m EC 50 m + D m ( CRA ) ( 25 ) φ ( V ^ ) =R 0 ( ⌀ ) R 0 ( ⌀ ) - 1 · D m IC 50 m + D m≈ R 0 ( ⌀ ) ≫ 1 D m EC 50 m + D m ( RTI ) ( 26 ) φ ( V ^ ) =R 0 ( ⌀ ) R 0 ( ⌀ ) - 1 · D m IC 50 m ( 1 ϑ ) + D m≈ R 0 ( ⌀ ) ≫ 1 D m EC 50 m + D m ( InI ) ( 27 ) where D denotes the concentration of the drug in the blood plasma , m is a slope parameter and IC50 denotes the plasma concentration of the drug that inhibits the targeted process ( co-receptor binding , reverse transcription or proviral integration ) by 50 percent . This parameter can typically be measured in vitro , e . g . using single-round turnover experiments [40] and is stated in Table 2 for various drugs . Parameters υ = C L ⋅ ρ rev , ⌀ C L ⋅ ρ rev , ⌀ + β ⋅ T u < 1 and ϑ = δ P I C + δ T 1 δ P I C + δ T 1 + k < 1 denote the respective probabilities , in the absence of drugs , that the virus is eliminated before entering a host cell , and that essential virus compartments get cleared intracellularly after reverse transcription and before provirus integration . The parameter EC50 denotes the plasma concentration of the drug that decreases the probability of infection by 50% , i . e . the prophylactic potency of the drug . R0 ( ⌀ ) denotes the basic reproductive number in the absence of drugs , i . e . the average number of viruses produced from a single founder virus [41] in a single replication cycle when no antivirals were present ( R0 ( ⌀ ) ≈ 67 according to the utilized model ) . When the target cell density is sufficiently high ( herein considered as a target cell environment ) , we have R0 ( ⌀ ) ≫ 1 and hence the left-side scaling factor in Eqs ( 25 ) – ( 27 ) will be close to one , R0 ( ⌀ ) / ( R0 ( ⌀ ) − 1 ) ≈ 1 . An analysis with low target cell densities is provided in S2 Text . In case of exposure to a single virus particle V ^ , the slope parameters in the right-most equations coincide with the slope parameter for the respective drug-targeted process m ( Eq ( 10 ) ) , stated in Table 2 . Notably , for RTIs , we have EC50 ≈ IC50 , i . e . the drugs potency measured in vitro in single-round turnover experiments [40] directly translates into its potency to prevent infection . Using parameters from Table 1 we observe EC50 > IC50 for CRAs and InIs , i . e . compared to their in vitro potency , they are less potent in preventing infection . This is largely due to the respective factors ϑ−1 , υ−1 > 1 , compare Fig 3A–3C . For InIs this observation is robust across a broad range of parameter values , as shown in S2 Text . Consequently , for InIs , higher concentrations are required to prevent infection than suggested after conducting the respective in vitro experiments . For CRAs , predictions are parameter dependent , S2 Text . Rearranging Eqs ( 25 ) – ( 27 ) allows to directly compute the drug concentration that prevents infection with x percent probability ( the ECx ) from the corresponding in vitro 50% inhibitory concentration IC50 ( derivations in S3 Text ) : In case of exposure to a single virus particle we get EC x = IC 50 · ( F · x 100 · C - x ) 1 / m , ( 28 ) where ECx is the drug concentration that achieves x percent of prophylactic efficacy and the term F ≥ 1 is a drug class specific factor F = { ( CL + β · T u ρ ) / CL , for CRA 1 , for RTI 1 / ϑ , for InI . ( 29 ) and C ≔ R 0 ( ⌀ ) R 0 ( ⌀ ) - 1 ≈ 1 , ( 30 ) if R0 ( ⌀ ) ≫ 1 . Importantly , when exposure to multiple viruses occurs , the concentration-prophylaxis profile is no longer an Emax equation for any inhibitor class , Fig 3A–3C . Furthermore , the slope parameter increases and the EC50 may exceed the in vitro measurable IC50 value . At large inoculum , the corresponding profiles become switch-like . For protease inhibitors ( PIs ) , we derive a power function to describe their prophylactic efficacy ( mechanistic derivation in S1 Text ) : φ ( V ^ ) = 1 R 0 ( ⌀ ) - 1 · D m IC 50 m = C · D m IC 50 m ( PI ) for 0 ≤ φ ≤ 1 , ( 31 ) where C << 1 is a constant . Moreover , for realistic ( large ) R0 ( ⌀ ) ≫ 3 their plasma concentration has to exceed their IC50 to decrease the probability of infection by at least 50% , Fig 3D . Similarly , we can rearrange the equation above and obtain EC x = IC 50 ⋅ ( G ⋅ x 100 ) 1 / m , ( 32 ) for the exposure to a to a single virus , where G ≔ R0 ( ⌀ ) − 1 . Again , in case of exposure to multiple viruses , the slope parameter and EC50 increase , making the prophylactic efficacy of PIs exhibiting a switch-like behaviour as can be seen in Fig 3D . This switch-like behaviour makes the prophylactic use of PIs vulnerable to non-adherence , as well as general variations in concentrations ( e . g . pharmacokinetics , inter-individual variability ) , and the prophylactic efficacy with these inhibitors may alternate between zero- or complete protection . The combination of the nucleoside reverse transcriptase inhibitors ( NRTIs ) emtricitabine and tenofovir ( Truvada ) is the only intervention approved for pre-exposure prophylaxis ( PrEP ) . According to our previous estimates [18] , Truvada provides 96% protection in fully adherent individuals , which is in line with clinical estimates of 86-100% protection in the IPERGAY study [1] , 58-96% in the PROUD study [2] and 96% in the Partners PrEP OLE study in apparently highly adherent individuals . The VOICE [3] and FEM-PrEP [52] studies indicated that Truvada may not prevent infection in poorly adherent individuals . Currently , a number of drugs are under investigation for PrEP repurposing [6] . Notably , all currently investigated compounds are patent-protected and may not be affordable in resource-constrained countries hit hardest by the epidemic . In this work , we wanted to unselectively assess the utility of treatment-approved antivirals for prophylaxis and to assess whether currently neglected ( patent-expired ) compounds may be cost-efficient alternatives to be further explored in non-profit prophylaxis programmes . We utilized comprehensive sets of drug-specific pharmacodynamic- and pharmacokinetic parameters ( Table 2 ) to parameterize Eq ( 10 ) and to predict the prophylactic efficacy of treatment approved CRAs , non-nucleoside reverse transcriptase inhibitors ( NNRTIs ) , InIs and PIs at clinically relevant concentration ranges ( the class of NRTIs have been analyzed in earlier work [18] ) . Moreover , we sampled the extent of viral exposure ( number of viruses transmitted and reaching a replication-enabling compartment; Eq ( 22 ) ) from a previously parameterized distribution [18] that accurately reflects transmitter virus loads and drug-free infection probabilities after sexual contact . The resultant benchmark is depicted in Fig 4 . Fig 4 allows for an initial screen of the utility of the various drugs for oral PrEP . Most analyzed drugs , except for maraviroc ( MVC ) , raltegravir ( RAL ) , elvitegravir ( EVG ) and nelfinavir ( NFV ) , potently prevent infection at concentrations ranges typically encountered in fully adherent individuals during treatment ( range between minimum- to maximum concentration , [Cmin; Cmax] ) . During prophylaxis , adherence to the dosing regimen is a major problem and we thus consider a lower bound concentration that would arise if the drug had not been taken for three days prior to exposure Clow ( thin dashed vertical line in Fig 4 ) to emphasise a ‘pharmacokinetic safety margin’ in case of poor adherence . Numerical values for the computed maximum prophylactic efficacy and the efficacy at the lower bound concentrations are reported in Table 3 alongside with estimated EC50 and EC90 values . While in Table 3 we report the EC50 and EC90 values after challenge with a single virus V ^ , the corresponding values after virus challenges sampled from the distribution for homosexual exposure Eq ( 24 ) were almost identical , see S4 Text for a comparison . Our simulations indicate a residual risk of infection for most analyzed drugs . Notably , most protease inhibitors may confer anything from none- to absolute protection within relevant concentration ranges , [Clow; Cmax] , which highlights a severe limitation to their PrEP use in the context of poor adherence or pharmacokinetic ( intra-/inter individual ) variability . An exception among this rule is darunavir ( DRV ) , which is predicted to be almost fully protective for the entire concentration range . Of the analyzed non-PI drugs , the NNRTIs efavirenz ( EFV ) , nevirapine ( NVP ) , etravirine ( ETR ) and rilpivirine ( RPV ) are extremely potent with regard to prophylaxis: These drugs prevent infection , even when the drug had not been taken for three consecutive days , Table 3 . Notably , NVP and EFV are patent-expired and may represent suitable candidates for use in resource-constrained settings ( price per day ≈ 0 . 1US$ ) . The co-receptor antagonist maraviroc ( MVC ) and the integrase inhibitor dolutegravir ( DTG ) retain some prophylactic efficacy ( 50 and 72% respectively ) at lower bound concentrations Clow . The CRA maraviroc ( MVC ) , the NNRTI rilpivirine ( RPV ) and the InI raltegravir ( RAL ) are currently investigated for use as PrEP compounds ( long-acting injections of RPV and RAL; oral- or topical application of MVC ) . In our simulations the predicted PrEP efficacy of these drugs would drop to 8% ( RAL ) and 50% ( MVC ) when the drug had not been taken for three consecutive days prior to virus exposure . Notably , RPV remained 100% effective . Lastly , we want to note that our predictions are based on viral dynamics parameters that may under-predict prophylactic efficacy , as indicated in S2 Text . The main purpose of this modelling study was to rule out drug candidates , based on lack-off- or uncertain- prophylactic efficacy . While some drugs’ prophylactic efficacy might be under-predicted , this conservative choice of parameters provides a more solid scientific basis for the remaining candidates that are predicted to be potent .
Our intent was to develop a tool to screen out unsuitable candidates for PrEP based on unfavourable pharmacokinetic and pharmacodynamic characteristics . Clearly , the attributes which make any compound favourable extend beyond PK/PD , and critically also depend on tolerability , ease of dosing , cost and acceptability . Nevertheless , screening antiretroviral agents based on their intrinsic antiviral activity , mode of action , duration of efficacy beyond the dosing interval , and tolerance for missed dosing is a logical starting point when assessing potential candidates for PrEP . Strikingly , we observed that in vitro measured drug potency may over-estimate PrEP potency in a drug-class specific manner . For all non-RTI drugs dosing schedules in clinical trials may have to be adjusted accordingly to reach the desired prophylaxis endpoints ( % protection ) . We provide an easy-to-use software tool to determine the corresponding target concentrations ( www . systems-pharmacology . org/prep-predictor ) . For non-PI drugs , we observed a more graded relationship between their prophylactic efficacy and drug concentrations . At low virus inoculum sizes , the slope of their concentration-prophylaxis profile is largely determined by the slope coefficient that describes their direct effects [40] . Notably , for PIs we observed a very steep concentration prophylaxis profile , suggesting that within clinically relevant ranges for oral PrEP ( Fig 4 ) their efficacy is likely to switch between zero- and complete protection , in an ‘either-or’ scenario . This characteristic renders PIs particularly vulnerable to poor adherence and drug-drug interactions . An intuitive explanation for this steep concentration-prophylaxis profile of PIs ( power function in Eq ( 31 ) ) is based on its unfortunate drug-to-target stoichiometry: A single late infected cell T2 produces hundreds of infectious viruses on average ( using parameters from Table 1 a6/a3 = 670 ) and a PI needs to prevent all of them from becoming infectious to fully prevent infection . By contrast , all other compounds only need to prevent a single viral entity from progressing , explaining the proportionality to the EMAX equation seen in Eqs ( 25 ) – ( 27 ) . By screening all treatment-approved antivirals for their PrEP utility , we predicted that efavirenz ( EFV ) , nevirapine ( NVP ) , etravirine ( ETR ) , rilpivirine ( RPV ) and darunavir ( DRV ) may fully prevent infection after oral application and in case of poor adherence ( Table 3 and Fig 4 ) . Notably , these compounds have favourable inhibitory quotients ( clinically achieved concentrations vastly exceed their EC50 ) and their long elimination half-lives guarantees that inhibitory quotients stay in that favourable range . The drugs maraviroc ( MVC ) and dolutegravir ( DTG ) potently prevent infection but may allow for HIV transmission when individuals poorly adhere to the medication . Notably , the NNRTIs EFV , NVP , RPV and ETR exhibit long elimination half-lives ( 30-40h ) and achieve concentrations required for PrEP to act quickly , and durably . However , there are some safety concerns with liver toxicity , which contraindicate e . g . the use of NVP in uninfected individuals . Liver toxicity to ETR remains to be elucidated in the context of prophylaxis . Skin reactions ( ETR and EFV ) and neuropsychiatric effects ( EFV ) have been reported in the context of HIV treatment that need to be evaluated in the context of potential PrEP applications . Likewise , skin reactions and rare liver toxicities with DRV need careful assessment in the context of PrEP repurposing . Moreover , the particular concentration-prophylaxis profile , as depicted in Fig 4 , argues for a form of DRV administration that is not dependent on daily dosing for maintaining drug levels ( e . g . slow release or nanoparticle formulations ) . For rilpivirine ( RPV ) , our simulations suggest that near complete protection can be achieved when concentrations exceed EC90 , Fig 4 . RPV is currently investigated as a long-acting formulation in HPTN076 using 1200mg injections every 2 month which yields tough concentrations ( median 186 nM ) well in excess of this target . However , significant variability is still observed related to gender , and between injections on different occasions [51] which could be incorporated into future model generations . Besides rilpivirine , maraviroc ( MVC; 300mg once daily ) , and raltegravir ( RAL ) are currently clinically investigated for oral PrEP . Our simulations suggest MVC may incompletely prevent infection even at maximum concentrations and that its efficacy steadily drops with declining levels down to 50% when the drug had not been taken for three days prior to exposure . Results from the NEXT-PrEP ( HPTN 069 ) phase II study observed that MVC may not be potent enough on its own and that among those acquiring HIV infection , MVC concentrations were low , absent or variable [53] . Our model prediction is consistent with the reported lack of efficacy of MVC as PrEP in animals and human explant samples [54] and suggests that the potency of MVC , against infection may be less than its potency in preventing HIV replication ( EC50 > IC50 , EC90 > IC90 ) . However , EC50 , EC90 estimates for co-receptor antagonists are highly parameter sensitive ( S2 Text ) warranting further research into elucidating the early infection dynamics . Using the parameters presented in Table 1 , we estimate that the EC90 may be around 350nM , which is approximately 70 times larger than its IC50 ( conversion formula provided in the results section ) . Notably , during the dose finding for MVC an IC90 of only 3 . 9nM ( 2ng/ml ) was considered and this estimate was taken directly to determine target concentrations providing 90% prophylactic efficacy . Other compounds currently under investigation ( HPTN-083 ) [55] , but not evaluated in our study are the novel long-acting integrase inhibitor cabotegravir . Our model has several limitations , but also a number of important advantages . Our simulations do not take into account drug concentrations at the site of mucosal exposure ( e . g . cervix , rectum ) [51 , 56] . These concentrations have , however , not been validated as targets for successful prevention or treatment , whereas data exist ( albeit limited ) for plasma drug concentrations . Instead , we modelled based on unbound concentrations , in line with the broadly accepted ‘free drug hypothesis’ . Under the ‘free drug hypothesis’ , the unbound concentrations are assumed to be available at the target site to exert pharmacological effects . For drugs highly bound to plasma protein ( > 90% ) , naturally , their total concentrations at sites other than the plasma are magnitudes lower [56] . Strikingly , however , the unbound concentrations are identical [57] . Therefore , throughout the work , we assumed , according to the ‘free drug hypothesis’ [58] that the unbound concentrations in plasma and at the target site coincide , where the latter exerts the antiviral effect [59 , 60] . All analyzed NNRTIs , InIs and PIs , except for raltegravir ( RAL ) , are highly lipophilic , enabling the unbound drug to rapidly cross cellular membranes , generating an equilibrium between the unbound drug on either side of the cellular membrane [61] . Even for the weakly lipophilic compound raltegravir , intracellular concentrations are proportional to plasma concentrations by a factor precisely resembling their unbound moiety [62 , 63] , strongly arguing for the validity of the ‘free drug hypothesis’ for all analyzed drugs . However , ultimate proof in terms of local measurements in humans are lacking currently and may be difficult to obtain experimentally . On the contrary , nucleoside reverse transcriptase inhibitors ( NRTIs ) , which we analysed in a previous work [18] are not expected to obey the ‘free drug hypothesis’ [17 , 30 , 64] . These compounds need to be actively taken up by cells and converted intracellularly into pharmacologically active triphosphates ( NRTI-TP ) . Since the expression of transporters and intracellular enzymes is likely cell-specific , different cell types may contain vastly different concentrations of pharmacologically active compound . It is therefore entirely unclear what relevance concentration measurements of NRTI-TPs in tissue homogenates [65] ( containing HIV target- and non-target cells ) from sites of viral exposure ( e . g . cervix , rectum ) have in terms of prophylaxis . Utilising the virus exposure model Eqs ( 23 ) and ( 24 ) , we estimated the probability of virus clearance ( and the prophylactic efficacy φ ) as a function of the number of viruses ultimately reaching a target cell environment , and not as a function of mucosal exposure . The quantitative role of a number of physiological processes underlying primary infection is currently not fully resolved and impossible to measure in humans ( e . g . the cells involved at the local site of exposure , their abundance , locations , their capabilities to transduce virus through physiological barriers and the respective R0s ) . It is known however , that the virus has to overcome a number of physiological bottlenecks/barriers to reach a compartment that permits viral expansion . Despite the mucosal barrier , the sub-mucosal target cell density might initially be low [66] , such that only a tiny fraction of viruses find a target cell before being cleared . It has also been reported [66–68] , that target cells are subsequently recruited to the site of initial exposure due to inflammation and seminal exposure , mitigating the ‘low target cell bottleneck’ subsequently . If the low target-cell bottleneck is only prevalent during the first replication cycle it can also be modelled by simply considering a smaller virus inoculum that reaches a target-cell environment . In our approach , to obviate model- and parameter uncertainties , we chose a minimal/parsimoneous , data-driven approach that treats all physiological barriers as a single bottleneck lumped in terms of the ‘success probability’ r in Eq ( 24 ) . The target cell environment herein is a compartment that is decisive for establishing- and shedding infection ( this compartment requires R0 > 1 ) . We also assumed that this compartment is well-perfused at the time scale of interest . Under this assumption , viral kinetic parameters measured in plasma coincide with kinetic parameters at the target cell environment , after converting the deterministic reaction parameters to their respective stochastic counterparts ( Table 1 ) . Notably , the model ( see Methods section ) is calibrated [18] to reflect the per-contact infection risks for typical transmitter virus loads and different modes of sexual exposure ( homo- and heterosexual ) , but can also be adapted to model intravenous exposure by e . g . injection . The calibrated virus exposure model [18] ( see Methods section ) predicts that either none- or a single infectious virus enter a replication enabling compartment in the majority of hetero-/homosexual contacts . Thus , we suggest that E C 50 ( V ^ ) , E C 90 ( V ^ ) values stated in Table 3 provide a good proxy for the drug-specific prophylactic potency after sexual exposure to HIV ( see also S4 Text for a comparison ) . Importantly , we also observed that increasing the inoculum size decreases the prophylactic efficacy of all drug classes considered ( as suggested by increasing EC50 and EC90 ) and increases the steepness of the concentration prophylaxis profile , Fig 3 . PIs in particular displayed an almost switch-like prophylactic profile in the case of large inoculum sizes . These observations strikingly indicate that preventive target concentrations can depend on the route of transmission . I . e . , intravenous exposure to HIV ( larger inoculum sizes compared to sexual exposure ) may require higher concentrations for HIV prevention . When R0 ( ⌀ ) is relatively large , we find that our predictions of prophylactic efficacy and -potency for or CRAs , RTIs and InIs are relatively invariant to parameter changes ( compare Eqs ( 25 ) – ( 30 ) ) . However , we find that when considering an extremely broad range of 1 . 7 < R0 ( ⌀ ) < 112 values , as in S2 Text , that the parameters used are rather conservative in the sense that they disfavour the drugs and may under-predict prophylactic efficacy . With regard to our work’s aim ( screen out candidates based on lack-off- , or uncertain potency ) such a conservative parameter choice should be preferred . For PIs , although highly sensitive to changes R0 ( compare Eqs ( 31 ) and ( 32 ) ) the qualitative statements made ( the prophylactic potency is less than suggested by IC50 , as well as the steep shape of the concentration-prophylaxis profile ) are unaffected for arbitrary , yet realistic parameters , as analysed in S2 Text . However , in the provided software tool ( www . systems-pharmacology . org/prep-predictor ) it is possible to freely change all virus dynamics parameters . Notably , Ribero et al . [69] have recently estimated R0 ≈ 8 during acute infection ( ≤ 10 days after exposure , virus is detectable ) , which is much lower than the value used by us R0 ≈ 67 , which considers viral replication immediately after exposure , in the so-called eclipse phase before virus becomes detectable . Our R0 value is relatively high because we assume a lower CL ( clearance of free virus ) during this early phase of infection , in line with other modelling approaches [14 , 33] and in line with the observation that adaptive immune responses develop only after about 14 days post exposure [70] . However , if we utilise CL = 23 ( 1/day ) , as in Ribero et al . [69] , we obtain similar values of R0 . The knowledge of concentration-prophylaxis relationships between drug classes , and for each component of a particular drug class allows for the intelligent design of PrEP regimens , including how quickly protection can be achieved after a loading dose and how forgiving the regimen is towards missed dosing events . In related article [27] we develop a sophisticated simulation framework that allows to make use of population pharmacokinetic models , to fully explore inter-individual pharmacokinetics and to assess sensitivity towards dosing , individual pharmacokinetic variability and timing of viral challenges . Our model can be adapted or developed in a number of ways . On a technical side , the analytical solutions provided in the article can be neatly integrated into hybrid stochastic-deterministic algorithms that consider time-varying drug concentrations ( pharmacokinetics ) , as outlined in an accompanying article [27] . In brief , therein we utilize analytic solutions for the extinction probability , Eq ( 22 ) , to define a set of states where extinction is feasible ( extinction simplex ) . Whenever trajectories leave the extinction simplex , simulations can be stopped and a hybrid stochastic-deterministic trajectory can be safely classified as an infection event . Regarding applications , the separate impact of treatment as prevention [71] ( in the case of the donor ) versus prophylactic efficacy in the exposed individual can be readily simulated by calibrating the virus load distribution in potential transmitter populations ( see ‘exposure model’ in the Methods section ) . The effect of PrEP on the transmission of resistance can be estimated by altering R0 ( ⌀ ) ( the fitness cost of resistance ) and by simultaneously increasing IC50 in Eq ( 10 ) ( extent of resistance ) . The fitness cost of resistance translates into a decreased transmissibility of resistance in the absence of drugs ( Eqs ( 19 ) – ( 21 ) ) , while the extent of resistance translates into an increased transmissibility relative to the wildtype at increasing drug concentrations , as e . g . illustrated in [18] ( Figure 3 therein ) . Consequently , provided any transmitted resistance confers some fitness defect , prophylaxis may increase the frequency of transmitted resistance relative to the wildtype , but not its absolute occurrence [18 , 72] . Since resistance to HIV drugs generally develops in a stepwise manner , the change in EC50 following acquisition of a resistance mutation can be introduced into this model , to identify a zone of selective pressure for the de novo evolution or spread of resistance under PrEP . However , during the early events when the virus infection can still be averted , the population size is too small for resistance to appear de novo: A single point mutation appears with probability 1 − ( 1 − μ ) k at a particular base , where μ ≈ 2 . 2 ⋅ 10-5 is the per base mutation rate of HIV during reverse transcription [73] and k is the number of reverse transcription ( = cell infection ) events . Thus , de novo resistance can be assumed to appear , if e . g . PrEP had not been taken at the time of exposure , such that the infection expanded exponentially , and when PrEP is ( re- ) initiated some time after this early infection has been established . De novo resistance development in the context of poor adherence can be modelled in analogy to the work conducted by Rosenbloom et al . [74] . It is well known that the establishment of a latent reservoir is the major barrier to viral extinction during treatment [75] and this reservoir may be established as early as 3 days post infection [76 , 77] . In the current framework , we computed viral extinction when t → ∞ , assuming drug concentrations stayed constant . Thus , extinction estimates are not affected by the inclusion of a long lived cellular compartment . In an accompanying article [27] we overcome this assumption , explicitly considering drug pharmacokinetics and e . g . short-course prophylaxis . In the accompanying article infection of long lived cells are considered as an algorithmic stopping criterium: I . e . , whenever long lived cells become infected , viral extinction is considered infeasible . In summary , we have developed a mechanistic modelling tool to a priori screen antivirals for their prophylactic utility . Our approach revealed that in vitro measured drug potency ( IC50 , IC90 ) should not be used directly to identify lower bound effective concentrations in PrEP trials: With the exception of reverse transcriptase inhibitors , PrEP potency may be less than in vitro drug potency , i . e . higher concentrations of drug are required for prophylaxis than suggested by their in vitro potency . Consequently , when clinical trial design is guided by in vitro drug potency , prophylactic dosing regimen may be selected that attain insufficient concentrations to adequately prevent HIV infection . Instead , we recommend to use the tool provided ( www . systems-pharmacology . org/prep-predictor ) to translate in vitro drug potency into prophylactic efficacy . We used the developed methods to assess the prophylactic utility of all treatment approved antivirals , allowing to rule out particular candidates by lack-of- , or uncertain prophylactic efficacy . To this end , we presented results using viral dynamics parameters that may under-predict prophylactic efficacy ( S2 Text ) . These preliminary screens indicated that darunavir ( DRV ) , efavirenz ( EFV ) , nevirapine ( NVP ) , etravirine ( ETR ) and rilpivirine ( RPV ) may fully prevent infection at concentrations typically achieved during treatment and with an adequate ‘pharmacokinetic margin’ . Notably , this prediction is robust across a wide range of ( uncertain ) parameters ( S2 Text ) . Moreover , we predicted that maraviroc ( MVC ) and dolutegravir ( DTG ) can potently prevent infection , but that these drugs do not provide a comparable ‘pharmacokinetic margin’ . Furthermore , predictions for MVC are uncertain with respect to viral dynamics parameters ( efficacy may both be over- or underpredicted ) . A next logical step is to further trim this candidate set by ruling out compounds with ominous safety profiles , followed by an assessment of different dosing ( roll-out ) schemes .
|
Pre-exposure prophylaxis ( PrEP ) is a novel , promising strategy to halt HIV transmission . PrEP with Truvada can substantially decrease the risk of infection . However , individuals often inadequately adhere to the once-daily regimen and the drug is expensive . These shortcomings may be overcome by next-generation PrEP compounds , including long-acting formulations . However , poor translatability of animal- and ex vivo/in vitro experiments , and difficulties in conducting long-term trials involving considerable sample sizes ( N > 1000 individuals ) make drug-candidate selection and optimization of administration schemes costly and often infeasible . We developed a simulation tool that mechanistically considers the mode-of-action of all antivirals . We used the tool to screen all available antivirals for their prophylactic utility and identified lower bound effective concentrations for designing PrEP dosing regimen in clinical trials . We found that in vitro measured drug potency may over-predict PrEP potency , for all antiviral classes except reverse transcriptase inhibitors . We could rule out a number of antivirals for PrEP repurposing and predicted that darunavir , efavirenz , nevirapine , etravirine and rilpivirine provide complete protection at clinically relevant concentrations . Further trimming of this candidate set by compound-safety and by assessing different implementation schemes is envisaged .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"drugs",
"immunology",
"microbiology",
"retroviruses",
"viruses",
"preventive",
"medicine",
"pre-exposure",
"prophylaxis",
"immunodeficiency",
"viruses",
"rna",
"viruses",
"pharmacology",
"reverse",
"transcriptase",
"inhibitors",
"public",
"and",
"occupational",
"health",
"white",
"blood",
"cells",
"animal",
"cells",
"medical",
"microbiology",
"hiv",
"microbial",
"pathogens",
"t",
"cells",
"viral",
"replication",
"pharmacokinetics",
"prophylaxis",
"cell",
"biology",
"virology",
"viral",
"pathogens",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"antivirals",
"lentivirus",
"organisms"
] |
2019
|
Mechanistic framework predicts drug-class specific utility of antiretrovirals for HIV prophylaxis
|
The cell wall is a dynamic structure that is important for the pathogenicity of Candida albicans . Mannan , which is located in the outermost layer of the cell wall , has been shown to contribute to the pathogenesis of C . albicans , however , the molecular mechanism by which this occurs remains unclear . Here we identified a novel α-1 , 6-mannosyltransferase encoded by MNN10 in C . albicans . We found that Mnn10 is required for cell wall α-1 , 6-mannose backbone biosynthesis and polysaccharides organization . Deletion of MNN10 resulted in significant attenuation of the pathogenesis of C . albicans in a murine systemic candidiasis model . Inhibition of α-1 , 6-mannose backbone extension did not , however , impact the invasive ability of C . albicans in vitro . Notably , mnn10 mutant restored the invasive capacity in athymic nude mice , which further supports the notion of an enhanced host antifungal defense related to this backbone change . Mnn10 mutant induced enhanced Th1 and Th17 cell mediated antifungal immunity , and resulted in enhanced recruitment of neutrophils and monocytes for pathogen clearance in vivo . We also demonstrated that MNN10 could unmask the surface β- ( 1 , 3 ) -glucan , a crucial pathogen-associated molecular pattern ( PAMP ) of C . albicans recognized by host Dectin-1 . Our results demonstrate that mnn10 mutant could stimulate an enhanced Dectin-1 dependent immune response of macrophages in vitro , including the activation of nuclear factor-κB , mitogen-activated protein kinase pathways , and secretion of specific cytokines such as TNF-α , IL-6 , IL-1β and IL-12p40 . In summary , our study indicated that α-1 , 6-mannose backbone is critical for the pathogenesis of C . albicans via shielding β-glucan from recognition by host Dectin-1 mediated immune recognition . Moreover , our work suggests that inhibition of α-1 , 6-mannose extension by Mnn10 may represent a novel modality to reduce the pathogenicity of C . albicans .
Candida albicans is a common fungal microorganism that colonizes the oral , genital and gastrointestinal surfaces of most healthy individuals . The maintenance of colonization is the result of a complex balance between fungal proliferation and host immune recognition . Despite host immune defenses aimed at clearing pathogens , C . albicans has developed numerous strategies to evade host immune detection [1] . In immunocompromised patients , C . albicans may disseminate into bloodstream , causing life-threatening systemic candidiasis [2 , 3] . The associated mortality rates of systemic infection are reported to be greater than 30% , highlighting the potential critical impact of C . albicans on global health burden [4–6] . The mature cell wall of C . albicans is a complex structure of cross-linked polysaccharides and glycosylated proteins . The cell wall is not only required for maintaining cell shape and stability , but also is critically related to immunogenicity and virulence of C . albicans . The outer layer of the cell wall is comprised of glycosylated mannoproteins that are post-translationally modified by N- and O-linked mannosides [7] . N-linked mannan chains are specifically required for cell morphology , phagocytosis , and immune recognition of C . albicans by host dendritic cells [8] . The core structure of N-mannan is a dolichol pyrophosphate anchored oligosaccharide comprised of three glucose , nine mannose and two N-acetylglucosamine residues ( Glc3Man9GlcNAc2 ) . The outer chain branched mannan is attached to the N-mannan core through an α-1 , 6-backbone . Addition of the first α-1 , 6-mannose is catalyzed by mannosyltransferase Och1 . Notably , Och1 mutant strains of C . albicans demonstrate attenuated virulence in animal models with systemic infection [9] . Extension of α-1 , 6-mannose backbone by mannose residues is performed by the enzyme complexes mannan polymerase I ( M-Pol I ) and II ( M-Pol II ) [10] . The α-1 , 6-backbone is then further modified with additional α-1 , 2-mannose units by Mnn2 family and Mnn5 , which similarly , are critical for C . albicans virulence in mice or Galleria mellonella [11 , 12] . The outer side chains are further capped with either α-1 , 3-mannose or β-1 , 2-mannose units via Mnn1 family and β-1 , 2-mannosyltransferases ( BMTs ) . The C . albicans MNN1 gene family contains six members , of which only MNN14 represent a critical factor for pathogenicity in vivo [13] . Bmt1 and Bmt3 , which are required for the addition of the first and second β-1 , 2-mannose units respectively , are not associated with the virulence of C . albicans [14] . Although a variety of C . albicans mannosylation mutants have been found to be less pathogenic in vivo , the mechanisms of host clearance associated with abnormal mannan structures remains unclear . The cell wall polysaccharides of C . albicans are mainly composed of multiple layers of carbohydrates , including mannans , β-glucans , and chitins [3] . These polysaccharides serve as pathogen-associated molecular patterns ( PAMPs ) that can be recognized by host-expressed pattern recognition receptors ( PRRs ) to initiate an innate immune response [1] . Several PRRs , such as toll-like receptors ( TLRs ) , spleen tyrosine kinase ( Syk ) -coupled C-type lectin receptors ( CLRs ) , and nucleotide binding oligomerization domain ( Nod ) -like receptors ( NLRs ) , can recognize PAMPs on the surface of C . albicans [15–17] . The PRRs engagement by C . albicans PAMPs triggers innate immune cells to respond and renders antigen-presenting cells competent to prime T cells . A complex signaling cascades , including nuclear factor-κB ( NF-κB ) and mitogen-activated protein kinase ( MAPK ) pathways , among others , lead to Th1 and Th17 activation and an adaptive immune response [18–21] . Dectin-1 , a myeloid-expressed Syk-coupled receptor , can recognize β- ( 1 , 3 ) -glucan carbohydrates on the surface of various fungi [22–24] . Clinical studies have demonstrated that patients with Dectin-1 Y238X mutation are highly susceptible to mucosal C . albicans infection [25] . However , live C . albicans , including yeast and hyphae forms , binds to Dectin-1 in highly specialized pattern in vitro , except in the region between the parental yeast cell and the mature bud [26] . During infection , β- ( 1 , 3 ) -glucan of C . albicans is completely masked in earlier stages , while large percentages are exposed at later stages in a morphotype-independent fashion [27] . Shielding of β- ( 1 , 3 ) -glucan favors fungal survival and persistence by escaping Dectin-1 mediated immune recognition [28] . Previous studies have indicated that unmasking C . albicans β- ( 1 , 3 ) -glucan elicits a stronger host immune response towards C . albicans via several experimental manipulations such as drug treatment and several key genes deletion [29–31] . Mnn10 , an important subunit of cis Golgi mannan polymerase , was identified as an α-1 , 6-mannosyltransferase which is responsible for mannan backbone extension in non-pathogenic fungal species such as Saccharomyces cerevisiae and Kluyveromyces lactis [32 , 33] . In the present study , we first characterized the role of α-1 , 6-mannose backbone in C . albicans pathogenicity . We demonstrated that inhibition of α-1 , 6-mannose backbone extension can block the development of invasive C . albicans infection , and suggested α-1 , 6-mannose backbone extension is essential for the evasion of host Dectin-1 mediated immune response towards C . albicans .
To analyze enzymatic activity , we used a bacterial expression system to produce MBP-fused Mnn10 protein from a pMAL-p5X vector and purified the protein by amylose magnetic beads ( S1 Fig ) . The purified proteins were subjected to a mannosyltransferase assay system , which required α-1 , 6-linked mannobiose as an acceptor and GDP-mannose as a donor for proper mannosyltransferase activity [12 , 34] . Our results indicated that Mnn10 catalyzes the transformation of GDP-mannose to α-1 , 6-mannobiose to form mannotriose and mannotetraose ( Fig 1A ) . Consistently , the reaction products could be digested to mannose via α-1 , 6-mannosidase treatment ( Fig 1B ) . Thus our results demonstrate that Mnn10 is able to transfer mannose from the donor onto the acceptor substrates to form α-1 , 6-linked oligomannose . To further confirm the role of MNN10 in α-1 , 6-mannose backbone extension , we generated mnn10Δ/Δ null mutant strain and mnn10Δ/Δ::MNN10 revertant strain using the homologous recombination method . The genotype was confirmed by PCR and the expression of MNN10 was determined by quantitative RT-PCR ( S2 Fig ) . When compared to the parental strain SN152 , transmission electron microscopy ( TEM ) analysis noted a significantly shortened external layer of mannan fibril surrounding the cell wall in mnn10Δ/Δ . Notably , MNN10 gene rescue was sufficient to restore the expected length of mannan fibril ( Fig 1C ) . The phosphomannan of C . albicans cell wall , characterized by Alcian Blue dye binding , is attached to the branched mannan from the α-1 , 6-mannose backbone [35] . Therefore , the content of phosphomannan reveals the length of α-1 , 6-mannose backbone . The Alcian Blue assay demonstrated significantly decreased binding of the dye to mnn10Δ/Δ as compared to the parental strain , which could be rescued by reintegration of MNN10 into mnn10Δ/Δ , confirming the role of MNN10 in α-1 , 6-mannose backbone extension ( Fig 1D ) . Inhibition of α-1 , 6-mannose backbone extension via deletion of MNN10 also affected the surface property of C . albicans such as hydrophobicity ( Fig 1E ) . Furthermore , the change of cell wall structures impaired the resistance of mnn10Δ/Δ to various stresses including calcofluor white , fluconazole , miconazole , and caspofungin ( S3 Fig ) . Taken together , our results suggest that Mnn10 protein possesses α-1 , 6-mannosyltransferase activity and is crucial in α-1 , 6-mannose backbone extension in C . albicans . The cell wall polysaccharides of C . albicans consist of an inner layer of chitin and β-glucan , and an outer fibrillar layer of mannan . To examine the effects of MNN10 deletion on cell wall polysaccharides biosynthesis and organization , C . albicans yeasts or hyphae were stained with Concanavalin A ( ConA ) , anti-β-glucan antibody and calcofluor white ( CFW ) . We observed and quantified the polysaccharides by confocal laser scanning microscopy and flow cytometry , respectively . ConA staining indicated that mnn10 mutant , either in yeast cells or hyphae , had a markedly decreased fluorescence intensity which is suggested lower mannose content as compared to the parental or revertant strain ( Figs 2A , 2D and S4A ) . β- ( 1 , 3 ) -glucan , a well-characterized PAMP of C . albicans , is buried underneath the outer layer of the cell wall . We found a remarkable exposure of β- ( 1 , 3 ) -glucan on the cell surface of mnn10 mutant ( Figs 2B , 2E and S4B ) . However , no significant difference of CFW binding was visualized in mnn10 mutant strain , suggesting normal chitin content ( Fig 2C and 2F ) . The highly glycosylated cell wall proteins ( CWPs ) of C . albicans often act as virulence factors and contribute to cell wall integrity , promote biofilm formation , mediate adherence to host cells , and promote invasion of epithelial layers [36–38] . To further evaluate the effect of α-1 , 6-mannose extension in C . albicans CWPs anchorage , we analyzed CWPs extraction by LC-MS/MS . Twenty representative CWPs were identified in the cell wall of the parental and mnn10 mutant strain ( S1 Table ) . We found that MNN10 deletion had almost no effect on the anchorage of CWPs . Notably several of the proteins are considered to mediate specific roles in the development of invasion , such as cell-surface antigens involved in virulence ( Tdh3p , Eno1p , Met6p , Hsp70p ) , proteins involved in cell wall biosynthesis and assembly ( Pga4p , Utr2p , Crh11p ) , and proteins involved in adhesions and cell wall morphogenesis ( Sap9p , Pga24p , Ecm33p ) ( Fig 2G ) . Taken together , our results demonstrate that inhibition of α-1 , 6-mannose backbone extension by MNN10 deletion resulted in abnormal cell wall polysaccharides biosynthesis and organization , but had no effect on the anchorage of CWPs including virulence factors . The composition and organization of the cell wall in C . albicans plays an important role in the initiation and maintenance of invasive infections . In the context of MNN10 deletion , we investigated the virulence of mnn10 mutant strain in vitro . Multiplication and yeast-to-hypha transition are the prerequisites of C . albicans invasive disease [39] . The growth curves obtained demonstrated that MNN10 did not impact the growth of C . albicans in vitro ( Fig 3A ) . Moreover , mnn10 mutant also did not show defective filamentation in either liquid ( RPMI1640 , 10% serum liquid medium ) or solid medium ( Spider and Lee’s agar medium ) favoring hyphal growth ( Figs 3B and S5A ) . Adhesion to the epithelium , such as oral and intestinal epithelial cells , and hyphal penetration are the first steps of C . albicans invasion [40] . Our results indicate that deletion of MNN10 did not affect the ability of C . albicans to adhere Caco-2 intestinal epithelial and KB buccal epithelial cells ( P > 0 . 05 , S5B Fig ) . Following adhesion , we used scanning electron microscopy ( SEM ) to evaluate the ability of hyphal penetration and invasion of mnn10 mutant strain . After a 2 h co-culture , we observed hyphal forms , of all the indicated strains , penetrated Caco-2 and KB cells at the apical face and microvillis attached to the hyphae at the point of penetration , indicating that MNN10 gene deletion had no effect on the invasion ability of C . albicans in vitro ( Fig 3C ) . Moreover , no significant difference , in epithelial cell damage , was observed among strains ( P > 0 . 05 , Fig 3D ) . Phospholipases and hemolysins secreted by C . albicans can also induce host cells damage [41 , 42] . We screened extracellular phospholipase and hemolytic activity by growing C . albicans strains on either egg yolk agar or sheep blood agar . As compared to the parental and revertant strains , mnn10 mutant induced similar zones of precipitation or clearance around the colonies , indicating that MNN10 deletion was not significantly associated with phospholipases and hemolysins secretion ( Fig 3E ) . As such , our results suggest that MNN10 deletion did not affect the invasive capacity of C . albicans in vitro . To examine the effects of α-1 , 6-mannose backbone of C . albicans on host infection in vivo , we compared the pathogenicity of SN152 , mnn10Δ/Δ::MNN10 and mnn10Δ/Δ in a murine systemic candidiasis model . Mice infected with mnn10 mutant strain had a much higher survival rate than those infected with the parental or revertant strain at a lethal dose ( 5×105 CFU ) . Over a 30-day observation period , only one mouse infected with the mnn10 mutant strain died . By contrast , all of the mice infected with the parental or revertant strain died within 30 days ( P < 0 . 01 , Fig 4A ) . The median survival analysis demonstrated that the 50% survival limit was attained at 10 days for mice infected with the parental strain and 18 days for the revertant strain , respectively . At day 2 or day 5 post-infection , mice infected with mnn10 mutant had significantly lower fungal burdens in the kidneys and livers as compared to those infected with either the parental or revertant strain ( P < 0 . 01 and P < 0 . 05 , respectively; Figs 4B and S6A ) . Moreover , Hematoxylin and eosin ( H&E ) staining revealed that during prolonged infections , inflammatory influx and tissue necrosis of the kidneys were aggravated in mice infected with either parental or MNN10 revertant strains ( Fig 4C , top panel ) . Periodic acid-Schiff ( PAS ) staining also identified more hyphae in the kidneys of parental or revertant strain infected mice as compared to the mnn10 mutant strain ( Fig 4C , bottom panel ) . After normalized to organ CFU burden of infection , IL-6 , GM-CSF , IFN-γ and IL-17 in the kidneys of mnn10 mutant infected mice were markedly higher than mice infected with parental or revertant strains ( P < 0 . 01 , Fig 4D ) ( the actual cytokine values were shown in S7 Fig ) . Furthermore , we performed a flow cytometry analysis to detect cellular inflammation in the kidneys of infected mice . Time-course analysis revealed that mnn10Δ/Δ infected mice recruited more SSChighCD11b+Ly-6C+Ly-6G+ neutrophils and SSChighCD11b+Ly-6C+Ly-6G- monocytes in the kidney than mice infected with the SN152 strain at day 2 and day 3 , respectively ( Figs 4E and S8 ) . The cellular inflammation in the kidneys of mnn10Δ/Δ infected mice reached a peak at day 3 post-infection , when compared to SN152 infected mice . With decreased fungal burden , neutrophils and monocytes were reduced in the kidney of mnn10Δ/Δ infected mice at day 5 post-infection as compared with SN152 infected mice ( Fig 4E ) . Taken together , the results suggest that inhibition of α-1 , 6-mannose extension by MNN10 deletion significantly impacted the pathogenesis of C . albicans by enhancing the host antifungal defense in vivo . To further explore whether host antifungal defense was crucial for the clearance of mnn10 mutant in vivo , we investigated the pathogenicity of mnn10 mutant strain in BALB/c mice and athymic nude mice ( BALB/c background ) . BALB/c mice infected with mnn10 mutant displayed significantly lower fungal burdens in the kidneys and livers as compared to those infected with either the parental or revertant strain ( P < 0 . 05 and P < 0 . 01 , respectively; Fig 5A ) . However , the results of kidneys and livers fungal burdens indicated mnn10 was not required for the C . albicans pathogenesis in athymic nude mice ( Fig 5B ) . To expand upon these findings , longer time course and survival experiments were performed . At day 10 post-infection , no significant difference in the levels of fungal burden of kidneys was observed between mice infected with mnn10 mutant strain versus those infected with the parental or revertant strains ( Fig 5C ) . Furthermore , there was no significant survival difference among these strains infected mice ( Fig 5D ) . To confirm the role of elevated IFN-γ and IL-17 in the infected mice ( Fig 4D ) in clearing mnn10 mutant strain , we performed an adoptive immunotherapy experiment . The combination of IFN-γ and IL-17 treatment significantly improved the survival of athymic nude mice infected with mnn10Δ/Δ ( P < 0 . 01 , Fig 5E ) . Moreover , mnn10 mutant infected mice treated with cytokines had a markedly higher survival rate than SN152 infected mice with the same treatment modality ( P < 0 . 05 , Fig 5E ) . Furthermore , we used neutralizing antibodies to IL-17 and/or IFN-γ to elucidate the importance of these cytokines in C57BL/6 mice infected with mnn10Δ/Δ . Compared with mice treated with isotype antibody rat IgG1 , mice receiving anti-IFN-γ , anti-IL-17A , or both antibodies exhibited significantly higher fungal burdens in the kidneys and livers ( S9 Fig ) ( P < 0 . 05 ) . These results suggest that IFN-γ and IL-17 played an important role in the enhanced host defense against mnn10 mutant strain in vivo . Our results suggested that the enhanced host antifungal immunity was the main factor that contributed to the diminished virulence of mnn10 mutant strain . Thus , we further investigated the myeloid cell recognition and response to mnn10 mutant using a macrophages-C . albicans interaction model . We found that mnn10 mutant yeast cells induced more nuclear translocation of NF-κB ( p65 ) , phosphorylation of Syk and IκBα , together with IκBα degradation in thioglycolate-elicited peritoneal macrophages ( Fig 6A ) . Consistently , mnn10 mutant yeast cells also induced more ERK phosphorylation , p38 phosphorylation , and JNK phosphorylation in macrophages than the parent parental or revertant strains ( Fig 6C ) . We also detected significantly higher levels of inflammatory cytokines including TNF-α , IL-6 , IL-1β and IL-12p40 in macrophages induced by mnn10Δ/Δ yeast cells ( Fig 6E ) . However , no differences in NF-κB and MAPK signaling activation were observed in macrophages stimulated by hyphal forms of the parental , mnn10 mutant and revertant strains , the same as the production of proinflammatory cytokines ( Figs 6B , 6D and S10 ) . We performed a macrophage phagocytosis assay to investigate whether mnn10 mutant was differentially taken up from the parental or revertant strain . Our results suggest that no significant difference of the macrophage phagocytosis was visually appreciated between mnn10 mutant , and the parental or revertant strains in the initial infection stage ( Figs 6F and S11 ) . However , we found that thioglycollate-elicited peritoneal neutrophils could produce significantly more ROS and thus destroy mnn10Δ/Δ more efficiently ( Fig 7A and 7B ) . While neutrophils can target pathogens in modalities such as myeloperoxidase ( MPO ) , no significant difference in the intracellular MPO activity of neutrophils stimulated by mnn10 mutant versus parental or revertant strain was observed ( Fig 7C ) . To determine whether the enhanced killing was dependent on ROS production , we scavenged ROS by 2 mM L-ascorbic acid in the co-culture medium , and found that the ability of neutrophils to eliminate mnn10Δ/Δ was diminished with decreased ROS ( Fig 7D and 7E ) . However , we found that removal of mnn10 mutant hyphae by neutrophils was similar in parental or revertant strains ( Fig 7F ) . We further explored how enhanced leukocytes recognition of mnn10 mutant strain affected inflammation response in vivo using a peritoneal infection model . Mice were injected intraperitoneally with C . albicans , and flow cytometry performed 4 h later revealed that mnn10Δ/Δ infected mice recruited more inflammatory cells in the peritoneum than SN152 infected mice , including SSChighCD11b+Ly-6C+Ly-6G+ neutrophils , SSChighCD11b+Ly-6C+Ly-6G- monocytes and SSChighCD11b+Siglec-F+ eosinophils ( Fig 8A and 8B ) . The enhanced inflammatory cell recruitment was also associated with increased production of specific cytokines and growth factors such as IL-6 , MCP-1 , MIP-1α , G-CSF and GM-CSF ( Fig 8C ) . However , this analysis failed to reveal significant difference in the inflammatory cells , including CD3-NK1 . 1+ NK cells , CD3+NK1 . 1+ NKT cells , and CD3+γ/δ T+ cells , between mnn10Δ/Δ and SN152 infected mice ( Fig 8D and 8E ) . Several PRRs , such as TLRs and CLRs , are involved in host defense during C . albicans infection . We hypothesized that the enhanced host immune response induced by mnn10 mutant may be attributed to the cell wall β- ( 1 , 3 ) -glucan exposure . We stimulated thioglycollate-elicited peritoneal macrophages from wild-type or Dectin-1-deficient mice with mnn10 mutant yeasts , and found that activation of NF-κB and MAPK signaling was defective in Dectin-1-deficient macrophages ( Fig 9A ) . Consequently , the mnn10 mutant yeasts could not significantly increase the production of inflammatory cytokines such as TNF-α and IL-6 in Dectin-1-deficient macrophage cells ( Fig 9B ) . The removal of mnn10 mutant by neutrophils from Dectin-1-deficient mice was similar in parental or revertant strains ( S12A Fig ) . Moreover , no significant difference in the inflammatory cytokines , including IL-6 , GM-CSF , IFN-γ and IL-17 , were detected between the kidneys of mnn10Δ/Δ and SN152 infected Dectin-1-deficient mice ( Figs 9C and S12B ) . The survival curves indicated that mnn10 mutant strain presented similar pathogenicity with the parental strain SN152 in Dectin-1-deficient mice ( Fig 9D ) . Dectin-1-deficient mice had similar kidney or liver fungal burdens when infected with the parental SN152 or mnn10Δ/Δ strain ( P > 0 . 05 , Figs 9E and S6B ) . However , our results demonstrated that other PRRs involved in antifungal immunity did not contribute to the enhanced immune responses elicited by mnn10 mutant . By example , mnn10 mutant strain presented similar pathogenicity in Dectin-2 deficient mice when compared to SN152 in the wild type mice ( P < 0 . 05 , Figs 9F and S6C ) . Dectin-2 deficiency had no effect on the inflammatory cytokines production such as TNF-α and IL-6 in macrophages , and TLR2 or TLR4 deficiency had no effect on the activation of NF-κB and MAPK signaling in macrophages when challenged with mnn10 mutant ( Fig 9G and 9H ) . Taken together , these data suggest that the enhanced immune response induced by α-1 , 6-mannose backbone inhibition in C . albicans was Dectin-1 dependent .
During C . albicans infection , both yeast cells and hyphae can be found in infected organs or tissues , and innate immune cells discriminate them using different PRRs to elicit a protective immune response [43] . Previous studies have shown that the mannan structure of C . albicans plays an important role in the development of invasive infection . Here we first determined that the cell wall α-1 , 6-mannose backbone maintained the pathogenicity of C . albicans by preventing host , Dectin-1 mediated , recognition of β- ( 1 , 3 ) -glucan . These results highlight a previously unappreciated relationship between cell wall mannan structure and pathogenicity of C . albicans . Several genes are involved in the biosynthesis of the cell wall mannan in C . albicans . Cell wall mannan structure mutant strains , induced by deletion of several genes including OCH1 , MNN2 and MNN5 , often represent a less pathogenic strain in vivo [9 , 11 , 12] . Herein , we determined that Mnn10 has α-1 , 6-mannosyltransferase activity , and is responsible for α-1 , 6-mannose backbone biosynthesis in C . albicans ( Fig 1 ) . We also highlighted the role of MNN10 in pathogenicity of C . albicans . Our studies using mnn10 null mutant strain demonstrated that MNN10 is required for C . albicans pathogenicity during a systemic candidiasis model in mice ( Fig 4 ) . Infection is mediated by the interplay between a pathogen’s ability to invade host , versus the host attempts to recognize and destroy the pathogen . Several cell wall proteins on the surface of C . albicans act as virulence factors to invade host [39 , 44] . Cell wall proteomic analysis indicated that MNN10 deletion had no effect on virulence factors of C . albicans ( Fig 2G ) . Several steps , including adhesion to the epithelium , epithelium penetration and invasion by hyphae , vascular dissemination , and endothelial colonization , are involved in the development of invasive candidiasis [2] . However , the data from our study in vitro indicated that the mnn10 mutant was not defective in its invasive capacity ( Fig 3 ) . Therefore , we hypothesize enhanced immune recognition of the mutant strain by the host , rather than decreased virulence , contributed to the attenuated pathogenicity . The normal pathogenicity of mnn10 mutant in athymic nude mice ( BALB/c background ) further confirmed our hypothesis ( Fig 5B , 5C and 5D ) . The difference of mnn10 mutant clearance in BALB/c mice versus athymic nude mice may be attributed to thymus , which is an important organ for the differentiation and maturation of T lymphocytes . Both Th1 and Th17 cells mediate host protection against C . albicans infection [19 , 45] . IFN-γ and IL-17 are the key cytokines produced by Th1 and Th17 cells , which recruit neutrophils and macrophages to destroy the pathogen . IFN-γ and IL-17 elevation , and corresponding neutrophil responses were observed in the kidneys of mnn10 mutant infected mice , indicating that the mnn10 mutant could stimulate stronger antifungal response ( Fig 4D and 4E ) . Although the source of IFN-γ and IL-17 can be from innate lymphocytes , our results suggest that elevated levels of IFN-γ and IL-17 elicited by mnn10 mutant in vivo are likely not derived from innate lymphocytes such as NK cells , NKT cells and γ/δ T cells ( Fig 8D and 8E ) . Intracellular cytokine staining analysis revealed that mnn10Δ/Δ infected mice induced more IFN-γ-producing and IL-17A-producing α/β T cells than SN152 infected mice ( S13 Fig ) . Therefore , our study suggested that inhibition of α-1 , 6-mannose backbone extension in C . albicans induced enhanced T lymphocyte mediated immune response in vivo . The host innate immune cells involved in invading pathogens recognition are predominantly monocytes and neutrophils in circulation and macrophages in infected tissues . Inflammatory cytokines and chemokines can recruit innate immune cells to infected tissues . In a peritoneal infection model , we demonstrated that mnn10 mutant strain could recruit more neutrophils and monocytes by inducing cytokines and chemokines including IL-6 , MCP-1 , GM-CSF , MIP-1α and G-CSF in the peritoneal cavity ( Fig 8A , 8B and 8C ) . IL-6 and G-CSF can promote neutrophil production and activation against C . albicans infection [46] . GM-CSF and MIP-1α are involved in potentiating neutrophil functions and maturation [47 , 48] . MCP-1 is a crucial mediator to recruit monocytes in inflammation in vivo [49] . Neutrophils contribute to the initial step to kill fungi , and are especially important in neutropenic and immunosuppressed individuals [50] . Our data also determined that neutrophils mainly eliminated mnn10 mutant strain in a ROS-dependent manner ( Fig 7 ) . These results suggest that inhibition of C . albicans α-1 , 6-mannose backbone extension by MNN10 deletion could enhance host innate immune recognition . Immune recognition could render antigen presenting cells competent to prime T cells , and thereby drive the adaptive Th1 and Th17 immune response . After encountering pathogens , the host macrophages secrete several cytokines , leading to the induction of Th cell differentiation [3] . Our study suggests that C . albicans mnn10 did not play an important role in the initial phagocytosis stage of macrophages ( Figs 6F and S11 ) . By contrast , mnn10 mutant yeast of S . cerevisiae was poorly taken up by primary macrophages , as compared to the parental strain [51] . We hypothesize that the differential macrophage phagocytosis of mnn10 mutant might be attributed to the pathogenicity and immunogenicity differences between C . albicans and S . cerevisiae . However , we demonstrated that mnn10 mutant strain elicited enhanced recognition by macrophages . The increased cellular responses of macrophages were associated with NF-κB and MAPK pathway activation , and inflammatory cytokine productions including TNF-α , IL-6 , IL-1β and IL-12p40 ( Fig 6A , 6C and 6E ) . TNF-α was involved in the innate immune response against Candida infection through promotion of neutrophil production and activation [52] . IL-6 and IL-23 ( consisting of IL-12p40 and p19 ) contributed to Th17 differentiation induced by C . albicans and Staphylococcus aureus , and IL-1β was essential pro-inflammatory regulators of Th17 cells both at priming and effect phase [53] . Therefore , we suggest that the enhanced recognition by innate immune cells could promote Th cell response and thus contributes to host clearance of mnn10 mutant strain in vivo . The special PAMPs on the surface of C . albicans could be recognized by PRRs of innate immune cells to initiate the host immunity . The skeletal component of C . albicans cell wall is based on a core structure of β- ( 1 , 3 ) -glucan that is covalently linked to β- ( 1 , 6 ) -glucan , chitin , and an outer layer of mannoproteins [3] . Recognition of β- ( 1 , 3 ) -glucan by Dectin-1 has been reported to be important in host antifungal defense [22] . However , β- ( 1 , 3 ) -glucan on the surface of C . albicans was normally shielded by the outer mannan layer from being recognized by Dectin-1 on innate immune cells [2 , 27] . Deletion of certain C . albicans genes , such as the phospholipids phosphatidylserine synthase gene CHO1 and the histidine kinase gene CHK1 , could unmask β-glucans of C . albicans , specifically recognized by Dectin-1 and leading to more host immune responses [30 , 31] . Antifungal compounds , such as caspofungin and gepinacin , can also cause the exposure of β- ( 1 , 3 ) -glucan in C . albicans and elicit a stronger host immune response [29 , 54] . Our results demonstrate that inhibition of α-1 , 6-mannose backbone extension by MNN10 deletion could unmask the concealed β- ( 1 , 3 ) -glucan in either yeast or hyphal form ( Figs 2B , 2E and S4B ) . The exposure of β- ( 1 , 3 ) -glucan may be due to the fact that the outer structure of cell surface do not adequately conceal the inner layer . The enhanced inflammatory responses stimulated by mnn10 mutant , including inflammatory signaling activation and cytokine secretion , were markedly down-regulated in Dectin-1-deficient macrophages , suggesting that they were Dectin-1 dependent ( Fig 9A and 9B ) . Moreover , our study indicates that mnn10 mutant restored its pathogenicity in Dectin-1-deficient mice , further confirming our hypothesis ( Fig 9D and 9E ) . While mnn10 mutant strain also stimulated enhanced response of macrophage cells from TLR-2 , TLR-4 deficient mice or showed less pathogenicity in Dectin-2 deficient mice , indicating the enhanced inflammatory responses induced by mnn10 mutant strain had little associations with other PRRs involved in host antifungal defense ( Fig 9F , 9G and 9H ) . We concluded that enhanced Dectin-1 dependent immune recognition of mnn10 mutant strain could induce increased inflammatory cytokines secretion in antigen presenting cells such as macrophages , which further regulated Th1 and Th17 cell differentiation . IFN-γ secreted from Th1 lymphocytes and IL-17 secreted from Th17 lymphocytes ultimately mediated the clearance of mnn10 mutant strain in vivo . Previous studies have reported that deficiency of Och1 or Mnn2 family involved in mannan biosynthesis of C . albicans could also lead to β-glucan exposure and decreased mannan [9 , 11] . However , Och1 and Mnn2 mutant elicited a host reduced immune response . As we know , PRRs can bind to short oligosaccharides and the precise carbohydrate epitopes to elicit antifungal immunity . We believe these discrepancies may be attributed to the differential effect of these genes on the mannan structure and the precise exposed carbohydrate epitopes of β-glucan . In addition , these previous studies concluded that decreased mannan , rather than β-glucan exposure , was the major PAMP recognized by host immune system in Och1 and Mnn2 mutant [9 , 11] . The present study not only indicates that deletion of MNN10 in C . albicans results in decreased mannose content and a more remarkable exposure of β- ( 1 , 3 ) -glucan on the cell surface , but also suggests that β- ( 1 , 3 ) -glucan of mnn10 mutant strain is the major PAMP and induced enhanced Dectin-1 dependent immune response . In conclusion , we first identified that Mnn10 as an α-1 , 6-mannosyltransferase , which is involved in the cell wall α-1 , 6-mannose backbone extension and maintained pathogenicity of C . albicans by evading host Dectin-1 mediated antifungal immunity . In addition , there are no mammalian homologs of Mnn10 protein , thus our results provide a new potential antifungal therapeutic strategy for modulating the host immune response to C . albicans .
All mouse experimental procedures were performed in accordance with the Regulations for the Administration of Affairs Concerning Experimental Animals approved by the State Council of People’s Republic of China . The protocol was approved by the Institutional Animal Care and Use Committee of Tongji University ( Permit Number: TJLAC-014-013 ) . Female C57BL/6 mice , BALB/c nude mice and athymic nude mice ( BALB/c background ) were obtained from Shanghai Laboratory Animal Center of the Chinese Academy of Sciences . TLR2 deficient and TLR4 deficient mice ( C57BL/6 background ) were purchased from Shanghai Biomodel Organism Science & Technology Development Company . Dectin-1-deficient ( Clec7a-/- ) mice were kindly provided by Dr . Gordon D . Brown ( the mice were backcrossed for nine generations on the C57BL/6 background ) and Dectin-2-deficient ( Clec4n-/- ) mice were kindly provided by Dr . Yoichiro Iwakura ( C57BL/6 background ) [55 , 56] . KB buccal epithelial carcinoma cell line and Caco-2 intestinal epithelial cell line were purchased from BIOK&KM . Alcian Blue , ANTS , HF-pyridine , percoll , calcofluor white ( CFW ) , fluorescein isothiocyanate-conjugated pisum sativum agglutinin ( PSA-FITC ) and dihydrorhodamine 123 were purchased from Sigma-Aldrich . Antibodies against phospho-ERK , p38 , phospho-p38 , JNK , phospho-JNK , phospho-IκBα , Syk , phospho-Syk , p65 , and PCNA were purchased from Cell Signaling Technologies . Antibodies against ERK , IκBα and β-actin were from Santa Cruz Biotechnology . Alexa-488-labeled and Cy3-labeled goat anti-mouse antibodies were purchased from Life Technologies . Antibody against β- ( 1 , 3 ) -glucan was purchased from Biosupplies Inc . The following antibodies , along with the appropriate isotype controls were used in flow cytometry: peridinin-chlorophyll-protein-complex anti-Ly-6C ( clone HK1 . 4 , Biolegend ) , Alexa Fluor 647 Siglec-F ( clone E50-2440 , BD Pharmingen ) , phycoerythrin-Cy7-conjugated anti-CD11b ( clone M1/70 , Biolegend ) , phycoerythrin-conjugated anti-Ly-6G ( clone 1A8 , Biolegend ) , fluorescein isothiocyanate-conjugated anti-CD3ε ( clone 145-2C11 , Biolegend ) , allophycocyanin anti-NK-1 . 1 ( clone PK136 , Biolegend ) and allophycocyanin anti-TCR γ/δ ( clone GL3 , Biolegend ) . All strains were maintained on SDA agar plates ( 1% peptone , 4% dextrose , and 1 . 8% agar ) and grown in YPD broth ( 1% yeast extract , 2% peptone , and 2% dextrose ) at 30°C . For hyphal growth , C . albicans yeast cells were cultured in RPMI 1640 medium [10 . 4 g RPMI 1640 ( Gibco BRL ) , 34 . 5 g morpholinepropanesulfonic acid ( Sigma ) , and 2 . 0 g NaHCO3 , pH 7 . 0 , in 1 liter double-distilled water sterilized by filtration] at 37°C for 3 h . To construct MNN10 null mutant strain ( mnn10Δ/Δ ) , the entire open reading frame of MNN10 was deleted from the parental strain SN152 by homologous recombination of auxotrophic markers HIS1 and LEU2 using a fusion-PCR-based strategy as previously described [57 , 58] . To construct MNN10 revertant strain ( mnn10Δ/Δ::MNN10 ) , the fusion fragment containing MNN10 ORF and C . albicans SAT1-flipper cassette was transformed into mnn10Δ/Δ and the SAT marker was subsequently looped out as described previously [59] . All of the strains and the primers used in this study were listed in S2 and S3 Tables . The non-transmembrane region of C . albicans MNN10 encoding amino acid residues 70 to 335 was cloned into pMAL-p5X ( NEB ) including MBP tag . And then the plasmid was transformed into BL21 ( DE3 ) pLysS cells for expressing MBP-fused Mnn10 protein . The transformants were cultured overnight at 37°C and diluted 1:100 in fresh LB culture . When the medium OD600 was up to 0 . 6 at 37°C , IPTG at a final concentration of 0 . 1 mM was added and the cells were grown overnight at 16°C . MBP-fused Mnn10 protein was purified by amylose resin ( NEB ) according to the protocols as previously described [60] . The supernatant was passed through a 0 . 45 μm filter and bound to amylose resin by gravity flow . Unspecific proteins were washed off by applying 10 column volumes ( CVs ) of column buffer . The protein of interest was then eluted by elute buffer ( column buffer added with 10 mM maltose ) . The eluate was finally dialyzed against buffer ( 25 mM Tris-HCl , pH7 . 5 ) for 2 h at room temperature . The mannosyltransferase activity assay was performed according to the method as described previously [61 , 62] . The purified MBP-fused Mnn10 protein ( 8 μg ) was subjected to the buffer ( 10 mM MnCl2 , 50 mM HEPES , 10 mM α-1 , 6-D-mannobiose and 5 mM GDP-D-Mannose , pH 7 . 2 ) . The α-1 , 6-mannose extended reaction was incubated at 30°C for 1 h . After labeled with 0 . 75 μmol 8-Aminonaphthalene-1 , 3 , 6-trisulfonic acid ( ANTS ) , the products were separated by electrophoresis on 40% polyacrylamide gels and detected by UV light on a transilluminator ( 3500 R , Tanon ) . The reaction products were then digested with α-1 , 6-mannosidase ( NEB ) overnight at 37°C according to manufacturer’s instructions , and analyzed by FACE . The Alcian Blue binding assay was carried out as described previously [35] . 1 . 5×107 exponentially growing C . albicans cells were washed with 0 . 02 M HCl , and then incubated with 30 μg/ml Alcian Blue for 10 min at room temperature . The supernatant was measured at OD600 and the concentration of Alcian Blue was determined by reference to a standard curve . The amount of dye bound to C . albicans cells were calculated by subtracting the amount of dye in the supernatant . Exponentially growing C . albicans cells were washed and resuspended in PBS buffer ( OD600 = 1 . 0 ) , and 0 . 75 ml cyclohexane was then added to the above 3 ml cell suspension . The mixtures were vortexed for 3 min and settled for 20 min at room temperature , the OD600 of the aqueous phase was measured . The relative hydrophobicity was measured as [ ( OD600 of the control minus OD600 after octane overlay ) /OD600 of the control] × 100% [63] . The covalent and non-covalent CWPs were isolated as previously described [64 , 65] . The major types of covalently linked CWPs are glycosylphosphatidylinositol anchored proteins ( GPI-APs ) and proteins with internal repeats ( Pir proteins ) . GPI-CWPs were released by resuspending the cell wall debris in undiluted HF-pyridine and incubated at 0°C for 3 h . Pir proteins were specifically released by incubating cell wall debris with 30 mM NaOH at 4°C for 16 h . The non-covalent CWPs of C . albicans were extracted by SDS buffer ( 50 mM Tris-HCl , 2% SDS , 100 mM EDTA , and 10 mM DTT , pH 8 . 0 ) . The whole CWPs were then mixed and further digested by trypsin for the analysis of LC-MS/MS on high-resolution instruments ( LTQ-Orbitrap XL and Velos , Thermo Fisher ) . Raw files were processed by MaxQuant soft for peptide/protein identification and quantification . To stain β- ( 1 , 3 ) -glucan of the cell wall , exponentially growing C . albicans yeast cells were washed in PBS ( for hyphal form assays , 1×106 C . albicans cells were cultured in RPMI 1640 medium at 37°C for 3 h on a microscope slide in a six-well plate ) , and then incubated with anti-β- ( 1 , 3 ) -glucan antibody overnight at 4°C and then stained by Cy3-labeled antibody for 1 h at 30°C . To stain mannan and chitin of the cell wall , C . albicans yeast cells or hyphae were washed in PBS and incubated in the dark with 50 μg/ml ConA to stain for α-mannopyranosyl or 30 μg/ml CFW for chitin for 30 min . The above stained cells were washed and scanned at 63 × magnification with confocal laser scanning microscope ( TCS SP5; Leica ) . Micrograph pictures were then acquired and analyzed by LAS AF Lite program . 5×107 exponentially growing C . albicans cells were washed in PBS and then fixed in 4 ml fixative solution ( 3% paraformaldehyde , 3 . 6% glutaraldehyde , pH 7 . 2 ) for 24 h at 4°C . After post-fixation of samples with 1% phosphotungstic acid for 2 h , they were washed by distilled water , block-stained with uranyl acetate , dehydrated in alcohol , immersed in propylenoxide , and embedded in glycide-ether . Ultrathin sections were observed under a transmission electron microscope ( Hitachi H-800 , Japan ) at 120 kV . To measure the growth curve of C . albicans , exponentially growing cells were washed and resuspended in fresh YPD broth ( OD600 = 0 . 1 ) , and then the optical density was determined at the indicated time point . To observe the hyphal growth , C . albicans cells were sub-cultured at 37°C in either RPMI 1640 medium plus 10% ( vol/vol ) heat-inactivated fetal calf serum ( FCS ) or spider solid medium . Phospholipase activity and hemolytic activity of C . albicans strains were screened as described previously [66 , 67] . Briefly , the suspension of yeast cells were spotted on egg yolk agar or sugar-enriched sheep blood agar and incubated at 37°C for 3 days . The phospholipase activity of each strain was observed by measuring the width of zone of precipitation around the colony . The presence of a distinct translucent halo around the colony indicated positive hemolytic activity . The Caco-2 or KB cells were grown as approximate 80%-90% confluent monolayer in MEM medium with 20% ( vol/vol ) heat-inactivated FCS . For SEM , the cells were grown on 8 mm diameter glass coverslips . Each coverslip was infected with 1×106 live C . albicans yeast cells . After 2 h of infection , the cells were gently washed with PBS prior to 1% OsO4 and then examined using a XL-30 scanning electron microscope ( Philips , Holland ) as described previously [68] . For the cell damage assay , 80%-90% confluent monolayer of Caco-2 or KB cells was infected with 1×105 live C . albicans yeast cells for 12 h , respectively . Lactate dehydrogenase ( LDH ) in the medium released from control or infected epithelial cells was determined by LDH Assay kit ( Beyotime , China ) according to the manufacturer’s instructions . Maximal LDH release was obtained by adding 0 . 1 ml of 1% Triton X-100 to each well and vigorously disrupting the epithelial layers 1 h before the end of incubation period . The relative LDH activity was measured as [ ( OD490 of infected cells minus OD490 of the control ) / ( OD490 of maximal LDH release minus OD490 of the control ) ] × 100% . Thioglycollate-elicited peritoneal macrophages and neutrophils were isolated as previously described [69] . Briefly , C57BL/6 mice were injected intraperitoneally with 2 ml 3% ( wt/vol ) thioglycollate ( Merck ) . Peritoneal cells were collected by washing with PBS containing 0 . 5 mM EDTA 14 h later and 72 h later , for neutrophils and macrophages isolation , respectively . The cells were cultured in RPMI1640 plus 10% ( vol/vol ) heat-inactivated FCS . The killing assay was carried out as described previously [70] . Thioglycollate-elicited peritoneal neutrophils were mixed with live C . albicans [multiplicity of infection ( MOI ) = 1:20] in a 24-well plate , and were kept for 1 h at 4°C to settle the cells before being transferred to 37°C for another 1 h . Control plates were kept in parallel at 4°C during the incubation . Then the cells were mixed and plated on SDA agar for counting live C . albicans colonies for 48 h at 30°C . For analysis of reactive oxygen species ( ROS ) , the inflammatory cells were co-cultured with C . albicans ( MOI = 1 ) in RPMI medium containing 10 μM dihydrorhodamine 123 for 1 h at 37°C . After incubation the fluorescent intensity of the oxidized dihydrorhodamine 123 was measured by a multi-mode microplate reader ( excitation wavelength , 485 nm; emission wavelength , 538 nm ) . Cells loaded with dihydrorhodamine 123 but not treated with C . albicans were used to assess background of ROS production . For analysis of myeloperoxidase ( MPO ) , the neutrophils were lysed by 1% Triton X-100 for 10 min and the MPO activity in neutrophil lysates was measured using an enzyme assay as described previously [71] . C . albicans yeast cells or hyphae were exposed to four doses of 100 , 000 μjoules/cm2 in a CL-1000 Ultraviolet Crosslinker ( UVP ) , with agitation between each dose to treat cells evenly [29] . The thioglycollate-elicited macrophages were stimulated with the UV-inactivated C . albicans yeasts ( MOI = 5 ) or hyphae ( MOI = 1 ) for the indicated time . Macrophage phagocytosis assay was performed as previously described , exponentially growing C . albicans cells were washed in PBS buffer and added to the monolayer macrophages ( MOI = 5 ) at the indicated time ( 30 min , 60 min , 90 min and 120 min ) . CFW staining was performed for C . albicans and PSA-FITC staining was performed for macrophages . CFW/PSA-FITC stained samples were scanned immediately at 63 × magnification with confocal laser scanning microscope . Micrograph pictures were acquired and analyzed by LAS AF Lite program . The cells were lysed in lysis buffer ( 250 mM NaCl , 50 mM HEPES , 1 mM EDTA , 1% NP-40 , protease inhibitors , pH 7 . 4 ) for total cell lysates . For nuclear extracts , cells were lysed in lysis buffer ( 10 mM KCl , 10 mM HEPES , 0 . 1 mM EDTA , 0 . 4% NP-40 , protease inhibitors , pH 7 . 9 ) . The nuclear pellets were harvested , washed with the lysis buffer and resuspended in the extraction buffer ( 20 mM HEPES , 400 mM NaCl , 1 mM EDTA , protease inhibitors , pH 7 . 9 ) , and then incubated with vortexing at 4°C for 30 minutes . The cell lysates were subjected to SDS-PAGE , blotted with the indicated primary antibodies and secondary antibodies , and then developed with the chemiluminescence method according to the manufacturer’s instructions ( Millipore ) using the ECL detection system ( GE Healthcare ) . The densitometry of indicated blot was quantified using Image J software ( National Institutes of Health , USA ) . Concentrations of tumor necrosis factor alpha ( TNF-α ) , interleukin-6 ( IL-6 ) , IL-1β and IL-12p40 in cell culture supernatant , gamma interferon ( IFN-γ ) and IL-17 in tissue homogenates , IL-6 , chemokines monocyte chemotactic protein-1 ( MCP-1 ) , macrophage inflammatory protein-1α ( MIP-1α ) , and granulocyte and granulocyte-monocyte colony-stimulating factors ( G-CSF/GM-CSF ) in peritoneal lavage fluid , were measured with commercially available Ready-Set-Go cytokine kits ( eBioscience ) or cytokine multiplex kits ( R&D systems ) in triplicate times according to the manufacturer’s instructions . For the C . albicans infection in vivo , groups of C57BL/6 female mice or BALB/c female mice ( 6–8 weeks ) were injected via lateral tail vein with 200 μl of a suspension containing indicated live C . albicans ( 5×105 cells for C57BL/6 mice and 3×105 cells for BALB/c mice ) in sterile saline . Mice were monitored daily and were killed after 2 days or 5 days of infection . The kidneys and livers were removed , and then homogenized in 0 . 5 ml PBS for fungal burdens measurement or fixed in 10% neutral formalin for H&E and PAS staining . Supernatants of kidney and livers homogenates were harvested and stored at -80°C for the measurement of cytokine production . C57BL/6 mice were injected with 5×105 CFU C . albicans and sacrificed at 12 h , 1 day , 2 days , 3 days , 4 days or 5 days post-infection and the kidneys were removed . The kidneys were minced into tissue pieces and digested for 1 h at 37°C . Then the digested tissues were passed through a 70-mm filter , washed and centrifuged in a 40%/70% percoll gradient for leukocytes isolation [72] . The leukocytes at the interphase were then analyzed by flow cytometry ( BD FACS ) . C57BL/6 mice were injected intraperitoneally with 5×105 CFU C . albicans and were killed after 4 h or 2 days . The peritoneal infiltrate was collected by lavage with ice-cold PBS containing 0 . 5 mM EDTA , and then the red blood cells were lysed . The inflammatory cells were counted and blocked with PBS containing 5% heat-inactivated FCS and 1 mM sodium azide at 4°C . The populations of the cells were analyzed by flow cytometry to determine the leukocyte composition as described before [22] . At least three biological replicates were performed for all experiments unless otherwise indicated . Log-rank test was used to evaluate the equality of survival curves . The two-tailed Student’s t-test was used for analysis of two groups and multiple groups were analyzed by one-way analysis of variance with Bonferroni post-tests . For analysis of nonparametrically distributed data , the Mann-Whitney test or Kruskal-Wallis test was used . Statistical significance was set at a p-value in the figures as: * , P < 0 . 05; ** , P < 0 . 01 .
|
Mannan plays a crucial role in cell wall structure and virulence of the opportunistic pathogen Candida albicans . Both the invasive ability of the pathogen and the host defense against the pathogen contribute to the outcome of invasive infection . In the present study , we identified a novel α-1 , 6-mannosyltransferase , which was responsible for cell wall α-1 , 6-mannose backbone extension in C . albicans . We determined that α-1 , 6-mannose backbone is necessary for the pathogenesis of C . albicans due to its ability to shield β- ( 1 , 3 ) -glucan from the host Dectin-1 recognition and Th1/Th7 response . Our study highlights a novel strategy to enhance the host immune response towards C . albicans .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"blood",
"cells",
"cell",
"walls",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"cytokines",
"pathogens",
"immunology",
"microbiology",
"mutation",
"fungi",
"developmental",
"biology",
"model",
"organisms",
"molecular",
"development",
"kidneys",
"cellular",
"structures",
"and",
"organelles",
"fungal",
"pathogens",
"neutrophils",
"research",
"and",
"analysis",
"methods",
"mycology",
"white",
"blood",
"cells",
"mutant",
"strains",
"animal",
"cells",
"medical",
"microbiology",
"microbial",
"pathogens",
"pathogenesis",
"yeast",
"immune",
"system",
"candida",
"cell",
"biology",
"anatomy",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"cellular",
"types",
"macrophages",
"renal",
"system",
"organisms",
"candida",
"albicans"
] |
2016
|
Mnn10 Maintains Pathogenicity in Candida albicans by Extending α-1,6-Mannose Backbone to Evade Host Dectin-1 Mediated Antifungal Immunity
|
The corn smut Ustilago maydis establishes a biotrophic interaction with its host plant maize . This interaction requires efficient suppression of plant immune responses , which is attributed to secreted effector proteins . Previously we identified Pep1 ( Protein essential during penetration-1 ) as a secreted effector with an essential role for U . maydis virulence . pep1 deletion mutants induce strong defense responses leading to an early block in pathogenic development of the fungus . Using cytological and functional assays we show that Pep1 functions as an inhibitor of plant peroxidases . At sites of Δpep1 mutant penetrations , H2O2 strongly accumulated in the cell walls , coinciding with a transcriptional induction of the secreted maize peroxidase POX12 . Pep1 protein effectively inhibited the peroxidase driven oxidative burst and thereby suppresses the early immune responses of maize . Moreover , Pep1 directly inhibits peroxidases in vitro in a concentration-dependent manner . Using fluorescence complementation assays , we observed a direct interaction of Pep1 and the maize peroxidase POX12 in vivo . Functional relevance of this interaction was demonstrated by partial complementation of the Δpep1 mutant defect by virus induced gene silencing of maize POX12 . We conclude that Pep1 acts as a potent suppressor of early plant defenses by inhibition of peroxidase activity . Thus , it represents a novel strategy for establishing a biotrophic interaction .
The basidiomycete smut fungus Ustilago maydis establishes a biotrophic interaction with its host plant maize which leads to the formation of plant tumors on all aerial parts of the host plant [1] , [2] . After penetration of the leaf surface , pathogenic U . maydis hyphae proliferate inside host cells that stay alive and do not show any obvious defense responses [3] . Prior to establishment of biotrophy , U . maydis infection causes a transient defense response [1] , [4] . This induction is most likely triggered by recognition of conserved pathogen-associated molecular patterns ( PAMPs ) through the maize immune system . With the onset of biotrophy 24 hours post infection ( hpi ) , defense gene expression is attenuated . In line with the model of necrotrophic pathogens inducing primarily SA-dependent cell death responses including expression of defense genes like PR1 [5] , biotrophic pathogens like U . maydis mainly induce the antagonistic JA and ethylene responses during compatible interactions [4] , [6] , [7] . Reactive oxygen species ( ROS ) are key molecules in plant defense [8]–[10] . The production of ROS is a hallmark of successful recognition of a pathogen and results in activation of plant defense responses , including oxidative burst , papilla formation , hypersensitive response ( HR ) and expression of PR genes [8] , [11]–[14] . ROS can directly act toxic at the site of infection or function indirectly as second messengers . The origin of ROS in plant defense is largely attributed to two major sources: membrane bound NADPH-oxidases and apoplastic/cell-wall associated peroxidases ( POX ) [13] , [15]–[17] . POX catalyze dehydrogenation of various phenolic and endiolic substrates by hydrogen peroxide ( H2O2 ) , resulting e . g . in the synthesis of lignin , suberin and the decomposition of IAA [8] , [18] , [19] . In addition , POX can exhibit oxidase activity , mediating the reduction of O2 to superoxide ( O2•− ) and H2O2 by substrates such as NADH or dihydroxyfumarate [8] , [18] . Moreover , in vitro studies of horseradish peroxidase showed the generation of hydroxyl radicals ( •OH ) from reduction of hydrogen peroxide [18] , [20] . The secretion of effector proteins by the pathogen that interact with targets of the host cell is a crucial aspect for the establishment of biotrophy . Effectors may mask the pathogen from recognition by the host immune system . For example , the LysM effector Ecp6 from Cladosporium fulvum sequesters chitin oligomeres originating from the fungal cell wall and therefore prevents PAMP-triggered immunity [21] . When suppressing an already triggered plant immune response , fungal effectors can either directly interact with a defense related protein or inhibit signaling pathways leading to defense responses . A direct inhibition of host defense proteins was shown for Cladosporium fulvum Avr2 that binds the host protease RCR3 and PIP2 to suppress host immunity [22]–[24] . On the other hand , the Pseudomonas syringae effector AvrPto blocks intracellular downstream signaling cascades by interfering with the flg22-triggered receptor FLS2 [25] . The effector protein AvrPiz-t , which is secreted by pathogenic hyphae of the hemibiotrophic ascomycete Magnaporthe oryzae suppresses BAX-induced programmed cell death in tobacco leaves [26] . In U . maydis , several gene clusters encoding putative effector proteins have been identified [27] , [28] . However , the modes of action of these clustered effectors and the way they contribute to fungal virulence still remain elusive . Recently , the secreted effector Pit2 has been identified as an essential virulence factor of U . maydis [29] . The pit2 gene resides in a small cluster next to pit1 , which encodes a transmembrane protein . Both Pit1 and Pit2 are required for U . maydis induced tumor formation but the function of these proteins in the host plant still remains unclear [29] . A single effector of U . maydis that is essential for the establishment of biotrophy is Pep1 [30] . The pep1 gene was found to be specifically expressed during pathogenic development of U . maydis [30] . Pep1 deletion mutants form normal penetration structures but plant infection is stopped immediately upon epidermal penetration and this coincides with the elicitation of strong defense reactions in the host plant . Pep1 was found to localize to the plant apoplast where it particularly accumulates at sites of cell-to-cell passages of biotrophic U . maydis hyphae [30] . In addition , it became evident that Pep1 is not only essential for virulence of U . maydis but also for the related barley smut fungus Ustilago hordei , indicating a conserved function of Pep1 in other fungal biotrophs besides U . maydis [30] . However , so far it remained unclear why this effector plays such a crucial role for Ustilago infections . In this study , we present the functional characterization of Pep1 and demonstrate that it acts as inhibitor of host peroxidases .
In a previous study we observed various plant responses such as autofluorescent cell wall depositions being induced by the U . maydis pep1 deletion mutant ( SG200Δpep1 ) [30] . To get additional insight in Pep1 function the defense activation , which is elicited by the pep1 deletion mutant has now been studied in more detail . Maize leaves infected with SG200Δpep1 were stained with aniline blue , which selectively marks callose ( 1-3-β-D-glucose ) that is synthesized as a defense response triggered by PAMP-induced elicitation [31]–[33] . Infection sites of the virulent U . maydis strain SG200RFP , which expresses cytoplasmic RFP [30] , revealed only marginal callose accumulations surrounding the penetrating hypha ( Figure 1A , S1A ) . In contrast , large depositions of callose were observed at points of attempted penetration by the RFP-expressing Δpep1 mutant strain SG200Δpep1RFP [30] ( Figure 1B ) . Additionally , we performed cerium chloride ( CeCl3 ) staining that allows the visualization of hydrogen peroxide ( H2O2 ) by transmission electron microscopy ( TEM ) studies [34] , [35] . TEM of maize epidermis cells infected by U . maydis strain SG200 revealed no ROS accumulation in these cells ( Figure S1A ) . Intracellular hyphae were surrounded by an intact plant plasma membrane , reflecting the established biotrophic interface ( Figure 1C ) . In contrast , around hyphae of SG200Δpep1 that penetrated epidermal cells we observed CeCl3 accumulations , indicating formation of ROS ( Figure 1D ) . The H2O2 signal found in SG200Δpep1 infections particularly localized to the plant cell wall at sites of penetration ( Figure S1B ) . CeCl3 staining also circumcised the penetrated cells , suggesting an ongoing hypersensitive response ( HR ) in these cells ( Figure S2 ) . The penetrated cells exhibited additional signs of HR such as ruptures of the tonoplast membrane and internal disintegration [36] , ( Figure 1D , S2 ) . Our previous microarray data on SG200Δpep1 infected maize leaves showed a strong induction of defense-associated host genes [30] . To further differentiate the transcriptional responses of maize to U . maydis wild type versus infections by the pep1 deletion mutant , expression of typical JA-associated marker genes and SA-responsive transcripts was determined by qRT-PCR . As JA markers we used a Bowman Birk trypsin inhibitor [4] and the maize Cystatin-9 [37] . Both genes were strongly induced after SG200 infection , while their expression was only weakly induced in SG200Δpep1 infected maize leaves ( Figure 2B ) . In contrast , the SA marker PR1 [5] as well as the SA-induced metal-ion binding protein ATFP4 [38]–[40] were upregulated specifically after SG200Δpep1 infection . Most interestingly , a gene encoding the maize peroxidase-12 ( POX12 ) was highly induced upon SG200Δpep1 infections compared to wild type infections ( Figure 2A ) . POX12 belongs to the class III peroxidases ( NCBI: cd00693 ) of the plant heme-dependent peroxidase superfamily ( Figure S3 ) . Peroxidases of this class have also been reported to be involved in plant responses to pathogen attack and were involved in ROS production during the initial phase of the oxidative burst [8] , [41] , [42] . Summarizing our previous data with the present cytological and molecular observations we conclude that the Δpep1 mutant induces an oxidative burst response , eventually leading to plant cell death . In light of the cell death-induction by the pep1 deletion mutant , the major challenge was to elucidate how Pep1 interferes with the maize immune system . To functionally characterize Pep1 , the open reading frames of pep1 and gfp , which was used as a recombinant control protein , were fused to an N-terminal 6×His-tag and expressed in E . coli ( see methods for details ) . As a test for the capability of Pep1 to suppress early plant defense responses , an oxidative burst was induced in maize leaf disks by the fungal elicitor chitosan ( Figure 3A ) , or by heat inactivated U . maydis cells , respectively ( Figure S4 ) . In each case , the generation of ROS was visualized using a luminol-based readout . While the chitosan induced oxidative burst appeared transiently within 15 minutes after elicitation ( Figure 3A ) , treatment with heat inactivated U . maydis caused a continuous burst that did not decrease within a measuring period of 60 minutes ( Figure S4 ) . Strikingly , recombinant Pep1 almost completely blocked the chitosan-induced oxidative burst ( Figure 3A , B ) . Similarly , ROS production induced by heat inactivated U . maydis cells was blocked by Pep1 ( Figure S4 ) . In contrast , neither E . coli expressed GFP nor heat inactivated Pep1 protein inhibited the oxidative burst ( Figure 3B , S4 ) . These results show that Pep1 acts as an inhibitor of the elicitor-triggered oxidative burst , suggesting that Pep1 interferes with an essential component of this early defense reaction . Next , we tested the origin of the PAMP triggered oxidative burst in maize leaves . To this end , maize leaf discs were treated with chitosan and the resulting H2O2 production was visualized in vivo by xylenol orange staining [43] ( Figure 3C ) . Potential sources of ROS production under these experimental conditions were tested using either salicylhydroxamic acid ( SHAM ) , an inhibitor of peroxidases [44] , [45] , or the NADPH-oxidase-inhibitor diphenylene iodonium chloride ( DPI ) [46] , [47] . ROS production could be inhibited by both DPI and SHAM to similar extents ( Figure 3C ) , suggesting that both NADPH-oxidase and peroxidase activity contributed to the PAMP-triggered oxidative burst in maize . While Pep1 did not interfere with the H2O2-induced color-change of xylenol orange ( Figure S5 ) , this assay showed an inhibition of the oxidative burst by Pep1 to similar levels as caused by treatment with either SHAM or DPI , respectively ( Figure 3C ) . The ability of Pep1 to suppress the oxidative burst response is in line with the phenotype of the Δpep1 mutant ( Figure 1; [30] ) . To test whether this particular function of Pep1 is relevant for U . maydis infection , pathogenic development of SG200Δpep1 was tested under conditions where ROS were scavenged through application of ascorbate . To this end , 5 mM ascorbate was applied to maize seedlings at the sites of SG200Δpep1 infections 12 and 24 hours after fungal inoculation , respectively . Confocal microscopy revealed that this treatment resulted in a drastic decrease in callose depositions at penetration sites ( Figure 4A , B ) . Most importantly , under these conditions SG200Δpep1 was able to enter the leaf tissue with intracellular hyphae reaching up to 100 µm of length without eliciting a visible defense response ( Figure 4B ) . In rare cases , branching of intracellular hyphae was observed , indicating proliferation of biotrophic SG200Δpep1 hyphae ( Figure 4B ) . Quantification of infectious hyphae revealed an average increase of 600% of intracellular hyphal length after ascorbate treatment , compared to mock-treated control plants ( Figure 4E ) . In addition , the rate of penetrated epidermis cells undergoing cell death was tested using a maize line expressing a YFP-tagged version of the auxin transporter PIN1 as a plasma membrane marker [30] . It got evident that ROS-scavenging significantly reduced Δpep1-induced cell death compared to mock treated maize leaves . While 75% of SG200Δpep1 infected maize cells collapsed in the controls ( Figure 4C ) , only about 30% of penetration events caused such a response when ascorbate had been applied ( Figure 4D , F ) . Maize leaves infected with SG200Δpep1 exhibited clusters of dead cells surrounding the infection site , while the addition of ascorbate led to reduced symptoms and the tissue stayed mostly alive and showed signs of chlorosis ( Figure S6 ) . The strong transcriptional induction of pox12 in Δpep1 infections as well as the observed oxidative burst inhibition by Pep1 led us to consider the possibility that peroxidases could be a potential target of Pep1 . To test this , a quantitative in vitro peroxidase assay was performed using commercial horseradish peroxidase ( HRP ) reacting with diaminobenzidine ( DAB ) in the presence of H2O2 . In this assay , peroxidase activity of HRP results in the formation of a brown DAB precipitate . This reaction was also observed , when recombinant GFP or heat inactivated Pep1 were added to the assay ( Figure 5A ) . In contrast , native Pep1 efficiently inhibited HRP activity ( Figure 5A ) . This Pep1-driven peroxidase inhibition was found to depend on the Pep1 concentration , as well as on the pH ( Figure 5B ) . To address the question whether Pep1 directly interacts with the peroxidase , a Far Western blot experiment was performed . Different amounts of Pep1 were blotted on a nitrocellulose membrane and incubated with HRP ( Figure 5C ) . As a negative control , similar concentrations of recombinant GFP were blotted on the same membrane . Specific chemiluminescence signals showed that HRP was binding to Pep1 ( Figure 5C ) . Furthermore , the intensity of chemiluminescence signals correlated with the amount of blotted Pep1 ( Figure 5C ) , suggesting a direct interaction of Pep1 with HRP . Given the finding that Pep1 resides in the plant apoplast [30] , one could hypothesize that during U . maydis penetration , Pep1 might suppress ROS formation by inhibiting apoplastic peroxidase activity . To test the impact of Pep1 on maize peroxidases , apoplastic fluid of maize leaves was isolated . Extracted apoplastic fluids were tested for peroxidase activity using DAB precipitation . High levels of peroxidase activity were detected in the apoplastic fluid and this activity was inhibited specifically by native Pep1 in a concentration dependent manner ( Figure 5D ) . The finding that Pep1 inhibits HRP as well as apoplastic maize peroxidases suggested a rather unspecific interaction of Pep1 with peroxidases . The maize POX12 , however , is not transcriptionally induced by H2O2 directly ( Table S2 ) , but displayed a particularly strong transcriptional activation upon Δpep1 infection ( Figure 2 ) and no other maize peroxidase besides POX12 was transcriptionally induced upon SG200Δpep1 infection [30] . Furthermore , both HRP and POX12 belong to the type-III class heme-peroxidases , sharing 37% identity on the amino acid level and are highly conserved in the active domain ( Figure S3 ) . We therefore investigated whether Pep1 and POX12 physically interact inside the plant . To visualize the protein interactions in vivo , a modified split-YFP system was established , which allowed a microscopic localization of proteins also in cases where no fluorescence complementation took place . To this end , an mCherry tag was fused to the C-terminus of the N-terminal part of YFP ( pSPYNE_R ) . Similarly , a CFP-tag was added to the C-terminal part of YFP ( pSPYCE_C ) . Both constructs also contained an N-terminal secretion signal ( for details see methods section ) to facilitate apoplastic localization of the fusion proteins . Using Agrobacterium tumefaciens mediated transformation , the constructs were transiently expressed in Nicotiana benthamiana under the control of the 35S promoter . N . benthamiana cells expressing both pSPYCE_C and pSPYNE_R fused to pep1 ( pSPYNE_Pep1 ) showed apoplastic fluorescence signals for mCherry and CFP , indicating secretion of the fusion proteins ( Figure 6A , S7A ) . However , expression of only the fluorescence fusion proteins with Pep1 did not result in any detectable YFP signal , demonstrating that no unspecific protein dimerization occurred ( Figure 6A ) . Similarly , no YFP fluorescence was detected when pSPYNE_mCherry was co-expressed with pSPYCE_C fused to POX12 ( pSPYCE_POX12 ) ( Figure S7A ) . This shows that neither Pep1 nor POX12 caused any unspecific fluorescence . In contrast , cells that co-expressed pSPYNE_Pep1 and pSPYCE_POX12 showed a complementation of YFP fluorescence , which co-localized with the mCherry and CFP signals ( Figure 6B ) . Similarly , co-expression of pSPYNE_POX12 ( pSPYNE_R fused to POX12 ) and pSPYCE_Pep1 ( pSPYCE_C fused to Pep1 ) resulted in YFP fluorescence complementation ( Figure S7B ) . In addition , a yeast two hybrid experiment was performed to test interaction of Pep1 and POX12 . Confirming the results obtained by fluorescence complementation , simultaneous expression of the two proteins in yeast restored growth on selection medium , indicating interaction of Pep1 and POX12 ( Figure 6C ) . This interaction was also evident when the putative active site of POX12 [48] was mutagenized ( see materials and methods for details ) , suggesting that Pep1 does not bind to the catalytic site of POX12 ( Figure 6C ) . This specific fluorescence complementation by co-expression of Pep1 and POX12 fusion proteins confirms a direct physical interaction of Pep1 and POX12 in vivo , substantiating a biological function of Pep1 as a peroxidase inhibitor . To test whether the Pep1-POX12 interaction is relevant for U . maydis interaction , the peroxidase gene was silenced in maize plants using a recently established virus-induced gene silencing ( VIGS ) assay that allows systemic gene silencing in maize during U . maydis interaction [49] . For POX12 silencing , two fragments of the coding region ( see methods for details ) were integrated into RNA3 of the Brome mosaic virus ( BMV ) . Maize seedlings were inoculated with the resulting construct BMV-POX12si for subsequent U . maydis infection . Control plants were inoculated with a BMV silencing construct for the non-plant gene YFP ( yellow fluorescent protein; BMV-YPFsi ) , which does not influence the maize – U . maydis interaction [49] . BMV inoculated plants were infected with SG200Δpep1 and fungal infection was monitored by confocal microscopy 48 hours after fungal infection . Similar to non-treated maize leaves ( Figure 1B ) , BMV-YFPsi plants formed large papillae at sites of SG200Δpep1 infection and the mutant hyphae were stopped during epidermal penetration ( Figure 7A ) . In contrast , silencing of POX12 led to a substantial increase in penetration efficiency as well as a reduction of visible plant defense responses ( Figure 7A–C ) . In the POX12 silenced plants , biotrophic SG200Δpep1 hyphae were found to pass from cell to cell and they even colonized mesophyll cells ( Figure 7A ) , a phenotype that has never been observed in control plants infected with SG200Δpep1 . qPCR analysis confirmed an average reduction of POX12 transcript levels of 85% in the BMV-POX12si inoculated plants compared to the BMV-YFPsi controls ( Figure 7C ) . From these results we conclude that POX12 activity contributes to maize resistance to the Δpep1 mutant and that inhibition of this peroxidase by Pep1 is crucial for U . maydis infection .
The U . maydis Δpep1 mutant induces cell death of the infected epidermal cells . This coincides with H2O2 formation , which particularly accumulates in the apoplastic space and around the penetrating hyphae . Therefore , the ability of Pep1 to suppress the oxidative burst is perfectly in line with the mutant phenotype . The inability of the Δpep1 mutant to establish biotrophy is reflected by the lacking induction of JA-responsive genes , which are markers for a compatible biotrophic interaction [4] , [7] . In contrast , the Δpep1 mutant strongly induced the expression SA-marker genes , which does not occur in U . maydis wild type infections [4] . This does also include the POX12 , as defense-associated peroxidases typically are induced by SA [60] . An important finding for understanding the Pep1 function was that scavenging of ROS by an antioxidant suppresses callose deposition and cell death response upon epidermal penetration of the Δpep1 mutant . Although ROS-scavenging might be considered as a rather unspecific intervention , its effect on Δpep1-induced plant responses suggests that suppression of apoplastic oxygen stress is sufficient to complement absence of Pep1 during epidermal penetration . Main sources of ROS formation in the plant apoplast are the plasma membrane bound NADPH-oxidases [61] , [62] and cell wall- or membrane-bound apoplastic peroxidases [11] , [17] , [63] . A couple of elegant studies , mainly done on the model plant Arabidopsis , demonstrated an important impact of NADPH-oxidase activity in ROS formation , cell death induction and thereby host immunity to microbial infection [13] , [64]–[66] . However , there is evidence that depending on the host-pathogen interaction , different oxidative burst profiles relying on the activation of the NADPH-oxidase and/or the peroxidase systems are established [67] , [68] . In barley , silencing of the NADPH-oxidase HvRBOHF2 increased susceptibility to the barley powdery mildew Blumeria graminis f . sp . hordei as well as a reduction in wound induced cell death [69] . Interestingly , this phenotype was not associated with altered H2O2 production . The HvRBOHF2 silencing barley plants showed normal H2O2 accumulation at penetration sites , papilla formation and hypersensitive reacting cells . In addition , the mutant plants showed a fully developed oxidative burst response after PAMP-treatment [69] . These findings indicate that in barley either functional redundancy among NADPH-oxidases , or action of additional components , i . e . peroxidases are involved in basal defense responses including the PAMP-triggered oxidative burst . Apoplastic peroxidases were found to be crucial during incompatible plant interactions [11] and their direct secretion was proposed to the sites of attempted pathogen invasion [63] . In wheat , the overexpression of a secreted class III peroxidase increased resistance to powdery mildew infection by potentiating the epidermal cell death response [70] , which supports the role of defense-related peroxidases in plant defense to biotrophic pathogens . Using in vitro assays , a physical interaction of Pep1 and HRP became evident , which correlated with a concentration-dependent inactivation of peroxidase activity by Pep1 . In addition , Pep1 largely suppressed maize apoplastic peroxidase activity . To demonstrate interaction in vivo we selected the maize POX12 because of its strong transcriptional activation upon infection by the pep1 deletion mutant , which was not found for any other peroxidase genes [30] . Our improved split-YFP method showed that apoplastic co-localization of either Pep1 or POX12 fusion proteins with only the fluorescence markers did not result in fluorescence signals , demonstrating specificity of the assay . Direct interaction with POX12 in planta verified the in vitro data of Pep1 physically binding to peroxidase . Regarding the Far Western approach it should be mentioned that Pep1-bound HRP apparently was not completely inactivated , because the assay is based on POX activity . This may result from the highly sensitive detection system that will visualize even marginal activity . However , together with the finding that Pep1 binds to POX12 as well as to HRP it indicates a rather unspecific interaction of Pep1 to peroxidases . This is also in line with the partial inhibition of the maize apoplastic peroxidase activity . The maize genome encodes for about 150 peroxidases [71] , which makes it very likely that this apoplastic activity results from a mixture of different peroxidases that are sensitive to Pep1 . However , a part of the apoplastic activity was not inhibited by Pep1 , which might indicate that not all H2O2 producing enzymes present in the maize apoplast are targets of Pep1 . Despite the presence of multiple apoplastic peroxidases in maize , the POX12 appeared to be relevant for U . maydis infection . VIGS of POX12 lead to a significant rescue of the Δpep1 mutant phenotype . Particularly defense responses during epidermal penetration were suppressed in POX12 silencing maize leaves and Δpep1 mutant hyphae were able to enter the mesophyll tissue . However , we did not observe the massive hyphal proliferation that is found in U . maydis wild type infections . Despite POX12 being the only maize peroxidase transcriptionally induced after Δpep1 infection , one must assume presence of other peroxidases that were not silenced in our approach . In silico prediction of siRNA formation ( http://bioinfo2 . noble . org/RNAiScan/RNAiScan . htm ) deriving from the used POX12 silencing constructs revealed a potential co-silencing of only three closely related peroxidases ( Genbank accessions: BT036551; BT036456; BT037744 ) . However , none of the corresponding genes was induced in SG200Δpep1 infections [30] . Therefore , we consider it likely that remaining peroxidase activity in POX-silenced plants prevented a full complementation of the Δpep1 phenotype . Another possibility might be that Pep1 holds an additional function that adds to U . maydis virulence , which cannot be ruled out at present . Here we have shown that Pep1 has a crucial role for oxidative burst suppression via a direct inhibition of apoplastic peroxidases . This has uncovered an elementary virulence mechanism of U . maydis . In contrast to known effectors that interfere with specific plant signaling pathways , Pep1 suppresses PAMP-triggered immunity by scavenging one of its core components . The targeted peroxidases are a conserved , integral part of the first layer of plant defense responses . In line with this , Pep1 is also highly conserved in related pathogens such as the maize smut Sporisorium reilianum or the barley covered smut fungus Ustilago hordei [28] , [30] . However , in published genome sequences of plant pathogens outside the group of Ustilaginales , no homologous proteins can be identified . Therefore it is an intriguing question whether other biotrophs developed analogous inhibitors of plant peroxidases , or if alternative mechanisms evolved to suppress basal plant defense .
For VIGS experiments , Zea mays L . cv Va35 [72] plants were grown in phytochambers at 28°C during the light period ( 26 , 000 lux; 14 . 5 h ) and at 22°C during the dark period ( 9 . 5 h ) . For all other experiments , Zea mays cv . Early Golden Bantam was grown in a green house at 28°C during the light period ( 26 , 000 lux , 14 . 5 h ) and 22°C during the dark period ( 9 . 5 h ) . Nicotiana benthamiana plants were grown at 22°C during the light period ( 26 , 000 lux; 14 . 5 h ) and at 20°C during the dark period ( 9 . 5 h ) . For infections with U . maydis , a liquid culture of the strain SG200Δpep1 [30] was grown in YEPSL ( 0 . 4% yeast extract , 0 . 4% peptone and 2% sucrose ) at 28°C shaking at 200 rounds min-1 ( rpm ) to an optical density ( OD600 ) of 0 . 6–0 . 8 . Cells were centrifuged at 900 g for 5 min , resuspended in H2O to an OD600 of 1 . 0 and used for infection of 17 day old maize seedlings ( 11 days after BMV inoculation ) . To scavenge reactive oxygen species in SG200Δpep1 infections 5 mM ascorbate solution was applied to the infection site 12 and 24 h after U . maydis inoculation . Confocal observations were carried out 3 days after fungal infections . For staining of callose , aniline blue ( Sigma , Taufkirchen , Germany ) was used . Leaf samples were washed two times in 50% EtOH , followed by 2 washing steps in 100 mM Sodium-Phosphate buffer pH 9 . 0 . Subsequently the samples were incubated in 0 . 05% ( w/v ) aniline blue in 100 mM Sodium-Phosphate buffer pH 9 . 0 for 1 h in the dark . Confocal microscopy of the samples was carried out in the staining solution . The S . cerevisiae strain AH109 ( Clontech ) was used for yeast two-hybrid interaction studies . Yeast cultures were grown in YPD full medium ( 1% yeast extract , 2% peptone , 1% glucose ) or SD-Glucose minimal medium ( 0 , 67% yeast nitrogen base , 2% glucose ) to an OD600 of 0 . 6–0 . 8 at 28°C shaking at 200 rpm . Yeast two-hybrid experiments were carried out following the Clontech Matchmaker GAL4 Two-Hybrid user manual . Selection of yeast transformants was conducted on SD-Glucose –Trp –Leu plates , selection for protein interaction was performed on SD-Glucose –Trp –Leu –His plates . To quantify generation of reactive oxygen species from maize leaves in response to elicitor treatment , Luminol-based assays have been performed using maize leaf discs of 6 mm diameter . Luminol reaction solution was prepared as follows: 80 µM Luminol and 0 . 15 U/ml HRP ( Sigma , Taufkirchen , Germany ) were dissolved in 50 mM Sodium-Phosphate buffer pH 8 . 0 . The assay volume was set to 1 . 5 ml , consisting of 1 ml Luminol reaction solution and 500 µl of elicitor solution . Luminescence measurements were carried out in a Tecan Infinite M200 plate reader ( Tecan Group Ltd . , Männedorf , Switzerland ) . Firstly a base line measurement was taken for 5 min before injection of the elicitor . After elicitation , measurements were continued for 180 min . Elicitor concentrations were 2 mg/ml glycol chitosan ( f . c . ) ( Sigma ) and 2×108 cells/ml heat inactivated SG200 cells . For usage in the luminol assay , proteins were dissolved in 100 mM Sodium-Phosphate buffer pH 7 . 5+150 mM NaCl . For the colorimetric visualization of H2O2 , each 12 leaf discs of 6 mm diameter were floated in 2 ml H2O . An oxidative burst was elicited by addition of 2 . 5 mg/ml chitosan to the assay . 5 min after elicitation , 120 µl of the water were harvested and H2O2 production was quantified with a xylenol orange based readout according to [17] , [73] . Inhibitors/proteins were added 10 min prior to elicitation . Absorption measurements of the assay solution were undertaken in a Tecan Safire plate reader ( Tecan Group Ltd . , Männedorf , Switzerland ) . Background measurements were taken in the respective buffers and substracted from sample values . In vitro HRP activity was visualized by DAB staining which was carried out in a clear 96 well micro titer plate ( Greiner Bio-One , Frickenhausen , Germany ) . The assay solution consists of 2 . 7 mM DAB ( Sigma , Taufkirchen , Germany ) , 0 . 375 U/ml HRP and 50 mM Sodium-Phosphate buffer ( pH 6 . 5 or 7 . 5 ) in a total volume of 150 µl . DAB precipitation was initiated by the addition of 2 µl of 0 . 1% H2O2 . If applicable , purified proteins were pre-incubated with the assay solution before the addition of H2O2 . After 10 min the micro plate was scanned in an Epson V700 Photo flat bed scanner ( Seiko Epson Inc . , Tokyo , Japan ) and staining intensity was quantified using Adobe Photoshop CS2 V . 9 . 02 ( Adobe Systems Inc . , San Jose , CA , USA ) as follows . The image of the scanned plate was converted to gray scale and inverted subsequently . Now average grey values of each well were measured using the histogram tool . Blank wells filled with unstained buffer were measured as well and resulting grey values subtracted from sample values to eliminate the background . For apoplastic fluid extraction , 100 Early Golden Bantam maize plants were grown under green house conditions at 28°C . After 7 days the seedling leaves were harvested and cut into pieces of 2–3 cm length , followed by evacuation under water in a vacuum chamber for 3×15 min at 400 mbar . The evacuated leaf sections were then stacked into packs of 20–30 and squeezed into the barrel of a 50 ml syringe so that the cut edges of the leaves faced the ends of the barrel . The barrel was then put into a 50 ml falcon tube with the needle hub facing downwards and spun for 20 min at 2000 g and 4°C . Afterwards the extracted apoplastic fluid was collected from the falcon tube and stored at −20°C . The POX activity of maize apoplastic fluid was visualized via in vitro peroxidase activity assay as described above , with the following modifications: The assay solution consists of 2 . 7 mM DAB ( Sigma , Taufkirchen , Germany ) , 2 µl of 0 . 5 mg/ml maize apoplastic fluid and 50 mM Tris buffer ( pH 7 . 5 ) , in a total volume of 100 µl . Confocal images were taken on a TCS-SP5 confocal microscope ( Leica , Bensheim , Germany ) , as described previously [30] . Fluorescence of YFP was elicited at 514 nm and detected at 520–540 nm , mCherry fluorescence was excited at 561 nm and detected at 590–630 nm , for detection of aniline blue ( 0 . 05% w/v in 0 . 1 M phosphate buffer pH 9 . 0 ) and cell wall autofluorescence , an excitation of 405 nm and detection at 435–480 nm were used . Sample preparation for transmission electron microscopy ( TEM ) was performed by a modified protocol according to [34] , [35] . Briefly , small pieces of leaves ( 1 mm2 ) were incubated for 1 h with a 5 mM cerium chloride ( CeCl3 ) solution dissolved in 50 mM MOPS-buffer ( 3- ( N-morpholino ) propanesulfonic acid ) at pH 6 . 5 . Samples were then fixed in a mixture of 2 . 5% paraformaldehyde and 2 . 5% glutardialdehyde , dissolved in buffer at pH 7 . 2 for 90 min , rinsed in buffer ( 4 times , 10 min ) and post fixed with 1% osmium tetroxide dissolved in buffer ( pH 7 . 2 ) . Dehydration was carried out in increasing concentrations of acetone ( 50% , 70% , 90% and 100% ) for 2 times , 10 min for each step . The acetone was then exchanged with propylenoxide and the samples were then infiltrated with increasing concentrations of Agar 100 epoxy resin ( 30% , 60% and 100% ) . Samples were polymerized at 60°C for 48 h . Ultrathin sections ( 80 nm ) were cut with a Reichert Ultracut S ultramicrotome and post stained for 5 min with a 2% lead citrate solution and for 15 min with 1% aqueous uranyl acetate before they were observed with a Philips CM10 TEM . For expression of recombinant proteins in E . coli , strain Rosetta-gami ( DE3 ) pLysS ( Novagen/Merck , Darmstadt , Germany ) was used . Expression vectors were based on pET15b ( Novagen , Madison/USA ) . Cells were grown in dYT medium , containing 100 µg/ml ampicillin , 50 µg/ml kanamycin , 2 . 5 µg/ml tetracycline , 34 µg/ml chloramphenicol and 1% glucose , to the mid-logarithmic growth phase at 37°C . For protein overexpression cells were shifted to 28°C , followed by induction with 100 to 400 µM Isopropyl-β-D-thiogalactopyranosid ( IPTG ) ( Sigma , Taufkirchen , Germany ) . After 4 hours cells were pelleted by centrifugation . Pellets were stored at −20°C . Conditions for the purification of all the recombinant His-tagged proteins were optimized for maximal yield and purity by nickel affinity chromatography . The frozen cell pellet was resuspended in binding buffer ( 20 mM Tris , 500 mM NaCl , 20 mM imidazole , pH 7 . 9 ) supplemented with 500 µg/mL Lysozyme ( Merck , Darmstadt , Germany ) , 0 . 1% Triton X-100 ( Carl Roth , Karlsruhe , Germany ) and incubated at room temperature for 20 min . The cells were disrupted by five times sonication at 4°C for 1 min with 1 min resting periods . The cellular debris was pelleted by centrifugation at 15 , 000 rpm and 4°C for 30 min . The purifications were carried out following the QIAexpressionist handbook ( Qiagen , Hilden , Germany ) with very little modification . The supernatant was applied to a gravity flow column , containing a bed volume of 1 ml Ni–NTA beads ( Qiagen , Hilden , Germany ) , previously equilibrated with binding buffer . After incubation for 30 min at 4°C , the Ni–NTA column was washed with 5 bed volumes of binding buffer , followed by 5 bed volumes of washing buffer ( 20 mM Tris , 500 mM NaCl , 60 mM imidazole , pH 7 . 9 ) . Recombinant proteins were then eluted from the Ni–NTA column with elution buffer ( 20 mM Tris , 500 mM NaCl , 500 mM imidazole , pH 7 . 9 ) . For further assays the buffer was exchanged using illustra Nap25 columns ( GE Healthcare , Buckinghamshire , United Kingdom ) . Depending on the subsequent assay , proteins were stored in Tris based storage buffer ( 100 mM Tris , 100 mM NaCl , 7 . 5 ) , or in a sodium phosphate based storage buffer ( 100 mM sodium phosphate , 150 mM NaCl , pH 6 . 5 or 7 . 5 ) . Proteins were concentrated using Amicon Ultra tubes ( Milipore , Tullagreen , Ireland ) with an exclusion size of 3 kDa . Finally 10% glycerol were added and proteins were stored at −20°C . The different stages of purification were monitored by SDS–PAGE . Protein concentrations were determined by Bradford assay employing BSA as a standard . VIGS using BMV was performed as described previously [47] . To obtain BMV RNA1 , RNA2 and RNA3 , the plasmid pF1-11 , pF2-2 and the different pB3-3 constructs were digested , individual transcripts were synthesized and the RNA integrities were tested as described previously [48] . To silence POX12 , two siRNA-fragments were designed . POX12si-fragment-1 was corresponding to the bases 557–798 of the 1086 bp coding region of the pox12 open reading frame . POX12si-fragment-2 was corresponding to the bases 766–937 of pox12 open reading frame . Both fragments were individually integrated into two pB3-3 vectors as described [48] . To produce BMV containing POX12si-fragments , Nicotiana benthamiana plants were infected as described [70] . After inoculation , the leaves were harvested and ground in 0 . 1 M phosphate buffer , pH 6 . 0 ( 1∶10 , w/v ) . The BMV titer was quantified by qPCR using primers specific for the minus strand of RNA1 ( Table S1 ) . All N . benthamiana extracts were adjusted by addition of 0 . 1 M phosphate buffer , pH 6 . 0 , to the same virus titer of 2000 relative expression units compared with non-inoculated tobacco extracts , POX12si-fragmentswere mixed in a ratio of 1∶1 ( v/v ) and applied to the maize plants as described previously [47] . To determine the efficiency of pox12 silencing , samples were taken two days after U . maydis infection of BMV inoculated plants and used for qRT-PCR as described previously [48] . Samples were cut diagonally to provide two samples for both qRT-PCR and microscopic analysis . Samples were shock frozen in liquid nitrogen and stored at −80°C . For RNA isolation , samples were ground to powder in liquid nitrogen and RNA was extracted with Trizol ( Invitrogen , Karlsruhe , Germany ) and purified using an RNeasy Kit ( Qiagen , Hilden , Germany ) . After extraction , the First Strand cDNA Synthesis Kit ( Fermentas , St . Leon-Rot , Germany ) was used to reverse-transcribe 300 ng of total RNA with oligo ( dT ) primers for qRT-PCR . The qRT-PCR analysis was performed , using an iCycler machine ( Bio-Rad , Munich , Germany ) in combination with iQ SYBR Green Supermix ( Bio-Rad ) . Cycling conditions were as follows: 2 min at 95°C , followed by 45 cycles of 30 sec at 95°C , 30 sec at 61°C and 30 sec at 72°C . Gene expression levels were calculated relative to gapdh as described in [49] . Error bars in all figures that show qRT-PCR data give the standard deviation that was calculated from the original CT ( cycle threshold ) values of three independent biological replicates . Pep1 and GFP were overexpressed and purified as described above . Protein samples were prepared using 10 mM DTT and SDS loading buffer and boiled for 5 min . 25 , 15 and 7 . 5 µg of total protein were separated by SDS-PAGE and transferred to a nitrocellulose membrane . After electroblotting , membranes were saturated with 5% non-fat dry milk in TBS-T ( 50 mM Tris-HCl , 150 mM NaCl , pH 7 . 6 , 0 . 1% Tween-20 for 1 h at room temperature ( RT ) . After blocking , the membrane was washed five times with TBS-T . Subsequently , the membrane was incubated over night at 4°C with horseradish peroxidase ( HRP ) ( Sigma , Taufkirchen , Germany ) dissolved in TBS-T at a concentration of 25 µM . Then , the membrane was washed five times with TBS-T and signals were detected by chemiluminscence detection using ECL Plus Western Blot detection reagent ( GE Healthcare ) . Standard molecular biology methods were used according to [74] . All restriction enzymes used in this study were purchased from New England Biolabs ( Frankfurt/Main , Germany ) . For protein overexpression of Pep1 and GFP the E . coli expression vector pET15b ( Novagen , Schwalbach , Germany ) was used . pep1 was amplified from U . maydis SG200 DNA , using Primers O19 and O20 ( Table S1 ) . gfp was amplified from the p123 vector [75] . PCR products were cloned into the pET15b vector via NdeI and BamHI restriction sites . Isolation of genomic U . maydis DNA was performed as described previously [76] . PCR was performed using Phusion High-Fidelity ( NEB , Frankfurt/Main , Germany ) . The PCR products of the different genes were cleaned up before digestion , using the Wizard SV Gel and PCR Clean-Up System ( Promega , Mannheim , Germany ) and ligated into the pET15b expression vector . The vectors were transformed into DH5α cells ( Invitrogen , Karlsruhe , Germany ) , and then plated on YT-agar plates containing 100 µg/ml ampicillin . At least three colonies were picked and grown over night in 2 ml dYT medium containing 100 µg/ml ampicillin . Plasmids were extracted using QIAprep system ( Qiagen , Hilden , Germany ) and cleaned by Wizard SV Gel and PCR Clean-Up System . After sequencing , one correct construct was transformed into E . coli Rosetta-gami ( DE3 ) pLysS ( Novagen/Merck , Darmstadt , Germany ) . Constructs for the microscopic interaction studies via BiFC were based on pUC-SPYNE-35S and pUC-SPYCE-35S [77] . Plasmids were modified as follows . To obtain the additional fluorescence tag , genes encoding for CFP and mCherry respectively were amplified using primers O23-O25 ( Table S1 ) adding a RSIATA spacer sequence . PCR products were cloned into the BiFC-vectors via XhoI and XmaI restriction sites . ORFs from the BiFC vectors were then digested with HindIII and EcoRI and ligated into pGreen0000 [78] . To remove excess restriction sites from the obtained vectors , two inverse PCR steps were added using primers O26-O29 ( Table S1 ) . PCR fragments were digested with NdeI and HindIII , followed by religation resulting in the vectors pGreen_SPYCE-CFP and pGreen_SPYNE-mCherry . Subsequently a codon optimized pep1 gene that carries a secretion signal from the legumine B4 gene from Vicia faba ( GenBank: X03677 . 1 ) was then amplified using primers O30 and O31 ( Table S1 ) and inserted into pGreen_SPYNE-mCherry and pGreen_SPYCE-CFP via restriction sites BamHI and XhoI . The POX12 gene from Z . mays was amplified using primers O32 and O33 ( Table S1 ) and inserted into pGreen_SPYCE-CFP via restriction sites BamHI and XhoI . Control vectors pSPYCE_C and pSPYNE_R were generated by removing the pep1 gene from pGreen_SPYNE-mCherry-Pep1 and pGreen_SPYCE-CFP-Pep1 via inverse PCR using primers O34 and O35 ( Table S1 ) and religation , leaving the legumine B4 signal peptide from V . faba in the vector , resulting in secreted BiFC constructs as negative controls . Created BiFC vectors were transformed into A . tumefaciens GV3101 cells by electroporation ( 1 . 5 kV ) . Agrobacterium mediated transient transformation of N . benthamiana was carried out following the protocol of [79] . For confocal microscopy , leaf discs with a diameter of 6 mm were cut from the infected leaf areas and immediately observed in the microscope . Constructs for Yeast-Two-Hybrid interaction studies were based on the vectors pGBKT7 and pGADT7 ( Clontech ) . The pep1 gene was cloned into pGADT7 using the primer pair O36/O37 and the restriction sites XmaI and XhoI . The pox12 gene was cloned into pGBKT7 using the primers O38/O39 and the restriction sites NdeI and EcoRI . The putative active site of POX12 [48] was mutagenised by exchanging Arg74 , His78 and His207 to Alanine . Phe77 was exchanged by Valine . Respective point mutations were introduced to the pox12 gene at positions using the primers O40 and O41 with the Quik Change Multi Site Directed Mutagenesis Kit ( Agilent Technologies , Santa Clara , CA ) . To generate the pB3-3/POX12si constructs , the primers O13-O16 ( Table S1 ) were designed for two RNAi antisense fragments of POX12 ( GenBank: EU964425 . 1 ) . All primers contained a HindIII restriction site for integration into pB3-3 ( Table S1 ) . Cloning and integration into pB3-3 vector was done as described in [49] . The following maize gene fragments were inserted in antisense orientation in pB3-3 vector to RNA3: for POX12-fragment 1 , 248 bp; and for POX12-fragment 2 , 178 bp . As a silencing control pB3-3/YFPsi of our previous study was used [49] . For in silico analysis of siRNA formation and the silencing specificity of maize sequences , the software tool SIRNA SCAN ( http://bioinfo2 . noble . org/RNAiScan . htm ) was used . Predictions were made using data from the J . Craig Venter Institute maize tgi v16 database . pF1-11 and pF2-2 were provided by X . S . Ding & R . S . Nelson [72] . For RNA isolation , infected leaf areas of each 30 maize seedlings were pooled , ground in liquid nitrogen , extracted with Trizol ( Invitrogen , Karlsruhe , Germany ) and purified using an RNeasy Kit ( Qiagen , Hilden , Germany ) . Expression of Zm-pr1 ( ZMU82200 ) , Zm-pox12 ( ACG36543 ) , Zm-atfp4 ( NP_001152411 . 1 ) , Zm-cc9 ( BN000513 . 1 ) and Zm-bbi ( EU955113 . 1 ) were analyzed by qRT-PCR using primers O1–O12 ( Table S1 ) . After extraction , the First Strand cDNA Synthesis Kit ( Fermentas , St . Leon-Rot , Germany ) was used to reverse-transcribe 1 µg of total RNA with oligo ( dT ) primers for qRT-PCR . The qRT-PCR analysis was performed as described above . Gene expression levels were calculated relative to gapdh expression levels as shown previously [49] . Sequence alignments of conserved domains among class III heme-peroxidases according to [48] was done using clone manager 9 . 1 software ( Sci-Ed software , Cary , USA ) . For sequence assembly pox12 ( EU964425 . 1 ) served as reference sequence .
|
The maize pathogen U . maydis establishes a biotrophic interaction with its host plant and causes the formation of plant tumors . The U . maydis infection is initiated by a direct penetration of the plant epidermis and relies on living plant tissue . Therefore , suppression of the host immune system is essential for successful infection . Previously we identified the secreted effector Pep1 , which is essential for U . maydis pathogenicity . pep1 deletion mutants are blocked by host defense responses immediately upon penetration . In the present study we identified the molecular function of Pep1 and explain its crucial role for fungal virulence . We found that Pep1 inhibits the plant oxidative burst , which is characterized by the accumulation of reactive oxygen species ( ROS ) such as hydrogen peroxide . A conserved component of the plant ROS generating system are peroxidases . We could show that Pep1 directly inhibits plant peroxidases . One specific maize peroxidase ( POX12 ) , which was strongly induced by infection of the pep1 deletion , directly interacts with POX12 in vivo . Moreover , POX12 silenced plants are penetrated by the pep1 deletion mutant , indicating functional relevance of the Pep1-POX12 interaction . Together , these findings show that Pep1 directly interferes with the ROS-generating system of the host plant to suppress immune responses .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"plant",
"science",
"plant",
"biology",
"plant",
"pathogens",
"microbial",
"pathogens",
"plant",
"pathology",
"biology",
"microbiology",
"host-pathogen",
"interaction",
"pathogenesis"
] |
2012
|
The Ustilago maydis Effector Pep1 Suppresses Plant Immunity by Inhibition of Host Peroxidase Activity
|
Meiotic chromosomes assemble characteristic “axial element” structures that are essential for fertility and provide the chromosomal context for meiotic recombination , synapsis and checkpoint signaling . Whether these meiotic processes are equally dependent on axial element integrity has remained unclear . Here , we investigated this question in S . cerevisiae using the putative condensin allele ycs4S . We show that the severe axial element assembly defects of this allele are explained by a linked mutation in the promoter of the major axial element gene RED1 that reduces Red1 protein levels to 20–25% of wild type . Intriguingly , the Red1 levels of ycs4S mutants support meiotic processes linked to axis integrity , including DNA double-strand break formation and deposition of the synapsis protein Zip1 , at levels that permit 70% gamete survival . By contrast , the ability to elicit a meiotic checkpoint arrest is completely eliminated . This selective loss of checkpoint function is supported by a RED1 dosage series and is associated with the loss of most of the cytologically detectable Red1 from the axial element . Our results indicate separable roles for Red1 in building the structural axis of meiotic chromosomes and mounting a sustained recombination checkpoint response .
Meiosis is a specialized developmental process , in which a diploid cell undergoes two chromosomal divisions without an intervening S phase to produce haploid gametes for sexual reproduction . The reduction in ploidy occurs during meiosis I , when homologous chromosomes are segregated . To enable this unique segregation pattern , meiotic cells must identify and physically link homologous chromosome pairs . In most sexually reproducing organisms , such linkages are achieved by meiotic recombination . In addition to increasing genetic variation in the progeny , recombination leads to crossover exchanges that , together with sister chromatid cohesion , stably connect homologous chromosomes [1] . Meiotic crossover formation relies on the controlled introduction and repair of a large number of programmed DNA double-strand breaks ( DSBs ) . DSBs are formed in a non-random manner across the genome by the conserved topoisomerase-like enzyme Spo11 [2–7] . In the budding yeast Saccharomyces cerevisiae , 160 DSBs are estimated to occur on average per meiosis [8] . Following DSB formation , nucleolytic processing releases Spo11 from break ends along with short covalently-linked oligonucleotides [9] . Exonucleases resect the break ends to produce 3’ single-stranded DNA ( ssDNA ) tails [10] that are used by the recombinases Rad51 and Dmc1 to preferentially invade the homologous chromosomes [11 , 12] . Some of the resulting strand-invasion intermediates are stabilized and processed to produce crossovers [13–17] . Meiotic recombination occurs in the context of highly conserved chromosome architecture , characterized by linear arrays of chromatin loops anchored to a proteinaceous axis known as the axial element [18–20] . As meiotic prophase progresses , axial elements of homologous chromosomes in many organisms co-align with the help of transverse filament proteins , such as Zip1 in yeast and Sycp1 in mice [21] . The axial elements subsequently form the lateral elements of the synaptonemal complex ( SC ) , a ladder-like structure connecting homologous chromosome pairs in the later stages of meiosis [21] . Like the rest of the SC , axial elements are defined cytologically; they appear as electron dense linear structures by electron microscopy and can be visualized by immunofluorescence analysis of proteins localizing to these structures [20 , 22–26] . The structural axes of meiotic chromosomes are also defined functionally , with mutants that lose axial integrity exhibiting numerous defects in meiotic recombination , including reduced DSB formation , aberrant DSB repair , and a loss of DSB surveillance [18 , 27 , 28] . These defects lead to inviable gametes in a variety of model organisms [20] and are associated with male infertility and premature ovarian failure in human patients [29 , 30] . The molecular structure of the axial elements is only beginning to be understood . In S . cerevisiae , the axial element comprises a specialized meiotic cohesin complex , containing the meiosis-specific subunit Rec8 [23] , as well as the meiotic proteins Red1 and Hop1 [22 , 24] . Rec8-cohesin is thought to form the chromatin anchor that recruits Red1 and Hop1 to axis-attachment sites along chromosomes [31 , 32] . Moreover , electron microscopy studies indicate that only Rec8-cohesin is essential for axial element formation , as red1 and hop1 mutants still exhibit dark-staining linear structures [23 , 33 , 34] . These structures , however , appear fragmented and irregular , suggesting that RED1 and HOP1 contribute to axial element formation or stability . RED1 and HOP1 exert essential regulatory functions in the context of the axial element . Deletion of RED1 or HOP1 reduces DSB levels to 10–30% of wild type [12 , 35–39] , likely because of the role of these genes in recruiting essential DSB factors near DSB hotspots [31] . Moreover , the DSBs that do form in red1 and hop1 mutants are associated with unusually long resection tracts [40 , 41] . Loss of RED1 or HOP1 function also causes severe synapsis defects [33 , 34] , although linear stretches of the central SC component Zip1 remain detectable in some red1 mutant cells [24] . Finally , red1 and hop1 mutants fail to activate the meiotic checkpoint network [42 , 43] , which helps block repair from the sister chromatid to bias repair events toward the homologous chromosome [44 , 45] , and acts to arrest cells in meiotic prophase in response to unrepaired DSBs [27 , 38 , 46 , 47] . Surprisingly , mutants lacking REC8 recapitulate only some of the phenotypes of red1 and hop1 mutants . Although rec8Δ mutants are unable to synapse [23] and exhibit excessive resection [41] , their defects in DSB formation are more nuanced than in red1 and hop1 mutants , affecting some chromosomal regions while sparing others [32 , 48 , 49] . In addition , rec8Δ mutants are proficient in mounting a meiotic checkpoint response [23 , 50] . This disparity may be explained by the fact that Red1 and Hop1 are recruited to a limited number of chromosomal regions in the absence of Rec8 [31 , 32] . Although their distribution pattern is highly abnormal , Hop1- and Red1-rich regions exhibit close to wild-type DSB levels [32 , 48 , 49] , implying that the axis-associated DSB and checkpoint activities remain functional in these regions . Red1 and Hop1 recruitment to the axial element is affected in mutants of the condensin complex . The condensin complex is a conserved regulator of chromosome architecture that localizes to axial elements and functions in chromosome compaction and removing cohesin from chromosomes at the end of meiotic prophase [51 , 52] . The intensity of Red1 and Hop1 on chromosome spreads is decreased in temperature-sensitive condensin mutants [51] , and nearly undetectable in the meiosis-specific condensin allele ycs4S , which is linked to a C-terminal 12xMYC tag on the condensin subunit Ycs4 [51] . The reason for the more severe meiotic defects of ycs4S has remained unclear , in particular as other condensin functions , including cell survival and chromosome compaction , appear unaffected by this allele [51] . Here , we show that the ycs4S allele is in linkage with a promoter mutation in the nearby RED1 gene that reduces RED1 expression to about 25% of wild-type levels and explains most , if not all , ycs4S phenotypes . Intriguingly , the reduced Red1 levels cause a differential loss of RED1 activities . Analysis of a Red1 dosage series including the ycs4S allele shows that DSB formation , deposition of the synapsis protein Zip1 , and spore viability are all buffered against substantial reductions in Red1 levels . By contrast , the ability to maintain a checkpoint arrest in meiotic prophase is very sensitive to Red1 dosage and abolished in the ycs4S mutant . These data suggest separable activities of Red1 in regulating axis integrity and the maintenance of checkpoint activity , and imply that the majority of Red1 is primarily involved in the maintenance of meiotic checkpoint activity .
The low Red1 and Hop1 signals localizing to meiotic chromosomes of ycs4S mutants [51] prompted us to test whether total axis protein levels are reduced by this allele . To this end , wild-type and ycs4S cells were harvested in a synchronous meiotic time course , and the levels of Red1 and Hop1 were analyzed by western blotting . We found that total Hop1 levels are similar in both strains , although we noted a reduction in the slower migrating , phosphorylated forms of Hop1 [53] in the ycs4S mutant ( Fig 1A ) . By contrast , Red1 protein levels are strongly reduced in the ycs4S mutant relative to wild-type meiotic cells . Quantitative western analyses showed that ycs4S mutants express on average approximately 20% of the Red1 protein levels of wild type during meiotic prophase ( Fig 1B ) . This decrease in Red1 protein levels differentiates the ycs4S mutant from the temperature-sensitive ycg1-2 and ycs4-2 condensin mutants , which do not exhibit a discernable reduction of Red1 levels at the restrictive temperature of 34°C ( Fig 1C ) . Coincident with the decrease in Red1 protein in the ycs4S mutant is a reduction of the slower migrating band , which is the result of Red1 phosphorylation [54] ( compare late time points in Fig 1A ) . The biased loss may be the result of preferential degradation of phosphorylated Red1 in the ycs4S mutant or reflect reduced phosphorylation of Red1 as a result of reduced Red1 levels . To test if the loss of Red1 abundance occurs at the transcriptional or post-transcriptional level , we analyzed relative RED1 mRNA levels at the 3-hour time point by qRT-PCR . These measurements revealed that RED1 mRNA levels in ycs4S mutants are about 25% of wild-type levels ( p < 0 . 05; Student t-test , Fig 1D ) . By contrast , no significant difference was observed for the mRNA levels of the axial element components HOP1 and REC8 , or the transverse filament component ZIP1 ( p = 0 . 79 , p = 0 . 43 , and p = 0 . 10 , respectively ) . Together , these data indicate that the ycs4S allele is associated with a specific loss of RED1 mRNA expression . Because ycs4S differs from other condensin mutants , both in its ability to compact its chromosomes [51] and in its decreased concentrations of Red1 protein , we hypothesized that the meiotic phenotypes displayed by ycs4S mutants could be primarily caused by low Red1 expression levels as opposed to a direct effect of condensin malfunction . To test this possibility , we placed RED1 under the control of the HOP1 promoter , which is unaffected in the ycs4S mutant ( Fig 1D ) . Western blot analysis indicated that the pHOP1-RED1 construct leads to an overexpression of Red1 , although this effect was less obvious in the ycs4S mutant ( Fig 2A ) . In addition , Hop1 levels were slightly higher , including an increase in Hop1 phosphorylation signal ( Fig 2A , bracket ) . Importantly , the pHOP1-RED1 construct completely rescued the spore viability defect of ycs4S cells ( Student t-test: p-value = 1 . 1 x 10−4 , Fig 2B ) . The ability of the pHOP1-RED1 construct to rescue the ycs4S spore viability defect indicates that the RED1 promoter has a role in causing the ycs4S phenotypes . This initially puzzling result is explained by the fact that the YCS4 gene ( YLR272C ) is located in close proximity to RED1 ( YLR263W ) , with the RED1 promoter being separated from the C-terminal 12xMYC tag of ycs4S by less than 14kb . Sequence analysis revealed approximately 100 SNPs across this region that differ from the SK1 strain background and reflect the fact that the YCS4-12xMYC construct was introgressed from YPH499 into SK1 ( Fig 2C ) [51] . The preservation of YPH499-derived SNPs throughout this region over multiple crosses indicates a close genetic linkage between RED1 and YCS4 . We identified a single SNP in the RED1 promoter of the ycs4S mutant , a change from G to A at position -162 from the RED1 open reading frame . This SNP overlaps a conserved residue of a URS1 consensus site for the transcriptional regulator Ume6 [55] , which is required for wild-type levels of RED1 expression [56] . Disruption of the corresponding residue along with a second residue in the URS1 sequence of the meiotic SPO13 gene causes a 6-fold reduction in expression [57] . Introduction of the YPH499 SNP into an otherwise wild-type SK1 strain resulted in decreased Red1 protein levels ( Fig 2D , S1A Fig ) and spore viability defects ( Student t-test: p-value = 1 x 10−3 , Fig 2B ) comparable to those of ycs4S mutants . Introduction of a 13xMYC tag at the C-terminus of YCS4 had no effect on spore viability ( Student t-test: p-value = 0 . 104 , Fig 2B ) . These findings together with the rescue mediated by the HOP1 promoter suggest that the ycs4S phenotypes are caused by a linked mutation in the RED1 promoter that causes low RED1 expression . We note that providing the YPH499 genome in trans in a SK1/YPH499 hybrid diploid did not improve the spore viability defects caused by the red1-pG162A mutation ( S1B Fig ) , indicating that YPH499 has not adapted to this mutation . For the remainder of this paper , we will refer to the introgressed YPH499 genomic fragment containing the YCS4-12xMYC construct and the red1-pG162A mutation as red1ycs4S . If the low levels of Red1 protein in the red1ycs4S mutant are responsible for its axis assembly defect , chromosomal levels of Red1 and Hop1 should be diminished but still distributed at the expected sites along meiotic chromosomes . To test this prediction , we probed Red1 and Hop1 recruitment in red1ycs4S mutants using immunofluorescence and ChIP-seq analysis . Consistent with published results [51] , we found that Red1 and Hop1 levels on chromosome spreads are greatly reduced and no cytological axis structures are observed in red1ycs4S mutants ( Fig 3A and 3B ) . Notably , whereas Red1 signals are nearly undetectable , nuclei with discernible Hop1 foci still exist . However , the intensity of total Hop1 signal on chromosomes in these nuclei was two-fold lower than in wild-type nuclei ( Wilcoxon Sign Rank Test: p-value = 3 . 9 x 10−10 , Fig 3C ) . The incongruity in the amount of Hop1 versus Red1 foci detected in our studies may be the result of different antibody sensitivities . Alternatively , given that Red1 is required for Hop1 recruitment to chromosomes [24 , 32] , it may indicate that multiple Hop1 proteins are recruited per Red1 protein . Importantly , although no axis structures were observed cytologically , ChIP-seq analysis indicated that Red1 and Hop1 are nevertheless recruited in a wild-type pattern to meiotic chromosomes ( Fig 3D , S2A Fig , Spearman correlation: 0 . 87 and 0 . 66 , respectively ) . We note that the apparently similar signal intensities of wild-type and red1ycs4S profiles are a consequence of the necessary internal normalization , which precludes comparison of absolute intensities between ChIP-seq datasets [58] . Accordingly , ChIP-qPCR analysis of several loci revealed a general decrease in Red1 binding in the red1ycs4S mutants ( S3B Fig ) . Together , these analyses show that axis protein binding is reduced in red1ycs4S mutants , but that the low chromosomal amounts of Red1 and Hop1 are distributed normally to axis recruitment sites , suggesting that their chromosomal binding per se is unaffected . To determine if reducing the Red1 levels in red1ycs4S mutants affects cohesin-independent recruitment of axis proteins , we introduced a rec8Δ mutation . ChIP-seq analysis revealed that the patterning of Red1 and Hop1 is highly concordant in red1ycs4S rec8Δ mutants , a characteristic also observed in rec8Δ mutants [32] ( Fig 3D , S2A Fig , Spearman correlation: 0 . 94 and 0 . 93 , respectively ) . However , in the red1ycs4S rec8Δ mutant , the genomic regions with relative enrichment of Hop1 and Red1 exhibited a substantial dampening relative to the other strains , even though Hop1 and Red1 protein levels are not substantially different from the red1ycs4S strain ( S4A and S4B Fig ) . Quantification of total chromosome-associated Hop1 signal on chromosome spreads indicated no further loss of axis proteins in the red1ycs4S rec8Δ double mutant compared to the red1ycs4S mutant ( Wilcoxon Sign Rank Test: p-value = 0 . 66 , Fig 3C ) , but this result is likely caused by the already low signal in the red1ycs4S mutant . Indeed , ChIP-qPCR analysis revealed a further loss of Red1 enrichment in the double mutant ( S3A and S3B Fig ) . Some binding of Hop1 and Red1 likely remains even in the red1ycs4S rec8Δ mutant , especially on chromosome III , as indicated by a general enrichment compared to a mock ChIP-seq experiment ( S2B and S2C Fig ) and supported by ChIP-qPCR analysis ( S3B Fig ) . In addition , a few sites with strong Hop1 and Red1 enrichment persist in the double mutant ( e . g . Chr VIII , open triangle , Fig 3D , S2A , S2B , S3A and S3B Figs ) , but their significance is unclear as they are not associated with any obvious chromosomal landmarks . Analysis of Hop1 signal on chromosome spreads also indicated that the red1ycs4S allele rescues a characteristic axis defect of rec8Δ mutants . While rec8Δ mutants form distinctive clumps of Hop1 and Red1 on chromosomes spreads [23] , red1ycs4S rec8Δ mutants do not ( Fig 3E , S4C Fig ) . These data indicate that these clumps are dependent on Red1 protein abundance . During our analysis of the red1ycs4S rec8Δ mutant , we noticed that these cultures produce a substantial number of spores ( Fig 4A ) . This behavior is in contrast to rec8Δ single mutants , which arrest in meiotic prophase and fail to sporulate due to defects in DSB repair [23 , 50] , suggesting that red1ycs4S rec8Δ mutants are unable to mount a sustained checkpoint-mediated arrest response . Consistent with this notion , phosphorylation of Hop1 , one of the earliest markers of recombination checkpoint activation [45] , and phosphorylation of histone H3 threonine 11 ( H3T11 ) , a marker for activity of the downstream checkpoint kinase Mek1 [45 , 59 , 60] , are reduced in red1ycs4S cells ( Fig 4B ) . To investigate the kinetics of this arrest bypass , we followed the rate of meiotic spindle pole separation , which initiates after cells exit meiotic prophase [61] . In a rec8Δ mutant , few cells produced separated spindle poles as seen previously [49] , whereas red1ycs4S mutants formed spindles at the same rate as wild type ( Fig 4C ) . Wild-type kinetics of spindle pole separation were also observed in red1ycs4S rec8Δ double mutants , indicating the reduced levels of Red1 in the red1ycs4S mutant allow a complete bypass of the rec8Δ checkpoint arrest ( Fig 4C ) . We considered two non-exclusive explanations for this bypass . First , red1ycs4S rec8Δ may be able to repair DSBs using an alternative pathway . This possibility is supported by the fact that the barrier to sister chromatid repair is weakened in red1ycs4S mutants [51 , 62] . Second , the signaling response triggering the arrest could be weakened by the red1ycs4S allele , either by reduced DSB initiation or poor signal transduction . To investigate arrest signaling without the confounding effects of DSB repair , we analyzed spindle pole separation in a dmc1Δ rad51Δ background , which prevents essentially all homologous meiotic DNA repair [63] . We confirmed that red1ycs4S leads to a similar reduction in Red1 levels in the presence of these mutations ( S5A and S5B Fig ) . Importantly , consistent with the model that the reduced Red1 levels interfere with checkpoint signaling , the dmc1Δ rad51Δ red1ycs4S mutant completely bypassed the checkpoint arrest and proceeded through meiosis at a rate similar to wild-type cells ( Fig 4D ) . This defect is specifically caused by the red1-pG162A mutation , as introduction of the SNP is sufficient to cause a bypass ( Fig 4E ) . Moreover , the bypass is partially rescued by increasing RED1 expression ( pHOP1-RED1ycs4S; Fig 4F ) , indicating that reduced RED1 levels are responsible for the checkpoint defects . We note that the pHOP1-RED1 construct itself caused a slight defect in arrest activity , possibly due to RED1 overexpression , which is known to affect exit from meiotic prophase [64] . Furthermore , because dmc1Δ rad51Δ mutants do not completely arrest like dmc1Δ mutants [63] , we also confirmed that the red1-pG162A mutation was able to bypass the arrest of a dmc1Δ mutant ( S5C Fig ) . These data indicate that arrest signaling is defective in red1ycs4S mutants . Our analysis showed that deletion of REC8 also allowed a partial bypass of the arrest expected in the dmc1Δ rad51Δ mutants ( Fig 4D ) , which has been attributed to a reduced number of DSBs [50] . To investigate if reduced DSBs might explain the bypass of the dmc1Δ rad51Δ arrest , we determined total DSB activity in red1ycs4S mutants by immunoprecipitating Spo11 and quantifying the amount of associated oligonucleotides as a measure of Spo11 activity [9 , 65] . This analysis revealed that maximal DSB levels in red1ycs4S mutants are reduced by approximately 20% compared to wild type ( Fig 5A ) , in line with Southern measurements at individual hotspots [51 , 62] . To further support this measurement , we compared chromosome breakage in red1ycs4S mutants to a collection of three spo11 mutants with defined DSB activities [65 , 66] . Quantification of several full-length chromosomes after pulsed-field gel electrophoresis ( PFGE ) revealed that red1ycs4S mutants undergo meiotic chromosome breakage at levels significantly below the 100% of wild type but significantly above the ~70% of DSB activity of the spo11-HA allele ( paired Student t-test: p-value = 3 x 10−4 and 6 . 8 x 10−5 respectively , Fig 5B ) . red1Δ mutants exhibited ~30% of wild-type breakage levels . These trends were also observed when measuring DSB levels on chromosome XVI by Southern blotting ( S6A and S6B Fig ) . Importantly , Spo11-oligo analysis indicated that total DSB levels in the red1ycs4S mutant are at least as high as in rec8Δ mutants , which have a nearly intact checkpoint response ( Fig 4D ) . Moreover , DSB formation was not further reduced in a red1ycs4S rec8Δ double mutant ( Fig 5A ) . Accordingly , DSB levels of the double mutant remained considerably higher than in red1Δ mutants , as indicated by pulse-field gel analysis of chromosome VIII ( S6C and S6D Fig ) . It is possible that dmc1Δ rad51Δ red1ycs4S mutants fail to efficiently activate the checkpoint arrest because of a failure to produce sufficient ssDNA . ssDNA coated with the ssDNA-binding protein RPA is the major signal that activates the checkpoint kinase ATR/Mec1 [67] . To investigate the amount of ssDNA formed by hyper-resection in checkpoint-activated dmc1Δ rad51Δ mutants , we used Southern blotting to analyze the CCT6 hotspot , which exhibits approximately 70% of wild-type DSB levels in all three mutants ( Fig 5C and 5D ) . Resection can be seen at this hotspot as a smearing of the DSB band that progressively shifts to faster migrating species , corresponding to resection less than 5kb from the DSB site ( Fig 5C and 5E ) . Resection is also detectable as fragments that are larger than the parental band ( Fig 5C , Res1-4 ) . These larger fragments are caused by the inability of a restriction enzyme to digest ssDNA once resection tracts have passed the respective restriction sites [40] . At 3h after induction of meiosis , all three mutants exhibited more extensive smearing of the DSB band compared to dmc1Δ rad51Δ control ( Fig 5E ) , indicating more extensive resection . Moreover , the signal of the DSB band diminished in the mutants at later time points , as slower migrating fragments became apparent . Resection product Res4 ( Fig 5C ) , which is produced when resection results in an ssDNA tract of at least 8kb , became detectable in red1ycs4S dmc1Δ rad51Δ and rec8Δ dmc1Δ rad51Δ mutants 5h after induction of meiosis and increased in intensity over time ( Fig 5F ) , suggesting that resection tracts persist in the red1ycs4S dmc1Δ rad51Δ mutants even though a large portion of these cells have progressed out of meiotic prophase ( Fig 4D ) . The Res4 signal in the red1ycs4S rec8Δ dmc1Δ rad51Δ mutant appeared earlier and at a higher levels than either of the single mutants ( Fig 5F ) , suggesting that hyper-resection is faster and more sustained . These experiments are consistent with previous analyses showing that loss or RED1 or REC8 function leads to more extensive resection [40 , 41] , and argue that insufficient resection is not the cause of the observed bypass of the checkpoint arrest by red1ycs4S . The phenotypes described above suggest that different meiotic processes are differentially affected by changes in Red1 dosage . Specifically , the Red1 levels associated with the red1ycs4S mutant strain support sufficient DSB formation and homolog-directed crossover repair to yield nearly 70% viable spores ( Fig 2B ) , yet are unable to trigger any detectable prophase delay in response to unrepaired DSBs ( Fig 4D ) . To test this differential sensitivity in a more systematic manner , we took advantage of the cis-encoded reduction of Red1 protein levels of the red1ycs4S strain to create a Red1 dosage series ( Fig 6A ) . Quantitative western analysis indicated that red1ycs4S/RED1 and red1Δ/RED1 mutants have approximately 50% Red1 protein compared to wild type ( Fig 6B ) , an estimate supported by direct comparison with a titration series of wild-type protein extracts ( Fig 6A ) . The homozygous mutant red1ycs4S/red1ycs4S strain and the red1Δ/ red1ycs4S strain have approximately 25% and 15% of wild type Red1 protein , respectively . Consistent with the data shown in Fig 1A , increased Red1 protein levels correlated well with an increased proportion of the slower migrating phosphorylated form of Red1 ( R2 = 0 . 762 , Fig 6C ) , suggesting that Red1 phosphorylation is stimulated by Red1 abundance . Analysis of this dosage series revealed that strains expressing at least 50% of wild type Red1 protein trigger a meiotic delay in a dmc1Δ rad51Δ background ( Fig 6D ) . By contrast , no delay is observed in strains with equal or less than 25% Red1 protein . We note that a phosphorylation-deficient red1 mutant fully maintains a dmc1Δ arrest [54] , indicating that the observed loss of arrest activity is not caused by the dosage-dependent loss of Red1 phosphorylation . Importantly , despite the loss of checkpoint-dependent arrest activity , spore viability of red1ycs4S/red1ycs4S mutants remains around 70% , with only mild meiosis I non-disjunction ( Fig 6E , S7A–S7C Fig ) , suggesting that the ability of the meiotic checkpoint network to promote homolog-directed crossover formation remains at least partially active at around 25% Red1 levels . This interpretation is consistent with analysis of the HIS4LEU2 locus , revealing a partial block to sister repair and only a limited reduction in crossover levels in this mutant [62] . Indeed , even 15% Red1 levels of wild type support substantial spore viability , unlike red1Δ or hop1Δ null mutants , which do not form viable spores [33 , 34] ( Fig 6E ) , or homolog bias-defective hop1-scd mutants , which form less than 7% viable spores [45] . These data indicate that mounting a meiotic prophase delay requires substantially higher doses of Red1 than promoting largely faithful meiotic chromosome segregation . In an attempt to investigate DSB formation and resection cytologically in this dosage series , we analyzed chromosome spreads of dmc1Δ rad51Δ mutants using an antibody against the RPA subunit Rfa2 . Spread nuclei of all strains analyzed exhibited foci and short tracks of Rfa2 indicative of resected DSBs . However , cells expressing lower levels of Red1 had a higher probability of accumulating large clumps of Rfa2 in low-DAPI regions of the nucleus ( S7D Fig ) . These Rfa2 aggregates do not co-localize with Zip1 aggregates and are also formed at high frequency in resection-defective rad50S mutants ( S7E Fig ) , indicating that they are not representative of ssDNA exposed at Spo11-dependent DSB ends . RPA aggregates have been reported in other meiotic DSB repair mutants , including rad52Δ [68] , but their significance remains unclear . We note , however , that their appearance correlated well with loss of spore viability in the RED1 dosage series . Red1 is also an important regulator of chromosome synapsis , as red1Δ mutants form little to no SC , although occasional cells with more elaborate SC structures can be observed [24] . Therefore , we investigated the formation of lateral and central elements of the SC in the Red1 dosage series by immunofluorescence analysis . Red1 binding patterns on chromosomes reflect the reduced Red1 levels , with only the wild-type and red1ycs4S/RED1 strains exhibiting at least partial Red1 tracks on chromosome spreads ( Fig 7A ) . By contrast , Zip1 deposition along meiotic chromosomes is only mildly affected by decreased RED1 dosage . Linear Zip1 tracks formed even in red1ycs4S/red1ycs4S cells at levels similar to wild type ( Fig 7B ) . Spreads with Zip1 tracks in these mutants had almost no cytologically detectable Red1 ( Fig 7C and 7D ) , indicating that the majority of Red1 along lateral elements is dispensable for Zip1 deposition . Moreover , Zip1 tracks co-localize with SUMO in all strains ( Fig 7E ) , suggesting the formation of mature central elements of the SC [69] . We did note an increase in late leptotene/early zygotene nuclei , characterized by the co-existence of abundant Zip1 puncta ( leptotene configuration ) and short Zip1 stretches ( zygotene configuration ) , in red1ycs4S/red1ycs4S mutants ( Fig 7B ) . This class is rarely observed at higher Red1 dosage , and suggests that Zip1 deposition along chromosomes may be delayed at low Red1 levels . Consistent with this notion , we also observed an increasing incidence of Zip1 aggregates ( polycomplexes; Fig 7F ) as Red1 levels were reduced , although we note that polycomplex formation did not strictly correlate with Red1 levels , possibly because of the Ycs4-12xMYC tag in the red1ycs4S background . These data suggest that abundant binding of Red1 to lateral elements is not required for the formation of cytologically normal SCs .
Our analyses show that the majority of ycs4S phenotypes can be explained by a linked point mutation in the RED1 promoter that reduces RED1 expression to ~25% of wild type . Deletion of RED1 causes spore lethality because of numerous meiotic defects , including reduced DSB levels , hyper-resection , failure to block repair from the sister chromatid , defective chromosome synapsis , and an inability to arrest cells in response to persistent DSBs [33 , 38 , 40 , 41] . The ycs4S mutant recapitulates all of these phenotypes , albeit to varying extents ( this study and [51 , 62] ) . Consistent with the RED1 promoter mutation being the cause of most ycs4S phenotypes , a strain carrying this point mutation without the 12xMYC tag on YCS4 phenocopies the spore viability and arrest defects of the ycs4S mutant , while boosting RED1 expression levels in ycs4S meiotic cells using a heterologous HOP1 promoter essentially rescues the meiotic defects exhibited by ycs4S mutants . The YCS4-12xMYC construct itself likely causes only minor meiotic effects . We observed a slightly higher rate of polycomplex formation in the red1ycs4S/RED1 mutant compared to the red1Δ/RED1 mutant , even though the red1ycs4S/RED1 mutant appeared to have more Red1 protein on its chromosomes . This elevation in polycomplex formation may be attributable to the YCS4-12xMYC construct . In addition , Red1 protein expression is slightly reduced in the red1ycs4S background even when RED1 is placed under control of the HOP1 promoter . Tagging YCS4 de novo with a C-terminal MYC tag does not affect spore viability ( Fig 2B ) , suggesting that this effect may be caused by one of the additional SNPs in close linkage with the red1 promoter mutation . Our analyses indicate that reduced RED1 levels impact some meiotic processes more severely than others . The ability to mount an arrest response to unrepaired breaks is abolished at a Red1 dosage of approximately 25% of wild-type levels . By contrast , DSB formation and SC assembly are only mildly affected , and sufficient inter-homolog crossovers form to support ~70% spore viability . The mild effect on axis-associated processes , such as DSB formation and SC assembly , indicates that the structural chromosome axis remains largely intact at these levels of Red1 . This interpretation is consistent with the observation that red1ycs4S mutants exhibit wild-type levels of chromosomal compaction [51] , and that establishment of the crossover interference pattern , which requires SUMOylated Red1 , is unaffected in red1ycs4S mutants [70] . The differential sensitivity of meiotic processes to altered Red1 levels may reflect differences in how long the presence of Red1 is needed to execute a given process . In particular , mounting a checkpoint delay or arrest that lasts for several hours is expected to require substantially longer residence time of Red1 on chromosomes than the formation of DSBs , which would require Red1 only until a DSB has formed . It is more difficult to explain the ability of cells with reduced Red1 levels to form synapsis-competent chromosome axes in this model , as stable axes are likely needed to support chromosome synapsis . It is possible , however , that Red1 is primarily involved in the formation chromosome axes and less in their structural maintenance . This model is consistent with the formation of axial structures in the absence of Red1 [33] and may explain the coexistence of leptotene and zygotene configurations in the same nucleus when Red1 levels are reduced ( Fig 7B ) . It is also possible that the differential sensitivity to altered Red1 levels reflects functionally distinct populations of Red1 on meiotic chromosomes that act in the assembly of DSB- and synapsis-competent chromosome axes and in checkpoint arrest maintenance , respectively . Indeed , Red1 not only interacts with multiple chromosomal proteins , including Rec8 and Hop1 , but also multimerizes and is modified by both phosphorylation and SUMOylation , which could create functionally distinct forms of Red1 [32 , 37 , 42 , 54] . Finally , it is also possible that checkpoint signaling requires increased amounts of Red1 because of the need for signal amplification . This model is appealing because meiotic DSB resection tracts are kept relatively short by meiotic axis proteins [40 , 41] , such that even resection tracts of irreparable breaks are maintained at lengths of about 1kb for several hours . By contrast , irreparable breaks in mitotic yeast cells are continuously resected , which is necessary for maintaining checkpoint activity of the CHK2-like kinase Rad53 [71 , 72] . In meiotic cells , Rad53 is largely inactive , as Spo11-dependent DSBs do not lead to Rad53 phosphorylation [73] . Instead , checkpoint maintenance is linked to the continued activity of CHK2-like kinase Mek1 , which requires Red1 and Hop1 for activation [45 , 46 , 53 , 74] . Thus , the abundant binding of Red1 along chromosomes may amplify the binding and recruitment sites for Hop1 and Mek1 activation to support a sustained checkpoint signal without the associated risks of extensive resection . We note that Red1 is likely the limiting factor in this context because deletion of one copy of HOP1 in a red1ycs4S/RED1 strain background has no effect on spore viability ( S7F Fig ) . The notion that axial element-associated Red1 provides a means for signal amplification may also explain why rec8Δ mutants are able to mount a strong checkpoint arrest despite having little Red1 bound along chromosomes and showing no axial elements by electron microscopy [23] . Although DSB formation in rec8Δ mutants is largely restricted to genomic regions that are able to recruit Red1 independently of Rec8-cohesin [32] , these mutants display large clumps of Red1 and Hop1 on chromosomes ( Figs 3E , S3D ) . These clumps may support the checkpoint-dependent prophase arrest of these mutants despite the absence of detectable axial elements . Consistent with this interpretation , rec8Δ red1ycs4S mutants do not form Hop1 clumps on chromosomes and fail to arrest .
All strains had an SK1 background unless indicated otherwise . A complete list is located in S1 Table . The original ycs4S strain was a gift from D . Koshland [51] , SPO11-6HIS-3FLAG-loxP-KanMX-loxP was provided by K . Ohta [48] , and YPH499 was a gift from A . Strunnikov [75] . Gene disruption and tagging were carried out using a PCR-based protocol [76] . For pHOP1-RED1 and pHOP1-RED1ycs4S , the HOP1 promoter ( -200 to -2 ) as defined by Vershon et al . [77] was placed in pFA6a-kanMX6-pCLB2 , replacing pCLB2 . It was inserted into the indicated genomes using Longtine F4 and RHop1 [AAT TCC TGA CCT TTC TGA AA] primers as described [76] replacing the 25bp immediately upstream of the RED1 start codon with the HOP1 promoter construct . To transfer the red1-pG162A mutation from the ycs4S background into a clean SK1 background , an HphMX4 cassette [78] was inserted at -400 relative to the RED1 start site . The entire promoter region ( -600 to +90 ) was transferred into a wild-type SK1 background by PCR amplification and selection on hygromycin . Transfer of the mutation was confirmed by sequencing . An HphMX4 insertion without associated promoter mutation was constructed in SK1 to serve as a wild-type control . Synchronous meiosis was induced as previously described [32] . Whole cell extraction by trichloroacetic acid precipitation , SDS-polyacrylamide gel electrophoresis and western blotting were completed as described [79] . Hop1 and Red1 ( Lot#16441 ) were detected using rabbit serum at 1:10 , 000 ( kind gifts of N . Hollingsworth ) . Hop1 phosphorylation was detected using affinity-purified pT318-Hop1 antibody at 1:100 as described [80] . Phosphorylation of histone H3 threonine 11 ( H3T11 ) was detected with MC83 rabbit antibody ( Millipore ) at 1:2000 . Nsp1 was detected using 32D6 mouse antibody ( ThermoFisher Scientific ) at 1:2500 . Pgk1 was detected using 22C5D8 mouse antibody ( ThermoFisher Scientific ) at 1:500 . Fpr3 was detected with anti-rabbit serum ( kind gift from J . Thorner ) at 1:5000 . For all westerns , the secondary antibody was either anti-mouse HRP or anti-rabbit HRP used at 1:5000 ( GE Healthcare ) . For quantitative westerns , Nsp1 was detected using 32D6 at 1:25 , 000 and Red1 was detected using an anti-Red1 antibody ( Lot#16440; kind gift of N . Hollingsworth ) at 1:4000 . The secondary antibodies anti-rabbit IRDye800CW and anti-mouse IRDye680RD ( Li-Cor Biosciences ) were used at 1:10 , 000 . Membranes were scanned on a Li-Cor Odyssey imager under non-saturating conditions . Data were quantified using pixel intensities with the Odyssey software according to the protocols of the manufacturer ( Li-Cor Biosciences ) . Meiotic nuclear spreads were performed as described [80] . Hop1 was detected using anti-Hop1 rabbit serum at 1:200 in blocking buffer and Alexa Fluor anti-rabbit ( Jackson ImmunoResearch ) at 1:200 . Zip1 was detected using Zip1 yC-19 goat antibody ( Santa Cruz Biotechnology ) at 1:200 and anti-goat Cy3 at 1:200 ( Jackson ImmunoResearch ) . Red1 was detected using anti-Red1 rabbit serum ( Lot#16441 ) at 1:100 and Alexa Fluor anti-rabbit at 1:100 . Rfa2 was detected using anti-Rfa2 rabbit serum ( kindly provided by S . Brill ) at 1:1000 and Alexa Fluor anti-rabbit at 1:200 . Images were obtained as described [80] and analyzed using softWoRx 5 . 0 software . Scatterplots were created using Prism 6 . For Fig 7E , affinity purified rabbit anti-Zip1 ( raised at YenZym Antibodies , LLC , against a C terminal fragment of Zip1 as described [81] ) was used at 1:100 . Affinity purified guinea pig anti-SUMO was used at 1:200 ( gift from G . S . Roeder [82] ) . Secondary antibodies Alexa Fluor 488 and Alexa Fluor 594 were used at 1:200 ( Jackson ImmunoResearch ) . Microscopy and image processing were carried out using a Deltavision RT or a Deltavision Elite imaging system ( Applied Precision ) adapted to an Olympus IX17 microscope . All cultures were collected at the 3-hr time point . Chromatin immunoprecipitation was performed as described [83] . Samples were immunoprecipitated with 2 μL of either anti-Red1 ( Lot#16440 ) or anti-Hop1 serum per IP . For qPCR , input samples were diluted 50X more than ChIP samples . qPCR was completed as described previously [84] . Library preparation was completed as described [32] . Library quality was confirmed by Qubit HS assay kit and Agilent 2100 Bioanalyzer or 2200 TapeStation . 51-bp single-end sequencing was accomplished on an Illumina HiSeq 2500 instrument . Sequencing reads were mapped to SacCer3 ( S288C ) using Bowtie [85] . The one condition adjusted was to only collect information about reads that mapped to a single position in the genome . Reads were also mapped to the SK1 genome ( only allowing perfect matches ) with similar results . Reads were extended towards 3’ ends to a final length of 150 bp using MACS-2 . 1 . 0 ( https://github . com/taoliu/MACS ) [86] . Normalization of read tag density was completed as described [32] . Plots shown are an average of two replicates . Datasets are available at GEO , accession number GSE87060 . Genomic DNA for DSB hotspot analysis was purified as described [79] . For the CCT6 hotspot , samples were digested with NsiI and XhoI , and run on a 0 . 6% agarose gel for ~18 hr . The probe was created as described [8] . Pulse-field gel electrophoresis and Southern blotting were performed as described [3] . Hybridization signal was detected using a Typhoon FLA 9000 . DSB levels on PFGE were estimated assuming a Poisson distribution as described [66] . Specifically , DSB levels were calculated as -ln ( uncuttimepoint/uncut0h ) . End-labeling of Spo11-oligonucleotide complexes was completed as described [65] with a few changes . In brief , 100mL SPO cultures were lysed with glass beads in 10% trichloroacetic acid . Lysed cells were centrifuged , and then resuspended in 1 . 5mL of 2% SDS , 0 . 5M Tris-HCl pH 8 . 0 , 10mM EDTA , and 2% β-mercaptoethanol . After boiling the samples , soluble protein was diluted 2X in 2% Triton X100 , 30mM Tris-HCl pH 8 . 0 , 300mM NaCl , 2mM EDTA , and 0 . 02% SDS . Immunoprecipitation of the Spo11-oligo complexes was completed using 2 . 5μg of monoclonal mouse anti-Flag M2 antibody ( Sigma ) . Precipitated complexes were end-labeled with 5μCi of [α-32P]dCTP and 5 units of terminal deoxynucleotidyl transferase ( Enzymatics ) . End-labeled complexes were run on a Bolt 4–10% bis tris plus acrylamide gel ( ThermoFisher Scientific ) , blotted onto a PVDF membrane using an iBlot2 gel transfer device ( ThermoFisher Scientific ) , and visualized using a Typhoon FLA 9000 ( GE Healthcare ) . Blots were also probed with mouse monoclonal anti-FLAG M2-HRP ( Sigma ) . Spindle formation was followed by anti-tubulin immunofluorescence as described previously [87] . For each time point , 200 cells were classified based on whether or not they had undergone spindle pole separation , indicative of exit from meiotic prophase . Spore viability was determined by dissection of tetrads into individual spores unless otherwise stated . RNA was extracted as described [88] . Reverse transcription and qPCR was performed as previously described , but using RiboLock RNase inhibitor ( Thermo Scientific ) to prevent RNA degradation [84] . Primers used in this study are listed in S2 Table .
|
Meiosis is a specialized cellular division that reduces chromosome copy by half to produce four gametes . To ensure proper chromosome segregation , a meiotic cell must induce DNA double strand breaks , repair them with the homologous chromosome , and fully align the homologs . These aspects of meiosis are dependent on specialized meiotic chromosome axes and the proteins that control this structure , such as Red1 in S . cerevisiae . Here we analyze the effects of reduced Red1 levels on meiotic processes . Our results suggest that a comparatively small amount of Red1 is sufficient to promote DNA breakage and repair , whereas most of the cytologically detectable Red1 is necessary for controlling the ability of the cell to delay meiotic progression when double strand breaks remain unrepaired .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"phosphorylation",
"meiosis",
"medicine",
"and",
"health",
"sciences",
"homologous",
"chromosomes",
"chromosome",
"structure",
"and",
"function",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"meiotic",
"prophase",
"surgical",
"and",
"invasive",
"medical",
"procedures",
"synapsis",
"chromosome",
"biology",
"proteins",
"surgical",
"resection",
"biochemistry",
"cell",
"biology",
"post-translational",
"modification",
"biology",
"and",
"life",
"sciences",
"chromosomes"
] |
2017
|
Reduced dosage of the chromosome axis factor Red1 selectively disrupts the meiotic recombination checkpoint in Saccharomyces cerevisiae
|
One of the characteristics of prions is their ability to infect some species but not others and prion resistant species have been of special interest because of their potential in deciphering the determinants for susceptibility . Previously , we developed different in vitro and in vivo models to assess the susceptibility of species that were erroneously considered resistant to prion infection , such as members of the Leporidae and Equidae families . Here we undertake in vitro and in vivo approaches to understand the unresolved low prion susceptibility of canids . Studies based on the amino acid sequence of the canine prion protein ( PrP ) , together with a structural analysis in silico , identified unique key amino acids whose characteristics could orchestrate its high resistance to prion disease . Cell- and brain-based PMCA studies were performed highlighting the relevance of the D163 amino acid in proneness to protein misfolding . This was also investigated by the generation of a novel transgenic mouse model carrying this substitution and these mice showed complete resistance to disease despite intracerebral challenge with three different mouse prion strains ( RML , 22L and 301C ) known to cause disease in wild-type mice . These findings suggest that dog D163 amino acid is primarily , if not totally , responsible for the prion resistance of canids .
Prion diseases or transmissible spongiform encephalopathies ( TSEs ) , a group of fatal neurodegenerative disorders , have been described since the XVIIIth century when clinical signs of scrapie in sheep were reported in England [1] . Prion disorders are caused by misfolding of a protein , the cellular prion protein ( PrPC ) , which is expressed abundantly in the central nervous system , into an aggregated , self-propagating , disease-associated isoform known as PrPSc [2 , 3] . Prion diseases occur worldwide and affect many different mammalian species including humans . Natural prion disorders include scrapie in sheep and goats , bovine spongiform encephalopathy ( BSE ) in cattle , transmissible mink encephalopathy ( TME ) in mink , chronic wasting disease ( CWD ) in cervids , and Kuru , Creutzfeldt-Jakob disease ( CJD ) , Gerstmann-Straussler-Scheinker syndrome ( GSS ) , fatal familial insomnia ( FFI ) and variably protease-sensitive prionopathy ( VPSPr ) in humans , which arise either sporadically–putatively spontaneous misfolding of PrPC–or are caused by mutations in the PrP encoding gene that are inherited as an autosomal dominant trait [4 , 5] . Additionally , as a consequence of the BSE epidemic which occurred in the United Kingdom during the 1990’s , several mammalian species became infected with BSE prions . Such natural interspecies prion transmission , which had never been reported previously , gave rise to several new prion diseases including feline spongiform encephalopathy ( FSE ) in several feline species , TSE in a small number of non-human primates ( NHP ) and exotic ungulate spongiform encephalopathy ( EUE ) in several species of exotic ruminants of the Bovidae family kept in captivity [6 , 7] . These disorders appeared as a consequence of the consumption of BSE , originating from cattle , contaminated food in a way similar to that of variant CJD ( vCJD ) in humans [8] . The number of species known to be susceptible to TSE increased also due to experimental infections performed during scientific investigations which demonstrated that many rodent species , such as mouse , rat , hamster , guinea pig and bank vole can be infected experimentally by prions , as well as several non-human primates , e . g . chimpanzees , marmoset monkeys , stump-tail macaques , gibbons , spider monkeys , sooty mangabey , pigtail , cynomolgus and rhesus macaques , and squirrel monkeys [4 , 9 , 10] . Pigs , also of interest , due to their widespread consumption by humans , have also been shown to be susceptible to infection by experimental challenge [11 , 12] . Even rabbits , a species thought resistant to prion diseases , have been infected experimentally with TSE [13] . Thus , a wide variety of animal species are susceptible to prion disease including members of several mammalian families such as Bovidae , Cervidae , Muridae , Mustelidae , Felidae , Cricetidae , Caviidae , Leporidae , Suidae and Hominidae along with other primate families . PrP amino acid sequences are highly conserved among mammals with approximately 85% sequence , or greater , identity with human PrP except for marsupials , such as the opossum or wallaby , which have approximately 70% sequence identity . Reptile , fish and avian PrPs have sequence identities ranging from 18% to 35% compared to humans . Among the animal species susceptible to TSEs , the most similar sequence identity to human PrP is the chimpanzee with a 99 . 2% and the most distant is the cat with an 84 . 7% similarity [14–19] . However , the susceptibility of each species to TSEs has no correlation with PrP sequence divergence or its misfolding proneness as it is determined structurally [20] . TSE susceptibility predictions are complicated further as the differences in species related PrPC misfolding capacity depend on the disease-associated prion strain to which they are exposed . Natural selection may have succeeded in making some individuals or a whole species highly or completely resistant to some diseases . Although there is evidence to suggest that complete resistance to TSEs may be a utopia [21] , the susceptibility of some species is so low—due to a PrP sequence that is highly resistant to misfolding–that they become of great scientific interest . This is the case for rabbits [22] , horses [23] and dogs [24] which were considered prion resistant species based on the lack of natural cases reported and negative results in those challenged experimentally with prions [25] . However , the susceptibility of every mammalian species has not been evaluated and therefore , theoretically , the existence of resistant species sympatric with susceptible ones remains a possibility such as species not present in UK zoos during the BSE epidemic or that were never challenged [e . g . pronghorns ( Antilocapra americana ) ] . The absence of naturally occurring TSE cases or unsuccessful experimental challenges is not enough to classify any mammalian species as resistant to prion diseases as shown when rabbits and transgenic mice expressing rabbit PrP were proven definitively susceptible to TSE [13 , 26] . Also , equine PrP has been shown capable of misfolding into a pathogenic isoform but only transiently or unable to be transmitted between individuals [27] . Absolute resistance of a species to TSEs should not be considered as such , rather , different levels of susceptibility should be considered , and these can be used to study the molecular mechanisms behind the low susceptibility of certain species to prion disorders . Given the idiosyncrasies of the causal agent of TSEs , the mechanism of the low susceptibility in some species may be due to a specific feature ( s ) of each PrP sequence and its misfolding proneness [28] . Comparative alignments of PrP sequences—taking into account the disease susceptibility of each species—is one of the preferred methods to search for low susceptibility determinants [19 , 29–33] . Leaving aside rabbits and horses as their PrPs can be converted in to pathogenic isoforms [13 , 27] , members of the Canidae family and in particular domestic dogs ( Canis lupus familiaris ) , are the most interesting species to study as their PrP has never been shown to misfold in vivo [34] . Despite some studies or reports related to the susceptibility of dogs to prion disorders [35–37] , most of them were never confirmed , were simple reports with no scientific procedure involved or were purely speculative . All the members of the Canidae family , which includes wolves , foxes , jackals , dingoes and domestic dogs , share an almost identical PrP sequence with few polymorphic variants . Among the polymorphisms , the presence of either Aspartic acid or Glutamic acid in position 163 draws specific attention as it is highly characteristic and almost unique to the Canidae family [19] . Moreover , when compared to susceptible species or to those of unknown susceptibility , dog PrP presents several characteristics of interest as possible determinants for its low misfolding proneness [38 , 39] . Thus , domestic dog PrP was chosen to study the possible causes of apparent resistance to misfolding of PrP from Canidae . Several attempts to obtain an in vitro misfolded dog PrP by Protein Misfolding Cyclic Amplification ( PMCA ) [40] using a large number of prion strains and extensive PrP sequence comparison led to the identification of a single amino acid residue that may be responsible for its low misfolding proneness . Generation of transgenic mice over expressing mouse PrP with the identified substitution from dog PrP confirmed the implication of this amino acid residue in vivo . These transgenic mice showed complete resistance to TSEs after intracerebral inoculation with several mouse-adapted prion strains , corroborating the importance of certain key amino acid substitutions on TSE susceptibility . These results , besides helping to understand the mechanisms behind prion disease susceptibility and interspecies transmission barriers , may be useful for the design of new dominant negative PrPs able to block prion propagation and result in new therapeutic approaches for prion diseases [41] .
In order to evaluate the in vitro misfolding ability of dog PrPC , comparative in vitro seeding studies were performed using brain homogenates of two different breeds ( Cocker Spaniel and German Wirehaired Pointer ) . Four distinct procedures with modifications on the propagation conditions and six prion strains of diverse origins ( from mouse , sheep , cattle and deer ) were used as individual or pooled seeds . After 10 serial PMCA rounds no misfolded dog PrP was detected in any case except when seeded with classical BSE ( BSE-C ) , or a derivative of this one such as sheep BSE ( Fig 1 ) . Classical BSE and sheep BSE were the only seeds able to misfold dog PrPC but requiring a modified propagation protocol ( Fig 1B ) . To make sure that the signals observed correspond to real propagation and not to the remnant signal of the original seed or to the conversion of bovine PrPC present in the brain homogenate of the seed , additional experiments were performed ( S6 Fig ) . These results clearly suggest that dog PrPC shows a very high resistance to be misfolded , despite the fact that it could be forced to misfold in vitro resulting in a protein with the biochemical characteristics , such as PK resistance and electrophoretic migration pattern , expected for a prion ( Fig 1C ) and with conserved biological features with respect to the original seed [42] . A comparison of PrPC expression levels in the brains of different species ( dog , rabbit , cat and mouse ) did not show any significant differences between them . Thus , the possibility of a low expression level of PrPC in dog brain being the responsible for its low misfolding proneness in vitro could be ruled out ( S1 Fig ) . An amino acid sequence comparable with dog PrP from a mammalian species with a different level of susceptibility to prion infection was identified ( Fig 2A ) . The region 91–230 of the PrP sequence with the greatest similarity is cat PrP with just 6 distinct amino acids different . As felids are susceptible to infection with several different prion strains ( e . g . BSE , CWD and CJD ) [43–47] , the 6 identified amino acid residues present in dog PrP in other species were examined in correlation with their reported susceptibility to TSEs ( Fig 2B ) . From the initial six residues , just two , D163 and R180 , were virtually exclusive to canids or present in uncommon mammalian species in which no prion disease has been reported; the nilgai ( Boselaphus tragocamelus ) , Californian big-eared bat ( Corynorhinus townsendii ) and anteater ( Myrmecophaga tridactyla ) . The other 4 amino acids differing between canine and feline PrPs also appear in species shown to be susceptible to prion infection ( Fig 2B ) . Among canids , dogs showed two polymorphisms in positions 101 and 163 of PrP; Ser/Gly and Asp/Glu , respectively . While the polymorphism in position 101 appears in different canine species , the Asp/Glu163 has been described in domestic dogs only , probably as a consequence of intensive selective breeding during the development of distinct breeds ( S2 Fig ) . For position 163 there are only three mammalian species , besides canids , known to have an amino acid different to Asn . Two of them , pine marten ( Martes martes ) and wolverine ( Gulo gulo ) , belong to the Mustelidae family and curiously enough have the same amino acid residue as the canids ( S3 Fig ) . A representation of the phylogenetic tree could explain the possible origin of this polymorphism ( S3B and S3C Fig ) . One other mammalian species is known to have an amino acid different to Asn or to Asp/Glu at position 163 , the Squirrel monkey ( Saimiri sciureus ) , which can have a Ser residue and is known for its susceptibility to prion diseases [48–50] . To determine if the presence of the particular amino acid residue present in canids affects the structural arrangement of PrP from a TSE susceptible species an in silico structural analysis was performed . Two structural models based on mouse prion protein were generated which contained the apparently critical canine substitution ( N158D , equivalent to position 163 in dog PrP ) using as templates the averaged NMR structure from PDB ID 2L39 ( Model01 ) and X-ray crystallography structure from PDB ID 4MA7 ( Model02 ) . Cα superposition of Model01 onto Model02 overlays 103 residues ( out of 114 in Model01 ) with a root-mean-square deviation ( R . M . S . D . ) = 1 , 420 Å , showing no significant structural rearrangements in the backbone between them ( Fig 3A ) . Similarly , overall folding of the mutants is the same as those observed in the native mouse PrP structures used as templates . Cα backbone superposition of Model01 onto PDB ID 2L39 overlays 105 residues ( out of 114 in Model01 ) with an R . M . S . D . = 0 , 396 Å and Cα superposition of Model02 onto PDB ID 4MA7 overlays 106 residues ( out of 110 in Model02 ) with an R . M . S . D . = 0 , 055 Å . Superposition of both models Cα backbone onto canine prion protein ( PDB ID 1XYK ) show some differences although overall folding remains similar ( Fig 3B ) . Model01 onto canine prion protein overlays 93 residues ( out of 114 in Model01 ) with an R . M . S . D . = 2 , 052 Å and Cα superposition of Model02 overlays 97 residues ( out of 110 in Model02 ) with an R . M . S . D . = 2 , 170 Å . A closer analysis of the residue Asn158 ( N158 ) from mouse PrP in all reported structures and also both models , show a conserved hydrogen bond to Met133 ( M133 ) stabilizing the Cα backbone and different conformations in the prion surface . Nevertheless , while Asp158 ( D158 ) from Model02 does not establish other interactions with surrounding residues ( Fig 3C ) , D158 conformer from Model01 exhibits a change in the side-chain conformation leading to the formation of a salt bridge interaction with Arg135 ( R135 ) side chain ( Fig 3D ) at the loop linking strand-β1 and helix α1 . Analogous hydrogen bond can be observed between N158 and R135 also in mouse PrP PDB ID 2L39 structure but , surprisingly , no similar interaction can be found with surrounding residues in Canine-PrP ( PDB ID 1XYK ) ( Fig 3E ) . Analysis of the local environment at N158/D158 , when reported prion structures are compared , shows slight mobility in residue 158 but a higher one on R135 , suggesting that the R135-D158 interaction depends mainly on the R135 side chain rotamer which is comparable with those observed in wild-type mouse prion structures ( S4 Fig ) . The most notable change due to the introduction of D158 was detected through the study of electrostatic potentials on the surface of PrP which revealed larger differences when polar amino acid D158 was situated within an area with four Arg/Lys residues ( Fig 4A ) , Arg135 ( R135 ) , Arg150 ( R150 ) , Arg155 ( R155 ) and Lys119 ( K119 ) . Acidic properties of D158 in canine PrP ( Fig 4B ) introduces a positively charged residue in that region , as occurs in Model01 ( Fig 4C ) and Model02 ( Fig 4D ) , changing the local charge distribution of mouse wild-type PrP ( Fig 4E ) . To further explore the findings presented above , this study focused on amino acid residue 163 ( dog PrP numbering ) and how the presence of an Asp/Glu instead of an Asn could affect the ability of certain PrPs to misfold in vitro . Mouse PrP was chosen as a model and genetic constructs were prepared to produce mouse PrP with either Asn ( N158 in mouse PrP numbering ) , Asp ( D158 ) or Glu ( E158 ) in cells . In all the cases , expression level of mutated PrPs in cell cultures was checked by Western blotting to make sure that the relative ratios of cell PrPC versus brain PrPC were between 1:1 and 1:3 , being the cell PrP amount equal or minor than the brain PrP amount in all the experiments ( see an example of cell PrP expression levels in S7 Fig ) . The three proteins were subjected to two types of study based on cell-PMCA . The first type of cell-PMCA evaluated the in vitro propagation ability of each of the proteins that , besides the mutation in position 158 , contained a 3F4-tag . These proteins mixed with mouse whole brain homogenates ( mouse PrPC does not contain the 3F4 epitope ) were subjected to a single cell-PMCA round seeded with two different mouse prions ( RML and 22L ) ( Fig 5A ) . Regardless of the seed used , only the N158 protein was able to misfold in vitro , unlike proteins E158 and D158 ( Fig 5B ) . The second study evaluated the inhibitory capacity shown by each of the three proteins over the misfolding of wild-type mouse PrPC ( Fig 5C ) . In this case , RML was used as seed for a single cell-PMCA round in which the inhibitory effect of proteins D158 and E158 on the propagation of wild-type mouse PrP was clearly demonstrated ( Fig 5D ) . Both experiments suggested that D158 and E158 substitutions impede the misfolding of mouse PrP as well as inhibited the propagation of wild-type mouse PrP . The previous results , obtained in vitro , led to the generation of a new transgenic mouse model expressing mouse PrP with the substitution N158D . Two hemizygous transgenic mouse lines were generated , Tg402 and Tg403 with a PRNP0/0 background and with transgene expression levels of 1X and 2X , respectively compared to wild-type mouse endogenous PrPC levels ( S5 Fig ) . In order to demonstrate that the mutant PrP is properly expressed in the transgenic animals , a comparative immunohistochemistry was performed in wild-type and transgenic mouse lines . A fine granular neuropil immunolabeling ( corresponding to PrPC on the dendrite cell membrane ) and absence of labeling within the pericarion were observed ( S8 Fig ) . Six hundred days old animals of both transgenic lines had a normal phenotype with no neurological clinical signs . Brain homogenates from animals of both lines were used to evaluate their capacity to propagate mouse prions in vitro . Three different mouse prion strains ( RML , 301C and 22L ) were used as seeds for PMCA using the brain homogenates of the two transgenic lines as substrate and a brain homogenate of wild-type C57/BL6 mouse as positive propagation control . None of the brain homogenates containing D158 PrP was able to propagate any of the mouse prions used ( Fig 6 ) . Animals from both transgenic lines , together with C57/BL6 wild-type animals as TSE susceptible positive controls , were inoculated with three different mouse-adapted prion strains ( RML , 301C and 22L ) . While the wild-type mice succumbed to each prion disorder with the expected incubation times for each strain ( Table 1 ) , none of the transgenic animals showed any clinical signs even after 550 days post-inoculation ( dpi ) . Biochemical analysis of the brains of inoculated transgenic animals confirmed the absence of PK-resistant PrP ( PrPres ) depositions , while the infected C57/BL6 brains showed high amounts of PrPres with electrophoretic migration patterns characteristic of each of the strains used for inoculation ( Fig 7 ) . Histopathological characterization of the brains confirmed the presence of typical TSE related neuropathological findings , i . e . spongiform change and PrP immunolabeled deposits , only in wild-type mice but not in Tg402 or Tg403 mice brains ( Fig 7B ) .
To establish TSE resistance in any given species requires that several features are examined before a definitive conclusion can be reached . Since TSEs are a group of neurodegenerative disorders , sometimes with extremely low prevalence , the absence of TSE cases reported for certain species may be due to reduced size populations or because sporadic deaths in wildlife species are rarely investigated . For example , sporadic BSE cases in cattle affect around 2–5 individuals per million per year and were only detected when millions of animals were analyzed as a consequence of the “mad cow” crisis [51] . Therefore , to determine the susceptibility of certain species to TSEs experimental inoculations in a statistically representative number of individuals are required . Performing experimental infections in some mammalian species is relatively easy but for others difficult or even impossible due to a number of technical and ethical/cultural hindrances , and the latter is the reason of the absence of literature reporting prion inoculations in dogs . Although there are reports of suspected TSE cases in canids [35–37] , none of the results were conclusive and there is no definitive proof of prion disease in this genus . BSE contaminated feed was undoubtedly fed to several wild canine species in UK zoos and domestic dogs during the BSE epidemic which suggests that , if not completely resistant , canids have a very low susceptibility to prion infection , as many other mammals in zoological collections fed contaminated food succumbed to the disease [7] . Given the lack of rigorous experimental challenge studies in dogs , the definitive proof of their TSE resistance may be derived by inoculation of experimental transgenic mice models expressing dog PrP ( ongoing experiments ) . However , based on the in vitro results presented here , including several attempts to misfold dog PrP by PMCA , it is reasonable to assume that dog PrP is highly resistant to misfolding as the same in vitro methodology misfolds the PrP of species previously considered to be TSE resistant such as the rabbit [13] . Dog PrP could eventually be misfolded in vitro but using only BSE and BSE-derived strains , which retained their ability to infect bovine PrP expressing transgenic mice [42] . This is strong evidence of the powerful misfolding capacity of PMCA and endorses the use of this system to evaluate the misfolding ability of different species’ PrPs . However , regarding the evaluation of the degree of transmission barrier , PMCA is qualitative or semi-quantitative at best . In order to semi-quantify the strength of an interspecies transmission barrier , the in vitro process should be done in a controlled manner since ultimately PMCA might convert any mammalian PrP through the seeding with any kind of prion inocula if enough rounds during the in vitro process are performed . The number of cycles in each round and the number of rounds are empirical data and should be established by comparison with an standard [52] . In this case , the degree of canine PrP resistance to misfolding was not specifically evaluated in comparison to other species . Nonetheless , a low susceptibility to misfolding was concluded due to the requirement of modified PMCA conditions which increased the possibility of misfolding and because from the seeds tested just BSE and derived strains were able to induce misfolding . These issues were not observed when rabbit PrP was misfolded in vitro , although its low susceptibility to prion infection in vivo is well known [13] . Assuming a high but not complete resistance of dog PrP to misfolding , we focused on the identification of the amino acids and their positions in the prion protein that could be responsible . A PrP sequence alignment with species phylogenetically close to canids showed that feline PrP was the most similar in terms of primary structure . Cats are highly susceptible to three distinct prion strains ( BSE , CWD and CJD ) [43–47] and this led to the comparison of a small number of differing residues with PrPs from other species . From these comparisons , three specific amino acids were identified , although one of them was readily discarded due to its presence in PrPs of TSE susceptible animals . Of the other two identified residues detected in canine PrP , the Asp/Glu in position 163 was chosen as the most relevant for canine resistance . The presence of an Asn in that position is a highly conserved residue in mammalian PrP sequences from different species and located in the loop of residues linking helix-α1 and strand-β2 . Therefore , from all the PrPC structures described ( i . e . human , mouse , rabbit , sheep , cow , horse , hamster , cat , bank vole , pig and elk [23 , 24 , 53–60] ) , only the structural model of canine wild-type PrP [24] allows studying the role of Asp/Glu polymorphism in the overall structure of PrP . When N158 is replaced by D158 in the mouse PrP , no structural changes in the Cα backbone at the mutated position are observed in our in silico models except those stabilizing the backbone . This is probably due to D158 ( D163 in dog PrP numbering ) being highly exposed to the solvent and like most charged residues in proteins , it could play a role in the folding of the protein . A hydrogen bond present between N158 and R135 is not present in canine PrP ( between residues D163 and R140 in dog PrP numbering ) . Nevertheless , a salt bridge could be established between D158 and R135 as observed in Model01 due to the high mobility of the R135 side chain present in mouse prion structures . Both non-covalent interactions are relatively weak but the presumptive salt bridge R135-D158 could contribute to D158-mouse PrP and canine-PrP overall increased structural stability . The definitive selection of this substitution as a candidate for prion resistance came from its role on the surface of the PrP molecule and its influence over the basic area comprised of R135 , R150 , R155 and K119 . Assuming these residues participate in the PrP conversion to the disease associated isoform , D158 probably disturbs that PrP region as well as establishing a salt bridge with R135 limiting the latter’s role in conversion . This is supported by recent findings in transgenic Drosophila expressing mouse PrP with N158D substitution where it impairs the locomotor dysfunction developed when wild-type mouse PrP is expressed [61] . Taking advantage of the PMCA system that allowed examining if this amino acid could cause a significant change in the misfolding and propagation ability of known prion strains , two assays were performed . Based on the expression of mouse PrP with the desired N158D or N158E substitutions ( equivalent to position 163 in dog PrP numbering ) in cell cultures from neuronal origin , cell-PMCAs were performed to test their misfolding ability . Both studies showed clearly that the presence of an Asp or a Glu in position 158 of the PrP significantly hindered its misfolding propagation ability . Therefore , as predicted by surface charge distribution analysis , the presence of a negatively charged amino acid ( Asp or Glu ) is needed for a significant alteration in misfolding ability . As well as being unable to misfold , these mutants were able to block the propagation over wild-type mouse protein . This data highlights the potential for using these proteins as dominant negatives , blocking the propagation of prions , which has been a successful strategy for other dominant negative PrPs in cell cultures [62] , and becoming an efficient anti-prion therapy . Although the mechanism by which this proteins block prion propagation over wild-type proteins is unknown , our data suggests that dominant negative proteins may act by binding prion seeds or misfolded proteins and out-competing the PrP that can be misfolded . The results obtained in cell-PMCA experiments did not clearly distinguish which of the substitutions , N158D or N158E , gave rise to the PrP with lowest misfolding capacity as they showed almost identical behaviour . However , there was a faint PrPres signal when 22L was used to seed N158E PrP propagation , suggesting that at least for certain strains , N158D substitution may exert a major blocking effect than N158E . Both are negatively charged residues with similar molecular weights , so a similar effect would be expected in PMCA for N158D and N158E mutants . Nonetheless , a classification based on the environment at protein structures for the 20 amino acids [63 , 64] sets several groups for residues , in which both negatively charged residues show distinct features regarding their propensity to be involved on protein surfaces or in binding regions . Asp ( D ) shows similar tendency to both while Glu ( E ) shows a strong preference to be exposed to solvent on protein surfaces [65] . Analysis of electrostatic potentials on the surface for N158D PrP shows that it is located in a region surrounded by positively charged residues . Although the negative charge of both Asp ( D ) and Glu ( E ) suggests they could play a similar role in terms of surface charge distribution , the shorter side-chain of Asp ( D ) makes it more rigid within protein structures than Glu ( E ) , with a larger and more flexible side-chain . Alterations in surface electrostatic potential due to the negatively charged residues could reduce PrPc to PrPSc conversion efficiency , but also could affect the structure of the fibrils decreasing their formation . The rigidity of Asp ( D ) compared to Glu ( E ) , could therefore explain the slight differences observed in the PMCA results with N158D and N158E PrPs . The replacement of N158 by Asp ( D ) , with its restricted mobility , would possibly lead to a major disturbance of the surrounding area in order to accommodate the negative charge . In contrast , the larger side-chain of Glu ( E ) could allow the adoption of a wider range of conformers , possibly reducing the effect observed for 158D and allowing some conversion of N158E PrP in PMCA . Thus , we decided to choose Asp ( D ) instead of Glu ( E ) for the substitution in the PrP of the transgenic mouse model presented herein . Also taking into account that all the dog brain homogenates used in previous PMCA studies contained the D163 polymorphism . Thus , our results on the effect of D163 on the high resistance of canines to prion disorders would be applicable to some breeds of domestic dogs , although the similarities of Asp ( D ) and Glu ( E ) and the results obtained in brain-cell PMCA studies ( Fig 5 ) indicate that it might be applicable also to the domestic dog breeds bearing E163 polymorphism ( S2 Fig ) . The expression levels of the canine PrP transgene in both mouse lines was close to PrPC expression in the brain of wild-type animals , thus , the results obtained from their experimental challenges are unlikely to be affected due to over expression , which is known to accelerate prion disease development [66] . The three different mouse-adapted prion strains used ( RML , 22L and 301C ) did not cause clinical disease nor histological brain lesions in any of the challenged transgenic animals , indicating that this specific amino acid substitution prevented mouse PrP from misfolding in vivo , in agreement with the in vitro results . Additionally , the complete absence of PK resistant PrP , by Western blotting ( WB ) and immunohistochemistry ( IHC ) , compared to the extensive PrPres accumulation in wild-type mice inoculated with the same prion strains further supports the protective properties of this specific amino acid substitution . Despite the possible existence of subclinical carriers with minimal amounts of PrPSc , undetectable by conventional methods ( WB and IHC ) among the inoculated transgenic mice , a significant delay in the disease development was clearly demonstrated . Although the possibility of overcoming this polymorphic transmission barrier through serial in vivo passages cannot be ruled out from the results shown here , the high resistance observed to misfolding both in vivo and in vitro shows the significant effect of this substitution on reducing the misfolding proneness of the protein . Although performed in transgenic mice , these results probably explain the purported resistance canids have to TSEs despite having been exposed to infectious prions by consuming TSE affected animals such as sheep , cattle or cervids . Definitive proof of their resistance will require the generation of transgenic mice expressing whole dog PrP to allow the appropriate bioassays . The potential dominant negative effect of the mutated protein presented in this work over the wild-type protein ( ongoing study ) may determine their future as anti-prion therapies .
Brain homogenates ( 10−1 in PBS ) for use as seeds for PMCA or direct intracerebral inocula were prepared manually using a glazed mortar and pestle from brains of animals clinically affected by various TSEs: scrapie ( six isolates ) , BSE-C ( three isolates ) and atypical L-type BSE field cases supplied by the Laboratorio Central de Veterinaria ( Algete , Madrid , Spain ) ; SSBP/1 ( scrapie ) , ME7 , 22F , 22L , 87V , 22A , 79A and 139A supplied by Animal and Plant Health Agency ( APHA ) ( New Haw , Addlestone , Surrey , UK ) ; CWD ( 2 mule deer and 2 elk isolates ) supplied by the Department of Veterinary Sciences ( Laramie , WY , USA ) ; RML supplied by Rocky Mountain Laboratories ( Hamilton , MT , USA ) and sheep BSE supplied by Ecole Nationale Vétérinaire ( Toulouse , France ) . In vitro prion propagation and PrPres detection of amplified samples was performed as described previously with minor modifications [13 , 67] . Briefly , two different dog breed ( Cocker Spaniel and German Wirehaired Pointer ) brains ( provided by Veterinary Faculty from Autonomous University of Barcelona ) used for substrates were perfused ( immediately after euthanasia ) using PBS + 5 mM EDTA and the blood-depleted brains were frozen immediately until required for preparing the 10% dog brain homogenates ( PBS + NaCl 1% + 1% Triton X-100 ) . 50–60 μl of 10% dog brain homogenates , either unseeded or seeded with the corresponding prion strain were loaded onto 0 . 2-ml PCR tubes and placed into a sonicating water bath at 37–38°C without shaking . Tubes were positioned on an adaptor placed on the plate holder of the sonicator ( model S-700MPX , QSonica , Newtown , CT , USA ) and subjected to different procedures of 24 h serial rounds of PMCA: A1: Standard 1–10% seeded PMCA with serial 1:10 dilution rounds . B1: 30–50% seeded PMCA with serial 1:10 dilution rounds . B2: same as B1 with the addition of 1 mm zirconia/silica beads ( BioSpec Products ) , that have been shown to improve PrPC to PrPSc conversion in PMCA [52] . C1: 10% seeded PMCA with serial 1:2 dilution rounds . D1: 30–50% seeded PMCA with serial 1:2 dilution rounds . All the samples were incubated in cycles of 30 min followed by a 20 s pulse of 150–220 watts sonication at 70–90% of amplitude . An equivalent number of unseeded ( 6–12 duplicates ) tubes containing the corresponding brain substrate were subjected to the same number of rounds of PMCA in order to control cross-contamination and/or the generation of spontaneous PrPres . The detailed protocol for PMCA , including reagents , solutions and troubleshooting , has been published elsewhere [68] . Protease resistance assay: PMCA treated samples and 10% brain homogenates from prion inoculated mice were incubated with 85–200 μg/ml of Protease-K ( PK ) ( Roche ) for 1 h at 42°C with constant agitation at 450 rpm ( Thermomixer comfort Eppendorf ) as described previously [40] . Samples were mixed previously ( 1:1 , v/v ) with 10% Sarkosyl ( Sigma-Aldrich ) digestion buffer and the digestion was stopped by adding electrophoresis Laemmli buffer NuPAGE ( Invitrogen Life Technologies ) . PK-resistant PrP detection: Protein inmunodetection by Western blotting ( WB ) was performed after separating proteins by sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( SDS-PAGE ) . 4–12% NuPAGE Midi gels ( Invitrogen Life Technologies ) were used . The proteins were electroblotted onto nitrocellulose membranes ( Protran BA85 , GE Healthcare ) . Membranes were probed with mouse monoclonal antibody D18 ( 1:5 , 000–10 , 000 ) , POM1 ( 1:10 , 000 ) or Saf83 ( 1:400 ) ( Cayman Chemical ) and visualized with horseradish peroxidase–conjugated seconday antibody and chemiluminescence using the Super Signal West Pico kit ( Thermo Scientific Pierce ) . The digital images were displayed by FluorChem Q ( Alpha Innotech ) . In silico PrP mutant N158D models based on the mouse prion protein structures available at the Protein Data Bank archive ( www . rcsb . org ) [69] were generated using the SWISS-MODEL homology-model [70–74] server at ExPASy . Briefly , a template search with Blast [74] and HHBlits [75] was performed against the SWISS-MODEL template library and models were built using ProMod3 and alternatively PROMOD-II [76] . Following loop and side chains modeling , a last energy minimization step is performed using the OpenMM [77] molecular mechanics library . For structural analysis purposes two different models were chosen based on different template structures obtained from different experimental methods . Firstly , a NMR-based model ( Model01 ) from mouse prion protein fragment 121–231 at 37°C , PDB ID: 2L39 [78] and a second one based on X-ray crystallography ( Model02 ) from mouse prion protein structure complexed with Promazine PDB ID: 4MA7 [79] . Both models were those reporting best Global Model Quality Estimation ( GMQE ) [80] . Figures and molecular surfaces with electrostatic potential were produced with PyMol [81] . The genetic construct containing the mouse N158D substitution was carried out by PCR site-directed mutagenesis which first uses internal primers for the specific substitution: 5’ CATGTACCGCTACCCTGACCAAGTGTACTACAGGCC 3’ and 5’ GGCCTGTAGTACACTTGGTCAGGGTAGCGGTACATG 3’ . After isolation by PCR amplification using 5’ CCGGAATTCCGGCGTACGATGGCGAACCTTGGCTAC 3’ and 5’ CTAGTCTAGACTAGGCCGGCCTCATCCCACGATCAGGAAG 3’ as primers , the mouse N158D-PrP ORF was cloned into the pGEM-T vector ( Promega ) and excised from the cloning vector by using restriction enzymes BsiWI ( Thermo Fisher Scientific Inc . ) and FseI ( New England Biolabs Ltd . ) , and then inserted into a modified version of MoPrP . Xho vector [82] , as described previously [83] , which was also digested with BstWI and FseI . This vector contains the murine PrP promoter and exon-1 , intron-1 , exon-2 and 3’ untranslated sequences . The transgene was excised using NotI and purified with Invisorb Spin DNA Extraction Kit ( Inviteck ) according to the manufacturer’s recommendations . Transgenic mouse founders were generated by microinjection of DNA into pronuclei following standard procedures [83] . DNA extracted from tail biopsies was analyzed by PCR using specific primers for the mouse exon 2 and 3’ untranslated sequences ( 5’ GAACTGAACCATTTCAACCGAG 3’ and 5’ AGAGCTACAGGTGGATAACC 3’ ) . Those which tested positive were bred to mice null for the mouse PRNP gene in order to avoid endogenous expression of mouse prion protein . Absence of the mouse endogenous PRNP was assessed using the following primers: 5’ ATGGCGAACCTTGGCTACTGGC 3’ and 5’ GATTATGGGTACCCCCTCCTTGG 3’ . The mouse N158D PrP expression levels of brain homogenates from transgenic mouse founders were determined by Western blot using anti-PrP MAb Saf-83 ( 1:400 ) and compared with the PrP expression levels from wild-type mouse brain homogenates . To evaluate the effect of in vitro propagation of the residue 158 in mouse PrP , a standard 24 h PMCA , based on mouse brain homogenate mixed with cellular 3F4-tagged PrP as substrate , was performed using 1:40 dilution of RML or 22L as seeds . Four different cell lines expressing different types of 3F4-tagged PrPs were used: mouse N158 , D158 , E158 or without PrP ( Knock-out PrP cells ) . All of them were obtained by transient expression of the following plasmids on a PrP knock-out cell line ( PRNP0/0 mouse cell line ) , specifically , the PRNP0/0 hippocampus-derived HpL3-4 cell line ( except for the cells devoid of PrP , which were not transfected ) [84] . Briefly , mouse genomic DNA was extracted from the brain of a C57BL6 mouse using NucleoSpin tissue kit ( Macherey-Nagel ) and following the instructions . The ORF from wild-type mouse PrP was obtained by PCR using primers 5’ CCGGAATTCCGGCGTACGATGGCGAACCTTGGCTAC 3’ and 5’ CTAGTCTAGACTAGGCCGGCCTCATCCCACGATCAGGAAG 3’ . The PCR product was cloned into a pCMV vector ( Thermo Fisher Scientific Inc . ) by using EcoRI and NotI ( Thermo Fisher Scientific Inc . ) restriction enzymes . Site-directed mutagenesis was used to introduce the 3F4 tag in wild-type mouse PrP using internal primers 5’ CAAACCAAAAACCAACATGAAGCATATGGCAGGGGCTGCGGC 3’ and 5’ GCCGCAGCCCCTGCCATATGCTTCATGTTGGTTTTTGGTTTG 3’ . N158D and N158E mutations were also introduced by site-directed mutagenesis using internal primers 5’ GAAAACATGTACCGCTACCCTG 3’ and 5’ CTGGCCTGTAGTACACTTGGTC 3’ and 5’ GAAAACATGTACCGCTACCCTGAG 3’ and 5’ CTGGCCTGTAGTACACTTGCTC 3’ , respectively . For all the mutagenesis , the same external primers than those used for wild-type mouse PrP were used and all mutated PrP were cloned into pCMV by using EcoRI and NotI restriction enzymes . Non-PMCA amplified samples and samples subjected to one 24 h single round of PMCA were digested with 20 μg/ml of PK and analyzed by Western blot using monoclonal antibody 3F4 ( 1:10 , 000 ) . To evaluate the inhibitory effect of the residue 158 in mouse PrP during the in vitro propagation of a wild-type mouse PrP , a standard 24 h PMCA , based on mouse brain homogenate mixed with cellular PrP as substrate , was performed using 1:5 , 000 and 1:10 , 000 dilutions of RML as seed . Four different cell lines expressing different type of mutated PrPs were used: mouse N158 , D158 , E158 or without PrP ( Knock-out PrP cells ) . Non-PMCA amplified samples and samples subjected to one 24 h single round of PMCA were digested with 20 μg/ml of PK and analyzed by Western blot using monoclonal antibody D18 ( 1:10 , 000 ) . C57BL6 and TgN158D mice of 42–56 days of age were inoculated intracerebrally under gaseous anesthesia ( Isoflurane ) through the right parietal bone . A 50 μl SGC precision syringe was used with a 25 G gauge needle and coupled to a repeatability adaptor fixed at 20 μl . Buprenorphine ( 0 . 3 mg/kg ) was injected subcutaneously before recovery to consciousness to reduce post-inoculation pain . Mice were kept in a controlled environment at 22°C , 12 h light -darkness cycle and 60% relative humidity in HEPA filtered cages ( both air inflow and extraction ) in ventilated racks . The mice were fed ad libitum , observed daily and their clinical status assessed twice a week until neurological clinical signs appeared , after which they were examined daily . The presence of ten different TSE-associated clinical signs [82] was scored . Diseased animals were culled at the terminal stage of the disease by exposure to a rising concentration of carbon dioxide . Survival time was calculated as the interval between inoculation and culling or death . Brains were removed immediately after death , if euthanized/culled , and divided into two parts; a portion was stored at -80°C and the other fixed in 10% formalin for histological studies . Fixed tissues were dehydrated through increasing alcohol concentrations , through xylene and then embedded in paraffin-wax . Four micrometer sections were mounted on glass microscope slides and stained with hematoxylin-eosin for morphological evaluation . Further slides were mounted on 3-trietoxysilil-propilamine-coated glass slides for immunohistochemical studies . Immunohistochemistry ( IHC ) for disease associated PrP was performed . Briefly , deparaffinised sections were immersed in formic acid and boiled at low pH in a pressure cooker , endogenous peroxidases were blocked , the sections were pre-treated with Protease-K ( PK ) and incubated overnight with the primary antibody ( anti-PrP mAb 6H4 , Prionics AG , 1:1000 diluted in DAKO background reducing antibody diluent ) . Finally , the DAKO EnVision system was used along with 3 , 3’diaminobenzidine ( as the chromogen substrate ) to visualize any PrP deposits . For cellular PrP staining , the formic acid step was omitted and a standard heat induced epitope retrieval step ( 20 min at 95°C ) was performed using DAKO Target Retrieval Solution . The same antibody was used but at a 1:100 dilution . All experiments involving animals adhered to the guidelines included in the Spanish Legislative Decree “Real Decreto 1201/2005 de 10 de Octubre” on protection of animals used for experimentation and other scientific purposes , which transposed the European Directive 86/609/EEC on Laboratory Animal Protection . The project was approved by the Ethical Committee on Animal Welfare of the Laboratorio Central de Veterinaria ( project code assigned by the Ethical Committee CEBA-07/2010 ) was performed under its supervision .
|
Detection of individuals or whole species resistant to any infectious disease is vital to understand the determinants of susceptibility and to develop appropriate therapeutic and preventative strategies . Canids have long been considered resistant to prion infection given the absence of clinical disease despite exposure to the causal agent . Through extensive analysis of the canine prion protein we have detected a key amino acid that might be responsible for their universal resistance to prion disease . Using in vitro and in vivo models we demonstrated that the presence of this residue confers resistance to prion infection when introduced to susceptible animals , opening the way to develop a new therapeutic approach against these , at present , untreatable disorders .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"biotechnology",
"animal",
"types",
"animal",
"diseases",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"ruminants",
"immunology",
"vertebrates",
"pets",
"and",
"companion",
"animals",
"dogs",
"animals",
"mammals",
"animal",
"models",
"animal",
"prion",
"diseases",
"model",
"organisms",
"experimental",
"organism",
"systems",
"protein",
"structure",
"antibodies",
"zoology",
"genetic",
"engineering",
"research",
"and",
"analysis",
"methods",
"monoclonal",
"antibodies",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"zoonoses",
"genetically",
"modified",
"organisms",
"sheep",
"proteins",
"mouse",
"models",
"molecular",
"biology",
"genetically",
"modified",
"animals",
"biochemistry",
"eukaryota",
"physiology",
"biology",
"and",
"life",
"sciences",
"amniotes",
"organisms",
"macromolecular",
"structure",
"analysis",
"prion",
"diseases"
] |
2017
|
Unraveling the key to the resistance of canids to prion diseases
|
The HCV NS5A protein plays multiple roles during viral replication , including viral genome replication and virus particle assembly . The crystal structures of the NS5A N-terminal domain indicated the potential existence of the NS5A dimers formed via at least two or more distinct dimeric interfaces . However , it is unknown whether these different forms of NS5A dimers are involved in its numerous functions . To address this question , we mutated the residues lining the two different NS5A dimer interfaces and determined their effects on NS5A self-interaction , NS5A-cyclophilin A ( CypA ) interaction , HCV RNA replication and infectious virus production . We found that the mutations targeting either of two dimeric interfaces disrupted the NS5A self-interaction in cells . The NS5A dimer-interrupting mutations also inhibited both viral RNA replication and infectious virus production with some genotypic differences . We also determined that reduced NS5A self-interaction was associated with altered NS5A-CypA interaction , NS5A hyperphosphorylation and NS5A subcellular localization , providing the mechanistic bases for the role of NS5A self-interaction in multiple steps of HCV replication . The NS5A oligomers formed via different interfaces are likely its functional form , since the residues at two different dimeric interfaces played similar roles in different aspects of NS5A functions and , consequently , HCV replication . In conclusion , this study provides novel insight into the functional significance of NS5A self-interaction in different steps of the HCV replication , potentially , in the form of oligomers formed via multiple dimeric interfaces .
Hepatitis C virus ( HCV ) is a main causative agent associated with chronic liver diseases including chronic hepatitis , cirrhosis and hepatocellular carcinoma [1 , 2] . It is an enveloped , positive-stranded RNA virus belonging to the genus hepacivirus within the flaviviridae family [3] . A single polyprotein translated from the viral genome encodes structural proteins , including core , E1 , and E2 at the N-terminal domain followed by the viral assembly accessory proteins p7 [4 , 5] and NS2 [6–10] . The C-terminal domain encodes five different nonstructural proteins including NS3 , NS4A , NS4B , NS5A and NS5B , which comprise viral replicase complexes [11] and regulate viral assembly [12–17] . NS5A associates with membrane through its N-terminal amphipathic helix ( AH ) domain [18] . Following the AH domain are three major domains called domain I ( DI ) , DII , and DIII . These domains are separated by two low-complexity sequences ( LCS ) called LCSI and LCSII . In general NS5A-DI and DII were shown to play roles in HCV RNA replication [19–21] , and DIII was associated with virus particle assembly [16] . NS5A is a phosphoprotein expressed as a hypophosphorylated form , which is further phosphorylated to a hyperphosphorylated form [22 , 23] . The clusters of highly conserved residues at LCSI served as a target of casein kinase I-α ( CKI-α ) -mediated hyperphosphorylation , and blocking this inhibited the NS5A localization to the lipid droplets ( LD ) -associated , low-density membranes and impaired infectious virus production [17] . The casein kinase II-mediated NS5A-DIII phosphorylation was also shown to affect HCV particle assembly by regulating NS5A and core interaction [24 , 25] . Highly effective anti-HCV therapies are composed of different combinations of antiviral compounds targeting viral enzymes , such as NS3/4A protease and NS5B polymerase , and NS5A [26] . Since NS5A lacks enzyme activity , NS5A inhibitors were discovered via high throughput screening ( HTS ) of chemical libraries by using HCV replicon systems [27 , 28] . NS5A inhibitors are one of the most potent antivirals to date , inhibiting NS5A function during HCV replication at a picomolar range in cell culture-based systems ( reviewed in [29] ) . NS5A inhibitors impaired HCV RNA replication by preventing the formation of double membrane vesicles ( DMV ) that constitute HCV RNA replication factories [30–32] . NS5A inhibitors also blocked intracellular HCV particle assembly [33] . Both AH and DI domains of NS5A were indispensable for DMV formation [19] . Therefore , it is no surprise that NS5A inhibitors , which were shown to inhibit DMV formation , target these two domains , as evidenced by the accumulation of drug-resistant mutations in these areas [34] . Interestingly , many NS5A inhibitors are bivalent ( have two identical pharmacophores ) , and the bivalent form of compounds such as the iminothiazolidinone-based compound from Bristol-Myers Squibb ( BMS-824 ) are much more potent than the corresponding monovalent forms [27 , 35] . These observations fueled the speculation that bivalent NS5A inhibitors target the dimeric form of NS5A . In fact , multiple lines of evidence from different studies support that NS5A functions as a dimer or multimer . These include its dimeric crystal structures [36–38] and the detection of NS5A-NS5A interaction in vitro by using purified proteins [20 , 39] or in vivo in the NS5A ectopic expression system [40 , 41] and HCV-replication system [30 , 42] . In addition , Lim et al . showed that mutations introduced to four cysteine residues in NS5A ( C39 , C57 , C59 and C80 ) , shown to be involved in Zinc binding and required for HCV RNA replication [43] , inhibited purified NS5A self-interaction [20] . However , while these studies clearly demonstrated the correlation between NS5A dimerization and HCV RNA replication , the exact roles of NS5A dimerization in HCV life cycle is still unclear . In this study , we performed a structure-function analysis of highly conserved , surface-exposed NS5A-DI residues located at two different dimer interfaces , predicted from NS5A-DI crystal structures , on NS5A protein interactions and HCV replication . Our data indicate that these dimer interface residues are involved in NS5A self-interaction in Huh-7 cells and that NS5A self-interactions through these residues are critical for different steps of HCV replication by regulating NS5A hyperphosphorylation , subcellular localization and interaction with host protein CypA .
The genotype 1b ( gt1b ) NS5A-DI crystal structures depicted in Fig 1 , designated as 1ZH1 and 3FQM , suggested that the NS5A self-interaction occurs via at least two different interfaces [36 , 37] . To determine the relevance of these two dimeric forms of NS5A in its self-interaction in cells , we performed alanine-scanning mutagenesis of selected dimer-interface residues in full- length NS5A derived from gt1a ( H77 strain ) and gt2a ( JFH1 strain ) and determined the impact of these mutations on NS5A self-interaction . NS5A-DI residues 36 , 37 and 38 were chosen since they provide a continuous surface patch involved in 1ZH1-specific dimer interaction ( Fig 1A , see also S1A Fig showing the 1ZH1 residue interaction networks for detail ) . Residues 112 and 148 were selected since they could form a salt bridge at the 3FQM-specific dimer interface ( Fig 1B , see also S1B Fig showing the 3FQM residue interaction networks for detail ) . As shown in Fig 1C , residues 38 serine ( 38S ) , 112 arginine ( 112R ) and 148 aspartic acid ( 148E ) are completely conserved among 672 HCV sequences derived from all genotypes deposited in the Los Alamos HCV database . Residue 36 encodes major variant phenylalanine ( F ) or minor variant leucine ( L ) . Residue 37 encodes hydrophobic residues , including valine ( V ) , leucine ( L ) , phenylalanine ( F ) or isoleucine ( I ) . We determined the NS5A self-interaction by using a mammalian two-hybrid system as previously described [40 , 41] ( Fig 2 ) . In brief , NS5A was fused to the GAL4 DNA-binding domain ( GAL4/BD ) in pBIND vector and the herpes simplex virus VP16 activation domain ( VP16/AD ) in pACT vector . Then , these two vectors plus a third vector , pGL4 . 3[luc2P/Gal4UAS/Hygro] , encoding five GAL4 binding sites upstream of the firefly luciferase ( F-Luc ) gene , were transfected to Huh-7 cells . Two days following the transfection of these plasmids , cells were lysed to detect NS5A self-interaction efficiency by measuring F-Luc activity normalized by that of Renilla reniformis luciferase ( R-Luc ) expressed from a pBIND vector to adjust transfection efficiency . Under these experimental conditions , stronger protein-to-protein interaction resulted in higher F-Luc/R-Luc ratio . These ratios derived from the basal vectors ( pACT and pBind ) and the known interacting-pair [pACT-MyoD ( myogenic regulatory protein ) and pBind-ID ( negative regulator of myogenic differentiation ) ] were used as a protein-protein interaction negative ( - ) control and positive ( + ) control , respectively ( Fig 2A ) . Robust interaction was detected between a wild-type ( wt ) NS5A pair derived from a gt1a HCV H77 ( designated as H-wt ) , evidenced by the higher F-Luc/R-Luc value from H-wt pair than that from the ( + ) control . Interestingly , NS5A from a gt2a HCV JFH1 ( designated as J-wt ) showed an even stronger self-interaction than did H-wt ( Fig 2A ) . Next , we determined the impact of alanine mutations in the residue lining the 1ZH1 and 3FQM NS5A dimer interfaces , including 36A/37A/38A ( mutation group a ) and 112A/148A ( mutation group b ) , respectively . These H77 NS5A mutants , designated as H-a or H-b , showed a significantly reduced NS5A-NS5A interaction compared to that of H-wt ( Fig 2B ) . The mutants having both mutations , designated as H-ab , showed a stronger reduction in NS5A self-interaction ( Fig 2B ) . Similarly , JFH1 NS5A-NS5A interaction was also significantly reduced when these mutations were introduced separately or combined ( Fig 2C ) . In summary , these data suggest that NS5A from gt1a H77 and gt2a JFH1 form dimers/oligomers via at least two different dimeric interfaces in Huh-7 cells . To narrow down the critical residues involved in H77 NS5A self-interaction , we introduced individual alanine mutations to residues 36 , 37 , 38 , 112 or 148 in H77 NS5A two-hybrid vectors and designated them as H-36A , H-37A , H-38A , H-112A or H-148A , respectively . Then the effects of these individual mutations on NS5A self-interaction were determined by measuring the dual luciferase ( F-Luc and R-Luc ) activities from Huh-7 cells transfected with these plasmids as described above . As shown in Fig 3A , all mutations , with the exception of 38A , significantly reduced NS5A self-interaction down to the levels between ~10 to 50% of wt H77 NS5A-NS5A interaction efficiency on average . Interestingly , the 38A mutation significantly enhanced NS5A self-interaction ( ~30% above wt level on average ) . We speculate that 38A mutation altered the non-covalent interactions of a residue 157 arginine ( 157R ) , which was a wt 38S interactor , inadvertently stabilizing the NS5A inter-molecular interaction ( see S1 Fig ) . Next , we determined the effect of mutations modulating NS5A self-interaction on infectious gt1a H77D replication [44] , following electroporation of wt and mutant H77D RNAs to Huh-7 cells , then , determining the expression of viral proteins and RNAs at different time points post electroporation by western blot and quantitative RT-PCR analyses ( Fig 3B and 3C ) . Results indicated that altered H77 NS5A self-interaction inhibited viral replication ( Fig 3B and 3C ) . In detail , the 37A and 148A mutations caused moderate-to-substantial reductions in viral protein expression and RNA replication ( Fig 3B and 3C ) . The 36A and 112A mutations completely blocked viral replication as evidenced by undetectable viral protein expression and the lack of viral RNA increase over the replication-defective H77 mutant AAG [45] ( Fig 3B and 3C ) . The 38A mutation , which moderately enhanced NS5A self-interaction , also reduced HCV RNA replication ( Fig 3C , left panel ) . These data suggest that optimal mode of H77 NS5A self-interaction is critical for efficient gt1a virus RNA replication . Alternatively , 38A mutation could have disrupted other functions of H77 NS5A unrelated to NS5A self-interaction , or altered overall conformation of NS5A , resulting in H77D replication defect . To better understand the exact role of H77 NS5A residue 38 in HCV RNA replication , we have tried to obtain the revertant of H77D/38A mutant . However , our multiple attempts to obtain H77D/38A revertant were unsuccessful . Next , to test the effect of H77 NS5A self-interaction-altering mutations on infectious virus production , we determined the intracellular and extracellular infectivity titers following electroporation of H77D RNA having these individual mutations to Huh-7 cells . As expected , replication-defective H77D/36A or H77D/112A mutants showed no evidence of virus production ( Fig 3E ) . Other mutants , including H77D/37A , H77D/38A and H77D/148A showed severely impaired virus production ranging between ~10- to ~1 , 000-fold reductions in intracellular infectivity and ~3- to ~100-fold reductions in extracellular infectivity compared to those of the wt at the 72-h time point ( Fig 3E ) . None of the virus-producing H77 NS5A mutants impaired the secretion of virus particles , since the percentage of intracellular infectivity per total ( intracellular plus extracellular ) infectivity was similar between the wt H77D and these mutants ( Fig 3F ) . Supporting this , the ratio of extracellular and intracellular HCV RNA was similar between the wt H77D and its virus-producing NS5A mutants ( Fig 3D ) . On the other hand , virus particle-assembly efficiency was significantly reduced in these H77 NS5A mutants , since the relative levels of total infectivity per total HCV RNA in these mutants were significantly lower than that from wt H77D ( Fig 3G ) . Next , we performed a density gradient centrifugation of extracellular virus present in the cell culture supernatant to determine whether H77 NS5A mutations , including 37A , 38A and 148A , impaired the H77D infectivity by modulating virus particle density or specific infectivity . The results shown in Fig 4 indicate that the density distributions of wt H77D and three different NS5A mutant viruses are similar , since the relative percentages of HCV RNA and infectivity present in different density fractions are almost identical between the wt and mutants ( Fig 4B and 4D ) . We also did not detect any significant alteration in the specific infectivity of virus particles between the wt and mutants calculated as a ratio of infectivity per HCV RNA at different density fractions ( Fig 4E ) . Interestingly , the majority of infectious particles ( >95% ) from both wt and mutants banded at a density between 1 . 060 and 1 . 200 g/cm3 , centering around at 1 . 100 g/cm3 ( Fig 4B ) , whereas the peak of specific infectivity was detected at a density of ~1 . 060 g/cm3 ( Fig 4E ) . These data indicate that a small fraction of low-density gt1a virus particles is more infectious than the majority of virus particles , consistent with previous reports [46 , 47] . In summary , these results suggest that H77 NS5A self-interaction is critical for infectious particle-assembly efficiency , but has little impact on specific infectivity and density of infectious virus particles as well as their egress efficiency . We introduced individual alanine mutations to residues 36 , 37 , 38 , 112 or 148 in JFH1 NS5A two-hybrid vectors and designated them as J-36A , J-37A , J-38A , J-112A and J-148A , respectively . We transfected these two-hybrid plasmid sets for each mutant to Huh-7 cells to determine the impact of these individual mutations on JFH1 NS5A self-interaction . It is interesting that most of the mutations introduced individually to JFH1 NS5A showed relatively moderate defects in NS5A self-interaction , retaining between ~60 to 75% of wt JFH1 NS5A self-interaction on average , except for an 36A mutation , which caused severe defects in NS5A self-interaction ( ~20% of wt interaction ) . Unlike the 38A mutation in H77 NS5A , which increased NS5A self-interaction , the same mutation in JFH1 NS5A reduced NS5A self-interaction . In the structural perspective , the key inter-domain interactor of wt residue 38S in H77 NS5A is 157R , which provides both van der Waals and hydrogen-bond interactions to 38S ( S1A Fig ) . However , in JFH1 , the corresponding 38S interactor is 157 glutamine ( 157Q ) , which shows substantial physico-chemical differences from 157R . We speculate that different properties of residue 157 in H77 and JFH1 contributed a genotype-specific difference in the intermolecular interaction phenotypes of 38A mutation . Also contrary to 112A mutation in H77 NS5A , which caused severe defects in NS5A self-interaction , the equivalent mutation in JFH1 NS5A did not impair NS5A self-interaction in cells ( compare Figs 3A and 5A ) . The wt residue 112R in H77 NS5A is involved in a network of electrostatic interactions ( “salt bridge” pattern ) with the residues 48 arginine ( 48R ) and 148 glutamic acid ( 148E ) in the partnering NS5A resulting in intra/inter-molecular 48R-148E-112R interaction network ( S1B Fig ) . We speculate that the disruption of this electrostatic network mediated by the 112A mutation was substantial enough to impair the intermolecular H77 NS5A self-interaction . However , since the residue 48 alanine ( 48A ) in JFH1 NS5A does not support a similar kind of interaction network , it is reasonable to assume that a different kind of , genotype-specific interaction network surrounding the residue 112 may have negated the impact of 112A mutation on JFH1 NS5A self-interaction . Alternatively , R112 residue in JFH1 NS5A may not contribute to NS5A self-interaction unlike the same residue in H77 NS5A ( see discussion ) . To determine the effect of these individual JFH1 NS5A mutations on viral RNA replication , we introduced them to H77-JFH1 chimeric HCV ( HJ3-5 ) encoding JFH1 NS3-NS5B proteins [6 , 48] ( Fig 5B ) and then analyzed viral protein expression and HCV RNA levels following electroporation of these RNAs to Huh-7 cells . The results of these experiments could be summarized as follows . First , viral RNA replication was undetectable for HJ3-5/36A mutant ( Fig 5B and 5C ) . Thus , a 36A mutation in NS5A , which impaired both H77 and JFH1 NS5A self-interaction , blocked both H77 and JFH1 RNA replication ( Figs 3 and 5 ) . We also detected reduced replication of an HJ3-5/148A mutant , similar to the case of an H77D/148A mutant . The relatively moderate impact of 148A mutation on HJ3-5 replication , compared to its more severe effect on H77D replication , correlates with its weaker impact on self-interaction of JFH1 NS5A compared to that of H77 NS5A ( compare Figs 3 and 5 ) . Overall , these data indicate that JFH1 NS5A self-interaction is also critical for JFH1 replicase-mediated viral RNA replication . Second , both 37A and 38A mutants moderately increased the relative levels of intracellular HCV RNA compared to that of wt HJ3-5 at 72 h post electroporation ( Fig 5C , left panel ) . These results may indicate that these two mutations in JFH1 NS5A potentially enhanced HCV RNA replication . However , additional data indicate that this may not be the case . For example , we detected much lower levels of extracellular HCV RNAs from these mutants compared to those from the wt HJ3-5 ( Fig 5C , right panel ) . In addition , relative extracellular/intracellular HCV RNA ratios from these mutants were over 10-fold lower than that from the wt at the 72-h time point ( Fig 5D ) . Based on these data , we believe that decreased viral RNA secretion , rather than enhanced viral replication , has caused the relatively high levels of intracellular 37A or 38A mutant RNA accumulation . Third , an 112A mutation in JFH1 NS5A completely blocked HCV replication despite having no effect on its self-interaction . At a first glance , these data seem to contradict to the potential role of NS5A self-interaction in HCV replication However , previous study determined that HCV replication defect caused by 112A mutation could be attributed to the inhibition of NS5A-RNA interaction and dysregulation of NS5A’s role in HCV translation [49] . Fourth , the 37A , 38A and 148A mutations impaired the JFH1 NS5A hyperphosphorylation ( Fig 5B ) . This phenotype was also observed from the corresponding H77 NS5A mutants , although the hyperphosphorylation efficiency of H77 NS5A was quite low compared to that of JFH1 NS5A ( compare Figs 3B and 5B , see also [50 , 51] ) . In summary , these data indicate that JFH1 NS5A self-interaction also plays an important role in JFH1 replicase-mediated RNA replication , despite some genotype-specific differences . As expected from reduced viral RNA replication , the HJ3-5/148A mutant showed significantly reduced intracellular and extracellular infectivity during the entire time course of experiments ( Fig 5E ) . In the case of 37A or 38A mutants , while they also showed significantly lower intracellular and extracellular titers at the 24-h time point , by 48 h , their titers rapidly caught up with those from the wt HJ3-5 . However , by the 72-h time point , extracellular infectivity from these two mutants was significantly lower than that from the wt , while intracellular infectivity remained similar to that of the wt ( Fig 5E ) . The extracellular/intracellular RNA ratios of the 37A , 38A and 148A mutants were also significantly lower than that of wt HJ3-5 at the 48- and 72-h time points ( Fig 5D ) . In addition , the percentages of intracellular infectivity per total infectivity of all three mutants were significantly higher than that of the wt at 72 h ( Fig 5F ) . These results suggested a decreased egress of these mutant viruses compared to the wt virus . Interestingly , none of these mutants affected virus particle-assembly efficiency , since the relative ratios of total HCV infectivity per total HCV RNA were similar between wt and these mutants ( Fig 5G ) . Overall , it is remarkable that same mutations introduced to highly conserved residues 37 , 38 and 148 in NS5A showed genotype-specific effects on virus production in that H77 NS5A mutants impaired virus assembly , not viral egress , but JFH1 NS5A mutants impaired viral egress , not virus assembly ( Figs 3 and 5 ) . Next , we determined the density and specific infectivity of extracellular HJ3-5 wt as well as those of 37A , 38A and 148A mutant viruses by performing density gradient centrifugation . In general , comparable density distribution patterns were detected between wt HJ3-5 and mutant viruses , as judged from the relative density distributions of viral RNA and infectivity ( Fig 6 ) . However , the percentage of wt HJ3-5 in high-density fractions ( >1 . 201 g/cm3 ) was higher than those of mutants . In fact , ~12% of viral RNA from wt HJ3-5 was detected in these high-density fractions compared to ~3% from each of the JFH1 NS5A mutants ( Fig 6D , right panel ) . However , infectivity of wt HJ3-5 in these high-density fractions was low accounting for less than 2% of total infectivity ( Fig 6B , right panel ) . These results suggest that a significant portion of poorly infectious , high-density immature particles might have been secreted from wt HJ3-5-replicating cells , probably due to highly efficient virus egress ( Fig 5F , see discussion ) . On the other hand , relative titers of mutant viruses at low-density fractions ( <1 . 059 g/cm3 ) were 6- to 7-fold higher than those of the wt HJ3-5 ( Fig 6B , right panels ) . Due to this , the specific infectivity of mutant viruses at low-density fractions was relatively higher than that of the wt HJ3-5 ( Fig 6E ) . These results suggest that too efficient virus egress might have negative impact on infectious virus maturation ( see discussion ) . The interaction between NS5A and CypA is critical for HCV RNA replication [52] . Since most of the NS5A self-interaction mutants , especially those derived from gt1a H77D , significantly impaired HCV RNA replication , we asked whether reduced NS5A self-interactions in these mutants might have impaired NS5A-CypA interactions , resulting in decreased HCV RNA replication . To measure the interaction between NS5A and CypA quantitatively , we used a checkmate assay as this method successfully measured the interaction between NS5A and CypA in the previous study [41] . As shown in Fig 7A , the level of interaction between H77 NS5A and CypA was comparable to that of the positive ( + ) control . Interestingly , interaction between JFH1 NS5A and CypA was stronger than that between H77 NS5A and CypA ( Fig 7A ) . Next , we determined the interaction between CypA and H77 NS5A dimer-interface mutants . The results showed that the mutations that significantly reduced NS5A-NS5A interaction , including H-36A , H-37A , H-112A and H-148A ( Fig 3A ) , also significantly impaired the NS5A-CypA interaction ( Fig 7B ) . These results indicate that effective H77 NS5A self-interaction is critical for H77 NS5A and CypA interaction . These data also suggest that reduced NS5A-CypA interaction in H77 NS5A mutants was responsible for defective viral RNA replication ( Fig 3C ) . The NS5A self-interaction-enhancing H-38A mutant did not significantly reduce the NS5A-CypA interaction ( Fig 7B ) , which correlates with the relatively moderate effect of this mutation on viral RNA replication ( Fig 3C ) . The results of NS5A dimer-interface mutations on the interaction between JFH1 NS5A and CypA could be summarized as follows . First , the J-36A mutant that showed the most significant defect in JFH1 NS5A self-interaction ( Fig 5A ) also had the most substantial defect in the JFH1 NS5A and CypA interaction ( Fig 7C ) , which correlates nicely with the undetectable level of HJ3-5/36A RNA replication ( Fig 5C ) . Second , both J-37A and J-38A did not show any significant effect on JFH NS5A-CypA interaction ( Fig 7C ) , which also correlates with their relatively minor effects on NS5A self-interaction and viral RNA replication ( Fig 5A and 5C ) . Third , despite having no impact on JFH1 NS5A self-interaction ( Fig 5A ) , 112A mutation impaired the JFH1 NS5A-CypA interaction ( Fig 7C ) , suggesting that the role of J112R on NS5A-CypA interaction could differ mechanistically from other residues involved in NS5A self-interaction . Fourth , the J-148A mutant showed a reduced NS5A-CypA interaction ( Fig 7C ) , correlating with reduced J-148A self-interaction ( Fig 5A ) and impaired replication of the HJ3-5/148A mutant ( Fig 5C ) . In aggregate , these results suggest that NS5A self-interaction contributes to NS5A-CypA interaction , and that the impaired viral RNA replication observed from the majority of NS5A self-interaction-defective mutants could be due to a defective NS5A-CypA interaction . Hyperphosphorylation of NS5A was shown to contribute to infectious HCV production by regulating NS5A recruitment to low-density membranes in the vicinity of lipid droplets ( LD ) and facilitating NS5A-core interaction [17] . Since NS5A hyperphosphorylation , as well as infectious virus production , were reduced in NS5A self-interaction mutants , we asked whether these phenotypes were caused by impaired NS5A subcellular localization and/or its interaction with core protein . To facilitate the detection of the NS5A in an immunofluorescence assay and the NS5A-core interaction in a co-immunoprecipitation assay , we used HJ3-5/NS5AYFP , which encodes YFP-tag within the NS5A-DIII and is capable of virus production ( Fig 8A ) [6] . First , we confirmed that NS5A dimer interface mutations in HJ3-5/NS5AYFP also impaired NS5A hyperphosphorylation and virus production ( Fig 8A ) . To determine the NS5A subcellular localization , Huh-7 cells electroporated with either wt HJ3-5/NS5AYFP or its 37A , 38A and 148A mutants were subjected to confocal imaging analysis following a LipidTOX deep-red lipid staining to detect the LD . As shown in Fig 8B , we frequently detected a tight association between wt NS5A ( measured by YFP fluorescence ) and LD . However , in the case of NS5A mutants , a majority of NS5A was detected as the distinct foci in the cytoplasm without the tight LD association ( Fig 8B ) . In fact , significantly lower degrees of NS5A-LD co-localization were calculated from the mutants compared to those from the wt , based on Pearson’s correlation measurements derived from the confocal images obtained from ~30 different cells ( with the means of Pearson’s correlation coefficients for wt equaling 0 . 5206 versus 0 . 2598 , 0 . 2424 and 0 . 3245 for 37A , 38A and 148A mutants , respectively ) ( Fig 8C ) . The NS5A-core interaction was measured by two different methods: NS5A-core co-localization and co-immunoprecipitation ( co-IP ) . NS5A-core co-localization was determined by performing confocal imaging analysis following immunostaining of core by using core-specific antibody in cells replicating HJ3-5/NS5AYFP ( Fig 9A ) . As shown in Fig 9B , a strong degree of co-localization was detected between wt NS5A and core ( with a mean of Pearson’s correlation coefficient equaling 0 . 7726 ) . However , lesser degrees of co-localization between these two proteins were detected from the 37A , 38A and 148A mutants ( with the means of Pearson’s correlation coefficients equaling 0 . 5742 0 . 6564 and 0 . 6100 , respectively ) ( Fig 9B ) . Next , we determined the NS5A-core interaction by performing a GFP-pull down assay . As shown in Fig 9C , compared to wt , NS5A mutants showed reduced NS5A and core co-IP efficiency ( ~50% lower than wt ) expressed as a ratio of co-IP-core level per immunoprecipitated ( IP ) -NS5AYFP . These NS5A-core co-IP results correlate well with their co-localization data ( Fig 9A and 9B ) and indicate that NS5A-core interaction in NS5A mutants was reduced compared to that in wt . These results indicate that NS5A self-interaction regulates subcellular localization of NS5A and NS5A-core interaction . Since previous study showed the core-dependent recruitment of NS5A to LD-associated membranes [53] , it is possible that NS5A self-interaction is critical for its interaction with core , which then promotes NS5A localization to LD-associated membranes . Alternatively , NS5A self-interaction promoted the NS5A localization to LD-associated membranes , consequently enhancing the interaction between NS5A and core at these membranes . Interestingly , all three NS5A mutants defective in self-interaction also reduced the core localization to the LD ( Fig 9D ) . These results are consistent with recent study by Yin et al . , which showed the reduced core localization to the LD by using other NS5A mutants ( V67A or P145A ) , also defective in NS5A self-interaction [42] . To better understand the role of NS5A self-interaction in HCV replication , we attempted to isolate revertants with primary- or second-site mutations that could rescue viral replication . This was done by continuously sub-culturing the Huh-7 cells electroporated with H77D or HJ3-5 encoding different NS5A interface mutations and monitoring the viral replication every 3 days . Among mutants that showed no evidence of transient replication , including 36A and 112A mutants in H77D or HJ3-5 background , only the HJ3-5/36A mutant showed strong evidence of viral replication and infectious virus production by day 10 post electroporation of this viral RNA to Huh-7 cells . Sequencing of the entire coding region of secreted HJ3-5/36A-derived revertant collected at 28 days post-electroporation cell culture supernatants reveled a single mutation in JFH-1 NS5A at residue position 36 to valine ( 36V ) . This was not a wt reversion since the wt JFH1 NS5A residue in this position is a phenylalanine . Since , JFH1 NS5A/36A was defective in NS5A self-interaction ( Fig 5A ) and NS5A-CypA interaction ( Fig 7C ) , which , we believe , has caused defective HJ3-5/36A replication ( Fig 5B and 5C ) , we asked whether the 36V mutation is capable of restoring all of these defects associated with NS5A/36A mutation . To answer this question , we determined the effect of 36V mutation on the efficiency of JFH1 NS5A self-interaction and NS5A-CypA interaction by using checkmate assays as described above . As shown in Fig 10A and 10B , the 36V mutation in JFH1 NS5A significantly restored the 36A mutation-mediated impairments in NS5A self-interaction and NS5A-CypA interaction . Next , to verify that the 36V mutation in JFH1 NS5A was indeed responsible for the emergence of a replicable revertant from HJ3-5/36A mutant , we introduced the 36V mutation to HJ3-5 . The results shown in Fig 10C to 10F indicate that HJ3-5/36V substantially restored viral RNA replication and virus production . HJ3-5/36V also showed reduced virus secretion compared to wt HJ3-5 ( Fig 10E and 10G ) without affecting virus assembly efficiency ( Fig 10H ) . It is also notable that NS5A hyperphosphorylation in the HJ3-5/36V mutant was significantly lower than that in wt HJ3-5 , probably due to only a partial restoration of NS5A self-interaction by a 36V mutation in NS5A ( Fig 10A and 10C ) . In fact , all of these replication phenotypes of HJ3-5/36V revertant strikingly resemble those of HJ3-5/37A and HJ3-5/38A mutants . Overall , these results verified the concept that JFH1 NS5A self-interaction is critical for the NS5A-CypA interaction and NS5A hyperphosphorylation , which contribute to efficient HCV RNA replication and virus secretion . Next , we determined the density and specific infectivity of extracellular HJ3-5/36V . Overall , density profiles of viral RNA and infectivity of this mutant closely resembled to other replication-competent , NS5A self-interaction defective NS5A mutants , including 37A , 38A and 148A mutants ( compare Figs 6 and 11 ) , reflecting the fact that 36V mutant is a partial revertant showing significantly lower NS5A self-interaction than wt ( Fig 10A ) . Accordingly , similar to above three mutants , the relative titers of low-density ( <1 . 059 g/cm3 ) 36V revertant were ~10 folds higher than those of the wt HJ3-5 ( compare Figs 6B and 11B right panels ) . Due to this , the specific infectivity of 36V mutant viruses at low-density fractions was also higher than wt HJ3-5 ( Fig 11E ) , similar to other mutants . Interestingly , however , proportions of 36V mutant RNA and infectivity , respectively , detected at high density fractions ( >1 . 201 g/cm3 ) were much higher than those of other mutants , but comparable to those of wt HJ3-5 ( compare Fig 6B and 6D and Fig 11B and 11D , right panels , see below and also Discussion ) . To determine the effect of 36V mutation on NS5A subcellular localization , we introduced this mutation to HJ3-5/NS5AYFP , and then confirmed that HJ3-5/NS5AYFP /36V mutant is defective in NS5A hyperphosphorylation and virus production compared to wt , similar to the phenotypes of HJ3-5/36V ( Fig 12A and 12B ) . Confocal imaging analysis revealed that a majority of NS5AYFP/36V was detected as the distinct foci in the cytoplasm without the tight LD association ( Fig 12C ) , which is consistent with low degree of NS5AYFP and LD co-localization based on Pearson’s correlation measurements ( Fig 12D ) . Also we detected reduced degree of NS5AYFP and core co-localization as well as core-LD association ( Fig 12F and 12H ) . These NS5AYFP/36V phenotypes resembled those of other NS5AYFP mutants shown in Figs 8 and 9 , probably because they shared a common defect in NS5A hyperphosphorylation , which was shown to facilitate NS5A localization to the LD as well as NS5A and core co-localization [17] . However , uniquely to 36V mutant , we consistently detected small fraction of large NS5A foci , co-localizing with LD and core ( Fig 12C and 12E ) . Also , we detected wt level pull-down of core by NS5AYFP/36V ( Fig 12G ) , which , apparently , is contradictory to the reduced degree of NS5AYFP and core co-localization detected from this mutant compared to wt ( Fig 12F ) . Based on these data , we propose that 36V mutation in NS5A may have enhanced its affinity to core , partially compensating its LD localization defect caused by its defective hyperphosphorylation , resulting in fraction of NS5A recruitment to LD in core-NS5A interaction dependent manner . We speculate that LD-localized NS5A/36V may have contributed to near wt level of high-density particles detected from HJ3-5/36V mutant ( Fig 11B and 11D , right panels ) . However , further study will be needed to verify this point .
HCV NS5A is a multifunctional protein involved in both viral RNA replication and virus production [24 , 54 , 55] . Our study provided mechanistic insights regarding the roles of , crystal structure-defined , NS5A dimer interface residues in these two critical steps in the HCV life cycle . Specifically , our data revealed that these residues regulate NS5A self-interaction , NS5A-CypA interaction , NS5A hyperphosphorylation , NS5A localization to LD and NS5A-core interaction , promoting HCV replication and infectious HCV production . Among three domains of NS5A , DI plays a major role in NS5A self-interaction [39] . Currently three independent crystal structures of NS5A-DI , two from gt1b ( Con1 strain ) and one from gt1a ( H77 strain ) , are available [36–38] . While their monomeric structures were similar to each other with an average Cα RMSD ( root-mean-square deviation ) equal or less than 1 Å [37 , 38] , four distinct dimeric forms were detected in different crystal packing conditions , including the two forms from gt1b shown in Fig 1 [36–38] . Interestingly , these NS5A dimeric forms are not mutually exclusive but have a potential to form NS5A oligomers via multiple different interfaces [37 , 38] . Our data shown in Fig 2B support this possibility , since combining mutations located at two different interfaces additively reduced the level of NS5A self-interaction . Importantly , the fact that NS5A mutations located at different dimeric interfaces exhibited similar phenotypes , including their effects on NS5A hyperphosphorylation , NS5A-CypA interaction and NS5A subcellular localization , and , consequently , viral RNA replication and virus assembly/egress ( Figs 3 to 9 ) , strongly suggests the cooperative roles of different dimeric interactions within the same complexes . From a functional point of view , NS5A oligomerization is desirable for its role in promoting the formation of DMV [31 , 32] , which are sites of HCV RNA replication . In addition , the NS5A oligomerization model may be the best way to explain the high potency of NS5A inhibitors , which corresponds to one molecule of NS5A inhibitor impacting ~ 50 , 000 molecules of NS5A as in the daclatasvir example [56] , and a synergistic activity of different NS5A inhibitors in re-sensitization of drug-resistant NS5A variants [56] . The structure of gt2a JFH1 NS5A is currently unknown . However , a substantial NS5A-DI sequence difference exists between gt2a JFH1 and gt1b Con1 ( 69% amino acid homology ) . Thus , it was remarkable that two-to-five mutations introduced to JFH1 NS5A-DI residues located at positions corresponding to two different gt1b NS5A-DI dimer-interfaces significantly inhibited its self-interaction at the levels similar to those of gt1a H77 NS5A-DI ( compare Fig 2B and 2C ) , since gt1a H77 NS5A-DI is more homologous to gt1b NS5A-DI in sequence ( 82% amino acid homology ) and structure ( average Cα RMSD of 0 . 57 Å ) [36–38] . Interestingly , the impacts of individual NS5A-DI dimer-interface mutations on NS5A self-interaction were different between H77 and JFH1-derived NS5A ( Figs 3A and 5A ) . This difference was most apparent for R112A mutation , since this mutation severely impaired H77 NS5A self-interaction , while same mutation had no effect on JFH1 NS5A self-interaction . These results suggest that some intermolecular NS5A residue interactions might vary in these two HCV isolates due to differences in near neighbor residues . Supporting this interpretation , our preliminary data indicate that R112 residue in JFH1 NS5A may not participate in NS5A self-interaction , since neither ( similarly charged ) R112K nor ( oppositely charged ) R112E mutations affected this interaction ( S2 Fig ) . On the contrary , JFH1 NS5A self-interaction was severely impaired by E148R mutation , but unaffected by E148D mutation ( S2 Fig ) . These results suggest that E148 residue in JFH1 NS5A is involved in NS5A self-interaction via salt bridge formation , similar to E148 residue in H77 NS5A , but with different residue ( s ) instead of R112 . However , JFH1 NS5A-DI structure determination will be necessary to identify the exact residues involved in NS5A self-interactions at its dimer interface ( s ) . The NS5A mutation-mediated impairment in NS5A self-interaction correlated with HCV RNA replication-defects driven by either H77 NS5A or JFH1 NS5A-containing viral replicases ( Figs 3 and 5 ) . These results suggest that NS5A self-interaction is critical for HCV RNA replication regardless of HCV genotypes . Interestingly , we detected a strong correlation between NS5A self-interaction and NS5A-CypA interaction that was shown to be critical for HCV RNA replication [52] ( Figs 3 , 5 and 7 ) . Since the mutations we tested are located at NS5A-DI rather than at NS5A-DII that encodes the CypA interacting domain [57 , 58] , a direct role of these mutations in disrupting NS5A-CypA interaction is unlikely . Also , two mutations in NS5A ( D316E/Y317N ) , which conferred the CypA-independent HCV replication phenotype [58] , did not affect H77 NS5A self-interaction and slightly reduced JFH1-NS5A interaction ( S3A Fig ) . These results support that NS5A self-interaction is driving NS5A-CypA interaction , and not vice versa . Interestingly , NS5A-CypA interaction was detected only from GAL4/BD-CypA and VP16/AD-NS5A pairs and not in reverse configuration ( S4 Fig ) . We believe that these data support our interpretation that NS5A oligomers may interact with CypA , since we could easily envision a soluble VP16/AD-NS5A , not a DNA-bound GAL4BD-NS5A , forming a CypA-binding-competent oligomer . Now , how could NS5A self-interaction affect NS5A-CypA interaction ? We propose that NS5A self-interaction may modulate the orientation of CypA-binding region in NS5A-DII , likely within the context of oligomeric NS5A structure , allowing efficient CypA binding . Previous study by Lim et al . showed that mutating zinc-binding cysteines to alanine ( C to A ) within NS5A-DI ( C39A , C57A , C59A and C80A ) disrupted NS5A self-interaction and HCV RNA replication by using bacterially expressed and purified proteins [20] . These data support the role of NS5A self-interaction in HCV replication . However , NS5A-CypA interaction was not affected by any of eleven C to A mutations within full length NS5A ( including the four zinc binding residues mentioned above ) , regardless of their impact on NS5A self-interaction [20] . These findings are different from our data , which showed positive correlation between NS5A self-interaction and NS5A-CypA interaction . We speculate that potential difference in NS5A oligomeric states between current and previous experimental systems might have altered the availability of CypA-interacting region in NS5A-DII for CypA binding leading to different outcomes . Our data indicated that impaired gt1a H77 NS5A self-interaction resulted in virus particle assembly defects ( Fig 3G ) , while that of gt2a JFH1 NS5A resulted in virus secretion defects ( Fig 5F ) . Although this genotypic difference in regard to assembly versus secretion is difficult to understand , we could envision that more than 10 folds higher viral replication from HJ3-5 ( encoding JFH1 NS5A ) , compared to H77D ( encoding H77 NS5A ) , could have altered the relative roles of NS5A self-interactions in viral assembly/egress processes . Alternatively , H77 NS5A and JFH1 NS5A contributed to HCV assembly/egress via intrinsically different mechanisms . Regardless , this genotypic difference was not due to the chimeric nature of HJ3-5 , which encode H77 core to NS2 in the background of JFH1 , since NS5A dimer interface mutations introduced to full length JFH1/QL showed exactly same virus secretion defect as did HJ3-5 mutants ( S5 Fig ) . Interestingly , while H77 NS5A mutants showed no effect on virus density or specific infectivity ( Fig 4 ) , JFH1 NS5A mutants increased the proportion of low-density , high-infectivity particles ( Fig 6 ) . It seems paradoxical that the specific infectivity of JFH1 NS5A mutants was higher than that of wt considering the reduced overall infectivity of mutants . However , it is possible that slowed viral egress in JFH1 NS5A mutants allowed their enhanced lipidation , resulting in low-density , highly infectious viruses , while this type of slow maturation of wt virus was relatively decreased due to efficient virus secretion ( Fig 5D and 5F ) . Interestingly , LD localization of JFH1 NS5A dimer interface mutants was reduced , indicating that virus maturation into low-density particles may not strictly depend on a tight association of NS5A with the cytoplasmic LD . Alternatively , efficient LD localization of NS5A in wt HJ3-5-replicating cells could have enhanced virus secretion at a level to over-saturate the cells’ capacity for normal virus maturation , consequently , forcing the significant portion of viral particles to quickly egress as immature forms ( Fig 6 ) . It is important to note that not all viral RNA secreted to the supernatant of HCV replicating cells is associated with infectious virus . In fact , Gastaminza et al . showed that HCV replicating cells released the low-density particles , including the exosome-like large vesicles , and high density particles , most likely representing non-enveloped core particles , in addition to majority of intermediate density particles corresponding to enveloped HCV particles [59] . Accordingly , low infectivity ( relative to viral RNA ) of both high- and low-density particles detected in our study could be attributed to these defective-particles , including exosomes and non-enveloped core particles . In this context , it is interesting to note that relative proportion of secreted , minimally infectious , high density viral particles from 36V revertant was significantly higher than those from 37A and 38A mutants ( compare Figs 6 and 11 ) , despite that most of other phenotypes of 36V revertant were similar to those of 37A and 38A mutants consistent with their similarly defective NS5A self-interaction . The second phenotypic difference between 36V revertant and these other mutants was the affinity of core and JFH1 NS5A , which was unchanged in the 36V revertant , but reduced in the 37A and 38A mutants , compared to that of wt ( compare Figs 9C and 12G ) . Based on these data , we propose that high-affinity interaction between core and NS5A , independent from NS5A self-interaction , could promote the secretion of defective , high-density core particles . Previously Miyanari et al . demonstrated that two different triple alanine mutations introduced to the residues 99–101 and 102–104 within NS5A-DI reduced NS5A localization to LD , establishing the role of NS5A-DI on NS5A localization to the LD [53] . Subsequently , Masaki et al . showed that NS5A hyperphosphorylation promote its LD localization [17] . Now , our data suggest that NS5A dimer-interface residues in NS5A-DI contribute to NS5A localization to the LD by regulating NS5A hyperphosphorylation ( Figs 5 and 8 ) . Interestingly , our extended study by using HCV polyprotein expression system indicate that all of NS5A dimer interface mutants that we tested , including 36A , 37A , 38A , 112A and 148A , regardless of their impact on NS5A self-interaction , could impair JFH1 NS5A hyperphosphorylation ( S6 Fig ) . These data suggest that NS5A self-interaction per se may not be sufficient to promote its hyperphosphorylation . A recent study by Ross-Thriepland and Harris [23] indicated that JFH1 NS5A-DI residue 146 serine ( 146S ) is a target of phosphorylation and mutating this residue to phosphomimetic aspartic acid ( 146D ) led to decreased NS5A hyperphosphorylation . Since 146S is located near the 3FQM dimer interface in the vicinity of the 112R-148E salt-bridge , the authors predicted that phosphorylation at 146S might potentially regulate NS5A dimerization [23] . However , our data showed that 146A or 146D mutations did not significantly affect H77 or JFH1 NS5A self-interaction ( S3B Fig ) . These data indicate that 146D mutation impaired NS5A hyperphosphorylation without affecting NS5A self-interaction similar to the phenotypes of 112A mutation ( Fig 5A and S6 Fig ) . Consistent with these data , the NS5A inhibitors also reduced NS5A hyperphosphorylation [60] , yet they did not affect NS5A self-interaction [40] . Intriguingly , NS5A inhibitors were implicated to affect intermolecular NS5A conformation [56 , 61] , and phosphorylation of NS5A residue 146 has a potential to alter local dimeric conformation [23] . Thus , these data may indicate that hyperphosphorylation of NS5A is dependent on its specific conformation that allows its interaction with kinases [such as casein kinase I-α ( CKI-α ) ] involved in this process [17] . Accordingly , we propose that only a defined conformation of NS5A , mostly likely within the oligomeric complexes , permits the access of kinases to NS5A LCSI domain resulting NS5A hyperphosphorylation [17 , 62] , and disturbing the kinase-accessible conformation of NS5A either by different NS5A mutations or treatment with NS5A inhibitors impairs NS5A hyperphosphorylation . It is possible that all or some of our mutants may have impacted HCV RNA replication and virus assembly by altering NS5A conformation or other functions of NS5A , in addition to altering NS5A self-interaction-mediated functions . However , the ultimate proof supporting the role of NS5A self-interaction in HCV RNA replication and virus production was provided by a 36 valine ( 36V ) revertant mutation in JFH1 NS5A , which replaced the original alanine mutation that conferred a severe defect in JFH1 NS5A self-interaction . This JFH1 NS5A/36V mutation , not only significantly restored the NS5A self-interaction , but also restored NS5A-CypA interaction , and HJ3-5/36V RNA replication and infectious virus production . Interestingly , 36V mutation introduced to H77 NS5A also partially enhanced NS5A self-interaction as well as an NS5A-CypA interaction ( S7 Fig ) . These data support the notion that 36V mutation in JFH1 NS5A was indeed selected to rescue the NS5A self-interaction-defect caused by 36A mutation . However , the H77D/36V mutant did not show detectable level of viral RNA replication . The exact reason for impaired replication of H77D/36V is unclear . However , it is interesting to note that H77 NS5A residue F36 points toward lipid phase in the 1ZH1 structure ( Fig 1A ) , suggesting that F36 may contribute to NS5A and membrane interaction . Based on this , we speculate that 36V mutation may have failed to restore the F36’s additional function at the NS5A-membrane interface and , as a consequence , could not rescue H77D replication . This potential , additional role of F36 residue in H77D replication may also explain the reason H77D/36A mutant could not replicate when 148A mutant showed low level replication ( Fig 3C ) , despite their similar NS5A self-interaction defects ( Fig 3A ) . As illustrated in Fig 13 , we propose that high-order NS5A oligomerization stabilizes NS5A conformation at local domains , including the DII and LCSI , which would allow them to interact with host proteins including CypA [63 , 64] and CKI-α [17 , 65] . The CypA-induced modulation of NS5A would then promote DMV formation , which will harbor HCV replication complexes [19 , 30–32] . This early event will be followed by NS5A hyperphosphorylation by CKI-α [17 , 22 , 66] , which would be promoted by interaction between NS5A and other HCV nonstructural proteins in the replication complexes as demonstrated in previous reports [67 , 68] . Then the hyperphosphorylated NS5A ( in the replication complexes ) will move to low-density membrane domains near LD [53] , interact with core protein , and promote HCV assembly and/or virus egress [17] . A recent pulse-chase imaging study by Wang and Tai strongly supports this scenario , since they determined that the NS5A-associated organelles ( replication complexes ) are continuously generated de novo and NS5A in the aged version of these organelles tend to associate with LD , accompanied with increases in NS5A hyperphosphorylation [66] . In conclusion , our study provided novel insights indicating that NS5A may function as oligomers formed via multiple dimeric interfaces to promote HCV RNA replication and virus production . It is likely that NS5A functions requiring its oligomerization make it an excellent target of highly potent inhibitors [29] . Although NS5A inhibitors did not perturb NS5A dimerization per se [20 , 40 , 69] , it is suggested that they might have modulated its conformation [70] or higher-order NS5A oligomerization state [56] . In the future , understanding the detailed mechanistic function of NS5A oligomers during HCV replication , combined with determining the exact mode of action of NS5A inhibitors , will provide insights into understanding HCV replication mechanisms and for improving/identifying potent antivirals against HCV and other agents that utilize proteins functioning in an equivalent manner .
The construction of H77D and HJ3-5 was described previously [44 , 48] . To generate the pairs of pACT and pBind vectors expressing full-length NS5A from two different genotypes , H77 and JFH1 NS5A sequences were PCR amplified from H77D and HJ3-5 with the primer sets introducing SgfI and PmeI restriction enzyme sites at their N- and C-terminus , respectively , and then cloned into pFN10A ( ACT ) Flexi vector and pFN11A ( BIND ) Flexi Vector digested with SgfI and PmeI enzymes ( Promega , WI , USA ) . NS5A mutations were introduced by using the QuikChange II XL site-directed mutagenesis kit ( Agilent Technology , Santa Clara , CA ) . The sequences of regions manipulated within each plasmid were verified by DNA sequencing . Other plasmids used for vector control ( pACT and pBind ) , positive controls ( pACT-MyoD and pBind-ID ) and pGL4 . 3[luc2P/Gal4UAS/Hygro] were provided via a Checkmate/Flexi Mammalian Two-Hybrid System ( Promega , WI , USA ) . Huh-7 cell lines used in this study including Huh7 . 5 ( a clonal cell line of Huh-7 , kindly provided by Dr . Charles M . Rice at Rockefeller University [71] ) and FT3-7 ( a clonal cell line of Huh-7 as described in [72] ) were maintained in Dulbecco's modified Eagle medium ( DMEM ) ( Invitrogen , Carlsbad , CA ) containing 10% fetal bovine serum ( Invitrogen , Carlsbad , CA ) at 37°C in a 5% CO2 atmosphere . The interactions between NS5A-NS5A or NS5A-CypA were evaluated by using a Checkmate Mammalian Two-Hybrid System ( Promega , WI , USA ) . In brief , pACT and pBIND based plasmids along with the pGL4 . 3[luc2P/Gal4UAS/Hygro] reporter plasmid were co-transfected into FT3-7 cells by using TransIT-LT1 ( Mirus , Madison , WI ) reagent according to the manufacturer’s instructions at a ratio of 3 μl transfection reagent per 1 μg of plasmid DNA . At 48 h post transfection , cell lysates were prepared to assess the firefly and Renilla luciferase activities by using a Dual-Luciferase Reporter ( DLR ) Assay System ( Promega , WI , USA ) and GloMax DISCOVER instrument ( Promega , WI , USA ) according to the manufacturer’s instructions . HCV RNA was transcribed in vitro from linearized HCV cDNA by using the T7 MEGAscript Kit ( Life Technologies , Carlsbad , CA ) and purified by using an RNeasy RNA isolation kit ( Qiagen , Valencia , CA ) . In brief , 5 x 106 FT3-7 cells were mixed with 10 μg of HCV RNA in a 4-mm cuvette and pulsed once at 270 V and 950 μF by using a Gene Pulser System ( Bio-Rad , Hercules , CA ) . Electroporated cells were transferred into 12-well plates for HCV RNA analysis and or 6-well plates or 6 cm dishes for virus titration and protein analysis . Extracellular and intracellular HCV titers in clarified cell culture supernatants and 4-cycle , freeze-thaw cell lysates harvested from FT3-7 cells at different time points post-electroporation of HCV RNA were determined by performing an HCV core antigen immunfluorescence assay as described before [73] . HCV RNA in cell culture supernatants and gradient fractions ( see below-density gradient ultracentrifugation ) was harvested by using a QIAamp viral RNA mini kit ( Qiagen , Valencia , CA ) . Cell-associated HCV RNA was harvested by using an RNeasy RNA isolation kit ( Qiagen , Valencia , CA ) . To quantitate the level of HCV RNA , a real-time RT-PCR assay was performed by using a QuantiNova Probe RT-PCR Kit ( Qiagen , Valencia , CA ) and a CFX96 real-time system ( Bio-Rad , Hercules , CA ) with custom designed primer probe sets ( Sense primer: HCV84FP , 5’-GCCATGGCGTTAGTATGAGTGT-3’; antisense primer: HCV 303RP , 5’-CGCCCTATCAGGCAGTACCACAA-3’; and probe: HCV146BHQ , FAM-TCTGCGGAACCGGTGAGTACACC-DBH1 ) . Briefly , 10 μl 2x QuantiNova Probe RT-PCR Master Mix , 1 μl each of 20 μM sense- and antisense primers , 0 . 16 μl of 20 μM HCV-specific probe , 0 . 2 μl of 100x QuantiNova Probe RT Mix , 4 μl of template RNA and RNase-free water were combined to make 20 μl reaction mixtures . HCV RNA was reverse transcribed for 10 min at 45°C followed by 5 min incubation at 95°C to activate PCR polymerase , then PCR was performed for 30 cycles of 95°C for 5 seconds ( denaturation ) and 60°C for 30 seconds ( annealing and extension ) . Cell lysates were prepared in 1% CHAPS in PBS lysis buffer containing 1x protease- and phosphatase inhibitor cocktail mix ( GenDEPOT , Katy , TX ) , separated by SDS-PAGE and transferred onto PVDF membranes . The membrane was blocked and probed with primary antibodies to core protein ( 1:2 , 000 dilution of C7-50 , Thermo Scientific , Rockford , IL ) , NS3 ( 1:2 , 000 dilution of 9-G2 , ViroGen , Watertown , MA ) , NS5A [1:15 , 000 dilution of 9E10 ( kindly provided by Dr . Charles M . Rice at Rockefeller University ) or 1:2000 dilution of 2F6 , BioFront Technologies , Tallahassee , FL] , and NS5AYFP ( 1: 2000 dilution of anti-GFP , Life Technologies ) and tubulin ( 1:7000 dilution , EMD Millipore , Billerica , MA ) . Protein bands were visualized by incubating the membranes with IRDye Secondary antibodies ( Li-Cor Biosciences , Lincoln , NE ) , followed by imaging with an Odyssey infrared imaging system ( Li-Cor Biosciences , Lincoln , NE ) . Approximately 1 . 5 x 107 FT3-7 cells were electroporated with in vitro-transcribed RNAs and seeded into 175cm2 flasks . The HCV containing cull culture supernatants were collected from 48 to 72 h for every 4–6 h , pooled and centrifuged to remove cell debris . The clarified supernatants were loaded onto Centricon Plus-70 ( Millipore , Germany ) , concentrated by centrifugation at 3 , 500 x g at 4°C and subjected to discontinuous Optiprep gradient centrifugation ( 60 , 45 , 30 , and 15% ) for 16h at 120 , 000 x g at 4°C in a SW55Ti rotor ( Beckman , Indianapolis , IN ) . Each of 450 μl fraction was collected by aspiration from the top of the gradient and analyzed to determine its density , infectivity and amounts of HCV RNA as described above ( see also [12] ) . Cell lysates were prepared in 1 ml of lysis buffer [0 . 5% Triton X-100 , 10mM Tris-HCl ( pH 7 . 5 ) , 150mM NaCl] containing 1x protease- and phosphatase inhibitor cocktail mix ( GenDEPOT , Katy , TX ) and incubated on ice for 1h . Cell lysates were incubated with anti-GFP magnetic beads ( Miltenyi Biotech , Auburn , CA ) for 1h at 4°C with gentle mixing and applied to μ columns . Magnetic beads were washed 4 times each with lysis buffer , and wash buffer I ( 150mM NaCl , 1%NP-40 , 0 . 5% sodium deoxycholate , 0 . 1% SDS , 50mM Tris-HCl , pH 8 . 0 ) , respectively , followed by one wash with buffer 2 ( 20mM Tris-HCl , pH 7 . 5 ) . Bound immune complexes were eluted from columns by applying a preheated SDS sample buffer . HCV RNA electroporated cells were plated on 8-well chamber slides ( BD Bioscience , Bedford , MA ) at a density of 1x104 cells per well . Two to three days later , the slides were washed with PBS , fixed with 4% formaldehyde for 20 min at room temperature , and permeabilized with 0 . 2% Triton X-100 in PBS for 10 min , then incubated overnight at 4 oC with anti-core monoclonal antibody ( 1:2 , 000 dilution of C7-50 , Thermo Scientific , Rockford , IL ) , followed by Alexa Fluor 405-conjugated goat anti-mouse antibody ( 1:1000 dilution , Invitrogen , Carlsbad , CA ) for 1 h . Lipid droplets were stained with HCS LipidTOX deep red neutral lipid stain ( 1:1000 dilution , Molecular Probes Inc , Eugene , OR ) . The slides were examined with an Olympus FluoView FV10i confocal microscope ( Olympus America Inc , Waltham , MA ) . Pearson’s coefficient was obtained by using FV10i-ASW 4 . 2 viewer software . Student’s t-test ( unpaired ) was performed by using GraphPad Prism version 6 software to determine the significance in differences between paired values . A P value less than 0 . 05 was considered statistically significant .
|
HCV NS5A is a multifunctional protein involved in both viral RNA replication and infectious virus production , and is a target of one of the most potent antivirals available to date . However , the mode of action of NS5A inhibitors is still unclear due to the lack of mechanistic detail regarding NS5A functions during HCV life cycles . In this study , we have provided evidence that surface-exposed NS5A residues involved in two different dimeric interactions in crystal structures are indeed involved in NS5A self-interactions in cells . We also showed that these NS5A residues play critical role in HCV RNA replication and infectious virus production by regulating NS5A hyperphosphorylation , its subcellular localization and its interaction with host protein CypA . Overall , our data support the functional significance of “NS5A oligomers” formed via multiple interfaces in HCV replication . We speculate that the NS5A inhibitors exploited the NS5A oligomer-dependent functions during HCV replication , rather than targeting individual NS5A , which consequently resulted in their high potency .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"dimers",
"(chemical",
"physics)",
"microbial",
"mutation",
"medicine",
"and",
"health",
"sciences",
"crystal",
"structure",
"pathology",
"and",
"laboratory",
"medicine",
"intracellular",
"pathogens",
"pathogens",
"biological",
"cultures",
"condensed",
"matter",
"physics",
"microbiology",
"viral",
"structure",
"viruses",
"rna",
"viruses",
"cell",
"cultures",
"crystallography",
"research",
"and",
"analysis",
"methods",
"solid",
"state",
"physics",
"viral",
"core",
"chemistry",
"viral",
"replication",
"physics",
"virology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"chemical",
"physics",
"organisms"
] |
2018
|
HCV NS5A dimer interface residues regulate HCV replication by controlling its self-interaction, hyperphosphorylation, subcellular localization and interaction with cyclophilin A
|
Altered cell metabolism is inherently connected with pathological conditions including cancer and viral infections . Kaposi's sarcoma-associated herpesvirus ( KSHV ) is the etiological agent of Kaposi's sarcoma ( KS ) . KS tumour cells display features of lymphatic endothelial differentiation and in their vast majority are latently infected with KSHV , while a small number are lytically infected , producing virions . Latently infected cells express only a subset of viral genes , mainly located within the latency-associated region , among them 12 microRNAs . Notably , the metabolic properties of KSHV-infected cells closely resemble the metabolic hallmarks of cancer cells . However , how and why KSHV alters host cell metabolism remains poorly understood . Here , we investigated the effect of KSHV infection on the metabolic profile of primary dermal microvascular lymphatic endothelial cells ( LEC ) and the functional relevance of this effect . We found that the KSHV microRNAs within the oncogenic cluster collaborate to decrease mitochondria biogenesis and to induce aerobic glycolysis in infected cells . KSHV microRNAs expression decreases oxygen consumption , increase lactate secretion and glucose uptake , stabilize HIF1α and decreases mitochondria copy number . Importantly this metabolic shift is important for latency maintenance and provides a growth advantage . Mechanistically we show that KSHV alters host cell energy metabolism through microRNA-mediated down regulation of EGLN2 and HSPA9 . Our data suggest that the KSHV microRNAs induce a metabolic transformation by concurrent regulation of two independent pathways; transcriptional reprograming via HIF1 activation and reduction of mitochondria biogenesis through down regulation of the mitochondrial import machinery . These findings implicate viral microRNAs in the regulation of the cellular metabolism and highlight new potential avenues to inhibit viral latency .
Viruses are the etiological agents in approximately 12% of human cancers . Most of these cancers can be attributed to infections by human papillomavirus ( HPV ) , hepatitis B virus ( HBV ) , hepatitis C virus ( HCV ) , Epstein-Barr virus ( EBV ) , and Kaposi's sarcoma-associated herpesvirus ( KSHV ) [1] , [2] . KSHV is the etiological agent of Kaposi's sarcoma ( KS ) and it is also causally linked to primary effusion lymphoma ( PEL ) and a subset of multicentric Castleman's disease [3]–[5] . KSHV , like other herpesviruses , can enter a latent viral program after multiple rounds of replication and infection of new target cells [6] . In traditional models of herpesvirus-induced tumorigenesis , latency has the primary role in oncogenesis , promoting cell proliferation and impairing apoptosis . The lytic cycle is not considered to directly contribute to oncogenesis , but plays an earlier role by allowing viruses to disseminate in the host and to infect the target cells [6] . Consistently , in KS tumors and in PEL , the majority of cells are latently infected and express only a subset of viral genes located within the latency-associated region . This includes the viral-encoded cyclin ( vCyclin ) , FLIP ( vFLIP ) , latency-associated nuclear antigen ( LANA ) and 12 microRNAs ( miRNAs ) ( to express 17 mature miRNAs ) [7]–[12] . miRNAs are regulatory RNAs expressed by animals , plants and some viruses [13] , [14] . They are synthesized as precursors that fold into imperfect double strand RNA hairpins . This structure is cleaved in two steps , which are catalysed by Drosha and the endoribonuclease Dicer RNaseIII nucleases . This process results in a ∼22 base pair miRNA duplex . One strand of this duplex can be incorporated into the RNA induced silencing complex ( RISC ) as a mature miRNA . Within RISC , miRNAs are bound by Argonaute ( Ago ) proteins and induce repression of mRNAs bearing sequences with partial complementary to the miRNA . Ten out of twelve of the KSHV miRNAs ( miR-K12-1 to 9 , and 11 ) are located within the intron of K12 and are expressed as a cluster [15] . Although the expression levels of these miRNAs varies between different cells lines and KS samples , it has been consistently shown that all 10 miRNAs are expressed together in latently infected cells [16]–[19] . Recent studies have suggested that approximately half of all human miRNAs are expressed and function as clusters; targeting the same mRNA or different mRNAs involved in the same pathway [20]–[22] . Recent cross-linking immunoprecipitation ( CLIP ) experiments in PEL cell lines have indicated that this also might apply to the KSHV miRNAs [16] , [17] . This suggests that these miRNAs could have developed in the course of viral evolution to function as a cluster during latent infection . The characteristic metabolic hallmark of tumor metabolism is aerobic glycolysis; In contrast to normal differentiated cells , which rely primary on mitochondrial oxidative phosphorylation to generate energy , most cancer cells instead rely on aerobic glycolysis , a phenomenon termed as the Warburg effect . Cancer cells metabolism is a result of the modulation of intracellular signaling pathways that are disrupted by mutated oncogenes and tumor suppressors . Moreover alteration in cell metabolism may trigger tumorigenesis [23] . Viruses do not inherently have their own metabolic output . However , upon infection , viruses dramatically alter the metabolism of the host cell . This can provide substrates necessary for viral replication and is also likely to be important for pathogenesis . Human cytomegalovirus ( HCMV ) , HCV , human immunodeficiency virus ( HIV ) , herpes simplex virus ( HSV ) and KSHV have all been shown to alter cell host metabolism [24]–[30] . Mitochondria are unique and complex organelles that perform essential functions in many aspects of cell biology . The dominant function of mitochondria is the production of more than 90% of the cell's energy in the form of ATP through oxidative phosphorylation ( OXPHOS ) . The oxidative phosphorylation system consists of four multimeric complexes , coenzyme Q and cytochrome c forming the mitochondrial respiratory chain ( I–IV ) , which transfer electrons from reducing equivalents to water , creating a proton gradient across the inner mitochondrial membrane , which is used by a fifth complex , the F1F0 ATPase , to drive the synthesis of ATP [31] . In addition to their role in energy metabolism , mitochondria also perform various other functions , which make them absolutely indispensable to the cell . Among these , mitochondria are implicated in apoptosis , the regulation of various metabolism pathways , and signal transduction of antiviral responses [32]–[34] . This makes them a target of almost all invading pathogens , including viruses [34] . Although mitochondria possess a separate and independent genome , most of the mitochondrial proteins are encoded in the nucleus , translated in the cytosol and imported into mitochondria [35] . Therefore , mitochondrial genes are exposed to post-transcriptional regulation by miRNAs . The hypoxia-inducible factor 1 alpha ( HIF1α ) is one of the master regulators of cell metabolism and was shown to inhibit mitochondrial biogenesis [36] . Under normoxic conditions , HIF1α is hydroxylated by the three HIF prolyl hydroxylases ( HPHs ) EGLN1-3 ( PHD1-3 ) and this marks it for proteosomal degradation [37] , [38] . Under hypoxic condition , the HPHs are suppressed , and consequently HIF1α is stabilized [38] . Stabilized HIF1heterodimerizes with HIF1β and binds to the hypoxia response elements ( HREs ) in numerous target genes to activate their transcription [39] . HIF1 is known to mediate an active switch from oxidative phosphorylation to glycolytic metabolism [40]–[45] . Although the KSHV genome encodes 12 miRNAs , up to now , only a handful of their targets have been confirmed [46] . Here we show that the KSHV miRNAs function as a cluster to induce metabolic transformation during latent infection from oxidative phosphorylation to aerobic glycolysis . We identified and confirmed two specific targets for these miRNAs that suggest a possible mechanism for this metabolic transformation: the HIF prolyl hydroxylase EGLN2 and the heat shock protein HSPA9 . We also show that this metabolic conversion contributes to the growth of latently infected cells under hypoxia and promotes latency maintenance .
KSHV was shown to induce the Warburg effect in latently infected endothelial cells , including lymphatic endothelial cells ( LEC ) , and in KS [24] , [25] . However , the mechanisms behind these alterations are not clear , and it is not known whether KSHV has a direct effect on mitochondrial function . During latency KSHV expresses only a subset of genes including vcyclin , vFLIP , LANA and the miRNA cluster . This suggests that one or more of these genes are responsible for changing cellular glucose metabolism . In order to test this hypothesis , we examined the effect of each of the latent proteins and the miRNA cluster on glucose metabolism . Metabolic output is known to vary between different cell types and miRNA function is affected by cell context . Therefore , we focused on primary dermal microvascular LEC , which are believed to be the progenitor cells for KS [47]–[49] . We used lentiviruses to express each of the latent proteins or the miRNA cluster in LEC ( Figure S1A–E ) . When we expressed the miRNA cluster , all the individual miRNAs are expressed and intriguingly the expression level of each one of them is similar to its expression in KSHV infected LEC ( KLEC ) ( Figure S1E–F ) . Since over expression of exogenous miRNAs in cells can have non-specific effects , we created a mutated version of the miRK12-3 , which is one of the most expressed among the KSHV miRNAs ( Figure S1E and F ) , and used it as a control for expression of the KSHV miRNA cluster . There are two obvious characteristics that indicate a shift in glucose metabolism from oxidative phosphorylation ( OXPHOS ) to aerobic glycolysis: reduced oxygen consumption and acidification of growth media due to secretion of the glycolysis product lactate . We first tested oxygen consumption rate using the Seahorse XF24 analyzer . The Seahorse Extracellular Flux Analyzer determines oxygen consumption rate ( OCR ) , and extracellular acidification rate ( ECAR ) , in order to assess cellular functions such as oxidative phosphorylation and glycolysis . While expression of LANA , vcyclin or vFLIP did not have a significant effect on oxygen consumption ( Figure S1G ) , expression of the KSHV miRNA cluster ( miR-LEC ) reduced base line oxygen consumption to a similar level to this in KLEC ( Figure 1A ) . In concordance with this finding we also observed a ∼33% increase of secreted lactate from miR-LEC cells when lactate levels were measured directly from the growth media ( Figure 1B ) . Another known marker of aerobic glycolysis is increased glucose uptake . We monitored glucose uptake into cells by measuring the uptake of the fluorescently labeled deoxyglucose analog 6-NBDG and found increased uptake of ∼30% in cells expressing miR-LEC ( Figure 1C ) , but not in those expressing latent proteins ( Figure S1J ) . We also found that the glucose transporter GLUT1 is overexpressed by over 2-fold in these cells ( Figure 1D ) ; a second indicator of increased glucose uptake . KSHV express additional two miRNAs out of the cluster; miR-K10-10 and miR-K12-12 . Although these miRNAs expressed in a relative high levels in KLEC ( Figure S1F ) we could not find any significant change in oxygen consumption , lactate secretion or glucose uptake when expressing these miRNAs in miR-LEC ( Figure S1 H–I and K ) . To further rule out unspecific effects caused by expression of a miRNA cluster we used miR-132/212 cluster as a second control . As shown in figure 1A and figures S1L–M , expression of this cluster does not affect oxygen consumption or glucose uptake . Taken together , these results suggest that the KSHV miRNAs have the ability to shift host cell metabolism toward aerobic glycolysis . The hypoxia-induced factor alpha ( HIF1α ) is a known regulator of glucose metabolism [50] , [51] and can mediate the Warburg effect in cancer cells [52] . KSHV was also shown to activate HIF1 and HIF2 alpha during latency [53] . We therefore tested whether the miRNA-induced alteration in cell metabolism is related to HIF1α expression and activity . As shown in figure 1E , in miR-LEC HIF1α protein is overexpressed by ∼3 fold compared to the control cells . Correspondingly , we found increased expression of two known targets genes of HIF1α: VEGFA ( vascular endothelial growth factor A ) and ADM ( adrenomedullin ) ( Figure 1F ) . This concurred with increased HIF-1 transcriptional activity as shown by increased luciferase activity when the miRNAs were expressed together with a HIF-1 luciferase-reporter assay [54] ( Figure S1N ) . Taken together these results suggest that expression of the miRNA cluster in LEC is sufficient to change glucose metabolism from oxidative phosphorylation to aerobic glycolysis and this might occur due to HIF1α stabilization . Mitochondria are key players in normal glucose metabolism in aerobic conditions , and as part of the Warburg effect , many cancer types show altered mitochondrial activity [55] . Having found that expression of KSHV miRNAs reduces oxidative phosphorylation , we next tested their effect on mitochondrial function . As an initial assessment of mitochondrial function , we loaded cells with MitoTracker together with Calcein AM . MitoTracker is a fluorescent dye that labels mitochondria within live cells utilizing the mitochondrial membrane potential . It therefore allowed us to calculate mitochondrial volume ( MitoTracker staining ) relative to total cell volume ( Calcein staining ) . Although we could not detect any change in mitochondrial structure , we did find a significant decrease in mitochondria volume in miR-LEC ( Figure 2A ) . Calculation of mitochondria volume in KLEC showed again similar reduction to this in miR-LEC while the miR-132/212 did not have any effect ( Figure S2A ) . When we tested the expression of miR-K12-10 and miR-K12-12 in miR-LEC we found again no significant different ( Figure S2B ) . Since mitochondria have their own independent genome , one can calculate their number by qPCR analysis using specific primers for the mitochondrial genome [56] . Total DNA ( mitochondrial and genomic ) , extracted from miR-LEC and control cells , was used as template for this analysis . This showed that while the control cells had between 50–60 copies of the mitochondrial genome , in miR-LEC there was a 25% decrease in mitochondrial DNA ( Figure 2B ) . Measuring the expression of different OXPHOS complexes using an antibody cocktail , showed a similar decrease in their expression in miR-LEC ( Figure 2C and Figure S2C–D ) . Next , we examined whether this reduction in mitochondrial number is due to reduced mitochondrial biogenesis . For this we measured the mRNA levels of COX-IV ( Cytochrome c oxidase subunit IV ) and TFAM ( mitochondrial transcription factor A ) as expression of these genes has been shown to correlate with mitochondrial biogenesis [57] , [58] . Using qRT-PCR we found that expression levels of both genes are significantly decreased by ∼20% in miR-LEC ( Figure 2D ) . Overall these findings suggest that expression of the miRNA cluster leads to decreased mitochondrial number and activity . miRNAs act to alter the translation and mRNA stability of specific genes and consequently they regulate the cellular pathways those genes control . Therefore , we investigated potential genes that are regulated by the KSHV miRNAs . For this we combined data from three prediction algorithms ( PITA [59] , miRanda [60] and TargetScan [61] ) , together with published data for the KSHV miRNAs targetome from two different CLIP experiments [16] , [17] . This allowed us to identify a number of potential genes that might be regulated by these miRNAs resulting in altered cellular glucose metabolism . All three different HIF prolyl hydroxylases ( EGLN1-3 ) were predicted to be targets of the viral miRNAs . EGLN1 was suggested to be targeted by miR-K12-2 and miR-K12-11 in both BC-1 and BC-3 PEL cell lines through a CLIP experiment [16] , while EGLN2 is predicted by the different algorithms to be targets by multiple miRNAs ( miR-K12-2 , miR-K12-3 , miR-K12-6 , miR-K12-7 , miR-K12-8 , miR-K12-9 and miR-K12-11 ) . It has been shown that in different cell types these three HIF prolyl hydroxylases ( HPH ) have different expression profiles and differential functions in the regulation of HIF1α [62] . We therefore tested the expression profile of these genes in LEC and found that EGLN2 is the most expressed , whilst EGLN3 is hardly expressed in LEC ( Figure S3A ) . We next quantified the mRNA levels of these genes after expression of the viral miRNA cluster and found that only EGLN2 and EGLN3 were down regulated ( Figure S3B ) . Taking these findings into consideration , we decided to focus only on regulation of EGLN2 by the viral miRNAs for further study . Interestingly , both previous CLIP studies identified in their pathway enrichment analysis genes that are involved in protein targeting and localization[16] , [17] . In addition , the different prediction algorithms also predicted many of these genes to be regulated by the KSHV miRNAs . Since mitochondrial biogenesis depends on protein translocation , down regulation of proteins involved in this process might explain the reduced mitochondrial number and biogenesis induced by KSHV miRNAs . To investigate this potential regulatory axis , we chose 9 different genes , predicted to be regulated by the KSHV miRNAs , and tested their mRNA expression in miR-LEC ( Figure S3C ) . We found that the mRNA of 5 of these genes was indeed down regulated in the presence of the KSHV miRNA cluster . In order to further examine this finding , we cloned the 3′UTR of these genes into a reporter plasmid , downstream of the luciferase coding sequence . As shown in Figure S3D , these 3′UTRs are indeed targeted by the KSHV miRNA cluster , resulting in decreased luciferase activity . As the mitochondrial heat shock protein HSPA9 was suggested to be a target of multiple viral miRNAs in both CLIP studies ( miR-K12-3 and miR-K12-4 ) and by all three algorithms ( miR-K12-3 , 4 , 5 , 6 , 7 , 10 and 11 ) , and it was shown to be the most down-regulated among the genes we tested , we focused on it for further studies . HSPA9 ( mtHSP70 , mortalin ) is a central subunit of the matrix-exposed import motor playing a central role in the mitochondrial translocation system and is essential for efficient import and export of proteins [63] , [64] . It was shown that it is crucial for viability and mitochondrial biogenesis [65] , [66] . Therefore , its down regulation could contribute to the reduction of mitochondria in cells expressing the viral miRNAs . To confirm that the KSHV miRNA cluster targets EGLN2 and HSPA9 , we measured their expression in miR-LEC . EGLN2 and HSPA9 mRNA and protein levels were significantly reduced in miR-LEC comparing to the controls cells ( Figure 3A and B and Figure S3E ) . To confirm that these miRNAs specifically target the 3′UTR of these genes , we used a vector containing the luciferase coding sequence up-stream of their 3′UTRs . We expressed this construct in the presence of the miRNA cluster and observed ∼50% reduction in luciferase activity , while the control vector maintained luciferase activity in the presence of the miRNA cluster ( Figure S3F ) . To confirm these genes are regulated also during KSHV infection , we tested their expression in KLEC and found both to be down regulated ( Figure S3G ) . Next , we tested our hypothesis that the viral miRNAs function as a cluster , and dissected the contribution of each individual miRNA to the regulation of EGLN2 and HSPA9 . For this we tested the mRNA levels of EGLN2 and HSPA9 in LEC upon expression of the separate miRNAs ( Figure S3H and figure 3C ) . This showed that , as predicted , each one of these genes is targeted by more than one of the KSHV miRNAs ( miR-K12-1 , 3 , 8 , and 11 for EGLN2 and miR-K12-1 , 3 , 4 , 6 , 8 , 9 and 11 for HSPA9 ) . Expressing luciferase vectors harboring the 3′UTRs of EGLN2 or HSPA9 , together with the individual miRNAs , also showed that these 3′UTRs have target sites for more than one of the KSHV miRNAs ( Figure 3D ) . Some miRNAs down regulated the mRNA levels of these genes but did not have any effect on luciferase activity ( e . g . miR-K12-4 on HSPA9 ) . These miRNAs may target predicted sites outside the 3′UTR or regulate ELGN2 and HSPA9 indirectly . Going back to the prediction algorithm PITA for these genes we found that indeed many of these miRNAs are also predicted to target these genes either in their 3′UTR or coding sequence ( Table S1 ) . To test whether the KSHV miRNAs trigger a shift in cell metabolism through down regulation of EGLN2 and HSPA9 , we used the human GIPZ shRNAmir lentiviral clones ( Open Biosystems ) to specifically suppress each of these genes . We down-regulated each one of these genes ( Figure 4A and Figure S4A–C ) and tested the effect on cellular metabolism in LEC . We found that knockdown of each of these genes led to stabilization of HIF1α , decreased oxygen consumption and reduced mitochondrial volume ( Figure 4A–C ) . Because knockdown of EGLN2 and HSPA9 caused HIF1α stabilization , we next investigated whether the observed metabolic phenotype depends on HIF1α . We therefore expressed the HIF1α P402A/P564A stable mutant in LEC [67] ( Figure S4D ) . As shown in Figure 4D–E , expression of this mutant resulted in a similar decrease in oxygen consumption and mitochondrial volume as that caused by EGLN2 and HSPA9 silencing , suggesting that HIF signaling is playing a role in this phenotype . To further examine whether this metabolic transformation depends on EGLN2 and HSPA9 , we next tested whether overexpression of these proteins could counteract the effect of the miRNA cluster . We achieved overexpression of both EGLN2 and HSPA9 , alongside viral miRNA expression , by using constructs that lacked 3′UTRs ( Figure S5A ) . Expression of each miRNA-resistant gene reversed the effect of the viral miRNAs on its particular mRNA but not on the other gene's mRNA ( Figure S5B ) , showing that the miRNA cluster down-regulates these genes independently . As expected , overexpression of EGLN2 prevented HIF1α stabilization by the miRNA cluster ( Figure 5A ) . More importantly , we found that both EGLN2 and HSPA9 overexpression significantly reduced the effect of the viral miRNA cluster on oxygen consumption and mitochondrial volume ( Figure 5B and C ) . Interestingly , EGLN2 up-regulation had a bigger effect on both oxygen consumption ( Figure 5B ) , and mitochondrial volume ( Figure 5C ) , when compared to HSPA9 overexpression . When both proteins were overexpressed together , they showed an additive effect . The fact that the miRNA cluster phenotype could not be fully reversed by overexpression of either EGLN2 alone , HSPA9 alone , or the combination of both proteins , suggests that although they are part of the miRNA-induced metabolic transformation mechanism , other genes also play a role . Increased glycolytic flux , together with HIF1α stabilization , suggests that latent cells might be more adaptable to low oxygen conditions . We therefore tested whether miR-LEC proliferate faster than the control cells when subjected to such conditions . Cells were grown in high oxygen ( 20% ) and then moved to 1 . 5% oxygen and analyzed after 24 , 48 and 72 hours . In contrast to high oxygen conditions , where both control and miR-LEC growth rates were similar ( Figure S6A ) , when cells were exposed to hypoxia , miR-LEC had a significant growth advantage in the first 24 hours ( Figure 6A ) . This difference was less significant after 48 and 72 hours , suggesting that by these time points the control cells had adjusted to the low oxygen conditions . When we tested the effect of knocking down EGLN2 and HSPA9 , we found that only EGLN2 silencing improved the initial response of LEC to hypoxia ( Figure 6B and Figure S6B ) . There is evidence that in vitro three-dimensional ( 3D ) cell cultures more accurately reflect the complex in vivo microenvironment than simple two-dimensional cell monolayers . We therefore explored the effect of the KSHV miRNAs on cell growth under these conditions . To do this we used the recently published microplate-based method to create and measure spheroid growth [68] . To investigate the effect of the KSHV miRNAs in this context , we chose the human osteosarcoma U2OS cell line , which has previously been shown to form spheroids using this method . U2OS were infected with lentiviruses to express the miRNA cluster or the specific hairpins for EGLN2 and HSPA9 knockdown . As in LEC , expression of the KSHV miRNA cluster in U2OS led to up-regulation of HIF1α and down regulation of EGLN2 and HSPA9 ( Figure S6C ) . These cells were then plated in ultra-low attachment 96-well round-bottomed plates as previously described [68] . Spheroid growth was measured using the CellTiter-Glo Luminescent Cell Viability Assay ( Promega ) and spheroid area was measured using the EVOS cell imaging system . We found that , similarly to the growth advantage under hypoxic conditions , expression of the KSHV miRNA cluster and knock down of EGLN2 led to enhanced growth in 3D cultures , as shown by increased spheroid area ( Figure S6D–E ) and cell number ( Figure 6F ) . The fact that KSHV has evolved to alter cellular metabolism and to reduce mitochondrial biogenesis during latency suggests that these phenomena have a function during the life cycle of the virus . We therefore examined the importance of reduced mitochondrial biogenesis for latency maintenance . To test this , we treated the latently infected PEL cell line BCBL1 with Resveratrol , which has been shown to induce mitochondrial biogenesis in some cell types [69] , [70] . We found that in PEL cells , Resveratrol treatment leads to increased expression of COX-IV and the OXPHOS complexes ( Figure 6C–D ) , in a dose dependent manner . Strikingly , this increase in mitochondrial biogenesis caused extensive cell death ( Figure 6E ) . When we tested for viral gene expression we found that Resveratrol treatment caused substantial expression of lytic genes ( Figure 6F ) , suggesting that the Resveratrol-induced cell death is due to induction of the lytic cycle in these cells . In order to directly test for viral production we used the media from Resveratrol induced cells to directly infect 293T cells and measured the KSHV copy number in these cells . As shown in figure 6G , infection using media from cells treated with Resveratrol leads to increase of up to ∼15 times of KSHV copy number in the infected cells . In order to determine if the same holds true in primary infected cells , we used the recombinant virus , rKSHV . 219 , which expresses a red fluorescent protein under the KSHV lytic PAN promoter [71] . This virus can establish a latent infection in LEC , making it a valuable model for the study of latency and reactivation . We selected infected LEC to create KLEC . 219 , as previously described [71] , and treated them with Resveratrol to induce mitochondrial biogenesis . We first confirmed that Resveratrol does not have a toxic effect on LEC and that it induced mitochondrial biogenesis in them ( Figure S7A and B ) . Resveratrol treatment led to extensive KSHV reactivation , as indicated by a 42% increase in RFP positive cells ( Figure 6H and Figure S7C ) . Interestingly , when we overexpressed EGLN2 and HSPA9 in BCBCL1 and KLEC . 219 , we found that while HSPA9 did not show a significant effect , EGLN2 overexpression caused a small increase in lytic phase activation ( Figure S7D and F ) . Respectively , knocking down HIF1α gave similar results; activation of the lytic phase indicated by increase in RFP positive cells and KSHV copy number in cell infected with media from these cells ( Figure S7 E–F ) . These results suggest that reduced mitochondrial activity is important for maintaining latency in KSHV infected cells .
KSHV has been shown to highjack and manipulate various cellular pathways to promote its own survival and spread [72] . Among those it was shown that KSHV alter its host cell energy metabolism , but neither the mechanism by which this is achieved nor the biological benefit , are understood . Our study has revealed a functional role for the KSHV miRNAs in the regulation of cell metabolism . We report that expression of the KSHV miRNA cluster , in primary LEC , induces the Warburg effect; it reduces oxygen consumption , increases lactate secretion , increases glucose uptake and reduces mitochondrial biogenesis . In addition , expression of these miRNAs also leads to stabilization and activation of the transcription factor HIF1α , a master regulator of cell metabolism . We identified and confirmed two new targets for the KSHV miRNAs , EGLN2 and HSPA9 , and show that KSHV alters host cell energy metabolism through down-regulation of these genes . Nevertheless , the fact that overexpression of these genes ( without their 3′UTRs ) did not fully rescue the phenotypic effect of these miRNAs , suggests that there are more genes involved in this metabolic transformation . This is not surprising because the miRNAs within the cluster are predicted to target many other genes , some of which are involved in cell metabolism , and these might also contribute to this phenomenon . We suggest a two-armed mechanism by which KSHV changes cellular energy metabolism; the first is based on activation of the transcription factor HIF and the subsequent up-regulation of its metabolically relevant target genes . The second arm directly reduces mitochondrial biogenesis by impairing the import of proteins into mitochondria ( Figure 7 ) . We show that the KSHV miRNAs induced-metabolic shift contributes to infected cells proliferation in low oxygen and is important for latency maintenance . KS and PEL cells are not exposed to extensive hypoxia and KSHV associated solid lymphomas are rare . Nevertheless , in some cases KS can have a solid mass [73] , or PEL develops in hypoxic niches , as was shown in an artificial cavity related to the capsule of a breast implant [74] . In addition , adjustments to energy metabolism have been suggested to give cancer cells many advantages with respect to proliferation and growth . The prevalent theory is that , although aerobic glycolysis is an inefficient way to generate energy , it is a necessary adaption to facilitate the uptake and incorporation of nutrients into the biomass , which is needed to produce new cells . This modification also offers similar advantages for KSHV-infected cells . We have shown here for the first time that the KSHV miRNAs regulate cell metabolism . Nevertheless , regulating energy and cancer metabolism using miRNAs is not exclusive to viruses , and cellular miRNAs are also known to control energy metabolism . As we show for the KSHV miRNAs , cellular miRNAs also participate in manipulating cancer cell metabolism by regulating the expression of genes with protein products that either directly regulate the metabolic machinery or indirectly modulate the expression of metabolic enzymes , serving as master regulators [75]–[78] . Various types of cancer cells have been shown to overexpress HIF due to intratumoral hypoxia or as a result of genetic alterations that have occurred as part of their oncogenic program ( reviewed in [79]–[82] ) . We show a new miRNA-based mechanism by which KSHV regulates HIF1α levels in cells . The algorithm miRror [83] suggests that there are 35 cellular miRNAs predicted to target EGLN2 ( table S2 ) and 14 ( table S3 ) predicted to target all three HIF prolyl hydroxylases . This suggests that a similar miRNA-based regulation mechanism of HIF1α stability might exist in normal or cancerous cells . All previously performed AGO2 CLIP experiments in PEL , several different prediction algorithms and our results suggest that the KSHV miRNAs target HSPA9 , together with other proteins from the mitochondrial import machinery . We confirm that several of the KSHV miRNAs collaborate to target HSPA9 and consequently reduce mitochondrial activity . Very little is known about how cells regulate import into mitochondria and whether this might be a mechanism through which mitochondrial biogenesis and activity is regulated . Saccharomyces cerevisiae , which do not have miRNAs , regulate import into mitochondria using cytosolic kinases [84] . We propose a possible regulation of the mitochondrial import machinery by miRNAs . Our results suggest the KSHV miRNAs regulates key proteins in this import machinery such as HSPA9 , TOMM40 , TOMM22 and others . This could be the reason that over expression of EGLN2 and HSPA9 failed to rescue the miRNAs phenotype , since other proteins from the import machinery are still down regulated . Regulation of this process by miRNAs could allow fine-tuning of mitochondrial activity , which could be rapidly reversed . It has been previously shown that knockout of the mitochondrial HSP70 ( HSPA9 ) , as well as other essential proteins in the import machinery , is lethal to cells . Thus , fine-tuning achieved by miRNAs can allow viruses to reduce mitochondrial oxidative phosphorylation while keeping its host cell alive . Our findings provide further support , in the context of cell metabolism , for the concept that KSHV miRNAs operate as a functional cluster . We show that KSHV miRNAs are expressed together in infected cells and collaborate to regulate specific genes ( e . g . EGLN2 and HSPA9 ) within specific pathways ( e . g . the mitochondria import machinery ) . With regards to individual mRNA targets we observe synergy but also some redundant targeting . However , we find that no individual miRNA can fully reproduce the effect of the whole KSHV miRNA cluster on metabolism , suggesting that these miRNAs function synergistically . Moreover , the fact that they are not all expressed to same level suggests an internal regulation of their expression . We speculate that by expressing this cluster , the virus circumvents one of the host cell's main evolutionary tools for escaping unwanted miRNA targeting: alterations in 3′UTR sequences . Even if one miRNA target site is lost , the remaining miRNAs are still able to bind to and regulate the transcript . Although the specific contribution of each of the KSHV miRNAs to this phenomenon has not yet been dissected , through suppressing expression of multiple genes within a specific cellular pathway , the overall effect of the miRNA cluster is coherent , robust and quantitatively sufficient to lead to functionally relevant outcomes . Interestingly , this functional clustering of viral miRNAs around cellular metabolism seems to be conserved in other herpesviruses . EBV , also a gamma-herpesvirus , which expresses multiple viral miRNAs , regulates many of the same genes and cellular pathways as KSHV . Notably , an AGO2 CLIP experiment [16] pulled HSPA9 , together with other proteins involved in import into mitochondria , as targets of the EBV miRNAs , suggesting potential functional convergence . KSHV-infected cells share many characteristics with cancer cells , in particular altered energy metabolism . Therefore , in addition to the implications for KS biology , KSHV infection could be used as a model to study the role of the Warburg effect ( and other altered metabolic pathways ) during transformation . Although having a growth advantage in hypoxic conditions may be significant in certain environments , we speculate that the main evolutionary benefit gained by KSHV through alteration of cellular metabolism is the generation of optimal intracellular conditions for latency establishment and maintenance . Our results show that inducing mitochondria biogenesis interferes with the virus ability to maintain latency - a key step in the virus oncogenic program . Consequently these findings have translational-medicine implications with regards to KSHV-associated malignancies , but also in the broader context of pathologies etiologically linked to DNA viruses .
LEC were purchased from Promocell and grown in endothelial growth medium MV2 ( Promocell ) . LEC were used for experiments before passage 8 . BCLB-1 cells latently infected with recombinant GFP-KSHV [47] were cultured in RPMI 1640 ( Invitrogen ) supplemented with 10% FCS and 400 ng/mL Geneticin ( Invitrogen ) . 293T , U2OS and Vero cells were grown in DMEM ( Invitrogen ) , supplemented with 10% FBS . The KSHV miRNA cluster and individual miRNAs were cloned into the lentiviral vector pSIN-MCS as previously described [18] . The cluster and individual miRNAs were subcloned into the gateway entry vector pENTR/pTER+ [85] . These miRNAs were further cloned into the 3rd gen lentiviral promoter-less Gateway destination vectors pLenti X1 Puro DEST and pLenti CMV GFP DEST using the Gateway LR Clonase II enzyme mix ( Invivogen ) . HSPA9 was amplified using specific primers ( Table S4 ) and cloned into pENTR4 . The gene was further cloned into pLenti PGK Puro Dest [85] using the Gateway LR Clonase II enzyme mix ( Invitrogen ) . FLAG-EglN2-pLenti6 was as previously described [85] . Vesicular stomatitis virus-G envelope-pseudotyped lentiviral virions were produced by cotransfecting 6 µg lentiviral construct , 3 µg p8 . 91 , and 1 µg pMD . G into a 10-cm dish of ∼70% confluent 293T cells using the FuGENE ( Roche ) protocol . Five hours after transfection , the medium was changed , and 48 h after transfection , the medium containing the lentiviral virions was collected , passed through a 0 . 45 µm filter , and either aliquoted directly or concentrated and stored at −80°C . Lentiviral infections were done by incubating the desired amount of virus preparation with suspension LEC for 5 h , after which the medium was changed . Titration of each lentivirus preparation was done either by quantitative PCR [86] or Flow cytometry [87] . All experiments were performed in infected LEC showing more than 80% positive cells . Cells were seeded in XF 24-well cell culture microplates ( Seahorse Bioscience ) at 4×104 cells/well ( 0 . 32 cm2 ) in 200 µl growth medium and then incubated at 37°C/5% CO2 for 20–24 hours . Assays were initiated by removing the growth medium from each well and replacing it with 600 µl of assay medium pre-warmed to 37°C . The cells were incubated at 37°C for 30 minutes to allow media temperature and pH to reach equilibrium before the first rate measurement . Prior to each rate measurement , the XF24 Analyzer gently mixed the assay media in each well for 3 min to allow the oxygen partial pressure to reach equilibrium . Following mixing , OCR and ECAR were measured simultaneously for 4 min to establish a baseline rate . The assay medium was then gently mixed again for 3 min between each rate measurement to restore normal oxygen tension and pH in the microenvironment surrounding the cells . Uncoupled , maximal and non-mitochondrial respiration was determined after the addition of 5 µM oligomycin , 1 µM carbonyl cyanide 4- ( trifluoromethoxy ) phenylhydrazone ( FCCP ) and 2 µM antimycin-A . All chemicals were from Sigma-Aldrich . Cells were lysed in RIPA buffer ( 300 mM Sodium Chloride , 1% NP-40 , 0 . 5% Sodium deoxycholate , 0 . 1% Sodium dodecyl sulphate and 50 mM Tris pH 8 . 0 ) . Equal amounts of protein were resolved on Mini-PROTEAN TGX Precast gels ( Bio-Rad Laboratories ) . Antibodies against EGLN2 ( Novus Biologicals ) , HSPA9 ( Abgent ) , GLUT1 ( Alpha Diagnostic ) , HSP90 ( Cell signaling ) , α-tubulin and Flag ( Sigma-Aldrich ) were detected with IRDye secondary antibodies ( LI-COR ) . Antibodies against HIF1α ( BD Transduction Laboratories ) and the MitoProfile Total OXPHOS Human WB Antibody Cocktail ( Abcam ) were detected with HRP-conjugated secondary antibodies and were quantified using ECL ( GE Healthcare ) . Images were analysed using the Image Studio Lite ( LI-COR ) or the ImageQuant ( GE ) Western blot analysis softwares . Genomic DNA for qPCR was extracted using the QIAamp DNA mini-kit ( Qiagen ) . Total RNA was extracted using either the RNeasy mini-kit or the miRNeasy mini-kit ( Qiagen ) . Approximately 100 to 1000 ng of total RNA was used for cDNA synthesis using the ProtoScript II Reverse Transcriptase ( New England Biolabs ) . DNA and mRNA levels were quantified by qPCR and qRT-PCR using optimized primers ( Table S4 ) and SYBR Green PCR master mix ( Applied Biosystems ) . qRT-PCR quantification of VEGF and ADM was performed using Taqman gene expression assays ( Applied Biosystems ) . cDNA synthesis for qRT–PCR quantification of mature miRNAs was performed using the Exiqon Universal cDNA Synthesis Kit II according to the manufacturer's instructions . Detection of the mature KSHV miRNAs was performed using the kshv-miR LNA PCR primer sets ( Exiqon ) . Cellular small nucleolar RNA RNU66 was used as a reference RNA . Cells were counted and analyzed for viability using the Muse Count & Viability Reagent on the Muse Cell Analyzer ( Merck Millipore ) . Glucose uptake was measured by incubating cells with 30 µM glucose analogue 6-NBDG ( Invitrogen ) for 15 minutes . Cells where then washed and trypsinized and their fluorescence ( λex: 465 nm , λem: 540 nm ) was measured by flow cytometry . The 3′UTRs of the indicated genes were amplified using specific primers ( Table S4 ) and cloned into the psiCHECK 2 vector ( Promega ) . The reporter plasmids ( 50 ng ) were transfected into 293T cells , 48 hours after transfection with the individual miRNAs or the miRNA cluster . Cells were harvested 24 hours after transfection according to the Dual-Luciferase Reporter assay system ( Promega ) . Luciferase activity was measured using a Fluoroskan Ascent FL luminometer ( ThermoScientific ) . Firefly activity was normalized to internal Renilla luciferase levels . U2OS-HRE-luc cells were infected with the control or the miRNA cluster vector . 72 hours post infection cells were washed twice with ice-cold PBS and lysed using Glo Lysis buffer ( Promega ) . Luciferase activity was measured using the steady Glo Luciferase assay ( Promega ) . Cells were loaded with 5 µM Calcein-AM and 50 nM MitoTracker Red ( Invitrogen; 37°C , 30 minutes ) in growth media for 30 minutes and Z-series of images were acquired using a Zeiss LSM 510 system ( Carl Zeiss , Inc . , Cambridge , UK ) , as previously described [88] . Maximal projection of images was used to quantify the area of green ( Calcein ) and red ( mitoTracker Red ) signal . Mitochondrial area was defined relative to cytoplasmic area as ‘area red/area green’ . Images were analyzed using the MetaMorph Microscopy Automation & Image Analysis Software ( Molecular Devices ) . The two channels ( Calcein-AM and MitoTracker Red ) were separated and threshold in order to acquire two separate binary images . rKSHV . 219 stocks were prepared from the Vero cell as previously described [71] . Stably infected LEC were selected and maintained with 1 ug/ml puromycin ( Invivogen ) . Cells were grown in 96well plates and fixed using cold 10% Trichloroacetic acid for 1 hour at 4°C . Wells were then washed 5 times with distilled water and left to air dry before staining with of 0 . 4% w/v Sulforhodamine B ( Sigma-Aldrich ) for 30 minutes at room temperature . Wells were washed 5 times with 1% v/v acetic acid and left to air dry before adding 100 µl of 10 mM Tris base pH 10 . 5 . Plates were incubated on a plate shaker for 5 minutes and read at 564 nm using the Varioskan Flash plate reader ( Thermo Scientific ) . For spheroid generation , 200 µl of cell suspension ( 1×104 cells/ml ) were dispensed into ULA 96-well round-bottomed plates ( Corning ) using a multichannel pipette . Plates were incubated for 4 days at 37°C , 5%CO2 . Images were acquired by EVOS cell imaging system and analysed using the Adobe Photoshop CS6 extended version .
|
Kaposi's sarcoma ( KS ) is the most common cancer in HIV-infected untreated individuals . Kaposi's sarcoma-associated herpesvirus ( KSHV ) is the infectious cause of this neoplasm . The discovery of KSHV and its oncogenic enigmas has enlightened many fields of tumor biology and viral oncogenesis . The metabolic properties of KS significantly differ from those of normal cells and resemble cancer cells in general , but the mechanisms employed by KSHV to alter host cell metabolism are poorly understood . Our work demonstrates that KSHV microRNAs can alter cell metabolism through coherent control of independent pathways , a key feature of microRNA-mediated control of cellular functions . This provides a fresh perspective for how microRNA-encoding pathogens shape a cell's metabolism to create an optimal environment for their survival and/or replication . Indeed , we show that , in the case of KSHV , viral microRNA-driven regulation of metabolism is important for viral latency . These findings will evoke new and exciting approaches to prevent KSHV from establishing latency and later on KS .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"rna",
"micrornas",
"virology",
"biology",
"and",
"life",
"sciences",
"microbiology",
"metabolism",
"energy",
"metabolism"
] |
2014
|
Kaposi's Sarcoma Herpesvirus MicroRNAs Induce Metabolic Transformation of Infected Cells
|
Bone involvement in human cystic echinococcosis ( CE ) is rare , but affects the spine in approximately 50% of cases . Despite significant advances in diagnostic imaging techniques as well as surgical and medical treatment of spinal CE , our basic understanding of the parasite's predilection for the spine remains incomplete . To fill this gap , we systematically reviewed the published literature of the last five decades to summarize and analyze the currently existing data on epidemiological and anatomical aspects of spinal CE .
Hydatid disease or cystic echinococcosis ( CE ) , caused by the larval stage of the cestode Echinococcus granulosus , is a cosmopolitan parasitic zoonosis occurring on every continent except Antarctica . Hydatid ( Greek for ‘watery cyst’ ) disease was already recognized by Hippocrates over 2000 years ago and in 1807 Churrier made the first description of spinal hydatidosis , roughly 100 years after Bidloo ( 1708 ) discovered the existence of a bony form of the disease [1] . The parasite's lifecycle involves two hosts . The definitive host is usually the dog ( but may be another carnivore ) , where the adult parasite lives - attached by hooklets and suckers to the mucosa - in the proximal small bowel . The eggs of the parasite are shed with the host's feces into the environment where the intermediate host , usually a sheep ( but may be some other herbivore ) , gets infected when grazing on contaminated ground . After ingestion of the egg , the embryo ( oncosphere ) hatches , penetrates the intestinal mucosa , enters into the host's circulatory system ( via venous and lymphatic pathways ) , and ( if not destroyed by the host's immune response ) develops into the characteristic vesicular metacestode when reaching a suitable anatomical site . This stage of the parasite is typically a unilocular , fluid-filled cystic lesion ( ‘hydatid’ , ‘hydatid cyst’ ) , which grows expansively by concentric enlargement ( increasing in diameter from 1–5 cm per year ) within the affected organ and harbors the infective protoscolices . When the definitive host feeds on infected viscera , the cycle is complete [2] . In the accidental human intermediate host , the characteristic cystic lesions are mainly found in the liver ( ∼70% ) and the lungs ( ∼20% ) , but virtually any part of the body may be affected , including the bone ( ∼0 . 5–4% ) . The central nervous system ( which is involved in ∼3% of all cases ) and the vertebral column ( which is involved in ≥50% of the ∼0 . 5–4% of cases affecting the bone ) [3]–[6] are particularly vulnerable given the sequelae that result from their involvement . ‘Spinal CE’ ( involvement of the spinal cord , the spine , or both structures ) is associated with a high degree of morbidity , disability , and mortality and the prognosis has often been compared to that of malignancies ( ‘le cancer blanc’ [7] ) . We systematically reviewed all published case reports and case series of spinal CE from 1965 until 2012 to summarize and analyze the epidemiological and anatomical aspects of the disease and discuss the findings in light of the existing data .
We performed a PubMed ( MEDLINE ) search of the literature using the key words ‘spinal echinococcosis’ , ‘spinal hydatidosis’ , ‘spinal hydatid disease’ , ‘spinal echinococcal cyst’ , ‘spinal cystic echinococcosis’ and reviewed the obtained references published from 1965 until July 1st 2012 ( figure 1; references S1 ) . The year 1965 was chosen , as it proved difficult to obtain articles before this year . All publications on clinical cases and case series of human spinal echinococcosis published in English , French , German , Italian , and Spanish were collected . When the original article was not obtainable but the abstract contained data on anatomy , treatment approach or therapeutic outcome , the publication was included in the analysis . In addition , the reference lists of the collected publications were screened for supplementary ( not PubMed listed ) case reports on spinal CE eligible for analysis . The collected data included patient's age , sex , if applicable manifestations , interventions and time frame of previous spinal or extraspinal CE , cyst number , cyst location ( s ) , and involved anatomical structures . The extracted data was entered into Microsoft Excel-files ( Version 2002 ) and later transformed into SPSS-files ( Version 16 . 0 . 0 , 2007 ) for analysis . Data on the age of the patients , follow-up periods and recurrence periods was summarized as medians and ranges and , if applicable , analysed by using the Mann-Whitney U test . Nominal data was summarized as frequencies and percentages and analysed by χ2-test . A p-value <0 . 05 was considered statistically significant .
Of the 367 publications identified by electronic search , 189 publications ( on 467 cases of spinal CE ) were included in the analysis ( figure 1 ) . Individual data on the patient's age was available for 325 cases , on the gender for 408 cases ( 232 male , 176 females ) and on age and gender for 316 ( 186 male , 130 female ) cases . The discrepancy between data on age and on gender is due to case series , where data on gender was available but data on age was limited to the mean or median of the case series . The overall median age was 35 years ( range 3–77 years ) without significant difference between male ( median 35 years; range 3–76 years ) and female cases ( median 36 years; range 4–77 years ) ( figure 2 ) . Data on the number of cysts was available for 243 of the 467 spinal CE cases: 56 ( 23% ) presented with a single cyst , 187 ( 77% ) presented with multiple cysts . Data on the spinal level of the cyst ( s ) ( cervical , cervico-thoracal , thoracal , thoraco-lumbar , lumbar , lumbo-sacral , sacral ) was available for 465 of the 467 cases . In 303 of these cases , specific data on the involved vertebral level ( s ) was available and in 287 of these cases , it was possible to determine the exact number of involved vertebral levels ( the discrepancy of these figures is due to the fact that not in all cases with sacral involvement the exact number of involved sacral vertebral levels was reported ) . The frequency and distribution of the spinal level ( s ) and individual vertebral level ( s ) involved is shown in figure 3 . A subgroup analysis was performed concerning the possible difference in cyst location in cases with a history of extraspinal CE surgery ( table 1 ) . Spinal CE cases having a history of previous extraspinal CE surgery were principally operated on for pulmonary CE ( table 2 ) and showed a statistically significant association with upper ( thoracic ) spine involvement ( figure 4 ) . To evaluate the allocation of spinal CE to the different anatomical structures , we classified the cases according to the Dew/Braithwaite & Lees classification ( figure 5 ) and additionally collected data on the involvement of posterior vertebral elements ( pedicles , transverse processes , vertebral arch ) and intervertebral disc involvement . Complete data on the involved anatomical structures was available for 230 cases ( table 3 , 4 ) . Figure 6 shows the involvement of the different anatomical structures at the vertebral level . A frequently reported manifestation of spinal CE is a ‘dumbbell’-formation: a continuous lesion with an intraspinal-extradural and an intrathoracic-paravertebral part , communicating through one or more intervertebral foramina ( i . e . a combination of a BL type 3 and a BL type 5 lesion [± additional structures] ) ( figure 5 ) . As spinal CE presenting with a ‘dumbbell’-formation has frequently been described in the literature , we explored the collected data on the frequency of this manifestation ( table 4 ) . We observed a statistically significant difference in the age of patients presenting with intradural ( BL type 1 & 2 ) and extradural cysts ( BL type 3 , 4 & 5 ) : the median age of patients with extradural cyst location was 36 years ( range 3–77 years ) , while the median age of patients with intradural cysts location was 18 . 5 years ( range 4–67 years ) [p<0 . 0001] ( figure 7 ) .
CE is prevalent throughout most of the world and regional incidence rates of human infection differ widely , depending on the local interaction of man and the natural definitive and intermediate hosts . The greatest prevalence of CE in human and animal hosts is found in countries of the temperate zones , including several regions of Eurasia ( the Mediterranean regions , southern and central parts of Russia , central Asia , China ) , Australia , some parts of America ( especially South America ) and north and east Africa [8] . Spinal CE is seen in all age groups , with both sexes being affected ( figure 2 ) . The median age of 35 years is consistent with published data from larger case series , where the median age was 30 [9] , 33 [10] , 35 [11] , and 36 [12] years respectively . The overall gender distribution of 56 . 9% male/43 . 1% female is similar to the distribution found in a large review of 38 Turkish publications covering 111 cases ( 65 . 8% male/34 . 2% female ) [13] . However , in our review , case reports and case series originating from very different epidemiological settings were included . Therefore , the analysis may not necessarily apply to specific local situations , where age or gender distributions may differ according to local exposure patterns ( with different social , occupational and environmental factors influencing the local interaction pattern of the accidental human intermediate host with the definitive host and the environment ) . Considering that only 17 . 9% ( 120 cases ) of all reviewed spinal CE cases had a history of extraspinal CE or were found to have concomitant newly diagnosed extraspinal CE ( table 1 ) , it appears that primary spinal CE is more frequent than secondary spinal CE . Even when taking into account that the reviewed case reports often did not mention or provide data on screening investigations for extraspinal CE , and that cases and case series were included , which were published before ultrasound and cross-sectional imaging techniques ( CT , MRI ) became available , this assumption appears to be justified . Whether spinal CE in patients with a history of extraspinal CE results from simultaneous primary infection ( spinal and extraspinal infection acquired simultaneously on primary infection ) , arises from secondary hematogenous seeding of extraspinal CE or constitutes a new exogenous infection is difficult to say . Tapia and colleagues stated that osseous CE is probably acquired in childhood and remains clinically latent even for more than 40 years and typically manifests in adults [24] . Local recurrence of spinal CE has been reported to occur up to 29 years after surgery [25] . Discriminating dormant primary spinal infection from dormant secondary hematogenous seeding to the spine is impossible . The only way to prove exogenous reinfection would demand genotyping of the primary and secondary site of infection ( Note: genetic characterization of the parasite was not reported in any of the reviewed spinal CE cases ) . When reviewing the collected data , secondary spinal CE arising from spontaneous haematogenous seeding of extraspinal CE might exist , but is probably very rare: we found only 6 cases of spinal CE where concomitant asymptomatic extra-spinal CE was reported ( table 1 ) . In 5 of these cases , spinal seeding from visceral CE may have occurred ( in 2 cases the cyst stage supports the assumption that visceral CE anteceded spinal CE ) . The 6th case , an intraventricular cardiac cyst diagnosed after 2 previous surgical interventions for spinal CE , was the only case we found indicating haematogenous seeding following surgery of spinal CE . Haematogenous seeding following surgery of extra-spinal CE has been reported and is generally considered to be the most common route in spinal infection [26] . We found a history of previous surgical intervention ( s ) for extraspinal CE in 16 . 7% of the spinal CE cases ( table 1 ) . This figure is comparable to the extraspinal CE prevalence of 14 . 4% reported in a Turkish series of 111 spinal CE cases [13] . Bearing in mind that hepatic CE is typically more common than pulmonary CE ( see introduction ) , it is interesting that most cases of spinal CE with a history of surgery for extraspinal CE were operated on for pulmonary CE ( table 2 ) . Even though the available data is limited and does not permit deeper analysis , three possible explanations could be discussed: 1 . the risk for spinal seeding following surgery of pulmonary CE might be higher than in surgery of hydatid cysts at other locations; 2 . pulmonary CE might be an indicator for a porto-systemic route of primary infection rather than being causally related; 3 . pulmonary CE might be an indicator for an inhalative route of primary infection: in 1965 Borrie and colleagues demonstrated that inhalation of E . granulosus eggs can lead to pulmonary hydatid disease in sheep [27] . Therefore , even though never proven for the human host , inhaled eggs could theoretically enter the pulmonary circulation , disseminate systemically and reach the spine . The performed subgroup analysis of the vertebral level involvement in cases with and without a history of surgery for extraspinal CE show a statistically significant difference: previous surgery for extraspinal CE appear to be more frequently associated with thoracic vertebral involvement ( figure 4 ) . This observation would indirectly support the speculation that primary hematogenous spinal infection of especially the lower parts of the spine occurs via porto-vertebral shunts ( see above ) . Irrespective of the route of infection the putative ‘dormant’ period of spinal CE appears to be very long ( table 1 ) and emphasizes long-term follow-up . However , the available data does not support standard screening of patients with extraspinal CE for concomitant asymptomatic spinal CE . In most cases the exact primary implantation site of the parasite and the primary affected spinal structure remains unclear and the disease is only diagnosed after several anatomical structures become affected ( table 4 ) . The summarized data ( table 4; figure 6 ) suggests that the parasite's primary implantation site can be either the vertebral bone ( with secondary extra-osseous spread to the paravertebral and intraspinal space ) or the paravertebral or intraspinal soft-tissue ( with secondary infiltration of the vertebral bone ) . While all spinal structures can be infiltrated in the course of disease , no case has been published reporting dura infiltration or penetration . Secondary ‘per continuitatem’ spinal CE appears to be very rare: we found 2 case reports of spinal seeding following surgery of cerebral CE , but no case report of spontaneous spinal seeding from cerebral CE [13] , [28] . Despite significant advances in diagnosis and treatment of CE many aspects , including the parasite's predilection for the spine in osseous CE , remain poorly understood . Spinal CE primarily affects the thoraco-lumbar spine , involving the individual vertebral levels with gradually ascending decline . Contrary to common perception , primary spinal CE appears to be more frequent than secondary spinal CE . It appears that the affected vertebral level in spinal CE differs in patients with and without history of surgery for extraspinal CE . Previous surgery for extraspinal CE appears to be more frequently associated with thoracic vertebral involvement . Patients with intradural CE present at a younger age than patients with extradural CE . Possibly future studies will be able to identify parasite and/or host specific parameters to provide molecular genetic based explanations for the interindiviudal differences in local manifestation and evolution of CE .
|
Spinal cystic echinococcosis ( CE ) is a rare but malignant form of a truly neglected tropical disease . Despite significant advances in diagnostic imaging techniques as well as surgical and medical treatment of spinal CE , our basic understanding of the parasite's predilection for the spine remains poor at best . Information on the influence of parasite and host specific factors on anatomical manifestations and evolution of CE is currently lacking . We systematically reviewed all published case reports and case series of spinal CE from 1965 until 2012 to summarize and analyze the epidemiological and anatomical aspects of the disease and discuss the findings in light of the existing data .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2013
|
Spinal Cystic Echinococcosis – A Systematic Analysis and Review of the Literature: Part 1. Epidemiology and Anatomy
|
We have empirically assessed the distribution of published effect sizes and estimated power by analyzing 26 , 841 statistical records from 3 , 801 cognitive neuroscience and psychology papers published recently . The reported median effect size was D = 0 . 93 ( interquartile range: 0 . 64–1 . 46 ) for nominally statistically significant results and D = 0 . 24 ( 0 . 11–0 . 42 ) for nonsignificant results . Median power to detect small , medium , and large effects was 0 . 12 , 0 . 44 , and 0 . 73 , reflecting no improvement through the past half-century . This is so because sample sizes have remained small . Assuming similar true effect sizes in both disciplines , power was lower in cognitive neuroscience than in psychology . Journal impact factors negatively correlated with power . Assuming a realistic range of prior probabilities for null hypotheses , false report probability is likely to exceed 50% for the whole literature . In light of our findings , the recently reported low replication success in psychology is realistic , and worse performance may be expected for cognitive neuroscience .
Low power and selection biases , questionable research practices , and errors favoring the publication of statistically significant results have been proposed as major contributing factors in the reproducibility crisis that is heavily debated in many scientific fields [1–5] . Here , we aimed to get an impression about the latest publication practices in the closely related cognitive neuroscience and ( mostly experimental ) psychology literature . To this end , we extracted close to 30 , 000 records of degrees of freedom ( df ) and t-values from papers published between Jan 2011 to Aug 2014 in 18 journals . Journal impact factors ranged from 2 . 367 ( Acta Psychologica ) to 17 . 15 ( Nature Neuroscience ) . The data allowed us to assess the distribution of published effect sizes ( D ) , to estimate the power of studies , and to estimate the lower limit of false report probability ( FRP ) . The text-mining approach we used enabled us to conduct a larger power survey than classical studies . Low power is usually only associated with failing to detect existing ( true ) effects , and therefore , with wasting research funding on studies which a priori have a low chance to achieve their objective . However , low power also has two other serious negative consequences: it results in the exaggeration of measured effect sizes and it also boosts FRP , the probability that statistically significant findings are false [5–7] . First , if we use Null Hypothesis Significance Testing ( NHST ) , then published effect sizes are likely to be , on average , substantially exaggerated when most published studies in a given scientific field have low power [6 , 8] ( see S1A Fig for the mechanism of effect size exaggeration ) . This is because even if we assume that there is a fixed true effect size , actual effect sizes measured in studies will have some variability due to sampling error . Underpowered studies will be able to classify as statistically significant only the occasional large deviations from real effect sizes . Conversely , most measured effects will remain under the statistical significance threshold even if they reflect true relationships [9–11] . Effect size inflation is greater when studies are even more underpowered . Consequently , while meta-analyses may provide the illusion of precisely estimating real effects , they may , in fact , estimate exaggerated effects detected by underpowered studies while at the same time not considering unpublished negative findings ( see , e . g . , [12] ) . Secondly , from the Bayesian perspective , the long-run FRP of the NHST framework can be defined as the probability that the null hypothesis ( a hypothesis to be “nullified” ) is true when we get a statistically significant finding . The long-run True Report Probability ( TRP ) can be defined as the probability that the alternative hypothesis is true when we get a statistically significant finding [13 , 5] . Note that the concepts of FRP and TRP do not exist in the NHST framework: NHST only allows for the rejection of the null hypothesis and does not allow for the formal acceptance of the alternative hypothesis . However , here we do not apply NHST but rather , characterize its long-run ( “frequentist” ) performance from the Bayesian point of view . This approach allows us to talk about true and false null and alternative hypotheses ( see more on this in [13 , 5] ) . Computationally , FRP is the number of statistically significant false positive findings divided by the total number of statistically significant findings . TRP is the number of statistically significant true positive findings divided by the total number of statistically significant findings . FRP and TRP can be computed by applying Bayes theorem ( see S1 Text , Section 5 for details ) . The overwhelming majority [14] of NHST studies relies on nil–null hypothesis testing [15] where the null hypothesis assumes an exact value . In such cases , the null hypothesis almost always assumes exactly zero difference between groups and/or conditions . For these applications of NHST , FRP can be computed as FRP=OαOα+Power where O stands for prestudy H0:H1 odds and α denotes the statistical significance level , which is nearly always α = 0 . 05 . So , for given values of O and α , FRP is higher if power is low . As , in practice , O is very difficult to ascertain , high power provides the most straightforward “protection” against excessive FRP in the nil–null hypothesis testing NHST framework [5–7] ( see further discussion of our model in the Materials and Methods section ) . Because published effect sizes are likely to be inflated , it is most informative to determine the power of studies to detect predefined effect sizes . Hence , we first computed power from the observed degrees of freedom using supporting information from manually extracted records to detect effect sizes traditionally considered small ( d = 0 . 2 ) , medium ( d = 0 . 5 ) , and large ( d = 0 . 8 ) [16–18] . Second , we also computed power to detect the effect sizes computed from t-values published in studies . Given that many of these published effect sizes are likely to be inflated compared to the true ones ( as explained above ) , this enabled us to estimate the lower limit of FRP [5 , 13] .
In summary , a computer algorithm searched through each paper for frequently occurring word and symbol combinations for reporting degrees of freedom and effect sizes provided as Cohen’s d . We extracted statistical information about t tests and F tests ( t-values , F-values , degrees of freedom , p-values , and effect sizes ) . Only t-test data is used in this paper , so here we limit data extraction description to t-tests . In psychology and cognitive neuroscience , full t-test records are typically reported in the text as , for example , 't ( df ) = x . xx; p = y . yy' . D-value reports are often added to these reports as , e . g . , 't ( df ) = x . xx; p = y . yy; d = z . zz' . Hence , in a first text parsing phase , the algorithm opened each PDF file from each journal and identified each point of text which contained a “t ( ” character combination or a “t” character . If these characters were identified , then a line of 65 characters were read out from the PDF file starting at the “t ( ” character combination or at the “t” character . Spaces between letters and symbols were removed from these lines of text . That is , it did not matter how many spaces separated relevant statistical entries . Lines of text were kept for further analysis if they contained the characters “=“ , “<” , or “>” and an additional “p =“ , “p<” , or “p>” character combination . This parsing phase identified lines potentially containing independent full t-test records . In building this parsing phase , the performance of the algorithm was initially evaluated by reviewing identified lines of text and extracted data from the first 30 papers analyzed for each journal . If specific journals used special characters ( as identified by the PdfToolbox package ) for typesetting some information ( e . g . , equation signs ) , then this was identified and taken into account in the code . In a second parsing phase , Matlab regular expressions were used to identify full t-test records using the templates noted above ( e . g . , “t ( df ) = x . xx” or “d = z . zz” ) . All text searches were done after converting lines to lowercase characters , so upper- or lowercase differences did not matter in searches . After data extraction , some error checks were done . First , the algorithm detected a few records twice . This may have happened if for any reason an additional “t” appeared within the statistical reporting text ( e . g . , if researchers used the ‘t’ character very close to a statistical record , then that record may have been picked up twice ) . So , records which had identical statistical information to preceding records were removed . Second , records where negative degrees of freedom ( two records ) and/or negative p-values ( one record ) were detected were removed . These may have occurred in response to odd character sets or to errors in the text . After cleaning the data , several informal spot-checks were run: hundreds of lines of extracted text were visually compared with the numerical records extracted from the text . A limitation is that the algorithm only extracted information from the text but not from tables . Further , in order to limit false positive detections ( see also later ) , we restricted our initial search for full p-value records , so some reported nonsignificant results and stand-alone t-values may have been missed ( e . g . , t < 1; t = 0 . 23 ) . It is important to note that we only assured that our extraction algorithm works fine for the journals and publication years analyzed here . It has not been validated as a more “universal” extraction algorithm like statcheck [19] , for example , which we did not know about when starting this project . The extraction algorithm is published as supporting material ( S1 Code ) . In a formal validation procedure , we randomly selected 100 papers with t-value , df , and effect size reports . The selected papers were manually checked for all statistical records . The content of the identified records was then compared to the content of automatically extracted records . This was done to see the accuracy of the computer algorithm and to gather information on the data . Validation results showed that the automatic extraction algorithm had highly satisfactory performance . The randomly selected papers for validation included 1 , 478 records of data . The algorithm correctly identified about 95% of t-values and degrees of freedom in these records . The algorithm missed only 76 records ( 5 . 14% ) , usually due to atypical punctuation or line breaks within a statistical record . There were no false alarms; that is , all data extracted really belonged to t-value records . This is plausible because regular expressions had to fulfill several conditions in order to be identified as potential t-test records . For example , it is unlikely that an expression like “t ( df ) = x . x” would stand for anything else than a t-value record . The good performance of the extraction algorithm is also reflected in the similarity between the distributions of automatically and manually extracted degrees of freedom shown in Fig 1 ( two-sample Kolgomorov-Smirnov test comparing the distributions: test statistic = 0 . 04; p > 0 . 127 ) . This suggests that the degrees of freedom distribution underlying our effect size analysis was extracted accurately . Using the validation data , we found that the overwhelming majority of extracted two sample t-test records reported close-to-equal group numbers ( median ratio of group numbers = 1 ) . The ratio of the participant numbers in the larger group to the participant numbers in the smaller group was smaller than 1 . 15 in 77% of records . We also established that with degrees of freedom of ten or less , about 94% of tests were one sample or matched t-tests , whereas about 72% of records with higher degrees of freedom were one-sample or matched t-tests . t-test data was used for effect size , power , and FRP analysis as it is straightforward to estimate effect sizes from published t-values . After checks for reporting errors , seven records with degrees of freedom > 10 , 000 were excluded from analysis as outliers . This left 27 , 414 potential records . Of these records , 26 , 841 from 3 , 801 papers had both degrees of freedom and t-values reported . We used this data for the effect size analysis . 17 , 207 t-test records ( 64 . 1% ) were statistically significant ( p ≤ 0 . 05 ) and 9 , 634 ( 35 . 9% ) t-test records were statistically nonsignificant ( p > 0 . 05 ) . 2 , 185 t-test records also reported Cohen's d as a measure of effect size ( 1 , 645 records with p ≤ 0 . 05 [75 . 3%] and 540 records with p > 0 . 05 [24 . 7%] ) . As it is not possible to establish the exact participant numbers in groups for our large sample size , making a few reasonable assumptions is inevitable . First , based on our validation data from 1 , 478 records , we made the assumption that participant numbers in two-sample t-test groups were equal . The number of participants in groups was approximated as the upwards rounded value of half the potential total number of participants in the study , i . e , Nsubgroup = roundupper ( ( df+2 ) /2 ) , where df = degree of freedom . This formula even slightly exaggerates participant numbers in groups , so it can be considered generous when computing power . Second , regarding matched t-tests , we assumed that the correlation between repeated measures was 0 . 5 . In such a case , the effect sizes can be approximated in the same way for both one-sample and matched t-tests . These assumptions allowed us to approximate effect sizes associated with all t-tests records in a straightforward way [20–21] . Computational details are provided in S1 Text , Section 2 . Considering the validation outcomes , we assumed that each record with a degree of freedom of ten or less had a 93% chance to be related to a one-sample or matched-sample t-test , and other records had a 72% chance to be related to a one-sample or matched-sample t-test . Hence , we estimated the effect sizes for each data record with an equation assuming a mixture of t-tests where the probability of mixture depended on the degrees of freedom: D=pr ( t1|df ) ∙Dt1+pr ( t2|df ) ∙Dt2 where pr ( t1|df ) and pr ( t2|df ) refer to the respective probabilities of one sample and matched t-tests ( t1 ) and independent sample t tests ( t2 ) and Dt1 and Dt2 refer to the respective effect sizes estimated for these tests . df refers to the degrees of freedom . The power of t-tests was computed from the noncentral t distribution [22] assuming the above mixture of one-sample , matched- , and independent-sample t-tests . Computational details are provided in S1 Text , Section 3 . Power was computed for each effect size record . ( Note that NHST is amalgamation of Fisher’s significance testing method and the Neyman-Pearson theory . However , the concept of power is only interpreted in the Neyman-Pearson framework . For extended discussion , see [23–24] ) . First , we calculated power to detect small , medium , and large effect sizes . Power was computed for each extracted statistical record , taking into account the extracted degrees of freedom , a fixed ( small , medium , or large ) effect size with a significance level of α = 0 . 05 . Second , we also calculated power to detect the published effect sizes . Importantly , these published effect sizes are likely to be highly exaggerated . Using these exaggerated effect sizes for power calculations will then overestimate power . Hence , if we calculate FRP based on power calculated from published effect size reports , we are likely to estimate the lower limits of FRP . So , we estimated the lower limits for FRP , using the probably highly inflated effect sizes ( computed from published t-values ) to calculate power for various H0:H1 odds and bias values and with α = 0 . 05 . ( The computation of FRP is laid out in detail in S1 Text , Section 5 . ) In order to get the expected value of FRP for the whole literature , we weighed the FRP computed for each degree of freedom ( df ) and effect size ( D ) combination by the probability of that particular ( df , D ) combination occurring in the research literature and summed the results for all ( df , D ) combinations: E[FRP]=∑i=1;j=1i=n;j=mFRP ( dfi , Dj ) pr ( dfi , Dj ) An issue worth mentioning is that our model for FRP solely characterizes nil–null hypothesis testing , which is by far the most popular approach to statistics in biomedical science [14] . A very serious drawback of nil–null hypothesis testing is that it completely neglects effect sizes and exclusively directs attention to p-values . In addition , it will inevitably detect very small effects as “statistically significant” once statistical power is high enough . However , these small effects can be so close to zero that one could argue that they are practically meaningless . So , from this perspective , if studies with high power detect small effect sizes as statistically significan't this will only increase FRP . Hence , in such cases , paradoxically , increasing power can be thought to lead to increased FRP . This could be taken into account by modifying our basic model described in the introduction as: FRP=Oα+PS∙pr ( S ) Oα+PS∙pr ( S ) +PL∙pr ( L ) Where PS stands for power to detect small effects , PL stands for power to detect Large effects , pr ( S ) and pr ( L ) stand for the probability of small and large effects , respectively ( pr ( S ) + pr ( L ) = 1 ) . O stands for prestudy H0:H1 odds , and α denotes the statistical significance level as before . A difficulty in computing FRP in this way is that the threshold between small and large effects is arbitrary and strongly depends on subjective decisions about what effect size is and is not important . Hence , explicitly modeling the negative impact of detecting very small effect sizes as statistically significant would be fairly arbitrary here , especially as we have collected effect sizes from many different subfields . Most importantly , factoring in very small but statistically significant effect sizes as false reports into our calculations would only further increase FRP relative to the nill–null hypothesis testing model outlined above . That is , our calculations here really reflect a best-case scenario , the lowest possible levels of FRP when researchers use NHST .
The extracted degrees of freedom distributions are shown in Fig 2A ( degrees of freedom reflect the sample sizes of the studies , e . g . , for an independent sample t-test , the degrees of freedom are the sample size minus two ) . The median of degrees of freedom was 20 for statistically significant and 19 for statistically nonsignificant results ( mode = 15 for both ) . During the validation process , we assessed the proportion of one-sample , matched , and two-sample t-tests , which enabled us to use a mixture model to compute published effect sizes and power . The distribution of the effect sizes computed from 26 , 841 t-value records showed an excellent match to the effect size distribution determined from 2 , 185 records where effect size ( D ) -values were reported ( Fig 2B ) . This suggests that the mixture model we used is likely to well approximate the proportions of one-sample , matched , and two-sample t-tests . The computed D-value distribution was more spread out to the right relative to the reported D-value distribution , but both the medians ( computed = 0 . 654; reported = 0 . 660 ) and means ( computed = 0 . 938; reported = 0 . 889 ) were very similar . Fig 2C shows the bivariate distribution of the 26 , 841 computed D-values and degrees of freedom and represents the mean and median effect sizes for statistically significant and nonsignificant records and for the whole dataset . Most statistically significant results were reported in the df 10–20 range , and the density of nonsignificant results also increased in this range . The effect size discrepancy between statistically significant and nonsignificant results is clear ( medians , 25th and 75th quantiles for statistically significant and nonsignificant D-values , respectively: d = 0 . 932 [0 . 637–1 . 458]; d = 0 . 237 [0 . 106–0 . 421] ) . Assuming 1:1 H0:H1 odds [5] , it is apparent that many statistically nonsignificant results are missing from the data; with 1:1 H0:H1 odds , a large density of nonsignificant t-values could be expected on the left of the significance threshold and an even higher density of nonsignificant than significant results can be expected if H0:H1 odds are larger than 1 ( see the extracted t-value distribution in S2 Fig and compare the shape of this extracted t-value distribution to the expected shapes shown in S1A and S1B Fig ) . Some nonsignificant results are missing because our extraction method could not pick up stand-alone p-values . However , the bias towards having mostly significant records in the data ( amounting to three quarters of the records here ) is also consistent with strong selective reporting biases . Such biases have been demonstrated in distributions of p-values reported in abstracts and full texts of biomedical papers [14] . Overall , effect sizes computed from the extracted data are biased towards larger effect sizes . Again , this means that the FRPs we estimate here represent lower limits . For a certain effect size , power is determined by sample size , which determines degrees of freedom . Subfields showed large differences in degrees of freedom with most records having much lower degrees of freedom in cognitive neuroscience than in psychology and medicine ( Fig 3A; 25th and 75th centiles for all records for cognitive neuroscience journals: df = 10–28; psychology: df = 17–60; medical journals: df = 15–54 ) . Reported effect sizes also differed markedly by subfields ( 25th and 75th centiles for all records for cognitive neuroscience journals: d = 0 . 34–1 . 22; psychology: d = 0 . 29–0 . 96; medical journals: d = 0 . 23–0 . 91 ) . The larger reported effect sizes in cognitive neuroscience may well be the consequence of effect size exaggeration due to having smaller sample sizes ( as shown above ) and consequential low power as the following power analyses suggest . Taking into account the reported degrees of freedom , we computed power ( at α = 0 . 05 ) for effect sizes , which are traditionally considered small ( d = 0 . 2 ) , medium ( d = 0 . 5 ) , and large ( d = 0 . 8 ) [16–18] . The cumulative probability of records reaching a certain level of power is shown in Fig 3B ( for power calculation details see S1 Text , Section 3 ) . Median and mean power for subfields are shown in Table 1 . Under the assumption that standardized effect sizes are similar in all subfields tested , it is apparent that cognitive neuroscience studies had the lowest level of power . For example , to detect a small true effect ( d = 0 . 2 ) , 90% of cognitive neuroscience records had power < 0 . 234 . This is a much worse chance to detect a true effect than relying on flipping a coin [17] . A comparison to prominent older surveys of power estimates 25 and >50-y-ago showed that median power to detect medium-sized effects has increased slightly in psychology journals but remained about the same for small and large effects ( see Table 1 [16–18] ) . Power for cognitive neuroscience and for all subfields together was lower than median and mean power reported in 1962 , more than half a century ago . Median degrees of freedom and median effect sizes for each journal are depicted in Fig 4 . It is apparent that cognitive neuroscience journals report the largest effect sizes but at the same time have the smallest degrees of freedom . Consequently , they also have the lowest power levels assuming similar true effect sizes across fields and are most subject to effect size exaggeration . As a further consequence , journal impact factors negatively correlated with median power because , on the average , cognitive neuroscience journals had the largest impact factors in our sample ( correlation for small , medium , and large effect sizes , respectively with 95% accelerated and bias corrected bootstrap confidence intervals [105 permutations]: r = −0 . 42 [−0 . 63; −0 . 09]; −0 . 46 [−0 . 71; −0 . 09]; −0 . 45 [−0 . 77; −0 . 02] ) . The somewhat higher power in the journals we classified as more medically oriented was driven by the Journal of Psychiatry Research ( JPR in Fig 4; median power to detect small , medium and large effects: 0 . 23 , 0 . 74 , 0 . 86 ) , which includes more behavioral studies than the other two journals we classified as “medical . ” These other two journals , more focused on neuroimaging , still performed better than cognitive neuroscience journals and at about the same level as psychology journals ( median power to detect small , medium , and large effects: 0 . 14 , 0 . 53 , 0 . 78 ) . FRP depends on power ( which depends on sample size and effect size ) , the prestudy odds of true H0 to H1 data , and on reporting bias [5] . In this context , we use the term “bias” in a general abstract sense as a model parameter as defined by Ioannidis [5] . That is , bias stands for any kind of implicit or explicit technique , manipulation , or error which can result in the outcome that a certain proportion of results which would otherwise be reported as statistically nonsignificant will be reported as statistically significant ( see details and mathematical definition in S1 Text , Section 5 ) . For example , if the bias parameter equals 0 . 1 , that means that 10% of results which would be reported as statistically nonsignificant in the absence of bias will be now reported as statistically significant . Such bias can easily appear due to data dredging techniques even if formal NHST parameters are maintained [5] . For example , if the main ( prespecified ) analysis does not yield a formally significant result , investigators may remove or add cases [25] , change the model specification [26] and/or data preprocessing parameters in neuroimaging [27] , change the statistical analytical method , report on a different outcome , or report a statistically nonsignificant result as significant ( e . g . , reporting p = 0 . 058 as p < 0 . 05; [28] . Altogether , there are many ways that nonsignificant results may become significant . Frank publication bias ( suppression/nonpublication of nonsignificant results ) , and the rarer fraud with fabrication of nonexistent data or distorting data , yielding significant results will all lead to an excess of reported statistically significant results [29] . The continuous lines in Fig 5 estimate lower limits for FRP , using the probably inflated effect sizes computed from published t-values , for various prestudy H0:H1 odds and bias values and for α = 0 . 05 . H0:H1 odds are difficult to determine empirically . First , the nil–null hypothesis is never exactly true . From this perspective , it could be argued that even a very small deviation from the null hypothesis , i . e . , a very small effect size , could be considered not only statistically but also practically “significant . ” However , very small effect sizes are practically meaningless ( see the Materials and Methods section on further elaboration on this ) . So , when considering H0:H1 odds , it makes more sense to think about these in the context of effect sizes which could be considered practically meaningful/useful to know about . From such a perspective , it would be unrealistic to assume that most tested hypotheses are really correct ( i . e . , that they are associated with reasonable effect sizes; [5] ) ; and a recent reanalysis of the Open Science Collaboration replication project [1] also suggests that H0:H1 odds are likely to be as high as 13:1 ( 93% true H0 situations ) , at least in psychology [30] . Fig 5A represents the lower limits of FRP computed from our data for a wide range of H0:H1 odds on a 10-based logarithmic scale . Observe that FRP is under 10% only if H0:H1 odds are smaller than one . In such a case , researchers would mostly come up with correct alternative hypotheses . This is perhaps possible in conservative , very incremental research . On the contrary , in the range of explorative research where H0:H1 odds are larger than 100 , FRP is above 90% . Fig 5B shows FRP zooming into the 1–30 H0:H1 odds range for better visibility . This range of H0:H1 odds would still represent a relatively high proportion of correct alternative hypotheses but would keep the ratio of true null hypotheses slightly or moderately higher than the ratio of true alternative hypotheses . Hence , this range of H0:H1 odds represents a kind of compromise between those who would assume that most null hypotheses are false and those who would assume that most null hypotheses are correct . In the best case of having H0:H1 odds = 1:1 = 1 and zero bias , FRP is 13 . 5% . A 10% bias pushes this to 23% . Staying in the optimistic zone when every second to every sixth of hypotheses work out ( 1 ≤ H0:H1 odds ≤ 5 ) and with relatively modest 10%–30% experimenter bias , FRP is 23%–71% ( median = 51% ) . That is , between one- to three-quarters of statistically significant results will be false positives . If we now move into the domain of slightly more exploratory research where even more experimental ideas are likely to be false ( 5 < H0:H1 odds < 20; bias = 10%–30% ) , then FRP grows to at least 60%–91% ( median = 77% ) . Notably , if we consider the recent estimate of 13:1 H0:H1 odds [30] , then FRP exceeds 50% even in the absence of bias . It is important to note that here we use a single α = 0 . 05 threshold for FRP calculations because this is the rule of thumb α level used in the science fields we analyzed . That is , even if a record reports , for example , p < 0 . 001 , it does not mean that the a priori α level was α = 0 . 001 . Rather , most probably , the result would have been reported as statistically significant as long as the condition p ≤ ( α = 0 . 05 ) would have been valid ( Note that α = 0 . 05 is an assignment , p ≤ α is a test of inequality , and the parentheses are important for correct interpretation . This notation aims to emphasize the crucial difference between the p-value and the α level which are often confused . [15] ) . That is , using a single α = 0 . 05 threshold here provides the most accurate estimates about the lowest expected limits of FRP in the cognitive neuroscience and psychology literature .
The trustworthiness of statistically significant findings depends on power , prestudy H0:H1 odds , and experimenter bias [5 , 7 , 13] . H0:H1 odds are inherent to each research field , and the extent and types of biases can vary from one field to another . The distribution of the types of biases may also change within a field if focused efforts are made to reduce some types of major bias ( like selective reporting ) , for example by preregistration of studies . However , power can in principle be easily increased by increasing sample size . Nevertheless , contrary to its importance for the economic spending of research funding , the accurate estimation of effect sizes , and minimizing FRP , our data suggest that power in cognitive neuroscience and psychology papers is stuck at an unacceptably low level . This is so because sample sizes have not increased during the past half-century [16–18] . Results are similar to other fields , such as behavioral ecology where power to detect small and medium effects was 0 . 13–0 . 16 and 0 . 4–0 . 47 , respectively [31] . Assuming similar true effect sizes across fields , we conclude that cognitive neuroscience journals have lower power levels than more psychologically and medically oriented journals . This confirms previous similar inference asserting that FRP is likely to be high in the neuroimaging literature [6 , 32] . This phenomenon can appear for a number of reasons . First , neuroimaging studies and other studies using complex and sophisticated measurement tools in general tend to require more expensive instrumentation than behavioral studies , and both data acquisition and analysis may need more time , investment , and resources per participant . This keeps participant numbers low . A related issue is that science funders may have reluctance to fund properly powered but expensive studies . Second , data analysis is highly technical , can be very flexible , and many analytical choices have to be made on how exactly to analyze the results; and a large number of exploratory tests can be run on the vast amount of data collected in each brain imaging study . This allows for running a very high number of undocumented and sometimes poorly understood and difficult to replicate idiosyncratic analyses influenced by a large number of arbitrary ad hoc decisions . These , in their entirety , may be able to generate statistically significant false positive results with high frequency [27 , 33–35] , especially when participant numbers are low . Hence , sticking to low participant numbers may facilitate finding statistically significant publishable ( false positive ) results . It is also important to consider that complicated instrumentation and ( black box ) analysis software is now more available , but training may not have caught up with this wider availability . Third , in relation to more medical journals , the stakes at risk are probably lower in cognitive neuroscience ( no patients will die , at least not immediately ) , which may also allow for more biased publications . That is , researchers may be more willing to publish less reliable findings if they think that these are not directly harmful . The power failure of the cognitive neuroscience literature is even more notable as neuroimaging ( “brain-based” ) data is often perceived as “hard” evidence , lending special authority to claims even when they are clearly spurious [36] . A related concern is the negative correlation between power and journal impact factors . This suggests that high impact factor journals should implement higher standards for pre-study power ( optimally coupled with preregistration of studies ) to assure the credibility of reported results . Speculatively , it is worth noting that the high FRP allowed by low power also allows for the easier production of somehow extraordinary results , which may have higher chances to be published in high impact factor journals [37] . Standardized effect sizes depend on the largeness of effects and the noise level they are embedded in ( effect size is larger if signal to noise ratio is better ) . In behavioral psychology studies , measurement imprecision and variability ( e . g . , test–retest replicability and reliability , stableness of participant characteristics , etc . ) introduce noise . In cognitive neuroscience studies , physiological noise ( e . g . , various physiological artefacts generated externally or internally to participants ) will further contribute to measurement imprecision , while the physiological signals of interest are usually small . Hence , we could expect that measurable standardized effect sizes are in general smaller in cognitive neuroscience than in psychology because both behavioral and physiological noise may contribute to measurements ( however , note , as explained before , that due to reliance on NHST , typically only statistically significant exaggerated effect sizes are reported in papers ) . Were effect sizes really smaller , power would be even worse in cognitive neuroscience relative to psychology than indicated here . Good quality cognitive neuroscience studies may try to counteract physiological noise by increasing trial numbers in individual measurements . A larger number of trials in individuals will then decrease the standard errors of means in these individuals , which may result in smaller group level standard deviations if there is an “ideal” mean measurement value not depending on individuality ( but note that individual differences are usually neglected in group studies ) . This , in turn , will increase group-level t-values and effect sizes . Hence , consequences of individual trial numbers have already been taken into account in the calculations reported here when calculating the lower limits of FRP . Here , we have not explicitly factored in the impact of specific questionable research practices ( see , e . g . , [26 , 38] ) . Rather , we have factored in their potential joint impact through the general “bias” parameter when calculating FRP . Nevertheless , it would be important to see the individual contribution of various data dredging techniques to increasing FRP . For example , researchers may neglect multiple testing correction [39–41]; post hoc select grouping variables [42 , 26]; use machine-learning techniques to explore a vast range of post hoc models , thereby effectively p-hacking their data by overfitting models ( http://dx . doi . org/10 . 1101/078816 ) ; and/or liberally reject data not supporting their favored hypotheses . Some of these techniques can easily generate 50% or more false positive results on their own while outputting some legitimate looking statistics [25–26] . In addition , it is also well documented that a large number of p-values are misreported , indicating statistically significant results when results are , in fact , nonsignificant [41 , 43–45] . With specific respect to functional magnetic resonance imaging ( fMRI ) , a recent analysis of 1 , 484 resting state fMRI data sets have shown empirically that the most popular statistical analysis methods for group analysis are inadequate and may generate up to 70% false positive results in null data [46 , 47] . This result alone questions the published outcomes and interpretations of thousands of fMRI papers . Similar conclusions have been reached by the analysis of the outcome of an open international tractography challenge , which found that diffusion-weighted magnetic resonance imaging reconstructions of white matter pathways are dominated by false positive outcomes ( http://dx . doi . org/10 . 1101/084137 ) . Hence , provided that here we conclude that FRP is very high even when only considering low power and a general bias parameter ( i . e . , assuming that the statistical procedures used were computationally optimal and correct ) , FRP is actually likely to be even higher in cognitive neuroscience than our formal analyses suggest . Some limitations need to be mentioned for our study . First , given the large-scale automation , we cannot verify whether the extracted data reflect primary , secondary , or even trivial analyses in each paper . In the absence of preregistered protocols , however , this is extremely difficult to judge , even when full papers are examined . Evaluation of biomedical papers suggests that many reported p-values , even in the abstracts , are not pertinent to primary outcomes [3] . Second , some types of errors , such as nondifferential misclassification ( measurement error that is not related to the outcome of interest ) , may lead to deflated effect sizes . However , in the big picture , with very small power , inflation of the statistically significant effects is likely to be more prominent than errors reducing the magnitude of the effect size . Third , given the large scale automated extraction , we did not record information about characteristics of the published studies , e . g . , study design . It is likely that studies of different designs ( e . g . , experimental versus observational studies ) may have different distribution of effect sizes , degrees of freedom , and power , even within the same subdiscipline . Hence , we could not take into account the impact of the quality of experimental design on power . Fourth , here we only estimated power for a mixture model of t-tests based on the extracted degrees of freedom . Nevertheless , it is very likely that the extracted degrees of freedom give a good indication of participant numbers in studies . These participant numbers would then be strongly correlated with the statistical power of any other analyses done besides t-tests . Fifth , we could not extract all nonsignificant relevant p-values that are often reported on their own . This biased the observed effect sizes towards larger values . However , this means that the FRPs we computed really reflect lower estimates . Finally , generalizations need to be cautious , since there can be large variability in the extent of these potential biases within a given subfield . Some teams and subfields may have superb , error-proof research practices , while others may have more frequent problems . In all , the combination of low power , selective reporting , and other biases and errors that have been well documented suggest that high FRP can be expected in cognitive neuroscience and psychology . For example , if we consider the recent estimate of 13:1 H0:H1 odds [30] , then FRP exceeds 50% even in the absence of bias . The low reproducibility rate seen for psychology experimental studies in the recent Open Science Collaboration [1] is congruent with the picture that emerges from our data . Our data also suggest that cognitive neuroscience may have even higher FRP rates than psychology . This hypothesis is worth evaluating with focused reproducibility checks of published studies . Regardless , efforts to increase sample size and reduce publication and other biases and errors are likely to be beneficial for the credibility of this important literature . Some promising avenues to resolve the current replication crisis could include the preregistration of study objectives , compulsory prestudy power calculations , enforcing minimally required power levels , raising the statistical significance threshold to p < 0 . 001 if NHST is used , publishing negative findings once study design and power levels justify this , and using Bayesian analysis to provide probabilities for both the null and alternative hypotheses [12 , 26 , 30 , 48] .
|
Biomedical science , psychology , and many other fields may be suffering from a serious replication crisis . In order to gain insight into some factors behind this crisis , we have analyzed statistical information extracted from thousands of cognitive neuroscience and psychology research papers . We established that the statistical power to discover existing relationships has not improved during the past half century . A consequence of low statistical power is that research studies are likely to report many false positive findings . Using our large dataset , we estimated the probability that a statistically significant finding is false ( called false report probability ) . With some reasonable assumptions about how often researchers come up with correct hypotheses , we conclude that more than 50% of published findings deemed to be statistically significant are likely to be false . We also observed that cognitive neuroscience studies had higher false report probability than psychology studies , due to smaller sample sizes in cognitive neuroscience . In addition , the higher the impact factors of the journals in which the studies were published , the lower was the statistical power . In light of our findings , the recently reported low replication success in psychology is realistic , and worse performance may be expected for cognitive neuroscience .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"applied",
"mathematics",
"social",
"sciences",
"neuroscience",
"simulation",
"and",
"modeling",
"algorithms",
"cognitive",
"neuroscience",
"mathematics",
"statistics",
"(mathematics)",
"neuroimaging",
"research",
"and",
"analysis",
"methods",
"statistical",
"distributions",
"meta-research",
"article",
"imaging",
"techniques",
"probability",
"theory",
"psychology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"cognitive",
"science",
"behavioral",
"neuroscience",
"statistical",
"data"
] |
2017
|
Empirical assessment of published effect sizes and power in the recent cognitive neuroscience and psychology literature
|
A national survey in 1997 demonstrated that trachoma was endemic in Mali . Interventions to control trachoma including mass drug administration ( MDA ) with azithromycin were launched in the regions of Kayes and Koulikoro in 2003 . MDA was discontinued after three annual rounds in 2006 , and an impact survey conducted . We resurveyed all districts in Kayes and Koulikoro in 2009 to reassess trachoma prevalence and determine intervention objectives for the future . In this paper we present findings from both the 2006 and 2009 surveys . Population-based cluster surveys were conducted in each of the nine districts in Koulikoro in 2006 and 2009 , whilst in Kayes , four of seven districts in 2006 and all seven districts in 2009 were surveyed . Household members present were examined for clinical signs of trachoma . Overall , 29 , 179 persons from 2 , 528 compounds , in 260 clusters were examined in 2006 and 32 , 918 from 7 , 533 households in 320 clusters in 2009 . The prevalence of TF in children aged 1–9 years in Kayes and Koulikoro was 3 . 9% ( 95%CI 2 . 9–5 . 0% , range by district 1 . 2–5 . 4% ) and 2 . 7% ( 95%CI 2 . 3–3 . 1% , range by district 0 . 1–5 . 0% ) respectively in 2006 . In 2009 TF prevalence was 7 . 26% ( 95%CI 6 . 2–8 . 2% , range by district 2 . 5–15 . 4% ) in Kayes and 8 . 19% ( 95%CI 7 . 3–9 . 1% , range by district 1 . 7–17 . 2% ) in Koulikoro among children of the same age group . TT in adults 15 years of age and older was 2 . 37% ( 95%CI 1 . 66–3 . 07% , range by district 0 . 30–3 . 54% ) in 2006 and 1 . 37% ( 95%CI 1 . 02–1 . 72% , range by district 0 . 37–1 . 87% ) in 2009 in Kayes and 1 . 75% ( 95%CI 1 . 31–2 . 23% , range by district 1 . 06–2 . 49% ) in 2006 and 1 . 08% ( 95%CI 0 . 86–1 . 30% , range by district 0 . 34–1 . 78% ) in 2009 in Koulikoro . Using WHO guidelines for decision making , four districts , Bafoulabe in Kayes Region; and Banamba , Kolokani and Koulikoro in Koulikoro Region , still meet criteria for district-wide implementation of the full SAFE strategy as TF in children exceeds 10% . A community-by-community approach to trachoma control may now be required in the other twelve districts . Trichiasis surgery provision remains a need in all districts and should be enhanced in six districts in Kayes and five in Koulikoro where the prevalence exceeded 1 . 0% in adults . Since 1997 great progress has been observed in the fight against blinding trachoma; however , greater effort is required to meet the elimination target of 2015 .
Trachoma , a blinding bacterial disease of the conjunctiva , is targeted for elimination as a public health problem by the year 2020 , yet an estimated 8 . 2 million people remain at immediate risk of blindness or visual impairment due to the disease [1] . To achieve the elimination target , the World Health Organization ( WHO ) recommends member states implement an integrated strategy of interventions known as SAFE: surgery to correct trachomatous trichiasis; mass administration of antibiotics to treat current trachoma infections and reduce the infectious reservoir; promotion of hygiene and facial cleanliness; and water and sanitation as environmental improvements aimed at interrupting transmission of the infection . Based on WHO guidelines , districts are categorized for intervention based on the prevalence of clinical signs of disease: trachomatous inflammation follicular ( TF ) in children aged 1–9 years and trachomatous trichiasis ( TT ) in adults aged 15 years and older [2] , [3] . Following a national trachoma prevalence survey in 1997 , The National Blindness Prevention Program in Mali initiated a trachoma control program . Mapping of trachoma in Mali identified trachoma to be of public health significance throughout the country , including the regions of Kayes and Koulikoro where the prevalence of TF in children less than 10 years of age was 42 . 5% and 33 . 5% respectively [4] . The highest levels of TT among women 15 years of age and older were observed in Kayes ( 3 . 3% ) and Koulikoro ( 3 . 9% ) [4] . From 2002 to 2006 all sixteen districts in Kayes and Koulikoro received SAFE interventions to control trachoma . The interventions implemented in each region are listed in Table 1 . Interventions were conducted in several ways: trained ophthalmic nurses moved from village to village offering free trichiasis surgery; mass distribution of oral azithromycin and tetracycline ophthalmic ointment occurred in annual campaigns for three consecutive years in each district following pilot distributions in target areas; facial hygiene , latrine construction and use , and the utilization of water for hygiene were promoted over local and regional radio stations; and persons in each region were trained to deliver health education and promote behavior change . The number of doses distributed and population coverage with azithromycin by district and year is shown in Table 2 . In 2006 , after three years of intervention and in accordance with the WHO guidelines , an impact evaluation was conducted to assess the effect of the SAFE activities [2] . The Ministry of Health withdrew A , and support for F and E interventions from partner organizations was limited . The Ministry of Health concentrated efforts to scale up the SAFE strategy in other regions yet to initiate interventions . The purpose of this study was to re-evaluate the prevalence of trachoma three years after SAFE interventions were discontinued in Kayes and Koulikoro . Here we present the data from the first impact evaluation in 2006 and the recent 2009 evaluation . We also aimed to quantify any need for additional interventions .
These prevalence surveys were conducted in accordance with WHO guidelines as part of the ongoing effort of the Ministry of Health to eliminate blinding trachoma in Mali and were necessary to evaluate the impact of interventions . In addition to the Ministry of Health , the survey protocol was approved by the Emory University IRB under protocol 079-2006 . Informed verbal consent and assent was received according to the principles of the Declaration of Helsinki . Written consent was not obtained in these surveys due to the low literacy rate , ranging from 3% in rural Mali to 38% in Bamako ( Enquête Démographique et de Santé 2001 ) . Emory University IRB approved the use of informed verbal consent . Oral informed consent was sought first from village chiefs before surveys were conducted in the randomly selected villages . Consent was then obtained from household heads of randomly selected households and finally oral consent was obtained from each adult examined and consent from a parent or caretaker was obtained to examine children . Verbal assent was obtained also from children 6–14 years of age . Survey participants were informed of the purpose of the trachoma examinations and their rights not to participate or to stop the examination at any time . Choosing not to participate did not affect any decision in determining the need for interventions . Verbal consent was documented on a standard survey data collection tool . All children presenting signs of TF or trachomatous inflammation intense ( TI ) were offered free tetracycline eye ointment and instructed to apply it twice daily for 6 weeks . Persons identified with TT were recorded , counseled , and offered free consultation and surgery with a trained TT surgeon . Kayes Region is located in the extreme west of Mali bordering Mauritania to the north , Senegal to the west , and Guinea to the southwest ( Figure 1 ) . The region is divided into seven health districts with an estimated combined population of 1 , 763 , 987 persons ( Mali National Demographic and Statistical Institute 2009 population projection ) . The primary ethnic groups are the Sarakole ( Soninke ) and Bambara . Koulikoro Region is located in the western interior of Mali directly east of Kayes Region . It also borders Mauritania to the north and Guinea to the southwest . Koulikoro is divided into nine health districts with an estimated population of 2 , 072 , 185 persons ( Mali National Demographic and Statistical Institute 2009 population projection ) . The primary ethnic groups are the Bambara and Malinke . In both 2006 and 2009 , population-based cross-sectional household surveys were conducted at the district level . Each survey in 2006 was done at least 6 months after the last round of antibiotic distribution following the implementation plan of the national program . Thus some districts in Kayes and Koulikoro were surveyed during the period between March and May and some during November and December . In 2009 , all districts were surveyed during the period between March and May . Twenty villages ( clusters ) were selected from each district with a probability of selection proportional to the total population of the village . All villages with less than 5 , 000 total population from each district were eligible for selection . In 2006 , 4 of 7 districts in Kayes ( Bafoulabe , Diema , Kita and Nioro du Sahel ) and all 9 districts in Koulikoro were assessed . In each of these districts , concessions ( household compounds ) were systematically selected using the random direction method [5] . All residents aged 1–9 years of age and 15 years of age and older from all households within selected concessions were examined for clinical signs of trachoma until approximately 60 qualifying children had been examined . In 2009 , all 16 districts in the two regions were surveyed for a total of 320 clusters . Households within a cluster were randomly selected following the method of sketch mapping and segmentation which aimed to survey 24 households per cluster [6] . With the assistance of village leaders , survey teams drafted a list of all households , dividing households into segments of four . Village chiefs selected segments via lottery . All households in a selected segment were surveyed and all consenting persons over six months of age in each household were examined for trachoma . From the sampling methodology used in both surveys we assumed that the data was self-weighted . Residents of selected households in 2009 were enumerated and designated as either present or absent . Absent was defined as not being physically present in the village on the day of the survey . Enumerated persons who were not at home , but in the village were found and recruited to the survey . Teams made at least one attempt on the same day to find residents marked absent during the first visit to the household . Absent residents in the 2006 survey were not enumerated . Clinical signs of trachoma were assessed using the WHO Simplified Grading System [3] . Examiners recorded the presence or absence of all trachoma grades in both eyes of survey participants using a ×2 . 5 binocular loupe and adequate light . The findings from the worst affected eye were reported . At examination in 2009 , children were assessed for a clean face , defined as the absence of both ocular and nasal discharge . In 2009 , each child 6–15 years of age was asked about their attendance in school , defined as public or private non-religious school . Attendance of Koranic schools or non-formal education was not assessed . Examined persons were asked about their participation in the most recent round of antibiotic distribution for trachoma control , defined as whether a person took oral azithromycin , applied tetracycline eye ointment , or did not participate . Estimates for participation in antibiotic distribution included only those persons present to give a response . Additionally in 2009 , one adult respondent was interviewed in each household to determine the presence and use of a household latrine and the location of the main water source used by the household . The presence of a latrine was confirmed by direct observation and ‘use’ was defined as the observation of feces in the pit . The location of the water source was designated as within the household compound , within the village , or outside the geographical village boundaries as a proxy for distance and availability of water . Household interviews were not conducted in the 2006 surveys . Prior to the surveys , ophthalmic nurses were trained to use the WHO Simplified Trachoma Grading System through repetitive grading of digital photographs in a classroom setting and assessment of individual patients in the field . In 2009 , these exercises were followed by a formal inter-observer reliability test of trachoma grading against a standardized set of 50 slides presented on a computer and a field exam of 50 children in which SB , DS and JDK were considered the reference examiners . Reference grading was supplemented with digital photographs . Eight out of ten ophthalmic nurses met the criteria of achieving greater than 80% reliability and a kappa statistic of 0 . 6 and above for grade TF and were selected as examiners for the survey . Survey teams were trained to randomly select households within a cluster , conduct household interviews , and record findings on standardized forms . A survey team consisted of one data recorder and one ophthalmic nurse . Formal inter-observer reliability tests were not conducted for the examiners in 2006 . Data were double-entered , compared and corrected . Based on the survey design used , we adjusted confidence intervals for the prevalence estimates and odds ratios to account for correlation among the data due to clustering using SAS SURVEY procedures ( SAS version 9 . 2 , SAS Institute Inc ) [7] , [8] . Regional prevalence estimates accounted for population differences between districts . We calculated differences between prevalence estimates from 2006 and 2009 and tested the equality of the estimates using the Z statistic with α = 0 . 05 . The ultimate intervention goal considered for achieving blinding trachoma elimination is the presence of less than 1 TT case per 1 , 000 population [9] . We calculated the total backlog of persons with TT in need of surgery by multiplying the 2009 point estimate and confidence limits of the population prevalence of TT by the estimated total population to give a point estimate and lower and upper bounds of the total number of people to be operated . According to WHO guidelines , where district-level prevalence of TF in 1–9 year-old children exceeds 10% at baseline , A , F and E activities are warranted district-wide and thus the total population living in these areas is targeted [2] . Where SAFE activities have been implemented , all areas that remain above 5% TF prevalence among children should continue antibiotic distribution [2] . The target prevalence by which mass antibiotic interventions to control trachoma is not needed is below 5% TF [2] . We calculated the number of household latrines required to achieve goal 7c of the United Nations Millennium Development Goals ( MDGs ) ; halve by 2015 the proportion of people who do not have access to improved sanitation [10] .
In 2006 , a total of 29 , 179 persons were examined from 29 , 779 persons available in 2 528 selected concessions . The mean number of concessions per district was 194 . 5 with a range by district of 110 to 312 concessions . A mean of 9 . 7 concessions were surveyed per village ( range by village 1–25 ) . The mean number of households per concession was 1 . 9 ( range by concession 1–17 ) . In the four surveyed districts of Kayes , 4 , 168 adults over 14 years of age and 4 , 808 children 1–9 years of age were examined for clinical signs of trachoma . In Koulikoro , 9 , 679 adults and 10 , 524 children were examined . Among examined adults over 14 years of age , 68 . 9% were women and among examined children 1–9 years of age , 51 . 5% were girls . The prevalence estimates of clinical signs of trachoma in Kayes and Koulikoro in 2006 are presented in the first three columns by district in Table 3 . Among adults 15 years of age and older , the prevalence of TT in Kayes was 2 . 37% ( 95%CI 1 . 66–3 . 07% , range by district 0 . 30–3 . 54% ) and in Koulikoro , 1 . 75% ( 95%CI 1 . 31–2 . 23% , range by district 1 . 06–2 . 49% ) . The prevalence of trachomatous scaring ( TS ) among adults was 10 . 33% ( 95%CI 8 . 6–12 . 0% , range by district 3 . 0–18 . 4% ) and 4 . 18% ( 95%CI 3 . 5–4 . 8% , range by district 0 . 6%–9 . 2% ) in Kayes and Koulikoro respectively ( data not shown ) . Prevalence of trachomatous corneal opacity ( CO ) in Kayes was 0 . 38% ( 95%CI 0 . 14–0 . 61 , range by district 0 . 0–0 . 9% ) and 0 . 31% ( 95%CI 0 . 12–0 . 51% , range by district 0 . 0–0 . 9% ) in Koulikoro . Women were more likely than men to have TT ( OR = 1 . 61 , 95%CI 1 . 16–2 . 23 , p = 0 . 004 ) . Adults 50 years and older were more likely to have TT than adults aged 15–49 years ( OR = 6 . 73 , 95%CI 4 . 99–9 . 07 , p = <0 . 0001 ) . District-level prevalence of TF among children 1–9 years of age had reduced to below the 10% intervention threshold in all surveyed districts . Among children 1–9 years of age , the prevalence of TF was 3 . 9% ( 95%CI 2 . 9–5 . 0% , range by district 1 . 2–5 . 4% ) in Kayes and 2 . 7% ( 95%CI 2 . 3–3 . 1% , range by district 0 . 1–5 . 6% ) in Koulikoro . The prevalence of trachomatous inflammation intense ( TI ) among children aged 1–9 years of age was 1 . 0% ( 95% CI 0 . 6–1 . 5% , range by district 0 . 3–1 . 8% ) in Kayes and 0 . 4% ( 95% CI 0 . 2–0 . 5% , range by district 0–2 . 0% ) in Koulikoro . Active trachoma ( TF and/or TI ) prevalence was 4 . 53% ( 95%CI 3 . 3–5 . 7% , range by district 1 . 2–6 . 5% ) in Kayes and 2 . 96% ( 95%CI 2 . 5–3 . 4% , range by district 0 . 2–5 . 6% ) in Koulikoro . In 2009 , from all districts in both regions a total of 42 , 128 persons were enumerated in 7 , 533 households and 32 , 918 were examined . In Kayes , a total of 13 , 576 persons were examined for signs of trachoma out of 17 , 127 persons enumerated from 3 , 287 households for a response rate of 79 . 3% . In Koulikoro , 19 , 342 persons were examined out of 25 , 001 persons enumerated from 4 , 246 households ( a response rate of 77 . 4% ) . The response rate in women was 83 . 1% ( 17 , 771/21 , 386 ) and 73 . 0% ( 15 , 147/20 , 742 ) in men . The majority of adult men unable to be examined were absent from the home at the time of the household visit . Children 1–9 years of age composed 33 . 9% of the total enumerated population . Adults 15 years of age and older were 51 . 0% of the total population . The proportion of enumerated children 1–9 years of age who were examined was 88 . 2% and 77 . 1% of enumerated adults were examined . Among examined adults over 14 years of age , 58 . 1% were women and among examined children 1–9 years of age , 49 . 7% were girls . Among children 6–15 years of age the proportion that reported attending school was 42 . 6% in Kayes and 54 . 1% in Koulikoro . The prevalence estimates of clinical signs of trachoma in Kayes and Koulikoro for 2009 are presented by district in the last four columns of Table 3 . The prevalence of TT in the total population of Kayes region was 0 . 69% ( 95%CI 0 . 53–0 . 85% , range by district 0 . 20–0 . 91% ) . In Koulikoro , TT prevalence in the total population was 0 . 56% ( 95%CI 0 . 43–0 . 69% , range by district 0 . 25–0 . 85% ) . Among adults 15 years of age and older , the prevalence of TT in Kayes was 1 . 45% ( 95%CI 1 . 10–1 . 79% , range by district 0 . 37–1 . 87% ) and in Koulikoro , 1 . 10% ( 95%CI 0 . 84–1 . 35% , range by district 0 . 34–1 . 78% ) . The prevalence of trachomatous scaring ( TS ) among adults was 4 . 22% ( 95%CI 3 . 7–4 . 8% , range by district 0 . 3–5 . 3% ) and 4 . 68% ( 95%CI 4 . 1–5 . 2% , range by district 1 . 4–8 . 1% ) in Kayes and Koulikoro , respectively ( data not shown ) . Prevalence of CO in Kayes was 0 . 11% ( 95%CI 0 . 03–0 . 19 , range by district 0–0 . 42% ) and 0 . 21% ( 95%CI 0 . 13–0 . 29% , range by district 0–0 . 73% ) in Koulikoro . Odds of TT among adults 50 years of age and older were ten times higher than adults 15–49 years of age ( OR = 10 . 61 , 95%CI 7 . 62–14 . 78 , p<0 . 0001 ) . Women were nearly two times more likely to have TT than men ( OR = 1 . 85 , 95%CI 1 . 40–2 . 46 , p<0 . 0001 ) . At the regional level , the prevalence of TF was 6 . 6% ( 95%CI 5 . 7–7 . 5% , range by district 2 . 5–15 . 4% ) among children 1–9 years of age in Kayes and 8 . 7% ( 95%CI 7 . 5–9 . 9% , range by district 1 . 7–17 . 2% ) in Koulikoro . The prevalence of TI among children aged 1–9 years of age was 1 . 5% ( 95% CI 1 . 1–1 . 8% , range by district 0 . 3–3 . 3% ) in Kayes and 0 . 6% ( 95% CI 0 . 4–0 . 8% , range by district 0–1 . 9% ) in Koulikoro . Active trachoma ( TF and/or TI ) prevalence was 7 . 34% ( 95%CI 6 . 4–8 . 3% , range by district 2 . 7–16 . 8% ) in Kayes and 8 . 91% ( 95%CI 7 . 7–10 . 1% , range by district 2 . 0–17 . 9% ) in Koulikoro . A total of 7 , 533 households were surveyed ( range by district 423–480 ) . The mean number of persons living in each household was 5 . 2 ( SD = 2 . 7 , range by district 4 . 5–5 . 9 ) in Kayes and 5 . 9 ( SD = 3 . 0 , range by district 4 . 9–6 . 9 ) in Koulikoro . Indicators of uptake of the A , F and E components of the SAFE strategy are listed by district in Table 4 . The proportion of examined household residents reporting taking azithromycin or using tetracycline eye ointment in the most recent round of distribution was 86 . 1% ( 95%CI 84 . 2–88 . 0 , range by district 54 . 6–99 . 8% ) in Kayes and 83 . 9% ( 95%CI 81 . 6–86 . 3% , range by district 59 . 9–96 . 8% ) in Koulikoro . Among children 1–9 years of age , 76 . 5% ( 95%CI 74 . 3–78 . 7% , range by district 46 . 7–95 . 2% ) and 75 . 0% ( 95%CI 71 . 8–78 . 1% , range by district 52 . 1–86 . 8% ) in Kayes and Koulikoro , respectively , had a clean face at examination . Basic sanitation ( a household latrine ) was evident in over 80% of the households in 12 out of the 16 districts . The presence of a latrine with evidence of use was observed in 88 . 1% ( 95% CI 85 . 2–91 . 1% , range by district 50 . 4–100% ) of surveyed households in Kayes and 87 . 2% ( 95%CI 84 . 5–89 . 9% , range by district 37 . 4–99 . 8% ) in Koulikoro . A water source inside the compound was observed in 9 . 7% ( 95%CI 5 . 5–13 . 9% , range by district 0 . 6–17 . 2% ) of surveyed households in Kayes and 3 . 2% ( 95%CI 0 . 5–6 . 0% , range by district 0–11 . 8% ) reported having to travel outside the geographical boundaries of the village to collect water . In Koulikoro , 19 . 8% ( 95%CI 15 . 6–24 . 0% , range by district 0–31 . 9% ) of households had a source of water within the compound and 5 . 6% ( 95%CI 3 . 5–7 . 7% , range by district 0–17 . 2% ) reported having to collect water from a source outside of village boundaries . Overall , there was no difference in the prevalence of clean faces between children living in households with water access inside the compound and children in households where the water source was outside the compound; 76 . 3% compared to 75 . 7% , Z = 0 . 48 , p = 0 . 633 . The regional estimates of prevalence of TT and CO among women 15 years of age and older from 2006 and 2009 are plotted in Figure 2 with the same estimates reported in 1997 for a comparison to baseline prevalence . In this age group , the difference in prevalence of TT between 2006 and 2009 for Kayes ( Z = −1 . 33 , p = 0 . 1829 ) and Koulikoro ( Z = −1 . 78 , p = 0 . 0744 ) regions was not statistically significant . There was no statistically significant difference in regional estimates of CO among adult women from 2006 to 2009 ( Kayes Z = −1 . 01 , p = 0 . 3117; Koulikoro Z = −0 . 44 , p = 0 . 6626 ) . However , among adults of both genders , the prevalence of TT in 2009 was less than the estimate in 2006 for both Kayes ( Z = −2 . 06 , p = 0 . 0396 ) and Koulikoro ( Z = −2 . 79 , p = 0 . 0052 ) . The prevalence of CO among all adults between 2006 and 2009 did not differ in Kayes ( Z = −1 . 21 , p = 0 . 2245 ) or Koulikoro ( Z = −0 . 55 , p = 0 . 5838 ) . The regional prevalence of TF in 2009 was statistically greater than that observed in 2006 for both regions ( Kayes Z = 8 . 13 , p<0 . 0001; Koulikoro Z = 16 . 20 , p<0 . 0001 ) . The prevalence of TF for the region in 1997 is plotted with the district level estimates of TF from 2006 and 2009 in Figure 3 . The differences in district level estimates between 2009 and 2006 with confidence intervals are listed in Table 5 along with Z statistic and p-values . The prevalence of TF observed in 2009 was statistically greater than that observed in 2006 for Bafoulabe , Nioro du Sahel , Banamba , Dioila , Fana , Kati , Kolokani and Koulikoro districts . The prevalence of TF in 2009 was the same or less than that observed in 2006 in the districts of Diema , Kita , Kangaba , Nara and Ouelessebougou . Regional estimates of TI among children from 1997 , 2006 and 2009 are shown in Figure 4 . Also for both regions , the prevalence of TI in this study was greater than that observed in 2006 ( Kayes Z = 2 . 06 , p = 0 . 0198; Koulikoro Z = 1 . 86 , p = 0 . 0316 ) . The prevalence of TI observed in 2009 was statistically greater than that observed in 2006 for Bafoulabe , Banamba , Dioila and Fana districts . Based on 2009 estimates , the total number of persons with TT who remain in need of surgery in Kayes is 10 967 ( lower and upper bounds: 7 , 144 to 14 , 123 ) and 10 , 726 ( bounds: 9 , 932 to 16 , 487 ) in Koulikoro . TT prevalence among adults exceeded 1% in 11 of 16 districts warranting continued , enhanced efforts to provide surgery to affected patients . While TT surgery may not be a priority in Yelimane , Fana , Kangaba , Nara and Oulessebougou where TT among adults is less than 1% , eye care facilities with the capacity to operate presenting TT cases should exist . Mass distribution of antibiotics should resume in Bafoulabe , Banamba , Kolokani and Koulikoro where the prevalence of TF exceeds 10% among children . Additionally , according to WHO guidelines , mass distribution of antibiotics should continue in areas where after three years of intervention , the prevalence of TF remains greater than 5% among children under 10 years of age [2] . This would include communities within all of the 12 other districts of Kayes and Koulikoro . If this WHO guideline is interpreted at the district level , 10 districts in Kayes and Koulikoro regions warrant ongoing mass distribution of antibiotics , targeting a total population of approximately 2 , 637 , 492 persons . The promotion of facial hygiene and environmental improvements should resume in all districts . Access to water within village boundaries and household latrine coverage was not lacking in most districts . Fana , Dioila , Yelimane and Nara had the highest proportion of households reporting having to collect water outside village boundaries . An estimated 11 , 526 households in Kayes and 19 , 718 households in Koulikoro must collect water outside village boundaries . The construction and maintenance of water points could be targeted to communities where access to water is lacking . Kenieba in Kayes and Nara in Koulikoro had the lowest proportion of households with a latrine . To ensure that every household has access to basic improved sanitation , 41 , 712 latrines need to be built in Kayes and 33 , 928 in Koulikoro . Building half of these by 2015 would meet MDG 7c .
The national survey conducted in 1997 providing baseline regional prevalence estimates were very useful in establishing the widespread nature of trachoma in Mali . In response to results from the national survey , trachoma control interventions were initiated in Kayes and Koulikoro Regions . Interventions were focused mostly on the S and A components of SAFE . F and E interventions were implemented , but had less geographical coverage of the target population than S and A . Until 2006 , monitoring of interventions was limited to program reports and did not include rigorous field evaluations . According to reports of azithromycin distributed , antibiotic coverage was not consistent between districts or years with reported district level coverage ranging from 20 . 9% to 108 . 6% after the pilot phase in 2002 . Several districts failed to reach the desired minimum of 80% coverage of total population in any one year and only one averaged above 80% over three years . These inconsistencies in coverage were either due to problems with the distribution or estimates of the target population . For example , the total population registered prior to MDA may have exceeded the census estimate of total population where coverage was greater than 100% . An evaluation of antibiotic distribution in Southern Sudan demonstrated the limitations of using distribution reports alone to calculate population coverage , as population estimates and treatment records can lead to inaccurate coverage estimates [11] . In Kayes and Koulikoro , we have defined “distributed” as the total number of doses reported to have been given out to individuals during mass drug administration campaigns , and caution that they have not been validated with coverage surveys . The figures are those reported by the district to the national program . The first impact assessment in 2006 found prevalence of TF among children to be below 5% in 9 of 13 districts and below 10% in all districts . The programmatic decision was made to focus the available resources to other endemic regions that had not initiated SAFE interventions . This resulted in stopping mass antibiotic distribution and limited ongoing promotion of facial cleanliness and environmental improvements through schools and radio . Follow-up on the progress of latrine construction and new water points targeted for trachoma control stopped . Surgical services to correct trichiasis were maintained . Between the surveys in 2006 and 2009 , it appears as though clinical signs of active trachoma returned in eight out of the thirteen districts . The current data have several possible interpretations . There may have been a true decline in active trachoma from baseline to the present and this decline is associated with interventions from 2003–2006 . Although prevalence of active trachoma signs are higher in some districts now than observed in 2006 , the prevalence remains well below the 34% and 42% TF reported in Koulikoro and Kayes respectively during the 1997 baseline survey . National programs do not have control groups and it is not possible to determine whether the decline is due to the intervention , or to a secular decline , as has been described elsewhere [12] , [13] , [14] , [15] . Prevalence of active trachoma has been observed to decline in the absence of a trachoma control program in the dessert region of Kidal [16] . We may also consider that there has been an heterogeneous effect of the interventions with some districts showing a sustained reduction in the prevalence of TF ( Diema , Kita , Fana , Kangaba , Nara and Oulessebougou ) and others showing a rapid rebound after initial control ( Bafoulabe , Nioro du Sahel , Banamba , Dioila , Kati , Kolokani and Koulikoro ) . Such random effects are assumed possible by chance at the community level according to a stochastic model of trachoma transmission [17] . Models also suggest that trachoma endemicity at baseline is predictive of return of infection after antibiotic intervention [18] , yet we have no district-level estimates at baseline on which to make assumptions . Antibiotic coverage is an important factor in the return of infection after treatment and thus the elimination of trachoma [17] , [18] , [19] . It is possible that high-risk marginalized sections of the population are systematically missed in mass drug administration leaving them untreated and able to repeatedly reintroduce infection into treated communities . Coverage surveys performed immediately following the mass distribution campaigns at least once during the three years of intervention may have identified any such problem . Alternatively , there may be no difference between prevalence estimates of active trachoma in 2006 and 2009 due to the differences between the survey methods used and season of assessment in some districts , although this is unlikely given the scale of the observed differences and that seasonality of trachoma in Mali has not been established . In the 2006 survey , household selection methods may have biased the samples in villages where only a few large concessions were selected . The starting points were markets or mosques , structures typically at the center of a community and often surrounded by more populated concessions . Some clusters in the 2006 survey were composed of persons examined from households within a few , very large concessions , rather than the randomly distributed sample of households obtained using the sketch mapping and segmentation in the 2009 surveys . The household sampling in 2009 was more similar to that used in 1997 where a systematic random sample was taken from a listing of households within clusters . Both evaluations began with training ophthalmic nurses in the WHO simplified trachoma grading system . However , in 2009 , the grader's reliability to diagnose TF was assessed rigorously and nurses not meeting a certain criteria were excluded from serving as a grader . This type of field assessment should improve the validity and reliability of a grader's findings . The observed reduction in the prevalence of TT may have been a direct effect of the ongoing surgical services provided to TT patients . The diagnosis of TT is straightforward and allows less room for subjectivity than TF since the grade is based on one or more lashes touching the eye , rather than 5 or more follicles greater than 0 . 5 mm in diameter in the central part of the tarsal conjunctiva [3] . The grader's ability to identify TT is assessed in the classroom using slides but not in the field reliability assessment [2] . It may have been possible that graders under diagnosed TT in the field , but the possibility of this type of misclassification should not have differed from 2006 and 2009 . Additionally , a greater proportion of adult males were absent from the household than females . TT impairs vision and thus compromises mobility; therefore men with healthy eyes may be more likely to be absent and not examined . Only if the reverse is true , men present in the household are more likely to have unhealthy eyes , would any bias in the prevalence of TT in men have masked any gender difference in TT . In both impact evaluations , women were more likely to have TT than men , which is consistent with findings from a recent review on the association of gender and trichiasis [20] . The statistically significant difference observed between TT prevalence among adults of both genders , but not among adult females from 2006 to 2009 , may suggest a gender disparity in benefit from ongoing surgical services with men being more likely than women to present for surgery . Eliminating the backlog of trichiasis patients needing surgery remains a priority in both regions . Surgical services may need modification to specifically target women . The indicators for A , F and E uptake ( Table 4 ) obtained from the household surveys have several limitations . Although antibiotic coverage obtained from personal reports from household residents appears high , these results should be interpreted with caution . Residents were asked whether they had taken azithromycin during the most recent mass distribution campaign , which was in 2006 . It is unlikely that residents could recall specifically taking drugs for trachoma given that mass drug distribution campaigns for other NTDs had occurred in more recent years . Additionally , only responses from residents available to respond were taken . These residents may have been more likely to have been available to receive antibiotics during campaigns than those residents absent from the household at the time of the survey , potentially inflating the coverage estimate . Not surprisingly , these personal reports are higher than coverage estimated by district distribution reports ( Table 2 ) . More than 75% of households surveyed in each district , except Nara and Keneiba , had access to a household latrine with evidence of use . The evidence of use was determined by the presence of feces in the pit , which may be incorrectly interpreted as latrine use by all persons within the household . A latrine will be categorized as ‘in use’ if only a proportion of the household is using it . The role of latrines in reducing trachoma transmission assumes that where latrines are used , no open human feces is available for flies to utilize as a medium for egg and larval development; reducing successive fly populations and reducing the number of fly to eye contact . However , if use of a latrine is limited to only certain groups or if certain groups choose not to use the latrine , open defecation will continue . Further evaluation may be needed to assess actual behavior and potentially explain conflicting outcomes of endemic trachoma in the presence of high sanitation coverage as seen in Bafoulabe and Banamba districts . Assessing behavior is also necessary in determining the influence of water on trachoma . In this survey , there was no association of a clean face and having access to a water source within the boundaries of the household compound or having access to a water source outside of village boundaries . Additionally , greater than 80% of children were observed to have a clean face in only four districts; indicating that the practice of face-washing has not been fully accepted and adopted among residents in the two regions . On the contrary , our findings may also indicate that the ability of F and E components to control trachoma may not be as effective as anticipated . However , a recent analysis of factors associated with active trachoma in Mali supports the utility of face washing and environmental improvements [21] . One of the criteria for the certification of the elimination of trachoma is to demonstrate the sustained reduction of prevalence of TF among children below 5% for a period of three years after interventions have ceased [8] . In only six districts did the point estimate of the prevalence of TF remain below 5% at the district level from 2006 to 2009 . In the 2009 survey the prevalence of TF among children was above 10% in four districts and above 5% in another six districts . With the recent global expansion of mass distribution of antibiotics for trachoma elimination , national programs may soon face a need to prioritize a limited quantity of drug [22] . Given such circumstances , Mali is facing unique programmatic decisions . Currently , WHO guidelines suggest the district be the implementation unit , but for certification of elimination , no community must have more than 5% TF among children [2] , [9] . This suggests a community-by-community approach to trachoma elimination even in districts where district-level estimates of TF prevalence are below 5% . There are no recommendations or guidelines as to how a country such as Mali should attempt to demonstrate each and every community throughout the vast landscape has reached the elimination target . An acceptable level of TF prevalence at which the risk of developing blinding trachoma has been eliminated is unknown if the acceptable level is not zero . TF is not closely correlated with the presence of Chlamydia trachomatis DNA on ocular swabs and is thought to linger in the absence of infection [23] , [24] , [25] , [26] . However , TI is better associated with DNA positive ocular swabs and is also linked to increased likelihood of progression to scarring , so is considered a more severe form of the disease [23] , [27] . Ocular Chlamydia infection may have been significantly reduced by the interventions as evidenced by prevalence of TI in 9 districts of less than 1 child per 100 . TI is more closely correlated to current infection with Chlamydia trachomatis than residual TF and has been suggested as a potential marker of infection post treatment [28] . In this setting , microbiological supporting evidence of the presence of bacteria would be useful , yet no guidelines exist for the use of laboratory diagnostics on a programmatic scale and it is perceived that costs of adding such tests to impact evaluations are prohibitive . Using TI for a proxy of infection , the prevalence of TI in 5 of the 10 districts in Kayes and Koulikoro qualified to receive mass distribution of antibiotics based on TF , indicates that less than 1 per 100 children would receive trachoma-specific benefits from the antibiotic . Achieving less than 5% TF at the district level is achievable and can be feasibly determined on a programmatic scale through the cluster random survey design as demonstrated in this study and in Ghana [29] . Not considering the differences in survey methodology , a district level prevalence of less than 5% TF after three continuous years of heavy antibiotic intervention did not equate in all districts to a sustained reduction of TF below 5% . No surveillance activities were implemented after stopping AFE interventions in these districts . Doing so may have identified resurgence in districts with an apparent rebound in active trachoma and allowed immediate intervention . Results from these surveys provide evidence in the setting of a national program that antibiotics alone are not enough to eliminate trachoma . An analysis of associations between the components of the SAFE strategy demonstrates clearly that changes in hygiene behavior and improved sanitation can have protective effects against active trachoma [30] , which argues for equal emphasis on hygiene and environmental improvements . Indicators used in the 2009 survey suggest very high access to sanitation in the two regions , but the indicators fail to capture actual behaviors . The promotion of facial cleanliness and good hygiene behavior should be reintroduced in all districts of Kayes and Koulikoro . Surgical services to correct trichiasis should also be continued , but where and for how long to continue mass distribution of antibiotics is not as clear . Currently , the 4 districts with TF above 10% among children are priority for mass distribution of antibiotics . More guidelines from the international community are urgently required to help prioritize the limited quantity of donated antibiotic in addition to recommending appropriate evaluation methodology for determining when certification targets have been achieved .
|
Trachoma , a blinding bacterial disease , is targeted for elimination by 2020 . To achieve the elimination target , the World Health Organization ( WHO ) recommends member states implement the SAFE strategy; surgery , mass administration of antibiotics , promotion of hygiene and facial cleanliness and water and sanitation as environmental improvements . We present results from evaluation surveys conducted in 2006 and 2009 from the regions of Kayes and Koulikoro , Mali . Prevalence of active trachoma in 2006 was below baseline intervention thresholds in all surveyed districts and the national program stopped antibiotic distribution . The prevalence of trachoma in 2009 remained well below levels in 1998 . However , in 8 of 13 districts compared , the prevalence of active trachoma was higher in 2009 than 2006 . Three years of antibiotic intervention did not equate in all districts to a sustained reduction of active trachoma . No surveillance activities were implemented after stopping interventions . Surgical interventions may have reduced the burden of blinding trachoma but there is an ongoing need for surgeries specifically targeting affected women . Four districts meet the WHO criteria for resuming district-wide mass antibiotic distribution . A community-by-community approach to elimination may be needed in other districts . The promotion of facial cleanliness and good hygiene behavior should be reintroduced .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"public",
"health",
"and",
"epidemiology/global",
"health",
"public",
"health",
"and",
"epidemiology/infectious",
"diseases"
] |
2010
|
Where Do We Go from Here? Prevalence of Trachoma Three Years after Stopping Mass Distribution of Antibiotics in the Regions of Kayes and Koulikoro, Mali
|
Self-association is a common phenomenon in biology and one that can have positive and negative impacts , from the construction of the architectural cytoskeleton of cells to the formation of fibrils in amyloid diseases . Understanding the nature and mechanisms of self-association is important for modulating these systems and in creating biologically-inspired materials . Here , we present a two-stage de novo peptide design framework that can generate novel self-associating peptide systems . The first stage uses a simulated multimeric template structure as input into the optimization-based Sequence Selection to generate low potential energy sequences . The second stage is a computational validation procedure that calculates Fold Specificity and/or Approximate Association Affinity ( K*association ) based on metrics that we have devised for multimeric systems . This framework was applied to the design of self-associating tripeptides using the known self-associating tripeptide , Ac-IVD , as a structural template . Six computationally predicted tripeptides ( Ac-LVE , Ac-YYD , Ac-LLE , Ac-YLD , Ac-MYD , Ac-VIE ) were chosen for experimental validation in order to illustrate the self-association outcomes predicted by the three metrics . Self-association and electron microscopy studies revealed that Ac-LLE formed bead-like microstructures , Ac-LVE and Ac-YYD formed fibrillar aggregates , Ac-VIE and Ac-MYD formed hydrogels , and Ac-YLD crystallized under ambient conditions . An X-ray crystallographic study was carried out on a single crystal of Ac-YLD , which revealed that each molecule adopts a β-strand conformation that stack together to form parallel β-sheets . As an additional validation of the approach , the hydrogel-forming sequences of Ac-MYD and Ac-VIE were shuffled . The shuffled sequences were computationally predicted to have lower K*association values and were experimentally verified to not form hydrogels . This illustrates the robustness of the framework in predicting self-associating tripeptides . We expect that this enhanced multimeric de novo peptide design framework will find future application in creating novel self-associating peptides based on unnatural amino acids , and inhibitor peptides of detrimental self-aggregating biological proteins .
In nature , proteins and peptides self-assemble and associate to produce a variety of diverse structures such as cellular nanomachines and multimeric structures , including cellular pumps , cytoskeletal filaments , and fibrils [1] . These complex biological structures can serve as templates for the design of novel bioinspired nanomaterials , as well as for the exploration of the underlying mechanisms of self-assembly [2] , [3] . The self-assembly of proteins is associated with the formation of amyloid fibrils that is implicated in the onset of Alzheimer's disease and other degenerative diseases [3]–[6] . While the causes of the onset of the formation of the disruptive fibrillar macrostructure has been well studied , the exact mechanism of self-assembly is not fully understood [6] , [7] . It is known that even in large self-assembling peptides , the association can be driven by only a few key interacting residues [8]–[12] . For this reason , the de novo design and discovery of small peptides that self-assemble will have major implications for the understanding of the determinants of self-assembly , as well as for providing insights that can be used to disrupt such associations . In addition to the medical relevance of self-assembling peptides and proteins , self-assembly in nature provides interesting and potentially fruitful avenues for biomaterial production , a field that has been amply covered in a variety of reviews [1] , [13]–[25] . Small , self-assembling peptide structures are of particular interest as they are relatively inexpensive to produce by standard chemical synthesis [26] and provide tunability of properties through substitution of individual amino acids [27]–[29] . This allows for a “bottom-up” approach to creating novel self-assembled biomaterials [19] , [20] . Several notable small associating peptides have been discovered by derivation of natural systems ( e . g . , Alzheimer's β-amyloid protein ) and through rational design [13] , [14] , [25] . The design of self-assembling peptides for biomedical and biomaterial purposes has most commonly been performed through rational design and large-scale screening . The discovery of a self-assembling dipeptide [30]–[32] has demonstrated the applicability of methods to such a problem . However , the size of the peptide is limiting in this design process , since the immense sequence space ( 20N possible designed sequences , where N is the number of design positions ) that must be searched may , in many cases , overstretch the combinatorial capabilities of such experimental methods . Due to the considerable cost and time involved in synthesizing and testing a large number of candidate peptides , it is highly desirable to screen computationally for self-assembly properties prior to experimental testing of peptides . For this reason , the application of computational methods to the design of self-assembling peptides is highly desirable . Computational protein design methods have become increasingly prevalent in the field of protein engineering . These design methods include those that employ probabilistic algorithms like Monte Carlo ( MC ) methods [33]–[36] and genetic algorithms [37] , as well as deterministic algorithms like dead end elimination ( DEE ) [38]–[42] , self-consistent mean field ( SCMF ) methods [43]–[48] , or quadratic assignment-like global optimization for sequence selection followed by fold specificity and approximate binding affinity [49]–[55] . Such computational methods allow for the consideration of large numbers of amino acid-amino acid interactions simultaneously . Computational design has been used to design inhibitors against H1N1 influenza hemagglutinin [56] , to switch cofactor specificity of an enzyme [57] , for generalized antibody design for recognition of a target epitope [58] , for the design of entry inhibitors of HIV-1 gp41 [59] , for the design of C3a receptor agonists for medicinal use [54] , and for the design of inhibitors of the histone methyltransferase EZH2 [55] . See Fung et al . [51] , Pantazes et al . [60] , Samish et al . [61] , and Khoury et al . [62] for reviews of the recent advances and successes in the area . As computational methods for single peptides and protein-peptide complexes have improved , the general interest in the design of multimeric protein assemblies for therapeutic and biomaterial applications [63]–[65] has also increased . Recently , there have been a number of successful computational designs carried out to create unique multimeric protein structures [66]–[69] , Here we present a de novo protein/peptide design framework applicable to multimeric systems and its application to the design of self-associating tripeptides . This framework utilizes a computationally-generated multimeric assembly [51] , [70] of the self-assembling tripeptide Ac-IVD [11] as the template for an optimization-based Sequence Selection method [49] , [50] , [71] , [72] . Selected sequences are then computationally screened via a Fold Specificity calculation [70] and/or calculation of Association Affinity via molecular dynamics ( MD ) simulations . The Association Affinity metric is based on statistical mechanics [59] , [73] and is used to select a small set of high confidence peptide sequences from the candidate set . To experimentally validate the framework , six in silico designed sequences were selected for experimental assessment based on the metrics described . We found that two of these tripeptides ( Ac-VIE , Ac-MYD ) formed hydrogels on time scales and at concentrations comparable to the template peptide Ac-IVD . Shuffled control sequences of these designed hydrogelating peptides were further experimentally and computationally assessed to validate the approach . Remarkably , Ac-YLD was capable of rapidly associating into large crystals under ambient conditions , which led to the elucidation of its crystal structure . The structural data obtained from the crystal are invaluable in refining the framework for improved accuracy in the design of self-associating systems .
The outcomes of the optimization and simulation ( Stages One and Two ) are tabulated in Tables 1–3 ( full results provided in Table S1 ) . The Sequence Selection table shows that there is a high frequency of double aromatic residues ( Trp , Tyr ) present in the top ten sequences exhibiting the lowest potential energies ( Pot . E ) , whereas there is a high frequency of Met and Ile being present in the last ten tripeptides with the highest potential energies ( Table S2 ) . Stage One calculates the pairwise interaction energies between residues . A fully extended polypeptide chain would result in side-chains of adjacent residues being on opposite planes of the polypeptide backbone . The fact that double aromatic residue sequences have been calculated to possess the lowest potential energies suggests that the backbones of these tripeptides are twisted to promote pairwise interactions between residues . Aromatic residues are known to associate via π-π/CH-π stacking , a prominent example being diphenylalanine [30]–[32] , so the high ranking of double aromatic residue sequences ( lowest potential energies ) enhances confidence in Stage One results . The high frequency of linear aliphatic residues ( Met , Ile ) in the sequences of highest potential energies reflects that van der Waals interactions between adjacent aliphatic side-chains of Met/Ile are weak compared to the aromatic residues . In order to improve confidence in the sequences to be selected for experimental validation , the full set of sequences was screened by Fold Specificity ( FSpec ) and ranked again . In turn , the top twenty sequences ranked by Fold Specificity were also assessed by the Approximate Association Affinity metric , K*association . This double-ranked set of peptide sequences is shown as “Run 1” in Table 2 . In order to separately assess the capabilities of the newly developed metric , 109 of the 128 tripeptide candidates were also directly assessed by K*association . 19 peptides were excluded since their Sequence Selection and/or Fold Specificity rankings were among the lowest ranking and thus did not warrant re-evaluation . The set of top ten sequences using this metric is given as “Run 2” in Table 3 . Unlike for Sequence Selection in which sequences with double aromatic residues dominate the top of the rank , no outstanding trends were observed with regard to the residues of the top-ranked sequences for either Fold Specificity or Approximate Association Affinity , despite the expectation that sequences with aromatic residues might exhibit higher association affinity . This illustrates the ability of the Approximate Association Affinity metric to discern tripeptides that are strong candidates for association , but would have otherwise been difficult to identify through rational design . The top-ranked tripeptide in both Runs 1 and 2 , Ac-LVE , was selected for validation . Compared to Ac-LVE , Ac-YLD has similar Pot . E and FSpec , but different K*association , so it was also selected . Similarly , Ac-LLE and Ac-YYD were selected because they have similar FSpec and K*association , but different Pot . E . The Ac-LVE/Ac-YLD and Ac-LLE/Ac-YYD pairings might allow the respective effects of K*association and Pot . E on self-association outcomes to be discerned . Lastly , Ac-MYD and Ac-VIE were selected as they have similar Pot . E to Ac-IVD . This allows the effects of FSpec and K*association on hydrogelation to be assessed . Thus , the tripeptides chosen for experimental validation were Ac-IVD , Ac-LVE , Ac-YYD , Ac-LLE , Ac-YLD , Ac-MYD , and Ac-VIE ( Table 4 ) . It should be noted that the Fold Specificity and Approximate Association Affinity are used strictly as metrics for selecting which peptides should be experimentally tested . We are not attempting to compare the calculated values to exact , experimental Fold Specificity or Association Affinity values . Rather , we aim to produce metrics capable of ranking a set of peptides to increase the probability that the top ranked peptides are positive hits , in this case , self-associating peptides . For this reason , it is of little concern whether the properties of the produced peptides match exactly the ranking shown in the tables . The inter-peptide interactions that are observed in the simulations of favorably self-associating sequences are predicted to have a higher tendency to self-associate ( form hydrogels or crystals ) and this forms the hypothesis being tested in this work . In the computational calculations , there is currently no metric that can distinguish whether the peptides could potentially form crystals or hydrogels . The six high-ranking tripeptides that were chosen to be evaluated based on their predicted abilities to self-associate can be divided into two classes: ( 1 ) the aliphatic class of Ac-LVE , Ac-VIE , Ac-LLE and ( 2 ) the aromatic class of Ac-MYD , Ac-YLD , and Ac-YYD . The ability of the tripeptides to associate was assessed across a concentration range from 5 mg/mL to the upper limit of 40 mg/mL , in steps of 5 mg/mL . Such a concentration series enables one to compare the association properties of the evaluated tripeptides at 20 mg/mL ( concentration at which the simulations were run ) , as well as bracket the concentration in which there is a change in the association state of the tripeptide . Of the three aliphatic tripeptides , the top-ranked sequence , Ac-LVE , was able to form a gelatinous precipitate between 5 and 10 mg/mL . This precipitation persisted up to 30 mg/mL , with hydrogelation of Ac-LVE observed at 35 and 40 mg/mL . The second-ranked sequence , Ac-VIE was able to form a hydrogel at 5 mg/mL over 48 h; at 10 mg/mL , hydrogelation proceeded within 10 min ( Figure 1A ) . The third-ranked sequence , Ac-LLE , formed a clear solution up to 40 mg/mL , even after standing for two weeks . This indicates that either there is no self-association , or that any association formed by Ac-LLE is still soluble in water . Of the three aromatic tripeptides , the top-ranked sequence , Ac-MYD , was able to form a hydrogel at 5 mg/mL over 24 h; at 10 mg/mL , hydrogelation proceeded within 1 min ( Figure 1A ) . The second-ranked sequence , Ac-YLD , spontaneously crystallized in water , even at the lowest concentration of 5 mg/mL , to furnish large crystals of diffraction quality under ambient conditions ( Figure 1B ) . This indicates that the self-association of Ac-YLD proceeded in an orderly manner to produce the well-defined packing of a crystal . The third-ranked sequence , Ac-YYD , was readily soluble in water , but over time , a small amount of gelatinous precipitate was observed . The amount of gelatinous precipitate scales approximately with concentration up to 40 mg/mL . These observations indicate that the propensity of Ac-YYD to aggregate and entrap water is low . The viscoelasticity of the hydrogels formed from Ac-VIE and Ac-MYD were assessed experimentally at 20 mM . Ac-MYD formed the stiffer hydrogel with a storage modulus ( G′ ) of 20 kPa compared to Ac-VIE ( G′ = 8 kPa ) ( Figure 2 ) . The loss modulus graph also shows that Ac-MYD possessed the larger loss modulus . The loss modulus is a measure of the viscosity of the system , so a substrate with large loss modulus would be very viscous , and less likely to “slip” . Indeed , while the hydrogels of Ac-VIE ( G″ = 1 kPa ) collapsed within two days , the hydrogel of Ac-MYD ( G″ = 9 kPa ) was able to maintain its physical form over more than 10 months ( Figure 2 inset: note the hydrogel suspended on the wall ) . The storage modulus values are comparable to those previously reported for the template tripeptide , Ac-IVD [74] . Control experiments were performed to illustrate the ability of the procedure to compare the relative self-association of analogous tripeptides . Calculations of K*association for analogous tripeptides of Ac-MYD and Ac-VIE , based on shuffling the amino acid residues in the tripeptide sequence , were performed . The calculations show that the optimal position of the polar headgroup is at the C-terminal position , which was previously proposed by Hauser et al . [11] . Four tripeptides ( Ac-YMD , Ac-DMY , Ac-IVE , Ac-EVI ) were chosen from the shuffled sequences and assessed experimentally . As Table 5 shows , the shuffled sequences of Ac-MYD , i . e . Ac-YMD and Ac-DMY , formed clear solutions with no signs of self-association , in agreement with the computed lower K*association values ( Table 5 ) . While the shuffled sequences of Ac-VIE ( i . e . Ac-IVE and Ac-EVI ) precipitated with fibrillar nanostructures , hydrogels were not formed ( Table 5 and Figure S1 ) . This could be related to the ability of the de novo protein design method to predict differently for aliphatic and aromatic tripeptides . It should be noted that the peptides Ac-EVI and Ac-DMY are the only two cases where the self-association motif detailed in Hauser et al . [11] is not incorporated in a tested peptide . The fact that both such peptides are predicted and experimentally validated to not form self-associating structures supports the use of the motif in this and future studies .
When evaluating a newly developed multimeric de novo peptide design framework that relies on several validation stages , it is important to be able to critically assess each stage separately . The experimental results aim to confirm/disconfirm the predictions that the proposed computational framework makes , thus providing an essential test of the approach . Run 1 utilizes Sequence Selection , Fold Specificity and Approximate Association Affinity to select sequences for experimental validation , whereas Run 2 utilizes only Sequence Selection and Approximate Association Affinity . In order to utilize the framework for reliable prediction of self-associating peptides , it is pertinent to understand the properties that each of Sequence Selection ( Pot . E ) , Fold Specificity ( FSpec ) , and Approximate Association Affinity ( K*association ) may influence . The potential energy used in the Sequence Selection stage , Pot . E , which measures the pairwise interaction energies of residues within the tripeptide , may be indirectly related to the extent to which the tripeptide interacts with the solvent . For instance , if the residues of the tripeptides interact in a highly favorable manner with each other ( large negative Pot . E ) they may correspondingly interact to a lower extent with the solvent . The converse would also be true . Such substrate interaction with the solvent is known to critically determine the nano-/microstructural form adopted by the substrate . The tripeptides can be grouped into three potential energy classes: low ( Ac-LLE; Pot . E = −0 . 0618 ) , medium ( Ac-LVE , Ac-YLD , Ac-YYD; Pot . E = −0 . 0324 , −0 . 0340 , −0 . 036 , respectively ) , and high ( Ac-MYD , Ac-VIE , Ac-IVD; Pot . E = −0 . 0151 , −0 . 0173 , −0 . 0153 , respectively ) . Ac-LLE ( FSpec = 3 . 54 , K*association = 4 . 31×10−64 ) and Ac-YYD ( FSpec = 3 . 89 , K*association = 2 . 65×10−70 ) have similar FSpec and K*association , so the effect of Pot . E on their self-association can be gleaned . With the lower Pot . E , Ac-LLE can interact to a lower extent with water , which may account for the formation of bead-like microstructures . With a higher Pot . E , Ac-YYD can interact to a greater extent with water , which accounts for its high water solubility . The high Pot . E of Ac-MYD/Ac-VIE/Ac-IVD suggests they can interact to the relatively highest extent with water , which accounts for their ability to entrap water in forming hydrogels . FSpec , which is derived from an ensemble of 500 models with varying backbone conformations , can be construed as sampling conformations that are amenable to self-associating into nano- and microstructures . Indeed , the chosen tripeptides , which all have FSpec more than one , are capable of self-associating into either fibrillar structures ( Ac-LVE , Ac-YYD , Ac-MYD , Ac-VIE ) , crystals ( Ac-YLD ) , or bead-like microstructures ( Ac-LLE ) to varying extents . This illustrates the capability of the new Fold Specificity metric for multimeric systems . The Approximate Association Affinity ( K*association ) reflects the affinity of the tripeptide to self-associate into multimeric structures . By comparing Ac-LVE ( Pot . E = −0 . 0324 , FSpec = 6 . 09 ) and Ac-YLD ( Pot . E = −0 . 0340 , FSpec = 5 . 18 ) , which have similar Pot . E and FSpec , the effect of K*association on self-association can be assessed . With a greater K*association , Ac-LVE ( 1 . 66×10−3 ) has a higher tendency to self-associate than Ac-YLD ( 4 . 97×10−50 ) . The higher tendency of Ac-LVE to associate might pre-dispose it to form disorderly aggregates whereas the lower tendency of Ac-YLD to associate could allow it to pack orderly and form crystals . The effect of K*association is also borne out by an inspection of the K*association of the seven tripeptides: Ac-YYD , which has the smallest K*association ( 2 . 65×10−70 ) relative to the other six tripeptides , certainly exhibited the lowest affinity to self-associate . Given that Ac-YYD possesses the highest aromatic content of the seven tripeptides , and that aromatic residues are known to self-associate readily via either π-π or CH-π stacking , it is surprising that Ac-YYD would have the lowest tendency to self-associate . Additionally , the negative controls for Ac-VIE and Ac-MYD ( i . e . , Ac-IVE , Ac-EVI , Ac-YMD , and Ac-DMY ) presented in the results demonstrate how the self-association properties of peptides with similar amino acid content can be adequately predicted by the calculated Approximate Association Affinity . These two examples aptly illustrate the capability of the new Approximate Association Affinity metric presented here . However , it would be remiss to consider that Pot . E , FSpec , and K*association independently impact on the self-association outcome of the tripeptides . The group of Ac-MYD/Ac-VIE/Ac-IVD ( Pot . E = −0 . 0151 , −0 . 0173 , −0 . 0153 , respectively ) provides a case in point . Ac-MYD ( K*association = 3 . 05×10−15 ) was observed to possess a higher tendency to gel than Ac-IVD ( K*association = 4 . 87×10−32 ) , and this could be related to the larger K*association of the former . However , although Ac-VIE ( K*association = 5 . 39×10−64 ) has a smaller K*association than Ac-IVD , it was also observed to gel faster than Ac-IVD . The larger FSpec of Ac-VIE ( FSpec = 2 . 69 ) compared to Ac-IVD ( FSpec = 1 . 00 ) suggests that Ac-VIE may adopt conformations that are more amenable to self-association than Ac-IVD , leading to faster gelling . These considerations illustrate how the interplay between FSpec and K*association influences the self-association outcome . Naturally , it can be expected that Pot . E would also influence self-association outcome although this is not exemplified in this case . These results demonstrate that both the filtered ( Run 1 ) and unfiltered ( Run 2 ) stages produced experimentally validated tripeptide sequences . With an interpretation of Pot . E , FSpec , and K*association , the effects of point mutations in ( Ac-LVE↔Ac-LLE ) and ( Ac-MYD↔Ac-YYD↔Ac-YLD ) might be assessed . In all four cases , all three metrics change drastically upon the point mutations . As our results indicate , switching the amino acid from Val to Leu in ( Ac-LVE→Ac-LLE ) caused the tripeptide to convert from fibrillar structures to bead-like microstructures . Switching the amino acid from the aliphatic methionine ( Ac-MYD ) to the aromatic tyrosine ( Ac-YYD ) abolished hydrogelating ability of the tripeptide . This is unlike the aliphatic-to-aromatic residue switch of the amyloid-forming fragment of the human islet polypeptide , in which changing the residue from alanine ( NAGAIL ) to the native phenylalanine ( NFGAIL ) led to a gain in amyloid-forming ability [76] . Conversely , switching the amino acid residue from the aromatic tyrosine ( Ac-YYD ) to the aliphatic leucine ( Ac-YLD ) led to the facile crystallization of the tripeptide . It is remarkable that such apparently small changes can result in major effects on Pot . E , FSpec , K*association , and physical properties of the designed peptides . It is tempting to suggest that these changes affect the multimeric structures of the tripeptides , which in turn affect the interaction of the multimeric structures with water [77] . There could be two reasons for the change observed in ( Ac-YLD→Ac-YYD ) : ( 1 ) the ( 4-phenol ) methylenyl side-chain of Tyr2 in Ac-YYD would hinder the tight packing of the tripeptide and ( 2 ) hydrophobic interactions among the ( 2-methyl ) propyl side-chain of Leu2 in Ac-YLD facilitate the lateral packing of Ac-YLD . Such lateral association of aliphatic side-chains has been noted to be important in the self-assembly of β-hairpin structures that form hydrogels [78] . From the crystal structures of diphenylalanine [79] , [80] , it can be observed that both intramolecular CH-π interactions [81] and intermolecular π-π stacking [82] , [83] are involved in the formation of the nanotubular structure of diphenylalanine . It has often been considered that aromatic groups play a critical role in the key interactions that drive peptide self-assembly , however the extent to which this is true is still unknown [12] . Analysis of the crystal structure of Ac-YLD in comparison to known crystal structures of small self-associating peptides allows for detailed analysis of the interactions that are important for self-association , and more specifically , those interactions that lead to the formation of ordered crystals . Crystal structures of small , self-associating peptides are rare in the PDB . A total of 96 structures in the PDB are classified as “Protein Fibril” . Of these structures , many have characteristics that make it difficult to compare to the crystal structure of Ac-YLD , such as the presence of modified amino acids , peptide lengths greater than 20 amino acids , presence of stabilizing small molecules , and elucidation by NMR rather than crystallography . Removing structures that contain these characteristics , we are left with 35 PDB structures of associating peptides ( Table S3 ) [9] , [10] , [84]–[90] . Through analysis of these structures we can identify a consistent motif for crystal stabilization that is also present in the newly determined crystal structure of Ac-YLD . A clear pattern of alternating hydrophobic zipper-like regions and hydrophilic regions stabilized through immobilized water molecules can be found throughout the crystal structure of Ac-YLD ( Figure 6 ) . Figure 7A–D provides examples of peptide fibril crystals showing similar patterns , despite their difference in peptide length , sequence , associating properties , and backbone orientation ( parallel or antiparallel β-sheet ) . This suggests that sequences that are amenable to forming such patterns may have a higher tendency for crystal formation . Additionally , the importance of the immobilized water molecule in such peptidic crystals points to the possibility that the inclusion of explicit water molecules in the approximate association energy simulations could improve the prediction of whether a peptide will self-associate into a hydrogel or crystal structure . The simulation trajectory of Ac-YLD was compared to the crystal structure of Ac-YLD using VMD [91] . Specifically , key intra- and inter-chain atom distances present in the crystal structure were compared with those sampled in the simulation trajectory . In Figure 8A , one periodic cell consisting of four peptides was extracted from the crystal structure of Ac-YLD . The intra-chain Tyr1∶OH to Asp3∶OD2 distance was 3 . 09 Å , the inter-chain Tyr1∶OH to Tyr1∶N distance was 5 . 02 Å , and the inter-chain Leu2∶CG to Asp3∶CG distance was 4 . 75 Å . Each of these distances were assessed for each of the 5000 frames in the 10 ns trajectory and are shown in Figures 8B , 8C , and 8D , respectively . In the calculation of the inter-chain distances , the corresponding atom on each chain that is closest to the starting chain was used for the calculation . Generally the intra-chain contacts between Tyr1 and Asp3 observed in the crystal structure were not sampled in all of the chains . Conversely , the inter-chain contacts were sampled for a subset of the chains ( Figures 8C and 8D ) . The overall structure at the beginning of the simulation was in a “box-like” configuration with an RMSD to the native of 9 . 35 Å ( Figure 8E ) . Throughout the simulation trajectory the states sampled became closer to the crystal reaching the minimum distance of 5 . 47 Å before finding another stable configuration which the multimeric system remained until the end of the simulation at 7 . 05 Å from the crystal ( Figure 8E ) . The differences in the models sampled and the crystal structure may be due to the initial configuration , or because the models were sampled at a constant temperature . Since we were assessing their strength of interactions , the simulations provide fair comparisons between different sequences of the same length . It is possible that enhanced sampling techniques such as replica-exchange [92] may have allowed for a larger sampling population and should be explored in future work . While computational de novo design methodologies have advanced in their ability to use simulated structures as input models , as was carried out in this study , it is highly preferable to use experimentally determined structures for design . For this reason the elucidation of a crystal structure for Ac-YLD provides an exciting opportunity for future de novo design studies; in particular , for the potential design of inhibiting peptides that may prevent the observed crystal formation . Designs of this category have biomedical implications for the design of inhibitors of amyloid formation . If the formation of such structures can be prevented by the addition of another small peptide , then the interactions important for such inhibition can be determined and exploited for research into the prevention of the onset of degenerative diseases . In this study , we have introduced a new computational de novo peptide design framework for multimeric systems and demonstrated its capability to predict self-associating tripeptides based on the metrics of Pot . E , FSpec , and K*association . Out of the six tripeptides that were computationally predicted to self-associate , all tripeptides formed aggregates of different forms and to different extents , as illustrated by self-association and electron microscopy studies . Two of the six proposed tripeptides , Ac-VIE and Ac-MYD , formed hydrogels at concentrations and on time scales comparable to the template peptide , Ac-IVD . The hydrogel of Ac-MYD showed surprising stability , remaining intact after 10 months , as perhaps might be expected by the computed large association affinity . We were able to use the experimental results to determine how the metrics devised in this work could potentially be used to discriminate between peptides that can and cannot self-associate . Additionally , several negative controls were used to demonstrate the strength of the Approximate Association Affinity metric in distinguishing between closely related peptide sequences that have different self-associating behaviors in nature . These negative controls also support the use of the self-association sequence motif detailed in Hauser et al . [11] as biological constraints in designs of this kind . It is also important to highlight that the aforementioned successful predictions were obtained having as a starting point a simulated initial multimeric structure of IVD and not an experimentally elucidated structure . Importantly , Ac-YLD produced large crystals at ambient conditions and low concentrations . It is often advantageous to use an experimentally elucidated protein structure as the starting template , rather than a simulated multimeric structure in peptide design . Hence , the Ac-YLD crystal structure can serve as a template basis for the design of additional crystal forming peptides or alternately to design peptidic inhibitors of its crystal formation . The use of a crystal structure as a template in future design will improve the accuracy of the first stage and increase the confidence in the designs produced through the subsequent stages . The crystal structure also provided direct observation of the important interactions for the peptide self-association and common packing features between the crystal structure of Ac-YLD and crystal structures of other small , fibril-forming peptides . It was observed that particular intra-molecular interactions were observed in both the MD simulation and the crystal structure , which may point to which interactions are important for crystal formation and can be used to predict which peptides will form crystals . Furthermore , it was determined that a pattern of alternating hydrophobic and water-stabilized hydrophilic regions exists in many small , peptidic crystals , which may indicate that the inclusion of explicit waters in the simulations may improve the accuracy of the simulations used in the calculation of the Approximate Association Affinity . These types of observations can be used as a guide in refining the de novo design framework which currently has no metric to determine whether gelation or crystal formation takes place .
The de novo protein/peptide design framework applicable to multimeric systems consists of two stages [49] , [50] , [52] , [53] , [59] , [70]–[72] . The framework has been developed to handle flexible backbone templates , since experimental structures are not often available for multimeric systems . As such , a flexible backbone template must be created through simulation . In the current design of self-associating tripeptides , MD simulations were performed for this purpose , which produced many snapshots of the plausible multimeric complex . These snapshots were then used to produce a flexible backbone template . The flexible backbone template was subsequently used as the input for the design framework . The first stage of the framework is Sequence Selection , which is based on a global optimization method that minimizes the potential energy of a designed sequence in the flexible template structure . The potential energy used can either be based on an 8-bin Cα-Cα force field or an 8-bin centroid-centroid force field [93] , [94] . A novel aspect of this method is the mathematical connection of residues in the design framework , so that identical chains in the template structure remain identical throughout the design procedure . The optimized sequences are then subjected to a Fold Specificity calculation and screening . Fold Specificity assesses how energetically favorable it is for the designed sequence to adopt the target multimeric structure in comparison to the native sequence . In cases where the native sequence is known to associate , Fold Specificity aims to produce designed sequences that are more energetically favorable in the target multimeric structure than the native sequence . In cases where the native sequence does not assemble , sequences with higher Fold Specificity are considered to have a higher chance of adopting the novel multimeric structure . Finally , a subset of high confidence sequences is subjected to an additional validation step whereby MD simulations are used to dynamically assess the energetics of each designed sequence and its potential to self-associate . In this type of design problem , the binding of several peptides into a multimeric structure has to be considered , which is tackled by the novel Association Affinity metric . All the steps in the design framework , which are presented in a workflow diagram ( Figure 9 ) , are defined in full detail in the following section . This framework is a general methodology that can be applied to a variety of multimeric protein/peptide design problems . PyRosetta [95] was used to generate the initial tripeptide models for the template sequence Ac-IVD through a Monte Carlo ( MC ) conformational search . The function “make_pose_from_sequence” was used in conjunction with the “fa_standard” Rosetta force field [96] . A SmallMover object was constructed with the backbone being allowed to move , with 5 MC perturbations per cycle . The model was subjected to 60 , 000 MC cycles , with the Metropolis criterion determining whether a move was accepted or rejected . This procedure was used to generate 200 low energy decoys for the template . The models were then clustered in Rosetta . The four lowest energy models from the densest cluster were centered at the origin . In CHARMM , the four tripeptides were translated 8 Å in both the y- and z- directions so as to form a square box with the distance from the center of the box to the center of each peptide being 11 . 31 Å . Each tripeptide was rotated randomly . “Hbuild” was used to construct the hydrogen atoms . Periodic boundary conditions , which determine the length and ( consequently ) volume of the box , were applied in CHARMM so that the concentration of the system was 20 mg/mL . The nonbonded cutoffs in CHARMM were set using the following options and values: ctonnb 20 , ctofnb 20 , cutnb 24 , and cutim 24 . Implicit solvent was invoked using the generalized Born with simple switching model [97] . A half smoothing length of 0 . 3 Å , a non-polar surface tension coefficient of 0 . 03 , and a grid spacing of 1 . 5 Å were used . The system was subjected to 2 , 000 steps of steepest descent , followed by 2 , 000 steps of adopted basis-set Newton-Raphson , and finally an additional 2 , 000 steps of steepest descent minimization . The system was heated to 300 K over 10 ps ( stepsize: 0 . 5 fs ) and with harmonic constraints on all heavy atoms with force constant 5 . 0 . The system ( N , V , T ensemble ) was equilibrated for 1 ns ( stepsize: 1 fs ) with a force constant of 1 . 0 on all heavy atoms . The system was subjected to 10 ns of molecular dynamics ( MD ) at 300 K , with SHAKE constraints applied to all bonds involving a hydrogen atom with a tolerance of 1–10 . All simulations were performed using Langevin dynamics with the leapfrog integration scheme . The last 5 ns of the simulation trajectory were processed into pdb files , which were used as the flexible template for design . In accordance to the amphiphilic profile of the template sequence Ac-IVD , the motif [hydrophobic]-[hydrophobic]-[E/D] was applied in the computational method . The hydrophobic residues were allowed to mutate to Leu , Ile , Val , Ala as utilized in the self-assembling hexapeptide Ac-LIVAGD [11] , as well as Met , Phe , Tyr , and Trp . Aromatic moieties have been observed to be important for association due to π-stacking , so aromatic residues were included to expand the scope of tripeptides available for comparison . This resulted in a total of 128 candidate tripeptide sequences , which is a small enough pool that no further biological or mutational constraints were required . Previously developed 8-bin Cα-Cα and centroid-centroid force fields [93] , [94] were available for use as the potential energy function in the Sequence Selection stage . In this study , the 8-bin centroid-centroid force field was employed , since unlike the Cα-Cα force fields , the centroid-centroid force field implicitly includes side-chain directionality in the potential energy calculation . Since the template for tripeptide association was determined through MD simulations , the optimization-based Sequence Selection method was used with a flexible template rather than a rigid backbone template . Two previous methods were developed for flexible template protein design: a Weighted-average method and a Distance Bin method [70]–[72] . The Weighted-average method takes into account each flexible template , such that the potential energy of the system is the average of all the determined templates in the flexible ensemble . The Distance Bin method allows for the optimization method to design not only for the sequence , but also for the optimal interaction distances for each residue-residue interaction . The Distance Bin method represents the most rigorous way of designing with a flexible backbone . For this reason , the Distance Bin Sequence Selection framework for multimeric system design was the chosen framework that was used in this study . Subject to ( 1 ) The model minimizes the summation of pairwise interaction energies , which is the interaction between residue types j and l in residue positions i and k whose distance apart falls in distance bin d . The binary variable equals 1 if residue type j is in residue position i , and 0 otherwise . The binary variable equals 1 if , and only if , and are both equal to 1 , and is 0 otherwise . This represents an exact linearization of the problem . The final binary variable is allowed to equal 1 , if and only if , the distance between positions i and k fall into distance bin d in at least one of the flexible models in the template . In this way , the model is allowed to select one , and only one , distance bin in which two residues can fall , from the set of distance bins observed in the flexible template . A new element of this model is the addition of a mathematical parameter denoted here as to connect design constraints between multiple chains . This parameter is defined as 1 if two design positions ( i and k ) are identical positions in a design system . For example , in the design of a dimer , two identical positions in the two proteins will not allow the model to design for one of the positions without designing for the other position as well . It is also important to emphasize that the objective function is a pairwise interaction potential energy , which takes into account the possible structural flexibility and mutational constraints through a series of linear constraints . The minimization of this objective function aims to improve the stability of the designed sequences in the target structure . This model was used to energetically evaluate all possible tripeptide sequences that fit to the defined design motif . This constituted a total of 128 possible designed sequences , a small enough pool to allow for an exhaustive design search and validation for each sequence . This provides an ideal test system for the new design method as all possible design sequences could be evaluated at each stage before experimental validation . To further validate and analyze the 128 possible tripeptides , a method capable of calculating the Fold Specificity [70] for sequences in multimeric systems was developed . This method uses a constrained annealing simulation in CYANA [98] , [99] to produce a set of initial models . A local AMBER energy minimization using TINKER [100] is then performed on each model to produce a set of 500 final models , along with corresponding AMBER ff94 energy values [101] . Using these AMBER energy values the Fold Specificity value is calculated as follows: ( 2 ) where “New” is the set of models produced for the new sequence , “native” is the set of models produced for the reference sequence Ac-IVD , and is the AMBER energy value calculated for model i . Physically , Fold Specificity assesses how energetically favorable it is for the designed sequences to adopt the target multimeric structure in comparison to the native sequence . The aim is to assess the specificity of the designed sequences for the target structure using a more detailed , atomistic potential energy than in the Sequence Selection stage . Since the metric compares the energy values of the designed sequence directly to those of the native sequence , a “favorable” sequence is one that has a Fold Specificity value greater than 1 . PyRosetta was used to construct initial coordinates of the subset of the 128 tripeptide sequences that ranked highly in the Fold Specificity metric . The MC and MD protocols , which were described previously in the Template Generation section , were used to generate a trajectory for each candidate sequence . The ensemble of models generated through this dynamics run could then be used in the calculation of an Approximate Association Affinity of four tripeptides associating together in the simulations . Since the tripeptides have high flexibility , they do not have a single stable state . Thus , the simulations did not attempt to reproduce the three-dimensional structure of an associate precursor [11] , but to provide an estimation of the affinity of a particular sequence to itself through physics-based intermolecular interactions . For the equilibrium association of two species A and B in solution , the binding affinity can be calculated as: ( 3 ) Lilien et al . [73] proposed an approach for the calculation of approximate binding affinities of protein-ligand complexes . It was based on generating rotamerically-based ensembles of the protein , ligand , and protein-ligand complex . These ensembles were used then to calculate partition functions . This approximate binding affinity was denoted as and defined by Eq . 4: ( 4 ) ( 5 ) where is the partition function of the protein-ligand complex , is the partition function of the free protein , and is the partition function of the free ligand . The partition functions are defined in Eq . 5 , where the sets B , F , and L contain the rotamerically-based conformations of the bound protein-ligand complex , free protein , and free ligand , respectively . The value is the energy of conformation n , R is the gas constant , and T is the temperature . A similar metric can be defined for the association of 4 monomeric peptides into a homogeneous multimeric system . This metric , referred to as the Approximate Association Affinity , is defined as: ( 6 ) This metric was used in conjunction with the Jacobi logarithm [102] to avoid numerical overflow in the calculation . K*association was calculated for each candidate sequence and rank-ordered from the highest ( most favorable spontaneous association ) to the lowest ( Table S1 ) . The simulation of each design was then visually inspected to assess whether the tripeptides associated during the simulation and thus fit to the model . Sequences that did not associate were not considered regardless of the value of the metric . The final set of designed sequences picked for experimental assessment was selected via a combination of Potential Energy , Fold Specificity , and Approximate Association Affinity . The criteria for this selection are provided in more detail in the Results .
|
The self-association of peptides and proteins plays an important role in many serious diseases , such as Alzheimer's disease . A complete understanding of how peptides and proteins self-associate is important in creating therapeutics for such diseases . Additionally , self-associating peptides can be used as templates for bioinspired nanomaterials . With these goals in mind , we have proposed a de novo peptide design methodology capable of producing peptides that self-associate . We have experimentally tested the framework through the design of several self-associating tripeptides . Using the framework we designed six self-associating peptides , including two peptides , Ac-MYD and Ac-VIE , which readily formed hydrogels and one peptide , Ac-YLD , which readily formed a crystal . An X-ray crystallographic study was performed on Ac-YLD to determine its crystal structure . The top-ranked designed sequences were shuffled and computationally and experimentally characterized in order to validate that the approach can differentiate the self-associating of tripeptides , which are derived from the same amino acids . Through the analysis of the experimental results we determine which metrics are most important in the self-association of peptides . Additionally , the crystallographic structure of the tripeptide Ac-YLD provides a structural template for future self-association design experiments .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biotechnology",
"computer",
"applications",
"medicine",
"and",
"health",
"sciences",
"engineering",
"and",
"technology",
"computer-aided",
"design",
"social",
"sciences",
"optimization",
"systems",
"science",
"mathematics",
"alzheimer",
"disease",
"bioengineering",
"computer",
"and",
"information",
"sciences",
"biomaterials",
"dementia",
"mental",
"health",
"and",
"psychiatry",
"economics",
"operations",
"research",
"neurology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences"
] |
2014
|
De Novo Design and Experimental Characterization of Ultrashort Self-Associating Peptides
|
Disease manifestations in neurocysticercosis ( NCC ) are frequently due to inflammation of degenerating Taenia solium brain cysts . Exacerbated inflammation post anthelmintic treatment is associated with leakage of the blood brain barrier ( BBB ) using Evans blue ( EB ) staining . How well EB extravasation into the brain correlates with magnetic resonance imaging ( MRI ) using gadolinium ( Gd ) enhancement as a contrast agent and pericystic inflammation was analyzed in pigs harboring brain cysts of Taenia solium . Three groups of 4 naturally infected pigs were assessed . The first and second groups were treated with both praziquantel plus albendazole and sacrificed two and five days post treatment , respectively . A third untreated group remained untreated . Pigs were injected with EB two hours prior to evaluation by Gd-enhanced T1-MRI , and euthanized . The EB staining for each cyst capsule was scored ( EB grades were 0: 0%; 1: up to 50%; 2: over 50% but less than 100%; 3: 100% ) . Similarly , the Gd enhancement around each cyst was qualitatively and quantitatively scored from the MRI . The extent of pericystic inflammation on histology was scored in increasing severity as IS1 , IS2 , IS3 and IS4 . Grade 3 EB staining and enhancement was only seen in treated capsules . Also , treated groups had higher Gd intensity than the untreated group . Grades of enhancement correlated significantly with Gd enhancement intensity . EB staining was correlated with Gd enhancement intensity and with IS4 in the treated groups . These correlations were stronger in internally located cysts compared to superficial cysts in treated groups . EB staining and Gd enhancement strongly correlate . The intensity of enhancement determined by MRI is a good indication of the degree of inflammation . Similarly , EB staining highly correlates with the degree of inflammation and may be applied to study inflammation in the pig model of NCC .
Neurocysticercosis ( NCC ) , infection of the brain by the larval stage of the parasite Taenia solium , is a common cause of epilepsy in endemic countries [1] . Imaging studies are essential to diagnose NCC . Magnetic resonance imaging ( MRI ) is more sensitive than computed axial tomography ( CT ) , providing better anatomical differentiation , superior visualization of small lesions , edema , vascular enhancement and tissue changes; CT scan is more sensitive for the detection of calcified lesions [2 , 3] . Cysts that have formed in the brain parenchyma can remain quiescent for a period of months to years . At some point , either as a result of the natural course of disease or because of cysticidal treatment , the host mounts a focal inflammatory immune response to the cyst resulting in parasite degeneration [1] . A contrast-enhanced MRI at this time shows a hyperdense ring in the adjacent capsule surrounding the cysticercus , reflecting blood brain barrier ( BBB ) disruption [4] . Previous experiments by our group using a T . solium naturally-infected pig model [5 , 6] measured the degree of capsular inflammation and the extravasation of Evans blue dye ( EB ) into the inflamed capsules of degenerating cysts . The presence of EB capsular staining , a measure of BBB disruption , correlates with the degree of pericystic inflammation [6] . Among imaging techniques , fluid-attenuated inversion recovery ( FLAIR ) and contrast-enhanced T1-MRI using gadolinium as contrast agent are the most useful for diagnostics and follow up , as they provide more detail about the stage of the inflammatory response and the evolution of the damage suffered by the parasite [3] . In the evaluation of human NCC , the degree of enhancement in the contrast T1-MRI protocol is commonly assumed to be a measure of the amount of inflammation present in specific lesions [4] . Since both contrast enhancement and EB staining reflect BBB dysfunction , we assessed whether these markers correlate between themselves and to pericystic brain inflammation as determined on histology . A significant correlation would support the use of enhancement on MRI as a measure of inflammation in swine and by analogy to human NCC .
We compared pericystic EB staining , gadolinium enhancement on MRI , and histological findings following cysticidal treatment of pigs naturally infected with T . solium brain cysts . An individual cyst was considered the unit of analysis for this study . Twelve pigs naturally infected with T . solium cysts from endemic Peruvian highland villages , confirmed by positive tongue examination [7] , were imaged by MRI to confirm brain infection and then randomly divided in three groups of four pigs each . One group remained untreated as a untreated , and the other two groups were treated with albendazole ( Zentel , GlaxoSmithKline , Peru ) at 15 mg/kg daily until sacrifice plus a single day treatment of praziquantel ( Helmiben , Farmindustria , Peru ) at 75 mg/kg divided into three doses of 25 mg/kg every two hours on the first day [7] . Pigs in the treatment groups were sacrificed two ( n = 4 ) and five ( n = 4 ) days after treatment . Immediately before sacrifice , all pigs were anesthetized with an intramuscular injection of a mixture of ketamine ( Ket-A-100 50 mg/kg , Agrovet Market SA , Peru ) and xylazine ( Dormi-Xyl 2mg/kg , Agrovet Market SA , Peru ) through a catheter inserted into the marginal ear artery and kept latent by very slow normal saline drip . Then they were infused through the ear catheter with EB dye first [5] , and after 45 minutes with Gadolinium diethylene triaminopentaacetic acid ( Gd-DTPA ) at 0 . 1 mmol/kg for contrast-enhanced brain MRI . Shortly after the MRI , the pigs were perfused intraaortally for 15 minutes with normal saline solution using a peristaltic pump and euthanized with a lethal IM dose of 120 mg/kg of pentobarbital . The pigs’ brains were retrieved at necropsy and examined macroscopically and by histology . A 2% EB ( Sigma–Aldrich , St . Louis , MO ) solution in normal saline was administered by the EB catheter as reported [5 , 6] and allowed to circulate for 2 hours under additional sodium pentobarbital sedation ( Halatal , AgrovetMarket SA Peru , at 25 mg/kg every 45 min ) . Sedated pigs were placed on a surgical table in a side-lying position . Pre and post-contrast MRI were performed on a 3-Tesla scanner ( Philips Achieva , Best , The Netherlands ) including axial , coronal , and sagittal TFE ( Turbo field echo ) . T1-weighted gradient-echo images were taken with 1–4 mm section thickness , 7 milliseconds ( ms ) of repetition time , 4 ms echo time , 8° flip angle , 270kHz pixel bandwidth and matrix = 256–480 pixels . Immediately upon extraction , brains were placed on dry ice , which helped with the slicing of the samples into 1-cm coronal sections . The anterior and posterior surfaces of each section were photographed . Spearman´s rank correlation was used to evaluate the correlation between the qualitative and quantitative measure of Gd enhancement and the relationship between the grades of EB staining ( ordinal variable ) with quantitative measure of Gd enhancement in each treatment group , and the correlation between each inflammatory stage ( IS1 to IS4 extension expressed in percentage as a four continuous variables ) with the grades of EB staining and quantitative enhancement measure . The non-parametric Wilcoxon-Mann Whitney test was used to compare Gd enhancement intensity and EB grades between the different locations of the cyst in each individual treatment group . All the analyses were performed using the R v3 . 2 . 2 and graphs were created using ggplot package [8] . Differences were considered significant at p<0 . 05 . The study was conducted in accordance with the National Institutes of Health/AALC guidelines , and was reviewed and approved by the Institutional Animal Care and Use Committee Animal Ethics of the Universidad Peruana Cayetano Heredia ( assurance Number: A5146-0 ) .
A total of 328 brain cysts from 12 naturally infected pigs were assessed in this study . The parasitic load per pig brain was widely distributed , ranging from 1 to 45 cysts in the untreated group , from 10 to 29 cysts in PZQ+ABZ 2d and from 4 to 152 cysts in the group allocated to PZQ+ABZ 5d . Pigs in the PZQ+ABZ 5d group had more cysts ( n = 192 ) than did pigs in the untreated group ( n = 73 ) or in the PZQ+ABZ 2d group ( n = 63 ) . The distribution of cysts in both hemispheres was similar in each pig , a total of 165 cysts were found in the right hemispheres of study pigs and 163 cysts were found in the left hemispheres . All brain cysts ( n = 328 ) were used for macroscopic assessment of Evans blue uptake and also for Gd enhancement on MRI ( Table 1 ) . Cysts with clear capsules ( grade 0 ) were seen only in untreated animals; whereas EB grade 3 capsules ( completely blue ) were seen only in animals treated with antiparasitic drugs ( Table 2 ) . Capsules from treated groups had significantly more EB staining than untreated capsules ( PZQ+ABZ 2d versus untreated group , p = 0 . 02 and PZQ+ABZ 5d versus untreated group , p< 0 . 001 ) . However the capsules from PZQ+ABZ 2d had more EB staining than those in the PZQ+ABZ 5d group ( p< 0 . 001 ) . The apparent decrease in EB staining from day 2 to day 5 ( proportional increase in EB grade 2 capsules with fewer grade 3 capsules ) may be due to one pig contributing 128 out of the 144 grade 2 capsules in day 5 ( S1 Table ) . Similarly to the EB staining findings , cyst capsules without Gd enhancement ( grade 0 of enhancement ) were seen only in the untreated group , whereas completely enhanced capsules ( grade 3 ) were seen only in treated animals . Capsules in treated groups had more Gd enhancement than capsules from the untreated group ( p< 0 . 001 ) , however there were marginal differences between both treated groups ( p = 0 . 06 ) ( Table 3 ) . Again , although the data suggest a maximum effect on day 2 , this may have been caused by many grade 2 cysts in the PZQ+ABZ 5d group that came from the same pig ( S2 Table ) . Quantitative measurements of Gd enhancement by cyst were highly correlated to the above-described qualitative assessment ( Table 4 ) . From here on , all analyses shown refer to the quantitative measurement of enhancement , which is less subjective . We analyzed the relationship between the different grades of EB staining and Gd enhancement intensity at the level of individual cysts per treatment group . Gd enhancement intensity had a significant tendency to increase with EB grades in all groups ( Table 5 , Fig 4 ) . Since the inflammatory response can vary according to cyst location , we performed a stratified analysis between deep and superficial cysts ( S3 Table ) . Considering the location of all cysts ( 328 ) , there were 214 superficial and 114 deep cysts . Superficial cysts were more frequent in the PZQ+ABZ 5d group , ( 153 of 192 ) , while deep and superficial cysts had similar frequency in the untreated ( 30 of 73 ) and the PZQ+ABZ 2d groups ( 31 of 63 ) . Within strata of treatment group , EB staining grades were significantly higher in deep cysts than in superficial cysts . These differences were statistically significant in the untreated group ( p<0 . 001 ) , where the majority of deep cysts were grade 2 ( 26/43 , 60% ) whereas superficial cysts were grade 1 ( 19/30 , 64% ) . The same significant difference was observed in PZQ+ABZ 5d group ( p<0 . 001 ) , where the majority of deep cysts were grade 3 ( 30/39 , 77% ) , whereas the superficial cysts were grade 2 ( 135/153 , 88% ) . In PZQ+ABZ 2d group we found a marginal significance ( p = 0 . 052 ) between deep ( 25/32 , 78% were grade 3 ) and superficial cysts ( 17/31 , 55% were grade 2 ) . Gd enhancement intensity ( quantitative measure ) was significantly higher in deep cysts than in superficial cysts in the PZQ+ABZ 2d group ( median 40 . 01 versus 33 . 02 , p<0 . 001 ) and in the PZQ+ABZ 5d group ( 40 . 7 versus 35 . 07 , p<0 . 001 ) , but not in the untreated group ( 34 . 38 versus 33 . 57 , p = 0 . 213 ) . Since the left brain hemispheres were reserved for molecular biology studies , only those cysts located in right brain hemispheres ( n = 165 ) were available for microscopic assessment . Right hemisphere cysts were similar to those in the left hemisphere in terms of grade of EB staining or intensity of Gd enhancement ( S4 Table ) . From these , 113 cysts ( 69% ) from 11 pigs had available slides showing complete parasite structures ( cyst wall and scolex with adjacent brain with immune response ) and were thus used for histopathological studies in order to correlate inflammatory findings with the corresponding EB staining and MRI findings . The distribution of inflammatory stages by cyst and treatment group was consistent with EB staining and Gd enhancement in the entire cyst population , with no extent of IS4 in untreated cysts and few extents of IS1 and IS2 in cysts from treated pigs ( Table 6 ) . When assessed within this population of 113 cysts , we first correlated each inflammatory stage with EB staining and intensity of Gd enhancement in each treatment group . A positive and significant correlation was found only between intensity enhancement and IS4 ( Spearman rank correlation; r = 0 . 363 , p = 0 . 041 ) in the PZQ+ABZ 2d group . There were significant and positive correlations between IS3 and Gd enhancement intensity in the PZQ+ABZ 5d group ( Spearman rank correlation; r = 0 . 358 , p = 0 . 006 ) and between IS4 and EB staining ( Spearman rank correlation; r = 0 . 578 p<0 . 001 ) and intensity enhancement ( Spearman rank correlation; 0 . 50 , p<0 . 001 ) in the PZQ+ABZ 5d . Also , IS4 was significantly higher in deep cysts than in superficial cysts in the PZQ+ABZ 5d group ( median 10 versus 0 , p<0 . 001 ) ( Fig 5 ) .
The study of inflammation due to degeneration of T . solium cysts ( natural or induced by antiparasitic treatment ) in NCC is limited by the lack of suitable animal models . Brain cysts can be found in the natural intermediate host , the pig , and studied by brain imaging ( particularly MRI ) , or histopathology ( macroscopically by assessing EB leakage , or by standard histological or immunohistochemistry techniques ) . MRI has been used in pigs naturally infected with NCC , providing important information on lesion characteristics ( localization , parasite load , localization and stage of lesions , inflammation and perilesional edema , etc ) [10–14] , suggesting its usefulness to evaluate treatment schemes in the pig model . Disruption of BBB is associated with high levels of pro-inflammatory markers and inflammatory cell recruitment into pericystic tissue [6] . The injection of gadolinium as an MRI contrast agent reveals alterations in the integrity of the BBB in various neurological diseases ( multiple sclerosis , strokes , acute ischemic brain injury , brain tumors , encephalomyelitis , etc . ) [15–17] . The same mechanism is observed when the EB stain is used to evaluate alterations in BBB permeability . Our group has applied EB staining to mark BBB disruption in porcine NCC , demonstrating that treatment with praziquantel induces inflammation , cyst damage and BBB leakage at 2 and 5 days ( 48 and 120 h ) ; we also have demonstrated that the BBB leakage is accompanied with expression of pro-inflammatory and immunoregulatory cytokines [5 , 6] . In this study we evaluated the correlation between EB staining and contrast enhancement on MRI in regards to the inflammation and BBB leakage in the pig model of NCC . Our results show that EB extravasation around cysts as a marker of BBB disruption is equivalent to gadolinium enhancement on contrast T1 MRI , reflecting the location and intensity of pericystic inflammation areas . There was a strong correlation between both techniques and the histopathological findings; the location and intensity of gadolinium enhancement , as well as EB staining , reflected the degree and areas of histologically assessed inflammation . Using grades of EB staining and Gd enhancement also allowed for the understanding of the evolution of BBB disruption in the pericystic tissue as the immune response and local inflammation increased in response to anthelmintic treatment . Gd enhancement and EB staining increased markedly at 2 and 5 days post-treatment , which strongly correlated with greater perilesional inflammation . These results are consistent with other studies that evaluated the disruption of the BBB in brain tumors [18] and may explain the increase of symptoms in patients after the onset of antiparasitic treatment [19] . Chronic inflammation by Taenia solium cysts may cause increased angiogenesis in the granuloma around the cysts [20] . These neo-vessels may become more susceptible to vascular leakage after anthelmintic treatment due to the stimulus by pro-inflammatory molecules , thus the correlation of EB staining and Gd enhancement with the severe inflammatory response in NCC . In brain tumors there is a strong association between contrast enhancement and tumor neovascularization [21–23] . The association between brain inflammation and BBB disruption was also described in neurodegenerative diseases as Alzheimer’s disease and multiple sclerosis . In these diseases , the disruption of BBB integrity is associated with extravasation of proinflammatory molecules into the brain that cause damage in the nervous cells [24–25] . The inflammatory response varies according to the location of the parasite in the brain in relation to the parenchyma or the meninges; it is more evident in cysts surrounded by parenchyma [9] . In this study , deep ( parenchymal ) cysticerci showed stronger contrast enhancement in relation to areas where the inflammatory response was severe ( IS3 or IS4 ) , whereas superficial cysts frequently had a thin , weak enhancement signal and moderate or less severe inflammation in histological examinations ( IS2 with some areas of IS3 ) . This suggests enhancement is a marker of more intense and destructive inflammation as seen in other studies [26] . These findings are also consistent with our previous histological findings in the same model [9] . Our study had some limitations . First , the parasite load by pig was very variable , leading us to adjust analyses to account for this variability in the multivariate models . Second , the correlation with histopathology studies only used cysts located in the right hemisphere , although we found no reason to suspect any systematic differences between cysts in the left and right brain hemispheres [27] , as seen in previous studies . Third , our analyses rely in two histological sections of cysts to determine the immune response; the two sections may not be representative of the inflammation around the whole cyst . Despite these limitations , the strong correlation between Gd enhancement on MRI and EB staining suggests that both methods are reliable in the evaluation of perilesional inflammation and BBB disruption in the porcine model of NCC and can be used alternatively or in combination .
|
Neurocysticercosis ( NCC ) is a frequent parasitic infection of the human brain in developing countries . The symptomatology of human NCC after antiparasitic treatment is generally related to inflammation . The presence and degree of enhancement after intravascular injection of the contrast agent gadolinium in magnetic resonance imaging ( MRI ) is commonly considered an evidence of blood brain barrier ( BBB ) leakage . Experimentally , the presence and degree of extravasation of Evans blue ( EB ) after intravascular injection into the tissues of the brain is a direct measure of blood brain barrier leakage . The BBB leakage of gadolinium in neurocysticercosis is commonly used as an indirect measure of inflammation but has never been experimentally proven . Here we evaluated the relationship between contrast T1-MRI , EB staining and histology findings in naturally infected pigs . There was a strong correlation between EB staining , contrast MRI and histopathology findings after antiparasitic treatment . This correlation was stronger when cysts were internally located in the brain than in superficial cysts partly located in the subarachnoid space ( meninges ) . Contrast-enhanced MRI is a non invasive tool used in diagnosis and follow up of NCC patients . This study shows that the use of EB staining allows for the same conclusions as when using MRI post-treatment , and that both techniques correlate with histopathology findings . These results support the use of EB staining to study NCC using the porcine model as well as validate MRI enhancement to assess brain inflammation in patients .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"livestock",
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"pathology",
"and",
"laboratory",
"medicine",
"pig",
"models",
"immunology",
"brain",
"vertebrates",
"cerebral",
"hemispheres",
"animals",
"mammals",
"left",
"hemisphere",
"magnetic",
"resonance",
"imaging",
"animal",
"models",
"histology",
"model",
"organisms",
"signs",
"and",
"symptoms",
"right",
"hemisphere",
"research",
"and",
"analysis",
"methods",
"swine",
"inflammation",
"imaging",
"techniques",
"chemistry",
"gadolinium",
"immune",
"response",
"agriculture",
"chemical",
"elements",
"radiology",
"and",
"imaging",
"diagnostic",
"medicine",
"anatomy",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"amniotes",
"organisms"
] |
2016
|
Perilesional Inflammation in Neurocysticercosis - Relationship Between Contrast-Enhanced Magnetic Resonance Imaging, Evans Blue Staining and Histopathology in the Pig Model
|
The robustness of complex biological processes in the face of environmental and genetic perturbations is a key biological trait . However , while robustness has been extensively studied , little is known regarding the fragility of biological processes . Here , we have examined the susceptibility of DNA replication and repair processes mediated by the proliferating cell nuclear antigen ( PCNA ) . Using protein directed evolution , biochemical , and genetic approaches , we have generated and characterized PCNA mutants with increased affinity for several key partners of the PCNA-partner network . We found that increases in PCNA-partner interaction affinities led to severe in vivo phenotypic defects . Surprisingly , such defects are much more severe than those induced by complete abolishment of the respective interactions . Thus , the subtle and tunable nature of these affinity perturbations produced different phenotypic effects than realized with traditional “on-off” analysis using gene knockouts . Our findings indicate that biological systems can be robust to one set of perturbations yet fragile to others .
Robustness , the ability to maintain performance in the face of environmental and genetic perturbations , is a fundamental trait of biological processes [1]–[5] . Accordingly , many design principles ensuring the robustness of biological processes , such as redundancy , modularity , and feedback mechanisms , have been described [2] , . However , robustness to one class of perturbations can render the same system fragile to other classes of perturbation . The concept of robust yet fragile is a well-known feature in the field of engineering and is one of the most common properties of complex systems [1] , [2] . In the case of complex biological processes , by contrast , very little is known regarding perturbations that result in enhanced sensitivity or fragility of a process . Understanding such perturbations could provide new mechanistic insight into biological processes mediated by complex hub-partner interactions and could elucidate relationships between the robustness and fragility of biological processes . In eukaryotes , DNA replication and repair processes are mediated by the proliferating cell nuclear antigen ( PCNA ) through the recruitment of various DNA-modifying enzymes to the replication fork [8] , including members of different families of DNA polymerases , helicases , exonucleases , and ligases [9]–[11] . PCNA forms a sliding platform to enhance the processivity and catalytic activity of many DNA-modifying enzymes by tethering them to the DNA template . Remarkably , many of the PCNA partners interact with a particular loop on PCNA through a conserved binding motif , suggesting that these partners bind and dissociate sequentially in order to perform their particular function . Switching of partners on the PCNA platform is crucial during different stages of DNA replication and repair , such as lagging strand replication , translesion synthesis ( TLS ) , mismatch repair ( MMR ) , and base excision repair ( BER ) [8] . In recent years , post-translational PCNA modifications have been shown to be an important control mechanism regulating partner switching on PCNA during DNA repair processes [12] , [13] . To investigate the importance of PCNA-partner interactions for DNA replication and repair , previous studies have focused on abolishing these interactions via mutational approaches [14]– . However , due to the functional redundancy exhibited by PCNA partners [17] , abolishing such interactions often results in relatively minor phenotypic defects . Hence , an alternative approach to study the regulation of PCNA-partner interactions during DNA replication and repair involving systematically strengthening specific PCNA-partner interactions is required . Due to the competitive nature of binding to PCNA , strengthening PCNA-partner interactions could result in prolonged PCNA-partner association , thereby hindering the binding of other partners required for the replication and repair processes . Tighter PCNA-partner interactions cannot , therefore , be suppressed by functional redundancy within the pool of network proteins and can thus reveal the importance of accurate regulation of PCNA-partner interactions for the progression of DNA replication . On the system level , this approach could shed new light on the robustness or fragility of DNA replication and repair in the face of such perturbations . In recent years , protein engineering methodologies , including directed protein evolution [18] , [19] , have proven to be highly effective for the generation of proteins with increased binding affinity for target protein partners [20] . In this study , we have examined the robustness or fragility of PCNA-mediated DNA replication and repair processes in the face of perturbations altering PCNA-partner interaction affinities . To do so we have generated and thoroughly characterized a collection of novel PCNA mutants exhibiting higher binding affinities for five different partners taking part in a variety of PCNA-mediated processes , including Pol32 , Rad27 , Rad30 , Msh6 , and Ung1 , participating in lagging strand replication , TLS , MMR , and BER , respectively ( Figure 1 ) . Surprisingly , in vivo analysis of these mutants revealed strong replication defects , indicating the high sensitivity of these processes to increases in PCNA-partner interaction affinities . Moreover , these defects illustrate the importance of a fine balance between different PCNA-partner interaction affinities for the progression of DNA replication and repair . The generation and in vitro and in vivo characterization of the PCNA mutants were performed using a newly developed integrated platform that includes directed evolution , biochemical , and genetic approaches ( Figure 2 ) .
To generate PCNA mutants with enhanced affinity for different partners exhibiting a variety of DNA-modifying activities ( Figure 1 ) , we utilized directed evolution methodologies . Directed evolution experiments are based on the principles of natural Darwinian evolution and consist of two major steps: ( i ) creation of genetic diversity in the target gene in the form of gene libraries and ( ii ) effective selection or screening of those libraries for the desired activity [19] , [21] . Accordingly , we first generated a large PCNA mutant library and displayed this library on the yeast cell surface ( Figure 2A , Step 1 ) [22] . To enrich the PCNA library for mutants with enhanced affinity for the target partner , the displayed PCNA library was incubated with biotinylated peptide derived from the target partner ( see below ) and streptavidin-conjugated allophycocyanin ( APC ) , in addition to a fluorescent antibody against the myc-tagged PCNA . The top fluorescent cell population was selected by fluorescence-activated cell sorting ( FACS; Figure 2A , Step 2 and Figure 2B ) . Next , the enriched libraries were sub-cloned , expressed , and screened in E . coli cells for mutants showing enhanced affinity for the target partners , using an enzyme-linked immunosorbent assay ( ELISA ) . The ELISA experiment for the detection of PCNA-PIP peptide interactions was performed with crude E . coli cell lysates containing the different mutants incubated with biotinylated PIP peptide-coated plates . The amount of bound PCNA was analyzed using antibodies against the 6×histidine-tagged PCNA ( Figure 2A , Step 3 and Figure 2C ) . To further characterize the binding profile of selected PCNA mutants toward an array of partners , yeast two hybrid ( Y2H ) [23] and surface plasmon resonance ( SPR ) [24] assays were used ( Figure 2A , Step 4 ) . Finally , to examine the in vivo activities of the selected PCNA mutants , these were reintroduced as a sole source of PCNA into yeast cells using plasmid shuffle of centromeric plasmids in a strain lacking the chromosomal POL30 gene . The resulting strains were subjected to a variety of DNA replication and repair assays ( Figure 2A , Step 5 ) . PCNA interacts with most of its partners through an inter-domain connecting loop ( IDCL ) that connects the two domains of the PCNA monomer ( Figure 1B ) [25] . Other sites of interactions include residues at the C-terminal and N-terminal regions of PCNA [26]–[29] . Accordingly , the majority of PCNA partners contain a conserved binding motif termed the PIP ( PCNA-interacting protein ) box , located in the N- or C-terminal region of the partner , distinct from its active site ( see Figure S1 for peptide sequences ) [9] , [10] . To generate a large PCNA mutant library , we focused on the diversification of the IDCL region while maintaining the conserved IDCL residues constant due to their specific interactions ( e . g . residues I126 and L128 ) with conserved residues in the PIP region [29] , [30] . We hypothesized that the non-conserved residues may control the specificity of the different PCNA-partner interactions . Therefore , we fully diversified the I121 , A123 , F125 , and E129 non-conserved positions ( Figure S2 ) , yielding a library including 160 , 000 different mutants . To establish a high-throughput screening system for the detection of PCNA binding to PIP peptides derived from the different partners ( Figure 1 ) , we efficiently displayed PCNA capable of binding the target PIP peptides on the outer membrane of yeast cells using YSD methodology ( Figure S3A ) [22] . To confirm that the observed PIP peptide binding was a result of specific PCNA-PIP interactions , we tested the binding of WT PCNA to a mutated Rad30 PIP peptide [15] , as well as the binding of an inactive PCNA mutant ( i . e . , PCNA79 ) [31] to the Rad30 PIP peptide ( Figure S3B–D ) . In both experiments , a dramatic reduction in binding affinity was observed , indicating a specific interaction between the IDCL and the PIP peptide ( Figure S3B–D ) . To enrich the PCNA library for PCNA mutants with enhanced affinity for the target partners ( Figure 1A ) , yeast cells expressing the PCNA library were independently incubated with the five PIP peptides and more than 5×106 cells were analyzed and sorted by FACS ( Figure 2B ) . Three to five iterative rounds of enrichment were performed until a significant enrichment for PCNA mutants with increased affinity for all PIP peptides was obtained ( Figure 3 ) . To identify single PCNA mutants with enhanced binding affinity for target partners , the five FACS-enriched libraries ( Figure 3 ) were sub-cloned into a bacterial plasmid , overexpressed in E . coli , and screened by ELISA ( Figure 2C ) . Using this approach , the crude cell lysates of 10–20 clones from each of the five enriched libraries were screened for binding to a given PIP peptide . The top performing 2–4 mutants from each of the five ELISA screens ( Table 1 ) were then taken for further in vitro and in vivo characterization , as described below . To verify that the selected PCNA mutants showing increased affinity for the PIP peptides also exhibit increased affinity for the full-length partner , a Y2H analysis was performed . For the Y2H assay , we used the YRG2 host strain , shown to be highly efficient in coupling the strength of protein-protein interactions with growth on media lacking histidine [32] . Selected PCNA mutants were characterized for their ability to bind each of 12 main PCNA partners [8] to obtain specificity profiles ( Table 1 , Table S1 , and Figures S4–S5 ) . Such profiling indicated , in most cases , that an increase in binding affinity for the target partner did not result in increased or decreased affinity for the other partners ( Table 1 , Table S1 ) . These results demonstrate the high flexibility of the IDCL in terms of increased binding specificity . However , in some cases a strong trade-off between the bindings of different partners was observed . For example , some PCNA mutants selected for high affinity to Rad27 exhibited reduced affinity for Rad30 and vice versa ( Table 1 ) . To examine whether increased PCNA-partner interaction affinities result in increased PCNA-partner complex formation in vivo in pol30 mutant strains , we analyzed the levels of these complexes extracted from yeast cells . We used strains expressing the pol30 mutants as a sole source of PCNA in the cell under the control of the native POL30 promoter ( see below ) . We immobilized PCNA from the crude yeast extracts onto ELISA plates and analyzed the amount of bound partner , relative to the amount of total PCNA immobilized on the plate ( Figure 4 ) . Using this approach , we successfully identified an increase of 70% and 17% in the amount of PCNA-Rad30 and PCNA-Rad27 complexes extracted from pol30 mutant strains with increased affinity for Rad30 and Rad27 , respectively , relative to the POL30 strain ( Figure 4 ) . However , we did not detect an increase in PCNA-Msh6 or PCNA-Ung1 complex formation , probably due to the relatively minor increase in affinities of these interactions ( 1 . 6- and 3 . 5-fold , respectively , see Table 1 ) and the transient nature of PCNA-partner interactions [33] . To quantify differences in binding affinities of the PCNA mutants relative to the WT , we characterized four PCNA mutants with increased affinity for PIP peptides derived from Rad30 , Pol32 , Msh6 , and Ung1 partners using SPR . These mutants were overexpressed in E . coli , purified by affinity chromatography , and immobilized on an SPR sensor chip for binding characterization [14] . SPR binding analysis enabled direct and sensitive measurement of interaction affinities relative to the Y2H system and indicated an up to 8-fold increase in binding affinity for the different PIP peptides , relative to WT PCNA ( Table 1 ) . Such analysis allowed the detection of a significant decrease in the Rad30 PIP peptide dissociation rate in the Pol30-Rad30E9p mutant relative to the WT , indicating a ∼9-fold increase in the lifetime of the PCNA-Rad30 complex ( see Table 1 ) . Collectively the ELISA , Y2H , and SPR assays validated and quantified the increase in mutant PCNA binding affinities , relative to WT PCNA . To validate that the PCNA mutations have not altered the ability of that mutant to form an intact PCNA structure , we characterized six different PCNA mutants showing increased affinity for different partners . We used gel filtration chromatography analysis of the purified PCNA mutants to examine the molecular mass of the proteins under non-denaturating conditions . Using this approach , we found that all mutants form intact trimers of molecular mass of ∼90 kDa , similar to WT PCNA ( Figure S6 ) . As a control , we analyzed purified PCNA-52 mutant , which was previously shown to be defective in trimer formation [34] , and detected a molecular mass of a monomer of ∼30 kDa ( Figure S6 ) . In addition , we examined the secondary structure content of the PCNA mutants , in comparison to the WT , using Circular Dichroism ( CD ) spectroscopy . This experimental approach allows examining whether the mutations in PCNA led to substantial structural alterations . We found that the CD spectra of the mutants are very similar to the CD spectrum of the WT PCNA ( see 6 ) . Overall , these results indicate that the PCNA mutations did not significantly alter the secondary structure of PCNA or its ability to form trimers . Sequence analysis of the selected mutants revealed the presence of 3–4 mutations in the IDCL region out of the 4 positions randomized in the naïve library ( Table 1 ) . Surprisingly , some of the PCNA mutants identified following Rad27 PIP selection were characterized by a deletion of two amino acids after the first aspartic acid of the IDCL region ( Figure 1B , Table 1 ) [25] . This deletion results in altered pattern of PCNA-partner specificity ( Table 1 ) due to shortening of the exposed and flexible IDCL loop that may result in new conformational diversity . To validate the effect of the deletion , we generated the same deletion on the background of WT PCNA and observed similar binding specificity , relative to the selected deletion mutants ( unpublished data ) . To study the in vivo ability of PCNA mutants to promote DNA replication and repair in yeast , we adopted a plasmid shuffling method to create haploid yeast strains carrying each mutant as the sole source of PCNA ( see Materials and Methods ) . This approach allowed us to identify PCNA mutants leading to cell death , indicating their inability to support essential DNA replication processes ( Figure 5A ) . In addition , many of the mutant strains exhibited high sensitivity to DNA damaging agents , such as hydroxyurea ( HU ) or methyl methanesulfonate ( MMS ) , drugs that cause global replication stress and DNA alkylation , respectively ( Figure 5B ) . Interestingly , the lethality or strong sensitivity displayed by our mutants presenting increased affinity for various partners far exceeds the sensitivity observed for the previously studied pol30-79 mutant , exhibiting a significant decrease in binding affinity for the majority of PCNA partners ( Figure 5B ) [31] . To verify that the replication defects in the mutant strains ( Figure 5 ) are not due to reduced expression levels of the PCNA proteins , we confirmed the expression level of PCNA using western blot analysis and observed similar expression levels of all PCNA mutants , relative to WT PCNA ( Figure S7 ) . Next , we examined the sensitivity of strains containing single deletions of each of the five different partners on a WT PCNA background . We found that these strains exhibit equal or lesser phenotypic defects , relative to the selected PCNA mutant strains with increased affinity for the respective partner ( Figure 5B ) . These results suggest that the cost for increasing PCNA-partner interaction affinity is equal to or much higher than the cost of that partner being absent ( Figure 5B , see Rad30 and Ung1 as prominent examples ) . To further test this idea and validate that the mutations in PCNA do not disrupt any critical PCNA function unrelated to PCNA-partner interactions , we examined whether deletion of different partners on the background of the pol30 mutant strains can suppress the growth sensitivity phenotypes . We first examined whether the rad27 deletion can suppress the lethality of a strain containing the pol30-Rad27L1 or pol30-Rad27L2 mutants ( Figure 6A ) . It was previously shown that rad27 deletion does not cause lethality [35] . Interestingly , rad27 deletion suppressed the lethality of the pol30-Rad27L1 and pol30-Rad27L2 strains probably due to the lack of PCNA-Rad27 complexes in these strains ( Table 1 , Figure 6A ) . This result demonstrates that the strong deleterious effect of PCNA mutants with increased binding affinity for Rad27 could be much higher than the effects of rad27 deletion and firmly correlates with our in vitro and in vivo analyses of the PCNA mutants . We also found that the rad27 deletion can suppress the phenotypes of other pol30 mutant strains ( Figure S8 ) , however such suppression could be due to indirect pathway activation [36] . In addition , we examined whether rad30 deletion , ung1 deletion , and msh6 deletion can suppress the phenotypes of the pol30 mutants with increased affinity to Rad30 , Ung1 , and Msh6 , respectively ( Figure 6B–D ) . We found that indeed such deletions suppressed the growth sensitivity phenotypes of the pol30 mutants with increased affinity for Rad30 or Ung1 and the high spontaneous mutation rate observed in the pol30 mutant with increased affinity to Msh6 ( Figure 6B–D ) . However , since the suppression of the strain phenotypes was not complete , additional factors , such as minor effects of the mutations on other PCNA-partner interactions , may play a partial role in the sensitivity phenotype of the examined pol30 mutants . Collectively the suppression of the phenotypes of the pol30 mutant strains by PCNA partner-deletions suggests that these phenotypes arise mainly as a result of specific enhancement of PCNA-partner interactions and that the pol30 mutants are not defective in any critical function unrelated to PCNA-partner interactions . We also examined whether overexpression of different partners can suppress the growth sensitivity of the pol30 mutants , however no suppression was observed ( Figure S9 ) . It was previously shown that PCNA contains another site of regulation ( i . e . K164 ) that is modulated by ubiquitination or SUMOylation ( Figure 1B ) and that is crucial for the recruitment of TLS polymerases [13] , [37] . To test whether the K164 regulatory site is active in the PCNA mutant strains , we examined the spontaneous mutation rate in these strains ( Table 1 ) . Increase in spontaneous mutation rate can indicate the recruitment of TLS polymerases to PCNA through K164 ubiquitination [13] , [37] . Alternatively , these mutants can indirectly affect the spontaneous mutation rate by reducing DNA replication processivity and causing replication fork stalling , thereby leading to TLS recruitment [37] . These mutants can also indirectly alter Pol-δ or Pol-ε proofreading leading to TLS recruitment [37] . We measured the spontaneous mutation rate using the CAN1 reporter assay by monitoring the ability of the PCNA mutant strains to grow in the presence of canavanine , a toxic analogue of arginine [38] . Interestingly , we observed a significant increase in the mutation rate in most of the PCNA mutant strains , relative to WT PCNA ( Table 1 ) . To examine whether Polζ , the major TLS polymerase [13] , [37] , is recruited in vivo to PCNA mutants showing increased affinity for Rad27 and Pol32 ( pol30-Rad27E6 and pol30-Pol32E5 , Table 1 ) upon K164 ubiquitination , we measured the spontaneous mutation rates of these mutant strains on the background of either mutated rev3 encoding the catalytic subunit of Polζ or mutated rad18 encoding the K164 ubiquitin ligase or K164R mutation ( Figure 7 , Table 2 ) . We observed a dramatic reduction in the mutation rates of all of these strains , indicating that the high mutation rate in the pol30 mutant strains is a result of Polζ recruitment to the PCNA mutants through K164 ubiquitination ( Figure 7 , Table 2 ) . Collectively , these results indicate the functional separation between IDCL and K164 ubiquitination in regulating partner binding to PCNA . Finally , to examine the mutation spectra at the CAN1 locus for pol30 mutant strains with increased affinity to Pol32 or Rad27 , we sequenced the CAN1 gene in individual canavanine resistance clones . We sequenced 37 and 26 CAN1-inactivating mutations isolated from the pol30-Pol32E5 and pol30-Rad27E6 mutants , respectively . We found that pol30-Pol32E5 mutant accumulated a broad range of mutations , including substitutions , frame-shifts , or deletions ( Table S2 ) . Interestingly , we observed that 20% of the mutations were characterized by a specific deletion of 67 nucleotides flanked by a 4 bp direct repeat ( Table S2 ) . This result suggests that polymerase slippage is one of the mutagenic mechanisms leading to the increased mutation rate of the pol30-Pol32E5 mutant strain [39] . In contrast , the pol30-Rad27E6 accumulated a very low frequency of deletions ( 4% ) but a high rate of substitutions and frame-shifts ( Table S2 ) . We also analyzed the CAN1-inactivation mutations accumulated in the pol30-Pol32E5 and pol30-Rad27E6 mutants generated on the background of rev3-deletion or pol30-K164R mutation . We found that these mutants accumulated low levels of substitution with no deletions or frame-shifts ( Table S2 ) , suggesting that the diverse mutation spectra in the characterized pol30 mutant strains is dependent upon REV3 and K164 modification .
In this study , we have generated and examined a novel collection of PCNA mutants with increased binding affinity for several partners , relative to WT PCNA . We have established an integrated approach that allows us ( 1 ) to generate PCNA mutants with increased affinity for different partners , ( 2 ) to perform binding characterization of the mutants for many different partners in order to profile changes in binding specificity , and ( 3 ) to perform detailed in vivo characterization of the mutants for the detection of defects in DNA replication and repair . The generation of PCNA mutants with increased affinity for five different partners revealed the high plasticity of PCNA for increases in partner interaction affinities , implying that the WT IDCL sequence naturally evolved to bind multiple partners with moderate affinity , rather than adopting higher binding affinities for specific partners . This property could be a selectable evolutionary trait designed to maintain the dynamic nature of PCNA-partner interactions and to facilitate partner switching on PCNA . The large number of mutations and the lack of conservation observed in the selected mutants ( Table 1 ) suggests that binding affinity for the PIP region involves diverse contributions from many IDCL residues and that multiple solutions exist for increases in PCNA-partner binding affinities ( Table 1 ) . In vivo analysis of the different mutants revealed severe phenotypic defects , ranging from non-viability to high sensitivity to DNA-damaging agents ( Figure 5 ) . In contrast , abolishment of PCNA-partner interactions by mutating conserved residues in PCNA [31] or in the PIP region of the different partners [14] , [16] or even by deletion of different partners results , in most cases , in relatively minor phenotypic defects ( Figure 5 ) [14] , [16] . Collectively , these results highlight the fragility of DNA replication and repair resulting from increased PCNA-partner interaction affinity , in contrast to the robustness detected in the face of abolishment of the same interactions . Functional redundancy in PCNA partners could be the major mechanism facilitating the robustness of the PCNA-partner interaction network in coping with an abolishment of PCNA-partners interactions [5] . Indeed , several publications have indicated that the exonuclease activity of Polδ can substitute for the 5′flap endonuclease activity of Rad27 in processing Okazaki fragments , thereby preventing genome instability [17] , [40] . What could be the mechanistic basis for the severe phenotypic defects observed in the PCNA mutant strains ? PCNA mutants showing increased binding affinities for different partners may experience prolonged PCNA-partner associations , thus altering partner switching at the PCNA IDCL region . Indeed , SPR analysis of the pol30-Rad30E9p mutant ( Table 1 ) indicates that the increase in the binding affinity of a given partner is due to a decrease in that partner's dissociation rate ( koff rate , see Table 1 ) . The phenotypic defects in the pol30 mutant strains ( Figure 5 ) were detected in strains containing PCNA mutants with relatively minor increases in affinity for different partners . SPR analysis indicated an up to 8-fold increase in binding affinity of PCNA for the different partners , indicating the high impact of such alterations on DNA replication and repair processes in vivo . Such defects indicate that affinity-based competition between different partners is a crucial factor for the regulation of PCNA-partner interactions during different stages of DNA replication and repair . This , moreover , suggests that these are highly dynamic processes that require multiple partner binding and dissociation events [41] . In support of this hypothesis , different stages of DNA replication and repair , such as lagging strand replication , TLS , and BER , require the sequential binding of multiple PIP-containing partners to the IDCL [8] . However , it is important to note that the mutations in the IDCL region of the pol30 mutants leading to increased affinity for various partners can still disrupt other important in vivo functions of PCNA . For example , such mutations can affect the PCNA in vivo localization or loading onto the DNA , thus contributing to the phenotypic defects observed in the pol30 mutant strains . Future work analyzing these in vivo PCNA properties would contribute to the characterization of the PCNA mutants described in this study . In addition , future in vitro assays would allow further analysis of the effects of increasing PCNA-partner interactions on partner switching on PCNA during the different steps of DNA replication and repair . One of the phenotypes of the PCNA mutant strains is a significant increase in the spontaneous mutation rate at the CAN1 gene , indicating the recruitment of TLS polymerases to PCNA ( Table 1 ) . To obtain deeper mechanistic insight into such recruitment , we have shown that this phenotype is suppressed by the deletion of either rev3 [42] or rad18 or by the K164R mutation ( Figure 7 ) [12] . These results indicate that the increase in PCNA binding affinity for Rad27 and Pol32 can trigger the recruitment of Polζ to the replication fork via K164 ubiquitination . Furthermore , our data indicate a functional separation between the IDCL and K164 regulatory sites and suggest that the recruitment of Polζ by K164 modification can provide a back-up mechanism by which to sustain replication in cases of regulation defects involving the IDCL region . Examination of the mutation spectra at the CAN1 locus of CAN1- resistant clones isolated from the pol30-Pol32E5 and pol30-Rad27E6 mutant strains indicated the accumulation of a broad range of mutations and a significant difference in the mutation spectra between these strains ( Table S2 ) . These results indicate that increased PCNA-partner interactions can lead to different mutations , demonstrating different characteristics of each mutant ( Table S2 ) . In addition , the pol30-Rad27E6 accumulated mutations were different from the mutations previously found in the rad27-deleted strain [35] . The mutations in the rad27-deleted strain were characterized by a high frequency of duplications , indicating severe impairment of lagging strand replication [35] . This comparison suggests that pol30-Rad27E6 affects the regulation of lagging strand replication , rather than leading to complete abolishment of Rad27 enzymatic activity . To facilitate the generation and examination of PCNA mutants with increased affinity for several partners , we have developed an integrated platform based on directed evolution and yeast genetic approaches . Currently , the most common manner of employing genetics for studying the robustness of cellular processes addresses the all-or-none effects generated by gene knock-outs ( 5 ) . The integrated approach that we have developed allows for the generation of much more subtle and controlled perturbations to reveal new properties of the DNA replication system . This approach does not directly affect the expression level of the proteins or the catalytic activities of the different partners , thus allowing for dissection of the effects of subtle alterations in PCNA-partner binding affinities on the replication process . In future studies , the mutants generated in this study could prove useful in efforts aimed at obtaining mechanistic insight into PCNA-partner binding and dissociation events during DNA replication and repair processes . In summary , using DNA replication and repair as a model system , we have shown that biological processes can be highly robust to one set of perturbations yet at the same time be highly fragile to completely different perturbations . Our data thus provide a more balanced view on the robustness of biological processes and reveal that similar to many man-made complex systems , these processes possess both properties of robustness and fragility . Finally , the approach developed in this study , which allows for the generation of a variety of minor perturbations in a protein-protein interaction network , can be applied to study the molecular basis , mechanism , and fragility of other networks promoting different biological processes , including signal transduction and gene transcription .
For E . coli expression , WT PCNA and the different mutants were cloned into plasmid pET28 ( Novagen ) using the NdeI and XhoI sites to yield a 6×Histidine-tagged version of the protein . For YSD , WT PCNA was cloned into plasmid pCTCON [22] using the NheI and BamHI sites to generate plasmid pCTCON-PCNA . For in vivo testing of PCNA mutants , a 200 bp PCNA-promoter region and a 300 bp PCNA-terminator region were amplified from genomic DNA using the fr-pro and rev-pro and the fr-ter and rev-ter primers for the promoter and terminator regions , respectively ( Table S3 ) . These fragments were cloned into the pRS315 and pRS316 centromeric plasmids using NotI and SpeI or HindIII and XhoI sites , respectively , to generate the pRS315-proterm and pRS316-proterm plasmids . WT and mutant PCNA genes were amplified using fr-pRS/PCNA and rev-pRS/PCNA primers and cloned into pRS315-proterm and pRS316-proterm plasmids by homologous recombination . PCNA partners were GFP-tagged at their natural locus using a GFP-cassette as previously described [43] . For the Y2H assay , the pAD-GAL4 and pBD-GAL4 plasmids ( Stratagene ) were used to clone WT or PCNA mutants and the various partners , respectively ( see Table 1 , Tables S1 and Table S3 for a list of partners and oligonucleotides , respectively ) . PCNA was displayed on the yeast cell surface of EBY100 strain cells ( see [22] for genotype ) and analyzed by flow cytometry , essentially as described [22] . Briefly , EBY100 transformed with plasmid pCTCON-PCNA were grown in SDCAA media to logarithmic phase and 2×106 of cells were washed , resuspended in SGCAA induction media , and grown at 20°C with shaking for an additional 18 h . Induced cells ( 1×106 ) were collected by centrifugation , washed with PBSF ( PBS+ 1 g/L BSA ) , and incubated for 1 h at 25°C with mouse α-Myc antibodies ( Santa Cruz Biotechnology , 1 µl/50 µl PBSF ) and 100–400 µM of biotinylated PIP peptide ( Peptron , see Figure S1 for peptide sequences ) . Subsequently , cells were washed and incubated with FITC-conjugated α-mouse IgG ( Sigma , 1 µl/50 µl PBSF ) and APC-conjugated streptavidin ( Jackson Immunoresearch , 1 µl/50 µl PBSF ) for an additional hour on ice , with frequent mixing . The labeled cells were washed , resuspended with PBSF , and analyzed by flow cytometry ( FACS Calibur , BD ) . Positions I121 , A123 , F125 , and E129 of the IDCL region were fully randomized by two fragments overlapping PCR using plasmid pCTCON-PCNA as template . The two PCNA gene fragments were amplified using two sets of primers ( fr-Lib1 and rev-NNS; rev-Lib1 and fr-NNS , Table S3 ) , assembled , and further amplified using nested primers . A naïve library was generated by in vivo recombination to obtain ∼1×106 colonies oversampling the PCNA library diversity . The naïve library was induced and labeled with different PIP peptides , as described above . EBY100 cells ( 1×107 ) displaying the PCNA library were labeled , analyzed , and sorted using a FACS ( Vantage , BD ) . Three to five iterative rounds of enrichment were performed . In each round , multiple “positive” events ( 3–5×104 ) , corresponding to cells found within the top 1%–2% of the green and red fluorescence intensity area , were collected into growth media and plated on agar plates for a new round of enrichment . For initial sorting of the naïve library , a sort gate of the top 5% of fluorescent cells was used . To increase the stringency of selection , a decreased peptide concentration was used in each subsequent round . Selection rounds were continued until no further enrichment was obtained . A pool of plasmids from the last round of FACS enrichment was PCR-amplified using the primers fr-pET/PCNA and rev-pET/PCNA ( Table S3 ) and cloned into plasmid pET28 . Single E . coli BL21 cells , transformed with the resulting plasmids , were inoculated into 10 ml LB media containing 50 µg/ml kanamycin , grown to OD600 0 . 6 , and induced with 1 mM of IPTG ( Calbiochem ) for 5 h at 30°C . The cells were then harvested , lysed in PBS supplemented with 0 . 1% Triton and 200 µg/ml lysozyme , centrifuged , and the cleared supernatant was collected . Total protein concentration of the different mutants was determined using a BCA protein assay kit ( Thermo Scientific ) and analyzed by SDS-PAGE to verify the similarity of PCNA expression levels . ELISA plates ( Griener Microlon 96W ) were coated with 0 . 2 µg/ml streptavidin ( Pierce ) and 0 . 1 µg/ml of biotinylated PIP peptides , as described [44] . Following peptide coating , the plates were incubated with the cleared lysate generated above at appropriate dilutions and shaken at 25°C for 1 h . Plates were then washed with PBS supplemented with 0 . 05% Tween-80 ( PBST ) and each well was incubated with mouse α-6×His-tag antibodies ( Santa-Cruz Biotechnology ) diluted by a factor of 1∶2000 and then with secondary HRP-conjugated goat α-mouse antibodies ( Jackson , 1∶5000 ) . The HRP chromogenic TMB substrate solution ( Dako ) was added and the reaction was stopped by the addition of 100 µL of 1 M sulfuric acid and recorded at 450 nm using a Tecan Infinite M200 plate reader . E . coli BL21 cells were induced and lysed as above in a volume of 0 . 5 L with minor modifications . Briefly , following centrifugation the cell pellet was sonicated in 20 ml of lysis buffer , centrifuged , and the cleared supernatant was loaded on a pre-equilibrated column containing 2 mL Ni-NTA resin ( Qiagen ) . The columns containing the lysates were gently shaken by inversion for 30 min at 25°C . The resin was then washed with 30 ml of wash buffer and PCNA was eluted in 1 ml fractions upon addition of elution buffer . Fractions containing PCNA were pooled and dialyzed against storage buffer . Protein concentration was determined with a BCA protein assay kit ( Pierce ) and analyzed by SDS-PAGE . The protein solutions were stored in 1 ml aliquots of 2 mg/ml at −20°C . Lysis , wash , elution , and storage buffers were derived from the activity buffer based on 300 mM NaCl , 50 mM Tris-HCl , pH 8 , and supplemented with imidazole , according to the manufacturer's recommendations . WT and mutant PCNA were purified as described above . Gel filtration chromatography was performed on a Superdex 200 10/300 GL column ( GE Healthcare ) using the ÄKTApurifier FPLC system . All proteins were run in activity buffer at monomer concentrations of 3–8 µM at which WT PCNA is a trimer and Pol30-52 is a monomer [34] . CD spectra were obtained for WT PCNA and six representative mutants using the Jasco J-810 CD Spectropolarimeter . All measurements were performed at room temperature in activity buffer . Data were obtained for the wavelength range of 204–260 nm and normalized to protein concentration to obtain molar ellipticity . Protein interaction assays were carried out using the ProteOn XPR36 ( Bio-Rad ) instrument . WT or PCNA mutants ( 0 . 4 to 5 . 2 fmol ) were immobilized on the surface of a GLM sensor chip by a carbodiimide-activated succinimide-coupling method , as specified by the manufacturer . All SPR experiments were performed by flowing 150 µl of the target peptide at a flow rate of 30 µl/min onto the PCNA-bound chip . Different concentrations ( 5–5 , 000 nM ) of PIP peptides ( Table 1 ) were injected over the PCNA chip , and binding parameters were determined using ProteOn XPR36 software ( Bio-Rad ) . The ligand ( PCNA ) and analyte ( peptide ) buffers were PBST and 150 mM NaCl , 1 mM EDTA , 0 . 01% Tween-80 , 30 mM Hepes , pH 7 . 5 , respectively . Y2H analysis was performed using the Yeast Two Hybrid Phagemid vector kit ( Stratagene ) , following the manufacturer's instructions . The pAD-PCNA-WT/mutant plasmids were used as bait while plasmids encoding 12 different PCNA partners ( Table 1 and Table S1 ) were used as prey . The YRG2 host strain ( Stratagene ) was cotransformed with pAD-PCNA WT/mutant and pBD-partner plasmids in all possible combinations using the LiAC method . Single transformants were grown in liquid SC-Leu-Trp to O . D600 10 , washed twice with DDW , and diluted to an initial OD600 of 0 . 3 . A series of 10-fold serial dilutions was then spotted onto selective SC-Leu-Trp-His plates and incubated at 30°C for 3 d . For quantitative Y2H , the generation time of the indicated mutants and their respective partners were calculated from their growth curves in liquid SC-Leu-Trp-His media . Cells were grown overnight in SC-Leu-Trp , washed twice with ddH2O , and diluted by a factor of 1∶50 into 10 ml of pre-warmed SC-Leu-Trp-His . O . D600 measurements of the cultures were taken at the indicated time; the generation time ( τ ) was calculated from the growth curves according to the equation ODt = OD0×2t/τ . The generation time calculated for each culture is an average of at least 3 independent experiments . Novel haploids containing PCNA mutants were generated using the plasmid shuffling method . Briefly , a pol30::KanMX magic marker heterozygote diploid strain BY4743 ( Open Biosystem ) was transformed with plasmid pRS316-POL30 . Following dissection of the diploid , a haploid containing the CAN+ gene and plasmid pRS316-POL30 as a sole source of PCNA was generated . This host strain was transformed with selected pRS315-pol30 mutants , plated on SC-Ura-Leu plates , followed by replica plating to SC-Leu+5FOA plates . Haploids , containing PCNA mutants as a sole source , were further verified by plating on either SC-Leu or SC-Ura plates . For testing selected PCNA mutants , a haploid containing rad27::HYG , rad30::HYG , pol32::HYG , msh6::HYG , or ung1::HYG were generated using plasmid pAG32 by conventional genetic approaches . Growth of the PCNA mutant strains in the presence of 120 mM HU ( Toronto Research Chemicals ) or 0 . 02% MMS ( Sigma ) was performed as described [45] . To examine the effects of partner overexpression on pol30 mutant strains , these strains were transformed with various plasmids containing GST-tagged PCNA partner encoding genes under the control of a GAL1/10 inducible promoter , as previously described [46] . Overnight cultures were plated in serial dilutions on SC-Ura containing either glucose or galactose with or without DNA-damaging agents , as described above . The mutation rates for the different pol30 mutant strains described in this study were determined by fluctuation test analysis using the Lea and Coulson method [47] , [48] . The different strains were plated as single colonies on SC-Leu plates and allowed to grow for 3 d at 30°C . At least 25 single colonies from each strain were excised from the plate and resuspended in 1 ml of sterile water to an O . D600 of 0 . 7 . Appropriate dilutions of the cells were then plated on SC-Leu and SC-Leu-Arg+canavanine ( 60 mg/ml ) to obtain the number of viable cells ( Nt ) and the number of canavanine-resistant cells ( r ) , respectively . Using the Lea and Coulson method [47] , the number of mutations ( m ) per colony was derived from the number of canavanine resistant-colonies ( r ) across parallel cultures , using the following equation: m/r-ln ( m ) −1 . 24 = 0 . The m values were then used to calculate the mutation rate , M , using the following equation: M = m/Nt , where Nt is the average number of viable cells per plating . The different M values were sorted to obtain the median . The low and high values for the 95% confidence interval for each rate were obtained using the confidence interval median test . The m , M , and 95% confidence interval values were determined using the Fluctuation Analysis CalculatOR ( FALCOR ) program , with r and Nt as the input values ( http://www . keshavsingh . org/protocols/FALCOR . html ) [49] . The significance of differences between the mutation rates of the mutants and the WT was estimated by the Wilcoxon-Mann-Whitney test to obtain p values . To analyze the mutation spectra of pol30 mutant strains , genomic DNA was extracted from individual CAN1-resistant colonies . The CAN1 locus was PCR amplified using upstream and downstream primers and the PCR product was sequenced using 3 primers spanning the entire ORF . Analysis of sequences was performed using the Geneious program . Yeast cell extracts were generated from 0 . 5 L of logarithmic cultures using conventional methods . Briefly , cell pellets were lysed with Cell Lytic ( Sigma ) , supplemented with protease inhibitors ( Sigma ) and glass beads , as suggested by the manufacturer . Following centrifugation , cell extracts were collected and protein concentration was determined by the BCA method . ELISA plates coated with rabbit α-PCNA antibodies ( 1∶3000 , Adar Biotech ) were incubated with 100 µl of yeast cell extract at a protein concentration of 3 mg/ml for 1 h at RT . Following 3 washing steps with PBST , wells were incubated with either mouse α-His antibodies ( 1∶500 , Santa Cruz Biotechnology ) , to detect PCNA adsorption , or α-GFP antibodies ( 1∶500 , Roche ) , to detect the presence of GFP-tagged PCNA partners bound to PCNA ( see Figure 4 ) . Plates were then washed 3 times with PBST and incubated with secondary HRP-conjugated goat α-mouse antibodies ( 1∶2000 , Jackson ) . PCNA-partner complex levels were calculated as the ratio of the GFP signal to the PCNA signal detected for the same cell extract . Values represent averages of at least 5 independent repeats . Selected haploid PCNA mutants were grown to OD600 0 . 8 , centrifuged , and lyzed using cell lytic solution ( Sigma ) supplemented with protease inhibitor cocktail ( Sigma ) , following the manufacturer's instructions . Following TCA treatment , samples containing 10 µg of crude lysates were loaded on a 10% SDS-PAGE gel and subjected to western blot analysis using rabbit α-PCNA ( custom-made by Adar Biotech , 1∶2000 in PBS+1% BSA ) and mouse α-Pgk1 ( Invitrogen , 1∶7000 in PBS+1% BSA ) antibodies . Antibody binding was detected using either HRP-conjugated goat α-rabbit ( 1∶10 , 000 ) or HRP-conjugated goat α-mouse ( 1∶10 , 000 ) antibodies , respectively . The latter were used to detect the yeast Pgk1 protein that served as a loading control .
|
Many biological processes are mediated by complex protein-protein interaction networks . The most highly connected proteins in such networks , termed hub proteins , precisely regulate biological processes by the regulated and sequential binding and releasing of partner proteins . In the case of DNA replication and repair , proliferating cell nuclear antigen ( PCNA ) is a hub protein that encircles the DNA to dynamically bind and release a variety of DNA-modifying enzymes . In this work , we explored the impact of subtle alterations of PCNA-partner interaction affinities on DNA replication and repair in yeast . Using directed evolution approaches , we generated a large library of PCNA mutants and selected for those with enhanced affinity for five different PCNA partners . In vivo analysis of such mutants indicated the high sensitivity of DNA replication and repair processes to minor alterations in PCNA-partner interaction affinities . Importantly , we discovered that some of the defects observed in the strains with increased PCNA-partner protein interaction far exceed the defects observed when the same partner protein is deleted altogether . Our analysis suggests that the cost of misregulating biological processes through disruption of the carefully orchestrated action of hub-interacting proteins can be much higher than the cost of deleting parts of the network altogether , demonstrating both the fragility and robustness of biological processes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
] |
[
"biochemistry/molecular",
"evolution",
"molecular",
"biology/dna",
"replication",
"molecular",
"biology/dna",
"repair"
] |
2010
|
Subtle Alterations in PCNA-Partner Interactions Severely Impair DNA Replication and Repair
|
Few animals are known to lay eggs in the absence of ovulation or copulation , as it is presumably energetically wasteful and subjected to negative selection . Characterization of Smed-boule , a member of the DAZ family of germline RNA-binding proteins , revealed that egg capsule ( or capsule ) production and deposition occurs independently of the presence of gametes in the planarian flatworm Schmidtea mediterranea . Reduction of Smed-boule expression by RNA-interference ( RNAi ) causes ablation of spermatogonial stem cells and the inability of ovarian germline stem cells to undergo oogenesis . Although animals subjected to Smed-boule RNAi lose their gametes and become sterile , they continue to lay egg capsules . Production of sterile capsules is even observed in virgin Smed-boule ( RNAi ) and control planarians maintained in complete isolation , demonstrating that egg production in S . mediterranea occurs independently of ovulation , fertilization , or mating . Evidence suggests that this is a conserved feature amongst Platyhelminthes , and therefore relevant to the pathology and dissemination of parasitic flatworms . These findings demonstrate that Smed-boule functions at different stages during male and female germline stem cell development , and also demonstrate that egg capsule production by planarian flatworms occurs independently of signals produced by mating or ova .
The characterization of developmental processes involved in sexual reproduction has important implications towards reproductive medicine , stockbreeding , farming , and for controlling the dissemination of infectious disease . Evolutionarily conserved molecular processes involved in metazoan germline development have been identified through decades of research using model organisms . For example , post-transcriptional regulation of gene expression by conserved germline-specific RNA-binding proteins is one of the conserved molecular processes that ensure development of gametes [1–3] . On the other hand , there is great diversity in the processes that occur during and after fertilization , many of which are the outcome of speciation events [4 , 5] . Planarian flatworms belong to the phylum Platyhelminthes , and are well known for their extraordinary regenerative abilities , which are founded in the availability of a pluripotent stem cell population throughout their life [6–9] . The evolutionary history of these organisms has yielded extreme divergence of reproductive strategies , both between and within populations of different planarian species . For example , there are planarians that rely exclusively or temporally on asexual reproduction , which involves transverse fission and stem cell driven regeneration [7 , 10] . There are also populations of planarians that reproduce predominantly through parthenogenesis ( Pongratz et al . , 2003 ) . However , the default reproductive strategy of turbellarians is believed to be hermaphroditic sexual reproduction [11] , more specifically for planarians , through cross-fertilization and oviparity [12] . By contrast , some parasitic flatworms ( i . e . schistosomes or blood flukes ) have complex life cycles that involve dioecious and asexual reproductive phases during transitions between vertebrate and invertebrate hosts , respectively [13] . Since the complex life cycle of schistosomes complicates husbandry and experimentation in laboratory settings , researchers have begun to use planarian flatworms as a model to dissect the molecular mechanisms behind the extensive lifespan and reproduction of their parasitic cousins [14] . One aspect of particular interest is the continuous production of thousands of eggs that both facilitate dissemination and sustain the pathology of schistosomes by populating organs of their host [13 , 14] . Planarian flatworms have become useful models for the study of metazoan germline development [12 , 15] . In general , the specification of germline stem cells can occur through mechanisms that involve: 1 ) inherited material deposited in the cytoplasm of the maturing oocyte ( preformation ) ; or 2 ) embryonic stem cell differentiation in response to inductive cell-to-cell interactions ( epigenesis ) [16 , 17] . Inductive determination occurs in mice and is also observed in planarians , both initially and during regeneration of fragments that lack germ cells , and it occurs through differentiation of pluripotent somatic stem cells called neoblasts [18] . In the planarian species Schmidtea mediterranea , germline stem cells are first detected as dorsolateral clusters in the area where testes develop [18] . In other planarian species , such as Dugesia ryukyuensis , germline stem cells are first detected in the area of the ovaries [19 , 20] . Upon feeding and growth , planarians that reproduce sexually develop a hermaphroditic reproductive system and their gonads begin continuous production of gametes [12 , 19] . Germline stem cells in the ovary enter oogenesis and produce oocytes of approximately 40 μm in diameter that exit the ovary and are fertilized by sperm deposited in the tuba [12 , 15] . Even-though oocytes are large in comparison to other planarian cells ( e . g . neoblasts are ~8 μm diameter ) these do not hold the nutrients necessary for embryonic development , as is normally observed in eggs of insects , amphibians or fish ( to name a few ) . Instead , planarian yolk glands ( vitellaria in other flatworms ) produce separate cells that provide material required for egg capsule shell formation and nurturing embryonic development [21 , 22] . The development of planarian germline stem cells depends on conserved post-transcriptional regulators such as Nanos and Bic-C [15 , 18] . Boule is an RNA-binding protein encoded by the basal member of the Deleted in AZoospermia ( DAZ ) gene family , which is required for germ cell development in species ranging from sea anemone to humans [23 , 24] . How DAZ family homologs contribute to germline development in planarians remains unknown . In this study , we characterize a Boule homolog in the planarian Schmidtea mediterranea and demonstrate that it functions at different stages during male and female germline development . Functional analyses by RNA-interference ( RNAi ) revealed that Smed-boule is required for development and maintenance of spermatogonial stem cells , but disposable for the existence of their oogonial counterparts , uncovering the presence of sex-specific germline stem cells in planarian hermaphrodites . Long-term analysis of Smed-boule knockdowns revealed that egg capsule deposition in planarians is not triggered by gametogenesis , ovulation , oocyte activation , fertilization , or mating . These results demonstrate that egg capsule formation occurs regardless of signals from sexual activity or germ cell activity in S . mediterranea . These findings also provide a unique opportunity to identify internal mechanisms that influence capsule production in Platyhelminthes , which is central in the dissemination and pathology of parasitic members of this phylum .
We identified a boule homolog in the planarian flatworm S . mediterranea with a region of amino acid sequence 55% identical with that of the RNA recognition motif of human BOLL ( E-value = 1e-23; Fig 1A ) . The protein encoded by this gene shared highest homology with members of the Boule-like subfamily , as compared with other members of the DAZ family of proteins ( Fig 1B ) . Expression of this gene ( from here on referred to as Smed-boule or boule ) was detected by whole-mount in situ hybridization ( ISH ) in testes and ovaries of sexually mature planarians that are actively laying egg capsules ( Fig 2A–2D ) . Smed-boule expression was also detected in testis primordia of hatchlings and sexually immature animals ( Fig 2D and 2D’ ) . To better understand the distribution of Smed-boule expression in testes and ovaries , we performed detailed analysis by double fluorescent in situ hybridization ( FISH ) with the germline stem cell marker Smed-nanos [18] ( Fig 3 ) . Detection of Smed-boule mRNA overlapped with that of Smed-nanos in testes ( Fig 3A ) and partially in ovaries ( Fig 3B ) . The presence of Smed-boule mRNA was also robustly detected in the spermatogonial layer of the testes ( Fig 3A ) . Detection of Smed-boule expression was not apparent in the soma , and thus we conclude that expression of this gene is restricted to the germline in S . mediterranea . To test the function of Smed-boule in planarian germline development and sexual reproduction we subjected planarians to three months of RNAi . Planarians continuously turn over all cells in their body from a continuous population of pluripotent stem cells , which allowed us to assess whether Smed-boule is required for normal germline development in sexually mature adults using germ cell markers ( Fig 4A–4F; S1 and S2 Figs ) . Groups of seven sexually mature planarians were fed liver supplemented with 100 ng/μl of double-stranded RNA ( dsRNA ) twice per week . Smed-boule knockdowns ( Smed-boule ( RNAi ) ) were compared to control ( RNAi ) planarians . DsRNA corresponding to a planarian Cytoplasmic Polyadenylation Element Binding Protein 1 homolog ( CPEB1 ) , which is required for yolk gland development and egg capsule production ( below ) , was administered to an additional group ( CPEB1 ( RNAi ) ) as readout of RNAi effectiveness overtime . At the end of three months of RNAi , we observed that both oocytes ( Fig 4B; S2B Fig ) and sperm ( Fig 4E; S2B” Fig ) were absent in Smed-boule ( RNAi ) . No defects in oocyte or sperm development were observed in control ( RNAi ) planarians ( Fig 4A and 4D; S2A–S2A” Fig ) . The testes of CPEB1 ( RNAi ) samples were fully developed ( Fig 4F ) , but their ovaries displayed abnormal morphology and distribution of oocyte marker expression ( Fig 4C ) . From these results we concluded that Smed-boule is required for development of sperm and ova in S . mediterranea . Neoophoran flatworms rely on a particular approach to oviparity in which nutrients for the developing embryo ( yolk ) are not accumulated in the developing ova . Instead , nutritional support is contributed by yolk cells ( vitellocytes ) , which are transferred from yolk glands to the planarian uterus and encapsulated with embryos during egg capsule deposition . We checked for the presence of yolk glands using the yolk cell marker Smed-surfactant b ( S1 Fig ) , which proved to be of comparable abundance and distribution in control ( RNAi ) and Smed-boule ( RNAi ) animals ( Fig 4G and 4H ) . However , the presence of yolk glands in CPEB1 ( RNAi ) animals was severely reduced ( Fig 4I ) . We looked for other possible defects in the development of somatic reproductive structures but were unable to find any abnormalities other than the absence of accumulated sperm in the seminal vesicles of Smed-boule ( RNAi ) ( Fig 5 ) , which is due to their inability to produce sperm ( Fig 4E ) . There was also a noticeable difference in size of CPEB1 ( RNAi ) planarians , which were on average 30 . 8% larger than control animals maintained under the same conditions ( n = 14; unpaired two-tailed t-test , p < 0 . 05; S3 Fig ) . The normality and functionality of the accessory reproductive system in Smed-boule ( RNAi ) was further supported by quantitative analyses of egg capsule production ( below ) . As expected , from planarians with underdeveloped yolk glands ( Fig 4I ) , CPEB1 ( RNAi ) ceased laying eggs within a month of RNAi ( Fig 4J ) . The rate of egg production in Smed-boule ( RNAi ) was unaffected during the three months of RNAi treatment ( Fig 4J ) . Both the control ( RNAi ) and Smed-boule ( RNAi ) groups continuously laid eggs for the entirety of the experiment ( Fig 4J ) . In fact , an increase of 33% and 30% in egg capsule production was observed in Smed-boule ( RNAi ) when compared to control ( RNAi ) planarians during the second and third months of RNAi treatment , respectively ( unpaired two-tailed t-test , p < 0 . 05; Fig 4J ) . Given the surprising result that planarians devoid of gametes continued to deposit egg capsules , we monitored and quantified the number of fertile capsules ( yielding progeny ) produced by the different knockdown groups for two months after capsule deposition . From this , we discovered that egg capsules produced by animals subjected to two months of Smed-boule RNAi completely ceased to hatch ( Fig 4K ) . Egg capsules produced by control ( RNAi ) groups hatched 22% to 48% of the time ( Fig 4K ) . From these results , we concluded that Smed-boule function is required for germline development and sexual reproduction in S . mediterranea , but dispensable for production of egg capsules . Furthermore , the continuous production of egg capsules by groups of Smed-boule ( RNAi ) planarians ( Fig 4J ) suggested that production and deposition of egg capsules do not require fertilization , contributions from sperm , ovulation , or the presence of oocytes . Given the fact that Smed-boule ( RNAi ) planarians were capable of producing sterile egg capsules in the absence of germ cells ( and therefore fertilization events ) , we hypothesized that control animals would also produce sterile egg capsules in the absence of fertilization events . To test this hypothesis , we obtained ≤ 1 week-old hatchlings ( which lack ovaries , testes , yolk glands , and accessory reproductive organs ) and maintained them in isolation for four months under continuous RNAi regimens . Planarians were maintained in isolation throughout the experiment , which allowed us to test whether egg capsule production is independent of signals produced during mating or the presence of potential mates altogether . Since planarians in this experiment were subjected to Smed-boule RNAi within a week of being born , which is a point when no sperm has developed , this approach also allowed us to verify that lingering sperm in adult knockdowns used in the previous experiment was not contributing to egg capsule production . Two categories of isolated virgins were maintained on either liver containing Smed-boule dsRNA or control dsRNA and were fed twice per week . These animals were expected to grow and eventually reach sexual maturity under these husbandry conditions . The production of egg capsules would only occur if independent from stimuli produced during mating , fertilization , embryonic development and , in the case of Smed-boule ( RNAi ) , the absence of gametes . Indeed , both control and Smed-boule ( RNAi ) isolated animals produced egg capsules during the third and fourth months of the experiment ( Fig 6A ) . The number of capsules produced during the length of the experiment by isolated individuals from each category ranged from none to six ( Fig 6B ) . The average number of capsules deposited by individuals in the control category was slightly , but not significantly higher than those of Smed-boule ( RNAi ) ( unpaired two-tailed t-test , p = 0 . 25 ) ( Fig 6B ) . As expected from results observed in animals subjected to RNAi in the presence of potential mates ( Fig 4J and 4K ) , none of the egg capsules produced by Smed-boule ( RNAi ) individuals were fertile ( n = 0/28 capsules ) . Capsules produced by control RNAi animals were also completely sterile ( n = 0/43 capsules ) , suggesting that the production of egg capsules in these animals were not due to self-fertilization or parthenogenesis . We verified that normal gamete development was present in control animals at the end of the isolation experiment ( Fig 6C and 6E ) and absent in Smed-boule ( RNAi ) flatworms ( Fig 6D and 6F ) , which was expected from analyses of knockdowns not maintained in isolation ( Fig 4A , 4B , 4D and 4E ) . We also validated successful development of yolk glands in control and Smed-boule ( RNAi ) planarians raised in isolation ( Fig 6G and 6H ) . Collectively , these results demonstrate that production of egg capsules in S . mediterranea occurs in response to internal triggers that do not require the presence of a mate , mating , or fertilization events . Furthermore , the production of egg capsules by Smed-boule ( RNAi ) planarians suggests that this trigger is detached from signals originating from sperm and oocyte development or ovulation . We decided to evaluate the severity of germline development defects caused by Smed-boule RNAi . The most severe phenotype would be the loss of germline stem cells , which are specified and maintained post-embryonically through neoblast differentiation [18] . Germline stem cells in the planarian ovaries and testes can be identified by the characteristic expression of germinal histone H4 and nanos [18 , 20 , 25 , 26] . We tested for the presence of germline stem cells in control ( RNAi ) and Smed-boule ( RNAi ) by nanos ISH after 3–4 months of RNAi ( at the end of experiments in Figs 4J–4K and 6A–6B ) . Whole-mount ISH analysis of germinal histone H4 and nanos expression revealed the presence of germline stem cells in the testes region of control ( RNAi ) animals ( Figs 7A and 8A ) . However , germline stem cells were completely absent from the testes region of Smed-boule ( RNAi ) planarians ( Figs 7B and 8B ) . Surprisingly , germline stem cells in the ovary region of both control ( RNAi ) and Smed-boule ( RNAi ) planarians were readily detectable ( Figs 7A’ , 7B’ , 8C and 8D ) . Identical results were observed from hatchlings raised subjected to Smed-boule RNAi while maintained in isolation ( S4 Fig ) . Furthermore , analysis of germline stem cells in presumptive testis primordia present in asexual strains of S . mediterranea ( Wang et al . , 2007 ) were also lost after Smed-boule RNAi ( S5 Fig ) . From these results , we conclude that spermatogenesis defects in Smed-boule ( RNAi ) are due to the absence of male germline stem cells , whereas defects in oogenesis occur further downstream in the differentiation pathway . To better evaluate the progression of oogenesis in Smed-boule ( RNAi ) ovaries , we analyzed control and Smed-boule knockdowns stained with DAPI by confocal microscopy ( Fig 8E and 8F; S1 and S2 Movies ) . DAPI is retained by DNA and allowed for the visualization of numerous large oocytes with condensed chromosomes in the ovaries of control samples ( Fig 8E; S1 Movie ) . In contrast , neither oocytes , cells with condensed chromosomes , or otherwise recognizable mid- to late-oogenic intermediates , were detectable in ovaries of Smed-boule ( RNAi ) flatworms ( Fig 8F; S2 Movie ) . From these results we conclude that Smed-boule function is required during the initial stages of oogenesis , sometime before development of primary oocytes , but after specification of ovarian germline stem cells . The different outcomes observed on germline stem cells of testes and ovaries following Smed-boule ( RNAi ) reveal that these are two fundamentally distinct germline stem cell populations that require Smed-boule function at different developmental stages . Smed-boule function is necessary for neoblast differentiation into male germline stem cells and/or maintenance of male germline stem cells , whereas ovarian germline stem cells only require Smed-boule for progression through early stages of oogenesis ( Fig 8G ) . Furthermore , the severe defects in germline development observed after Smed-boule RNAi further support the hypothesis that egg capsule production and deposition occur independently of gametes , ovulation , parthenogenesis , fertilization , mating , or embryonic development in S . mediterranea .
Collectively , the data presented demonstrate that production and deposition of the egg capsules that ensure development of planarian embryos occur independently of fertilization events . Rather , it seems that egg capsule deposition , at least in S . mediterranea , is driven by intrinsic signals that are activated once these flatworms grow past a certain size and develop their yolk glands and other accessory reproductive organs . Given these findings , conclusions regarding planarian “fecundity” previously calculated from the rate of capsule production [27 , 28] , may need to be re-evaluated . Additionally , knowledge of the separation between capsule deposition and fertility should assist in the study of planarian germline and embryonic development , as well as in generation of methodologies for transgenesis , which have proven elusive to this point . Planarian reproduction can occur asexually through transverse fission , or sexually through post-embryonic development of a hermaphroditic reproductive system [7] . In planarians committed to sexualization , the development of gonads and gametes precedes formation of the oviducts , sperm ducts , and copulatory organs [19] . Yolk gland development in S . mediterranea , which is essential for production of egg capsules , is initiated towards the end of sexual development depending on sufficient nutritional intake and growth . The rate of egg capsule production observed in our experiments ( 1 to 5 egg capsules per animal per month; Fig 4J ) is comparable to those observed in different planarian species both in their natural habitat and raised under laboratory conditions following a similar liver-only diet [29 , 30] . Therefore , we believe that the conditions used for husbandry of S . mediterranea in the laboratory are conducive to normal egg capsule production rates , and that this is not the limiting factor in reproductive output . However , the low yield of fertile egg capsules observed from control animals in our experiments ( 22% to 48%; Fig 4J and 4K ) suggests that optimal laboratory husbandry conditions need to established to promote oocyte production , ovulation , or mating ( either of which may be rate limiting in actual reproductive output ) . How can triggering egg production independently of fertilization be an efficient approach to survival of planarian populations ? First , we must consider that in terms of sexual reproduction , S . mediterranea performs rather poorly under laboratory conditions . This is supported by the studies of Jenkins and Brown [29] who observed D . dorotocephala yield an average of 16 . 5 hatchlings per egg ( approximately 10-fold higher from what is observed in our laboratory for fertile egg capsules of S . mediterranea ) . Studies in S . polychroa have shown that siblings emerging from a single egg result from different fertilization events , which is possible because sperm from one or more partners can be stored for at least a month after insemination [31] . The ability to store sperm for an extended period of time after insemination , combined with the delay in development of yolk glands in comparison to the rest of the reproductive system , presents a scenario that would benefit from a mechanism that triggers capsule formation independently of copulation . In fact , it would be optimal if the activation of capsule formation also triggers ovulation of the many fully-grown oocytes present in ovaries of sexually mature planarians ( Consequential Model; S6A Fig ) . Since sperm can be stored in the tuba , massive ovulation could maximize the number of hatchlings generated per capsule . Alternatively , encapsulation of multiple embryos in a single egg capsule could also be facilitated by extended storage of zygotes prior to capsule deposition ( Complete Autonomy Model; S6B Fig ) . We are currently unable to differentiate between these two possibilities , or the possibility that passage of oocyte precursors ( e . g . oogonial or female germline stem cells ) may activate capsule formation . Indeed , oocytes were not detected in Smed-boule ( RNAi ) planarians ( n = 0/20; Figs 4B and 6D; S2B Fig ) , but ovaries and oogonial stem cells were readily observed ( n = 14/15; Figs 7B , 8D and 8F; S4D Fig; S2 Movie ) . It is possible that the release of early oocyte precursors from the ovary triggers capsule formation . However , this hypothesis is challenged by the fact that dozens of hatchlings often emerge from single capsules of different planarian species , and the observation that the rate of capsule production was not compromised in Smed-boule ( RNAi ) when compared to control planarians ( which contained both oocytes and precursors ) . Nevertheless , current and previous observations do support a model by which a sustainable approach to oviparity could rely on a trigger for capsule formation that is independent of mating , fertilization , or ovulation . Inside the phylum Platyhelminthes , free-living species ( such as S . mediterranea ) are evolutionarily distant from members of parasitic groups ( Trematoda , Monogenea , and Cestoda ) . However , the non-causative relationship between ovulation/fertilization and capsule production appears to be conserved in some cestodes and trematodes , whose dissemination and pathology depend on the continuous production of egg capsules . Parasitic flatworms of the genus Schistosoma have been reported to produce egg capsules from females after single-sex infections of mammalian hosts [32 , 33] . Although female schistosomes depend on interactions with a mate to fully grow and develop their gonads , they are also able to develop some vitelline cells and immature ovaries on their own . Shaw [33] observed the production of infertile capsules from females without male stimuli , probably through mechanisms conserved with those reported here for planarian flatworms . Similarly , parasitic flatworms belonging to the class Cestoda ( tapeworms ) have been reported to produce unviable egg capsules in the absence of fertilization events when cultured in vitro [34 , 35] . Thus , given that continuous production and deposition of egg capsules is central to dissemination and pathology of different types of parasitic flatworms , the molecular machinery involved in egg capsule production ( and not germline development alone ) becomes a desirable target for therapeutic developments .
A laboratory sexual strain of Schmidtea mediterranea [36] was used all experiments , except for those presented with asexual planarians [37] in S5 Fig . Planarian cultures were maintained in 0 . 75x Montjuïc Salts at 18°C under dark conditions , whereas 1x Montjuïc Salts and 21°C were used for asexuals as per [37] . Planarians were exposed to room temperature and light during weekly feedings of pureed organic beef liver ( Vantage USA , Oak Park , Illinois ) . Experimental animals were starved at least seven days before experimentation . Smed-boule was identified from a collection of S . mediterranea contig sequences assembled from RNAseq and conventional cDNA expressed sequence tag reads ( [38]; https://www . ideals . illinois . edu/handle/2142/28689 ) . A PCR product corresponding to Smed-boule ORF sequence was amplified from oligo ( dT ) -primed total RNA cDNA using 5’-GTTGTTTCAACGGTTCTACTGGCATC -3’ and 5’- GATTATTCCGGACAAAGCTGGACAAG -3’ forward and reverse primers ( respectively ) and ligated to pJC53 . 2 [39] after Eam1105I restriction digest . The insert sequence was validated and deposited into NCBI under accession number KT709533 . Fixation and preparation of samples for whole-mount in situ hybridization and DAPI staining were performed as per King and Newmark [40] . Colorimetric development for visualization of riboprobes was performed as described by Pearson et al . [41] . Smed-boule riboprobes were synthesized using SP6 RNA Polymerase . Smed-CPEB1 ( NCBI accession number KU990884 ) , Smed-nanos , were also synthesized from a pJC53 . 2-based construct [39] , whereas Smed-synaptotagmin XV , Smed-granulin ( grn ) , Smed-surfactant b , germinal histone H4 , and a tuba/oviduct marker were synthesized from pBluescript-based clones ( PL04017B1F10 , PL05005A1F08 , PL010001001D12 , pBS-gH4 , and PL04015A2A02 , respectively [18 , 36 , 42] ) using T3 RNA Polymerase . Colorimetric and low magnification analyses of DAPI signals from testes were imaged on a Zeiss Axio Zoom . V16 stereoscope . Confocal images were captured on an Olympus FluoView FV1000 confocal microscope hosted at Wright State University’s Microscopy Core Facility . Double-stranded RNAi feedings were performed twice every seven days and the protocol was followed as previously described [43] . DsRNA corresponding Escherichia coli ccdB sequence , which does not affect planarian development or behavior was used for unaffected control groups . For isolated RNAi samples , each planarian was fed individually and in isolation . For other experiments , planarians were maintained in groups of seven animals . Groups of seven sexual planarians of 0 . 5 to 0 . 7 cm size and with a visible gonopore were maintained in glass Petri dishes and subjected to dsRNA feedings as described above . For isolated experiments , single ≤ 1 week-old hatchlings were maintained in isolation in glass Petri dishes throughout the experiment , under the husbandry conditions described above . Isolated planarians were fed liver containing control or Smed-boule dsRNA twice per week , at which point any capsules present were collected and isolated . DsRNA corresponding to E . coli ccdB sequence was used for control samples . Egg capsules were monitored for hatchling events weekly for a period of three months after deposition .
|
Our work shows that production and deposition of egg capsules by planarian flatworms does not require fertilization , mating , ovulation , or even the existence of gametes . We also uncovered evidence for the existence of gender-specific germline stem cells in Schmidtea mediterranea , a hermaphroditic species of flatworm that develops germ cells post-embryonically . These findings surfaced from the characterization of Smed-boule , a member of the Deleted in AZoospermia gene family of RNA-binding proteins required for germline development in a broad range of animals . These findings lead to a better appreciation of the evolutionary diversity in approaches to oviparity . Additionally , discovering that egg capsule production occurs independently of germline or mating activities may carry a potential applied aspect with regards to regulating the dissemination and pathology of parasitic flatworms ( such as blood flukes and tapeworms ) , if conserved in these organisms .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"reproductive",
"system",
"rna",
"interference",
"animals",
"germ",
"cells",
"oocytes",
"stem",
"cells",
"epigenetics",
"planarians",
"sperm",
"genetic",
"interference",
"animal",
"cells",
"flatworms",
"gene",
"expression",
"biochemistry",
"rna",
"ovaries",
"testes",
"anatomy",
"nucleic",
"acids",
"cell",
"biology",
"ova",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms",
"genital",
"anatomy"
] |
2016
|
Germline Defects Caused by Smed-boule RNA-Interference Reveal That Egg Capsule Deposition Occurs Independently of Fertilization, Ovulation, Mating, or the Presence of Gametes in Planarian Flatworms
|
Vectors derived from human adenovirus type 5 , which typically lack the E1A and E1B genes , induce robust innate immune responses that limit their therapeutic efficacy . We reported previously that the E1B 55 kDa protein inhibits expression of a set of cellular genes that is highly enriched for those associated with anti-viral defense and immune responses , and includes many interferon-sensitive genes . The sensitivity of replication of E1B 55 kDa null-mutants to exogenous interferon ( IFN ) was therefore examined in normal human fibroblasts and respiratory epithelial cells . Yields of the mutants were reduced at least 500-fold , compared to only 5-fold , for wild-type ( WT ) virus replication . To investigate the mechanistic basis of such inhibition , the accumulation of viral early proteins and genomes was compared by immunoblotting and qPCR , respectively , in WT- and mutant-infected cells in the absence or presence of exogenous IFN . Both the concentration of viral genomes detected during the late phase and the numbers of viral replication centers formed were strongly reduced in IFN-treated cells in the absence of the E1B protein , despite production of similar quantities of viral replication proteins . These defects could not be attributed to degradation of entering viral genomes , induction of apoptosis , or failure to reorganize components of PML nuclear bodies . Nor was assembly of the E1B- and E4 Orf6 protein- E3 ubiquitin ligase required to prevent inhibition of viral replication by IFN . However , by using RT-PCR , the E1B 55 kDa protein was demonstrated to be a potent repressor of expression of IFN-inducible genes in IFN-treated cells . We propose that a primary function of the previously described transcriptional repression activity of the E1B 55 kDa protein is to block expression of IFN- inducible genes , and hence to facilitate formation of viral replication centers and genome replication .
A major obstacle to the therapeutic application and efficacy of adenoviral vectors is the induction of powerful innate and pro-inflammatory immune responses following systemic delivery [1]–[4] , independently of viral gene expression [5]–[9] . The constellations of chemokines produced in response to adenovirus vector infection depend on the host cell type and its species of origin , as do the mechanisms by which infection is detected by host cell pattern recognition receptors to activate signal transduction pathways and transcription of genes that encode these immunomodulators [2]–[4] . Nevertheless , production of several chemokines , including Rantes , Mip1-α and IL-8 , and such cytokines as interferon ( IFN ) α and β , Tnf-α and IL-6 has been observed upon infection of a wide variety of established and primary human and murine cells in culture and in vivo [4] , [7]–[19] . Interferon α and β , designated hereafter IFN , bind to the same heterodimeric receptor to establish a front line of anti-viral defense via stimulation of transcription of numerous genes [20]–[23] . The products of such interferon-stimulated genes ( ISGs ) inhibit replication of a wide variety of viruses by multiple direct or indirect mechanisms [22] , [24]–[27] . Proteins encoded by ISGs also reinforce synthesis of IFN and other cytokines , promote processing and presentation of antigens , and modulate the activity of important effector cells of the immune system [22] , [25]–[29] . The replication of human adenovirus type 5 ( Ad5 ) , from which nearly all vectors have been derived , is refractory to IFN in several lines of established human cells [30]–[32] , as a result of the actions of several viral gene products that counter the effects of the cytokine . The first to be identified , the small viral RNA , VA-RNA I [30] binds to , and prevents activation of , the interferon-induced , double-stranded RNA-activated protein kinase , which phosphorylates elF2-α to inhibit translation during the late phase of infection [33] . More recently , it has been established that the viral E4 Orf3 protein is required to prevent inhibition of viral DNA synthesis in type I IFN-treated cells [32] . This function of the E4 Orf3 protein correlates with reorganization of the promyelocytic leukemia protein ( Pml ) from the discrete , nuclear Pml bodies present in uninfected cells to track-like structure that also contain the viral protein , and is abrogated by shRNA-mediated knockdown of Pml or Daxx [34] . In addition , the viral E1A proteins suppress transcription of interferon-sensitive genes in infected cells [35]–[39] and both block activation of the Jak-Stat signaling pathway that induces transcription of ISGs and interact directly with Stat 1 co-activators [37]–[43] . The contributions of these viral products to modulation of innate immune responses in vivo have not been investigated intensively . Nevertheless , both the 243R E1A protein and the E3 gene , which encodes several proteins that inhibit inflammatory responses and apoptosis induced by binding of their ligands to the Tnfα and related receptors [44] , [45] , have been shown to decrease such responses to adenoviral vectors in various murine organs or tissues [46]–[49] . Comparison of induction of edema in mouse ears by vectors carrying different combinations of E1A , E1B and E3 coding sequences also implicated the E1B 19 kDa and 55 kDa proteins in inhibition of inflammatory responses [49] . The anti-inflammatory activity of the E1B 19 kDa protein was proposed to be the result of the anti-apoptotic activity of this viral Bcl-2 homologue [50] , [51] . The E1B 55 kDa protein makes an important contribution to optimizing the host cell environment for efficient viral replication [52] , [53] via formation of a virus-specific E3 ubiquitin ligase that also contains the viral E4 Orf6 protein , Cul5 and several other cellular proteins [54] , [55] . The activity of this enzyme targets the cellular proteins p53 , Mre11 , Rad50 , DNA ligase IV and integrin α3 for proteasomal degradation [54]–[60] . The destruction of Mre11 and Rad50 facilitates inhibition of the DNA double stranded break repair response and helps circumvent inhibition of viral DNA in infected cells [61]–[63] , while degradation of DNA ligase IV contributes to prevention of genome concatamerization [59] . Assembly of the virus-specific E3 ubiquitin ligase is also necessary for induction of selective export from the nucleus of viral late mRNAs [64] , [65] . One of the earliest functions ascribed to the E1B 55 kDa protein was repression of transcription of genes regulated by the tumor suppressor p53 in in vitro and transient expression assays [66] , [67] . This activity correlates with the ability of the E1B protein to cooperate with E1A proteins to transform rodent cells [68]–[71] . It has long been supposed that inhibition of p53-dependent transcription by the E1B 55 kDa protein in infected cells would contribute to preventing induction of cell cycle arrest or apoptosis upon stabilization and activation of p53 by the viral E1A proteins ( e . g . [52] , [72] , [73] ) . However , when p53 accumulates to high concentrations in cells infected by Ad5 mutants that cannot direct synthesis of this E1B protein ( E1B 55 kDa null-mutants ) , expression of p53-activated genes is not increased [74]–[76] . Indeed , as assessed by microarray hybridization , the reversal of the p53 transcriptional program is as complete in normal human cells infected by an E1B 55 kDa-null-mutant as in wild-type Ad5-infected cells [76] . However , in the absence of the E1B protein , expression of some 340 genes highly enriched for those associated with immune responses and anti-viral defense was increased significantly [76] . In particular , we observed that this set contained many interferon-sensitive genes , including GBP1-5 , IF1H1 ( MDA5 ) , IF1T2 , MX2 and TAP1 [76] . These observations suggested that repression of expression of such genes by the E1B 55 kDa protein might protect Ad5-infected cells against anti-viral measures induced by type I IFN . We now report the results of experiments designed to test this hypothesis , which demonstrate that the E1B 55 kDa protein represses expression of ISGs and blocks type I IFN-induced inhibition of viral DNA synthesis and replication in normal human cells .
In initial experiments to investigate whether repression of expression of interferon-stimulated genes ( ISGs ) by the E1B 55 kDa protein protects Ad5-infected cells from the anti-viral defenses induced by this cytokine , replication of Ad5 and the E1B 55 kDa null-mutant Hr6 were compared in HFFs treated with exogenous IFN . Cells were maintained in the presence of 500 units/ml IFN , or of vehicle only control , for 24 hrs . , prior to and during infection with 3 p . f . u . /cell Ad5 or Hr6 . They were harvested after increasing periods of infection , and viral yields measured by plaque assay on complementing 293 cells as described in Materials and Methods . This IFN treatment regimen decreased the yields of Ad5 by less than 3 fold ( Figure 1A ) , in agreement with previous observations ( see Introduction ) . In contrast , replication of Hr6 was reduced to a much greater degree , up to 500-fold . The Hr6 mutant was isolated after nitrous acid mutagenesis of Ad5 by virtue of more efficient replication in complementing 293 cells [77] , [78] than in non-complementing cells [79] . The mutation responsible for this phenotype was mapped to the E1B 55 kDa protein coding sequence by marker rescue and sequencing [80] . We have observed recently that the Hr6 genomes contain at least one additional mutation outside the E1B gene that substantially reduces the infectivity of virus particles ( S . Kato , J . C . and S . J . F . , manuscript in preparation ) . As described above , adenoviral VA-RNA I , E1A proteins and the E4 Orf3 protein have been reported previously to protect Ad5 replication against the inhibitory effects of IFN . It was therefore essential to determine whether the increased sensitivity of Hr6 replication to inhibition by IFN was the result of mutations in the E1B gene , or elsewhere in the genome . To this end , we exploited a mutant carrying the Hr6 E1B 55 kDa frameshift mutation ( deletion of base-pair 2347 ) [80] introduced into the genome of a phenotypically wild-type , E1-containing derivative of AdEasy [81] . As reported elsewhere [82] , no E1B 55 kDa protein can be detected in HFFs infected by this mutant ( AdEasyE1Δ2347 ) , and , as expected in the absence of this viral protein [80] , [83]–[86] , expression of viral late genes was impaired . HFFs maintained in the absence or presence of IFN were infected with 30 p . f . u . /cell AdEasyE1Δ2347 or its parent AdEasyE1 [82] , and yields determined after a single cycle of replication at 2 days p . i . Consistent with the results described above , replication of AdEasyE1 was inhibited by only a modest degree ( <4-fold ) in IFN treated cells ( Figure 1B ) . However , in the presence of IFN , the yield of AdEasyE1Δ2347 was inhibited nearly 300-fold ( Figure 1B ) , indicating that E1B 55 kDa protein prevents IFN-induced inhibition of replication in HFFs . As described in the Introduction , it has been reported previously that the E1B 55 kDa protein can repress transcription in simplified experimental systems . This property suggested that this protein is likely to inhibit transcription of ISGs in infected cells , but the microarray hybridization data collected previously [76] cannot distinguish among the multiple mechanisms by which RNA concentrations can be regulated . The concentrations of pre-mRNAs of representative ISGs increased in expression in Hr6- compared to wild type-infected HFFs [76] were therefore examined in the presence and absence of the E1B protein . HFFs that were not exposed to IFN were infected with 30 p . f . u . /cell AdEasyE1 or AdEasyE1Δ2347 for 30 hrs , and primary transcripts detected by using reverse transcription with random priming , followed by PCR with primers specific for three ISG pre-mRNAs , that is , spanning exon-intron junctions ( see Materials and Methods ) . To provide an internal control , GAPDH mRNA was examined in parallel . Pre-mRNAs transcribed from the IL6 , IFIT2 and STAT1 genes were present in mock-infected cells , and decreased significantly in concentration following infection with AdEasyE1 , whereas only minimal differences in GAPDH mRNA were detected ( Figure 2A ) . In contrast , synthesis of these ISG pre-mRNAs was not repressed in AdEasyE1Δ2347-infected HFFs , but rather these RNAs accumulated to higher concentrations than observed in uninfected or wild type-infected cells ( Figure 2A ) . For example , quantification of signals as described in Materials and Methods indicated that the concentration of IL6 pre-mRNA was 14-fold higher in mutant compared to wild-type-infected cells , whereas that of GAPDH mRNA was only 1 . 2-fold greater . These observations , which are consistent with our microarray hybridization data [76] , indicate that the E1B 55 kDa protein is a potent repressor of ISG pre-mRNA synthesis in infected cells . Although HFFs are permissive for adenovirus replication in tissue culture , we wished to confirm the sensitivity to IFN of E1B 55 kDa null-mutants in normal human bronchial/tracheal epithelial cells ( NHBECs ) , which better represent the host cell type encountered by serotype C adenoviruses in their natural site of infection , the upper respiratory tract [87] . To investigate the sensitivity of NHBECs to IFN , the concentrations of ISG pre-mRNAs were compared before and after IFN-treatment . Because phenotypes exhibited by E1B 55 kDa-null-mutants of Ad5 have been reported to be cell-type dependent [83] , [88]–[91] , we also examined the effects of infection in the presence or absence of the E1B 55 kDa protein on ISG expression . NHBECs treated with IFN or control as described in Materials and Methods were infected with 30 p . f . u . /cell AdEasyE1 or AdEasyE1Δ2347 for 24 hrs , and the concentrations of pre-mRNAs and of GAPDH mRNA examined by RT-PCR . As observed in HFFs ( Figure 2A ) , IL6 and STAT1 pre-mRNAs were detected in uninfected , untreated NHBECs , and accumulated to reduced concentrations in AdEasyE1- , but not in AdEasyE1Δ2347- , infected cells ( Figure 2B ) . The same pattern was observed for GBP1 mRNA . In these cells , IFIT2 pre-mRNA could be detected only following infection , and was present in significantly greater quantities in the absence of the E1B 55 kDa protein ( Figure 2B ) . With the exception of IL6 , the RNA products of these genes accumulated to increased concentrations in mock-infected cells exposed to IFN , indicating that NHBECs respond to this cytokine . Expression of all the ISGs examined was repressed in IFN-treated cells when infected by the wild-type virus , but not following AdEasyE1Δ2347 infection ( Figure 2B ) . To confirm that the sensitivity of E1B 55 kDa null-mutant virus replication to IFN was not specific to HFFs , the replication of the E1B 55 kDa-null-mutants Hr6 and AdEasyE1Δ2347 was compared to that of the corresponding wild-type virus in IFN-treated or untreated NHBECs . In these experiments , IFN had only a minor effect on replication of Ad5 or AdEasyE1 ( Figures 3A and B ) . Treatment with 250 U/ml IFN reduced Hr6 and AdEasyE1Δ2347 titers by between two and three orders of magnitude ( Figures 3A and 3B ) . Our previous studies have established that in these epithelial cells , in contrast to HFFs , E1B 55 kDa null-mutants do not exhibit defects in viral genome replication [92] . It is therefore unlikely that replication of the mutants would exhibit a lower degree of sensitivity to IFN at times later in the infectious cycle than examined in these experiments ( 36 hrs . p . i . ) . In toto these data indicate that exposure to exogenous IFN induces an antiviral state in HFFs and NHBECs that is detrimental to adenovirus replication , and blocked by the E1B 55 kDa protein . It is well established that exposure of cells to type I IFN can restrict replication of many types of virus by multiple mechanisms , including inhibition of viral gene expression , mRNA translation , and genome replication [22] , [24] , [93] , [94] . As a first step to identify the reaction ( s ) in the adenoviral life cycle that are protected by the E1B 55 kDa protein from inhibition induced by IFN signaling , the accumulation of viral early proteins was monitored in NHBECs infected with AdEasyE1 or AdEasyE1Δ2347 . Cells were treated with IFN as described above , and harvested after increasing periods of infection . Whole cell lysates were prepared , and the steady-state concentrations of the E1A proteins and the E2 DNA-binding protein ( DBP ) examined by immunoblotting ( Figure 4A ) . Somewhat reduced concentrations of E1A proteins were observed in IFN-treated compared to untreated cells at 12 hrs . p . i . , but the quantities of these proteins detected in AdEasyE1- and AdEasyE1Δ2347-infected cells exposed to the cytokine were similar at both 12 and 18 hrs . p . i . ( Figure 4A ) . The concentrations of DBP were reduced to a small degree by IFN-treatment of cells infected by AdEasyE1 or AdEasyE1Δ2347 , by approximately 40% and 55% , respectively . These results indicate that type I IFN induced modest decreases in the steady-state concentrations of viral early proteins , including E2 replication proteins epitomized by the DBP , regardless of whether the E1B 55 kDa protein was synthesized in infected cells . We therefore examined the next reaction in the infectious cycle , synthesis of viral genomes . NHBECs treated with IFN or untreated as described above were infected with 5 p . f . u . /ml AdEasyE1 or AdEasyE1Δ2347 for 24 hrs . Nuclear DNA was isolated , and viral DNA concentration was measured by qPCR as described in Materials and Methods . The concentrations of the viral genome at 24 hrs . p . i . , normalized to the input concentrations measured at 2 hrs . p . i . , were observed to be reduced upon IFN pretreatment by approximately 7-fold in AdEasyE1-infected cells , but by nearly 300-fold in AdEasyE1Δ2347-infected cells ( Figure 4B ) . We also monitored the formation of viral replication centers by using immunoflourescence to visualize DBP in IFN treated or untreated HFFs infected with the mutant or its wild-type parent , as described in Materials and Methods . The majority of untreated cells infected with AdEasy E1 or AdEasyE1Δ2347 contained discrete DBP-containing nuclear structures that appeared as discrete foci or ring-like or reticulated structures ( Figures 5A and 5B ) . Treatment with IFN induced little change in the formation of these structures in AdEasy E1-infected cells , whereas the majority of AdEasyE1Δ2347-infected cells exposed to the cytokine exhibited only diffuse nuclear DBP staining ( Figure 5A ) . Quantification of the different patterns of DBP staining indicated that approximately 25% of nuclei in untreated , AdEasyE1-infected cells stained positive for DBP , but exhibited diffuse nuclear localization with no replication center formation ( Figure 5B ) . Interferon treatment led to a less than a two-fold increase in the number of wild-type-infected cells without distinguishable replication centers . In contrast , when cells exposed to IFN were infected with AdEasyE1Δ2347 , 94% of infected nuclei exhibited only diffuse DBP staining ( Figure 5B ) . The reduced accumulation of viral genomes and failure to form replication centers after AdEasyE1Δ2347 infection of IFN treated cells could be attributed to either a defect in de novo DNA synthesis , and/or rapid degradation of input DNA . To distinguish between these possibilities , viral DNA concentrations were measured in IFN treated or untreated NHBECs , as described above , between 2 and 11 hrs . p i . to monitor degradation of viral DNA early in infection . In agreement with the results described previously , IFN treatment lead to only a modest decrease in the concentration of wild-type viral DNA by 22 hrs . p . i . , but a reduction of greater than 150-fold in AdEasyE1Δ2347-infected cells ( Figure 5D ) . In untreated cells infected with 5 p . f . u . /cell of either virus , modest decreases ( <50% reduction ) in viral DNA concentrations compared to the values measured at 2 hrs . p . i . were observed during the first few hours of infection ( Figure 5C ) . A similar pattern was observed in IFN-treated cells , and by 11 hrs . p . i . the concentrations of intranuclear viral DNA in AdEasyE1- and AdEasyE1Δ2347-infected , IFN-treated cells were similar . These data indicate that IFN does not induce degradation of viral genomes in the absence of the E1B 55 kDa protein . The previously identified set of genes repressed by the E1B 55 kDa protein upon infection of normal human cells [76] includes 130 listed as interferon responsive in the Monash Institute INTERFEROME database [21] ( Table S1 ) , 15 of which are associated with apoptosis ( Table 1 ) . We therefore wished to determine if the defect in viral genome replication observed when AdEasyE1Δ2347-infected cells were treated with IFN could be attributed to induction of apoptosis , and breaks in the viral genome . In one approach , IFN-treated and untreated HFFs were infected with AdEasyE1 or AdEasyE1Δ2347 for 34 hrs . , and numbers of annexin V- and propidium iodide ( PI ) -positive cells measured by flow cytometry , as described in Materials and Methods , to assess induction of early apoptosis and cell death , respectively . To provide a positive control , mock-infected cells were incubated with etoposide for 34 hrs in parallel . While etoposide treatment , which induces DNA damage and apoptosis , led to a dramatic increase in the numbers of cells positive for staining with annexin V or both annexin V and PI in control experiments , no significant increase was observed in IFN treated compared to untreated cells infected with either AdEasyE1 or AdEasyE1Δ2347 ( Table 2 ) . Analysis of HFFs by TUNEL assay similarly failed to detect apoptosis in AdEasyE1 or AdEasyE1Δ2347-infected cells , regardless of whether they were exposed to IFN ( Figure 6 and data not shown ) . These data indicate that inhibition of viral DNA synthesis in IFN treated cells infected by E1B 55 kDa null-mutants cannot be ascribed to induction of apoptosis . In Ad5 infected cells , nuclear Pml bodies , which are electron-dense spherical nuclear substructures defined by the presence of Pml proteins , are disrupted by the viral E4 Orf3 protein [95]–[97] . The relocalization of Pml into track-like structures that contain the E4 Orf 3 protein has been reported to be required to prevent inhibition of viral DNA synthesis induced by IFN-α or IFN-γ in Vero and IMR90 cells [32] , [34] . As previous studies have indicated that the E1B 55 kDa protein associates transiently with the E4 Orf3 protein and reorganizing Pml bodies [96] , [97] , we wished to determine whether the E1B protein is also required for Pml relocalization in IFN-treated cells . IFN-treated or untreated HFFs were infected with 30 p . f . u . /cell AdEasy E1 or AdEasy E1Δ2347 , fixed at 36 hrs . p . i . and stained for Pml and E4 Orf3 as described in Materials and Methods . Mock infected , untreated cells showed Pml staining in discrete nuclear puncta numbering on average 10 . 4±4 . 8 bodies per cell ( n = 80 ) ( Figure 7 , panels a–d ) . Exposure of mock-infected cells to IFN resulted in an increased number of Pml-staining foci of larger size ( 15 . 3±6 . 6 per cell , n = 80 ) , in agreement with previous observations ( see [98] ) . The concentration of Pml detected in wild-type infected cells was noticeably lower than that in mock-infected cells ( Figure 7 , compare panels b and j , and f and n ) , presumably because expression of the PML gene is repressed by the E1B 55 kDa protein [76] . As expected , the Pml protein was observed to be colocalize with the E4 Orf3 protein in distinct track-like or spherical structures ( Figure 7 , panels i–l , white arrows ) . The number of wild type-infected cells exhibiting elongated track-like structures that contained Pml and E4 Orf3 was reduced by IFN treatment , although Pml and E4 Orf3 remained colocalize ( Figure 7 , panels m–p ) . Pml bodies were also disrupted in cells infected with AdEasyE1Δ2347 , exhibiting a similar distribution of Pml and E4 Orf3 signals as observed in AdEasyE1-infected cells in the presence or absence of IFN ( Figure 7 , panels q–x ) . The concentration of Pml detected in these cells was , however , higher than that observed in wild type- infected cells ( Figure 7 , compare panels j and r , and r and v ) , as expected in the absence of the E1B protein repressor of PML transcription . Because the relocalization of Pml proteins occurred normally in AdEasyE1Δ2347-infected cells , E1B 55 kDa must block the IFN response by a mechanism that is distinct from the disruption of nuclear Pml bodies by E4 Orf3 . As discussed previously ( see Introduction ) , assembly of a virus-specific E3 ubiquitin ligase that contains the E1B 55 kDa and E4 Orf 6 proteins is required for many of the functions fulfilled by the E1B 55 kDa protein during the infectious cycle . Furthermore , it has been reported that >95% of the nuclear E1B 55 kDa protein present during the initial period of the late phase in Ad5-infected HeLa cells is assembled into this ligase [55] . Whether this partition of the E1B 55 kDa protein is representative of other periods in the infectious cycle , such as the early phase , or of Ad5 infection of other cell types is not known . We therefore wished to determine whether the ability of the E1B 55 kDa protein to block the IFN-mediated inhibition of viral replication in normal human cells is dependent upon formation of the virus-specific E3 ubiquitin ligase . Consequently , the sensitivity to IFN of replication of the E4 Orf6-null-mutant dl355 [99] and of a mutant carrying a 4 amino acid insertion in the E1B 55 kDa protein coding sequence , A143 [100] , that impairs its interaction with the E4 Orf6 protein were examined . Cells were infected with 5 p . f . u . /cell of these mutants , Ad5 , or the E1B 55 kDa-null-mutant Hr6 , harvested at 36 hrs . p . i . , and viral yields measured by plaque assay . In agreement with results described above , replication of Hr6 was inhibited to a significantly greater degree than that of Ad5 in IFN-treated cells ( Figure 8 ) . In contrast , yields of A143 and dl355 were reduced by only 1 . 4-fold and 2 . 6-fold , respectively , in cells exposed to the cytokine . These data indicate that neither the assembly of the virus specific E3 ubiquitin ligase nor the presence of the E4 Orf6 protein are required for the E1B 55 kDa protein to block inhibition of viral replication induced by IFN .
It is well established that the adenoviral E1B 55 kDa protein plays an important role in circumventing host cell mechanisms that limit viral replication . For example , the E1B- and E4 Orf6-protein-containing E3 ubiquitin ligase targets components of the MRN complex for proteasomal degradation [59] . When formation of this virus-specific enzyme and the ability of the E4 Orf3 protein to relocalize Mre11 are prevented by mutation , viral DNA synthesis is inhibited by a mechanism that is independent of formation of concatamers of the viral genome [62] , [101]–[103] . The studies reported here establish the E1B 55 kDa protein also provides an additional , previously unrecognized defense against host anti-viral measures: a frameshift mutation that prevents synthesis of this protein renders viral replication in normal human cells sensitive to exogenous type I IFN , with reductions in virus yield of greater than two orders of magnitude ( Figures 1 and 3 ) . Three smaller related E1B 55 kDa-proteins can also be produced from E1B transcripts by alternative splicing [104] . However , the 1 bp deletion ( of bp 2347 in the viral genome ) that prevents production of the E1B 55 kDa protein in the null-mutants studied here [80] , [82] lies downstream of the coding sequence for the N-terminal segment common to the E1B 55 kDa and its related proteins . Consequently , a function in subversion of inhibition of viral replication by IFN can be unambiguously ascribed to the E1B 55 kDa protein . Exposure of cells to type I IFN prior to and during infection resulted in modest decreases in the accumulation of viral immediate early ( E1A ) and early proteins , epitomized by the E2 DBP ( Figure 4A ) . However , the degree of inhibition of production of these proteins was the same in the absence as in the presence of the E1B 55 kDa protein ( Figure 4A ) . This observation indicates that all prior reactions in the infectious cycle , including attachment , entry , uncoating , transport of genomes to the nucleus and initial transcription within nuclei , proceed with the same or closely similar efficiencies in IFN-treated cells infected with WT and E1B 55 kDa null-mutant viruses . In contrast , accumulation of viral genomes was observed to be strongly reduced in IFN-treated cells in the absence of the E1B protein ( Figure 4B ) , and very few viral replication centers were formed ( Figures 5A and B ) . At this juncture , we cannot exclude the possibility that the protection against IFN-induced mechanisms of inhibition of viral replication afforded by the E1B 55 kDa protein also extends to one or more later reactions in the infectious cycle: transcription of viral late genes requires viral DNA synthesis of infected cells [105] , a reaction that is inhibited when E1B null-mutant-infected cells are exposed to IFN . The E1B 55 kDa protein does not participate in viral DNA synthesis [105] , nor is it required for this process in normal or transformed human cells [75] , [80] , [83]–[86] , unless the onset of viral early gene expression is delayed [92] , [106] . It must therefore act indirectly to allow viral genome replication and prior formation of replication centers in IFN-treated cells . The latter reaction is directed by entry of viral DNA into infected cell nuclei and does not require viral DNA synthesis . However , we found no evidence for enhanced degradation of viral DNA in IFN-treated cells when the E1B 55 kDa protein was not present ( Figure 5C ) , consistent with the similar efficiencies of early protein synthesis observed in WT and E1B 55 kDa null-mutant-infected cells . Furthermore , induction of apoptosis , which would lead to introduction of double-stranded breaks into entering viral DNA molecules , could not be detected in either WT- or E1B 55 kDa null-mutant-infected cells exposed to type 1 IFN ( Figure 6 , Table 2 ) , despite the repression of transcription of several apoptosis-associated ISGs by the E1B protein ( Table 1 ) . In toto , these observations suggest that formation of viral replication centers and genome replication depend on one or more alterations in intranuclear structures or components that can take place in IFN-treated cells only when the E1B 55 kDa protein is made . Several viruses with DNA genomes that are replicated in infected cell nuclei , including polyomaviruses , herpesviruses and adenoviruses , encode proteins that disrupt the intranuclear structures termed Pml bodies ( a . k . a . ND IOs ) , and in this way are thought to circumvent an intrinsic anti-viral defense [98] , [107] , [108] . In the case of Ad5 , the E4 Orf3 protein sequesters Pml in distinctive track-like structures [95]–[97] . Such relocalization of Pml is dispensable for viral replication , at least in established lines of human cells , as mutations that prevent synthesis of the E4 Orf3 protein , or its interaction with Pml , exhibit no growth defects following high multiplicity infection [109] , [110] . However , this reaction is necessary for efficient formation of viral replication centers in normal diploid fibroblasts or Vero cells exposed to IFN α [32] , [34] , which induces increased transcription of the genes that encode Pml and other proteins present in Pml bodies [98] . Inhibition of synthesis of Pml or the Pml body-associated co-repressor Daxx by RNAi restored formation of replication centers in E4 Orf3 mutant-infected cells exposed to exogenous IFNα , and the inhibitory effects of Daxx were shown to be independent of effects on expression or functions of viral early gene products . These observations establish a critical role for Pml bodies in IFN-induced inhibition of adenoviral replication . Although the mechanism of action of Pml body components remains unclear , Hearing and colleagues proposed that Pml and Daxx might function together to inhibit viral DNA synthesis in IFN-treated cells , or that Daxx functions as the effector of such inhibition upon Pml-dependent recruitment to Pml bodies [34] . The E1B 55 kDa protein has been reported to interact with both the E4 Orf3 protein during the initial period of the early phase of infection [97] and with Pml-bodies [96] . Furthermore , the defects in formation of viral replication-centers in IFN-treated cells observed in the absence of the E4 Orf3 [32] or the E1B 55 kDa ( Figures 5A , B ) proteins are very similar . Nevertheless , the E4 Orf3 protein sequesters Pml efficiently , even when this cellular protein is overproduced in IFN-treated cells in the absence of the E1B 55 kDa protein ( Figure 7 ) . This observation indicates that the E1B protein is dispensable for the protective reorganization of Pml bodies by the E4 Orf3 protein , and hence that these two viral early proteins block inhibitory effects of IFN by different mechanisms . The E1B 55 kDa protein has been reported to repress transcription of p53-regulated genes via binding to p53 and to contain a repression domain that inhibits transcription of reporter genes when fused to an heterologous DNA-binding domain [67] , [69] , [70] . This protein also induces inhibition of export of mature cellular mRNAs from the nucleus to the cytoplasm [80] , [86] . As such cellular mRNAs do not accumulate in infected cell nuclei [111] , this activity could contribute to repression of expression of interferon-sensitive genes by the E1B 55 kDa protein [76] . However , regulation of mRNA export in infected cells depends on assembly of the E1B- and E4 Orf6 protein-containing E3 ubiquitin ligase [64] , [65] , whereas protection of viral replication from IFN-induced inhibition does not ( Figure 8 ) . Rather , the E1B 55 kDa protein blocks accumulation of the primary transcripts of several ISGs in normal human cells , either untreated or exposed to exogenous IFN ( Figure 2 ) . It remains possible that the E1B 55 kDa protein inhibits splicing of such pre-mRNAs , or promotes their intranuclear turnover . However , this viral protein has not been implicated in regulation of pre-mRNA processing . Furthermore , we have observed recently that substitutions within the previously identified repression domain of the E1B 55 kDa protein [69] impair inhibition of synthesis of ISG pre-mRNAs in normal human cells ( J . S . C . , C . Gallagher and S . J . F . , manuscript in preparation ) . These observations are consistent with the conclusion that the E1B 55 kDa protein represses transcription of specific cellular genes during the productive cycle , and identify ISGs as an important target . We therefore propose that this viral protein permits viral DNA synthesis in IFN-treated cells by repressing transcription of one or more specific ISGs that encode a protein that either prevents formation of viral replication centers , or removes , or inactivates protein ( s ) essential for this process . An important implication of this hypothesis is that formation of replication centers is an active process essential for successful viral DNA synthesis and replication , rather than simply the result of association of viral DNA molecules and replication proteins within infected cell nuclei . It will therefore be of considerable interest to identify the cellular protein ( s ) that is targeted by the E1B 55 kDa protein to facilitate this reaction in the infectious cycle in IFN-treated cells . Although intranuclear sites of viral genome replication and transcription have been characterized in some detail [112] , very little is known about either the initial intranuclear localization and molecular interactions of Ad5 DNA , or dynamic changes that might be required for formation of viral replication centers . Furthermore , none of the host protein previously reported to be associated with viral replication centers , Sp100 [96] , Mdc1 [63] , and several proteins that participate in Atr-dependent signaling ( Atr itself , Atrip , Rpa32 , TopBP1 and E1B-Ap5/Hnrl1 ) [113]–[115] appear to be likely candidates: only the Pml body component Sp100 is encoded by an IFN-inducible gene [98] that is transcriptionally repressed by the E1B 55 kDa protein [76] , and , as discussed above , disruption of Pml bodies by the E4 Orf3 protein takes place normally in IFN-treated cells infected by the E1B 55 kDa null-mutant virus . As noted previously , it is well established that the E1B 55 kDa protein can act as a direct repressor of transcription in simplified experimental systems [66] , [67] . The mechanism of such repression remains incompletely understood , but has been proposed to reflect the recruitment of co-repressors to promoters by the viral protein [116] . The E1B 55 kDa protein has indeed been reported to interact with the cellular repressors Daxx [117] and via Sin3a , HdacI [118] , [119] , but the physiological consequences of these interactions have not been elucidated . At this juncture , it is not clear that this viral protein functions as a direct repressor of ISG transcription in infected cells . The virus-specific E3 ubiquitin ligase that contains the E1B 55 kDa protein is not required to prevent inhibition of viral replication by IFN ( Figure 8 ) . However , the E1B protein can also function as a SUMO-1 E3 ligase independently of the E4 Orf6 protein [119] , [120] , and could therefore modify and inactivate transcriptional regulators necessary for efficient transcription of IFN-inducible genes . Experiments are in progress to investigate whether the E1B kDa protein directly or indirectly represses transcription of cellular genes during the infectious cycle . The discovery of the previously unrecognized participation of the E1B 55 kDa protein in obstructing inhibition of viral replication by IFN indicates that at least four viral gene products help counter this host defense . Such a multiplicity is not unexpected , as many viruses that replicate in mammalian cells circumvent the anti-viral actions of IFN via multiple mechanisms and gene products [22] , [24] , [94] . The contributions of the Ad5 E1A , E1B 55 kDa and E4 Orf3 proteins and VA-RNA I to blocking inhibition of replication by IFN have not been evaluated systematically . However , the information currently available argues that these viral gene products function by different , non-redundant mechanisms . The E4 Orf3 protein and VA RNA I block the inhibitory effects of products of IFN-inducible genes , namely Pml [34] and Pkr ( aka Eif2ak2 ) [33] , respectively . In contrast , E1A proteins act prior to transcription of such genes by impairing assembly of the critical transcriptional activator [38] , [40] . Like E1A proteins , the E1B 55 kDa protein likely represses transcription of IFN-inducible genes . However , this protein also inhibits expression of several genes encoding proteins that induce synthesis of IFN in response to viral infection , including Rig 1 ( aka Ddx58 ) , Mda5 ( aka If1h1 ) , Irf7 and Myd88 [76] , suggesting that it also blocks the initial production of anti-viral cytokines . Experiments are in progress to test the hypothesis that the E1B 55 kDa protein is a broadly acting inhibitor of induction of the IFN response that complements the mechanism of E1A protein-mediated inhibition . The finding that the E1B 55 kDa protein is a potent inhibitor of induction of the IFN-mediated anti-viral defenses in Ad5-infected normal human cells is clearly of interest in the context of development of adenoviral vectors . It is likely that the deletion of the coding sequence for this E1B protein common to such vectors contributes to the induction of vigorous innate immune responses in vivo , particularly upon systemic delivery ( see Introduction ) . Typical vectors lacking E1A , E1B and E3 genes have been reported to induce increased expression of many genes in the liver of C57/B16 mice [12] , including genes repressed by the E1B 55 kDa protein in normal human cells [76] . However , it will be important both to identify genes repressed by the E1B 55 kDa protein and to investigate the role of this protein in blocking innate responses to infection in vivo . The identification of specific mutations that eliminate the transcriptional repression function of this E1B protein , but confer potentially advantageous properties such as tumor-cell selective replication , may also facilitate the design of more efficacious adenoviral vectors . Alternatively , this function may help account for the poorly understood tumor cell-selective replication of first generation oncolytic adenoviruses [88] , which carry mutations that prevent synthesis of the E1B 55 kDa and related proteins [105]: many human tumor cells carry mutations that result in defects in the production or response to IFN [121]–[123] .
293 cells were maintained in Dulbecco's modified Eagle's medium ( DMEM , GIBCO ) containing 5% bovine growth serum ( Thermo Scientific Hyclone ) and 5% calf serum ( GIBCO ) . Human foreskin fibroblasts ( HFFs ) were maintained in DMEM containing 7 . 5% bovine growth serum . Primary human bronchial/tracheal epithelial cells ( NHBECs ) were obtained from BioWhittaker Inc . and maintained in bronchial epithelial growth media ( BEGM , Lonza ) , and passaged according to the manufacturer's recommendations . Universal Type I Interferon was obtained from PBL InterferonSource and diluted in sterile PBS containing 0 . 1% ( w/v ) BSA . Cells were pretreated with the indicated concentration of IFN or vehicle only ( PBS plus 0 . 1% BSA ) for 24 hrs . prior to infection . The construction of a phenotypically wild-type derivative of AdEasy [81] containing the E1A and E1B genes ( AdEasyE1 ) and the introduction of the Δ2347 mutation into this background to create AdEasyE1Δ2347 have been described [82] . These viruses , Ad5 and the E1B 55 kDa-null-mutant Hr6 [79] were propagated and titered in 293 cells . To monitor viral replication , cells at <90% confluency were infected for the periods indicated , harvested and washed once with PBS , and cell pellets resuspended in 0 . 01 M Tris–HCl , pH 7 . 4 , containing 0 . 15 NaCl , 0 . 005 M KCl , 10 mM MgCl2 , and 0 . 1% ( w/v ) dextrose . Samples were freeze-thawed 4 times and debris removed by centrifugation at 13000×g for 5 minutes at 4°C . Concentrations of infectious particles units ( P . f . u . s ) were measured by plaque assay on complementing 293 cells as described [124] . Plaque assays were performed at least in triplicate , and standard deviations plotted for each data point indicate the combined propagated standard deviations of these assays and biological replicates of the experiments . Virus- or mock-infected cells were harvested 24 hrs . ( NHBECs ) or 30 hrs . ( HFFs ) after infection , washed once with PBS , and lysed in 20 mM Tris pH 7 . 5 , containing 2 mM EDTA , 0 . 15 M NaCl , and 0 . 65% ( v/v ) NP-40 . NaCl was added to a final concentration of 0 . 5 M and samples diluted with 0 . 6 volumes water pre-treated with diethyl pyrocarbonate . One volume of 2× proteinase K buffer ( 20 mM Tris pH 7 . 5 , containing 0 . 15 M NaCl , 2% ( w/v ) SDS , 2 mM EDTA ) , and 200 µg proteinase K ( NEB ) were then added , and samples incubated for 30 min at 37°C . Samples were extracted with ( 1∶1 ) phenol-CHCl3 and ethanol precipitated . Following resuspension in DNase I digestion buffer ( Roche ) , solutions were incubated with 10 U DNase I ( Roche ) at 37°C for 30 min prior to phenol:CHCl3 extraction and ethanol precipitation . RNA samples were resuspended in 10 mM Tris , pH 7 . 5 , containing 5 mM NaCl , and 0 . 5 U/µl RNasin ( Promega ) . RNA concentrations were determined from A260 reading made using a NanoDrop ND-1000 spectrophotometer . cDNA was synthesized from 1 µg of RNA by priming with 200 ng random hexamers ( Roche ) and extension with SuperScript II reverse transcriptase ( Invitrogen ) using the conditions recommended by the manufacturer . ISGs were detected by PCR with the following primers ( 5′ to 3′ ) and reaction conditions: IFIT2 primary transcript , fwd: GAGTGCAGCTGCCTGAACCGAGCC , rev: GCAACTCAACTCCCCCAGGCGTGC , 60 . 9°C annealing , 67°C extension , 32 cycles; IL6 primary transcript , fwd: GCCCACCGGGAACGAAAGAGAGC , rev: CCTGGGCCACACACCCCTCC , 59°C annealing , 66°C extension , 32 cycles; STAT1 primary transcript , fwd: CTCGACAGTCTTGGCACCTAACG , rev: CATTAAGCCCTTCCATCTTTGAACATA , 53°C annealing , 60°C extension , 25 cycles; GBP1 mRNA , fwd: GTCAACGGGCCTCGTCTAGA , rev: CCCACTGCTGATGGCAATG , 50°C annealing , 65°C extension , 30 cycles; GAPDH mRNA transcript , fwd: CTGTTGCTGTAGCCAAATTCGT , rev: ACCCACTCCACCTTTGAC , 50°C annealing , 65°C extension , 20 cycles . PCR products were resolved by electrophoresis in 8% polyacrylamide gels . Signals were quantified using Image J . NHBECs in 6-well dishes were infected with 5 p . f . u . /cell AdEasyE1or AdEasy E1Δ2347 and harvested after the periods of infection indicated . DNA was isolated from nuclei as previously described [106] . Quantitative real-time PCR was carried out using the ABI PRISM 7900HT sequence detection system with SYBR Green Master Mix ( Applied Biosystems ) to detect an amplicon within the ML transcription units , 90 base pairs long ( nucleotides 7128 to 7218 ) . The primers used were as follows: fwd: ACT CTT CGC GGT TCC AGT ACT C , rev: CAG GCC GTC ACC CAG TTC TAC . 20 µl reactions contained 2 µl sample DNA , diluted 1∶100 for ML amplification , and undiluted for the detection of genomic β-actin DNA as an internal cellular control , using the primers: fwd: TCCTCCTGAGCGCAAGTACTC , rev: ACTCGTCATACTCCTGCTT . Experiments were carried out in biological duplicate . PCR cycles were programmed as follows: two initial steps at 50°C for 2 min and 95°C for 10 min , and then 40 cycles of 95°C for 15 sec and 60°C for 60 sec . Relative DNA concentrations were determined by the standard curve method using the plasmid pTG3602 , which contains the Ad5 genome sequence , as reference standard for ML detection , and a recombinant HCMV BAC containing the genomic human β-actin sequence , ( a kind gift of Thomas Shenk ) as standard for the internal control . All qPCR measurements were performed in triplicate . Mean ML values were corrected with respect to the mean β-actin values for each sample before normalization to the 2 hrs . p . i . input value . Standard deviations plotted for each data point represent the combined propagated standard deviations of the qPCR assay and biological replicates of the experiment . NHBECs at 80–90% confluence were infected with AdEasyE1or AdEasy E1Δ2347 as described above . Cells were harvested at the times after infection indicated , washed with phosphate-buffered saline ( PBS ) , and extracted with 25 mM Tris-HCl , pH 8 . 0 , containing 50 mM NaCl , 0 . 5% ( w/v ) sodium deoxycholate , 0 . 5% ( v/v ) Nonidet P-40 ( NP-40 ) and 1 mM phenylmethylsulfonyl fluoride for 30 min at 4°C . Extracts were sonicated for a total of 30 s or incubated with 125 units Benzonase nuclease ( Sigma ) for 30 minutes at 37°C , and cell debris removed by centrifugation at 10 , 000×g for 5 min at 4°C . The extracts were analyzed by sodium dodecyl sulfate ( SDS ) -polyacrylamide gel electrophoresis and immunoblotting as described previously [125] . The E1A and E2 DBP proteins were detected with the monoclonal antibodies M73 [126] , and B6 [127] respectively . β-actin was examined using a horseradish peroxidase-conjugated monoclonal antibody ( Abcam ) to provide an internal loading control . HFFs grown to no more than 90% confluence on sterile coverslips were mock infected , or infected with AdEasyE1or AdEasy E1Δ2347 for 36 hrs . , and the cells processed for immunoflourescence as described previously [125] . To examine replication centers , the viral E2 DBP was visualized using the B6 antibody [127] and Alexa 488 anti-mouse IgG ( Invitrogen ) , and DNA was stained with DAPI ( Invitrogen ) . Coverslips were mounted on glass slides in Aqua Polymount ( Polysciences Inc . ) and images acquired using a Zeiss Axiovert 200 M fluorescence microscope and AxioVision software . Cellular Pml proteins were detected using the monoclonal antibody PAB14682 ( Abnova ) , with Alexa 488 anti-rabbit IgG ( Invitrogen ) , and the E4 Orf3 protein by using the rat monoclonal antibody 6A11 [128] and Alexa 568-conjugated goat anti-rat IgG ( Invitrogen ) . Coverslips were mounted as described above , and samples examined by confocal microscopy using a Zeiss LSM 510 confocal system . All images were organized using Adobe Photoshop 7 . 0 . HFFs grown to 80% confluency in 6-well dishes were pretreated with 500 U/ml IFN or BSA-only for 12 hrs prior to mock infection or infection for 34 hrs with 200 p . f . u . /ml of the viruses indicated . IFN treatment was resumed after adsorption . To provide a positive control , HFFs exposed to 200 µM etoposide for 34 hrs . were included in the analysis . Medium was collected and pooled with trypsinized cells . Cells were pelleted by centrifugation at 16 , 000×g for 3 min , washed once with PBS , and resuspended in 0 . 5 ml 10 mM Hepes-NaOH , pH 7 . 4 , containing , 0 . 14 M NaCl , and 2 . 5 mM CaCl2 ( binding buffer ) . AlexaFluor 488-conjugated annexin V ( 1∶20 final dilution ) ( InVitrogen ) , and propidium iodide ( 5 µg/ml final concentration ) was added to 100 µl aliquots of the cell suspensions , and incubated for 15 min at room temperature . Volumes were brought up to 500 µl with binding buffer , and samples were analyzed by flow cytometry using a BD LSRII Multi-Laser Analyzer . Experiments were carried out in biological duplicate . Terminal deoxynucleotidyl transferase dUTP nick end labeling ( TUNEL ) assays were performed using a Click-iT TUNEL AlexaFluor 488 Imaging Assay kit ( Invitrogen ) . HFFs grown on coverslips were pretreated with 500 U/ml IFN for 12 hrs and were mock-infected or infected for 34 hrs . with 200 p . f . u . /ml of the virus indicated , and IFN treatment resumed after adsorption . As a positive control , uninfected HFFs were treated for 34 hrs . with 200 µM etoposide . Cells were fixed , and TUNEL reactions and staining were performed exactly according to the manufacturer's protocol . DBP and DNA staining ( Hoechst , Invitrogen ) were performed after TUNEL as described above , except that the secondary antibody used to visualize DBP was AlexaFluor 555-conjugated anti-mouse IgG ( Invitrogen ) .
|
The most frequently used therapeutic vectors for gene transfer or cancer treatment are derived from human adenovirus type 5 ( Ad5 ) . We have observed previously that the E1B 55 kDa protein encoded by a gene routinely deleted from these vectors represses expression of numerous cellular genes regulated by interferon ( IFN ) α and β , which are important components of the innate immune response to viral infection . We therefore compared synthesis of pre-mRNA from IFN-inducible genes , viral yields and early reactions in the infectious cycle in normal human cells exposed to exogenous IFN and infected by wild-type or E1B 55 kDa null-mutant viruses . We report that the E1B 55 kDa protein is a potent repressor of expression of IFN-regulated genes , and protects viral replication against anti-viral actions of IFN by blocking inhibition of formation of viral replication centers and genome replication . These observations provide the first information about the function of the transcription repression activity of E1B during the infectious cycle . Importantly , they also suggest new design considerations for adenoviral vectors that can circumvent induction of innate immune responses , currently a major therapeutic limitation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"viral",
"immune",
"evasion",
"immunity",
"virology",
"innate",
"immunity",
"immunology",
"biology",
"microbiology",
"immune",
"response"
] |
2012
|
The Human Adenovirus Type 5 E1B 55 kDa Protein Obstructs Inhibition of Viral Replication by Type I Interferon in Normal Human Cells
|
Human metapneumovirus ( HMPV ) , a member of the Paramyxoviridae family , is a leading cause of lower respiratory illness . Although receptor binding is thought to initiate fusion at the plasma membrane for paramyxoviruses , the entry mechanism for HMPV is largely uncharacterized . Here we sought to determine whether HMPV initiates fusion at the plasma membrane or following internalization . To study the HMPV entry process in human bronchial epithelial ( BEAS-2B ) cells , we used fluorescence microscopy , an R18-dequenching fusion assay , and developed a quantitative , fluorescence microscopy assay to follow virus binding , internalization , membrane fusion , and visualize the cellular site of HMPV fusion . We found that HMPV particles are internalized into human bronchial epithelial cells before fusing with endosomes . Using chemical inhibitors and RNA interference , we determined that HMPV particles are internalized via clathrin-mediated endocytosis in a dynamin-dependent manner . HMPV fusion and productive infection are promoted by RGD-binding integrin engagement , internalization , actin polymerization , and dynamin . Further , HMPV fusion is pH-independent , although infection with rare strains is modestly inhibited by RNA interference or chemical inhibition of endosomal acidification . Thus , HMPV can enter via endocytosis , but the viral fusion machinery is not triggered by low pH . Together , our results indicate that HMPV is capable of entering host cells by multiple pathways , including membrane fusion from endosomal compartments .
Human metapneumovirus ( HMPV ) , first isolated in 2001 [1] , is a leading cause of lower respiratory infection in infants and children worldwide [2–13] . Similar to the closely related respiratory syncytial virus ( RSV ) , HMPV causes inflammation , sloughing , and necrosis of the airway epithelium [14] . Despite a significant burden to human health , there is limited knowledge about how HMPV initiates infection of airway epithelial cells . All enveloped viruses must merge viral and cell membranes to establish infection . Paramyxovirus membrane fusion is thought to occur at the plasma membrane , largely based on observations that paramyxoviruses fuse in a pH-independent manner and often induce cell-cell fusion or syncytia formation in cell culture . In general , enveloped viruses are divided into two types , those for which membrane fusion is triggered by low pH and those that fuse at neutral pH , presumably at the plasma membrane . For influenza virus and vesicular stomatitis virus ( VSV ) , a drop in pH triggers conformational changes in the viral fusion proteins [15 , 16]; thus , endosomal entry and acidification are required for productive infection . In contrast , paramyxoviruses and most retroviruses are resistant to ammonium chloride , a weak base that blocks vacuolar acidification , suggesting that these viruses induce membrane fusion at neutral pH and do not require endocytosis . However , there is evidence that while capable of mediating fusion at the cell surface , HIV-1 also is capable of productively entering cells via endocytosis and fusing with endosomes in a pH-independent manner [17–20] . Thus , pH-independent virus fusion can occur either at the cell surface or after internalization into endosomes , and resistance to acidification inhibitors does not necessarily indicate where virus-cell membrane fusion occurs . Receptor engagement on the cell surface can influence virus entry mechanisms [21] . Paramyxovirus binding to cell surface receptors is thought to induce conformational changes in the fusion ( F ) protein that drive virus fusion at the plasma membrane [22] . However , several recent studies have reported evidence for endocytic entry of RSV [23–25] . The HMPV F protein contains a conserved arginine-glycine-aspartate ( RGD ) motif that serves a critical function for infection by engaging RGD-binding integrins at the cell surface , utilizing them as entry receptors [26 , 27] . However , HMPV virus-cell fusion is not triggered by integrin engagement [27] . Because integrin engagement by other viruses is known to induce endocytosis [28] , we speculated that HMPV may engage RGD-binding integrins as a means of internalization . In contrast to other paramyxoviruses that rapidly initiate virus fusion with cells [29 , 30] , HMPV exhibits a delay in membrane fusion with adherent cells [27] . Moreover , HMPV infection is impaired by chemical inhibitors of endocytosis [31] . Thus , we investigated whether the delay in HMPV fusion kinetics is related to virus internalization and sought to identify the cellular site of virus-cell fusion . To determine the entry pathway of HMPV into human bronchial epithelial cells , we developed fluorescent fusion and endosomal content mixing assays . We complemented these with inhibitors , RNAi , and confocal microscopy to study binding , entry , fusion , and infectivity . Our results demonstrate that HMPV is capable of entering and infecting host cells in a pH-independent manner and a significant proportion of virions do not fuse at the plasma membrane , but are internalized by clathrin-mediated endocytosis in a dynamin-dependent mechanism that is followed by virus fusion within endosomes .
We used an R18-dequenching assay to monitor HMPV fusion kinetics in real time [27] . This approach measures lipid mixing during virus fusion . Self-quenched R18 dye inserted into the virus membrane dilutes into cell membranes with a concomitant increase in R18 fluorescence . This assay allowed us to investigate the kinetics of HMPV fusion with adherent human bronchial epithelial ( BEAS-2B ) cells . PIV5 and VSV were used as controls . Sucrose-purified , R18-labeled virus was bound to cells on ice , and fluorescence was monitored after shifting to 37°C . HMPV fusion began after a discernable delay; R18 fluorescence did not increase until ~15 min after virus fusion was initiated at 37°C ( Fig 1A ) . Moreover , R18 fluorescence decreased significantly for 10 to 15 min before a sharp linear increase was observed . R18 fluorescence increased continuously for about 2 h , but most rapidly between 15 and 60 min . VSV , which enters cells via endocytosis [32 , 33] , exhibited an early decrease in R18 fluorescence and similar delayed fusion kinetics to HMPV . In contrast , during parainfluenza virus 5 ( PIV5 ) entry , which fuses at the plasma membrane [29] , the increase in R18 fluorescence started in ~5 min and increased linearly for 2 h . Thus , the initiation of HMPV fusion and the overall kinetics were more similar to VSV , a virus that requires endocytosis during entry , than the related PIV5 that fuses at the cell surface . These observations led us to hypothesize that the initial delay and decreased R18 fluorescence observed during HMPV fusion was due to virus internalization . If HMPV is internalized before membrane fusion , we reasoned that virus enclosed within endosomes should not be neutralized by F-specific antibodies added to the cell surface . To test this hypothesis , we added HMPV neutralizing antiserum to cells at different times after virus attachment . HMPV escape from antibodies added to the cell surface ( Fig 1B ) was more rapid than virus fusion kinetics ( Fig 1A ) . We found that the majority ( 60% ) of infectious virus escaped neutralization by 10 min postbinding ( Fig 1B ) . We observed no increase in R18 fluorescence during the first 10 min of virus entry ( Fig 1A ) , suggesting that escape from neutralization was not due to virus fusion . After the first 10 min of virus entry , ~20% of the remaining bound virus particles were neutralized by antibodies at the cell surface , although neutralization kinetics were slow ( Fig 1B ) . This observation may reflect a saturation of particle internalization , where not all infectious particles are internalized immediately , or may represent a subset of HMPV particles that are capable of fusing at the cell surface . Based upon these results , we concluded that some infectious virus particles were quickly internalized into bronchial epithelial cells , which allows these particles to escape neutralization before virus fusion occurs with an intracellular membrane . To test whether HMPV particles are internalized during entry , we investigated the cellular distribution of HMPV particles during a time course of cell entry . After HMPV binding at 4°C , BEAS-2B cells were incubated at 37°C for various intervals before cells were fixed and stained for HMPV particles as described in Experimental Procedures . Briefly , we used a differential F-specific antibody staining protocol to distinguish HMPV F protein on the cell surface ( green or yellow in the merged images ) from that on intracellular membranes ( red in merged images ) . Representative images show F protein detected at the cell surface ( Fig 1C , green or yellow ) and on intracellular membranes ( Fig 1C , red ) . We focused on the first hour of virus entry and reasoned that fusion at the plasma membrane would result in predominantly F staining at the cell surface , while internalization before fusion should result in F protein contained in endosomes . We imaged cells using confocal microscopy , enumerated total HMPV particles per cell ( Figs 1C and S1 ) , and determined the percentage of internalized particles at each time point ( Fig 1D–1F ) . We observed a loss of ~50% of total particles between 20 and 30 min ( Fig 1D ) . The timing corresponds to after HMPV escaped neutralizing antibodies at the cell surface ( Fig 1B ) and a linear increase in R18 fluorescence intensity in the fusion assay ( Fig 1A ) . Moreover , the significant loss of total particles at 30 min ( Fig 1D ) correlated with a significant increase in the amount of endosomal F protein detected by differential staining at 30 min ( Fig 1E ) . These results show that HMPV particles are internalized during infection . We observed peak internalized virus at 30 min , regardless of the multiplicity of infection ( MOI ) ( Fig 1E and 1F ) . From 30 to 60 min , internal particle number decreased while total particle number remained largely the same ( Fig 1D and 1E ) . This result is consistent with endosome maturation and F protein degradation during virus entry . The data also indicate that the increase in R18 fluorescence observed from 30 to 60 min ( Fig 1A ) might result from virus fusion at the plasma membrane . Although we always detected a sharp increase in internalized F protein at 30 min , the majority of F protein remained at the cell surface during the first 4 h of infection . F protein staining at the cell surface varied in area . It was not possible to distinguish an aggregate of unfused virus from a site where virus fused at the plasma membrane because the F-specific antibodies do not distinguish between prefusion and postfusion F protein . The percentage of total particles detected inside the cell at 30 min was inversely related to MOI ( Fig 1F ) . This is consistent with virus saturation of the endocytosis pathway , i . e . a smaller percentage of HMPV can be internalized in 30 min when too many particles are bound at the cell surface . However , at high MOI HMPV may enter cells by fusing at different membrane sites , e . g . , plasma membrane and endosomes . These experiments suggest that HMPV entry likely occurs by multiple pathways and that a percentage of infectious HMPV particles are internalized into cells before virus fuses with endosomes . To visualize the site of HMPV-cell fusion , we developed a confocal microscopy assay to monitor HMPV internalization and fusion ( Fig 2A ) . We modified a method used by [17] to distinguish between fusion at the plasma membrane or at an intracellular site . We co-labeled virus particles with a red , diffusible cytoplasmic dye and the self-quenching membrane DiD dye ( DiD-Red-MPV ) . For endosome fusion , we expected that unfused virus particles ( red only ) would transition during virus fusion at the time of lipid mixing ( blue + red ) , and then virus content would be released from the DiD+ endosome ( blue only fluorescence ) . We monitored DiD-Red-MPV entry into BEAS-2B cells; images from representative time points are shown ( Fig 2B ) . After attachment , red virus particles were visible at the cell surface ( Fig 2B , t = 0 min ) . Over time , we observed the expected transition of red-only virus particles to DiD+ endosomes . DiD+ endosomes containing the red virus content marker were observed during the first 2 h of virus entry ( Fig 2B , t = 30 min , arrows ) , indicating that these viruses had initiated but not completed membrane fusion within the endosome . DiD+ endosomes without red fluorescence were also observed . By 4 h , very few red viruses and mostly DiD+ endosomes without red virus content dye were present in cells ( Fig 2B , t = 240 min ) . Lipid mixing at the plasma membrane would be expected to result in no DiD+ endosomes or in the appearance of blue fluorescence at the plasma membrane , neither of which was observed . Thus , the appearance of cytoplasmic dual-color and DiD+ endosomes indicates that HMPV fusion occurs within endosomes . Fusion events were quantified by counting the number of red , blue , or dual-color puncti in multiple cells . The percentage of puncti with DiD fluorescence increased rapidly over time , and the data shown in Fig 2C define the kinetics of virus-endosome membrane fusion during entry . Intracellular HMPV fusion events ( Fig 2C ) occurred with approximately the same kinetics as total virus entry was neutralized by antibodies at the cell surface ( Fig 1B ) . This finding suggests that productive entry of a substantial proportion of infectious particles involves virus fusion inside endosomes . DiD+ endosome number increased from 10 to 60 min , and no DiD+ endosomes were observed in the presence of neutralizing antibodies ( Fig 2D ) . The number of unfused HMPV particles decreased over time ( red circles ) , while the number of DiD+ endosomes ( blue squares ) increased over time ( Fig 2E ) . Unfused virus particle number changed significantly between 5 min and 1 h , most rapidly during the first 30 min of entry . DiD+ endosome number increased linearly during the first hour , most rapidly between 5 and 20 min . Thus , unfused virus disappeared with roughly the same kinetics as DiD+ endosomes appeared in the cell . DiD+ endosomes with and without red virus content dye were observed ( Fig 2F ) . Fig 2F shows the number of dual-color ( purple triangles ) and DiD+ only ( blue squares ) endosomes visible in cells over time . Dual-color and DiD+ only endosomes increased in number with slightly different kinetics . Dual-color endosomes , representing fusion intermediates , increased most rapidly between 5 and 20 min; the steepest slopes were observed in the first 5 min and between 10 and 20 min . After 20 min , dual-color endosomes increased and decreased in a cyclic pattern . DiD+ only endosome number increased according to a logarithmic growth model , most rapidly in the first 10 min of entry followed by a slower rate . These observations are consistent with the following model for HMPV intracellular fusion: virus particles are rapidly internalized and fuse with endosomes; new fusion events result in a dual-color endosomes that convert to DiD+ only endosomes when virus content is delivered into the cell; and finally , these DiD+ endosome membranes traffic inside the cell to common endosome compartments . Collectively , the results indicate productive entry of HMPV particles can occur via internalization and fusion inside endosomes . HMPV F attachment to RGD-binding integrins mediates virus binding and is required for a subsequent entry step after fusion , but before virus transcription , and full infectivity [27] . To determine whether RGD-integrins are required to complete membrane fusion and deliver virion contents into the cytoplasm , BEAS-2B cells were incubated with function-blocking integrin-specific mAbs before virus was adsorbed at 4°C and cells warmed to 37°C . Fusion was quantified using the DiD-Red-MPV microscopy assay . As a negative control , we used an α2 integrin-specific blocking mAb , which binds to collagen-binding integrins . To block all available RGD-integrins , we used a combination of anti-αV , anti-α5 , and anti-β1 mAbs . Representative images from these experiments are shown ( Fig 3A ) . HMPV binding was reduced by ~50% ( Fig 3B ) and fewer particles fused over time ( Fig 3C ) , in the presence of RGD-integrin-specific mAbs . However , while fewer particles underwent fusion , the rate of hemifusion was similar to that observed for the anti-α2 mAb treated cells ( Fig 3D ) , consistent with previous results [27] . Interestingly , hemifused particles appeared to be trapped near the cell periphery during RGD-integrin blockade and incapable of completing membrane fusion ( Fig 3A , arrows ) . After 4 h , significantly fewer blue vesicles ( indicative of complete virus fusion ) were observed during RGD-integrin blockade compared with anti-α2 mAb treated cells ( Fig 3A , arrow heads , and 3E ) . Concordantly , treatment with the α2-specific mAb did not alter infectivity , while blocking RGD integrins results in ~90% inhibition of HMPV infectivity [34] . These results suggest that RGD-binding integrin engagement is required for efficient completion of membrane fusion , whether at the cell membrane or with intracellular vesicles , and that RGD-binding integrin blockade reduces infectivity by inhibiting entry . Actin dynamics are essential for most known cellular internalization pathways [35] . To determine whether actin function was required for HMPV entry , we treated cells with well-characterized inhibitors of actin polymerization or depolymerization and measured HMPV infectivity or internalization . Pretreatment of BEAS-2B cells with latrunculin A did not affect HMPV binding ( Fig 4A ) . However , we found that latrunculin A ( Fig 4B ) , jasplakinolide ( S2A Fig ) , and cytochalasin D ( S2B Fig ) treatment impaired HMPV infection in a dose-dependent manner . Actin-based cytoskeletal rearrangements appeared to be required for HMPV entry because treatment with latrunculin A at 4 h postbinding did not significantly impair infectivity , except at the highest dose ( Fig 4C ) . After 30 min , intracellular vesicles containing HMPV particles in untreated ( Fig 1C ) or vehicle treated ( Fig 4D ) cells were diffusely located within the cytoplasm . Latrunculin A treatment reduced the total number of HMPV particles internalized at 30 min ( S2C Fig ) . Moreover , internalized particles were located in vesicles that remained attached to the plasma membrane ( Fig 4E ) . These results suggest that actin polymerization is required for vesicles containing internalized HMPV particles to traffic away from the plasma membrane into the cytoplasm and initiate infection . ( C ) The effect of latrunculin A treatment on HMPV infectivity when added to cells at 4 h postbinding . ( D-G ) Cells pretreated with DMSO ( vehicle; D ) , latrunculin A ( E ) , anti-α2 mAb ( 20 μg/mL; F ) , or RGD-integrin function-blocking mAbs [anti-αV ( 20 μg/mL ) plus anti-α5 ( 7 μg/mL ) plus anti-β1 ( 7 μg/mL ) ] ( G ) were incubated with HMPV ( MOI = 1 ) at 4°C and transferred to 37°C for 30 min before fixation . Cells were processed for confocal microscopy as described in Experimental Procedures and images of single Z-planes are shown . Green ( yellow in merge ) particles are on the cell surface , red only particles are internal , and actin is shown in white . White bar in image indicates 10μm . Results ( mean ± SEM ) in panels A-C and representative images in panels D-G are from 3 experiments . * p <0 . 05 , ANOVA with Dunnett’s test using DMSO as the reference . To determine whether RGD-binding integrins are required for HMPV internalization , we incubated BEAS-2B cells with function-blocking integrin-specific mAbs before HMPV binding and visualized HMPV internalization at 30 min post-adsorption . Integrin blockade did not abolish HMPV particle internalization ( Fig 4G ) . However , the distribution of internalized HMPV particles at 30 min postbinding was consistently altered by RGD-binding integrin blockade . In the presence of control α2 integrin blocking mAb , vesicles containing internalized HMPV particles were located diffusely throughout the cytoplasm ( Fig 4F ) . In contrast , blocking RGD-binding integrins led to retention of HMPV particles in intracellular vesicles that remained close to the plasma membrane ( Fig 4G ) . Furthermore , blocking RGD-binding integrins with mAbs inhibits HMPV infection by ~90% [34] . These results suggest that RGD-binding integrins are required for vesicle trafficking during HMPV entry . Taken together with the latrunculin A experiments , these results indicate that vesicles containing HMPV particles are internalized from the plasma membrane in an actin- and RGD-integrin-dependent manner , and that this trafficking leads to productive infection . Next , we sought to elucidate the cellular mechanism of HMPV internalization by human bronchial epithelial cells . In Vero cells , HMPV infection is sensitive to chlorpromazine treatment [31] . Chlorpromazine treatment prevents the formation of nascent endosomes and inhibits clathrin-mediated endocytosis [36] . To determine whether chlorpromazine impairs virus entry in BEAS-2B cells , we tested the effect of chlorpromazine on HMPV internalization and fusion . Chlorpromazine treatment restricted HMPV particles to the cell surface , significantly impairing internalization ( Fig 5A and 5B ) , and increased HMPV particle susceptibility to neutralizing antibodies added to the cell surface ( Fig 5C ) . These results suggest that arrest of HMPV particle internalization impairs infection . To determine whether chlorpromazine diminished HMPV fusion , we treated cells with chlorpromazine and measured R18-MPV binding and fusion . Pre-treatment of cells with chlorpromazine modestly impaired virus binding and fusion ( S3 Fig ) . To eliminate the possibility that chlorpromazine treatment was partially preventing virus attachment by reducing HMPV receptor expression at the cell surface , we added chlorpromazine after incubating cells with R18-MPV and confirmed that addition of the inhibitor postbinding did not decrease the amount of bound virus ( Fig 5D ) . Chlorpromazine treatment after virus binding led to significantly delayed HMPV fusion , diminished fusion rate , and reduced fusion extent at 4 h ( Fig 5E and 5F ) . Furthermore , chlorpromazine treatment ( 50 μM ) reduced HMPV infectivity by ~90% ( Fig 5G ) . These results indicate that chlorpromazine inhibits HMPV infectivity by preventing virus internalization , resulting in diminished virus fusion and infection . Results from the chlorpromazine experiments suggested that HMPV was internalized by clathrin-mediated endocytosis ( CME ) . To directly test whether cellular uptake was mediated by the CME pathway , we performed an siRNA screen targeting genes involved in common cell entry pathways . We transfected BEAS-2B cells with a non-targeting ( Scramble ) siRNA or siRNA pools specific for the genes listed in Table A in S1 Text . Reduced target protein expression levels were confirmed by immunoblotting only for genes with minimal or no effects on HMPV fusion; representative blots are shown ( S4 Fig ) and the quantification is shown ( Table A in S1 Text ) . After 72 hours , cells were inoculated with HMPV , PIV5 , or VSV . At 20 h post-adsorption , the percentage of infected cells was quantified by flow cytometry . PIV5 is thought to mediate fusion at the plasma membrane [37] , while VSV enters cells via clathrin-mediated endocytosis and requires endosomal acidification to trigger fusion [16 , 38] . Along with the VSV infection control , we also tested whether siRNA treatment affected the uptake of fluorescently labeled transferrin , which is endocytosed by CME in a dynamin-dependent mechanism [39] . The siRNAs targeting endocytosis components Eps15 , dynamin-1 , dynamin-2 , dynamin-3 , and clathrin heavy chain significantly impaired HMPV infection ( Fig 6A ) . Reduction in HMPV infectivity by clathrin heavy chain knockdown was similar to that observed for both VSV infectivity ( Fig 6C ) and transferrin uptake ( Fig 6D ) . Unexpectedly , we observed a reduction of PIV5 infectivity with dynamin-1 knockdown ( Fig 6B ) . Because PIV5 or HMPV F protein expression at the cell surface was used to identify virus-infected cells , siRNA treatment could have either blocked virus entry and subsequent infection of a single cell or impaired F protein trafficking to the plasma membrane of an infected cell . To distinguish between these two possibilities , we quantified cell-surface viral F protein expression on infected cells by flow cytometry . The level of F expression on siRNA-transfected , HMPV-infected cells was similar to the control siRNA ( Fig 6E ) , suggesting that HMPV F trafficking was not affected by these siRNAs . Thus , while Eps15 , dynamin-1 , dynamin-2 , dynamin-3 , and clathrin heavy chain siRNAs inhibited HMPV entry and diminished infection , cells that did become infected exhibited normal amounts of surface F protein . In contrast , PIV5 F expression on cells transfected with either caveolin-1 or dynamin-1 siRNAs was consistently reduced compared to the Scramble siRNA ( Fig 6F ) , suggesting that siRNA treatment affected PIV5 F protein trafficking to the cell surface rather than virus entry . To confirm this hypothesis , we stained for intracellular expression of the PIV5 phosphoprotein ( P ) . Neither the caveolin-1 or dynamin-1 siRNAs significantly altered P expression ( Fig 6G ) , suggesting that the siRNAs did not impair the initiation of PIV5 infection but altered F expression on the surface of an infected cell . Some siRNAs were associated with increased surface expression of HMPV or PIV5 F ( Fig 6E and 6F ) , likely through perturbation of normal trafficking . As a complementary approach to dynamin-specific siRNA , we tested whether dynasore hydrate , a small molecule inhibitor of dynamin [40] , diminished HMPV infection . Pre-treatment of BEAS-2B cells with dynasore significantly impaired both R18-MPV binding and fusion ( S5A–S5C Fig ) . However , dynasore treatment post-adsorption did not reduce the amount of bound virus ( S5D Fig ) or time of onset of fusion ( S5E Fig ) but blocked the extent of fusion in a dose-dependent fashion ( S5E and S5F Fig ) . Furthermore , dynasore treatment ( 50 μM ) of BEAS-2B cells inhibited HMPV infection by ~70% ( S5G Fig ) , greater than dynasore treatment of Vero cells [31] . These results suggest that dynamin-dependent endocytosis facilitates virus entry into BEAS-2B cells . However , some virus appears to fuse and initiate infection in a dynamin-independent manner , likely at the plasma membrane . The siRNA targeting caveolin-1 did not affect HMPV infection ( Fig 6A ) . The siRNAs targeting PKCα , which is required for macropinocytosis , did not impair HMPV infectivity ( Fig 6A ) . Moreover , treating BEAS-2B cells with EIPA , a chemical inhibitor of macropinocytosis , did not alter HMPV infectivity under conditions where dextran uptake was significantly impaired ( S6 Fig ) . These results suggest that HMPV internalization does not occur by caveolin-mediated uptake or macropinocytosis . Thus , siRNA and inhibitor results indicate that HMPV internalization by CME leads to productive infection . Next , we sought to determine whether HMPV internalization and fusion required components of the CME pathway . We transfected BEAS-2B cells with a non-targeting ( Scramble ) siRNA or siRNA pools targeting caveolin-1 , clathrin heavy chain , dynamin-1 , or Eps15 . At 72 h post transfection , we quantified HMPV particle internalization after 30 min using confocal microscopy ( Fig 7A–7F ) . HMPV internalization was significantly impaired in cells transfected with the clathrin heavy chain , dynamin-1 , or Eps15 siRNAs but not the caveolin-1 siRNA ( Fig 7F ) . None of the siRNAs diminished binding ( S7A Fig ) . We also found that R18-MPV fusion was significantly impaired in BEAS-2B cells transfected with clathrin heavy chain , dynamin-1 , or Eps15 targeting siRNAs ( Figs 7G , S7B and S7C ) . These results suggest that HMPV is internalized via CME and that endocytosis in a clathrin- , dynamin- , and Eps15-dependent manner is required for fusion . Taken together with the effect of clathrin heavy chain , dynamin-1 , or Eps15 siRNAs on HMPV infectivity ( Fig 6A ) , these data indicate that CME of HMPV during entry leads to productive infection of human bronchial epithelial cells . The majority of viruses that enter cells by CME require acidic pH to activate fusion proteins [41] . ATP6VOC is a subunit of the vacuolar ATPase that mediates endosomal acidification [42] . The ATP6VOC siRNA reduced HMPV infectivity , although the effect was modest ( ~20% inhibition , Fig 6A ) and not as potent as the expected reduction in VSV infectivity ( ~70% inhibition , Fig 6C ) . HMPV F proteins from a few strains require low pH to induce cell-cell fusion , while most induce cell-cell fusion or infection at neutral pH [31 , 43–45] . To determine whether low pH is required for HMPV fusion and infection , we treated cells with endosomal acidification inhibitors bafilomycin A or ammonium chloride and quantified binding , fusion , and infectivity of HMPV and VSV . Using the same A2 genotype strain of HMPV used in the siRNA experiments , we found that neither bafilomycin A nor ammonium chloride impaired HMPV binding or fusion ( Fig 8A–8D ) . High concentrations of bafilomycin A ( 100 to 400 nM ) , which completely blocked VSV infection , inhibited infectivity of all four HMPV genotypes ( A1 , A2 , B1 , and B2 ) by 40 to 50% , but the effect was not dose-dependent ( Fig 8E ) . Ammonium chloride ( 2 . 5 to 10 mM ) impaired A2-HMPV and VSV infectivity in a dose-dependent manner , but not infection by other genotypes of HMPV ( Fig 8F ) . These results suggest that exposure to low pH may enhance HMPV infectivity for some strains , but is not required for triggering virus fusion .
Paramyxovirus binding to cell-surface receptors is thought to induce conformational changes in the F protein that trigger virus fusion at the plasma membrane . Our data suggest that HMPV binding alone is not sufficient to trigger fusion , as we consistently observed a ~15–20 min delay before the onset of R18-MPV fusion at 37°C . The fusion kinetics of HMPV in our assay were very similar to those of VSV and notably slower than the fusion of PIV5 . Other studies examining the fusion kinetics of R18-labeled viruses provide evidence that a fusion lag correlates with entry mechanisms that separate binding from fusion by an internalization event . For example , VSV requires endocytosis before fusion , and R18-VSV fusion occurs after a discernable delay [32] . In contrast , R18 dequenching of PIV5 and Sendai virus , which are thought to fuse at the plasma membrane , initiates nearly immediately and reaches a plateau within 20 min [29 , 30] . Thus , the lag in HMPV fusion resembles that seen for a virus that enters cells before fusing with endosomal membranes , and indeed the fusion kinetics of HMPV were similar to those of VSV in this assay . Concordantly , we found that HMPV particles are internalized by human bronchial epithelial cells and escape antibody neutralization at the cell surface during the first 10 min postbinding , before the initiation of fusion . We found that HMPV particles are internalized and transfer a fluorescent membrane dye to intracellular vesicles before releasing a virus content dye into the cytoplasm . Collectively , these data show that a proportion of HMPV particles are internalized before virus fusion . Thus , the site of HMPV fusion is not exclusively at the cell surface , as predicted by analogy to other paramyxoviruses . In addition , HMPV is capable of being internalized by clathrin-mediated endocytosis in a dynamin- , Eps15- , and actin-dependent manner . Interestingly , recent data suggest that RSV fuses with intracellular vesicles after internalization by macropinocytosis [24] and caveolin-mediated entry has been suggested for Newcastle disease virus [46 , 47] . Thus , within the paramyxovirus family , HMPV and RSV are capable of using endocytic pathways during entry in addition to fusing at the cell surface . It is important to note that HMPV and other enveloped viruses likely are capable of entering cells through more than one pathway . Similarly , influenza virus can enter cells both by CME and macropinocytosis [48–50] . Notably , inhibition of one pathway may enhance entry by an alternate pathway , and entry mechanisms of some viruses may be cell type-specific [21 , 51] . RGD-binding integrins are entry receptors for HMPV , serving both to engage the F protein during virus binding and a postbinding role during entry [26 , 27] . Reverse-engineered HMPV strains with mutations in the F RGD motif exhibited reduced cell-cell fusion , diminished replication in vitro , and attenuation in rodents , confirming a role for RGD integrin engagement in vivo [52] . However , integrins are not the only receptors for HMPV , as HMPV has been shown to bind to heparan sulfate via the F protein [53] . Moreover , blocking RGD-binding integrins reduced binding by ~50% but infectivity by ~90% [27] . Thus , the precise function of integrins in HMPV entry was not known . Here , we show that RGD-binding integrins are required for functional HMPV internalization . Blocking integrin engagement during binding does not prevent HMPV particle internalization , but endosomes containing HMPV particles remain in close proximity to the cell surface and do not lead to productive infection . Although HMPV hemifusion can initiate in vesicles trapped at the cell surface , we present three lines of evidence to suggest that these vesicles traffic elsewhere in the cell for complete fusion and subsequent infection . First , blocking RGD-integrin engagement results in intracellular HMPV particles that are retained at the cell surface ( Fig 4 ) , can initiate hemifusion but cannot complete membrane fusion ( Fig 3 ) , and do not productively infect cells [34] . Second , inhibiting actin polymerization impedes intracellular HMPV particle trafficking and infection ( Figs 4 and S2 ) . Third , dynasore treatment , which inhibits intracellular vesicle release from the plasma membrane , does not completely inhibit HMPV fusion but significantly impairs infectivity ( S5 Fig ) . Thus , HMPV appears to engage RGD-binding integrins to gain access to an intracellular compartment where fusion occurs . Integrins are adhesion receptors that bind extracellular proteins and associate with cytoskeletal proteins , adaptors , and kinases , allowing them to transduce bidirectional signals between intra- and extra-cellular environments [54] . The association of integrins with cell-signaling cascades and endosomal sorting pathways makes them a common receptor for mammalian viruses , including adenovirus , hantavirus , herpesvirus , picornavirus , and reovirus ( reviewed in [28] ) . Integrin engagement and integrin-mediated signaling are required for internalization of viral and bacterial pathogens including adenovirus [55] , simian virus 40 ( SV40 ) [56] , reovirus [57] , Yersinia species [58] , Staphylococcus aureus [59] , and Neisseria meningitidis [60] . Our results suggest that HMPV F engagement of RGD-binding integrins during attachment leads to virus internalization and productive infection . Whether HMPV binding induces integrin-mediated signaling that influences actin-dependent trafficking of HMPV particles requires further investigation . Virus endocytosis during entry for many viruses is correlated with a requirement for low pH exposure . Low pH either directly activates virus fusion proteins or is required for enzymes that activate fusion proteins [61] . In our experiments , neither ammonium chloride nor bafilomycin A treatment altered HMPV fusion kinetics , suggesting that HMPV fusion is triggered by a pH-independent mechanism . Others found modest and non-dose-dependent effects of ammonium chloride , bafilomycin A , and concanamycin A on HMPV infectivity [31 , 45] . Inhibition of vacuolar ATPase ( V-ATPase ) modestly reduced HMPV infection in a non-dose dependent manner; since disrupting V-ATPase activity also impairs endosomal trafficking [41 , 62] , the reduction in HMPV infectivity may result from vesicles not arriving at the appropriate compartment for HMPV fusion to occur . HMPV F-mediated cell-cell fusion of a few subgroup A strains is enhanced by exposure to low pH; however , subgroup B strains fuse in a pH-independent manner [31 , 43–45] . One study found that a glycine at residue 294 determined the low pH effect in vitro [44]; however , 294G is present in only 4–6% of >500 HMPV A sequences and no HMPV B sequences [44 , 63 , 64] . A tetrad of residues at positions 294 , 296 , 396 , and 404 in the F protein ( GKRN ) conferred a low-pH-dependent cell-cell fusion phenotype to HMPV from all lineages [45] . We used plaque-purified , fully sequenced prototype strains from each subgroup in these experiments [63] . The A strains we tested here encode EKRN at these four residues , while the B strains encode ENRP . Thus , none of the strains we tested for pH dependence encode the tetrad of residues associated with low pH enhancement of fusion . The majority of sequenced HMPV F genes from circulating strains do not encode this tetrad [44 , 45 , 64] , indicating that low pH enhancement of fusion is an uncommon and strain-specific phenomenon . This does not exclude the possibility that low pH may contribute to entry of some strains by destabilization of the F protein via protonation of histidine residues predicted to lie near the GKRN tetrad [31 , 45 , 65] . Why does HMPV use the endocytic pathway ? A clue may come from studies of RSV that indicate the F protein can be triggered by changes in ionic strength [66] . The ionic nature of endosomes varies considerably as endosomes mature , with significant changes in the endosomal concentration of Ca2+ , Cl- , H+ , K+ , and Na+ , among others [67 , 68] . HMPV F may use the ionic constituents in endosomes as a fusion trigger . It is also possible that diminished Ca2+ ion concentrations in early endosomes lead to reduction in integrin affinity that induce conformational changes in the bound HMPV F protein , as integrins require divalent cations to maintain their active conformation [54] . Other possible reasons for HMPV to utilize the endosomal pathway for entry could be endosomal proteases , as for the related Hendra and Nipah viruses [69 , 70] or bypassing the dense cortical actin cytoskeleton [71–73] . As noted , for at least some rare strains , lower pH may contribute to fusion triggering but is not required . Our results show that HMPV can be internalized via clathrin-mediated endocytosis and that virus fusion initiates and completes within intracellular vesicles . While endosomal pH may affect the efficiency of HMPV infection for rare strains , the HMPV F protein does not require exposure to low pH to initiate fusion and infection . Thus , it appears that HMPV entry is pH-independent . Our data suggest that HMPV is capable of entering human bronchial epithelial cells by endocytosis as well as fusion at the plasma membrane; these results do not confirm which pathway , if either , is preferred during natural human infection . The key to understanding F triggering inside endosomes might be to define the nature of the compartment where fusion occurs . Endosomes are dynamic vesicles that could provide a variety of unique signals that might influence fusion protein activity . Future discoveries about how HMPV F protein enters and initiates membrane fusion will potentially identify new therapeutic targets .
LLC-MK2 ( ATCC CCL-7 ) and BEAS-2B ( ATCC CRL-9609 ) cells were maintained in Opti-MEM I medium containing 2% fetal bovine serum ( FBS ) . Suspension 293-F cells were maintained as recommended by the manufacturer ( 293 Freestyle expression system; Invitrogen ) . BHK-21 ( ATCC CCL-10 ) , caveolin-1 wt ( ATCC CRL-2752 ) and Cav-1-/- ( ATCC CRL-2753 ) MEFs were maintained in DMEM medium containing 10% FBS . HMPV strain TN/94-49 ( subgroup A2 ) was used for all experiments , except to determine pH sensitivity of infection when TN/96-12 ( subgroup A1 ) , TN/98-242 ( subgroup B1 ) , and TN/89-515 ( subgroup B2 ) were also used . All HMPV strains were propagated and titrated using LLC-MK2 cells as described previously [74] . R18-labeled virus ( R18-MPV , R18-PIV5 , R18-VSV ) was prepared as described previously [27] . DiD-Red-MPV was prepared by metabolically labeling HMPV-infected 293-F cells . At 12 h post virus inoculation , cells were incubated with 10 μM CellTracker Orange CMRA and 5 μM DiD for 45 min at 37°C , washed extensively to remove unincorporated dyes , resuspended in virus growth medium , and incubated at 37°C with 5% CO2 with shaking for 4 days . Virus in cell supernatant was clarified by centrifugation and sucrose-purified as described previously [27] . PIV5 ( ATCC VR-288 ) was propagated using LLC-MK2 cells in Opti-MEM I medium containing 2% fetal bovine serum ( FBS ) . VSV G-complemented VSVΔG-GFP virus was generated in BHK-21 cells as previously described [75] , and the complementation system was a generous gift from Michael Whitt . See Supplemental Experimental Procedures . At 24 hours postinoculation , BEAS-2B cell monolayers were fixed , immunostained , and infected cells were enumerated as described previously [27] . For flow cytometry experiments , HMPV-infected cells were immunostained for HMPV F surface expression and the percentage of infected cells was quantified with a BD LSRII flow cytometer . BEAS-2B cells were seeded on glass cover slips coated with a thin layer of Matrigel ( BD Biosciences ) . HMPV was adsorbed for 1 h at 4°C , unbound virus washed away , and pre-warmed ( 37°C ) cell culture medium added to initiate virus entry . Cells were incubated at 37°C for various time intervals , fixed with 5% buffered formalin , washed , and stained . HMPV particles at the cell surface were detected by staining with a polyclonal HMPV antiserum followed by anti-guinea pig IgG Alexa Fluor 647 antibody , cells were permeabilized with 1% Triton X-100 for 5 min , and total ( surface and internal ) particles were detected with anti-HMPV F mAb DS7 at 5 μg/mL followed by anti-human IgG Alexa Fluor 546 antibody . DS7 recognizes both prefusion and postfusion F conformers [76] . Cells were incubated with phalloidin Alexa Fluor 488 to detect actin , fixed on glass slides using AquaPolyMount ( Polysciences ) , and imaged by confocal microscopy . HMPV particle number and co-localization of red and blue pixels were quantified with MetaMorph using individual z-planes . The percentage of internalized particles was calculated as the number of red only particles / total particles x 100 . Images were obtained on a Zeiss inverted LSM510 confocal microscope using a 63x oil objective lens . R18-MPV ( MOI ~1 ) was adsorbed to BEAS-2B cells grown in black , transparent-bottom 96-well plates for 1 h on ice , unbound virus washed away , and binding quantified as previously described [27] . For endocytosis inhibitor experiments , dynasore hydrate and chlorpromazine hydrochloride dissolved in dimethyl sulfoxide ( DMSO ) at 10 mM were diluted into medium immediately before use . Cells were incubated with medium containing DMSO or inhibitor for 1 h at 37°C followed by 30 min at 4°C and removed during R18-MPV binding ( pretreatment ) , or R18-MPV was bound to cells before medium containing DMSO or inhibitor was added and incubated with cells for 30 min on ice ( treated postbinding ) . R18-MPV fusion was measured and quantified as previously described [27] . R18-MPV ( MOI ~1 ) was adsorbed to BEAS-2B cells grown in black 96-well plates , unbound virus washed away , ice-cold fusion medium added , plates transferred to a preheated ( 37°C ) plate reader , and R18 fluorescence monitored for 4 h with readings collected every 5 min . Endocytosis inhibitor treatments were performed as described above . To identify the cellular site of HMPV fusion , we modified a method used by [17] . The principle of the assay relies on visualizing fusion of viruses co-labeled with the lipophilic dye DiD and a red , diffusible content dye ( CellTracker Orange CMRA ) . Similar to the R18 hemifusion assay , quenched DiD dye is transferred to cell membranes during virus hemifusion . HMPV particles were labeled with a 10-fold lower concentration of DiD such that DiD contained within particles is only partially quenched . The DiD concentration is sufficient to observe DiD dequenching if virus particles fuse with a limited membrane , such as an endosomal vesicle . Fusion at the cell surface should lead to the disappearance of virus particles due to dilution of DiD into the plasma membrane and the content dye into the cytoplasm . However , virus fusion with an intracellular vesicle leads to an increase in the DiD ( blue ) fluorescence intensity , the appearance of particles that are red plus blue because they contain the content dye and become DiD-positive during hemifusion , and an appearance of blue vesicles that have lost the content dye after virus-cell membrane fusion is complete . BEAS-2B cells , grown on Matrigel-coated glass cover slips , were loaded with CellTracker Green CMFDA at 0 . 5 μM for 30 min and DiD-Red-MPV ( MOI ~0 . 2 ) was adsorbed to cells for 1 h at 4°C . Unbound virus was washed away , pre-warmed medium added , and cells incubated at 37°C for various time intervals before formalin fixation . Cover slips were mounted onto glass slides with AquaPolyMount . For inhibitor treatments , HMPV-specific neutralizing antiserum ( dilution 1:20 ) was added to the cell medium after virus binding ( Fig 2 ) . or integrin function-blocking antibodies were incubated with cells for 30 min at 4°C before virus binding ( Fig 3 ) . Z-stacks for random fields of cells were collected for each time point using confocal microscopy . The confocal imaging parameters were kept constant for each time point in an experiment for comparative analysis of fluorescence intensity . For all images , the entire z-stack was analyzed and the number of red , blue , or red plus blue particles was enumerated . HMPV ( MOI ~1 ) was adsorbed to BEAS-2B cells on ice to prevent fusion , unbound virus removed by washing with PBS , and pre-warmed ( 37°C ) medium added to initiate virus entry . Cells were incubated at 37°C for time intervals before medium supplemented with neutralizing HMPV-specific antiserum was added . Cells were incubated on ice for 30 min to allow antibody binding before cells were returned to 37°C and incubated for 20 to 24 h . HMPV-infected cells were quantified by flow cytometry as described above . For the experiment shown in Fig 5C , cells were pretreated with DMSO or chlorpromazine for 30 min at 37°C before virus binding , the treatment removed during virus binding , added to the incubation medium during virus entry , and washed away before anti-HMPV antibody binding . For the flow cytometry experiments , 1 x 105 BEAS-2B cells were reverse transfected with Lipofectamine RNAiMax ( Invitrogen ) and 1 . 5 pmol of siRNA . All siRNAs used in this study ( see Table A in S1 Text ) were siGENOME SMARTpool ( ThermoScientific ) . Knock down of target gene expression was confirmed by immunoblot ( see S4 Fig ) . Cells were infected with HMPV , G-complemented-VSVΔG-GFP , or PIV5 at 72 h post transfection to achieve ~50% infected cells for the control siRNA-treated cells . After 18 h , HMPV and PIV5 infected cells were immunostained for viral F protein expression on the cell surface , fixed and enumerated by flow cytometry . Alternatively , PIV5-infected cells were fixed , permeabilized and intracellular PIV P protein was detected by immunostaining . VSV-infected cells express GFP and were fixed and enumerated by flow cytometry . For transferrin uptake experiments , transferrin Alexa Fluor 488 ( 20 μg/mL ) was added to siRNA-transfected cells for 30 min at 37°C . Cells were detached with trypsin , treated with 30 mM citric acid buffer ( pH 4 . 5 ) for 2 min to remove residual surface transferrin , fixed , and analyzed by flow cytometry . For HMPV internalization experiments , 1 x 105 BEAS-2B cells seeded on Matrigel-coated glass coverslips were transfected with Lipofectamine RNAiMax and 10 pmol of siRNA . For R18-MPV fusion experiments , adherent BEAS-2B cells seeded in 96-well plates were transfected with Lipofectamine RNAiMax and 3 pmol of siRNA . At 72 h post transfection , HMPV ( MOI ~1 ) internalization after 30 min at 37°C or R18-MPV fusion was analyzed as described above . Statistical analyses are described in figure legends . A P value ≤ 0 . 05 was considered statistically significant .
|
Human metapneumovirus ( HMPV ) is a paramyxovirus that causes severe lower respiratory tract infections . HMPV infection is initiated by the viral surface fusion ( F ) glycoprotein . HMPV F attaches to cellular receptors , including RGD-binding integrins , and catalyzes virus membrane fusion with cellular membranes during virus entry . Although most paramyxoviruses enter cells by coupling receptor binding to membrane fusion at the cell surface , the entry mechanism for HMPV is largely uncharacterized . In this study , we sought to determine the cellular site of HMPV fusion . We show that HMPV particles are internalized by clathrin-mediated endocytosis and fuse with endosomal membranes . Furthermore , HMPV engages RGD-binding integrins for endosomal trafficking and full virus membrane fusion with intracellular membranes , suggesting that HMPV uses integrins to facilitate movement into target cells rather than as a trigger for fusion at the cell surface . Inhibition of endosomal acidification had only a modest strain-specific effect , suggesting that low pH exposure is not required for HMPV fusion . These results expand knowledge of mechanisms of HMPV entry and suggest new potential therapeutic interventions against this medically important virus .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Human Metapneumovirus Is Capable of Entering Cells by Fusion with Endosomal Membranes
|
Heme crystallization as hemozoin represents the dominant mechanism of heme disposal in blood feeding triatomine insect vectors of the Chagas disease . The absence of drugs or vaccine for the Chagas disease causative agent , the parasite Trypanosoma cruzi , makes the control of vector population the best available strategy to limit disease spread . Although heme and redox homeostasis regulation is critical for both triatomine insects and T . cruzi , the physiological relevance of hemozoin for these organisms remains unknown . Here , we demonstrate that selective blockage of heme crystallization in vivo by the antimalarial drug quinidine , caused systemic heme overload and redox imbalance in distinct insect tissues , assessed by spectrophotometry and fluorescence microscopy . Quinidine treatment activated compensatory defensive heme-scavenging mechanisms to cope with excessive heme , as revealed by biochemical hemolymph analyses , and fat body gene expression . Importantly , egg production , oviposition , and total T . cruzi parasite counts in R . prolixus were significantly reduced by quinidine treatment . These effects were reverted by oral supplementation with the major insect antioxidant urate . Altogether , these data underscore the importance of heme crystallization as the main redox regulator for triatomine vectors , indicating the dual role of hemozoin as a protective mechanism to allow insect fertility , and T . cruzi life-cycle . Thus , targeting heme crystallization in insect vectors represents an innovative way for Chagas disease control , by reducing simultaneously triatomine reproduction and T . cruzi transmission .
Chagas disease ( CD ) is a chronic and debilitating illness caused by the protozoa Trypanosoma cruzi [1] , afflicting about 6 million people predominantly in Latin America [2] . Despite low relative mortality of CD patients , development of chronic manifestations , such as digestive , cardiovascular and neurological disorders , pose a huge burden to the patients . Trypanosoma cruzi is usually transmitted to mammalians through the contact of infected feces of triatomine insects with mucosa or skin lesions [1] . Strategies designed to directly eliminate parasites in human hosts are still inefficient , making the control of the triatomine vector population the most useful method to prevent CD dissemination . Trypanosoma cruzi replicates and differentiates in the digestive tract of triatomine insects , which ingest about five times their own weight on vertebrate blood to meet their energy demands . As a result , huge amounts of toxic “free” heme are released in triatomine digestive tract [3] . On one hand , Trypanosoma cruzi lacks a functional heme biosynthetic pathway [4] , making this parasite strictly dependent on heme to support proliferation mediated by specific redox-dependent signaling pathways [5–8] . On the other hand , reactive oxygen species ( ROS ) , propagated by "free" heme affects T . cruzi differentiation to infective trypomastigote forms [7] . This suggests that T . cruzi adapted to a unique environment in the triatomine midgut where "free" heme levels are tightly regulated . Indeed , excessive levels of “free” heme are deleterious , not only to trypanosomes [7–9] , but also to triatomine vectors [3 , 10 , 11] . The endogenous release of massive amounts of heme in the triatomine midgut under normal circumstances is counteracted by a very effective array of protective adaptations to deal with this toxic molecule [3 , 12 , 13] . Heme crystallization into Hemozoin ( Hz ) represents the main protective mechanism against heme toxicity in many organisms that digest hemoglobin , such as malaria parasites , Schistosoma worms , and triatomine insects [14–16] . Indeed , heme crystallization is a very efficient mechanism for heme disposal in these organisms , accounting for about 95% of the iron derived from blood intake [14 , 15] . Hz crystals are produced by the interaction of "free" heme with amphiphilic structures , including food vacuole membranes in malaria parasites , extracellular lipid droplets in Schistosoma gut , and phospholipid membranes in triatomine insects [10 , 17–19] . Regardless of the organism model , aminoquinoline drugs can either form stable complexes with heme , and directly interact with Hz crystals [20–22] which ultimately impair Hz formation , building up heme and oxidized products levels [10 , 23–25] . As the cytotoxic properties of "free" heme were extensively explored [9 , 11 , 26 , 27] , the mechanism for the antimalarial effects of aminoquinolines are currently explained through a redox imbalance process , and the resultant molecular damage from impaired Hz formation [28] . Although the regulation of heme homeostasis is critical for both triatomine vectors [3 , 12 , 13] , and T . cruzi parasites [4–9] , the physiological significance of Hz for these organisms remains unknown . We hypothesized that pharmacological blockage of Hz formation in the triatomine insect Rhodnius prolixus might dysregulate the heme/redox homeostasis and disrupt vector/parasite physiology ultimately , with potential effects in CD transmission . To fill this gap of knowledge , we have investigated here the biochemical and physiological effects of the aminoquinoline drug quinidine ( QND ) to the triatomine insect Rhodnius prolixus . Selective blockage of heme crystallization in vivo by QND caused a systemic heme overload and redox imbalance in distinct insect tissues . QND treatment activated compensatory defensive heme-scavenging mechanisms to cope with excessive heme . Importantly , egg production and laying , and total T . cruzi parasite counts in R . prolixus were significantly reduced by QND treatment . The physiological effects of QND were partially reversed by the major insect antioxidant urate . Taken together , these results indicate that heme crystallization represents the prime redox regulator for triatomine vectors , highlighting the dual role of Hz as a protective mechanism to allow insect reproduction , and T . cruzi infection , opening new possibilities for effective CD control .
To establish optimal conditions for oral administration of QND , insect survival , blood intake and digestion were assessed using individuals from two distinct colonies ( Federal University of Rio de Janeiro , hereafter named "IBqM" , and Fundação Oswaldo Cruz-Rio de Janeiro , hereafter named "Fiocruz" ) . We observed that dietary QND supplementation at 100 μM caused no significant effects in any of these outputs ( S1 Fig ) . However , Hz formation in vivo was strikingly inhibited in the midgut of adult triatomines in a dose ( Fig 1A ) , and time ( Fig 1B and S2A Fig ) dependent manner . The highest inhibitory activity of QND on heme crystallization ( >70% ) occurred around 4 days after blood meal in both colonies ( Fig 1B and S2A Fig ) . Inhibitory effects of QND on Hz formation were also observed in young nymphs ( S2B Fig ) . Since hemoglobin digestion , heme release and crystallization take place in the lumen of posterior midgut , we postulated that this tissue would be the first line of defense against “free” heme overload . Conversely , the posterior midgut would also be the most directly affected tissue by impaired Hz formation . As the pro-oxidant and toxic effects of heme are strongly reduced upon its crystallization as Hz [26] , we anticipated that impaired Hz formation would increase heme-derived oxidant levels . Indeed , feeding adult insects with 100 μM QND shifted posterior midgut redox balance towards oxidation , as revealed by 3 . 7 folds increase in the fluorescence intensity of the oxidant-sensitive dye dihydroethidium ( DHE ) ( Fig 1C and 1D ) . Therefore , selective inhibition of Hz formation in vivo promotes posterior midgut redox imbalance . Given the higher oxidant levels resulted from impaired Hz formation in the posterior midgut , we further investigated the potential consequences to tissue and cell architecture . Histological observations of posterior midgut cross sections from QND-treated insects by light microscopy revealed that the midgut lumen was completely washed-out , with barely noticeable Hz crystals , along with extensive cytosolic vacuolization , and numerous intracellular lipid droplets in midgut cells ( Fig 2A and 2B ) . At the cellular level , QND treatment caused severe ultrastructural changes in the posterior midgut , with remarkable disappearance of organelles resembling macroautophagy , as numerous vacuoles , lipid droplets , residual bodies , and autophagosomes were detected ( Fig 2D , 2G and 2H ) . Interestingly , the mitochondria demonstrated clear structural changes in the midgut of QND-treated insects , but observations were plagued by low abundance of this organelle ( Fig 2D , 2G and 2H ) . These effects include swollen and washed-out mitochondrial matrix , suggesting increased permeability transition of the inner membrane , a pre-requisite for mitophagy [29] ( Fig 2G and 2H ) . Indeed , mitochondria were frequently observed within autophagosomes in posterior midguts of QND-treated insects , as shown in Fig 2H . Despite the massive changes in midgut architecture , the homogeneous and pronounced distribution of microvilli along luminal epithelial border ( see "asterisks" on Fig 2C , 2D and 2E ) , and the preserved capacity to digest blood ( S1E and S1F Fig ) , indicate that midgut integrity , and function were preserved upon QND treatment . Contrasting with the general organelle disappearance , extensive cytosolic electron-dense structures that resemble residual bodies ( hemoxisomes , [30] ) accumulate in the midgut cells of QND-treated insects ( Fig 2D and 2E ) . Thereby , maintenance of low "free" heme levels in the midgut lumen through its crystallization into Hz preserves midgut cellular architecture . To determine the consequences of impaired Hz formation at systemic levels , we evaluated in the next set of experiments hemolymphatic heme properties and redox homeostasis markers . The results showed that the hemolymph from QND-fed insects were reddish in color compared to controls ( Fig 3A Inset ) , exhibiting higher levels of "free" and protein-bound heme , as measured by light absorption at 365 nm and 412 nm , respectively ( S2C and S2D Fig ) . Direct quantification of total heme showed that inhibition of Hz formation in the midgut caused a remarkable increase in hemolymph ( Fig 3B ) and heart ( S3E Fig ) heme levels . The high hemolymphatic levels of heme after QND treatment increased the expression of Rhodnius Heme Binding Protein ( RHBP ) in about 7 . 5 times in fat bodies ( Fig 3C and S2F Fig ) , which is consistent with the antioxidant role of this protein by preventing the pro-oxidant effect of heme [31 , 32] . Higher RHBP levels conferred improved hemolymph buffering capacity against the "free" heme , as revealed by higher resistance of hemolymph to undergo blue shift the Soret peak after heme titration ( Fig 3D ) . Moreover , heme overload did not significantly affect biliverdin production in the heart ( S2G Fig ) , suggesting that Rhodnius heme oxygenase ( HO ) activity [33] might be saturated , thus not significantly contributing to heme detoxification . Despite the improved heme buffering capacity by RHBP ( Fig 3D ) , we observed systemic redox imbalance , as measured by higher lipid peroxide levels ( Fig 3E ) , and reduced concentrations of the main low molecular weight antioxidant urate in the hemolymph of adults ( Fig 3F and S2H Fig ) and of young nymphs ( S2I and S2J Fig ) from both colonies upon QND treatment . Reduction in urate levels by QND is directly related to defective heme crystallization , and not related to inhibition of urate synthesis , as this effect was only observed in blood fed insects ( Fig 3G ) . Supporting this proposal , reductions in urate levels were more pronounced exactly at the times of highest Hz production rates ( Fig 1B and S2A Fig ) . The data presented here indicate that systemic heme overload upon QND treatment causes redox imbalance , while activating a compensatory heme-scavenging mechanism ( RHBP ) in the hemolymph . As shown above , impairment of Hz formation in vivo promotes redox imbalance ( Figs 1C , 1D , 3E , 3F and 3G , S2H , S2I and S2J Fig ) , which is a key mechanism to control fertility in different models [11 , 34 , 35] . Then , we determined the potential consequences of QND treatment on R . prolixus reproduction , by assessing the number of eggs produced and laid per female . Unlike control insects , blockage of heme crystallization by QND caused a remarkable effect on ovary development and egg production , with no apparent pinkish-colored eggs observed four days after a blood meal ( Fig 4A ) . This remarkable effect on oogenesis was sustained along the full blood digestion and reproductive cycle , in both insect colonies ( Fig 4B and S3 Fig ) . Interestingly , the reduced egg production caused by QND involves redox imbalance , as supplementation in the diet with the major antioxidant urate restored oogenesis in QND-treated insects ( Fig 4B ) . This indicates that heme crystallization in the midgut determines oogenesis in triatomines by systemically regulating redox balance . Interestingly , QND strongly reduced T . cruzi parasite loads in R . prolixus digestive tract ( by ~ 75% ) 15 days after blood meal ( Fig 4C ) . Similar effects were also observed at earlier ( S4A Fig ) , and later ( S4B Fig ) times of blood digestion . To demonstrate whether parasites were affected by the pharmacological inhibition of Hz formation , we assessed the effects of QND and heme on in vitro epimastigotes and trypomastigotes cultures of T . cruzi . Although incubation of epimastigotes with 30 μM heme promoted parasite proliferation by ~ 55% , co-incubation with 50–100 μM QND completely blunted this effect ( S4C Fig ) . On the other hand , heme exerted a strong cytotoxic effect to trypomastigotes , reducing parasite counts by ~ 50% , which was not affected by the presence of 50–100 μM QND ( S4D Fig ) . These data indicate that the effect on the parasites is not due to the QND administration per se . On the other hand , dietary supplementation with urate restored parasite count in vivo to levels similar to the control group ( Fig 4C ) . In summary , these results indicate that heme crystallization in triatomine midgut represents a key mechanism to allow T . cruzi proliferation and survival , by keeping heme at strict concentrations , enough to boost epimastigote growth , but without leading to an excessive redox imbalance which culminates in parasite death .
We describe here an unprecedented function for Hz production in the triatomine insect R . prolixus that efficiently reduces redox heme reactivity allowing insect fertility and T . cruzi life-cycle . To our knowledge , this is the first comprehensive description of disrupted heme crystallization in a CD vector and identifies the unique role of Hz as a key redox-protective mechanism for both triatomine vector and for T . cruzi parasites . A summary of the effects resulted from impaired Hz formation in R . prolixus is schematically depicted in Fig 5 . As obligatory blood feeders , triatomine insects ingest copious amounts of blood to meet their energy demands , and hemoglobin is degraded into peptides and amino acids , releasing ~ 10 mM heme in the midgut lumen [3 , 36] . This massive heme release poses a major threat to triatomines [10 , 11 , 13 , 31] , owed the cytotoxic properties of heme [9 , 23 , 25 , 27 , 37] . The primary and major defensive mechanism against heme toxicity in triatomines consists on its crystallization into Hz . When this process was inhibited by QND treatment , “free” heme accumulates in the midgut lumen , a proposal that is experimentally sustained by the following observations: i ) blood intake ( S1C and S1D Fig ) , and digestion ( S1E and S1F Fig ) were not affected by QND , implying that heme supply for crystallization was not limited; ii ) Hz formation was significantly affected by QND both in adult ( Fig 1A and 1B , and S2A Fig ) , and nymph stages ( S2B Fig ) ; iii ) hemoxisomes/residual bodies increased in density with QND treatment ( Fig 2A and 2B ) ; and iv ) hemolymph ( Fig 3A and 3B , S2C and S2D Fig ) , and heart ( S2E Fig ) were overloaded with heme upon QND treatment . From the parasite side , although excessive “free” heme levels in the digestive tract are not harmful to T . cruzi epimastigote forms [6 , 7] , it exerts powerful cytotoxic effects to metacyclic trypomastigotes [7] . In fact , the proliferative effects of heme to epimastigote forms depend on oxidants generated by parasite mitochondria [6–8] , while trypomastigotes require reduced environments to allow differentiation [7] . In line with these observations , inhibition of Hz formation reduced total parasite counts in R . prolixus digestive tract ( Fig 4C , S4A and S4B Fig ) . Also , QND inhibited the proliferative effects of heme on epimastigotes ( S4C Fig ) , while did not potentiate the cytotoxic effect of heme on trypomastigotes ( S4D Fig ) . Therefore , we concluded that reductions in T . cruzi counts in triatomine midgut were most likely a consequence of the cytotoxic effect of excessive “free” heme on trypomastigotes rather than in epimastigotes . Despite the fact that QND promoted clear effects on triatomine heme/redox homeostasis , we cannot rule out the potential off-target effects of QND on T . cruzi parasite forms . In this sense , one might consider that reduced T . cruzi counts upon QND treatment results from lower urate levels , independently of its redox properties [7 , 38] . The differential susceptibility to excessive heme/oxidants may result from specific adaptations developed by T . cruzi forms to cope with environmental challenges that parasites face during their development [6–8 , 39] . Conceivably , Hz formation would provide a suitable and unique environment for trypanosomes to develop through the triatomine digestive tract , by maintaining “free” heme levels high enough to induce epimastigotes proliferation , without compromising metacyclogenesis and trypomastigotes survival . Although midgut represents the first line of defense against multiple stressors coming from the diet , its cells are not completely immune to handle “free” heme overload . In the process of transporting “free” heme from the midgut lumen to the hemolymph , midgut cells also facilitate the production and propagation of oxidants ( Fig 1C and 1D ) , causing redox imbalance . Supporting these observations , inhibition of Hz production in malaria parasites was also associated to redox imbalance [25 , 40] . A second phenotype observed is the remarkable architectural change undergone by midgut cells upon QND treatment ( Fig 2 ) . Extensive organelle disappearance ( Fig 2D ) , cytosolic vacuolization ( Fig 2B and 2D ) , enrichment of intracellular lipid droplets ( Fig 2B and 2E ) , and increase in structures with concentric membranes , similar to autophagosomes ( Fig 2H ) were all evident in QND treated insects . Moreover , mitochondrial disappearance was also a hallmark in QND treated insects , and the few detected exhibited remarkable structural changes , including clear matrix spaces ( Fig 2G ) , with fewer cristae compared to controls . Mitochondrial structural changes are indicative of reduced functionality , which might precede the extensive organelle elimination by autophagy . Similar ultrastructural observations were reported for Plasmodium and Schistosoma parasites under aminoquinoline treatment , including mitochondrial swelling , cytoskeletal disorganization , and autophagy [24 , 41 , 42] . Indeed , higher oxidant levels resulting from accumulation of porphyrins [43] , or iron [44] in different models point out to autophagy as a potential protective mechanism for cell quality control during iron/heme overload . Interestingly , given that proteases involved in hemoglobin degradation are produced by midgut cells [36] , and that blood digestion was not affected by QND treatment ( S1E and S1F Fig ) , the function of midgut cells was preserved despite the massive organelle removal , structural changes , and redox imbalance . Inhibition of Hz formation increased the flux of heme through the midgut leading to heme overload in the hemolymph ( Fig 3B , S2C and S2D Fig ) and the heart ( S2E Fig ) . As a consequence , systemic redox homeostasis shifted towards oxidation , as measured by higher lipid peroxide levels ( Fig 3E ) , and lower levels of the antioxidant urate in the hemolymph of QND treated insects ( Fig 3F and 3G , S2H , S2I and S2J Fig ) . Despite the altered heme homeostasis , biliverdin levels in the heart were not significantly changed by limited Hz production ( S2G Fig ) , suggesting that HO activity was not affected by QND . In line with these observations , the expression of the main heme scavenging protein RHBP significantly increased upon inhibition of Hz formation ( Fig 3C and S2F Fig ) , improving the buffering capacity of “free” heme in the hemolymph ( Fig 3D ) . Hemolymph light absorption at 412 nm , which reflects heme binding to RHBP [31 , 32] , significantly increased under QND treatment ( Fig 3A , S2C and S2D Fig ) , indicating that most of “free” heme was scavenged by apo-RHBP . Previous reports demonstrated that RHBP expression peaks ~ 2–4 days after blood meal , suggesting that heme regulates RHBP expression [45] , a proposal that is experimentally supported by the present work ( Fig 3B , 3C and 3D , S2C , S2D , S2E and S2F Fig ) . Despite the compensatory increase in heme buffering capacity provided by higher apo-RHBP levels under blockage of Hz formation ( Fig 3D ) , “free” heme accumulates in the hemolymph , as revealed by the higher absorption at 365 nm ( S2D Fig ) [46] . Thus , higher “free” heme titers in QND treated insects overwhelms the buffering capacity of RHBP , allowing heme to exert its pro-oxidant effects ( Fig 3E , 3F and 3G ) . Urate is a key antioxidant in triatomines [47] , and it is synthesized in the fat body upon heme signaling involving PKC activity [48] . In spite of heme overload , urate levels reduced upon QND treatment ( Fig 3F and 3G , S2H , S2I and S2J Fig ) , which might indicate urate overconsumption to counteract higher oxidant levels propagated by heme and to minimize redox damage in the hemolymph . Therefore , inhibition of Hz formation is not lethal to triatomines , as opposed to malaria parasites , but rather shifts the heme detoxification mechanisms from heme crystallization towards its scavenging by a chelating protein ( RHBP ) . In this study , we have observed a delay and a decrease of oogenesis and egg deposition as a consequence of impaired Hz formation ( Fig 4A and 4B and S3 Fig ) . This effect was partially reversed by dietary urate supplementation , suggesting that redox imbalance generated by heme overload represents an important mechanism that mediates reduced oogenesis ( Fig 4B ) . Due to the partial restoring effect of urate on egg production , we cannot completely exclude the possibility that QND might exert direct effects on oogenesis by an unknown redox-independent mechanism . In this regard , similar tradeoffs between redox homeostasis and reproduction were reported for distinct models , including blood feeding insect vectors [11 , 34] . Also , impairment in heme scavenging through the knockdown of RHBP in triatomines , caused lipid peroxidation , reducing mitochondrial function , and energy supply , with direct effects on embryogenesis , but not oogenesis [11] . The emerging picture is that RHBP serves not only as a preventive antioxidant , by sequestering "free" heme into an inert complexed form , but also as a mechanism for heme delivery to maintain a proper energy supply during embryogenesis . Thus , despite the fact that impairment of Hz formation and RHBP knockdown converge towards "free" heme overload in the hemolymph , the biological outputs were contrasting: while QND treatment impaired oogenesis ( Fig 4B and S3 Fig ) , RHBP silencing reduced embryogenesis [11] . These observations can be reconciled if we consider that the redox insult generated upon interference of Hz formation is much stronger than observed after RHBP knockdown , and , potentially , that embryogenesis would be a more sensitive process to oxidants than oogenesis in triatomines . Living organisms have evolved distinct mechanisms to avoid the potential cytotoxic effects of "free" heme overload [49] . In systems that are not physiologically adapted to high heme levels , such as mammals , heme is essentially detoxified through its enzymatic degradation by HO activity , or scavenged by heme binding proteins [49–51] . However , depending on the magnitude of heme exposure , these protective mechanisms might be saturated and result in tissue damage . Conversely , natural exposure to high levels of heme , as in the case of hematophagous organisms , posed a selective pressure to evolve efficient and unique mechanisms to cope with heme overload [3 , 12] . Given the central importance of heme homeostasis for both triatomine insect vector [3 , 12] , ( Figs 1–5 ) , and T . cruzi parasites [4–9] , ( Fig 4 and S4 Fig ) , interference on Hz formation disrupts vector-parasite interaction , with detrimental effects on CD transmission . In conclusion , we present here , to our knowledge , the first comprehensive description of physiological , biochemical , cellular , and molecular consequences of disrupted heme crystallization in a CD vector . Collectively , our data highlight the dual role of Hz as a key protective mechanism , with striking beneficial outputs for both triatomine vector and for T . cruzi . The possibility to target such a central redox process for both vector and parasite represents a major step towards the development of innovative and rational strategies for effective CD control .
R . prolixus were fed and raised according to the Ethical Principles in Animal Experimentation approved by the Ethics Committee in Animal Experimentation ( CEUA ) at Federal University of Rio de Janeiro , CEUA/UFRJ ) , and at Fundação Oswaldo Cruz ( CEUA/FIOCRUZ ) , under the approved protocols #IBQM050 , and P-54/10-4/LW12/11 , respectively . Experiments with citrated human blood were carried out using insects from Fiocruz colony and were fed by using an artificial apparatus according to the Ethical Principles in Animal Experimentation approved by the Ethics Committee in Animal Experimentation ( CEUA/FIOCRUZ ) under the approved protocol L-0061/08 . All blood donors provided informed written consent . All protocols are from CONCEA/MCT ( http://www . mctic . gov . br/mctic/opencms/institucional/concea/index . html ) , which is associated with the American Association for Animal Science ( AAAS ) , the Federation of European Laboratory Animal Science Associations ( FELASA ) , the International Council for Animal Science ( ICLAS ) and the Association for Assessment and Accreditation of Laboratory Animal Care International ( AAALAC ) . Technicians dedicated to the animal facilities at the Institute of Medical Biochemistry , and at Fundação Oswaldo Cruz , carried out all aspects related to rabbit husbandry under strict guidelines to ensure careful and consistent handling of the animals . Rhodnius prolixus insects used throughout this work were from two different well-established colonies: i ) Institute of Medical Biochemistry ( IBqM ) at Federal University of Rio de Janeiro ( UFRJ , Brazil ) and ii ) Laboratory of Insect Physiology at Oswaldo Cruz Institute ( Fiocruz , Rio de Janeiro , Brazil ) . The reason to use two colonies aimed the exclusion of potential external ( colony handling/maintenance , blood source ) , and internal ( microbiota , genetic background ) interference factors that might eventually bias the results . The trends observed throughout the data from both colonies were strikingly similar , only differing in the magnitude of the responses . All insects were kept at 28°C and 60–80% relative humidity , with a photoperiod of 12 h of light/12 h of dark , as previously described [52] . Insects from IBqM colony were maintained by feeding directly on rabbit ears , while individuals from Fiocruz colony were maintained by feeding artificially with defibrinated rabbit blood using an artificial feeder [53] . T . cruzi Dm28c strain ( CT-IOC-010 ) was provided by the Trypanosomatid Collection of the Oswaldo Cruz Institute , Fiocruz , Brazil . Epimastigote parasite forms were grown at 28°C for 7 days in brain-heart infusion medium ( BHI , BD Bacto , BD Biosciences , USA ) + 10% fetal bovine serum ( Vitrocell , Brazil ) , and were further incubated for 5 days with 30 μM heme , or 30 μM heme + 50–100 μM QND in 96-well plates . Metacyclogenesis was performed as described elsewhere [54] , and resulting trypomastigotes were incubated for 2 days with 30 μM heme , or 30 μM heme + 50–100 μM QND in 96-well plates . Parasites in the culture supernatant were collected , and survival was determined by cell counting of viable cells in a Neubauer chamber . For both insect colonies , treatments were performed in adult mated R . prolixus females from their second cycle after the last ecdysis , while young nymph stages were from the fifth cycle before the last ecdysis . To inhibit Hz formation in vivo , all insects were artificially fed with rabbit blood supplemented with either 0 . 6% ( v/v ) ethanol ( control ) or different quinidine concentrations ( 0–500 μM QND ) previously prepared from 30% ethanol or 5 mM QND in 30% ethanol stocks and maintained for up to 24 days after blood meal . To control the volume of blood ingested by each group , the insects were weighted before and after blood feeding , and no significant changes were reported in the volume of engorged blood ( S1C and S1D Fig ) . For the T . cruzi infection experiments , insects from Fiocruz colony at their first cycle after the last ecdysis were fed with heat-inactivated citrated human blood supplemented with 108 epimastigotes/mL ( Dm28c strain ) as described elsewhere [55] . After 28 days , insects were artificially fed with rabbit blood ( control ) , or blood supplemented with 100 μM QND . Viable parasite counts were determined in the whole midgut and monitored 10 , 15 and 30 days upon feeding with QND by direct counting of parasites in a Neubauer chamber . Four days after feeding , anterior midguts were homogenized in PBS , and general protease inhibitor cocktail ( Sigma , catalog number P2714 ) and stocked at -80°C until the measurement . Hemoglobin in anterior midgut was quantified by using a commercially available kit ( Bioclin , Brazil ) . Reagent solution was diluted 1:100 in water and 2 . 5 mL were added to 10 μL of sample . Experiments were performed in 96-well plates and the absorbance at 540 nm was determined after 5 min at room temperature in a Molecular Devices Spectra Max M5spectrophotometer and hemoglobin concentration was calculated using hemoglobin ( Bioclin , Brazil ) as standard . Posterior midguts were dissected under PBS and the luminal contents were collected , homogenized and centrifuged at 11 , 000×g for 10 min . The pellet was re-suspended in 0 . 1 M sodium bicarbonate buffer , pH 9 . 1 and 2 . 5% SDS . Samples were centrifuged at 11 , 000×g for 10 min and the pellet was washed five times with the same buffer , followed by two washes with deionized water . Hz quantification was carried out by adding 1 mL of 0 . 1 M NaOH to the pellet , vortexing the samples for 30 min , followed by determination of heme content at 400 nm in a GBC-UV/Vis-920 spectrophotometer , using a standard curve made with hemin ( Frontier Scientific , USA ) dissolved in 0 . 1 M NaOH . To assess oxidant levels , the wings , legs and dorsal plaques were dissected from the insects , and the hemocoel was filled with a 50 μM solution of oxidant-sensitive fluorophore dihydroethidium ( hydroethidine , DHE ) ( Invitrogen , USA ) diluted in L15 culture medium ( Gibco , USA ) containing 5% ( v/v ) fetal bovine serum . The samples were incubated in the dark at 28°C . After 20 min of incubation , the midguts were washed with 0 . 15 M NaCl , and immediately transferred to a glass slide for fluorescence microscopy analysis , as previously described [13] . Each condition was registered in at least three different areas for fluorescence , and differential interference contrast ( DIC ) . Quantitative evaluation of fluorescence levels was performed by acquiring images under identical conditions using an objective of 20 x and 100 ms exposure time in each experiment . The images were acquired in a Zeiss Observer Z1 fluorescence microscope with a Zeiss Axio Cam MrM , and data were analyzed using AxioVision version 4 . 8 software . The #15 filter set ( excitation BP 546/12 nm; beam splitter FT 580 nm; emission LP 590 nm ) was used . Posterior midguts of adult females were collected and fixed in 2 . 5% of glutaraldehylde in 0 . 1 M sodium cacodylate buffer pH 7 . 2 for at least 24 h at 4°C , followed by a post-fixation in 1% osmium tetroxide , 0 . 8% potassium ferricyanide and 2 . 5 mM calcium chloride in the same buffer for 1 h at room temperature . The dehydration steps were performed by incubations of 15 min for each acetone concentration ( 30% , 50% , 70% , 90% and 100% ) . Then the samples were gradually embedded in epoxy polybed resin ( Polybed 812 , Polysciences , Germany ) . Infiltration was performed by incubating samples with 1:3 , 1:2 , 2:3 of epoxy polybed resin:acetone for 24 h to each step . After that , samples were infiltrated with epoxy polybed resin 100% , incubated at room temperature for 4 h and then incubated at 60°C for 4 days to complete polymerization . Then , semi-thin sections ( 0 . 5 μm ) were obtained , stained with toluidine blue and observed in a Zeiss Axioplan bright field microscope for histological analysis . Alternatively , for ultrastructural analysis , ultrathin sections were obtained with an ultramicrometer Ultracuts ( Leica ) collected in copper grids , stained in uranyl acetate for 20 min and lead citrate for 2 min , and sections were observed in a JEOL JEM1011 transmission electron microscope at Oswaldo Cruz Institute electron microscopy platform . Hemolymph was collected before blood meal and 1 , 2 , 3 , 4 , 7 and 15 days after feeding , in tubes containing a few crystals of phenylthiourea , by cutting a leg , and applying a gentle pressure to the insect abdomen . Qualitative analyses of heme levels were assessed 4 days after feeding by diluting a hemolymph aliquot ( 15 μL ) in 485 μL PBS pH 7 . 4 , and the light absorption spectra was analyzed between 300 nm and 800 nm in a Shimadzu UV-2550 spectrophotometer . Absolute quantification of total heme was determined by the alkaline pyridine-hemochrome method , using the reduced minus oxidized spectra as described elsewhere [56] . Quantification of "free" and RHBP complexed heme were determined in hemolymph specifically at 365 nm and 412 nm , respectively [32 , 46] . Heme binding to Rhodnius heme binding protein ( RHBP ) was measured by progressively adding 2 μL of 0 . 1 mM heme solution ( Frontier Scientific , USA ) as previously described [32] using a Shimadzu UV-2550 spectrophotometer . Lipid peroxidation was assessed by quantifying the levels of thiobarbituric acid reactive substances assay ( TBARS ) at 532 nm in a Molecular Devices Spectra Max M5 spectrophotometer as described elsewhere [10] . To determine urate levels , 3 μL of hemolymph was diluted 1:5 in water and subsequently urate concentration was enzymatically determined using a commercially available kit ( Doles , Brazil ) following the manufacturer’s instructions and using urate ( Doles , Brazil ) as standard as previously described [47] . Ten heart samples per group were collected four days after feeding , homogenized with 200 μL 5% acetonitrile and 0 . 05% TFA as solvent , pH 2 . 0 ( 1:2 v/v ) , centrifuged for 15 min at 12 , 000×g and the supernatant was applied onto a Shimadzu CLC-ODS C18 column ( 15 mm × 22 cm ) equilibrated with the same solvent , using a flow rate of 0 . 4 mL/min . After 10 min , an acetonitrile linear gradient ( 5–80% ) was applied for 10 min , followed by 20 min of 80% acetonitrile with 0 . 05% TFA , pH 2 . 0 , and heme and biliverdin peaks were identified as previously described [33] . The experiments were performed at room temperature . Fat bodies from fed females were dissected 4 days after feeding . Total RNA was extracted from 1 fat body per tube containing 1 mL of TRIzol reagent ( Invitrogen , USA ) , according to the manufacturer’s instructions and at least 3 insects were used per experiment . RNA concentrations were determined spectrophotometrically at 260 nm on a Nanodrop 1000 spectrophotometer v . 3 . 7 ( Thermo Fisher Scientific , USA ) . Following treatment with 1 U RNase-free DNase I ( Fermentas International Inc . , USA ) for 5 min at 37°C , 1 μg of RNA was used for cDNA synthesis with a high capacity cDNA reverse transcription kit ( Applied Biosystems , USA ) and random hexamers according to the manufacturer’s instructions . Quantitative PCR was performed in a 7500 real time PCR system ( Applied Biosystems , USA ) using SYBR Green PCR Master Mix ( Applied Biosystems , USA ) under the following conditions: one cycle for 10 min at 95 °C , followed by fifty cycles of 15 s at 95 °C and 45 s at 60 °C . PCR amplification was performed using the following primers: qRHBPF ( 5′-TCCTTCACACTCTCCGCAAC-3′ ) ( forward ) and qRHBPR ( 5′-GTACGCTTGGTACGCCACTT-3′ ) ( reverse ) [11] . Three independent biological replicates were conducted , and all PCRs were performed in triplicate . R . prolixus actin gene ( accession number EU2337941 ) expression was used as an internal control for normalization . Primers used for actin PCR amplification were RpActRT F ( 5′-CCATGTACCCAGGTATTGCT-3′ ) ( forward ) and RpActRT R ( 5′-ATCTGTTGGAAGGTGGACAG-3′ ) ( reverse ) [11] . ΔΔCt values were calculated from Ct ( cycle threshold ) values obtained on quantitative RT-PCR and were used to calculate relative expression and perform statistical analysis [57] . Differences were considered significant when p<0 . 05 . The relative expression values based on 2-ΔΔCt were used only for graphic construction . Oogenesis and egg laying were determined along 24 days after blood feeding , by direct counting laid eggs .
|
Chagas disease is a fatal illness caused by Trypanosoma cruzi parasites , which are transmitted by blood sucking triatomine insect vectors . Although blood is a natural food source for these insects , its digestion releases toxic products , which poses a dietary challenge for both triatomine insects and trypanosomes . To overcome this , triatomines eliminate these toxic blood products by a unique process of heme crystallization into hemozoin that take place in their digestive tract . Here we describe that this detoxification process represents the major mechanism for redox balance control , and is necessary to allow triatomine insect reproduction , and Trypanosoma cruzi infection . Disruption of heme crystallization in triatomine insects thus represents a new venue for Chagas disease control , by targeting at the same time insect reproduction and parasite transmission .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"heme",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"condensed",
"matter",
"physics",
"parasitic",
"protozoans",
"animals",
"protozoans",
"crystallography",
"insect",
"vectors",
"zoology",
"crystallization",
"oxidation-reduction",
"reactions",
"infectious",
"diseases",
"solid",
"state",
"physics",
"proteins",
"chemistry",
"insects",
"disease",
"vectors",
"arthropoda",
"chemical",
"precipitation",
"physics",
"biochemistry",
"trypanosoma",
"cruzi",
"animal",
"physiology",
"trypanosoma",
"eukaryota",
"blood",
"anatomy",
"post-translational",
"modification",
"electrochemistry",
"entomology",
"physiology",
"insect",
"physiology",
"invertebrate",
"physiology",
"biology",
"and",
"life",
"sciences",
"chemical",
"reactions",
"physical",
"sciences",
"species",
"interactions",
"organisms"
] |
2018
|
Heme crystallization in a Chagas disease vector acts as a redox-protective mechanism to allow insect reproduction and parasite infection
|
In an endeavor to find an orally active and affordable antileishmanial drug , we tested the efficacy of a cationic amphiphilic drug , imipramine , commonly used for the treatment of depression in humans . The only available orally active antileishmanial drug is miltefosine with long half life and teratogenic potential limits patient compliance . Thus there is a genuine need for an orally active antileishmanial drug . Previously it was shown that imipramine , a tricyclic antidepressant alters the protonmotive force in promastigotes , but its in vivo efficacy was not reported . Here we show that the drug is highly active against antimony sensitive and resistant Leishmania donovani in both promastigotes and intracellular amastigotes and in LD infected hamster model . The drug was found to decrease the mitochondrial transmembrane potential of Leishmania donovani ( LD ) promastigotes and purified amastigotes after 8 h of treatment , whereas miltefosine effected only a marginal change even after 24 h . The drug restores defective antigen presenting ability of the parasitized macrophages . The status of the host protective factors TNF α , IFN γ and iNOS activity increased with the concomitant decrease in IL 10 and TGF β level in imipramine treated infected hamsters and evolution of matured sterile hepatic granuloma . The 10-day therapeutic window as a monotherapy , showing about 90% clearance of organ parasites in infected hamsters regardless of their SSG sensitivity . This study showed that imipramine possibly qualifies for a new use of an old drug and can be used as an effective orally active drug for the treatment of Kala-azar .
The disease visceral leishmaniasis or Kala-azar is caused by the protozoan parasite Leishmania donovani ( LD ) and is widening its base in different parts of the world [1] , [2] . Pentavalent antimonial or SSG , which has long been the first line drug , is no longer recommended for use as high levels of resistance in the Indian subcontinent have been reported [3] . Other drugs like miltefosine ( hexadecylphosphocholine , a polyene antibiotic ) and amphotericin B ( an anti-fungal agent ) are in current clinical use . As miltefosine is orally active , it offers advantages in terms of reduced hospitalization but cannot be used during pregnancy and lactation [2] . Amphotericin B and its liposomal form are to be administered as an infusion and therefore the patients require hospitalization [4] . Unfortunately , treatment failure cases to miltefosine [5] and amphotericin B [6] are emerging , which raises serious concerns for their future use . There is a genuine need for an orally active and affordable drug for the treatment of relapsed Kala-azar cases . Imipramine , N- ( γ-dimethylaminopropyl ) -iminodibenzyl HCl , is a tricyclic antidepressant and belongs to the broad class of cationic amphiphilic drugs . The tricycle consists of two benzene rings fused with a seven member heterocycle . Imipramine is FDA ( Food and Drug Administration ) approved drug for treating depression and paediatric nocturnal enuresis [7] , and is sometimes used off-label to treat chronic pain in combination with other pain medications [8] . The dose range for treating depression is 100–200 mg daily and the recommended use for enuresis is 10–75 mg daily [9] . The selection of imipramine for therapy of experimental visceral leishmaniasis is based on the following past observations by others: ( i ) the drug alters the proton motive force of LD's membrane [10] , ( ii ) inhibits trypanothione reductase , an enzyme upregulated in SSG resistant LD parasites [11] , ( iii ) an effective immunomodulator as it induces the production of TNF-α , an important cytokine for antileishmanial defense [12] , ( iv ) cationic properties favor its absorption by phagocytic cells and accumulation in phagolysosomal bodies [13] , and ( v ) its metabolite desipramine is as effective as the parent drug against LD promastigotes [14] . These compelling attributes of imipramine towards Leishmania parasites led us to test its efficacy directly on LD and also in experimental infection induced by recent clinical isolates of SSG-S and SSG-R LD parasites with miltefosine as a reference oral drug . In this investigation , we endeavored to study the effect of oral administration of imipramine in LD infected hamster model . Our study done in hamster model very clearly showed that this drug is highly active in vitro as well as in vivo . Furthermore it plays a strong immunomodulatory role which also favored parasite clearance . Thus imipramine may be used orally in the treatment of visceral leishmaniasis . To our knowledge this is the first report on the therapeutic efficacy of imipramine in experimental visceral leishmaniasis .
BALB/c mice ( Mus musculus ) and hamsters ( Mesocricetus auratus ) were maintained and bred under pathogen free conditions . Use of both mouse and hamster was approved by the Institutional Animal Ethics Committees of Indian Institute of Chemical Biology , Kolkata , India . All experiments were performed according to the National Regulatory Guidelines issued by CPSEA ( Committee for the Purpose of Supervision of Experiments on Animals ) , Ministry of Environment and Forest , Government of India . All parasites for this study were received from European Union KaladrugR project consortium . These parasite samples are fully anonymized and study with these parasites is approved by Institutional Review Board of Institute of Medical Sciences , Benaras Hindu University , Varanasi , India . The details of the patients and the treatment profile of the patients from whom Leishmania donovani ( LD ) parasites were derived have been published previously [15] . Clonal population of LD parasites MHOM/IN/10/BHU816/1 ( BHU 816 ) and MHOM/IN/09/BHU777/0 ( BHU 777 ) are SSG sensitive ( SSG-S ) and strains MHOM/IN/09/BHU575/0 ( BHU 575 ) , MHOM/IN/10/BHU782/0 ( BHU 782 ) , MHOM/IN/10/BHU814/1 ( BHU 814 ) and MHOM/IN/10/BHU872/6 ( BHU 872 ) are SSG resistant ( SSG-R ) . LD promastigotes were maintained in M199 medium ( Sigma Aldrich , St . Louis , MO ) supplemented with 10% heat inactivated FBS ( Gibco ) , 100 IU/mL of penicillin and 100 µg/mL of streptomycin ( Gibco ) in a 22°C room as described elsewhere [15] . Imipramine hydrochloride ( Sigma Aldrich , St . Louis , MO ) and miltefosine ( Kindly provided by Aeterna Zentaris GmbH ( Germany ) to the KaladrugR project consortium , batch#1149149 ) solutions were prepared at 1 mg/ml in PBS ( Sigma Aldrich , St . Louis , MO ) , followed by sterile filtration using 0 . 22 µM filters ( Milipore ) as and when required . PECs were harvested from BALB/c mice by lavage , 48 h after i . p . injection of 2% ( w/v ) soluble starch ( Sigma Aldrich , St . Louis , MO ) . For convenience , PECs of BALB/c mice were defined as MΦ . MΦ were harvested on sterile 22 mm square coverslips ( Bluestar , India ) in 35 mm disposable petriplates ( Tarsons , India ) at a density of 105/cover slip in RPMI 1640 medium ( Sigma Aldrich , St . Louis , MO ) supplemented with 10% heat inactivated FBS , 100 IU/mL of penicillin , and 100 µg/mL of streptomycin , i . e . RPMI complete medium . The cells were left to adhere for 48 h at 37°C under 5% CO2 before infection . The MΦs were infected with stationary phase promastigotes at a ratio of 1∶10 [14] , [15] . After incubating the cultures at 37°C and 5% CO2 overnight or for 4 h , non-phagocytosed promastigotes were washed off with serum free medium RPMI 1640 and treatment provided as described [16] . MΦs were harvested on a 96-well tissue culture plate ( BD Biosciences ) in RPMI complete media and left to adhere for 48 h at 37°C under 5% CO2 . Successive increasing concentrations of imipramine were added in triplicate and incubated for 24 h . After completion of incubation , MTT ( Sigma Aldrich , St . Louis , MO ) was added and incubated for 4 h at room temperature . Solublizing agents [0 . 04 N HCl ( Merck ) in isopropanol ( Merck ) ] were added after incubation and the optical density ( OD ) was measured after 30 min in a plate reader at 570 nm . The relative number of live cells was determined based on the optical absorbance of the treated and untreated samples and of blank wells , as described previously [17] . Day 5 culture of parasites was used to determine the drug efficacy ( IC50 ) to kill promastigotes using MTT [18] . The LD parasites were plated on the 96-well cell culture plates at a density of 105 cells/well and kept in presence of imipramine for 48 h . Results were expressed as the concentration that inhibited parasite growth by 50% ( IC50 ) . Analysis was carried out using Graphpad Prism5 software ( version 5 . 03 ) . In order to determine EC50 ( Efficacy against intracellular amastigote ) , the drug was serially diluted in RPMI complete medium over six concentrations in triplicate at each concentration . Stock solutions and dilutions were freshly prepared for each use . Infected MΦs were incubated with drug dilutions for another 24 h at 37°C and under 5% CO2 . Untreated MΦs received medium alone and intracellular parasites were enumerated . At the endpoints , the coverslips were washed with PBS , dried , fixed with 100% methanol ( Merck ) , stained with 10% Giemsa ( Sigma Aldrich , St . Louis , MO ) and examined microscopically . One hundred MΦs/coverslip were scored and the amastigotes were enumerated [19] . The average of three untreated cultures was taken as 100% control against which the percentage inhibition of infected MΦs in treated cultures was calculated . The 50% effective concentration ( EC50 ) of imipramine for each of the isolates was estimated as described elsewhere [19] , [20] . The JC-1 dye ( Molecular Probes , Eugene , OR ) has been used routinely to monitor the mitochondrial potential [21] . The monomeric form has an emission maximum at 527 nm . The dye at higher concentrations or potentials forms red fluorescent J-aggregates with an emission maximum at 590 nm . The ratio of this red/green ( λ590/λ527 ) fluorescence is known to depend only on the membrane potential . A working solution of JC-1 was therefore prepared as per manufacturer's instruction . Imipramine and miltefosine treated and untreated LD promastigotes were incubated with the JC-1 working solution for 25 min in a 96-well plate and washed . Cell pellets were resuspended in assay buffer and analyzed under a fluorescent plate reader ( Fluorescence plate reader LS 55 , Perkin Elmer ) . Double staining for annexin V fluorescein isothiocyanate ( FITC ) -PI was performed with the Annexin-V apoptosis detection kit ( Molecular Probes , Eugene , OR ) [22] . In brief , untreated , imipramine-treated , or miltefosine-treated promastigotes were washed twice in cold PBS and centrifuged at 3000 rpm for 10 min . The pellets were resuspended in 100 µL of annexin V-FITC in the presence of PI according to the instructions of the manufacturer . After 15 min of incubation in the dark , the intensity of annexin V-FITC labeling was recorded on a flow cytometer ( FACSARIA II , Becton Dickinson , San Diego , CA ) and analyzed with FACSDIVA software , version 6 . 1 . 1; the percentage of positive cells was then assessed . The membrane fluorescence and lipid fluidity of MΦ under parasitized condition as well as after imipramine treatment were measured following the method described by Shinitzky and Inbar [23] . Briefly , the fluorescent probe DPH ( Molecular Probes , Eugene , OR ) was dissolved in tetrahydrofuran ( Merck ) at 2 mM concentration . To 10 ml of the rapidly stirring PBS solution ( pH 7 . 2 ) , 10 µL of 2 mM DPH solution was added . For labeling , 106cells were mixed with an equal volume of DPH in PBS ( cf 1 µM ) and incubated for 2 h at 37°C . Thereafter the cells were washed thrice and resuspended in PBS . The DPH probe bound to the membrane of the cell was excited at 365 nm and the intensity of emission was recorded at 430 nm in a spectrofluorometer ( LS 55 , Perkin Elmer ) . The FA value was calculated using the equation: FA = [ ( III−I⊥ ) / ( III+2I⊥ ) ] , where III and I⊥ are the fluorescent intensities oriented , respectively , parallel and perpendicular to the direction of polarization of the excited light [23] . MΦs , also defined as antigen presenting cells ( APCs ) , were harvested from peritoneal cavity of mice at 106 cells/well in a 48 well tissue culture plate , then incubated for 24 h with specific peptide Lambda repressor λR12–26 ( GenScript , USA ) and T cell hybridoma 9H3 . 5 ( kind gift of Professor Malcolm Gefter , Massachusetts Institute of Technology , Cambridge , Massachusetts ) in complete RPMI medium in a 37°C incubator . The culture supernatants were analyzed for the presence of IL 2 using mouse IL 2 ELISA kit ( BD Biosciences , San Diego , CA ) as per manufacturer's instruction . To monitor the level of reactive oxygen species ( ROS , including superoxide , hydrogen peroxide , and other reactive oxygen intermediates ) , the cell-permeable , non polar , H2O2-sensitive probe H2DCFDA ( Molecular Probes , Eugene , OR ) was used [24] . The extent of H2O2 generation was defined as the extent of ROS generation for convenience . For each experimental sample , fluorometric measurements were performed in triplicate and the results were expressed as the mean fluorescence intensity per 106 cells . Nitric oxide ( NO ) generation was monitored by using the Griess reagent ( Molecular Probes , Eugene , OR ) as described previously [25] , and the results are expressed in µM nitrite . To infect hamsters ( 6 weeks old ) , two SSG-S ( BHU 777 and BHU 816 ) and two SSG-R ( BHU 575 and BHU 814 ) LD amastigotes were purified as described [26] and inoculated ( 107 parasites in 200 µL ) via intracardiac routes as described previously [27] . Imipramine is usually used in human at a dose of 100–200 mg/day ( average of 150 mg/day ) for the treatment of depression [9] , [28] . Considering the average human body weight of 60 kg , the effective dose is 2 . 5 mg/kg/day . Based on the dose equivalence between human and rodents [29] , the dose of imipramine in mouse and hamsters would be 41 and 25 mg/kg/day respectively . In our investigation , the highest dose used was 5 mg/kg/day both in mouse and hamsters which is effectively 12 . 3 and 7 . 4 times lower than the equivalent human dose . Miltefosine is used in human at the dose 2 . 5 mg/kg [5] . Based on dose equivalent formula , we converted the normal human dose to hamster equivalent dose . Thus we treated hamsters with the maximum dose 17 . 5 mg/kg which is ∼7 times high then the normal human dose . The 8-week infected hamsters ( i . e . 14-week old hamsters ) were randomly divided into four groups ( groups I to IV ) . Group I received only saline , groups II to IV received imipramine at the dose levels of 0 . 05 , 0 . 5 and 5 mg/kg/day respectively for 4 weeks by oral route using a feeding needle as described by others [30] . Miltefosine treatment was carried out in 8-week infected hamsters for 4 weeks at a dose of 17 . 5 mg/kg/day orally . Two days after the completion of treatment , hamsters were sacrificed to determine splenic and hepatic parasite burdens by stamp smear method as described elsewhere [27] , [31] , as well as by the serial dilution method [27] . Blood was collected from hamsters and mice as described previously [27] and kept overnight at 4°C; serum was prepared by centrifugation . Soluble Leishmanial Antigen ( SLA ) was prepared from stationary phase LD promastigotes of LD following the published protocol [27] . Briefly , leishmanial lysate from washed promastigotes ( 109/ml ) was prepared by several cycles ( minimum six ) of freezing ( −70°C ) and thawing ( 37°C ) followed by 5 min incubation on ice . Partially lysed promastigotes were then disrupted in a sonicator ( Misonex , Farmingdale , NY ) thrice for 30 s each and centrifuged at 10 , 000 rpm for 30 min at 4°C . The supernatant containing soluble antigen was collected and the protein concentration was determined by Bradford Protein Assay method ( Bio-Rad , Herculis , CA ) . The prepared antigen was stored at −70°C until further use . Splenocytes from different experimental groups of hamsters were prepared after Ficoll ( Sigma Aldrich , St . Louis , MO ) density gradient centrifugation and then suspended in complete RPMI medium . Cells were plated in triplicate at a concentration of 105 cells/well in 96-well plates and allowed to proliferate for 3 days at 37°C in a 5% CO2 incubator either in the presence or absence of SLA ( 5 µg/ml ) ( 29 ) . For ConA ( Sigma Aldrich , St . Louis , MO ) induced proliferation , the mitogen was added at a concentration of 5 µg/mL as described previously [27] . Cells were treated with MTT ( 0 . 5 mg/mL ) 4 hr before harvest as described previously [18] and incubated again at the same condition for 4 more hrs . MTT crystals were then solubilised using Isopropanol-HCl mixture ( 0 . 04% ) and the absorbance at 570 nm was read at an ELISA plate reader ( DTX 800 multimode detector , Beckman Coulter , California ) . Serum samples were obtained from different groups of hamsters and mice ( five animals per group ) , and analyzed to determine the parasite SLA-specific antibody titer . 96-well Enzyme-Linked ImmunoSorbent Assay ( ELISA ) plates were coated with SLA ( 2 µg/ml ) in PBS for overnight at 4°C . The plates were blocked with 5% FCS in PBS at room temperature for 1 h to prevent nonspecific binding . Sera from different groups of hamsters were added at various dilutions , and incubated for 2 h at room temperature . These were diluted 10−1 , 10−2 , and 10−3 times for the determination of IgG1 and 10−3 , 10−4 , and 10−5 times for IgG2 . Biotin-conjugated mouse anti-hamster IgG1 ( BD Biosciences , San Diego , CA ) and mouse anti-Armenian and anti-Syrian hamster IgG2 ( BD Biosciences , San Diego , CA ) were added and incubated for 1 h at room temperature; this was followed by 1 h of incubation with the detection reagent ( streptavidin-conjugated horseradish peroxidase ) . As a peroxide substrate in citrate buffer ( 0 . 1 M , pH 4 ) , TMB ( Sigma ) was added along with 0 . 1% H2O2 ( Merck ) to a 96-well plate , and the absorbance at 450 nm was read with an ELISA plate reader [27] . RNA was isolated from the splenocyte of hamsters using Trizol ( Invitrogen ) as described previously [27] . The forward and reverse primers were used to amplify cytokine transcripts . All of these hamster-specific primers , except for the inducible NO synthase ( iNOS ) primer , were originally described by Melby et al . [31] . The following forward and reverse primers were used: for IL 10 , forward primer 5′ACAATAACTGCACCCACTTC3′ and reverse primer 5′AGGCTTCTATGCAGTTGATG3′ ( 432-bp product ) ; for IL 4 , forward primer 5′CATTGCATYGTTAGCRTCTC3′ and reverse primer 5′TTCCAGGAAGTCTTTCAGTG3′ ( 463-bp product ) ; for interferon gamma ( IFN γ ) , forward primer 5′GGATATCTGGAGGAACTGGC3′ and reverse primer 5′CGACTCCTTTTCCGCTTCCT3′ ( 309-bp product ) ; for tumor necrosis factor alpha ( TNF α ) , forward primer 5′GACCACAGAAAGCATGATCC3′ and reverse primer 5′TGACTCCAAAGTAGACCTGC3′ ( 695-bp product ) and for transforming growth factor beta ( TGF-β ) , forward primer 5′CCCTGGAYACCAACTATTGC3′ and reverse primer 5′ATGTTGGACARCTGCTCCAC3′ ( 310-bp product ) . To obtain specific amplification for iNOS , the following specific primers were used ( 6 ) : forward primer 5′ GCAGAATGTGACCATCATGG3′ and reverse primer 5′CTCGAYCTGGTAGTAGTAGAA3′ ( 198-bp product ) . For hypoxanthine phosphoribosyl transferase ( HPRT ) amplification the following primers were used [31]: forward primer 5′ATCACATTATGGCCCT CTGTG3′ and reverse primer 5′CTGATAAAATCTACAGTYATGG3′ ( 125-bp product ) . Degenerate bases are indicated above by International Union of Pure and Applied Chemistry designations ( Y = C or T; R = A or G ) . Details of the procedure have been described previously [27] . Densitometry analyses were done using the ImageJ software ( v1 . 41o ) , ethidium bromide staining , and visualization under a UV transilluminator . For densitometric calculations , the same band area was used to determine band intensity and normalized for HPRT . To evaluate long-term therapeutic ability , normal hamsters , infected hamsters and imipramine treated infected hamsters ( 30 hamsters per group ) were used to study survival kinetics as described previously [32] . Spleens and livers were fixed in 10% formalin ( Merck ) and embedded in paraffin . Tissue sections ( 5 µm ) were stained with hematoxylin-eosin to study their microarchitecture by light microscopy . Photomicrographs were taken with a Nikon Eclipse E200 microscope . The statistical significance of differences between groups was determined by the unpaired two-tailed Student's t test . Statistical significance was defined as a P value of <0 . 05 and the results were expressed as averages and standard deviations of triplicate measurements .
Details on the clinical isolates used in this investigation in terms of their sensitivity to Sodium stibogluconate ( SSG ) have already been published [15] . Out of these , two SSG sensitive ( BHU 777 and BHU 816 ) and four SSG resistant strains ( BHU 575 , BHU 782 , BHU 814 and BHU 872 ) were selected for this investigation . These isolates were subjected to imipramine treatment in vitro and ex vivo to measure the IC50 and EC50 ( Table 1 ) . It was observed that regardless of difference in SSG sensitivity , there was no significant difference in IC50 or EC50 ( Table 1 ) . For convenience , the rest of the study was carried out with two SSG-S ( BHU 777 and BHU 816 ) and two SSG-R ( BHU 814 & BHU 575 ) isolates , and were defined as BHU 816 ( S ) , BHU 777 ( S ) , BHU 814 ( R ) and BHU 575 ( R ) respectively . The transmembrane potential ( ΔΨm ) was evaluated using JC-1 , a lipophilic cationic dye as described [21] . For this investigation the drugs imipramine and miltefosine were used at a concentration of 75 µM [33] and 40 µM [22] respectively . We observed that both the SSG-R and SSG-S strains showed similar sensitivities to imipramine as evident from the significant decrease in ΔΨm after 8 h of treatment ( Figure 1A ) . On the other hand , miltefosine failed to induce any change in ΔΨm at 8 h of treatment ( Figure 1A ) . After 8 h of imipramine exposure to BHU 575 ( R ) , 60% of parasites were apoptotic whereas miltefosine induces apoptosis in only 5 . 5% parasites ( Figure 1B ) . However at 24 h and 48 h after miltefosine treatment the extent of apoptotic BHU 575 ( R ) was 32 . 8% and 60 . 7% respectively ( Inset Figure 1B ) . Similar studies were performed with lesion derived purified amastigotes , BHU 575 ( R ) and BHU 777 ( S ) to find that 8 h treatment with imipramine , but not miltefosine induced a significant decrease in ΔΨm ( Figure 1C ) . The replication of intracellular LD in the presence of imipramine was studied in in vitro infected MΦ . It was observed that intracellular LD replication was inhibited very efficiently as a function of the imipramine concentration regardless of the SSG sensitivity ( Figure 2 ) . The dose required to clear 100% of the intracellular parasites was around 60 µM of imipramine . To show that 60 µM of the drug has no toxic effect; MΦs were incubated with increasing concentration of imipramine . It was observed that almost 100% MΦs remained viable upto 90 µM imipramine ( Figure 2 , inset ) . It is known that infected MΦs are more fluid than their normal counterpart , and this is associated with defective T cell stimulating ability [32] . For convenience BHU 575 ( R ) infected MΦ were defined as MΦ-575 ( R ) . To show that imipramine restores membrane rigidity , we treated MΦ-575 ( R ) with increasing dose of imipramine and observed that there was a gradual increase in fluorescence anisotropy ( FA ) value in a dose dependent manner ( Figure 3A ) . To show that imipramine treatment restores the antigen presenting ability , MΦ-575 ( R ) were used as antigen presenting cells ( APC ) with and without imipramine treatment . This showed that the T-cell stimulating ability of MΦ-575 ( R ) is improved as a function of imipramine concentration as evident from the increase in resulting IL-2 production from I-Ad restricted T-cell hybridoma ( Figure 3B ) . ROS ( Reactive Oxygen Species ) and NO ( Nitric oxide ) are two very important leishmanicidal molecules [34] . Generation of these molecules was found to be enhanced in a time and dose dependent manner in imipramine treated MΦs ( Figure 4 ) . ROS generation reached a plateau at around 8 h in the presence of 75 µM imipramine treatment ( Figure 4A ) , whereas maximum NO generation was observed after 20 h exposure at the same concentration of imipramine ( Figure 4B ) . The effect of graded doses of orally administered imipramine on the splenic and hepatic parasite load in infected hamsters was investigated . Hamsters were infected with BHU 816 ( S ) , BHU 777 ( S ) , BHU 814 ( R ) , or BHU 575 ( R ) LD isolates . We performed a microscopic evaluation of stamp smears and limiting dilutions to detect parasites in tissue samples of infected organs . Eight-week infected hamsters were divided into 4 groups ( I–IV ) for a given isolate . Groups I–IV received imipramine at doses of 0 , 0 . 05 , 0 . 5 and 5 mg/kg/day respectively for 4 weeks . The results were expressed as total parasite load in terms of LDU . There was no clearance of splenic and hepatic parasite load in group ΙΙ whereas about 50% clearance was observed in group ΙΙΙ animals and there were no detectable parasites in group ΙV animals ( Figure 5 A–D ) . The organ parasite clearance essentially showed similar trends after imipramine treatment for all the isolates regardless of their SSG sensitivity . To show that group IV hamsters were indeed infected with LD , antileishmanial antibodies were measured in the animals . The presence of antileishmanial IgG2 antibodies in the sera of these animals was detected together with a marginal increase in anti IgG1 titer ( Figure 5 A–D c ) . We used miltefosine as a reference drug and studied its effect on the organ parasite clearance in a similar set up . Here also eight-week infected hamsters were subjected to miltefosine treatment ( 17 . 5 mg/kg/day ) orally for 4 weeks and splenic and hepatic parasite load were determined 2 days after completion of the last treatment dose ( Figure 5 A–D Group V ) . The dose of miltefosine was selected as described elsewhere [5] . This showed that miltefosine treated hamsters infected with either BHU 575 ( R ) or BHU 814 ( R ) had low level of residual parasites in the spleen and liver ( Figure 5 A–B ) whereas miltefosine treated hamsters infected with BHU 816 ( S ) or BHU 777 ( S ) showed no residual parasites ( Figure 5 C–D ) . The presence of residual parasites was further confirmed by limiting dilution experiments with spleen tissue ( Figure S1 ) . To study the status of antileishmanial T cell repertoire in infected and imipramine treated infected hamsters , splenocytes were purified and stimulated either with SLA or ConA . The hamsters were infected either with SSG-S ( BHU 816 and BHU 777 ) or SSG-R ( BHU 814 and BHU 575 ) LD parasites . After completion of imipramine treatment in 8 week infected hamsters , animals were sacrificed and splenocytes were prepared . Fixed concentrations of SLA and ConA were used to stimulate splenocytes as described previously [27] . Splenocytes of group ΙΙ ( received 0 . 05 mg/kg body weight ) hamsters failed to mount any antileishmanial immune response but responded well to non specific mitogen ConA regardless of the phenotype of the input parasites for infection ( Figure 6 ) . The SLA specific proliferation was marginally improved in group ΙΙΙ ( received 0 . 5 mg/kg body weight/animal ) , but was further improved in group ΙV ( receiving 5 mg/kg body weight/animal ) . The antileishmanial T cell response was essentially similar regardless of the phenotype of SSG sensitivity . The response to the non specific mitogen ConA remained unaltered in infected and in imipramine treated animals regardless of the dose of imipramine ( Figure 6 ) . The ability of cytokine and iNOS gene expression in infected hamsters and imipramine treated infected hamsters was studied by profiling cytokine gene expression ( Figure 7A ) . The results generated from a densitometry analysis of each hamster were collectively expressed as mean±sd for each cytokine , and the statistical significance between groups was determined ( Figure 7A ) . Comparative cytokine analysis showed that the expression of IFN-γ , TNF-α and iNOS transcripts were 1 . 43 , 1 . 23 , and 1 . 35 times higher , respectively , in imipramine treated hamsters than in infected hamsters , whereas the levels of TGF-β and IL-10 transcripts were 1 . 65 and 1 . 13-fold lower , respectively . Studies of the ratio of IFN-γ to TGF-β or IL-10 revealed that the IFN-γ/TGF-β ratio was 1 . 78-fold greater in imipramine treated hamsters than in infected hamsters ( Figure 7B ) . Similarly , the IFN-γ/IL-10 ratio was 1 . 47-fold greater in imipramine treated animals than in infected ones ( Figure 7B ) . To show that imipramine treated hamsters are also protected in the long run , we studied the survival kinetics of infected hamsters and imipramine treated infected hamsters ( 5 mg/kg/day for 4 weeks ) , using normal hamsters as control . In each group 30 hamsters were used . Hamsters were infected at 6 weeks and infection was allowed to proceed for another 8 week , i . e . before initiating any treatment . We observed that 80% of the infected hamsters survived up to 14 weeks , 60% up to 18 weeks , 20% up to28 weeks , and the rest died by 34 weeks . On the other hand , 90% of imipramine treated infected hamsters remained alive until the termination of the experiment , i . e . 44 weeks ( Figure 8 ) . Remarkably , amastigotes could not be detected by microscopy in impressions of Giemsa-stained tissue stamp smears of spleen and liver or by limiting dilution experiments at 34th week and also at 44th week . The organ weights had returned to near normal ( Table 2 lower panel ) , and high titers of antileishmanial IgG2 persisted at 34th week ( Figure 8 inset ) . Evolution of granuloma formation in infected and imipramine treated infected hamsters was studied . Liver section of LD infected hamsters showed immature granuloma formation associated with Kupffer cells surrounded by less number of infiltrating lymphocytes ( Figure 9A ) . High resolution figure of the same showed the presence of parasitized Kupffer cells ( Figure 9B ) inside the cell assembly . Imipramine treated infected liver tissue shows fair number of lymphocyte infiltration in periportal area and the presence of fair number of mature and uniform granuloma ( Figure 9C ) . Parasites could not be seen in these mature granulomas . The organ weight at 18 weeks increased significantly in untreated infected animals as compared to uninfected animals , but the 4 week imipramine treatment in infected group showed only a marginal decrease in organ weight ( Table 2 ) . After 34 weeks , the organ weight continued to increase marginally in the untreated infected group whereas in the imipramine treated group the organ weight was reduced by 50% as compared to the 18 week time point . However , at 44 weeks the organ weight of the imipramine treated infected group was similar to that of the age matched normal ( Table 2 ) .
In an effort to find new orally active chemotherapeutics for visceral leishmaniasis , we evaluated the potential of the existing antidepressant drug imipramine against SSG-R and SSG-S LD parasites . The potent in vitro activity of imipramine against intracellular amastigotes ( EC50 = 16 . 2 µM ) [14] as well as promastigotes ( Low IC50 values , Table 1 ) coupled with the absence of obvious cytotoxicity on host MΦ formed the basis for advanced exploration of this promising lead . Our study showed that imipramine decreases the mitochondrial transmembrane potential of SSG-S and SSG-R LD promastigotes significantly by 8 h of treatment which continued to decrease with time . In contrast , miltefosine showed only a marginal change in the mitochondrial transmembrane potential only at 24 h of treatment . Similar observation was made in imipramine or miltefosine treated purified amastigotes . Oddly enough , imipramine induced 60% of apoptosis of LD promastigote at 8 h whereas miltefosine at that stage induced only 5 . 5% apoptosis . Similar level of apoptosis by miltefosine was noted at 48 h . This clearly indicates that imipramine induces apoptosis in LD parasites much faster than miltefosine . This is in agreement with reports that apoptosis and change in mitochondrial potential are linked phenomenon [35] , [36] . LD infection is associated with increase of membrane fluidity [37] as well as defective antigen presentation by APC's [37] . Therefore the antigen presenting ability of imipramine treated MΦs ( also defined as APCs ) with murine T cells was studied . We observed that successive doses of imipramine restored membrane rigidity , which was in turn coupled with improved antigen presentation ability ( Figure 3A & 3B ) . This may be attributed to the fact that due to the clearance of intracellular parasites upon imipramine treatment , MΦ may regain its normal fluidity . Due to structural similarities to some extent ( fused ring structure with side chain ) , imipramine can mimic cholesterol which acts as a cementing molecule to pack the lipid bilayer [38] . This may be another reason for the restoration of membrane fluidity . Imipramine has already been reported to mimic the action of cholesterol to regulate protein synthesis in SREB ( Sterol Regulatory Element Binding ) protein synthesis pathway [39] . We have observed that imipramine induces production of leishmanicidal molecules such as superoxide and nitric oxide in MΦ ( Figure 4A & 4B ) . This observation complements other studies that imipramine induced production of TNF-α [12] , an important cytokine for antileishmanial defense . The therapeutic role of imipramine in experimental VL infection may be most likely be attributed to a direct leishmanicidal activity both in vitro and in vivo , its capacity to modulate the host immune response in favour of the host and its ability to induce reactive oxygen species generation on MΦ . Since imipramine at a dose of 60 µM completely clears the intracellular amastigotes in an in vitro macrophage system ( Figure 2 ) , we tested the efficiency of imipramine in an in vivo model . At a dose of 5 mg/kg/day for 4 weeks , the drug indeed clears >99 . 5% parasites in hamster infected with either SSG-R or SSG-S LD . We chose to administer oral treatment of imipramine for 4 weeks based on the fact that in humans miltefosine is given orally for 4 weeks [40] . The drug imipramine was found to be equally effective in a murine model where infection was induced by SSG-R parasites ( Unpublished observation ) . The in vivo dose of miltefosine was determined based on the animal equivalent of the human dose as described elsewhere [29] . On the other hand , miltefosine treatment for 4 weeks at the dose 17 . 5 mg/kg/day showed complete clearance of SSG-S but not SSG-R parasites from the spleen and liver of infected hamsters ( Figure 5 A–D Group V ) . The residual SSG-R parasites after miltefosine treatment , even though their number was low , raise concern of cross resistance between miltefosine and SSG . This is perhaps not so surprising since infection with SSG-R parasites upregulates an ABC transporter in the host cells [14] that regulates efflux of both the drugs [17] , [41] . Recently there is a great deal of interest for short course of combination treatment regimens [42] . As such , a 10-day treatment protocol has been proposed as an ideal short term regimen by the Drugs for Neglected Diseases Initiative ( DNDi ) [43] . The DNDi also aims to study new indications for existing medicines in the field of the most neglected diseases . In tune with these recommendations , we also tested the efficacy of imipramine as a monotherapy for a shorter version of treatment , i . e . for 10 days . We again opted to select the dose that provided maximum protection , i . e . 5 mg/kg/day . This regimen showed to clear about 90% of the organ parasites in infected hamsters ( Data not shown ) . Importantly , there is still opportunity to increase the dose of the drug based on the rodent equivalent of human dose [27] , also to opt for combination treatment . Imipramine at 5 mg/kg/day for 4 weeks does not affect the hepatic enzymes or creatinine levels in hamsters ( Unpublished observation ) , suggesting that the dose might even be increased further . Elicitation of effective T cell based host immune response defines the success of antileishmanial chemotherapeutics [44] . Disease severity in experimental animal models infected with SSG-R strains was associated with significantly hampered antigen presentation; antigen-specific T cell activation , low expression of IL-12 , TNF-α and IFN-γ , and upregulation of suppressive cytokines IL-10 and TGF-β in murine and hamster models respectively . We therefore tested the immunological parameters associated with successful chemotherapy in SSG-R LD infected animals treated with various therapies . Enhanced antigen presentation by imipramine treated APCs are also reflected in antigen specific expansion of T cell repertoire in vivo . The skewing of the T cell repertoire towards a Th1 type population is substantiated by the elevated level of IFN γ mRNA expression in splenocytes derived from imipramine treated hamsters . In vivo treatment with imipramine markedly increased the level of IL-12 in mRNA . The established phase of VL is associated with deactivation of MΦ with severely reduced capacity for production of inflammatory mediators like IL-12 and TNF-α . IL-12 interacts with T cells and induces the initiation and maintenance of Th1 responses via IFN-γ production . IFN-γ and TNF-α are often reported to act synergistically to activate iNOS for the production of NO , the leishmanicidal effector molecule [45] . Strong IL-12 driven IFN-γ coupled with TNF-α triggering by imipramine treatment suggests that these cytokines might be acting in concert to produce NO to effectively kill the parasites . A growing body of literature correlates IL-10 and TGF-β with susceptibility to Leishmania infection [46]–[54] . Imipramine caused strong suppression of IL-10 and TGF-β production that correlated with successful resolution of infection . Recent reports suggest that resistant parasites modulate the host immunity to exacerbate the ongoing disease pathogenicity [55] . TGF-β is implicated as an important contributor to disease susceptibility or resistance to Leishmania by direct MΦ deactivation [56] and also by increased production of IL-10 [57] . TGF-β being a pleotropic cytokine also suppresses IFN-γ-induced MHC class II expression by inhibiting class II transactivator mRNA [58] . We found significant down regulation of TGF-β mRNA expression in imipramine treated hamsters compared to infected controls . While imipramine treatment attenuates TGF-β expression , it could be causally related to a simultaneous inhibition of IL-10 production [46] with concurrent rescue of MΦ deactivation , thus implying the importance of both the cytokines in disease progression . We like to emphasize the possible role of TGF-β in the outcome of LD infection in hamsters because the IFN-γ/IL-10 ratio changed ( 1 . 47 fold ) compared to the IFN-γ/TGF-β ratio ( 1 . 78 fold ) upon imipramine treatment . The study of the survival kinetics of infected hamsters showed clearly that imipramine treatment increased the life expectancy of the infected hamsters with 90% . In the remainder 10% of the hamsters , death occurred the early time point , the cause of which is not clear . The remaining 90% remained healthy until the termination of the experiment , i . e . 44 weeks post infection . We determined the organ weight ( spleen and liver ) at 34 weeks and 44 weeks ( Table 2 ) and the parasite burden of the imipramine treated group . Surprisingly , parasites could not be detected in imipramine treated hamsters at the 34th nor at the 44th week and the organ weights were close to normal at 44 weeks post infection . These hamsters at the time of termination of experiment failed to show any parasites but displayed the presence of antileishmanial antibodies , indicating that they were indeed exposed to parasites . Efficient immune response in the liver depends on the formation of granulomas [59] , which is associated with the resolution of hepatic parasite burden [60] , [61] . It is well documented that only mature granuloma can develop efficient leishmanicidal mechanism to kill parasites whereas developing immature granulomas lack that efficiency [62] , [63] . Parasite killing within the granulomas requires infiltrating monocytes and TNF α [64] , although their formation is independent of TNF α family of cytokines [65] . Our previous study with KMP-11 vaccinated hamsters reveals well formed granuloma formation and absence of LD infected Kupffer cells [27] . Our present study showed matured sterile granuloma formation in 4 week of imipramine treated infected hamsters , which was absent in the 12 week infected group ( Figure 9 ) and is associated with the protection . It may be recalled that enhanced maturation of granulomas represents a marker of vaccine induced protection [66] . Orally administered imipramine is rapidly absorbed in the gastro-intestinal tract [9] . Imipramine is a lipophilic compound , binds to albumin , and attains a peak plasma concentration within 2–6 h . This tertiary amine is typically metabolized by demethylation to the secondary and active metabolite , desipramine [9] . Both imipramine and its metabolite desipramine have been found to be equally effective against LD promastigotes [14] . Resistance to SSG is a major problem in the Indian subcontinent and MΦs upregulate both MRP-1 and P-gp upon infection with SSG-R LD leading to efflux of antimonials [15] . Furthermore , circulatory monocytes of kala-azar patients harboring SSG-R LD show over expression of P-gp and MRP-1 [67] . SSG in combination with pharmacological inhibitors of MRP-1 and P-gp favors killing of intracellular SSG-R LD [68] . The tricyclic imipramine is lipophilic and possesses a positive charge due to the nitrogen atom , characteristics that are important to affect the function of P-gp [69] In MDR gene transfected and also in human AML cells ex vivo , such drugs reverse the multidrug resistance phenotype [69] . Thus imipramine will offer an additional advantage since it is a selective P-gp inhibitor . Furthermore , cationic amphiphilic drugs that is basic ( pKa 7–8 ) concentrates on lysosomes [13] . Imipramine being a tertiary amine and weak base will remain as positively charged molecular entity in the body fluids , leading to an affinity towards lysosomes [13] , [70] . This unique property is of importance because phagolysosomes constitute the home for intracellular Leishmania parasites . In conclusion , our study clearly indicated that imipramine is more effective than miltefosine and has a strong potential to be considered as an orally active , highly effective , very cheap , affordable chemotherapeutic agent against kala-azar either alone or in combination .
|
The disease Kala-azar or visceral leishmaniasis is still a big problem in the Indian subcontinent . The antimonials were used for the chemotherapy of Kala-azar but with time its efficacy has reduced dramatically . The newer version of orally active drug miltefosine has been introduced , but its efficacy has decreased considerably as relapse cases are on the rise . Other drugs like liposomal form of amphotericin B is expensive and the patients require hospitalization . Thus there is a genuine need for an orally active antileishmanial drug . There are reports that the cationic amphiphilic molecule , imipramine , a drug used for the treatment of depression in humans , kills the promastigotes of Leishmania donovani . We tested the efficacy of imipramine in experimental infection in hamster and mouse model . Our study showed that the drug is highly effective against antimony sensitive and antimony resistant Leishmania donovani infected hamsters as well as mouse and offered almost sterile cure .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases"
] |
2012
|
Imipramine Is an Orally Active Drug against Both Antimony Sensitive and Resistant Leishmania donovani Clinical Isolates in Experimental Infection
|
The guidance cue UNC-6/Netrin regulates both attractive and repulsive axon guidance . Our previous work showed that in C . elegans , the attractive UNC-6/Netrin receptor UNC-40/DCC stimulates growth cone protrusion , and that the repulsive receptor , an UNC-5:UNC-40 heterodimer , inhibits growth cone protrusion . We have also shown that inhibition of growth cone protrusion downstream of the UNC-5:UNC-40 repulsive receptor involves Rac GTPases , the Rac GTP exchange factor UNC-73/Trio , and the cytoskeletal regulator UNC-33/CRMP , which mediates Semaphorin-induced growth cone collapse in other systems . The multidomain flavoprotein monooxygenase ( FMO ) MICAL ( Molecule Interacting with CasL ) also mediates growth cone collapse in response to Semaphorin by directly oxidizing F-actin , resulting in depolymerization . The C . elegans genome does not encode a multidomain MICAL-like molecule , but does encode five flavin monooxygenases ( FMO-1 , -2 , -3 , -4 , and 5 ) and another molecule , EHBP-1 , similar to the non-FMO portion of MICAL . Here we show that FMO-1 , FMO-4 , FMO-5 , and EHBP-1 may play a role in UNC-6/Netrin directed repulsive guidance mediated through UNC-40 and UNC-5 receptors . Mutations in fmo-1 , fmo-4 , fmo-5 , and ehbp-1 showed VD/DD axon guidance and branching defects , and variably enhanced unc-40 and unc-5 VD/DD axon guidance defects . Developing growth cones in vivo of fmo-1 , fmo-4 , fmo-5 , and ehbp-1 mutants displayed excessive filopodial protrusion , and transgenic expression of FMO-5 inhibited growth cone protrusion . Mutations suppressed growth cone inhibition caused by activated UNC-40 and UNC-5 signaling , and activated Rac GTPase CED-10 and MIG-2 , suggesting that these molecules are required downstream of UNC-6/Netrin receptors and Rac GTPases . From these studies we conclude that FMO-1 , FMO-4 , FMO-5 , and EHBP-1 represent new players downstream of UNC-6/Netrin receptors and Rac GTPases that inhibit growth cone filopodial protrusion in repulsive axon guidance .
The formation of neural circuits during development depends on the guidance of growing axons to their proper synaptic targets . This process relies on the growth cone , a dynamic actin based structure present at the tip of a growing axon . Growth cones contain a dynamic lamellipodial body ringed by filopodial protrusions , both important in guiding the axon to its target destination [1–4] . Guidance receptors present on the leading edge of the growth cone sense and respond to various extracellular guidance cues , which attract or repel axons enabling them to reach their proper target destination [5 , 6] . The secreted laminin-like guidance molecule UNC-6/Netrin mediates both axon attraction and axon repulsion and defines a dorsal-ventral guidance mechanism conserved from invertebrates to vertebrates [7–9] . Attractive or repulsive responses to UNC-6/Netrin depend on the receptors expressed on the growth cone . Homodimers of the UNC-6/Netrin receptor UNC-40/DCC mediate attraction , and UNC-5:UNC-40 heterodimers or UNC-5 homodimers mediate repulsion [10–12] . In C . elegans , UNC-6/Netrin is secreted by the ventral cells and along with its receptors UNC-40 and UNC-5 is required for the dorsal ventral guidance of circumferential neurons and axons [8 , 13 , 14] . Previous studies of repelled VD growth cones in Netrin signaling mutants revealed a correlation between attractive axon guidance and stimulation of growth cone protrusion , and repulsive axon guidance and inhibition of growth cone protrusion [15] . For example , in unc-5 mutants , growth cones were larger and more protrusive , and often displayed little or no directed movement . This is consistent with observation that increased growth cone size was associated with decreased neurite growth length [16] . Conversely , constitutive activation of UNC-5:UNC-40 signaling in repelled VD growth cones led to smaller growth cones with severely reduced filopodial protrusion [15 , 17] . Thus , directed growth cone repulsion away from UNC-6/Netrin requires a balance of pro- and anti-protrusive activities of the receptors UNC-40 and UNC-40:UNC-5 , respectively , in the same growth cone [15] . Genetic analysis has identified a cytoskeletal signaling pathway involved in stimulation of growth cone protrusion in response to the attractive UNC-40 signaling that includes CDC-42 , the Rac-specific guanine nucleotide exchange factor TIAM-1 , the Rac-like GTPases CED-10 and MIG-2 , as well as the cytoskeletal regulators Arp2/3 and activators WAVE-1 and WASP-1 , UNC-34/Enabled , and UNC-115/abLIM [18–23] , consistent with findings in other systems [7] . Mechanisms downstream of UNC-5 in axon repulsion are less well described , but the PH/MyTH4/FERM molecule MAX-1 and the SRC-1 tyrosine kinase have been implicated [24 , 25] . We delineated a new pathway downstream of UNC-5 required for its inhibitory effects on growth cone protrusion , involving the Rac GEF UNC-73/Trio , the Rac GTPases CED-10 and MIG-2 , and the cytoskeletal-interacting molecule UNC-33/CRMP [17] . Collapsin response mediating proteins ( CRMPs ) were first identified as mediators of growth cone collapse in response to the Semaphorin family of guidance cues [26] , and we have shown that UNC-33/CRMP inhibits growth cone protrusion in response to Netrin signaling [17] . This motivated us to consider other mediators of Semaphorin-induced growth cone collapse in Netrin signaling . In Drosophila , the large multidomain cytosolic protein MICAL ( Molecule Interacting with CasL ) is required for the repulsive motor axon guidance mediated by interaction of Semaphorin 1a and Plexin A [27 , 28] . MICAL proteins are a class of flavoprotein monooxygenase enzymes that bind flavin adenine dinucleotide ( FAD ) and use the cofactor nicotinamide dinucleotide phosphate ( NADPH ) to facilitate oxidation-reduction ( Redox ) reactions [27] . MICAL regulates actin disassembly and growth cone collapse in response to semaphorin via direct redox interaction with F-actin [29 , 30] . MICAL molecules from Drosophila to vertebrates have a conserved domain organization: an N-terminal flavin-adenine dinucleotide ( FAD ) -binding monooxygenase domain , followed by a calponin homology ( CH ) domain , a LIM domain , a proline-rich domain , and a coiled-coil ERM α-like motif [27 , 31] . The C . elegans genome does not encode a MICAL-like molecule with the conserved domain organization described above . However , it does encode five flavin monooxygenase ( fmo ) genes similar to the Flavin monooxygenase domain of MICAL: fmo-1 , fmo-2 , fmo-3 , fmo-4 and fmo-5 [32] . Like MICAL , the C . elegans FMO molecules contain an N-terminal FAD binding domain and a C-terminal NADP or NADPH binding domain [27 , 32] . The C . elegans gene most similar to the non-FMO portion of MICAL is the Eps-15 homology domain binding protein EHBP-1 [33] , which contains a CH domain as does MICAL . In this work , we test the roles of the C . elegans FMOs and EHBP-1 in Netrin-mediated axon guidance and growth cone protrusion . We find that fmo-1 , fmo-4 , fmo-5 and ehbp-1 mutants display pathfinding defects of the dorsally-directed VD/DD motor neuron axons that are repelled by UNC-6/Netrin , and that they interact genetically with unc-40 and unc-5 . We also find that VD growth cones in these mutants display increased filopodial protrusion , similar to mutants in repulsive UNC-6/Netrin signaling ( e . g . unc-5 mutants ) , and that transgenic expression of FMO-5 inhibits growth cone protrusion , similar to constitutively-activated UNC-40 and UNC-5 . We also show that FMO-1 , FMO-4 , FMO-5 and EHBP-1 are required for the growth cone inhibitory effects of activated UNC-5 , UNC-40 , and the Rac GTPases CED-10 and MIG-2 . Together , these genetic analyses suggest that FMO-1 , FMO-4 , FMO-5 , and EHBP-1 normally restrict growth cone protrusion , and that they might do so in UNC-6/Netrin-mediated growth cone repulsion .
Experiments were performed at 20°C using standard C . elegans techniques [34] . Mutations used were LGI: unc-40 ( n324 ) , unc-73 ( rh40 ) ; LGII: juIs76[Punc-25::gfp]; LGIII: fmo-3 ( gk184651 ) ; LGIV: fmo-1 ( ok405 ) , fmo-2 ( ok2147 ) , lqIs128[Punc-25::myr::unc-40] , unc-5 ( op468 and e152 ) , unc-33 ( e204 ) ; LGV: fmo-4 ( ok294 ) , fmo-5 ( tm2438 ) , ehbp-1 ( ok2140M+ ) ; LGX: lqIs182[Punc-25::mig-2 ( G16V ) ] . Chromosomal locations not determined: lqIs129[Punc-25::myr::unc-40] , lqIs296[Punc-25::myr::unc-5] , lqIs204[Punc-25::ced-10 ( G12V ) ] , lhIs6[Punc-25::mCherry] , lqIs311[fmo-5 genomic] by integration of lqEx1047 . Extrachromosomal arrays were generated using standard gonadal injection [35] and include: lqEx901 and lqEx931[Pehbp-1::gfp , Pgcy-32::yfp]; lqEx1014 , lqEx1015 , lqEx1016 , lqEx1045 , lqEx1046 and lqEx1047[Pfmo-5::fmo-5 , Pgcy-32::yfp]; lqEx949 , lqEx950 , lqEx951 , lqEx1053 , lqEx1054 and lqEx1055[Punc-25::fmo-1 , Pgcy-32::yfp]; lqEx1098 , lqEx1099 , lqEx1100 , lqEx1101 and lqEx1102[Punc-25::fmo-4 , Pgcy-32::yfp]; lqEx952 , lqEx953 , lqEx954 , lqEx1061 , lqEx1062 , lqEx1063 , lqEx1078 , lqEx1079 and lqEx1080[Punc-25::fmo-5 , Pgcy-32::yfp]; lqEx1146 , lqEx1147 , lqEx1148 and lqEx1149[Pdpy-7::fmo-5 , Pgcy-32::yfp]; lqEx1113 and lqEx1114[Pfmo-5::fmo-5::GFP , Pstr-1::gfp]; whEx28[Pfmo-4::gfp , pRF4/rol-6] . Multiple ( ≥3 ) extrachromosomal transgenic lines of Pfmo-5::fmo-5 for overexpression data of fmo-5 were analyzed with similar effect , and one was chosen for integration and further analysis . Genotypes containing M+ indicate that homozygous animals from a heterozygous mother were scored . The ehbp-1 ( ok2140M+ ) strain was balanced with the nT1 balancer , since ehbp-1 ( ok2140 ) homozygous animals are sterile . RNAi was performed by feeding as previously described [36] . Details about transgene construction are available by request . Punc-25::fmo-1 , Punc-25::fmo-4 and Punc-25::fmo-5 were made using the entire genomic regions of fmo-1 , fmo-4 and fmo-5 respectively . Expression analysis for fmo-5 was done by amplifying the entire genomic region of fmo-5 along with its endogenous promoter ( 1 . 2kb upstream ) and fusing it to gfp followed by the 3’ UTR of fmo-5 . VD neurons were visualized with a Punc-25::gfp transgene , juIs76 [37] , which is expressed in GABAergic neurons including the six DDs and 13 VDs , 18 of which extend commissures on the right side of the animal . The commissure on the left side ( VD1 ) was not scored . In wild-type , an average of 16 of these 18 VD/DD commissures are apparent on the right side , due to fasciculation of some of the commissural processes . In some mutant backgrounds , fewer than 16 commissures were observed ( e . g . unc-5 ) . In these cases , only observable axons emanating from the ventral nerve cord were scored for axon guidance defects . VD/DD axon defects scored include axon guidance ( termination before reaching the dorsal nerve cord or wandering at an angle greater than 45° before reaching the dorsal nerve cord ) and ectopic branching ( ectopic neurite branches present on the commissural processes ) . In the case of double mutant analysis with unc-40 and unc-5 only lateral midline crossing ( axons that fail to extend dorsally past the lateral midline ) were considered . Fisher's exact test was used to determine statistical significance between proportions of defective axons . In double mutant comparisons , the predicted additive effect of the mutants was calculated by the formula P1+P2- ( P1*P2 ) , where P1and P2 are the phenotypic proportions of the single mutants . The predicted additive effect of single mutants was used in statistical comparison to the observed double mutant effect . VD growth cones were imaged as previously described [15 , 22] . Briefly , animals at 16 h post-hatching at 20°C were placed on a 2% agarose pad and paralyzed with 5mM sodium azide in M9 buffer , which was allowed to evaporate for 4 min before placing a coverslip over the sample . Some genotypes were slower to develop than others , so the 16 h time point was adjusted for each genotype . Growth cones were imaged with a Qimaging Rolera mGi camera on a Leica DM5500 microscope . Projections less than 0 . 5 µm in width emanating from the growth cone were scored as filopodia . Filopodia length and growth cone area were measured using ImageJ software . Filopodia length was determined by drawing a line from the base where the filopodium originates on the edge of the peripheral membrane to the tip of the filopodium . Growth cone area was determined by tracing the periphery of the growth cone , not including filopodial projections . Significance of difference was determined a two-sided t-test with unequal variance .
The C . elegans genome lacks an apparent homolog of MICAL . However , it contains five flavin monooxygenase genes ( fmo-1 , 2 , 3 , 4 , 5 ) ( Fig 1A ) [32] . The C . elegans molecule most similar to the non-FMO portion of MICAL is EHBP-1 , the homolog of the mammalian EH domain binding protein 1 ( Ehbp1 ) protein [33] . We analyzed existing mutations in fmo genes and ehbp-1 ( Fig 1B ) for VD/DD axon guidance defects . fmo-1 ( ok405 ) was a 1 , 301-bp deletion that removed part of exon 3 and all of exons 4 , 5 and 6 . fmo-2 ( ok2147 ) was a 1070-bp deletion that removed part of exon 4 and 5 . fmo-4 ( ok294 ) was a 1490-bp deletion that removed all of exons 2 , 3 , 4 and 5 . fmo-5 ( tm2438 ) is a 296-bp deletion which removes part of intron 3 and exon 4 . These deletions all affected one or both predicted enzymatic domains of the FMO molecules . fmo-3 ( gk184651 ) was a G to A substitution in the 3’ splice site of intron 6 and may not be a null mutation . ehbp-1 ( ok2140 ) is a 1 , 369-bp deletion that removed all of exon 5 and 6 . Genotypes involving ehbp-1 ( ok2140 ) have wild-type maternal ehbp-1 activity , as homozygotes are sterile and must be maintained as heterozygotes balanced by nT1 . The 19 D-class motor neurons cell bodies reside in the ventral nerve cord . They extend axons anteriorly and then dorsally to form a commissure , which normally extend straight dorsally to the dorsal nerve cord ( Fig 2 and Fig 3B ) . On the right side of wild-type animals , an average of 16 commissures were observed , due to the fasciculation of some processes as a single commissure ( Fig 2C and Materials and Methods ) . fmo-1 , 4 , and 5 and ehbp-1 mutants showed significant defects in VD/DD axon pathfinding , including ectopic axon branching and wandering ( ~3–5%; see Materials and Methods and Fig 3A , 3C and 3D ) , although most crossed the lateral midline despite wandering . fmo-2 and fmo-3 mutations showed no significant defects compared to wild-type ( Fig 3A ) . We used RNAi directed against ehbp-1 , which has the potential to eliminate any maternally-supplied mRNAs , but not translated proteins . ehbp-1 ( RNAi ) resembled ehbp-1 M+ mutants ( Fig 3A ) , suggesting that maternal mRNAs are not involved in VD/DD axon guidance . Most double mutants showed no strong synergistic defects compared to the predicted additive effects of the single mutants ( Fig 3A ) . However , the fmo-2; fmo-3 and the fmo-2; fmo-4 double mutants showed significantly more defects compared to the predicted additive effects of the single mutants . The fmo-4; ehbp-1 double mutant displayed significantly reduced defects than either mutation alone . The fmo-1; fmo-4 fmo-5 triple mutant also showed no synergistic defects as compared to single mutants alone ( Fig 3A ) . Lack of extensive phenotypic synergy suggests that the FMOs do not act redundantly , but rather that they might have discrete and complex roles in axon guidance , as evidenced by fmo-4; ehbp-1 mutual suppression . In unc-40 ( n324 ) strong loss-of-function mutants , most axons ( 92% ) extended past the lateral midline despite wandering ( see Materials and Methods and Fig 4A and 4B ) . fmo-1 , fmo-4 , fmo-5 , and ehbp-1 displayed < 1% failing to extend past the lateral midline ( Fig 4A ) . fmo-1 , fmo-4 , and fmo-5 mutations significantly enhanced the VD/DD lateral midline crossing defects of unc-40 ( n324 ) ( Fig 4A and 4C ) . ehbp-1 did not enhance unc-40 ( Fig 4A ) . unc-5 ( e53 ) strong loss-of-function mutants display a nearly complete failure of VD axons to reach the dorsal nerve cord [13 , 15] . unc-5 ( e152 ) is a hypomorphic allele [38] and displayed 22% failure of axons to cross the lateral midline ( Fig 5A ) . The unc-5 ( op468 ) allele [39] also displayed a weaker lateral midline crossing phenotype ( 10% ) , indicating that it is also a hypomorphic allele ( Fig 5B ) . fmo-1 , fmo-4 and fmo-5 significantly enhanced the VD/DD axon guidance defects of both unc-5 ( e152 ) and unc-5 ( op468 ) , but ehbp-1 did not ( Fig 5 ) . While fmo mutations alone caused few midline crossing defects compared to unc-40 and unc-5 , they enhanced the midline crossing defects of unc-40 and unc-5 hypomorphic mutants . These results indicate that FMO-1 , 4 , and 5 might act with UNC-40 and UNC-5 in VD/DD axon pathfinding . Expression of the fmo-1 , fmo-4 , and fmo-5 coding regions were driven in VD/DD motor neurons using the unc-25 promoter . Punc-25::fmo transgenes significantly rescued lateral midline crossing defects in fmo; unc-5 ( op468 ) and fmo; unc-5 ( e152 ) ( Fig 6 ) . These data suggest that the axon defects observed in fmo mutants are due to mutation of the fmo genes themselves , and that fmo-1 , 4 , and 5 can act cell-autonomously in the VD/DD neurons in axon guidance . Previous studies showed that fmo-1 and fmo-5 promoter regions were active in intestinal cells and the excretory gland cell , whereas the fmo-4 promoter was active in hypodermal cells , duct and pore cells [32 , 40] . ehbp-1 is expressed in all somatic cells including neurons [33] . Furthermore , cell-specific transcriptome profiling indicated that fmo-1 , fmo-4 and fmo-5 were expressed in embryonic and adult neurons , including motor neurons [41–43] . We fused the upstream promoter regions of fmo-1 , fmo-4 , and fmo-5 to gfp . We could observe no fmo-1::gfp expression in transgenic animals , in contrast to previous studies using a LacZ reporter [32] . However , transcriptome profiling indicates neuronal expression of fmo-1 [43] . Our fmo-1::gfp transgene might be missing regulatory regions required for expression . fmo-4::gfp was expressed strongly in hypodermal cells , excluding the seam cells and vulval cells , consistent with previous studies [32] ( Fig 6D ) . We also observed fmo-4::gfp expression in cells in the ventral nerve cord ( Fig 6D and 6D’ ) . Pfmo-5::fmo-5::gfp was expressed strongly in the intestine as previously reported [32] ( Fig 6E ) . We also observed expression along the ventral nerve cord ( Fig 6E and 6E’ ) . In sum , previous expression studies combined with those described here suggest that fmo-1 , 4 , 5 and ehbp-1 are expressed in neurons , and that fmo-1 , 4 , and 5 can act cell-autonomously in the VD/DD motor neurons in axon guidance . The growth cones of dorsally-directed VD commissural axons are apparent in early L2 larvae ( Fig 2B ) . We imaged VD growth cones at 16 hours post-hatching , when the VD growth cones have begun their dorsal migrations , as described previously [15] . fmo-1 , fmo-4 , and fmo-5 mutant growth cones displayed longer filopodial protrusions compared to wild type ( e . g . 0 . 96 µm in wild type compared with 1 . 55 µm in fmo-5 ( tm2438 ) ; p < 0 . 001 ) ( Fig 7 ) . This effect was not significant in ehbp-1 ( ok2140 ) ( Fig 7 ) . Growth cone area was not significantly different in any mutant . These results suggest that fmo-1 , fmo-4 and fmo-5 normally limit growth cone filopodial protrusion length . This is consistent with ectopic axon branches observed in post-development VD/DD neurons in these mutants ( Fig 3 ) , as other mutants with increased growth cone filopodial protrusions ( e . g . unc-5 , unc-73 , unc-33 ) also display ectopic branches , likely due to failure of filopodial retraction and subsequent consolidation into a neurite [15 , 17] . Expression of the fmo-5 coding region driven in VD/DD motor neurons using the unc-25 promoter also significantly rescued axon guidance defects as well as the long filopodial protrusions seen in fmo-5 ( tm2438 ) ( Fig 8 ) . We expressed the fmo-5 coding region in the hypodermis using the dpy-7 promoter [44] and observed no significant rescue of axon guidance defects or filopodial protrusions ( Fig 8 ) . These data confirm that fmo-5 can act cell-autonomously in the VD/DD neurons in axon guidance and growth cone filopodial protrusion . Previous studies showed that UNC-6/netrin signaling via the heterodimeric UNC-5:UNC-40 receptor leads inhibition of growth cone protrusion important in UNC-6/Netrin’s role in repulsive axon guidance [15 , 17] . Constitutive activation of this pathway using expression of myristoylated versions of the cytoplasmic domains of UNC-40 and UNC-5 ( myr::unc-40 and myr::unc-5 ) results in small growth cones with few if any filopodial protrusions ( i . e . protrusion is constitutively inhibited by MYR::UNC-40 and MYR::UNC-5 ) [15 , 17 , 18] . Loss of fmo-1 , fmo-4 , fmo-5 and ehbp-1 significantly suppressed inhibition of filopodial protrusion and growth cone size caused by myr::unc-40 ( Fig 9 ) and myr::unc-5 ( Fig 10 ) . Notably , ehbp-1 did not enhance loss-of-function mutations in unc-5 or unc-40 ( Fig 4 ) , suggesting that myr::unc-5 and myr::unc-40 are sensitized backgrounds in which interactions can be determined that are not apparent in loss-of-function backgrounds . Expression of activated CED-10 ( G12V ) and MIG-2 ( G16V ) in the VD neurons results in reduced growth cone protrusion similar to MYR::UNC-40 and MYR::UNC-5 [17] . We found that fmo-1 , fmo-4 and fmo-5 suppressed filopodial protrusion deficits caused by ced-10 ( G12V ) and mig-2 ( G16V ) ( Fig 11 ) . ehbp-1 suppressed mig-2 ( G16V ) , but ehbp-1 ( ok2140M+ ) ; ced-10 ( G12V ) double mutants were inviable and could not be scored . Furthermore , fmo-4 and fmo-5 , but not fmo-1 , significantly suppressed growth cone size reduction caused by CED-10 ( G12V ) and MIG-2 ( G16V ) . ehbp-1 also suppressed growth cone size reduction of MIG-2 ( G16V ) . Taken together , these data indicate that functional FMO-1 , FMO-4 , FMO-5 , and EHBP-1 are required for the full effect of MYR::UNC-40 , MYR::UNC-5 , CED-10 ( G12V ) , and MIG-2 ( G16V ) on growth cone protrusion inhibition , including filopodial protrusion and growth cone size . fmo-5 loss-of-function mutant growth cones displayed excessively-protrusive filopodia ( Fig 7 ) and suppressed activated UNC-5:UNC-40 and Rac signaling ( Figs 9–11 ) . Transgenic expression of wild-type FMO-5 driven by its endogenous promoter rescued the axon guidance defects and long filopodial protrusions seen in fmo-5 ( tm2438 ) mutant VD growth cones ( Fig 12A–12E ) . In a wild-type background , fmo-5 transgenic expression resulted in growth cones with smaller area and shortened filopodia ( Fig 13A , 13B and 13E ) , indicating that wild-type FMO-5 activity can inhibit growth cone protrusion . This inhibition was not observed in the fmo-5 ( tm2438 ) background , possibly due to the decreased levels of FMO-5 compared to the wild-type background . Mutations in unc-5 , unc-73 , and unc-33 result in excessively large growth cones with increased filopodial length ( Fig 13A and 13B ) [15 , 17] . Transgenic fmo-5 expression significantly reduced growth cone size and filopodial protrusion in unc-5 ( e152 ) , unc-73 ( rh40 ) , and unc-33 ( e204 ) ( Fig 13A , 13B , 13F , 13G and 13H ) . However , fmo-5 expression only partially inhibited filopodial protrusion in unc-5 and unc-33 ( i . e . to wild-type levels , higher than fmo-5 transgenic expression alone ) ( Fig 13A ) . These data indicate that FMO-5 activity does not rely on UNC-5 , UNC-73 , or UNC-33 and that it might act downstream of them . However , the hybrid interaction of fmo-5 transgenic expression with unc-33 ( e204 ) could also indicate that FMO-5 and UNC-33 represent distinct , compensatory pathways downstream of UNC-5 and the Rac GTPases to inhibit filopodial protrusion . unc-33; fmo-5 double mutants did not show any significant increase in filopodial length ( Fig 13A and 13B ) , which would be expected if they act in parallel pathways . In contrast to unc-5 mutants , unc-40 single mutants display shortened filopodial protrusions and a relatively normal growth cone size ( Fig 13A and 13B ) [15] . This is likely due to the dual role of UNC-40 in both stimulating protrusion as a homodimer and inhibiting protrusion as a heterodimer with UNC-5 . fmo-5 transgenic expression had no effect on filopodial protrusions in unc-40 , but did reduce growth cone size , consistent with a role of FMO-5 in inhibiting protrusion . Finally , we also found that transgenic fmo-5 expression in fmo-1 ( ok405 ) suppressed growth cone area and filopodial length of fmo-1 ( ok405 ) mutants ( Fig 13A , 13B and 13H ) , indicating that fmo-5 activity can partially compensate for loss of fmo-1 .
fmo-1 , fmo-4 , fmo-5 , and ehbp-1 mutants display defects in dorsal guidance of the VD/DD motor axons that are repelled from UNC-6/Netrin ( Fig 3 ) . Double and triple mutant analysis did not uncover significant redundancy , suggesting that these molecules might have discrete and complex roles in axon guidance . Consistent with this idea , fmo-4 and ehbp-1 mutually suppress VD/DD axon guidance defects . Furthermore , transgenic expression of FMO-5 rescued excess growth cone and filopodial protrusions of fmo-1 mutants . This suggests that FMO-5 can partially compensate for loss of FMO-1 , and that the function of FMO-5 does not depend on FMO-1 . Combined with lack of phenotypic synergy , these data suggest that the FMOs act in a common pathway , where loss of one abolishes pathway function , and that FMO-5 might act downstream of FMO-1 in this pathway . fmo-2 and fmo-3 mutations displayed no significant defects alone , suggesting that they are not involved in axon guidance . fmo-2 did significantly enhance fmo-4 . Possibly , fmo-2 and fmo-3 might have roles in axon guidance that were not revealed by the mutations used . Drosophila and vertebrate MICAL regulate actin cytoskeletal dynamics in both neuronal and non-neuronal processes through direct redox activity of the monooxygenase domain [27 , 30 , 46–50] . In Drosophila , loss of MICAL showed abnormally shaped bristles with disorganized and larger F-actin bundles , whereas , overexpression of MICAL caused a rearrangement of F-actin into a complex meshwork of short actin filaments [29] . Here we show that loss of fmo-1 , fmo-4 , and fmo-5 resulted in longer filopodial protrusions in the VD motor neurons ( Fig 7 ) , suggesting that their normal role is to limit growth cone filopodial protrusion . Indeed , transgenic expression of wild-type FMO-5 resulted in VD growth cones with a marked decrease in growth cone filopodial protrusion ( Fig 13 ) . Growth cone size was not affected in any loss-of-function mutation , but growth cone size was reduced by transgenic expression of wild-type FMO-5 ( Fig 13 ) , suggesting a role of the FMO-5 in both filopodial protrusion and growth cone lamellipodial protrusion . Previous studies have shown that Drosophila MICAL may require both its FMO and CH domain to induce cell morphological changes; however , mammalian MICAL in non-neuronal cell lines requires only its FAD domain suggesting a difference in the mechanism of action in these MICALs [29 , 51] . These data suggest that in some cases , the FMO domain is sufficient for the function of MICAL . Thus , single domain FMOs as in C . elegans could function despite lacking the multi-domain structure of MICAL . Loss of the C . elegans MICAL-like molecule EHBP-1 , which contains a CH domain and is similar to the non-FMO portion of MICAL ( Fig 1 ) , also resulted in VD/DD axon guidance defects , but did not significantly affect growth cone filopodial protrusion . EHBP-1 might act with the FMOs in axon guidance . Phenotypic differences could be due to EHBP-1-dependent and independent events , or to the wild-type maternal contribution in ehbp-1 homozygous mutants derived from a heterozygous mother . It is also possible that EHBP-1 affects axon guidance independently of the FMOs . EHBP-1 is involved in Rab-dependent endosomal vesicle trafficking by bridging interaction of endosomal Rabs with the actin cytoskeleton [33 , 52] . MICAL has also been implicated in Rab-dependent endosomal biogenesis and trafficking [53–55] , suggesting that FMO/EHBP-1 and MICALs might share common functions , although it remains to be determined if FMOs in C . elegans regulate endosomal trafficking . MICAL has been shown to directly oxidize cysteine residues in F-actin , leading to actin depolymerization and growth cone collapse [29 , 30 , 56 , 57] . We speculate that FMO-1 , FMO-4 , and FMO-5 might act by a similar mechanism to inhibit growth cone filopodial protrusion . Previous studies have shown that the single calponin homology ( CH ) domain containing protein CHDP-1 promotes the formation of cell protrusions in C . elegans by directly binding to Rac1/CED-10 through its CH domain[58] . The role of EHBP-1 however , is less clear , but previous studies have shown that Drosophila MICAL might require both its FMO and CH domain to induce cell morphological changes [29] . Thus , in axon guidance , FMO-1 , FMO-4 , and FMO-5 might require the CH domain provided by EHBP-1 in some instances . Mammalian MICAL requires only the FMO domain [51] , suggesting that in some cases the CH domain is not required and the FMO domain can act alone . Future studies will be directed at answering these questions . Expression of full length fmo-1 , fmo-4 and fmo-5 coding regions under the control of the unc-25 promoter specific for GABA-ergic neuron expression ( including the VD/DD neurons ) rescued VD/DD axon guidance defects ( Figs 6 and 8 ) . Furthermore , the promoters of fmo-4 and fmo-5 were active in ventral nerve cord cells ( Fig 6 ) . Expression of full length fmo-5 coding region under the control of the unc-25 promoter rescued axon guidance defects as well as the long filopodial protrusions seen in fmo-5 ( tm2438 ) , whereas expression from the hypodermal dpy-7 promoter did not ( Fig 8 ) . Cell-specific transcriptome profiling indicated that fmo-1 , fmo-4 and fmo-5 were expressed in embryonic and adult neurons , including motor neurons [41–43] . Together , these results suggest that the FMOs can act cell-autonomously in the VD/DD neurons in axon guidance and growth cone filopodial protrusion . Our findings suggest that the FMOs act with the UNC-40 and UNC-5 receptors to mediate UNC-6/netrin repulsive axon guidance and inhibition of growth cone protrusion . fmo-1 , fmo-4 , and fmo-5 mutations enhanced axon pathfinding defects in unc-40 and hypomorphic unc-5 mutants ( Figs 4 and 5 ) . The axon guidance defects of the fmos were weaker than those of unc-40 and unc-5 mutants ( e . g . the fmos displayed few lateral midline crossing defects despite axon wandering ) . We speculate that the FMOs are but one of several pathways mediating the effects of UNC-40 and UNC-5 in axon pathfinding . ehbp-1 did not enhance unc-40 or unc-5 , suggesting discrete roles of these molecules or wild-type maternal ehbp-1 contribution . fmo-1 , fmo-4 , fmo-5 , and ehbp-1 mutations each suppressed the effects of activated MYR::UNC-40 and MYR::UNC-5 on inhibition of growth cone protrusion ( Fig 10 ) . In this case , both filopodial protrusion and growth cone area was restored , consistent with a role of these molecules in inhibiting both growth cone filopodial and lamellipodial protrusion . We also find the fmo-5 transgenic expression suppressed unc-5 ( e152 ) growth cone area and filopodial protrusions ( Fig 13 ) . That the FMOs and EHBP-1 were required for the effects of the constitutively active MYR::UNC-40 and MYR::UNC-5 suggest that they act downstream of these molecules in growth cone inhibition of protrusion . While the loss-of-function and gain-of-function data are consistent with acting downstream of UNC-40 and UNC-5 , it is possible that the FMOs define a parallel pathway in growth cone protrusion . Interestingly , mutations in these genes have very distinct penetrances ( e . g . the axon guidance and protrusion defects of unc-5 are much stronger than those of the fmo mutants ) . One explanation for this is that these molecules act in networks rather than simple linear pathways . The FMOs might be one of many mechanisms acting downstream of UNC-5 , and multiple pathways might converge on UNC-33 ( e . g . UNC-5 and UNC-33 are major “nodes” in this network ) . Further loss- and gain-of-function studies will be required to understand this signaling network . Similar to activated MYR::UNC-40 and MYR::UNC-5 , constitutively-activated Rac GTPases CED-10 ( G12V ) and MIG-2 ( G16V ) inhibit VD growth cone protrusion . We show that fmo-1 , fmo-4 , fmo-5 and ehbp-1 mutations suppressed activated CED-10 ( G12V ) and MIG-2 ( G16V ) ( e . g . double mutant growth cones displayed longer filopodial protrusions similar to fmo-1 , fmo-4 , fmo-5 and ehbp-1 single mutants ) ( Fig 11 ) . Furthermore , loss of the Rac GTP exchange factor UNC-73/Trio had no effect on the inhibited growth cone phenotype of FMO-5 transgenic expression ( i . e . the growth cones resembled those of fmo-5 over expression alone ) ( Fig 13 ) . UNC-73/Trio acts with the Rac GTPases CED-10 and MIG-2 in growth cone protrusion inhibition , and unc-73 mutants display excessive growth cone protrusion [17] . That FMO-5 transgenic expression could inhibit protrusion in the absence of the Rac activator UNC-73/Trio suggests that FMO-5 acts downstream of UNC-73/Trio , consistent with the FMOs and EHBP-1 acting downstream of the Rac GTPases . Previous studies have shown that the C . elegans CRMP-like molecule UNC-33 is required in a pathway downstream of Rac GTPases for inhibition of growth cone protrusion in response to UNC-6/Netrin [17] . unc-33 loss-of-function mutants with FMO-5 transgenic expression displayed a mutually-suppressed phenotype . The excessively-long filopodial protrusions of unc-33 mutants were reduced to wild-type levels , but were significantly longer than in animals with FMO-5 transgenic expression , and the growth cone area was reduced to resemble FMO-5 transgenic expression alone ( Fig 13 ) . This hybrid phenotype makes it difficult to determine if FMO-5 and UNC-33 act in the same pathway , in parallel pathways , or both . One proposed mechanism of cytoskeletal regulation by MICAL is the production of the reactive oxygen species ( ROS ) H2O2 by the FAD domain in the presence of NADPH [59] . Upon activation by Sema3A , MICALs generate H2O2 , which can , via thioredoxin , promote phosphorylation of CRMP2 via glycogen synthase kinase-3 , leading to growth cone collapse [60] . Thus , the FMOs have the potential to inhibit growth cone protrusion through direct oxidation of F-actin resulting in depolymerization , and through redox regulation of the activity of UNC-33/CRMP ( i . e . to act both in the UNC-33 pathway and in parallel to it ) . In summary , we present evidence of a novel role of the C . elegans flavin-containing monooxygenase molecules ( FMOs ) in inhibition of growth cone protrusion downstream of UNC-6/Netrin signaling . The FMOs acted downstream of the UNC-6/Netrin receptors UNC-5 and UNC-40 , and downstream of the Rac GTPases CED-10 and MIG-2 ( Fig 14 ) . Future studies will determine if the FMOs regulate UNC-33/CRMP , if they cause actin depolymerization , or both , to inhibit growth cone protrusion .
|
Mechanisms that guide axons to their targets in the developing nervous system have been elucidated , but how these pathways affect behavior of the growth cone of the axon during outgrowth remains poorly understood . We previously showed that the guidance cue UNC-6/Netrin and its receptors UNC-40/DCC and UNC-5 inhibit lamellipodial and filopodial growth cone protrusion to mediate repulsion from UNC-6/Netrin in C . elegans . Here we report a new mechanism downstream of UNC-6/Netrin involving flavin monooxygenase redox enzymes ( FMOs ) . We show that FMOs are normally required for axon guidance and to inhibit growth cone protrusion . Furthermore , we show that they are required for the anti-protrusive effects of activated UNC-40 and UNC-5 receptors , and that they can partially compensate for loss of molecules in the pathway , indicating that they act downstream of UNC-6/Netrin signaling . Based on the function of the FMO-containing MICAL molecules in Drosophila and vertebrates , we speculate that the FMOs might directly oxidize actin , leading to filament disassembly and collapse , and/or lead to the phosphorylation of UNC-33/CRMP , which we show also genetically interacts with the FMOs downstream of UNC-6/Netrin . In conclusion , this is the first evidence that FMOs might act downstream of UNC-6/Netrin signaling in growth cone protrusion and axon repulsion .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"caenorhabditis",
"nervous",
"system",
"enzymes",
"enzymology",
"neuroscience",
"animals",
"motor",
"neurons",
"animal",
"models",
"caenorhabditis",
"elegans",
"model",
"organisms",
"experimental",
"organism",
"systems",
"nerve",
"fibers",
"research",
"and",
"analysis",
"methods",
"developmental",
"neuroscience",
"animal",
"cells",
"axons",
"proteins",
"guanosine",
"triphosphatase",
"biochemistry",
"signal",
"transduction",
"cellular",
"neuroscience",
"hydrolases",
"cell",
"biology",
"axon",
"guidance",
"anatomy",
"nerves",
"neurons",
"axon",
"guidance",
"receptors",
"nematoda",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms"
] |
2017
|
Flavin monooxygenases regulate Caenorhabditis elegans axon guidance and growth cone protrusion with UNC-6/Netrin signaling and Rac GTPases
|
This work introduces a coordinate-independent method to analyse movement variability of tasks performed with hand-held tools , such as a pen or a surgical scalpel . We extend the classical uncontrolled manifold ( UCM ) approach by exploiting the geometry of rigid body motions , used to describe tool configurations . In particular , we analyse variability during a static pointing task with a hand-held tool , where subjects are asked to keep the tool tip in steady contact with another object . In this case the tool is redundant with respect to the task , as subjects control position/orientation of the tool , i . e . 6 degrees-of-freedom ( dof ) , to maintain the tool tip position ( 3dof ) steady . To test the new method , subjects performed a pointing task with and without arm support . The additional dof introduced in the unsupported condition , injecting more variability into the system , represented a resource to minimise variability in the task space via coordinated motion . The results show that all of the seven subjects channeled more variability along directions not directly affecting the task ( UCM ) , consistent with previous literature but now shown in a coordinate-independent way . Variability in the unsupported condition was only slightly larger at the endpoint but much larger in the UCM .
Although highly stereotyped , human movements performed with the same intention are never exactly the same , displaying large variability in consecutive trials . Rather than just ‘biological noise’ , many studies have pointed out how variability may in fact provide important clues on the underlying neural strategies . Analysis of structure in variability , and its changes , has therefore become an important tool for researchers in neuromotor control and learning , especially in presence of redundancy [1] . Here , we are interested in the analysis of accuracy of pointing tasks performed with hand-held tools , for applications such as surgery . The problem of variability in redundant motor tasks was formulated by Bernstein [2] , who studied the kinematics of skilled movements performed by professional blacksmiths while striking a chisel with a hammer . Bernstein observed how the variability of the trajectory of the hammer , at its tip , was in fact smaller than the variability of the trajectory of each of the joints of the arm holding the hammer . This suggests that the individual joints are not controlled independently and that the brain exploits kinematic redundancy to accurately control the endpoint . Redundancy and motion variability are important not only for blacksmiths but characterize virtually every daily activity , from grasping a cup to signing off a letter , where we typically have many more degrees-of-freedom ( dof ) than necessary to fulfil the task . We are particularly interested in tasks involving hand-held tools such as microsurgery , where noise induced by tremor , amplified by the visual magnification provided by the optical microscope , is a critical factor of performance [3] . In this work , we consider static pointing tasks , such as the one in Fig . 1 , where a subject is asked to keep the tip of a pen-like tool , e . g . a surgical scalpel , in steady contact with another object . In other words , the position of the tip , characterized by dof of mobility , is prescribed while the subject is free to choose among different postures , which include positioning of the torso , joint angles of the arm ( i . e . shoulder , elbow and wrist ) as well as grasping pattern of the hand . For postures away from biomechanical limits , there exists a task-equivalent manifold consisting of distinct postures which do not affect the task . A major challenge with real-world scenarios is that , due to the large number of degrees of freedom involved in the task , biomechanical analysis would be either intractable or oversimplified . A key aspect of our study is that the tool itself has redundant degrees of freedom with respect to the task , i . e . subjects are asked to control position and orientation of the tool ( 6dof ) while maintaining a steady tool tip position ( 3dof ) . On the one hand , this allows one to focus on the low-dimensional space of tool configurations , rather than the high-dimensional space of possible postures . On the other hand , by focusing on the tool , we can make use of the geometric properties of rigid bodies , as detailed in the following sections . Scholz and Schoner [4] hypothesized that movement variance across task repetitions projects differently onto the task-equivalent manifold than it does onto the orthogonal complement directly affecting the task . A larger variance projected in the task-equivalent space ( or nullspace ) is indicative of neural control and the task-equivalent manifold ( where larger variance is expected for skilled movements ) was named uncontrolled manifold ( UCM ) . Despite its appeal , the computational procedures behind UCM ( and principal component analysis in general ) have been recently criticized for being coordinate-sensitive [5] . A fundamental issue with UCM is that the orthogonal space is typically defined via the standard Euclidean metric applied to the space of joint angles . Such an implicit choice is coordinate-dependent and the results of the UCM analysis would change if , for example , we decide to represent postures via joint angles instead of a normalized angle ( e . g . the joint angle divided by the biomechanical range of motion ) . More importantly , covariance-based analysis may reach different results if the coordinates are transformed . Even the linear transformation of joint angle coordinates from absolute to relative leads to different conclusions [5] . More than a century ago , Physics undertook a geometrization process of its main theories in the effort to achieve descriptions of phenomena in a coordinate-independent way , using differential geometry . Computational modelling in motor control is still at an early stage although some attempts have been made in this direction , e . g . [6] , [7] , [8] , [9] . It is not clear which coordinate system might be adopted by the brain to represent postures , therefore any specific selection rests on an arbitrary choice of the researcher . In situations where the configuration space is the set postures determined by several joint angles , it is not clear how to introduce an appropriate metric structure which relates angles relative to different joints , with very different ranges of motion . For example , starting from a reference posture , can we say that two new postures achieved , respectively , via a of ankle rotation and of knee rotation are equally ‘distant’ from the reference posture ? Although internal representations are largely unknown , it is clear that the brain does take into account the geometry and physics of the external world . Early studies on reaching tasks in the horizontal plane showed how we consistently move along straight lines in the extrinsic , end-point space [10] . In the last three decades , similar studies have been conducted under different conditions , including visual distortion [11] and force perturbations [12] , producing evidence that the brain is able to learn and adapt in order to produce straight lines in the visually perceived space ( so , sometimes slightly curved in the actual space ) . Computational studies showed how this large body of experimental observations is compatible with optimization of kinematic and/or dynamic costs which are related to the dynamics of the task , e . g . inertial and geometric properties of the human limbs [13] , [14] , [15] . These studies make use of methods traditionally applied in Mechanics and Robotics . Furthermore , a very recent study by Danziger and Mussa-Ivaldi [16] supports the hypothesis that movement trajectories are based on the perceived geometrical properties of the object ( such as the hand-held tool , in our paper ) that is being controlled by the brain . The novelty of their approach is in the use of a virtual object controlled via the ( hyper-redundant ) posture of the fingers , an experimental paradigm designed to eliminate any bias due to limb dynamics and experience in reaching . The findings of Danziger and Mussa-Ivaldi [16] , along with the work of Biess et al . [8] , [9] , suggest that metric properties such as distance measures and geodesics ( i . e . ‘straight lines’ ) of the operational space ( be it virtual or real ) play an important role in shaping our motor strategies . In this sense , Riemmanian geometry represents the appropriate theoretical framework for analytical investigation in motor control . In this paper we study accuracy during pointing via a coordinate-independent analysis of variance based on a choice of metric structure suggested by the specific application , in our case manipulation via hand-held tools . The paper outline is as follows . Next section will present all the steps involved in the classical UCM approach , based on vector calculus . Then an overview of the Riemannian geometric framework required to extend such vector calculus steps to more general settings is presented , along with the detailed formulation required to compute intrinsic variance based on metric properties of rigid bodies motions . This approach is then applied to the analysis of data of healthy subjects performing static pointing tasks . Results from the experiments are presented and discussed .
The starting point for classical uncontrolled manifold ( UCM ) analysis is the definition of a forward kinematic model ( 1 ) relating -dimensional human joint variables to a particular -dimensional variable which is hypothesized to be directly controlled by the brain . For example , to investigate how the center of mass ( COM ) is controlled by children during quiet stance , Wu et al . [17] derived a forward kinematic model mapping seven joint angles ( i . e . the angles formed between the foot , shank , thigh , trunk , head/neck , forearm , and upper arm segments with the horizontal ) onto the anterior-posterior position of the center , a one-dimensional variable . The task is kinematically redundant as a 7D joint configuration space is mapped onto a 1D task space ( anterior-posterior sway of the center of mass ) . The UCM analysis is simplified by linearizing the nonlinear forward kinematics about the average posture , hereafter reference posture , computed across measurements: ( 2 ) Linearization around the reference posture via the Jacobian , a matrix mapping joint-space velocities into velocities of the COM , is used to map ‘small variations’ in the joint space into ‘small variations’ in the controlled variable space: ( 3 ) As an approximation of the UCM , the nullspace ( ) of the Jacobian is used , i . e . a linear subspace of the configuration space for which deviations from the reference posture produce no motion in the task space . ( 4 ) At the same time a linear subspace ‘orthogonal’ to UCM ( ) is also computed as the orthogonal complement to the nullspace ( 5 ) Experimental deviations from the reference posture are projected onto the two orthogonal subspaces named , respectively , UCM ( ‘null’ ) and orthogonal ( ‘orth’ ) component . ( 6 ) Variance is computed for each component and normalized by the dimension of each subspace ( 7 ) where represents the Euclidean norm . The so called UCM ratio is then computed as the logarithm of the ratio of the variance of the UCM and orthogonal components . Riemannian geometry allows generalizing to nonlinear spaces traditional concepts and tools from vector calculus , e . g all the steps behind the classical UCM approach . In this section , we shall try to build some intuition to help relating these new geometric tools with the classical ones . For a more comprehensive and detailed description , the reader is referred to [8] , [9] and references therein . To motivate this need of generalization , one can think of cartography: our rigid rulers and goniometers work well on flat sheets of papers but not on a globe . Therefore cartographers draw charts by mapping points on the globe onto a sheet of paper . Clearly , patching the globe requires multiple charts , each with its own coordinate system ( ‘squared’ paper ) , and some rule to reconcile overlapping charts . With reference to Fig . 2 , the starting point is to define a -dimensional configuration manifold ( ) and to patch it with coordinate charts . Any chart of coordinates is sufficient to describe the behavior of the system locally , i . e . around a reference point ( or posture ) in the configuration manifold . Around such a reference point , we can approximate the manifold with its tangent space ( ) , an -dimensional linear ( vector ) space , tangent to the manifold at the reference position . The tangent space will be also tangent , at , to any trajectory on the manifold passing through itself . For this , elements of the tangent space are naturally identified with velocity vectors , at . The central element in Riemannian geometry is the introduction of a metric , i . e . a symmetric , positive definite bilinear function ( or quadratic form ) . As a generalization of the inner product of Euclidean spaces , a Riemannian metric acts on pairs of vectors ( , ) tangent to the manifold at a given point to determine , for example , the length of a vector ( ) or the angle between them ( ) . As mentioned previously , a metric does not come with a configuration manifold , it is extra structure which is typically defined by the application . Therefore the investigator always has to make a choice , especially when defining experimental conditions . Once the choice is made , the Riemannian geometric framework ensures that the results of the analysis will not depend on the choice of coordinates . The general Riemannian geometric approach starts from the definition of the configuration manifold . When it comes to rigid body motions , the configuration manifold is more structured than the general case . The space of rigid body configurations is in fact a Lie group , a manifold with additional algebraic structures as we shall see next . With reference to Fig . 1 , consider a space-fixed frame and a moving frame attached to the hand-held tool at the gripping point . Let represent the 3D coordinates of the gripping point in space coordinates . The orientation of with respect to is determined by the coordinate axes , the latter being aligned with the tool major axis , pointing away from the tip . At all times , the orientation of relative to can be represented via a rotation matrix whose first , second and third columns represent , respectively , the space-fixed coordinates of the axes . The focus of our analysis is on pointing tasks performed via hand-held tools . The key aspect is that the tool itself has redundant degrees of freedom with respect to the task . This allows disregarding the complexity of the body postures ( which especially for modeling the grasping finger pattern would either lead to oversimplified models or intractable ones ) and focusing on the intermediate space of tool configurations . In fact , we can think of the forward kinematics as a concatenation of two maps: the first transforming bodily postures to tool poses , the second transforming tool poses to controlled variables , i . e . the tool tip . To determine the latter map , we should consider that the gripping point of the hand tool is at a constant distance from the tip , therefore ( 24 ) Differentiating with respect to time leads to ( 25 ) where and are , respectively , the velocities at the tip and at the gripping point , in space coordinates . As we seek a formulation which is spatial frame invariant , we will try to express everything in body coordinates . Body velocities and are related to space velocities via the transformations and , which leads to ( 26 ) where is the skew-symmetric matrix for the body angular velocity . Equation ( 26 ) is clearly invariant to changes of space frame as it only comprises constants and body velocities which are left-invariant , as shown in ( 14 ) . A more compact formulation in terms of generalized velocity leads to ( 27 ) where is the Jacobian matrix ( 28 ) In the kinematic relation ( 27 ) , the Jacobian projects the 6 dof generalized velocity of the hand-held tool into the 3 dof velocity of the task . This captures the redundancy of the tool with respect to the prescribed task . The nullspace is the sub-space of generalized velocities of the tool which produce no motion at the end-tip , i . e . . It is straightforward verifying that the following generalized velocities ( 29 ) are an orthonormal basis of the nullspace of , i . e . produce zero end-tip velocity for ; are mutually orthogonal whenever ; and have unitary length for . Similarly , it can be verified that ( 30 ) is an orthonormal basis of the orthogonal complement of ( 31 ) Given a set of rigid body poses , where , an intrinsic definition of mean pose , see [24] , [22] and reference therein , iswhere is the distance between the poses and as in ( 23 ) . It can be shown that the intrinsic mean pose can be expressed as , where ( 32 ) For further details , the reader is referred to [24] , [22] . In the classical UCM approach [4] [17] , where postures are typically parameterized via a vector of joint angles , deviations from the reference position are directly computed as differences between vectors and projected onto the nullspace of the Jacobian and onto its orthogonal complement . For nonlinear spaces as for rigid body motions , this is not possible and we will extend the classical UCM approach with the concept of geodesics as proposed by Fletcher et al . [24] . Starting from a given point ( e . g . a reference position ) , geodesic curves are completely specified once the initial velocity is given and allow connecting sufficiently close points via minimal paths . Therefore , geodesics are a natural way to define the deviation of a point B from a point A as the initial velocity for a geodesic curve to start in A and reach B in a unit time . Recalling ( 22 ) , the ‘difference’ between a pose and the reference pose can be determined as the ( constant ) body velocity required to join the two poses via a geodesic in unit time: ( 33 ) By definition , the body velocity is a vector of the tangent space and can be projected onto the UCM and its orthogonal complement via the available inner product ( 20 ) : ( 34 ) As in the classical UCM approach , variance-per-dof can be computed as ( 35 ) where is the number of measurements and 3 is the dimension for both the nullspace and its orthogonal complement . The definition in eq . ( 35 ) corresponds to the geometric framework proposed by Fletcher et al . [24] where ( based on the early work of Frechet [25] ) the variance of a random variable in a metric space is defined as the expected value of the squared distance from the mean . This section first introduced the general steps involved in the classical UCM approach and then derived an intrinsic definition of each of these steps for the case of static pointing with hand-held tools . In particular , an intrinsic definition of deviations eq . ( 33 ) from an average pose eq . ( 32 ) was constructed by means of geodesics . In this way , the variance eq . ( 35 ) of these deviations on nullspace and on its orthogonal complement can be carried out independently of the choice of coordinates . To analyze variability during static pointing tasks with hand-held tools , experiments were conducted with 7 healthy subjects without any known history of neuromuscular impairment . All of them declared to be right-handed and gave their informed consent prior to the experiment . The study was approved by the institutional review board of Nanyang Technological University and was conducted according to the principles expressed in the Declaration of Helsinki . Each subject was asked to hold a sensorized stylus of a Polhemus Liberty system ( and resolution within range ) at a specific gripping point from the tip , onto which a hypodermic needle with luer connector ( Terumo ) was attached . The subject was then asked to touch the tip of a similar needle , firmly attached to a wooden table in a vertical position , with the tip of the stylus , as shown in Fig . 1 . The position of the tip and the orientation of the stylus , which are related to the position of the gripping point via eq . ( 24 ) , were acquired at 240 Hz via the Polhemus Liberty system and recorded onto a local PC for off-line data analysis . For both needles , only 1 mm of the tip is exposed while the remaining part is isolated with tapes . The setup is such that a beeping sound is produced when electrical contact between the exposed tips of the two needles occurs . The experimental protocol consisted of 20 consecutive trials . In each trial , the subject was asked to make a 15 seconds , steady contact between stylus and target tips separated each time by a large movement of the stylus approximately 20 cm away from the body . Only the inner most 10 seconds between two large movements were analyzed ( thick solid lines in Fig . 3 ) . Firstly , the furthest positions away from the target were detected ( corresponding to minima in Fig . 3 ) then a midpoint was calculated . For data analysis we considered only the data points within 5 seconds before or after the midpoint . One minute rest was given every 5 trials . No visual magnification was provided to the subject . The protocol was performed in two different experimental conditions: in Exp I the elbow of the right arm was supported on the table , while in Exp II the arm was unsupported , resulting in different noise conditions [26] . For every subject , both types of experiments were carried out in the same day , with Exp I preceding Exp II and one hour rest in between . The data relative to the three components of the tip position ( ) during one representative trial are shown in Fig . 3 . For each trial , only the inner most 10 seconds ( 2400 samples ) of steady contact were analyzed . To detect physiological tremor , the power spectral density ( PSD ) was estimated . To this end , the velocity components along each axis were estimated by numerically differentiating the tip position , component-wise . Then , for each ( 10 seconds ) trial , the pwelch ( ) function in the MATLAB environment was used to estimate the average PSD over ten non-overlapping time windows ( 1 second each ) . Finally , for each subject , the PSD estimates obtained for each trial were averaged . For each trial , the relative UCM components and were computed as in eq . ( 35 ) . The logarithm of their ratio , referred to as UCM ratio , was computed ( 36 ) The logarithm , instead of the ratio , was used in order to correct for non-normal distribution [27] , [28] . The UCM components for all trials of a representative subject are shown in Fig . 4 . Similarly to the UCM components and their ratio , also the variances of and and their ratio ( 37 ) were computed for each trial . According to the UCM theory [4] , a larger variance in the null space ( ) than in its orthogonal complement ( ) indicates that the position of the stylus tip is a variable directly under neural control . Therefore , we hypothesized that the UCM ratio ( ) will be significantly greater than zero . We also tested the influence of the experimental conditions ( Exp I and Exp II ) on the UCM ratio . Similar analysis was conducted for the variances of , and their ratio eq . ( 37 ) . To test whether the average UCM ratio eq . ( 36 ) is significantly different from zero , a Wilcoxon signed rank was run for the each subject on the values derived from every trial , separately for the two experimental conditions Exp I and Exp II . Similar tests were conducted for . A series of analysis of variance ( ANOVA ) tests with repeated measures was conducted to test the effect of experimental conditions and of UCM component on the variance-per-dof . The dependent variables are variance-per-dof , and . A three-way repeated measures ANOVA–2 ( experimental condition ) 2 ( UCM component ) 20 ( Trial ) was conducted on variance-per-dof . Two 2 ( Experimental condition ) 20 ( Trial ) repeated measure ANOVAs were conducted on and . All the ANOVAs used the MATLAB implementation RMAOV33 [29] and RMAOV2 [30] . A series of Kruskal-Wallis test were conducted to test the effect of experimental conditions on and for every individual subject .
For both experimental conditions , there was much more variability in the UCM subspace than in the orthogonal subspace ( i . e . , Fig . 6 ) as a Wilcoxon signed-ranks test showed a UCM ratio significantly different from zero ( ) . A three-way repeated measures ANOVA ( experimental condition UCM component trials ) conducted on the variance-per-dof indicated that all the main effects are significant: for the UCM component effect , for the trial effect ) and for the experimental condition effect . In addition , significant interactions were found between trial numbers and UCM component ( for all possible interactions ) . Effects due to experimental condition ( ) and trials ( ) were also found in a two-way repeated measures ANOVA on ( Experimental condition trial ) . No interaction effect was found between experimental condition and trial ( ) . As shown in Fig . 7 , there is a significant difference for the UCM ratio evaluated for the two experimental conditions in four subjects ( , , , and , Kruskal-Wallis test on subject 1 , 2 , 3 , and 7 ) . Similar analysis as for the UCM components and ratio , revealed , in general , more variability in than in , although this is not the case for all subjects ( Fig . 8 ) . Wilcoxon signed-rank tests on the showed that all the are significantly different from zero ( ) for all subjects and two experimental conditions , except subjects 1 , 3 , and 7 in Exp I ( , , and respectively ) . This indicates that for all subjects , the variance is larger at the gripping point of the hand tool than at the tip of the pen when the arm can move freely without support; while half of the subjects have similar variance at the gripping point and the tip when the arm is supported . As shown in Fig . 9 , a series of Kruskal-Wallis tests were applied to the between two experimental condition and for subjects 1 , 2 , 3 , and 7 there was a significant difference due to the experimental condition ( ) , for the remaining subjects . A two-way repeated measures ANOVA conducted on indicated that there were experimental condition effect ( ) and trial effect ( ) but no significant interaction ( ) .
Complexity of the human body typically leads to an excess of degrees of freedom for virtually every motor task we are routinely involved with . Redundancy is also adopted in the design of artificial systems , e . g . articulated robots , as extra dof can increase dexterity and robustness . However , redundancy also requires sophisticated control strategies , for example in devising control laws which guarantee repeatable postures [32] . In this lies one of most fascinating aspects of human motion: the apparent conflict between repeatability and variability of the movement itself . From an analytical perspective , repeatability and variability of movement have traditionally been distilled from experimental data via statistical approaches , by computing average and standard deviation estimates of movement properties derived from repeated trials . As pointed out by Newell and Slifkin [33] , the vast majority of motor control literature on normal human subjects has neglected movement variability , considering it as a reflection of ‘biological noise’ , while literature on motor disorders would interpret ‘low’ variability as a deficit , e . g . as in the case of stereotypies . Thresholds for standard deviations according to which the amount of variability should be considered large or small are often unreported , reflecting a bias relative to the theoretical views of the investigator . Furthermore , there is clearly more to movement variability than just standard deviation . In the last decade , various researchers have started exploring the structure of variability rather than just its amount . Structure in variability has been so far explored along two major avenues: its temporal or its geometric features . These two aspects are by no means exclusive and , in general , a combined temporal and geometric analysis is likely to provide more insight into human motor control . In this paper , we considered a pointing task and we focused on the geometric structure of variability , and also estimated the power spectral density to verify the frequency signature of physiological tremor . Consistent with previous literature on physiological tremor [31] , our spectral analysis confirmed , for all subjects , the presence of at least two peaks in the frequency ranges 0–7 Hz and 7–15 Hz . However , variability is not just tremor , in particular in goal directed tasks where voluntary control actions are expected to take place . Furthermore , despite being a static pointing task , temporal aspects are still present due to , for example , fatigue and learning effects . In our analysis of variance , a main effect of trials was always present . Given the simplicity of the task , we believe that fatigue rather than learning might have contributed to the trial effect . The geometric structure of variability was underlined by the pioneering work of Scholz and Schoner [4] who hypothesized that variability in redundant tasks is largely restricted to a subspace ( UnControlled Manifold , UCM ) of the configuration space which does not affect the task . Along this line , other related approaches have been proposed such as the Goal-Equivalent Manifold ( GEM ) [34] and Tolerance , Noise and Covariation ( TNC ) method [1] . While it is expected that the variability of the distal segments increases with the degrees of freedom , Morrison and Newell [31] showed that the contribution of more proximal upper limb segments to distal ( finger ) tremor is not simply additive , which is compatible with the hypothesis of neurally-driven compensatory strategies . In their study , Morrison and Newell [31] asked healthy subjects to minimize motion at the tip of the index fingers while standing with the arms parallel to the ground , with both index fingers fully extended and the remaining fingers fully flexed . The contribution of different joints to the distal tremor was analyzed by successively increasing the support of upper-limb segments from proximal to distal . Among other things , their study highlighted how the synergistic action of the wrist joint resulted in significantly smaller tremor at the index finger during a postural task . Similarly to Morrison and Newell [31] , we evaluated variability at a distal endpoint in two different conditions: supported ( Exp I ) and unsupported arm ( Exp II ) . In addition , our task was designed to be kinematically redundant as we wished to analyze changes in terms of ‘good’ and ‘bad’ variability , where ‘good’ refers to the variability which does not affect the task ( i . e . in the null space ) and ‘bad’ denotes the variability directly reflected in the task space ( i . e . in the orthogonal complement to the null space ) . As expected , in both experimental conditions , our results show that each subject shows a statistically significant difference ( ) between the UCM components , projecting more tremor along directions which do not affect the task . Furthermore , in the unsupported-arm condition ( Exp II ) , indeed more variability is introduced at the distal endpoint but that it is mostly channeled into ‘good’ variability ( Fig . 6 ) . This is also clear from Fig . 7 where the increase in both ‘good’ and ‘bad’ variability is analyzed when the experimental condition is changed from supported-arm ( Exp I ) to unsupported-arm ( Exp II ) . While , in terms of mean values , ‘bad’ variability increases but does not double for all subjects , ‘good’ variability increases significantly for most of the subjects ( for half of them , there is a five-fold increase ) . We performed a similar analysis comparing variability at the gripping point and at the tool-tip . Although leading to qualitatively similar results , UCM analysis leads to ‘crisper’ results in terms of statistical significance . This was expected since the goal of the task , clearly defined in pointing tasks , is fully captured by the UCM analysis . One of the most appealing aspects of the UCM method is the possibility to distinguish ‘good’ variance ( not affecting the task success ) from ‘bad’ variance ( affecting task performance ) and , therefore , to identify skillful performance . Subjects who are able to channel physiological tremor into movements which do not affect the task , might be deemed more skillful . Therefore one might be tempted to relate skills to the UCM ratio . However , our results suggest another possible explanation . From our experiments , there appear to be two groups of subjects: those who show a statistically significant improvement in terms of UCM ratio ( Fig . 7-top ) in relation to a change of experimental condition and those who do not . A possible interpretation is that , while all subjects perform well in the unsupported-arm condition , the former group ( formed by subjects ‘ma’ , ‘me’ and ‘ql’ ) performs equally well also in the supported-arm condition . The latter condition is characterized by a reduced number of redundant degrees of freedom . In this sense , when more dof are available , it might be easier to channel variability into motions which do not affect the task ( ‘good’ variability ) and , thus , skill might be related to the ability of performing equally well with a reduced number of dof . However , the mechanism behind this group difference is unknown and this requires further investigation . Despite its appeal , a weakness in the UCM analysis has been recently pointed out in relation to its coordinate dependence [5] . An issue with UCM is that the orthogonal space is usually defined via the standard Euclidean metric applied to the space of joint angles . This choice is coordinate-dependent , thus the results of the UCM analysis would change if , for example , we decide to represent postures via joint angles instead of a normalized angle ( e . g . the joint angle divided by the biomechanical range of motion ) . As also mentioned by Sternad et al . [5] , in computational motor control a distinction should be made between internal coordinates , which are assumed to be used by the brain to process information as well as plan and execute actions , and external coordinates used to describe and analyze behavior by the investigator . In the latter case , care must be exercised to ensure that the results are independent of the researcher's choice of coordinates . Differently from the UCM approach , Sternad et al . [5] proposed a method where variability is evaluated in the “space of the result” , a task-related space , making it less sensitive to coordinates in the configuration ( or execution ) space . This is done in recognition of the fact that for unambiguously defined tasks there should be a natural way to evaluate performance , possibly leading to a well-defined metric in the task-space . From this perspective , a main contribution of this paper is the use of task-specific features to construct an appropriate metric , which leads to a frame-invariant and objective analysis in the sense of [35] . In particular , manipulation via hand-held tools suggests the use of the scale-dependent left-invariant metric ( 18 ) , a particular type of kinetic energy metric especially suitable for kinematic rather than dynamic analysis , initially proposed by Park and Brockett [20] , [21] . Left-invariance guarantees independence of the inertial frame but the lack of a bi-invariant metric [21] implies dependence on the choice of body-fixed frame . Nevertheless , in the case of kinetic energy metrics for rigid body motions , left-invariance is in fact sufficient to guarantee the principle of objectivity [35] . From a mathematical perspective , it should be noted that one could have chosen a right-invariant metric to guarantee body-fixed frame indifference and forgo the left-invariance , i . e . bearing a dependence on the spatial frame . This would not be acceptable in our case , as the results would be dependent on the choice , for example , of the measuring system . Previous mathematical arguments are very general and do not take into account the specific requirements of the task . To describe the pose of a tool , the experimentalist needs to choose two coordinate frames: a fixed frame and a moving frame . Our left-invariant description does not depend on the fixed frame but , due to the nature of rigid body motions , necessarily depends on the moving frame , despite being an objective description in the sense of [35] . However , our experimental protocol explicitly requires the subject to grasp the tool at a prescribed position , i . e . at fixed distance from the tip as in Fig . 1 . Arguably , this induces a unique , natural choice for the position of the moving frame . It should be noted that our analysis only depends on the position of the moving frame and not on its orientation , as shown in supporting information Text S1 . Therefore , although mathematically there might be a general dependence on the choice of the body frame , a well defined task should always induce a natural choice of such a frame . This is consistent with the idea that unambiguously defined tasks should allow to measure performance in an unambiguous way , as also observed by Sternad et al . [5] .
|
Daily motor tasks typically involve more degrees-of-freedom than strictly required . For instance , pressing a button in the elevator only requires positioning the fingertip at a three-dimensional location in space . However , to move the arm we need to control many more degrees-of-freedom ( at least seven , only considering the shoulder , elbow and wrist ) than required by the task , each with its own variability due to physiological factors such as tremor . Variability at proximal joints ( e . g . shoulder or elbow ) is expected to be amplified and projected at the distal end ( fingertip ) . Remarkably , inter-joint coordination reduces the final variability at the fingertip position . Recent theories , such as the uncontrolled manifold ( UCM ) , distinguished between inter-joint variability that would not affect the finger position and variability that would affect the final task . A major issue is that traditional UCM methods rely on the coordinate system chosen to analyse the arm motion . Therefore , we introduce a coordinate independent UCM method for tasks performed with handheld tools , e . g . surgery . This paper describes a new method and demonstrates that it enables an accurate analysis of static pointing . The results clearly show that the subjects can channel variability in dimensions that do not affect the task outcome .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"motor",
"systems",
"biology",
"neuroscience"
] |
2013
|
Analysis of Accuracy in Pointing with Redundant Hand-held Tools: A Geometric Approach to the Uncontrolled Manifold Method
|
A key challenge in genetics is identifying the functional roles of genes in pathways . Numerous functional genomics techniques ( e . g . machine learning ) that predict protein function have been developed to address this question . These methods generally build from existing annotations of genes to pathways and thus are often unable to identify additional genes participating in processes that are not already well studied . Many of these processes are well studied in some organism , but not necessarily in an investigator's organism of interest . Sequence-based search methods ( e . g . BLAST ) have been used to transfer such annotation information between organisms . We demonstrate that functional genomics can complement traditional sequence similarity to improve the transfer of gene annotations between organisms . Our method transfers annotations only when functionally appropriate as determined by genomic data and can be used with any prediction algorithm to combine transferred gene function knowledge with organism-specific high-throughput data to enable accurate function prediction . We show that diverse state-of-art machine learning algorithms leveraging functional knowledge transfer ( FKT ) dramatically improve their accuracy in predicting gene-pathway membership , particularly for processes with little experimental knowledge in an organism . We also show that our method compares favorably to annotation transfer by sequence similarity . Next , we deploy FKT with state-of-the-art SVM classifier to predict novel genes to 11 , 000 biological processes across six diverse organisms and expand the coverage of accurate function predictions to processes that are often ignored because of a dearth of annotated genes in an organism . Finally , we perform in vivo experimental investigation in Danio rerio and confirm the regulatory role of our top predicted novel gene , wnt5b , in leftward cell migration during heart development . FKT is immediately applicable to many bioinformatics techniques and will help biologists systematically integrate prior knowledge from diverse systems to direct targeted experiments in their organism of study .
Defining the role of proteins in pathways is among the key challenges of human genomics . Many successful approaches have been developed for prediction of protein function and pathway membership [1]–[6] , however they rely on prior knowledge in the organism of interest to make new predictions ( i . e . at least some genes in the organism already annotated to the pathway ) [7]–[11] . These approaches rely on identifying characteristic behavioral patterns , in functional genomic datasets , phylogenetic profiles , or genomic feature studies of genes that are known to participate in a pathway , then use these patterns to predict additional pathway members [12]–[14] . For example , gene expression and protein interaction profiles can be used by machine learning methods to associate novel genes to pathways based on previously known pathway members [15] , [16] . The potential of such computational approaches to direct experiments has been demonstrated in studies investigating mitochondrial biogenesis [17] and seed pigmentation [18] . Other common exploratory methods , such as hierarchical clustering [19] , don't directly use known gene annotations to learn a prediction classifier , however they often use existing annotations to interpret the resulting cluster of genes ( e . g . gene enrichment analysis ) [20] . However in many organisms including human , pathways and processes where functional annotations of genes are most needed often have few or no prior experimentally confirmed annotations , making novel predictions of genes that may participate in such a process difficult or impossible . Thus , our study describes a method to robustly increase the set of prior gene annotations , which has the potential to improve all function prediction methods by increasing the accuracy of their predictions and enabling wider coverage of pathways and biological processes . Many of these processes are well studied in some model organism , but not necessarily in an investigator's organism of interest . Even when applying a conservative examination of only the closely related and heavily studied mammalian species human , mouse , and rat , processes represented in one species are often not well-characterized in another ( summarized in Figure 1 and a full list of processes available in Text S1 ) . For example , the process cellular glucose homeostasis , an increasingly important process due to the role of cellular metabolism in cancer development , has less than 5 gene annotations in human , yet has 31 in mouse , a commonly used model organism for cancer studies . These processes ( referred to in the text as understudied processes ) are not well studied in a particular organism of interest ( i . e . very few genes are annotated ) but might be well characterized in some other organism . A longstanding solution to improving the prediction accuracy of understudied processes has been to transfer functional annotations from organisms where the process is better characterized [21] . The critical challenge in accurately transferring functional knowledge between organisms is identifying the appropriate genes for the transfer: those genes that share not only sequence similarity , but also conserved pathway roles . Large-scale automated methods have so far exclusively used sequence homology to identify functionally conserved genes [22] , [23] . However , the relationship between sequence similarity and function is not trivial . For example , human angiopoietin-4 ( ANGPT4 ) , an important angiogenesis growth factor , has been shown to activate TEK ( tyrosine-protein kinase receptor ) , while the mouse sequence-ortholog ( Angpt4 ) has been shown to inhibit TEK [24] . In our previous work [25] , we developed a cross-organism gene functional similarity measure , which relied on the concept that functional genomics data can be used to resolve homologous relationships among closely related genes . The approach summarizes the compendium of genomics data in each organism into functional relationship networks to identify genes that do not simply share sequence similarity but also functional behavior in large collections of heterogeneous functional data , and are thus functionally analogous ( referred to in text as functional analogs ) . In this current study , we present a novel knowledge transfer method , Functional Knowledge Transfer ( also referred to in text as FKT and outlined in Figure 2 ) , which leverages the mapping of functional analogs to direct cross-organism annotation transfer for function prediction . FKT can be especially beneficial for existing and future machine learning methods studying biological processes with sparse annotations in any given organism of interest . By transferring experimental knowledge between genes that have been identified as functional analogs , our method extends beyond simple annotation transfer by sequence similarity . Experimental functional annotations are only transferred for genes that are not just similar in sequence , but also in their functional behavior derived from a large and relatively comprehensive compendium of genomic data . In this study , we show that FKT improves the prediction accuracy of machine learning algorithms , particularly for biological processes with few existing annotations in an organism of study . We compare FKT to annotation transfer by sequence similarity ( BLAST ) and demonstrate the superior performance of our method in improving gene function prediction performance . The consistent improvement and high performance across various state-of-the-art classification algorithms demonstrates our approach is robust to different learning models , which is crucial for wide applicability . We apply FKT to gene function ( i . e . biological process ) prediction in six metazoan organisms ( Homo sapiens , Mus musculus , Rattus novegicus , Drosophila melanogaster , Danio rerio and Caenorhabditis elegans ) and show that FKT is robust enough for the automated transfer of annotations among these diverse organisms and accurate function prediction . Finally , we demonstrate an application of FKT to discovering novel biology by coupling the knowledge transfer with a Support Vector Machine ( SVM ) to predict proteins involved in left-right asymmetry regulation during heart development in Danio rerio . We correctly predict several proteins in the pathway and experimentally confirm the first evidence of wnt5b's role in the process . A comprehensive application of FKT to 11 , 000 biological processes , along with the functional relationship networks for all six organisms , are available through the IMP web-server portal accessible at http://imp . princeton . edu [26] .
Most modern machine-learning methods that predict novel members of a biological pathway require a set of genes already known to participate in the pathway . These approaches are therefore limited to predicting genes to biological processes with sufficient prior knowledge in an organism [30] . For example , in the MouseFunc competition [7] ( a broad competition focused on the performance of biological process prediction approaches ) , terms with less than three gene annotations were considered infeasible to predict and not included . We address this constraint by leveraging knowledge across species , which allows us to take advantage of known biology from a model organism where the pathway of interest may be better studied . We applied our functional cross-annotation strategy ( FKT ) to biological processes with few known genes ( annotation sizes of < = 5 and < = 15 ) in six metazoans and evaluated the predictive performance of an SVM trained with these annotations . To evaluate our performance , we constructed a three-year temporal holdout of experimental annotations . We used only biological process annotations added to Gene Ontology [31] before 5/11/2008 ( all dates in mm/dd/yyyy format ) in learning the functional networks , transferring annotations across organisms , and predicting gene-process participation . New experimental annotations added to Gene Ontology between 5/11/2008 to 5/11/2011 were held-out and used for evaluation . In total , 3 , 207 GO biological process terms across the six organisms acquired new gene annotations in the subsequent three years . We evaluated the accuracy of our predictions with the gene-process assignments made during the hold-out time period in Figure 3 ( evaluation results of all GO terms in Text S3 ) . We observed substantial improvement using FKT when compared with only using the direct annotations for an organism . Improvement was evident across all six organisms , suggesting that even well characterized model organisms ( e . g . mouse ) can benefit from genomic-data-driven knowledge transfer . In addition , by holding out gene-process annotations acquired within the last three years , we could evaluate our ability to predict genes to processes which had no known genes in an organism prior to the hold out date ( i . e . before 5/11/2008 ) . Even though these processes were uncharacterized at that time , they subsequently became the focus of a directed experiment and thus were deemed biologically relevant and experimentally feasible in the organism . As shown in Figure 3 , FKT gene predictions to these processes performed competitively even compared to biological processes with known gene annotations . Furthermore , these results were robust to the evaluation timeframe ( 1-year temporal holdout shown in Figure S1 ) . We hypothesized that our transfer method could improve prediction performance for a wide range of machine learning methods . Machine learning algorithms are often based on distinct learning models and assumptions , thus any widely applicable annotation transfer method must be robust to not only the biological variability ( e . g . different organisms or pathways ) but also to this modeling variability . Thus in addition to SVM , we evaluated two widely used state-of-the-art learning methods: L1-regularized logistic regression [28] and Random forest [29] . We trained both classification methods with and without FKT and evaluated on the held-out set of annotations . FKT improved prediction accuracy across each machine-learning algorithm and organism ( Figure 4 ) . In particular , these improvements were consistent across biological process annotation sizes ( < = 5 and < = 15 ) . Altogether , these results indicated that FKT could recover biological processes that would be otherwise missed by most prediction methods , and that the transfer had wide applicability - improving performance across diverse organisms and machine learning algorithms . We coupled FKT with an SVM and applied the machine learning classifier to predicting novel gene functions in six organisms . These predictions included gene-process membership for 8 , 091 GO biological processes currently without experimental annotations in at least one organism . Supervised machine learning methods would be unable to predict novel genes to these biological processes without annotation transfer . They represent a wide range of biological pathways and processes ranging from development and metabolism to immune response and response to various stimuli ( a complete list of these GO terms is in Text S2 , categorization and specificity of these terms are shown in Figure S2 , S3 ) . For example , the biological process regulation of exit from mitosis ( GO:0007096 ) represents a crucial mitotic cell cycle process that enables cells to regulate their exit from M phase . This process had no experimental annotations in Danio rerio at the time of our study , however had been extensively studied in the model organisms Saccharomyces cerevisiae [32] , Mus musculus [33] and Drosophila melanogaster [34] . Our functional cross-annotation method has identified a total of 18 genes in Danio rerio with functional analogs annotated to this process ( 11 from yeast , 5 fly , 1 mouse and 1 rat ) , enabling novel predictions of gene membership to this process . Our top gene prediction for this process , cdh2 , has been experimentally confirmed in a recent study examining cell cycle progression in cdh2 mutant retina cells [35] . Interestingly , cdh2 is not only a novel prediction in Danio Rerio ( i . e . this gene function was unknown at the time of our study ) , but also no cdh2 homologs are known to be involved in the regulation of exit from mitosis in other organisms . Cdh2 is a member of the cadherin protein family , which are important transmembrane proteins that play a crucial role in cell adhesion in multi-cellular organisms . Methods that employ only sequence similarity would be unable to predict this because cdh2 homologs have not been annotated to this process in other model organisms . Furthermore , prediction methods without FKT will miss this finding because there are no existing Danio rerio annotations to this process . Only methods coupling FKT with a machine learning algorithm can take advantage of information from the single cell model organism Saccharomyces cerevisiae , where cell-cycle checkpoints have been extensively studied [36] , and successfully predict this finding in the multicellular model organism Danio rerio . This in vivo experimental result demonstrates FKT's utility for predicting novel genes to understudied processes . In addition , by coupling functional transfer to machine learning methods that leverage organism-specific functional data collections , we can make reliable gene-process predictions even without an annotated sequence-homolog . To compare our functional transfer method , which applied a more specific annotation transfer , to commonly used methods that used only sequence homology , we evaluated a method that did not leverage functional similarity and a baseline method without any cross-annotation . In this sequence-only method , all homologous gene pairs ( reciprocal BLAST best hit gene pairs ) were targets for annotation transfer - any biological process annotated to a gene was transferred to its reciprocal best-hit gene in all organisms . To obtain a representative set of gene-process annotations for evaluation , we conducted a threefold cross-validation on genes that had experimental biological process annotations , and evaluated the SVM classifier prediction performance on each corresponding held-out set of biological process annotations . The results of the comparison showed that although both methods improved performance for small processes , FKT showed greater performance gains ( Figure 5 , evaluation results of all GO terms in Text S3 ) . In a few organisms , the performance gains were substantial - for example , in human and mouse , the median performance ( precision at 10% recall ) increased more than fivefold . Upon examining the processes that improved the most when compared to a sequence-only method , many pathways and processes with transcriptional based regulatory control showed improved performance using FKT . Response to mechanical stimulus , ameboidal cell migration , regulation of neuron differentiation and striated muscle cell development were among the top improved processes in all organisms using FKT compared to sequence-only . Unsurprisingly , these processes have been well known to be tightly regulated through transcriptional programs ( e . g . stress response , developmental TF gradients ) [37]–[39] and have multiple datasets measuring the transcriptional profiles incorporated in our functional networks [40]–[42] . We expect that FKT will continue to improve as the functional genomics compendia for many organisms continue to grow , including expression and other types of measurements across multiple perturbations . An additional advantage of a functional genomics similarity approach , as shown in [25] , is the ability to differentiate functional differences in tissue specificity between sequence homologs . The example of mouse RNA polymerase II elongation factor Supt5h and its direct sequence-ortholog C . elegans spt-5 highlight this issue . FKT determined these sequence-orthologs as not being functional analogs . Indeed , mouse Supt5h is predominantly neuronal , while C . elegans SPT-5 is non-neuronal and primarily expressed in the intestine and pharynx [43]–[45] . Even though these sequence-orthologs have diverged in tissue specificity , they still share high sequence similarity and a sequence-only method would inappropriately transfer all functional annotations between them . In all vertebrates , the heart develops with a distinct left-right ( L-R ) asymmetry during embryonic morphogenesis . Deviations in left-right heart development can lead to complex congenital heart defects that are among the most common human neonatal diseases [46] , [47] . During cardiac morphogenesis in Danio rerio , two distinct stages of cell migrations lead to the final asymmetries of the heart . In the first stage , called “heart jogging” , myocardial cell migration within the cardiac cone place the ventricular cells to the left side , while atrial cells remain near the midline . In the second stage of “heart looping” , the heart tube bends and forms a loop that places the ventricle to the right of the atrium . Although the steps of cell migration progression leading to left-right heart asymmetry are beginning to be explored [48]–[51] , an understanding of how it is achieved mechanistically is still lacking . In Gene Ontology , the biological process term “determination of heart left right asymmetry” ( GO:0061371 ) represents the developmental pathways regulating heart jogging and looping . To validate our prediction method ( FKT coupled with SVM ) , we investigated the top five predicted genes that had not already been annotated to this GO term: sox32 , wnt5b , ndr1 , tbx1 and lft1 . We found existing literature evidence confirming the involvement of four of the five genes ( sox32 [52]–[55] , ndr1 [56] , tbx1 [57] , and lft1 [58]–[60] ) . Although there existed experimental results confirming the role of these genes in influencing heart asymmetry , these results had not yet been curated by GO annotators . For example , in a knock-out experiment of our top predicted gene ( sox32/casanova ) , Danio rerio embryos had fewer dorsal forerunner cells which led to defects in Kupffer's vesicle formation and subsequent left-right patterning of the heart , confirming that sox32 was required for proper establishment of heart asymmetry . The only gene among the top five without existing experimental support was wnt5b , our second ranked prediction after sox32 . Previous work had shown the involvement of wnt5b in cell migration during gastrulation [61] but the gene had not been specifically associated with heart left-right asymmetry regulation . To experimentally validate our prediction of wnt5b to left-right heart determination , we knocked down its function by means of morpholino antisense oligonucleotides ( MO ) [62] . A significantly greater proportion of embryos where wnt5b was knocked down with a morpholino ( Figure 6 ) had a defective heart jogging phenotype ( Fisher's exact test p-value<0 . 001 ) . In total , 48% of morpholino treated embryos showed either right-sided heart jog or midline jog comparable to previous genes known to be involved in this biological process [63]–[65] . Only 4% of wild type and control-MO treated embryos exhibited this phenotype . This phenotype is likely due to the disruption of asymmetric expression of the TGFbeta member Nodal ( data not shown ) , which is typically asymmetrically expressed on the left side of vertebrate embryos during somitogenesis . Left-sided Nodal in Danio rerio myocardial cells directs their subsequent migration during asymmetric cardiac morphogenesis [48] , [51] . Further investigation would be necessary to understand the mechanistic role of wnt5b in left-right heart determination , however our in vivo experiment confirmed the regulatory role of wnt5b in Danio rerio left-right asymmetry determination in heart development , as our method predicted .
This study demonstrates that state-of-the-art machine learning methods coupled with our functional knowledge transfer method can accurately prioritize novel genes of understudied processes . Previous methods have focused on incorporating functional genomic data primarily as input data [66]–[69] . In contrast , here we demonstrate that the prevalence of understudied processes and the abundance of genomic data provide an opportunity to improve the accuracy of cross-organism annotation transfer and extend prediction coverage to processes with no prior annotations . We now integrate FKT into our IMP web-server [26] . This makes IMP a web interface for exploratory analysis covering all organisms included in this study across 10 , 653 biological processes ( http://imp . princeton . edu ) . Functional knowledge transfer allows IMP to also include gene predictions for processes currently unannotated in an organism . Although in our current study we have experimentally followed up on our top predicted gene , all of our predictions in IMP are shown with estimated probabilities allowing biologist to draw a threshold dependent on how much the assay costs , and how important it is to find true positives ( versus not finding false positives ) . In addition , the website includes the Bayesian functional relationship networks that were used for mapping functional analogs and used as input features to the machine learning methods . In particular , to the best of our knowledge , we include the first zebrafish ( Danio rerio ) functional relationship network . We anticipate that our approach can be extended beyond the six organisms shown in this study , as it is especially beneficial in organisms that have high-throughput genomic data with sparse annotations ( e . g . frog , slime mold ) . Next-generation sequencing is further increasing the diversity of organisms that are measured on the genome-scale , and functional knowledge transfer can help us understand and annotate the roles of genes in such emerging model systems . Functional knowledge transfer allows for accurate hypothesis generation and experiment guidance even for pathways with no previous experimental knowledge in a given organism , thus benefiting human biology , broadly studied organisms such as mouse and fly , and newly adopted model systems .
In total , 10 , 653 GO biological process terms were predicted for new gene annotations covering six organisms , Homo sapiens , Mus musculus , Rattus novegicus , Drosophila melanogaster , Danio rerio and Caenorhabditis elegans . We limited the positive examples for each GO term to propagated experimental GO annotations with GO evidence codes EXP , IDA , IPI , IMP , IGI and IEP ( all “NOT” annotations were removed ) . In addition , to leverage the research strengths across organisms , we transferred gene annotations among six organisms plus yeast , first based on sequence similarity and second filtered by function similarity . In detail , we start with all sequence paralog and ortholog gene relations within each TreeFam [22] gene family . Next , based on our previous algorithm [25] , we filtered for functional analogs among all paralog and ortholog gene pairs using our functional relationship networks . We define a functional analog to be a gene pair that has a significant number of overlapping TreeFam gene families among its closest gene neighbors in the global functional relationship network ( a functional network is converted into a binary network by using a probability cutoff of 0 . 5 ) . We defined a gene pair's score as the following:where m and n are the number of TreeFam gene families in each gene G1 and G2's direct neighborhood in the functional network , k is the number of overlapping TreeFam gene families between gene G1 and G2 gene neighbors and N is the total number of TreeFam gene families around gene G1 and G2 . The functional similarity score is the probability of observing greater or equal to the number of overlapping gene families by chance , thus can be interpreted as a hypergeometric p-value . We used a score cutoff of < = 0 . 01 to consider a gene pair as functional analogs . Finally , all experimental annotations are propagated between functional analogs . In total , our supervised functional knowledge transfer allowed us to make predictions for 8 , 091 additional GO biological processes , thus extending our predictions beyond simply well-studied and well annotated processes and pathways . We used the augmented gold standard genes by functional knowledge transfer and functional relation network as features into state-of-art machine learning algorithm Support Vector Machine ( SVM ) to predict novel biological process gene annotations . Our functional relation network based SVM method has shown to outperform methods that directly input the raw data into the SVM or a simplistic sum of the functional networks to the positive examples [89] . For each biological process , the feature space was constructed as the weights in the functional relation network . Thus for each gene example , all gene edge weights connecting to the example gene were used to create the feature vector . Therefore , each organism feature count will be equal to the number of genes in the organism . The set of feature vectors for training examples were used to train a linear SVM according to the standard formulation:where n is the of training example genes , w is the gene weight vector , yi is the training label of gene i and xi is the edge weight vector connecting gene i to all genes in the functional network . Finally , the unbounded SVM prediction scores were transformed into probabilities based on a maximum likelihood sigmoid fit to the SVM outputs [90] . To validate that the observed performance improvement was not specific to any single learning algorithm , we applied the functional knowledge transfer to two additional widely used machine learning methods: L1-regularized logistic regression and Random forest . Regression analysis coupled with regularization has been a broadly used approach to control for the bias-variance trade-off [91] . In particular , L1-regularization has been successfully used in many methods for shrinkage and feature selection applications , most famously in the works of LASSO [92] . By coupling L1-regularization with a logit link function , conditional probabilities of a gene membership to a biological process can be computed based on selected genes of predictive power . The predictive gene weight vector w was obtained by the following regression problem:where λ>0 is the regularization parameter , yi is the training label of gene i and xi is the edge weight vector connecting gene i to all genes in the functional network . Random forest classifiers are a combination of decision trees that are aggregated to make a final prediction . Random forest algorithms have been shown to produce improved prediction accuracy compared to a single decision tree by better estimating the contribution of each predictor through random sampling [29] . In genomic applications , Random forest has gained interest due to the many high-dimensional genomic learning problems [93] . Formally , random forest is defined by the following:where the random forest RF is a set of decision tree functions , trained on training examples X and a bootstrap sample di from the original feature space of D . For classification , the votes of each n decision trees are averaged as shown in the following:where is the indicator function for the prediction class of interest . In our study , for each GO term 61 decision trees were trained on independent bootstrap samples of our original genomic training data . For performance evaluations for GO terms with no prior annotation , we used a three-year temporal holdout set of gene annotations for each GO biological process ( one-year holdout results shown in supplemental material ) . The held-out gene annotations were preserved throughout the prediction pipeline ( functional network integration and SVM predictions ) to avoid any overestimation of performance . Although we train our SVM classifiers using the augmented cross-annotated gold standard , only the non-transferred experimental GO term annotations were used for evaluation with all transferred annotations excluded . The GO gene association files used to create our gold standard was downloaded from Gene Ontology [31] on 5/11/2011 ( all dates in mm/dd/yyyy format ) . To generate an accurate temporal three-year holdout we downloaded the GO gene association version archived at 5/11/2008 . All annotations were propagated and only experimental examples newly annotated after 5/11/2008 for each GO term was used in the temporal evaluation . Accordingly , any GO term that had no gene annotations on 5/11/2008 , but subsequently accumulated new annotations were used to evaluate our performance in predicting terms with no-known prior annotations . To compare performance between knowledge transfer methods , we conducted an evaluation by performing a threefold cross-validation among genes that had experimental biological process annotations . This set of evaluation annotations represents a random sampling of the current knowledge base as of 5/11/2011 . Identical to our temporal holdout , all evaluation annotations for each holdout were withheld from our prediction pipeline to avoid any performance over-estimation . All software used in this study has been implemented in the open source and publicly available Sleipnir library [94] available from http://libsleipnir . bitbucket . org , which interfaces with the SVMperf library [95] for linear kernel SVM classifiers ( the error parameter C was set to 100 for these experiments through cross-validation ) . L1-regularized logistic regression used the LIBLINEAR [28] and Random forest used the MILK ( Machine Learning Toolkit ) python package implementation with 61 decision trees per GO term . The wnt5b morpholino ( MO ) and standard control MO were purchased from GeneTools . The sequence of the wnt5b MO used is as follows: 5′-GTCCTTGGTTCATTCTCACATCCAT-3′ . Morpholinos were injected at a concentration of 6 ng/uL into the yolk of one-cell stage embryos for whole knockdown in the embryonic cells . Initial assessment ( Figure 6 ) was performed via in situ hybridizations on fixed embryos using the standard protocol [96] with cmlc2/myl7 used as a probe . Images were captured at 4× or 10× magnification using a ProgressC14 digital camera ( Jenoptik ) on a Leica MZFLIII microscope . Heart laterality for each treatment ( wnt5b MO , control MO , wild type ) was evaluated in live Tg ( cmlc2::GFP ) embryos at 27 hours post fertilization . Embryos were scored as left/right/no jog based on expression of GFP driven by cmlc2's heart specific promoter using a Leica SP5 confocal microscope ( Figure S4 ) .
|
Due to technical and ethical challenges many human diseases or biological processes are studied in model organisms . Discoveries in these organisms are then transferred back to human or other model organisms . Traditional methods for transferring novel gene function annotations have relied on finding genes with high sequence similarity believed to share evolutionary ancestry . However , sequence similarity does not guarantee a shared functional role in molecular pathways . In this study , we show that functional genomics can complement traditional sequence similarity measures to improve the transfer of gene annotations between organisms . We coupled our knowledge transfer method with current state-of-the-art machine learning algorithms and predicted gene function for 11 , 000 biological processes across six organisms . We experimentally validated our prediction of wnt5b's involvement in the determination of left-right heart asymmetry in zebrafish . Our results show that functional knowledge transfer can improve the coverage and accuracy of machine learning methods used for gene function prediction in a diverse set of organisms . Such an approach can be applied to additional organisms , and will be especially beneficial in organisms that have high-throughput genomic data with sparse annotations .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genomics",
"functional",
"genomics",
"biology",
"computational",
"biology",
"genetics",
"and",
"genomics"
] |
2013
|
Functional Knowledge Transfer for High-accuracy Prediction of Under-studied Biological Processes
|
In order to extract retinal disparity from a visual scene , the brain must match corresponding points in the left and right retinae . This computationally demanding task is known as the stereo correspondence problem . The initial stage of the solution to the correspondence problem is generally thought to consist of a correlation-based computation . However , recent work by Doi et al suggests that human observers can see depth in a class of stimuli where the mean binocular correlation is 0 ( half-matched random dot stereograms ) . Half-matched random dot stereograms are made up of an equal number of correlated and anticorrelated dots , and the binocular energy model—a well-known model of V1 binocular complex cells—fails to signal disparity here . This has led to the proposition that a second , match-based computation must be extracting disparity in these stimuli . Here we show that a straightforward modification to the binocular energy model—adding a point output nonlinearity—is by itself sufficient to produce cells that are disparity-tuned to half-matched random dot stereograms . We then show that a simple decision model using this single mechanism can reproduce psychometric functions generated by human observers , including reduced performance to large disparities and rapidly updating dot patterns . The model makes predictions about how performance should change with dot size in half-matched stereograms and temporal alternation in correlation , which we test in human observers . We conclude that a single correlation-based computation , based directly on already-known properties of V1 neurons , can account for the literature on mixed correlation random dot stereograms .
Stereoscopic vision is possible because the left and right eyes receive slightly different images of the world . This geometric arrangement gives rise to retinal disparity which can be used to extract depth information from a visual scene . A key challenge in stereo vision is determining which elements in the left eye’s image correspond to those in the right eye’s image . This computationally demanding task is known as the stereo correspondence problem and has been extensively studied [1–5] . The initial step in the solution of the correspondence problem is generally thought to be something close to a cross-correlation of the image patches in the left and right eye [6–8] . This view arose from work by Ohzawa , DeAngelis and Freeman ( 1990 ) [9] who proposed the very successful binocular energy model ( BEM ) for disparity selective V1 complex cells . The BEM provides a physiological instantiation of a correlation-type computation which accounts for the ability of these neurons to signal disparity in random dot stereograms ( RDSs ) . Ohzawa et al [9] , recording from V1 in the cat , used line stereograms and showed that inverting the contrast in one eye ( i . e . making the line stereogram anticorrelated ) also inverts the profile of the disparity tuning curve . Similarly , Cumming and Parker [10] inverted the binocular correlation of random dot stereograms and demonstrated that the profile of the disparity tuning curves of macaque V1 neurons also invert . The original binocular energy model is an example of a pure correlation computation as it predicts that the amplitude of disparity-related modulation should be equal for correlated and anticorrelated patterns . In other words , the amplitude ratio between the correlated and anticorrelated responses should be equal to 1 . This is illustrated in Fig 1a , where the amplitude ratio is given as the ratio between the length of the black and red lines . However , many V1 neurons have amplitude ratios less than 1; the anticorrelated response is attenuated relative to the correlated response [10] . One explanation could be that an initial correlation computation is followed by a nonlinearity which results in responses that are not a linear function of correlation [11–13] . In a recent series of papers , Doi et al [14–16] have proposed that two distinct computations contribute to depth perception in cyclopean stimuli . They postulate a pure correlation mechanism , which depends linearly on interocular correlation , plus an additional “matching” computation , which in their most recent work [16] they have suggested may simply be the correlation mechanism plus an additional output nonlinearity . Under some circumstances , this is identical to a single mechanism with a nonlinear response to correlation ( see Fig 1b and 1c ) . However , Doi et al [14 , 15] propose that the two mechanisms have distinct spatiotemporal integration properties , and so may be differentially activated by different stimuli . These conclusions are motivated by a series of ingenious psychophysical experiments in which the authors mixed correlated and anticorrelated dots within a single random dot stereogram ( RDS ) . When half the dots in an RDS are correlated and half are anticorrelated ( half-matched RDSs; Fig 2a ) , the global binocular correlation of the stimulus is 0 . In this case , the authors argue , a pure correlation computation should not be able to detect depth . However , humans can perceive depth in such stimuli [14 , 15] . Doi et al [14 , 15] argue that this cannot be explained by a pure correlation mechanism , and propose an additional matching mechanism to account for these data . In addition to depth perception in half-matched RDSs , two pieces of evidence suggest that two separate mechanisms extract disparity in random dot stereograms . Doi et al [14] have reported that larger disparities tend to lead to decreased performance for half-matched RDSs and more reversed depth responses to anticorrelated RDSs . In a subsequent publication , Doi et al [15] reported a similar phenomenon in the temporal domain . They investigated dynamic random-dot patterns , in which the dot pattern is periodically replaced with a new random pattern with the same disparity and correlation . They showed that faster dot pattern refresh rates lead to poorer half-matched judgments and more reversed depth responses to anticorrelation . The authors argue that these results again reflect the weighted contribution of two separate mechanisms: one slow matching computation , responsible for fine disparity discrimination , and one rapid correlation computation , responsible for coarse disparity discrimination . Fig 2b illustrates schematically the performance they expect from these two computations in isolation . While it is true that half-matched RDSs—stereograms with equal numbers of correlated and anticorrelated dots—have a mean binocular correlation of 0 , it is possible that local fluctuations in correlation could be exploited to determine the stimulus disparity [15 , 16] . Doi et al [14 , 15] propose that these fluctuations are used by the matching computation , possibly in extrastriate cortex . However , the attenuated responses of V1 neurons to anticorrelated dots [10] makes it possible that even V1 neurons could encode disparity in these stimuli . V1 neurons respond more strongly to positive than negative binocular correlation [10] . If this attenuated response was generated by a simple output nonlinearity , then their responses to stimuli with high correlation variability may be greater than predicted from the mean correlation alone . In other words , the mean response of the cell may depend on the local correlation variability as well as the mean correlation . Doi & Fujita [16] explore this with a modified version of a cross-correlation computation which they refer to as “cross-matching” . Cross-matching computes the correlation between left and right images , but then follows this by half-wave rectification . Doi & Fujita [16] conclude that cross-matching has the necessary properties to serve as the computation underlying their putative match-based computations . Importantly , they still postulate that human stereo vision is subserved by separate pure correlation and match-based computations , with different spatiotemporal properties , and whose contribution to perception varies with properties of the stimulus . The proposition that these two mechanisms have different spatiotemporal properties is crucial if the two-mechanism model is to be distinguished from a single mechanism intermediate between the pure correlation and pure matching models . While differences in spatiotemporal properties are essential in order to separate “pure correlation” and “cross-matching” , changes in psychophysical performance with changes in spatiotemporal properties of the stimulus do not necessarily imply that there must be two mechanisms . In principle , changes in performance could be due to changes in the stimulus , even with a single-mechanism model . However , no one has yet explored whether a single-mechanism model can account for the results of Doi et al [14 , 15] . Here we explore the possibility that a single computation can explain depth perception in correlated , anticorrelated , and half-matched random-dot stereograms . Like Doi & Fujita [16] , we use a model that can describe the attenuation observed in V1 neurons—a binocular energy model with an additional output nonlinearity—and show that this can explain responses to half-matched stereograms . Our scheme differs from theirs in that it does not suppose two distinct computations operating in cortex , but rather uses a single mechanism to explain all the psychophysical data . To explore this model , we first investigate the model responses to half-matched RDSs . We then show that this model can account for a range of previously documented psychophysical phenomena , including effects which Doi et al [14–16] have suggested are diagnostic of either a pure correlation or a match-based computation . Finally , we confirm two new predictions made by the model in human observers: that psychophysical performance should become worse with 1 ) decreased dot size in half-matched stereograms and 2 ) in response to rapid temporal modulation in correlation .
We created binocular energy model ( BEM ) units by combining model binocular simple cells whose monocular receptive fields were in quadrature phase . This gives a model complex cell response C which is invariant to stimulus phase . As discussed above , the response of a BEM unit like this is a linear function of binocular correlation . We made the model nonlinear with respect to correlation by adding a static squaring output nonlinearity , giving a final response C2 ( see Materials and Methods ) . The same model was also explored by Read , Parker , and Cumming ( 2002 ) [13] . We computed the response of both models to correlated , half-matched and anticorrelated random dot stereograms ( RDSs ) of various disparities . Disparity tuning curves for the BEM with and without an output nonlinearity are shown in Fig 3 . For the binocularly linear BEM ( Fig 3a ) , the model’s responses to correlated and anticorrelated stimuli have a characteristic symmetry about a horizontal line ( the response to zero binocular correlation ) , i . e . the amplitude ratio between the correlated and anticorrelated responses is 1 . For half-matched stimuli , the model has no disparity selectivity . When the model has a static output nonlinearity ( Fig 3b ) , the correlated and anticorrelated tuning curves become asymmetric . This asymmetry leads to a modest , but very clear disparity tuning for half-matched RDSs with a peak at the neuron’s preferred disparity . In the linear BEM ( Fig 3a ) , the half-matched nature of the stimulus manifests itself as a variable firing rate , but not as a mean change [15] . The disparity tuning in Fig 3b arises because the expected value of a squared random variable depends on its variance: E[X2] = ( E[X] ) 2 + Var[X] ( other choices for the exponent , as well as other nonlinearities , such as thresholding , will also yield a dependence on variance ) . For the model with an output nonlinearity , therefore , the high correlation variability in half-matched RDSs gets converted into a mean change in the firing rate . Because the tuning in Fig 3b is the consequence of fluctuations in local correlation over the RF , stimulus parameters that affect the variability of these local measures also affect the disparity tuning . Increasing the dot density decreases the correlation variability as there are more dots within a neuron’s spatial receptive field . Following earlier studies [14 , 15] , dot density is defined as the proportion of the stimulus area that would be covered by dots if the dots were not allowed to occlude ( although dots were allowed to occlude ) , hence the units are in proportion coverage . Having fewer , larger dots ( while maintaining constant density ) generally increases the correlation variability as a single dot fills a larger fraction of the RF with pixels sharing the same correlation . With more , smaller dots , the cell is integrating across a greater number of samples ( since the cell is likely to see more independent dots within its receptive field ) and so the variability is reduced . Because increasing the dot size while holding density constant is the equivalent of reducing the receptive field size in the model , we use the relative receptive field size , defined as σ/r , where σ is the standard deviation of the monocular Gabors , and r is the dot radius . The effect of density and relative receptive field size on disparity tuning in our model can be seen in Fig 4a , where the strength of disparity tuning for half-matched stimuli is plotted as a proportion of the modulation produced by correlated patterns with the same spatial parameters . We define this normalized mean response from the responses to the preferred disparity: Rnorm=〈Chm2〉−〈Cuncorr2〉〈Ccorr2〉−〈Cuncorr2〉 . ( 1 ) where 〈Cuncorr2〉 , 〈Chm2〉 , and 〈Ccorr2〉 recepectively the mean response to uncorrelated , half-matched and correlated stimuli at the neuron's preferred disparity . As the relative receptive field size decreases ( i . e . smaller receptive field relative to the dot size ) , modified BEM cells signal disparity more vigorously to half-matched stimuli ( relative to equivalent correlated stimuli ) . It is worth noting that while the amplitude ratio is large for very low densities , this does not necessarily translate to better psychophysical performance on the task . This is because at low densities , the responses to both correlated and anticorrelated patterns are much weaker , so the signal-to-noise ratio is lower . ( The variations caused by the monocular image content dominate the model’s response ) . The effect of this is that there may be poorer performance at very low densities than at higher densities , despite the amplitude ratio being higher at low densities . Consider the case where the density is so low that on some trials no dots are presented in the RF . For these trials performance based solely on this neuron will be at chance , despite the high normalized response ratio averaged across many trials . Even performance based on many neurons would presumably suffer from a lower signal to noise ratio . We will explore the effects of signal and noise in simulations of a psychophysical task below . Just as increasing the spatial extent of the receptive field ( relative to dot size ) reduces variability in binocular correlation seen by the model , so increasing the temporal extent of the RF ( relative to the pattern update frequency ) reduces variability . We explore this using a fixed RF and a changing dot pattern refresh rate . As refresh rates increase , the model cell is integrating more dots within its temporal window , reducing fluctuations in the binocular correlation . Modified BEM cell responses are shown in Fig 4b for different receptive field sizes and pattern refresh rates . For low refresh rates and small RF sizes , the cell exhibits substantial disparity tuning to half-matched stimuli . As in Fig 4a , when the RF size increases relative to the dot size , the disparity tuning to half-matched stimuli decreases since the cell is integrating across more dots . Similarly , as the dot pattern refresh rate increases , the half-matched disparity tuning decreases since the cell is again integrating across more dots ( but now across time rather than across space ) . These simulations demonstrate two key properties: the model exhibits less disparity tuning to half-matched stimuli as 1 ) the receptive field increases in size relative to the dot size , and 2 ) the refresh rate increases relative to the temporal integration period of the neuron . The first finding is noteworthy because it was observed in Doi et al [14] that human observers are better at reporting depth in fine disparity half-matched RDSs than in coarse disparity ones . Given the effects of RF size we show , this observation might be accounted for by the well-known size-disparity correlation [7 , 17–20] , since coarse disparity detectors tend to have larger receptive fields than fine disparity detectors . The second finding is noteworthy because Doi et al [15] reported that performance to half-matched stimuli also decreased with increasing pattern refresh rate . The authors interpreted this as a shift from a matching computation at low refresh rates to a pure correlation computation at high refresh rates . In the current framework , the decreased disparity tuning with refresh rate reflects temporal integration within a single correlation-based computation , rather than differential activation of two distinct computations with different spatiotemporal properties . Although the match-based computation hypothesized by Doi et al [15 , 16] is similar to our modified correlation-based computation , in our framework , only a single computation is involved . Indeed , our model neurons could be described as the sum of a matching computation and pure correlation computation , just as illustrated in Fig 1c , but this is achieved by a single mechanism . Unless the two components differ in some other way ( e . g . temporal response ) , the two descriptions are identical . Clearly , the fact that our version of the BEM can signal disparity in half-matched stereograms makes it possible that this explains human psychophysical performance . Additionally , our model neurons lose disparity tuning to half-matched stimuli with increasing receptive field size and dot pattern refresh rate , which is also in agreement with the psychophysical literature [14–16] . Finally , the model neurons show weaker responses to anticorrelated dots , which produce weak or absent depth sensations [14 , 15 , 21–23] . Thus , at least qualitatively , the signal strength in these model neurons parallel all of the psychophysical phenomena that have been used to suggest two stereo mechanisms . However , these manipulations also influence the ratio of the signal to noise , so the response amplitudes described above cannot simply be compared to psychophysical performance . To test more formally whether our model can explain these psychophysical phenomena , we simulated responses of a small population of neurons and made the perceptual decision based on a straightforward linear readout of the population activity . The population and decision rule are described in detail in the Materials and Methods section . Briefly , our starting-point was the modified binocular energy model units discussed above , whose response is denoted by C2 in Eq 1 . For neurons with different RF sizes , we scaled these responses , such that the cells had equal mean responses to correlated stimuli at their preferred disparity . The disparity tuning curves of these cells are shown in Fig 5 in response to correlated RDSs . The peak of the disparity tuning curves are identical for all cells , but the uncorrelated baseline is slightly lower for the fine disparity neurons ( see Methods ) . We then applied Gaussian noise , independently for each unit , with variance proportional to the response at any time and the constant of proportionality being a free parameter . When using a fixed stimulus duration , higher pattern refresh rates mean that model responses average over a larger number of RDS images , reducing the stimulus-driven variability between trials . We therefore fit the noise parameter separately for each frequency . If an equal number of frames is used ( and hence varying stimulus duration ) , the same results can be reproduced with fixed noise across frequencies . For simplicity , we restricted ourselves to four preferred disparities ( −0 . 48° , −0 . 03° , 0 . 03° , 0 . 48° ) , and assumed that receptive field size scaled with disparity magnitude according to σ = 0 . 023 + 0 . 41|Δx| . In simulations using only one cell per disparity , random fluctuations in the monocular image content led to performance that was poorer than human subjects . We therefore included multiple cells for each disparity , differing only in their locations on the retina . Each cell had non-overlapping receptive fields that were otherwise identical . We found a good match to human performance using 40 cells per disparity , for a total of 160 cells in the model population . For the decision rule , we first created opponent cells by taking the difference of a squared energy model neuron and its “antineuron” , i . e . the neuron at the same location in the retina but with preferred disparity differing in sign . To make the decision , each opponent cell’s response was summed across time points to obtain R ( i ) the overall opponent ( squared ) energy for the ith neuron on a given trial ( Eq 7 ) . A negative value for any given neuron-antineuron pair means that the pair signals a negative ( near ) disparity , while a positive value means that the pair signals positive ( far ) disparity . To obtain a decision , we summed the activity across the pool of neuron-antineuron pairs . If this summed value was negative , the model reported that the stimulus had a negative disparity , and vice versa . We first consider the effect of disparity magnitude on stereo depth perception in half-matched RDSs . As Fig 4 shows , large dots ( relative to RF size ) produce stronger responses to half-matched stereograms than small dots . For a stimulus with a fixed dot size , this means that neurons with smaller RFs give stronger disparity signals in half-matched stereograms . Because we include a size-disparity correlation in our model , fine disparities will elicit responses predominantly from neurons with smaller RFs . These cells will see larger correlation fluctuations because of their small RF size and thus have a larger response in the half-matched condition . The results from the simulations are shown in Fig 6a . For half-matched stimuli ( correlation of 0 ) the model performs better in response to fine disparity stimuli than to coarse disparity stimuli . This remains true across a range of correlation values , because even for low correlation values there are local fluctuations in the correlation level . The fine disparity model cells are more sensitive to these fluctuations than the coarse disparity cells , which leads to a leftward shift in the psychometric function . The upward shift at fine disparities arises because here all cells contribute to the disparity judgment ( often with opposing signals , see Fig 5 ) . However , for coarse disparities , the coarse cells dominate the decision to anticorrelated stereograms . These shifts are similar to that observed by Doi et al [14] , but our account invokes only a single mechanism . Next we consider the effect of dot pattern refresh rate on model performance . We used the same model as for the disparity magnitude simulations , but used fine disparity stimuli . We presented RDSs at two different frequencies: 5 . 3Hz and 42 . 5Hz , with a stimulus duration of 1 . 5s . Fig 6b shows the psychometric functions for the same model in response to low and high refresh rates for fine disparities ( 0 . 03° ) . As the pattern refresh rate increases , the psychometric function moves rightward . This shift occurs because the neuron is integrating across more dot patterns ( in this case over time rather than space ) which reduces local fluctuations in correlation . A similar result was obtained using an equal number of stimulus frames rather than equal stimulus duration . Our model produced weak reversed depth in response to anticorrelated stimuli . This is similar to some human studies [14 , 15 , 22] , but the reversed depth reported by our model was also generally stronger than that reported in the literature . This partly reflects the fact that our model cells modulate their activity more than typical V1 neurons . A final output exponent greater than 2 would reduce this , but we present data for the simplest model as it is more tractable . Additionally , responses to anticorrelated stimuli are influenced by factors that are not readily incorporated into simple models . For example , when a zero-disparity annulus is also anticorrelated , depth perception is abolished [22 , 24] , but when the surround is correlated , depth is sometimes reported [14 , 15 , 23] . No existing models provide an account of this effect of the surround . While a sufficiently complicated model could doubtless explain this important phenomenon , adding additional parameters to the model would make it harder to interpret the success in explaining the results we discuss here . To this extent , the description of reversed depth in our model ( and all other extant models ) is a simplification . In our model , the extent to which reversed depth is reported is quite sensitive to the shape of the disparity tuning curves used for model neurons . Despite this , the model does make clear that the information available about negative correlations is influenced by these stimulus manipulations . Our configuration illustrates that it is possible to explain the results of Doi et al [14 , 15] , including reversed depth , using a single mechanism . Here we have have used a linear readout of population activity , which is a parsimonious method of neural decoding , and has frequently been used before [25–27] . It is possible that a more sophisticated decision model could yield performance even more similar to the human observers . Indeed , more complex models , such as those incorporating Bayesian priors [28] , or using maximum likelihood decoding [29] , have been shown to capture human psychophysical performance across a range of stimuli and tasks . However , given the simplicity of our model population ( only two cell types ) , and the fact that the model cells do not faithfully reflect the properties of V1 neurons ( which show stronger attenuation to anticorrelated stimuli ) , exploring more complex decision rules seems inappropriate . Our model neurons show reduced responses to half-matched stimuli as dot size ( relative to RF size ) is reduced . If this correctly captures the nature of signals used to perceive depth , we should expect human performance also to depend on dot size to half-matched RDSs , but not for 100% correlated stereograms . We therefore examined the effect of dot size on depth perception . Fig 7a shows the proportion correct as a function of dot size for 4 observers . For the smallest dot size ( 0 . 025° ) , performance is not significantly different from chance for three of the observers using 95% binomial confidence intervals . We used a Monte-Carlo simulation to test if dot size had a significant effect on the variance in performance and found that it did ( P = 6 . 3 × 10−5 ) . correlated stimuli , all observers performed at ≥ 95% correct for all dot sizes . Thus , the decreased performance in response to small dot half-matched RDSs cannot be attributed to changes in the spatial content of the monocular images . The decreased performance is consistent with predictions made by energy model units with a simple output nonlinearity . Fig 7b compares the average response of the human observers with the psychophysical decision model introduced earlier ( Fig 6 ) . Both the model and the average human performance show a similar decline for the smallest dot size . We showed above that the effects of pattern refresh rate on performance in half-matched stereograms can be explained by the effects of temporal integration on local fluctuations in correlation . Doi et al [15] propose a different explanation , which is that the matching process is slow and so the rapid changes in monocular patterns disrupt the matching computation , leaving the pure correlation computation to dominate at fast refresh rates . In order to provide an additional test of these hypotheses we introduce a new stimulus in which the monocular pattern refresh rate is always high , but binocular correlation changes at different rates . Each random dot pattern is either 100% correlated or 100% anticorrelated ( with the same disparity ) , and the correlation value alternates ( while monocularly every new monitor frame shows a new pattern , as illustrated in Fig 8 ) . We then explored the effect of changes in frequency with which the correlation alternates . By presenting dot patterns at very rapid pattern refresh rates , we should be able to keep the contribution of any putative slow matching computation to a minimum , independent of alternation rate . Doi et al [15] showed that the energy seen by their sustained , matched-based mechanism fell by a factor of 2 as the pattern refresh rate increased from 5 . 3Hz to 43Hz [15] . At 43Hz , the highest refresh rate they could present , their sustained and transient channels were seeing equal stimulus energy . In our CRT mirror stereoscope , we use a pattern refresh rate of 120Hz . According to Doi et al’s [15] model , this will ensure that the transient channel feeding into the pure correlation mechanism is driven far more strongly than the sustained channel feeding into the match-based mechanism . Perception in this stimulus should therefore be dominated by the pure correlation mechanism . However , Doi et al’s [15] model also predicts that even though the pure correlation mechanism is strongly driven , it must perform at chance in this task . Their definition of a pure correlation mechanism is one that outputs 100% veridical depth for 100% correlated stimuli , 100% reversed depth for anti-correlated stimuli , and is at chance ( 50% ) for half-matched stimuli . Now , at alternation rates which are slow compared to the temporal kernel of this mechanism , the observers are simply seeing a rapidly updating stimulus which periodically flips between being correlated and anticorrelated . Let’s say this stimulus has a near disparity . Doi et al’s [15] pure correlation mechanism will report alternately “near” and “far” as the correlation flips . Since we randomised the phase of our alternation , over many trials there is no way for their pure correlation mechanism to report “near” more or less often than “far” . On average , therefore , performance must be at chance . The situation is no better for alternation rates which are fast compared to this mechanism . There , both correlated and anticorrelated frames fall within the temporal integration window . The stimulus is effectively half-matched , and by definition , Doi et al’s [15] pure correlation mechanism must be at chance . Thus , a pure correlation mechanism , as defined by Doi et al [15] , cannot contribute to above-chance performance with this stimulus . Their match-based mechanism can contribute in principle , but their conclusions about its temporal properties—that it is temporally low-pass—makes depth discrimination in rapidly updating stimuli , such as the alternating-correlation stereograms , a particularly demanding task . According to Doi et al [15] , depth perception makes use of “a simple , correlation-based representation for more dynamic ( faster ) and coarser features , and a complex , match-based representation for less dynamic ( slower or stationary ) and finer features” [15] . The implication is that for very rapidly updating stimuli , the relative contribution of the matching computation to depth perception should be very small . Thus , this task presents a particular challenge because the near/far judgment must be based on weak signals from the slow matching computation , and stronger , but alternating and conflicting signals from the fast correlation computation . Yet as Fig 9 indicates , our human observers performed well above chance for alternation rates below 30Hz . We compared the responses of human observers to alternating-correlation RDSs of various alternation frequencies with that of the psychophysical decision model used earlier . Fig 9a shows psychophysical performance in response to alternating-correlation RDSs averaged across four subjects for the three disparities employed . Clearly , stimulus disparity had virtually no effect on this task , where task difficulty was manipulated by increasing the alternation rate . Individual psychometric functions , averaged across disparities , are shown for each subject in Fig 9b . In both plots , model responses are shown in magenta . At alternation rates below 4Hz , the human observers make accurate judgments , but as the alternation rate increases , performance decreases . At intermediate alternation rates , the human observers can still do the task , but crucially , as the alternation rate increases beyond about 20Hz , the correlation variability decreases and the performance of human observers falls to chance . We conclude that our model accounts excellently for human performance in this stimulus . Using the temporal kernel defined by Doi et al [15] for the matching computation , we find a slightly less good fit to the data ( see dotted line in Fig 9 ) . The main difference is at slow alternation rates , where subjects perform very well as does our mechanism with a bandpass kernel . This model is unable to produce good performance here because of the high monocular pattern refresh rate and its lowpass kernel . Presumably , performance would be even worse in a two-mechanism model where the weak signal from the low-pass matching computation competes with a much stronger , alternating signal from the pure correlation computation .
Here we show that all three of these observations can be explained by a model which uses a single nonlinear correlation-based mechanism ( the BEM with a static output nonlinearity ) . The nonlinearity we propose is similar to the nonlinear “cross-matching” model proposed by Doi & Fujita [16]—our proposed mechanism is well described by the linear sum of a “pure-correlation” mechanism and a “cross-matching” mechanism ( Fig 1c ) . Thus the two accounts are closely related . If the pure correlation filter and cross matching filter have the same monocular spatiotemporal RFs , the two accounts are indistinguishable . If the different computations are associated with filters that have different spatiotemporal properties it becomes possible to distinguish one mechanism from two , as two mechanisms will then predict different psychophysical performance . This is exactly what Doi et al [14 , 15] propose to explain psychophysical changes that occur with pattern refresh rate and disparity magnitude . We point out here that many of those changes in psychophysical performance might occur because of the effects of stimulus changes on the activity of a single mechanism . Doi et al [14–16] do not report any simulations with single-mechanism models , and so never test the null hypothesis that a single channel suffices . Our results suggest that it may . Importantly we do not claim to falsify a two-computation hypothesis—the isomorphism shown in Fig 1c means that any data described by a single computation can also be described with two . All three of these mechanisms ( pure correlation , cross-matching , and the BEM with an output nonlinearity ) could be described as “correlation-based” , since they all start by computing correlation . Adding an output nonlinearity to the BEM allows the neurons to signal disparity in half-matched stereograms because the correlation fluctuations are converted to a mean firing rate through this nonlinearity . ( When there is no nonlinearity , the correlation fluctuations manifest as a variable firing rate , but do not lead to a change in the mean ) . It follows from this that the larger the correlation fluctuations , the greater the disparity tuning to half-matched stimuli . We showed that the effect of disparity magnitude on half-matched depth perception can be explained if larger receptive fields are used to detect larger disparities ( the well-known size-disparity correlation [7 , 17 , 19 , 20] ) . A similar point was noted by Doi et al [15] , who found that larger receptive fields decrease the response variability of standard energy model cells , and by Doi & Fujita [16] who extended these findings to a “cross-matching” computation . Along the same lines , the observation that depth perception is compromised in rapidly changing half-matched RDSs is compatible with temporal integration within the correlation mechanism , and does not imply a qualitative shift to a different computation . This model correctly predicts our finding that psychophysical performance decreases with smaller dot size , and states that this is because smaller dots tend to decrease the local correlation variability . It also correctly predicts that alternating the correlation over time should decrease psychophysical performance because of a reduction in the effective variability in binocular correlation . This is particularly interesting since the alternating stimulus presents particular challenges for both of the mechanisms proposed by Doi et al [15] . Their pure correlation mechanism should be at chance since , by definition , it reports opposite depth sign for opposite correlations , and thus reports either depth sign with equal probability for our alternating stimulus . Our single-mechanism model can straightforwardly account for depth perception in these stimuli . Our model has a single , fixed temporal kernel , and yet it can account simultaneously for the effects both of pattern refresh rate ( Fig 6b ) and of alternation rate ( Fig 9 ) . Indeed our model describes the data somewhat better than the “cross-matching” mechanism , most tellingly at low alternation rates . Here subjects are close to 100% correct , whereas the sustained temporal properties of Doi et al’s [16] cross-matching mechanism means that it only reaches 80% correct . The differing contrast polarity of individual dots is a key feature that allows half-matched RDSs to be created . It may therefore be that the psychophysical performance with half-matched RDSs is related to an earlier observation , that human stereopsis performs better on random dot stereograms made up of mixed black and white dots than on stereograms made up of only one polarity [32 , 33] . No correlation-based model has yet provided an account of this phenomenon [33] . Like Doi et al [14–16] , Harris & Parker ( 1995 ) [32] explained their result in terms of a mechanism which matches same-contrast dots ( using independent “on” and “off” channels ) . However , it is not clear that independent on and off channels can explain the even greater benefit produced by mixed polarity dots at low correlation [33] . It may be that the inclusion of nonlinearities like those we use here , or the “cross-matching” of Doi & Fujita [16] would allow a single mechanism to explain the benefits of mixed-polarity dots , but this has yet to be demonstrated . It has long been recognized that in some situations stereo correspondence that is not based on correlation can be exploited . For example , it is well-known that monocular occlusion ( i . e . objects seen by one eye but occluded in the other ) can contribute to the perception of depth in humans ( so-called da Vinci stereopsis [34] ) . Additionally , patients with binocular vision disorders such as strabismus may show no depth perception with cyclopean stereograms , while having measurable stereoacuity in images with monocularly visible contours [35–37] . For isolated monocular targets , humans can correctly report depth for larger disparities than is possible in random dot patterns [38 , 39] . This suggests that the human visual system may also contain a separate algorithm , which enables a coarse form of stereopsis even when the correlation-based system is damaged [40] , at least for sparse images consisting of a small number of monocularly visible objects . There is some evidence suggesting that this system can use head-centric rather than retinotopic coordinates [40–42] , implying an extrastriate locus . Importantly , however , the match-based computation proposed by Doi et al [14–16] is quite different from this mechanism , as it would have to operate on dense random dot patterns . Additionally , in human observers , depth perception in half-matched and anti-correlated stereograms requires the presence of an adjacent correlated region [14 , 15] . This may be related to humans’ greater sensitivity to the relative disparity between adjacent surfaces than to the absolute disparity of a surface in isolation . However , the presence of a correlated surround is not sufficient for reversed depth in anticorrelated RDSs [21 , 43] . As these observations indicate complex interactions between different regions of the visual field , it is inevitable that they cannot all be explained by models like ours ( or that of Doi et al [16] ) that consider only the responses of a population of neurons at one location . That a local model exploiting a single mechanism successfully explains so many phenomena strongly suggests that a single mechanism is responsible . This mechanism closely resembles the known properties of disparity tuning in V1 . The critical property is that V1 neurons show weaker tuning to anti-correlated than to correlated stimuli . The reduction becomes more pronounced in later areas [44 , 45] , but is already present in V1 [10] . This asymmetry suggests that V1 neurons should , weakly , encode disparity in half-matched stimuli . This together with the tendency for V1 neurons tuned to large disparities to have larger receptive fields [46] can account for all the psychophysical data regarding half-matched stimuli . As we have shown , a model based on these ideas provides an excellent account of human performance from previous studies [14 , 15] ( Fig 6 ) and also predicts performance on new stimulus manipulations ( Figs 7 and 9 ) . Importantly , we use one model while only fitting a noise parameter to explain the psychophysical results in Figs 6 , 7 and 9 . It is important to note , though , that no one has yet examined the response of V1 neurons to half-matched stereograms . Our model of V1 neurons captures their weaker tuning to anti-correlated stereograms , and predicts that this results in weak tuning to half-matched stimuli . Yet until this prediction has been directly tested in V1 neurons , we cannot be sure it occurs . If V1 neurons do not show disparity tuning for half-matched stimuli , this would give much greater credence to the idea of a separate dot-matching computation in extrastriate cortex . Our laboratory is currently exploring this question .
We created dynamic random-dot stereograms with a varying number of correlated and anticorrelated dots as described by [14 , 15] . Black and white circular anti-aliased dots , 0 . 09° in radius , were painted on a gray background . When the stimulus was half-matched , half the dots were correlated , i . e . had the same luminance in both eyes , while the other half were anticorrelated , i . e . drawn black in one eye and white in the other . There were on average equal numbers of black and white dots for each correlation value . Dots had zero binocular disparity except within the central 2 . 5° of the stimulus . The surrounding annulus had a width of 1° . The stimulus thus depicted a disparate disk either in front of or behind a zero-disparity background . No subpixel disparity was used . Unless otherwise specified , the dot density was 24% , meaning that if none of the dots overlapped , they would have occupied 24% of the stimulus area . The dots were , however , allowed to occlude one another . The dots were painted in random order so as to prevent any cues arising from occlusion due to either correlated dots systematically occluding anticorrelated dots or the surround dots systematically occluding the center dots . The above applies to both the human psychophysics and the simulations . For the psychophysical investigations , the background was kept 100% correlated for consistency with [14 , 15] . For all simulations , 292 x 292 pixels were used to simulate the stimulus such that 1 pixel corresponded to 0 . 03 degrees . When time was incorporated into the model , the simulations were carried out at a temporal resolution of 1 ms . For the dot size experiment ( Fig 7 ) , we created half-matched ( binocular correlation of 0 , match level of 0 . 5 ) and correlated RDSs ( binocular correlation of 1 , match level of 1 ) . We used three different dot sizes: 0 . 025° , 0 . 05° , and 0 . 075° . The dots were circular and anti-aliased . The surrounding annulus was always correlated as per Doi et al [14 , 15] . The stimuli were presented at a dot pattern refresh rate of 21 . 25 Hz , for 500 ms , and at a disparity of ±0 . 075° . All other features of the RDSs were as previously described . Instead of manipulating binocular correlation over space , as in the half-matched stereogram , we manipulated the binocular correlation over time ( Fig 9 ) . The stimulus was presented at a constant refresh rate– 120 Hz—meaning that a new frame of the dynamic RDS was generated every 8 . 33 ms . The key manipulation was how often the binocular correlation of the dynamic RDS flipped . This could either be at 60 , 30 , 15 , 7 . 5 or 3 . 75 Hz . At 60 Hz , a correlation alternation cycle is completed after two frames , at 30 Hz after four frames and so on . Whether a trial started with a correlated or an anticorrelated frame was randomized . For the alternating-correlation RDSs , the surround had the same correlation as the disparity-defined region . We used a dot density of 200% and six disparities: ±0 . 2275° , ±0 . 1365° , and ±0 . 0455° . The stimulus was presented for 500 ms ( 60 frames ) and following the presentation , the observers were asked to report whether the central disk appeared near or far relative to the background using a mouse press . The stimuli were generated in Matlab and displayed using Psychtoolbox [47] . The stimuli were displayed on a 19” Dell Trinitron CRT monitor . For the dot size experiment , the refresh rate of the monitor was 85 Hz and the resolution was 1024 × 768 pixels . The monitor’s luminance output was linearized prior to the experiment . For all experiments , the Michelson contrast was >99% . For the alternating experiment , the refresh rate of the monitor was 120 Hz and the resolution was 800 × 600 . Dichoptic presentation was ensured through the use of a simple four-mirror haploscope . In both experiments , the observers indicated using a mouse button press whether the central disk appeared near or far relative to the background . For the statistical testing of the effect of dot size ( Fig 7 ) we used a Monte-Carlo method equivalent to 1-way ANOVA ( which does not apply here since the data are binomial proportions ) . For each observer and dot size we generated random draws from a binomial distribution with a fixed probability , equal to the mean across dot sizes for that observer . We then measured the variance in proportion correct across dot size , generating a distribution of values compatible with the null hypothesis . Four observers participated in the dot size experiment , three of whom were male . Six observers participated in the alternating experiment , four of whom were male . In the alternation experiment , two observers’ data were discarded as they were unable to reliably report depth in 100% correlated stereograms . For both experiments , one of the observers was an author , the rest were naive to the purpose of the experiment . All observers had normal or corrected-to-normal vision using spectacles or contact lenses . Both experiments were approved by the Faculty of Medical Sciences ethics committee at Newcastle University . The energy model has been described in detail elsewhere [6 , 9 , 13] . Briefly , we modeled the receptive fields of monocular subunits as two-dimensional Gabors with vertical orientation tuning , ρs ( x , y ) =exp ( − ( x−x0±Δx2 ) 22σx2− ( y−y0 ) 22σy2 ) cos ( 2πf ( x−x0±Δx2 ) + φ ) ( 2 ) x0 and y0 denote the horizontal and vertical receptive field centers , respectively , Δx denotes horizontal disparity and φ denotes phase . σx and σy denote the horizontal and vertical extent of the receptive fields , respectively , and f is the frequency of the Gabor . For all simulations carried out here we used σx = σy , and the receptive field centers x0 and y0 were placed randomly within the disparity-defined region of the stimulus . No phase disparity was used in any of the models . For Figs 3 , 4 and 6 there was no temporal component of the receptive field . For Figs 4b , 6 , 7 and 9 we incorporated time by giving each monocular subunit a biphasic temporal kernel as described by Qian and Freeman ( 2009 ) [48] ρt ( t ) = {1Γ ( α ) ταtα−1exp ( −tτ ) cos ( ωt+ϕ ) if t≥00 otherwise . ( 3 ) For all simulations where a temporal component was incorporated , we used α = 2 . 5 , ω = 4 × 2π , ϕ = − π , and τ = 0 . 035 , which gives a temporal kernel with peak response at approximately 4 . 3 Hz . We chose a biphasic kernel because most V1 neurons have temporal kernels that are bandpass [49] . In Fig 9 , we also used the monophasic temporal kernel from Doi et al [15] . The spatial receptive field and temporal kernel were separable giving ρ ( x , y , t ) =ρs ( x , y ) ρt ( t ) . ( 4 ) Two binocular simple cells were constructed by squaring the sum of two monocular inputs . This produces a binocular simple cell response S= ( VL+VR ) 2=VL2+VR2+2VLVR ( 5 ) where VL and VR denote the left and right monocular responses , respectively . The disparity tuning of the model arises from the cross-term 2VLVR . The BEM models a complex cell by combining simple cells whose receptive fields are π2 out of phase , generating response invariance to stimulus phase . Combining the simple cell responses , we now have C = S1 + S2 . In order to obtain a cell with an amplitude ratio < 1 , we added a static squaring output nonlinearity so that our final model is simply C2 . We computed the disparity tuning curve in Fig 3 by calculating the mean response of the model to 20 000 images displayed at 21 disparities , spanning the range of disparities covered by the neurons’ responses . For generating Fig 4a , we computed the correlated and half-matched response of 30 cells , whose RF size , parameterized by σ , was in the range [0 . 01° , 0 . 3°] . We used 11 dot density values , logarithmically spaced from 0 . 01 to 5 . 12 . The dot size was fixed at 0 . 09° . We computed the response of each of the 30 cells to 20 000 RDSs per density . For Fig 4b , we computed the correlated and half-matched responses for each cell to dynamic RDSs of 11 different frequencies , ranging from 1Hz to 100Hz . The RF sizes and dot size were the same as for Fig 4a . We obtained the model responses by averaging across 5000 trials per frequency-relative RF size combination , where each trial had a duration of 10s . Our model population consisted of 160 neurons tuned to four disparities ( ±0 . 48° , ±0 . 03° ) . Using only a single neuron at each disparity produced significantly poorer performance than human observers , because performance is limited by the fluctuations in image context from trial to trial . With 40 neurons in each group , and fitted noise levels , performance was comparable to humans . Disparity selectivity was introduced with a position disparity between left and right eyes , with a phase disparity of 0 . We made the assumption that receptive field size , parameterized by the standard deviation of the Gaussian envelope σ , scaled with disparity magnitude . Specifically , we had σ = 0 . 023 + 0 . 41|Δx| , where |Δx| is the absolute value of the cell’s preferred disparity , measured in degrees . This is similar to previous modeling work that incorporate disparity-size correlations [7] . The specific parameters we have chosen here are not critical—many different sizes and disparities will yield very similar results for the half-matched stimuli , though the exact shape of the psychometric function varies . The frequency of the monocular Gabors scaled inversely with the receptive field size: f = 0 . 3125/σ , in agreement with physiological estimates [46] . The resulting disparity tuning curves , shown in Fig 5 , were obtained from computing the average response of each cell to 5000 correlated RDSs per disparity . For the tuning curves , we used disparities 51 disparities spaced from −1° to 1° , with disparities near the peak of the fine tuning curves being sampled more finely . All other stimulus parameters were as previously described . Gaussian noise was included in the model , where the variance of the noise at any moment in time was proportional to the response of the cell at that time . The response of the ith neuron to the kth time point is given as Pik=Cik2〈C¯i2〉+κϵik . ( 6 ) where Cik2 is the squared energy model response of the ith neuron to the kth time point , and 〈C¯i2〉 is the mean response of the squared energy model cell to correlated RDSs at the preferred disparity of the cell , presented at 21 . 25 Hz . Dividing by the constant scaling factor 〈C¯i2〉 ensures that all cells , have the same maximum response to correlated stimuli at their preferred disparity . In other words , a value of 0 . 5 in this scheme means that the response was half the mean response to correlated RDSs at the cell’s preferred disparity . εik ∼ N ( μ , σ2 ) is the noise in the model , with μ = 0 and σ2=Cik2/〈C¯i2〉 ( i . e . the variance is proportional to the response magnitude of the cell at any given time ) . κ is a free parameter which governs the magnitude of the noise . This scaling reveals a subtle difference in disparity selectivity with RF size . Because of the final squaring , differences in variability of C lead to differences in mean response . For correlated RDS at the preferred disparity , these fluctuations are correlated in the monocular responses , whereas for uncorrelated RDS they are not . As a result the variability in C is greater for correlated stimuli than for uncorrelated stimuli , and this variability is greater for small RFs than larger ones . Consequently , when scaled by the response to the preferred disparity , smaller RFs show slightly weaker responses to uncorrelated stimuli . Note that the half-matched stimulus introduces additional variation in the binocular correlation , causing responses that are greater than those to uncorrelated dots . Each neuron also has an antineuron whose response is denoted by Nik . The antineuron response is defined exactly the same as Pik above , with the same retinal position , except its disparity is the opposite sign . That is to say , if a neuron P has a disparity preference of 0 . 03° , then its antineuron N would have a disparity preference of -0 . 03° . We created an opponent cell by taking the difference of a neuron and its antineuron . To make the decision , each opponent cell’s response was summed across time points R ( i ) = Σk ( Pik−Nik ) . ( 7 ) R ( i ) thus reflects the overall opponent ( squared ) energy for the ith neuron-antineuron pair on a particular trial . If this value is negative , the pair has a mean signal indicative of a negative disparity , and vice versa for positive values . Within each group of 40 neurons tuned to the same disparity , the models were given non-overlapping RF locations , so that they sampled independent regions of the image . In Fig 6 , we used 1 . 5s trials and summed the responses over time in each model neuron . To obtain the decision , we used a straightforward linear readout of the population response: if the summed activity across the neuron-antineuron pairs was negative , then the decision model would report that the stimulus disparity was negative , and vice versa for positive values: Ψ={1 if ΣiR ( i ) >0 , 0 otherwise . ( 8 ) R ( i ) is as defined in Eq 7 . Ψ = 0 and Ψ = 1 indicate near and far responses , respectively . The biphasic temporal kernel employed here had a peak response to 4 . 3 Hz , meaning that the on-phase of the kernel has a duration of approximately 125 ms . For reference , at 21 . 25 Hz , the dot pattern is updated every 47 . 06 ms and at 120 Hz every 8 . 33 ms . We computed the response of the model to a constant number of RDSs consisting of mixed correlated and anticorrelated dots . Correlation varied from -1 ( completely anticorrelated ) to +1 ( completely correlated ) in steps of 0 . 2 . A correlation of 0 in this scheme corresponds to half-matched RDSs . We presented 10 000 trials at each disparity-correlation combination; all other stimulus parameters were as previously described .
|
Relating neural activity to perception is one of the most challenging tasks in neuroscience . Stereopsis—the ability of many animals to see in stereoscopic 3D—is a particularly tractable problem because the computational and geometric challenges faced by the brain are very well understood . In essence , the brain has to work out which elements in the left eye’s image correspond to which in the right image . This process is believed to begin in primary visual cortex ( V1 ) . It has long been believed that neurons in V1 achieve this by computing the correlation between small patches of each eye’s image . However , recent psychophysical experiments have reported depth perception in stimuli for which this correlation is zero , suggesting that another mechanism might be responsible for matching the left and right images in this case . In this article , we show how a simple modification to model neurons that compute correlation can account for depth perception in these stimuli . Our model cells mimic the response properties of real cells in the primate brain , and importantly , we show that a perceptual decision model that uses these cells as its basic elements can capture the performance of human observers on a series of visual tasks . That is , our computer model of a brain area , based on experimental data about real neurons and using only a single type of depth computation , successfully explains and predicts human depth judgments in novel stimuli . This reconciles the properties of human depth perception with the properties of neurons in V1 , bringing us closer to understanding how neuronal activity causes perception .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"social",
"sciences",
"neuroscience",
"simulation",
"and",
"modeling",
"human",
"performance",
"vision",
"eyes",
"neuronal",
"tuning",
"research",
"and",
"analysis",
"methods",
"animal",
"cells",
"behavior",
"head",
"cellular",
"neuroscience",
"psychology",
"cell",
"biology",
"anatomy",
"neurons",
"psychophysics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"sensory",
"perception",
"ocular",
"system"
] |
2016
|
A Single Mechanism Can Account for Human Perception of Depth in Mixed Correlation Random Dot Stereograms
|
Despite years of intensive research , much remains to be discovered to understand the regulatory networks coordinating bacterial cell growth and division . The mechanisms by which Streptococcus pneumoniae achieves its characteristic ellipsoid-cell shape remain largely unknown . In this study , we analyzed the interplay of the cell division paralogs DivIVA and GpsB with the ser/thr kinase StkP . We observed that the deletion of divIVA hindered cell elongation and resulted in cell shortening and rounding . By contrast , the absence of GpsB resulted in hampered cell division and triggered cell elongation . Remarkably , ΔgpsB elongated cells exhibited a helical FtsZ pattern instead of a Z-ring , accompanied by helical patterns for DivIVA and peptidoglycan synthesis . Strikingly , divIVA deletion suppressed the elongated phenotype of ΔgpsB cells . These data suggest that DivIVA promotes cell elongation and that GpsB counteracts it . Analysis of protein-protein interactions revealed that GpsB and DivIVA do not interact with FtsZ but with the cell division protein EzrA , which itself interacts with FtsZ . In addition , GpsB interacts directly with DivIVA . These results are consistent with DivIVA and GpsB acting as a molecular switch to orchestrate peripheral and septal PG synthesis and connecting them with the Z-ring via EzrA . The cellular co-localization of the transpeptidases PBP2x and PBP2b as well as the lipid-flippases FtsW and RodA in ΔgpsB cells further suggest the existence of a single large PG assembly complex . Finally , we show that GpsB is required for septal localization and kinase activity of StkP , and therefore for StkP-dependent phosphorylation of DivIVA . Altogether , we propose that the StkP/DivIVA/GpsB triad finely tunes the two modes of peptidoglycan ( peripheral and septal ) synthesis responsible for the pneumococcal ellipsoid cell shape .
Bacterial cell growth and division are intimately linked . Complex webs of proteins interacting with each other temporally and spatially control the cellular events leading to the production of two identical daughter cells [1]–[3] . Most of the proteins required for cell division and elongation have been characterized in rod-shaped bacterial models like the Gram-negative bacteria Escherichia coli and Caulobacter crescentus or the Gram-positive bacterium Bacillus subtilis , and robust models depicting their division process are proposed . This knowledge has been beneficial for characterizing and understanding cell division of other bacteria . However , some aspects related to cell division , including the achievement of cell shape , are often hardly transposable and species-specific mechanisms exist to allow cells to divide , assume a given shape and/or cope with their environment [4] , [5] . In the Gram-positive human pathogen Streptococcus pneumoniae ( the pneumococcus ) , some conserved division proteins have been studied , but overall , little is known about the mechanisms governing cell division and those responsible for peptidoglycan ( PG ) synthesis , as well as how this species achieves its characteristic ellipsoid ( rugby-ball like ) shape [6]–[10] . Early studies have suggested that S . pneumoniae utilizes a combination of two PG synthesis modes , namely septal and peripheral [11] . Due to the absence of the actin-like protein MreB and any homologues in the pneumococcus , it is speculated that these two modes of PG synthesis are coordinated with and organized by FtsZ-ring formation [12] . A two-state model in which two different PG synthesis machineries are responsible for either septal or peripheral synthesis has been proposed . In this model , the PG transpeptidase PBP2x , a penicillin binding protein ( PBP ) that catalyzes PG cross-linking , and the lipid-flippase FtsW , that transports lipid-linked PG precursors from the inner to the outer leaflet of the cytoplasmic membrane , belong to a septal machinery and are exclusively required for septal PG synthesis and cell separation . On the other hand , the transpeptidase PBP2b and the lipid-flippase RodA would be exclusively associated with a peripheral machinery , and required for peripheral PG synthesis and cell elongation . However , it is unclear how S . pneumoniae would coordinate peripheral and septal synthesis . An interesting possibility comes from work in B . subtilis showing that cell elongation-division cycles are controlled by shuttling of PBP1 , a transpeptidase/transglycosidase class A PBP involved in peptidoglycan polymerization [13] . PG synthesis could thus be fine-tuned by a yet uncharacterized process to allow the alternate synthesis of septal and peripheral PG in pneumococcus . StkP , a membrane eukaryotic-type serine/threonine kinase , represents an attractive candidate to regulate septal and peripheral PG synthesis in S . pneumoniae . This kinase has been recently shown to play an important role in pneumococcal cell division and growth [14]–[16] . In vivo , only a few proteins appear specifically phosphorylated by StkP [14] , [17] , [18] . Among them , it is noteworthy to find the division protein DivIVA , which was shown to be phosphorylated by StkP on Thr-201 . Expression of the non-phosphorylatable form of DivIVA ( i . e . , a mutant in which Thr-201 is substituted for an alanine ) induced severe defects in cell shape , possibly by affecting pole maturation [14] . Interestingly , a DivIVA paralog , named GpsB [19] , was identified in B . subtilis and shown to be involved together with EzrA in PBP1 shuttling between elongation and division sites [13] . Global phosphoproteome analyses of B . subtilis and Streptococcus agalactiae indicated that GpsB is phosphorylated in these species [20] , [21] . GpsB is also found in S . pneumoniae , as in most Firmicutes . This situation prompted us to investigate the role of the two paralogs , GpsB and DivIVA , in PG synthesis and cell morphogenesis of S . pneumoniae , and to examine whether their phosphorylation by StkP could affect their role . Here , we establish that GpsB and DivIVA are both crucial for cell morphogenesis , and demonstrate that DivIVA is necessary for cell elongation whereas GpsB acts as a negative regulator of DivIVA to prevent cell elongation . Moreover , we show that GpsB is not phosphorylated , but required for StkP septal localization and subsequent phosphorylation of DivIVA . In light of these observations , we propose that the StkP/DivIVA/GpsB triad finely tunes the two modes of PG synthesis to achieve the ovoid shape of pneumococci and we discuss the relevance of this process in other bacteria . Our observation of similar localization patterns for the transpeptidases PBP2x and PBP2b as well as for the lipid-flippases FtsW and RodA in cells deficient for GpsB and/or DivIVA questions the existence of two distinct PG biosynthesis machineries .
To analyze the potential role of DivIVA in pneumococcal morphogenesis , we constructed a nonpolar , markerless divIVA-null mutant and investigated its cell morphology . As previously reported for the S . pneumoniae RX1 strain [6] , [22] , 99 . 8% of ΔdivIVA R800 cells exhibited a striking chain phenotype ( Figure 1A and Table S1 ) . When the divIVA mutation was repaired back to wild type ( WT ) by transformation , the morphology of the resulting strain was similar to that of the WT strain with a typical diplo-ovococcal shape ( compare Figure 1A with Figure S1 ) indicating that the chain phenotype is due to the deletion of divIVA . ΔdivIVA chains contained up to several dozen of tightly joined cells separated by well-defined membranes ( Figures 1A and S2A ) . Cells were clearly not ovoid but flattened at the poles , exhibiting a rounded shape . Analysis of individual cell parameters further confirmed this visual impression and showed that divIVA deletion resulted in reduced pneumococcal cell length ( Figure S2B and Table S1 ) . We also examined ΔdivIVA cells by scanning- and transmission-electron microscopy ( SEM and TEM ) . Using SEM , cells seemed to be interlocked into the neighboring ones ( Figure 1B ) . Nevertheless , TEM indicated that cells were clearly separated by membranes , consistent with efficient Z-ring constriction and closure , and suggesting that septal PG is efficiently produced ( Figure 1C ) . To confirm the latter , we applied the strategy described by Kuru and co-workers [23] and PG synthesis was visualized using Bodipy-FL containing fluorescent D-amino acid , namely Bodipy-FL-amino-D-alanine , or BADA [24] . More specifically , the exponentially growing cells are pulsed with BADA for 4 min corresponding to ca . 10–12% of the generation times of the WT and the mutants . As a control , we checked that BADA labeled the division site in WT cells as previously described using fluorescent vancomycin [14] ( Figure 2 ) . BADA labeling of ΔdivIVA cells revealed PG synthesis localizing exclusively as bands across the cells at the division septa in 99 . 4% of cells ( Figure 2 and Table S2 ) . Altogether , these results suggest that in the absence of DivIVA , cell elongation is hampered while septum closure still occurs . On the other hand , the last step of cell division allowing the final separation of daughter cells is affected . To unravel the role of GpsB in S . pneumoniae , we first constructed a nonpolar markerless deletion mutant of gpsB in strain R800 . In contrast to the deletion of divIVA , the deletion of gpsB severely affected growth ( Figure S3A ) and cell viability decreased to 60% suggesting that GpsB is crucial for the pneumococcus . Microscopy analysis revealed a striking phenotype characterized by the presence of very elongated cells ( Figures 1A , S3B and Table S1 ) . Morphometric measurements indicated that the length was below 1 . 4 µm for 90% of wild-type cells ( WT ) , whereas nearly 90% of ΔgpsB cells exhibited length greater than 1 . 3 µm . ΔgpsB cells seemed irregularly shaped and septal membranes across cells were lacking , indicating that cell constriction was seriously hampered ( Figure 1A ) . Examination of the ultrastructure of ΔgpsB cells by TEM and SEM confirmed that mutant cells displayed a strongly affected morphology with irregular width ( Figures 1B–C ) . The presence of several septal initiations positioned asymmetrically on each side of the long axis of the cells was detected by TEM ( Figure 1C ) . SEM images confirmed an irregular elongation of ΔgpsB cells which displayed a “twisted-towel” shape ( Figure 1B ) . One could further observe the presence of a helical groove at the surface of the cells that seemed to correspond to the asymmetric septal initiations detected by TEM , suggesting that the divisome is stretched upon cell elongation . Importantly , a wild-type diplo-ovococcal shape and normal growth were restored when the ΔgpsB strain was transformed back to gpsB+ confirming that the observed phenotype resulted from the inactivation of gpsB ( compare Figure 1A with Figure S1 ) . Deletion of gpsB was also attempted into four other well-characterized and widely used S . pneumoniae strains , the encapsulated D39 and TIGR4 , and the unencapsulated R6 and RX1 . The same elongated phenotype was observed with gpsB− derivatives of the unencapsulated strains , while our efforts to delete gpsB failed with both encapsulated strains ( Figure S3C ) . The latter observation is consistent with the recent report of Land and co-workers [25] . Altogether , these data show that deletion of gpsB triggers cell elongation and prevents proper pneumococcal cell division . We then analyzed the effect of the deletion of gpsB on FtsZ localization . For this purpose , we first constructed a C-terminal GFP fusion to FtsZ . Throughout this study and unless otherwise indicated , C-terminal and N-terminal fusions ( denoted respectively Protein-GFP and GFP-Protein ) were constructed at each native chromosomal locus , expressed under the control of the native promoter and represented the only source of protein . The FtsZ-GFP fusion seemed fully functional as cells grew as rapidly as WT cells and did not display any shape defect ( Figures 3A and S4A ) . As expected , FtsZ-GFP localized at midcell in exponentially grown WT cells as well as in ΔdivIVA rounded cells ( Figure 3A and Table S2 ) . FtsZ-GFP also appeared as several transversal bands in a majority of ΔgpsB cells . Interestingly however , a number of cells ( 25 . 2% ) displayed a zig-zag localization of FtsZ ( Figures 3A , S5 and Table S2 ) . To clarify the dimensional nature of this zig-zag structure , we carried out deconvolution microscopy ( Movie S1 ) that revealed a continuous FtsZ-helical organization . This unexpected helical localization of FtsZ in ΔgpsB cells was confirmed by immunostaining using anti-FtsZ antibodies ( Figure 3B ) indicating that the helical pattern is not an artifact due to the GFP tag on FtsZ . Control experiments also confirmed that FtsZ levels were unaffected in the absence of GpsB ( Figure S4B ) . To reconcile the dual FtsZ localization ( spirals vs . bands ) in ΔgpsB elongated cells , time-lapse microscopy was performed . During cell elongation , some Z-rings were replaced by helical structures which eventually split to generate several new Z-rings , whereas others continued to stretch ( Figure 3C and Movie S2 ) . However , in both cases cells ended up bursting and dying ( e . g . , cells marked with arrows in Figure 3C ) . We conclude from these observations that the Z-ring is replaced by a helical structure during elongation of ΔgpsB cells . To test whether cell elongation in ΔgpsB cells was accompanied by altered localization of PG synthesis , the latter was labeled using BADA . BADA labeling revealed a helical organization of neosynthesized PG in 28 . 7% of ΔgpsB elongated cells ( Figure 2 and Table S2 ) , which is comparable to the percentage of cells exhibiting GFP-FtsZ spirals . This prompted us to examine the localization of PBP2x , PBP2b , FtsW and RodA , which are involved in PG synthesis , using GFP-PBP2x , GFP-PBP2b , FtsW-GFP or RodA-GFP . All four GFP-fused proteins were functional as cells grew normally and displayed WT shape ( Figures 4A and S6A ) . Fluorescence microscopy indicated that PBP2x , PBP2b , FtsW , and RodA localize at midcell in WT cells as well as in ΔdivIVA cells ( Figures 4A and 4C ) . Strikingly , they all appeared mislocalized in elongated ΔgpsB cells , exhibiting a helical pattern ( Figure 4B and Table S2 ) reminiscent of that observed for FtsZ ( Figure 3A ) and PG synthesis ( Figure 2 ) . Western blot control experiments confirmed that the four GFP-fusions were produced at similar levels in WT , ΔdivIVA and ΔgpsB cells , excluding any artifact due to aberrant protein expression ( Figure S7 ) . These helical patterns suggested that the four GFP-proteins could co-localize with FtsZ . To directly assess this , we constructed double-labeled strains containing FtsZ fused to RFP and either PBP2x , PBP2b , FtsW , or RodA fused to GFP . Cells containing a pair of fusion proteins in an otherwise WT background exhibited a growth delay ( Figure S6B ) indicating that the combination of FtsZ-RFP with GFP-fused PBP2x , PBP2b , FtsW , or RodA is somehow detrimental . Nevertheless , microscopy analyses indicated that cell shapes were normal and that each pair of RFP/GFP-fusions co-localized properly at midcell ( Figure 5A ) . On the other hand , when we detected helical RFP-FtsZ in ΔgpsB elongated cells , PBP2x , PBP2b , FtsW and RodA also displayed an helical organization co-localizing with helical FtsZ ( Figure 5B ) . To examine DivIVA localization in ΔgpsB elongated cells , we first generated a WT strain producing DivIVA-GFP . WT cells expressing DivIVA-GFP grew normally and displayed a classic ovoid-shape ( Figures 6A and S8A ) establishing that the DivIVA-GFP fusion is functional . In agreement with a previous report [15] , DivIVA-GFP localized at both midcell and the cell poles ( Figure 6A ) . By contrast , DivIVA-GFP exhibited a helical organization in 20 . 1% of elongated ΔgpsB cells ( Figure 6A and Table S2 ) . This phenotype was not due to an aberrant expression of DivIVA-GFP since western blot analyses confirmed that the fusion protein was synthesized at similar levels in WT and ΔgpsB cells ( Figure S8B ) . DivIVA localization was thus comparable to that of FtsZ in ΔgpsB cells ( Figure 3A ) suggesting that the two proteins co-localize during cell elongation . To confirm this , we constructed WT and ΔgpsB strains expressing both DivIVA-GFP and FtsZ-RFP . As expected , FtsZ and DivIVA displayed respectively septal and septal/polar localization in WT cells ( Figure 6B ) . In ΔgpsB cells , DivIVA co-localized with FtsZ in a helical pattern ( Figure 6B ) . To further investigate the role of GpsB and DivIVA , we introduced the divIVA deletion in cells deficient for gpsB . The double mutant was readily obtained . FM4–64 membrane staining showed that ΔdivIVAΔgpsB cells exhibited the same cell shape and chain phenotype as ΔdivIVA cells ( 97 . 9% of cells ) , although a few cells were irregularly shaped ( Figure 1A and Table S1 ) , indicating that inactivation of divIVA suppressed the elongated cell phenotype typical of ΔgpsB cells ( the same was observed with a double mutant constructed by introducing the ΔgpsB mutation into ΔdivIVA cells; data not shown ) . However , cell viability decreased to 60% and was comparable to that of ΔgpsB cells . As controls , FtsZ-GFP still localized at the division septa in the absence of DivIVA and GpsB , and was produced to WT levels ( Figure 3A , Table S2 and compare Figures S4B with S9B ) . These observations suggest that GpsB interplays with DivIVA to coordinate cell elongation and cell division , and that GpsB is dispensable for septal PG synthesis when DivIVA is absent . To further test this hypothesis , we analyzed BADA labeling of PG in ΔdivIVAΔgpsB cells . We observed that PG is produced properly at the division site ( Figure 2 and Table S2 ) . Likewise , we analyzed the localization of PBP2x , PBP2b , FtsW and RodA in ΔdivIVAΔgpsB cells ( Figure 4D and Table S2 ) . All of them still localized to the division septa as observed in ΔdivIVA cells ( Figure 4C and Table S2 ) . Western blot control experiments confirmed that the four GFP-fused proteins were produced at similar levels in WT and ΔdivIVAΔgpsB cells ( Figure S7 ) . These observations indicate that deletion of gpsB is tolerated in a ΔdivIVA mutant without inducing further detectable cell shape and septum closure defects and does not impair PG synthesis at the division site . To gain an insight into a possible connection between GpsB , DivIVA and cell division , we first looked for physical interactions between these proteins and FtsZ using a bacterial two-hybrid screen [26] . Neither GpsB nor DivIVA were found to interact with FtsZ ( Figure S10A ) . However , as the cell division protein EzrA was found to bridge GpsB with the Z-ring in B . subtilis [13] , we also analyzed EzrA interactions with GpsB , DivIVA and FtsZ . Reproducible interactions were detected between EzrA and either GpsB , DivIVA or FtsZ by bacterial two-hybrid assays ( Figure S10A ) . These interactions were further analyzed by surface plasmon resonance ( SPR ) , which confirmed that EzrA interacts with GpsB ( KD = 770±230 nM ) , DivIVA ( KD = 530±75 nM ) and FtsZ ( KD = 295±60 nM ) ( Figures S10B and S10C–E ) . We also tested whether GpsB interacts with DivIVA . Reproducible interactions were first detected with the two-hybrid screen ( Figure S10A ) and SPR confirmed that GpsB interacts with DivIVA ( KD = 85±14 nM ) ( Figures S10B and S10F ) . We then analyzed the localization of EzrA and GpsB fused to GFP . WT cells producing GFP-GpsB or GpsB-GFP appeared elongated and displayed aberrant cell shapes indicating that both fusions were not fully functional ( Figure S11 ) . We therefore constructed a merodiploid strain carrying an ectopic gfp-gpsB fusion under the control of the zinc-inducible PZn promoter at the non-essential bgaA locus . Fluorescence microscopy indicated that GFP-GpsB localizes as bands across the short axis of the cells at the division septum in WT cells and the same observation was made in ΔdivIVA rounded cells ( Figure 7A and Table S2 ) . By contrast , while EzrA-GFP localized at midcell in exponentially grown WT cells as well as in ΔdivIVA and ΔdivIVAΔgpsB cells , EzrA-GFP formed helical structures that extended across the long axis of the cell in 19 . 9% of ΔgpsB elongated cells ( Figure 7B and Table S2 ) as was found for FtsZ-GFP and DivIVA-GFP ( Figures 3A and 6 ) . Western blot control experiments confirmed that GFP-GpsB and EzrA-GFP were produced at similar levels in WT and ΔdivIVA ( and ΔgpsB and ΔdivIVAΔgpsB cells for EzrA ) ( Figures S8B and S9 ) . These observations are consistent with EzrA serving as a connector between FtsZ and GpsB and/or DivIVA . The elongated phenotype with incomplete septa displayed by ΔgpsB cells was reminiscent of that reported for cells expressing the kinase-dead form of StkP and suggested a relationship between these two proteins [14] , [15] . Hence , we hypothesized that GpsB could be phosphorylated by StkP in S . pneumoniae . The general phosphorylation pattern of crude extracts of pneumococcal cells was thus analyzed using anti-phosphothreonine antibodies . We detected an intense phosphorylation signal around 15 kDa , which could be compatible with the phosphorylation of GpsB ( 13 kDa ) ( Figure 8A ) . Therefore , GpsB from S . pneumoniae cells was purified to examine its in vivo phosphorylation state using high-resolution based mass spectrometry ( Figures S12A–B ) . No phosphorylated sites were detected suggesting that GpsB is not phosphorylated in vivo . To confirm this result , we analyzed the phosphorylation pattern of cells expressing only GFP-GpsB . An intense phosphorylation signal at 15 kDa was still detected and no new phosphorylation signal appeared around 45 kDa , the predicted mass of GFP-GpsB ( Figure 8A ) , though GFP-GpsB was efficiently stained with anti-GFP antibodies ( Figure 8B ) . Altogether , these data show that GpsB is not phosphorylated in vivo and that the 15-kDa phosphorylation signal corresponds to another unidentified protein . In parallel , we analyzed the phosphorylation pattern of ΔgpsB cells using anti-phosphothreonine antibodies . Surprisingly , the deletion of gpsB abolished not only the phosphorylation of all the substrates of StkP , including DivIVA , but also StkP autophosphorylation itself ( Figure 8A ) . Nevertheless , StkP was expressed at similar levels in WT and ΔgpsB cells ( Figure S12C ) . Furthermore , SPR analysis showed that GpsB was able to interact with the inactive cytoplasmic domain of StkP-K42R ( KD = 500±60 nM ) ( Figure S10G ) . These data raised the question of whether GpsB could affect StkP septal localization [14] , [15] ( Figure 8C ) . Therefore , we constructed a ΔgpsB mutant harboring a GFP-StkP fusion [14] . Fluorescence microscopy revealed an intense signal distributed all around the cell , consistent with a diffuse membrane localization of GFP-StkP ( Figure 8C ) . StkP localization and general phosphorylation patterns of ΔdivIVA and ΔdivIVAΔgpsB cells were also analyzed . While StkP localized to midcell and was able to phosphorylate its targets in the absence of DivIVA , a deficient phosphorylation pattern as well as a StkP diffuse membrane localization were observed in the double mutant ( Figures 8A and 8C ) . In both mutant strains , StkP was produced to similar levels as in WT cells ( Figure S12C ) . These data establish that , although GpsB is not phosphorylated , it is crucial for StkP septal localization and its capacity to autophosphorylate and phosphorylate its substrates , notably DivIVA . By contrast , DivIVA is not required for StkP kinase activity and localization . Because GpsB is required for StkP septal localization and thus phosphorylation of DivIVA , we investigated the impact of either gpsB or divIVA deletion on the elongated morphology of cells producing the kinase-dead form of StkP , StkP-K42M [14] . As shown in Figure 8D , ΔgpsB-stkP-K42M cells still displayed an elongated phenotype supporting the idea that the main function of GpsB is to allow StkP to phosphorylate its targets . By contrast , the deletion of divIVA was found to abrogate stkP-K42M cell elongation . Indeed , ΔdivIVA-stkP-K42M cells formed chains of rounded cells similar to those of the ΔdivIVAΔgpsB mutant ( compare Figure 8D with Figure 1A ) . This observation confirms that DivIVA is crucial for cell elongation and the late step of cell separation , and further suggests that its non-phosphorylation likely results in aberrant elongation of stkP-K42M and ΔgpsB cells .
According to the current model of PG synthesis in pneumococcus , cell elongation is due to the peripheral PG synthesis . Our observations suggest that peripheral PG is impaired in ΔdivIVA cells ( Figures 1 and S2 ) . In addition , we show that DivIVA co-localizes with FtsZ in ΔgpsB elongated cells and that divIVA deletion suppresses ΔgpsB cell elongation ( Figures 1A and 6B ) . Therefore , we propose that one function of DivIVA is to switch from septal to peripheral PG synthesis to trigger cell elongation . DivIVA performs quite different functions in B . subtilis and Staphylococcus aureus [27] , but it has already been shown to participate in cell wall biosynthesis in bacteria such as Streptomyces coelicolor , Mycobacterium tuberculosis and Corynebacterium glutamicum that either lack or do not require MreB for their vegetative growth . In these bacteria , DivIVA is required for polar growth allowing tip extension and cell elongation [28]–[30] . Pneumococcus is devoid of MreB and any identifiable homologues . Altogether , these observations raise the possibility that DivIVA is crucial for cell elongation in those species in which vegetative growth is not dependent on MreB . The chaining displayed by ΔdivIVA cells also suggests that while septum closure leading to the separation of the daughter cell cytoplasms is normal , their final separation is somehow affected , as previously reported for the RX1 strain [6] . DivIVA has been previously found to interact or to contribute to the positioning of some PG hydrolases in the pneumococcus [31] , [32] or in autolysin secretion in other bacteria as Listeria monocytogenes [33] . The chain phenotype displayed by the ΔdivIVA mutant is consistent with impairment of PG hydrolysis and remodeling required for final separation of daughter cells . Previous studies using high-throughput gene disruption approaches have suggested that gpsB could be essential in pneumococcus [34]–[36] . In this study , we show that GpsB is actually not essential for pneumococcal laboratory strains ( Figures 1 and S3 ) . However , and in agreement with the previous observations , no ΔgpsB transformants could be obtained with the pathogenic strains D39 and TIGR4 indicating that the requirement for GpsB depends on the genetic background . The recent work of Land and co-workers also suggests that suppressive mutations are required for growth of unencapsulated derivatives of pathogenic strains expressing low level of GpsB [25] . The conditional essentiality of GpsB is reminiscent of the situation with MreC and MreD . These proteins are essential in D39 and TIGR4 pathogenic strains but not in the R6 laboratory strain due to suppressive mutations in PBP1a and in proteins of unknown function in the latter [9] . Inactivation of gpsB resulted in severely impaired cell division , with a large fraction of the population appearing as elongated cells with incomplete septa similar to cells producing the kinase-dead form of StkP [14] . This phenotype is accompanied by helical patterns for PG synthesis and FtsZ , along the long axis of the cell ( Figures 2 and 3 ) . Interestingly , Z-spiraling was not observed in the work published by Land and co-workers [25] . Rather , the authors detected multiple non-constricted rings of FtsZ in elongated cells . Because GpsB expression was under the control of an inducible fucose promoter , we tentatively attribute the absence of Z-spiraling to low level of gpsB expression in absence of fucose , likely preserving Z-ring formation in elongated cells . This hypothesis is consistent with our time-lapse analysis showing Z-spiraling upon cell elongation , eventually splitting to generate several new Z-rings ( Figure 3C and Movie S2 ) . The stimulation of cell elongation and aberrant helical organization of the divisome observed in our study in ΔgpsB cells suggest that GpsB is a negative regulator of cell elongation in WT cells . On the other hand , the deletion of gpsB has no effect on cell shape and septum closure in the absence of DivIVA ( Figure 1A ) , and septal localizations of PG synthesis , PBP2x , PBP2b , FtsW and RodA are not affected ( Figures 2 and 4D ) . However , the deletion of both divIVA and gpsB genes has a detrimental effect on cell viability . It could thus be proposed that in absence of DivIVA , GpsB is dispensable for septal PG synthesis but required for optimal cell survival . We also show that GpsB interacts with StkP and is crucial for both StkP localization at the division site and its ability to phosphorylate its targets , including DivIVA ( Figure 8 ) . This represents the first evidence of STPKs regulation by a cell division protein . Consequently , GpsB becomes the major determinant of pneumococcal cell division . Most of StkP targets remain to be identified but their phosphorylation is crucial for cell division [14] . On this basis , one cannot exclude that deficient phosphorylation of StkP targets in the absence of GpsB could favor FtsZ-ring spiraling . In other words , FtsZ could be prone to move in a spiral driving peripheral PG synthesis in the absence of GpsB , yet preventing septal PG synthesis and cell division . The function of GpsB could be more complex than promoting septal PG synthesis during septum closure and also involved in mediating proper condensation of the divisome at midcell . Analysis of protein-protein interactions revealed that GpsB and DivIVA do not interact with FtsZ but with the cell division EzrA , which itself interacts with FtsZ ( Figure S10 ) . Together with the helical organization of EzrA and the co-localization of DivIVA with FtsZ in ΔgpsB elongated cells , we propose that septal and peripheral PG synthesis are coordinated with and organized by FtsZ via EzrA , GpsB and DivIVA . Considering the opposing function of DivIVA and GpsB in cell elongation , and the finding that inactivation of divIVA in ΔgpsB cells results in the disappearance of elongated cells , we propose that GpsB is required to confine PG synthesis at the division site and to negatively control cell elongation promoted by DivIVA . GpsB and DivIVA are also found to interact . Therefore , we propose that GpsB and DivIVA constitute a molecular switch , connected to FtsZ via EzrA , that orchestrates the production of peripheral ( cell elongation ) and septal ( cell division ) PG to confer to the pneumococcus its characteristic ovoid shape . How could this switch operate ? The finding that inactivation of gpsB affects both StkP septal localization and kinase activity , and thus DivIVA phosphorylation , leads us to propose that cell elongation is stimulated by non-phosphorylated DivIVA and that DivIVA phosphorylation by StkP abolishes its ability to promote cell elongation . Suppression of the elongated cell shape of the stkP-K42M mutant upon divIVA deletion is consistent with this hypothesis ( Figure 8D ) . The current model of PG synthesis in S . pneumoniae , and more generally in ovococci , proposes that the two modes of PG synthesis depend on the action of two distinct machineries [12] , as described for rod-shaped bacteria . Recently , Land and co-workers have analyzed the localization of PBP2x and PBP1a over the cell cycle [25] . These two enzymes display similar localization patterns in pre- and mid-divisional cells , but not during septum closure . Indeed , PBP1a localized as a ring larger than that of PBP2x . This observation was interpreted as supporting the existence of two distinct PG synthesis machineries . While PBP2x is essential and participates in cell constriction ( septal PG synthesis ) , the role of PBP1a in PG synthesis remains elusive . A pbp1a mutant is affected both in length and width but cells grow normally with no viability defects and cells remain ovoid rather than being elongated or rounded [9] , [37] , [38] . In the two-machinery model , PBP1a would be involved in both elongation and constriction . That PBP2x and PBP1a display different localization dynamics during septum closure does not necessarily imply that they belong to two distinct machineries . Here we have analyzed the localization of PBP2x and FtsW as well as PBP2b and RodA , which are proposed to be specific for septal and peripheral PG synthesis , respectively , in the two-machinery model . We show that they all co-localize with helix-shaped FtsZ in elongated ΔgpsB cells ( Figure 5B ) . In addition , we failed to delete the genes encoding PBP2x , PBP2b , FtsW or RodA in ΔgpsB , ΔdivIVA or ΔdivIVAΔgpsB cells indicating that all these proteins remain essential even when septal or peripheral PG synthesis is impaired . Therefore , our data hardly fit with ( and challenge ) the two-machinery model . An exciting and promising alternative conciliating the data reported by Land and co-workers with ours would be that the four proteins are present in a same unique complex , ensuring both septal and peripheral PG synthesis , whose composition varies in the course of the cell cycle . A previous study has demonstrated that a first short step is dedicated to cell elongation ( around 300 nm ) ( peripheral PG synthesis ) [39] . This is followed by a second step , in which cell constriction ( septal PG synthesis ) occurs simultaneously with elongation at mid-cell of the forming daughter cells , and a third step dedicated to constriction . In the second step , PG synthesis is distributed along progressively constricting circles converging toward the future new cell pole to achieve both elongation and septation . These observations are consistent with a finely tuned single machinery allowing concomitant cell elongation and constriction , with components displaying different localization dynamics toward the future equatorial division site . Considering these constraints imposed by an ovoid cell shape , a unique machinery thus represents an attractive mean to achieve PG synthesis along progressively constricting circles . Deciphering these mechanistic questions will certainly require implementing higher resolution microscopy approaches than 3D-SIM , such as PALM or STORM , to assess the dynamics of each components of the division machinery over the pneumococcus cell cycle and particularly during the second step involving simultaneous cell constriction and elongation . Using a depletion approach , Berg and co-workers recently reported that lowering the amount of either PBP2b or PBP2x in pneumococcus results in lentil-shaped and lemon-shaped-cells , respectively [40] . These cell shapes are distinct from that of ΔdivIVA rounded cells and ΔgpsB elongated cells . An interpretation could be that while the catalytic activity of the PBPs is important to specifically achieve septal or peripheral synthesis , they are both structurally ( physically ) required for the two PG synthesis modes . This would be consistent with our observations and further supports our model in which PG synthesis would depend on a single machine responsible for both septal and peripheral PG synthesis ( Figure 9 ) . In such a model , we propose that the StkP/DivIVA/GpsB triad finely tunes this machine to dictate the type of PG ( septal or peripheral ) produced . Investigating the underlying regulatory mechanism , which might involve modification of DivIVA interactions with EzrA , GpsB , or other partners in the divisome presumably via StkP-driven phosphorylation , will likely improve the understanding of how septal and peripheral PG synthesis are coordinated . Phosphorylation of GpsB and/or DivIVA homologs has previously been detected in B . subtilis , S . coelicolor , S . agalactiae and M . tuberculosis [20] , [21] , [41] , [42] . However , phosphorylation sites are not conserved or occur in regions of poor amino acid conservation ( Figure 10A ) . In addition , phosphothreonines can be replaced by glutamic acids , as revealed by alignment of GpsB sequences from several streptococci ( Figure 10B ) . Interestingly , negatively charged amino acids ( Asp/Glu ) can mimic the phosphorylated state of a protein . A recent comparative genomic study indicated that nature uses this trick in reverse by evolving serine , threonine , and tyrosine phosphorylation sites from Asp/Glu residues [43] . It is thus possible that GpsB and DivIVA phosphorylation by StkP is a widespread means for finely tuning cell-wall synthesis and defining bacterial cell shape though the underlying mechanism may differ between species .
For growth experiments , S . pneumoniae strains were cultivated at 37°C in Todd-Hewitt Yeast ( THY ) broth ( Difco ) . For induction of PZn , ZnCl2 was added at the concentration of 0 . 15 mM . For construction of S . pneumoniae mutants , transformation was performed as described previously [44] , using precompetent cells treated at 37°C with synthetic competence stimulating peptide 1 ( CSP 1 ) to induce competence . Transformants were plated into THY-agar supplemented with 3% ( vol/vol ) defibrinated horse blood and then incubated for 120 min at 37°C . Selection was then performed by adding a 10 ml THY-agar overlay containing the appropriate antibiotic ( streptomycin 200 µg/ml , kanamycin 250 µg/ml , tetracyclin 2 , 5 µg/ml ) and overnight incubation at 37°C . For viability assays , several samples of exponentially growing cells were taken every 30 min , diluted appropriately and plated onto THY-agar supplemented with horse blood . After overnight incubation , colony-forming units ( CFU ) were counted and the percentage of viability of mutant strains was expressed relatively to the WT strain . The Escherichia coli XL1-Blue strain was used as a host for cloning . E . coli BL21 ( DE3 ) strain was used as host for overexpression . The E . coli BTH101 was used as host for bacterial two-hybrid analysis . Luria–Bertani ( LB ) broth and agar supplemented with appropriate antibiotic ( tetracyclin 15 µg/ml , ampicillin 100 µg/ml , and kanamycin 25 µg/ml ) were used for routine growth at 37°C . The nucleotide sequences of all synthesized DNA fragments were checked to ensure error-free amplification . Strains used in this study are listed in Table S3 . DNA fragments coding for GpsB , DivIVA , FtsZ , inactive StkP cytoplasmic domain and EzrA without its N-terminal transmembrane domain were obtained by PCR using chromosomal DNA from S . pneumoniae R800 strain as template and oligonucleotides described in Table S4 , section 2 . Site directed mutagenesis of StkP kinase domain was achieved by 2 successive PCRs using chromosomal DNA as template and primer pair IX/XI and then the resulting DNA fragment and primer X . The obtained DNA fragments were cloned between the NdeI and BamHI cloning sites of the pETPhos plasmid ( except for ezrA that has been inserted using NheI and BamHI ) [45] . To construct PZn-gfp-gpsB , gfp was first amplified using the primer pair XII/XIII ( Table S4 section 3 ) using pUC57-gfp as template [46] . After digestion with AgeI and NotI , gfp was cloned into pCM38 ( gift form C . Morlot , IBS , Grenoble ) previously opened with the same enzymes resulting in PZn-gfp . pCM38 is a modified version of pJWV25 [47] in which an AgeI restriction site has been inserted upstream of the gfp+ gene . Then , gpsB was amplified using the primer pair XIV/XV ( Table S4 section 3 ) using pneumococcus WT chromosomal DNA . The amplified fragment was then digested by SpeI and NotI and inserted in PZn-gfp resulting in PZn-gfp-gpsB . To construct plasmids for bacterial two-hybrid , DNA fragments were amplified by PCR using specific primers pairs presented in Table S4 , section 4 . The PCR DNA fragments were then digested by Acc65I and XbaI and ligated into either pKNT25 or pUT18 vectors [26] . The nucleotide sequences of all final PCR DNA fragments were checked to ensure error-free amplification . Plasmids and primers used in this study are listed in Tables S3 and S4 , respectively . S . pneumoniae strains were constructed by transformation in R800 and are therefore isogenic . We used a two-step procedure , based on a bicistronic kan-rpsL cassette called Janus [48] to delete , or replace the genes of interest by their gfp or rfp fusion forms . This procedure avoids polar effects and allows a physiological level of expression of GFP and RFP fusions . An exhaustive description of the procedure is provided in Supplemental Materials and Methods ( Text S1 ) . The genes encoding GFP and RFP were from [46] and [15] , respectively . Recombinant plasmids overproducing GpsB , FtsZ EzrA , DivVA and inactive StkP cytoplasmic domain ( StkP-K42R ) were transformed into the BL21 ( DE3 ) E . coli strain . The transformants were grown at 37°C until the culture reached an OD600 = 0 . 4 . Expression was induced by adding IPTG to a final concentration of 0 . 5 mM and incubation was continued for 3 h . Proteins were extracted , purified on a Ni-NTA agarose column ( Qiagen ) and dialyzed overnight at 4°C as previously described [14] . The concentration of protein was determined using a Coomassie Assay Protein Dosage Reagent ( Uptima ) and aliquots were stored at −80°C . To purify GpsB from S . pneumoniae cells , we constructed a strain in which gpsB is fused to a DNA fragment encoding for 6 histidines at the chromosomal locus . We checked that cells grew as the WT cells and displayed proper cell shape . This strain was cultured in THY medium at 37°C until OD550 reached 0 . 4 . After centrifugation , the pellet was suspended in buffer A ( 50 mM Tris-HCl pH7 . 5 , 10% ( v/v ) glycerol , 200 mM NaCl , 10 mM imidazole , 0 . 3% ( w/v ) SDS ) supplied with 1 mg/L lysosyme , 6 mg/L DNase/RNase , 1× cocktail of anti-protease ( Roche ) and 0 . 1% ( v/v ) anti-phosphatase ( Sigma ) . The cells were then incubated at 4°C for 10 min and opened by sonication . The lysate was supplied with 1% ( v/v ) Triton X-100 and further incubated at 4°C for 15 min . Then , the lysate was subjected to ultracentrifugation of 14 , 000× g for 30 min . Ni-NTA agarose was equilibrated with buffer A′ ( 50 mM Tris-HCl pH7 . 5 , 10% ( v/v ) glycerol , 200 mM NaCl ) and then incubated with the ultracentrifuged supernatant . The resin was washed twice with buffer A supplied with 0 . 1% ( v/v ) Triton X-100 and then twice with buffer B ( buffer A′ containing 20 mM imidazole and 0 . 1% ( v/v ) Triton-X100 ) . Elution was carried out with buffer C ( buffer A containing 300 mM imidazole and 0 . 1% ( v/v ) Triton-X100 ) . Eluted fractions were collected and added with 0 . 02% ( w/v ) deoxycholate and 8% ( w/v ) trichloroacetic acid and shake vigorously . After centrifugation of 13 , 200× g at 4°C for 30 min , the supernatant was discarded and the pellet was resuspended in SDS-PAGE loading buffer . pH was adjusted using 1 . 5 M Tris-HCl pH8 . 8 . The resulting samples were separated by 15% SDS-PAGE after boiling for 5 min . TEM , SEM , fluorescence and immunofluorescence microscopy were carried out as previously described [14] . Cells were grown at 37°C in THY broth and analyzed when the OD reached Abs550 = 0 . 1 Polyclonal antibody specific for FtsZ [49] was used at 1/200 . Slides were visualized with a Zeiss AxioObserver Z1 microscope fitted with an Orca-R2 C10600 charge-coupled device ( CCD ) camera ( Hamamatsu ) with a 100× NA 1 . 46 objective . Images were collected with AxioVision ( Carl Zeiss ) and analyzed with ImageJ ( http://rsb . info . nih . gov/ij/ ) . For TEM , cells were examined with a Philips CM120 transmission electron microscope equipped with a Gatan Orius SC200 CCD camera . For SEM , cells were observed with a Quanta 250 FEG ( FEI ) scanning electron microscope . For PG labeling with Bodipy-FL-amino-D-alanine ( BADA ) , the procedure used was adapted from [23] , [24] . Exponentially growing cells of ( OD550 = 0 . 1 ) were incubated for 4 min at 37°C with 500 µM of BADA . Cells were then washed three times with Phosphate Buffer Saline ( PBS ) pH 7 . 4 . Then , 0 . 7 µl of the mixture was placed on slides and observed under the microscope . Time-lapse microscopy was performed as described [50] using an automated inverted epifluorescence microscope Nikon Ti-E/B equipped with the perfect focus system ( PFS , Nikon ) and a phase contrast objective ( CFI Plan Fluor DLL 100× oil NA1 . 3 ) , a Semrock filter set for GFP ( Ex : 482BP35; DM : 506; Em : 536BP40 ) , a Nikon Intensilight 130W High-Pressure Mercury Lamp , a monochrome OrcaR2 digital CCD camera ( Hamamatsu ) and an ImagEM-1K EMCCD camera ( Hamamatsu ) . Briefly , after gentle thawing of THY stock cultures , aliquots were inoculated at OD550 = 0 . 006 in C+Y medium and grown at 37°C to an OD550 of 0 . 3 . These precultures were inoculated ( 1/100 ) in C+Y medium and incubated at 37°C to an OD550 of 0 . 1 unless otherwise specified . Two microliters were directly spotted on a microscope slide containing a slab of 1 . 2% C+Y agarose . The microscope is equipped with a chamber thermostated at 30°C . Images were captured every 5 minutes and processed using Nis-Elements AR software ( Nikon ) . All fluorescence images were acquired with a minimal exposure time ( exposure time: 2 seconds; camera gain: 50; light attenuation with neutral-density filters: 25% ) to minimize bleaching and phototoxicity effects . GFP fluorescence images were false colored green and overlaid on phase contrast images . Detection of in vivo phosphorylated proteins in crude extracts of S . pneumoniae strains was performed after SDS-PAGE by immunoblotting using an anti-phosphothreonine polyclonal antibody ( Cell Signaling ) at 1/2000 as described in [14] . A goat anti-rabbit secondary antibody HRP conjugate ( Biorad ) was used at 1/5000 . Detection of StkP and GFP fusions were performed using a rabbit polyclonal antibody specific for StkP [14] and rabbit anti-GFP ( AMS Biotechnology ) . To examine GpsB in vivo phosphorylation , GpsB was analyzed by SDS-PAGE after purification ( see Protein purification ) . An in gel digest using trypsin was performed , followed by a phosphorylated peptide enrichment procedure with TiO2 beads as previously described [51] , with minor modifications: TiO2 beads ( 10 µm ) ( MZ Analysetechnik , Mainz , Germany ) were incubated with 2 , 5 dihydrobenzoic acid in 80% acetonitrile ( final concentration 30 g/L ) prior to phosphopeptide enrichment . 5 mg of TiO2 beads were added to the sample and incubated at room temperature on a rotating carousel for 30 minutes . After washing in 1 mL 30% acetonitrile/3% TFA and 80% acetonitrile/0 . 1% TFA for 10 min each , the phosphopeptides were eluted from the TiO2 spheres with 3×100 µL of 40% ammonium hydroxide solution in 60% acetonitrile , pH 10 . 5 . The sample volume was reduced in a vacuum centrifuge at room temperature and brought to a final volume of 6 µL for nano-LC-MS/MS analysis . NanoLC-MS/MS-experiments were performed on an EASY-nLCt system ( Proxeon Biosystems , ) connected to an LTQ-Orbitrap XL or Elite . For proteome analysis , peptides were applied onto a 15 cm nano-HPLC column , in-house packed with reverse-phase 3 µm C18 spheres ( Dr . Maisch , Ammerbuch , Germany ) at a flow rate of 500 nL/min in 0 . 5% acetic acid . The peptides were eluted using a segmented 90 min gradient of 5–33% of Solvent B ( 80% acetonitrile in 0 . 5% acetic acid ) at a constant flow rate of 200 nL/min . Peptide were ionized via the electrospray ion source ( ESI ) ( Proxeon Biosystems , Odense , Denmark ) . The mass spectrometer was operated in the positive ion mode with the following acquisition cycle: one initial full scan in the Orbitrap analyzer ( MS ) was followed by fragmentation through rapid collision induced dissociation ( CID ) of the 20 most intense multiply charged precursor ions in the linear ion trap analyzer ( LTQ ) . The full scan was performed range of m/z 300–2 , 000 at a resolution of 120 , 000 ( defined at m/z = 400 ) . Target values were set at 1E6 and 5E3 charges for MS or MS/MS , respectively . Sequenced precursor ions were subjected to dynamic exclusion ( set for 90 seconds ) . The LTQ Orbitrap XL was used for the detection of phosphorylation sites in the same way as above but with slight modifications: CID was performed on the 5 most intense precursor ions . Multi stage activation ( MSA ) was applied in all MS/MS events when a neutral loss event was detected on the precursor ions depending on their charge state: singly ( −97 . 97 Th ) , doubly ( −48 . 99 Th ) and triply ( −32 . 66 Th ) . The full scan was set at 60 , 000 and the lock-mass option [52] was enabled for real time recalibration of MS spectra . All RAW files were processed with the MaxQuant software version 1 . 2 . 2 . 9 [53] . N-acetylation of protein ( N term+42 . 010565 Da ) , N-pyro-glutamine ( Gln _17 . 026549 ) , oxidized methionine ( +15 . 994915 Da ) and phosphorylation of serine , threonine and tyrosine ( Ser/Thr/Tyr +79 . 966331 Da ) were searched as variable modifications . The database used to search all submitted peak lists was uniprot S . pneumoniae ATCC BAA-255 R6 . Bacterial two-hybrid experiments were performed according to the manufacturer's instructions ( Euromedex ) . The picture was taken after 40 h of growth at 30°C onto LB-agar plates containing X-gal ( 40 µg/ml ) , 0 . 5 mM IPTG and appropriated antibiotics . For analyses using surface plasmon resonance ( SPR ) , real time binding experiments were performed on a BIAcore T100 biosensor system ( GE Healthcare ) . EzrA or GpsB ( ligand ) were covalently coupled through their amino groups to the surface of a CM5 sensorchip according to the manufacturer's instructions . Increasing concentrations ( 0 . 002 , 0 . 005 , 0 . 1 , 0 . 2 , 0 . 5 and 1 µM from bottom to top ) of DivIVA , EzrA , StkP , GpsB , or FtsZ ( analyte ) were injected over the surface of the sensorchip at a flow rate of 30 µL/min in 10 mM HEPES pH 7 . 4 , 150 mM NaCl , 0 , 005% surfactant . For all experiments , aspecific binding to the surface of the sensorchip was substracted by injection of the analytes over a mocked derivatized sensorchip . The resulting sensorgrams were analyzed using BIAevaluation software ( GE Healthcare ) . KD values were calculated from the equilibrium resonance signal ( Req ) as a function of the analyte concentration . Req values were estimated by extrapolation to infinite time using plots of resonance signal as a function of the reciprocal of time . Apparent KD were then calculated by nonlinear fitting to the expression Req = RmaxC/ ( KD+C ) , where Rmax is the maximum binding capacity of the surface and C is the analyte concentration . The goodness of the fit was assessed by inspecting the χ2 values . The measurements were made in triplicate .
|
Over the last decade , bacterial genomics have revealed the presence of eukaryotic-type serine/threonine protein kinases ( STKPs ) in many bacteria . However , their role and mode of action is still elusive . Recent studies have suggested that STKPs could play an important role in regulating cell division of some bacterial species but the underlying regulatory mechanisms are largely unknown . Considering that much remains to be discovered about the mechanisms by which the cell division machinery is assembled at the cell center and how the diversity of bacterial cell shapes is achieved and maintained , studying the role of STKPs represents a promising approach to decipher the inner workings of bacterial cell division . In this article , we show that the ser/thr-kinase StkP and the two cell division paralogs GpsB and DivIVA of Streptococcus pneumoniae ( the pneumococcus ) work together to finely tune peptidoglycan synthesis and achieve proper cell shape and division . We discuss the likelihood that similar mechanisms occur in other bacteria requiring protein-kinases for the cell division process . We propose that the interplay between protein-kinases and cell-division proteins like GpsB or DivIVA is of crucial importance to satisfy the modes of cell division and the cell shape displayed by streptococci and other bacteria .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"bacteriology",
"bacterial",
"physiology",
"medical",
"microbiology",
"streptococcus",
"microbial",
"pathogens",
"biology",
"and",
"life",
"sciences",
"microbiology",
"bacterial",
"pathogens"
] |
2014
|
Interplay of the Serine/Threonine-Kinase StkP and the Paralogs DivIVA and GpsB in Pneumococcal Cell Elongation and Division
|
Dengue often presents with non-specific clinical signs , and given the current paucity of accurate , rapid diagnostic laboratory tests , identifying easily obtainable bedside markers of dengue remains a priority . Previous studies in febrile Asian children have suggested that the combination of a positive tourniquet test ( TT ) and leucopenia can distinguish dengue from other febrile illnesses , but little data exists on the usefulness of these tests in adults or in the Americas . We evaluated the diagnostic accuracy of the TT and leucopenia ( white blood cell count <5000/mm3 ) in identifying dengue as part of an acute febrile illness ( AFI ) surveillance study conducted in the Emergency Department of Saint Luke's Hospital in Ponce , Puerto Rico . From September to December 2009 , 284 patients presenting to the ED with fever for 2–7 days and no identified source were enrolled . Participants were tested for influenza , dengue , leptospirosis and enteroviruses . Thirty-three ( 12% ) patients were confirmed as having dengue; 2 had dengue co-infection with influenza and leptospirosis , respectively . An infectious etiology was determined for 141 others ( 136 influenza , 3 enterovirus , 2 urinary tract infections ) , and 110 patients had no infectious etiology identified . Fifty-two percent of laboratory-positive dengue cases had a positive TT versus 18% of patients without dengue ( P<0 . 001 ) , 87% of dengue cases compared to 28% of non-dengue cases had leucopenia ( P<0 . 001 ) . The presence of either a positive TT or leucopenia correctly identified 94% of dengue patients . The specificity and positive predictive values of these tests was significantly higher in the subset of patients without pandemic influenza A H1N1 , suggesting improved discriminatory performance of these tests in the absence of concurrent dengue and influenza outbreaks . However , even during simultaneous AFI outbreaks , the absence of leucopenia combined with a negative tourniquet test may be useful to rule out dengue .
Dengue , the disease caused by four related but distinct dengue viruses ( DENV ) , is now considered the most important arthropod-borne disease worldwide . It is transmitted through the bite of an infected mosquito , usually Aedes aegypti or Aedes albopictus [1] , and is endemic to tropical and subtropical regions . Dengue affects 50–100 million people each year . In 2007 , more than 890 , 000 dengue cases were reported in the Americas . Globally , 500 , 000 patients with dengue , mostly children , require hospitalization and at least 12 , 500 die each year [2] . There is significant year-to-year variation in the incidence of dengue , with large outbreaks typically occurring in 3- to 5-year cycles . Island-wide epidemics of dengue have been identified in Puerto Rico since 1915 [3] but similar to other countries in the Western Hemisphere , dengue epidemics have increased in frequency and severity over the past 20 years [4] . During the last large island-wide dengue outbreak in 2007 , disease appeared to be more severe than in previous large outbreaks , with more hospitalizations and cases of dengue hemorrhage fever ( DHF ) , and a high proportion of cases with hemorrhagic manifestations [5] . In 2007 , a total of 10 , 508 suspected dengue cases were reported to the passive dengue surveillance system ( PDSS ) in Puerto Rico; 53% were hospitalized , 32% had hemorrhage and less than 1% died . One lesson learned from this outbreak was that most of the laboratory-confirmed fatal cases had a delay in diagnosis and treatment initiation , highlighting the importance of timely dengue diagnosis [5] . While morbidity and mortality has been linked to delayed provision of supportive treatment [6] , case fatality rates for severe dengue infections , including DHF and dengue shock syndrome ( DSS ) , can be reduced from 10% to 20% to less than 1% with early diagnosis and proper treatment [7]–[8] . However , there is no rapid , point-of-care diagnostic test available , and the clinical diagnosis of dengue may be challenging , as it usually presents with non-specific symptoms , including fever , headache and myalgia . Therefore , dengue is difficult to distinguish from other AFIs , such as influenza , leptospirosis , and enteroviral infections . In 2007 , a simultaneous outbreak of acute gastroenteritis and dengue at the height of the dengue outbreak made recognition of dengue difficult , especially among those with warning signs for severe dengue such as persistent vomiting and abdominal pain . The concurrent outbreaks threatened to overwhelm the emergency and inpatient capacity of many Puerto Rican hospitals . In addition , simultaneous outbreaks of dengue and leptospirosis [9] , dengue and measles [10] , and dengue and influenza [11] in Puerto Rico caused similar difficulties with the clinical recognition , diagnosis , and management of dengue in the past . The tourniquet test ( TT ) has been recommended as a tool to differentiate dengue from other AFIs [12] . Studies from Thailand found that a positive TT in combination with leucopenia could distinguish dengue from other AFI in children [13]–[15] . The aim of our study was to evaluate the accuracy and usefulness of the TT and leucopenia ( white blood cell count <5 , 000/mm3 ) in identifying dengue among patients with AFI in a setting with a large number of adult cases and during concurrent dengue and influenza outbreaks .
The study design and consent process was approved by the Institutional Review Boards ( IRB ) at the Centers for Disease Control and Prevention ( CDC protocol # 5638 ) and Ponce University School of Medicine . Written consent was not required by the IRB as data was analyzed anonymously , the project was considered enhanced public health surveillance and the risk to the patient was considered minimal to none . Data was stripped of personal identifiers for analysis purposes and the database did not include any information that could link back to individual patients . Dengue is reportable by law in Puerto Rico and this project used the same reporting form used in the PDSS and no additional blood specimens were obtained in this project beyond what are normally collected for routine surveillance . Verbal consent was obtained for performing nasopharyngeal testing for influenza and for conducting the tourniquet test , a standard non-invasive dengue diagnostic test . A patient information sheet in Spanish written at an elementary-school level explaining the purpose of the study and the rationale and nature of these tests was read to and given to the patient . Verbal consent for performance of the influenza testing was documented by recording the results of the rapid influenza test on a separate sheet . The bottom half of the sheet , with explanation of test results in Spanish , was given to the patient and the top half was kept in a locked cabinet inside a secure CDC facility with access available only to study investigators . The PDSS in Puerto Rico has been in operation for more than 30 years as previously described [5] . In 2009 , the PDSS at the Saint Luke's Episcopal Hospital , in Ponce , Puerto Rico , was augmented to include surveillance for all AFI . Children and adults presenting to the emergency department ( ED ) were eligible for enrollment if they met the following case definition of AFI: documented fever of ≥38 . 0°C at presentation to the ED or history of fever that persists for 2 to 7 days without identified source of fever . Those with an identifiable source of fever , including but not limited to diagnoses of otitis media , sinusitis , pneumonia , cellulitis , impetigo , wound infection , pyelonephritis , osteomyelitis or varicella were excluded . Saint Luke's Episcopal Hospital is a 425 bed tertiary care hospital that serves more than 54 , 000 patients in their ED each year . The Saint Luke's Episcopal Hospital is one of the largest hospitals outside of the San Juan metropolitan area and serves as the primary referral hospital for a population of more than 500 , 000 people in southern Puerto Rico [16] . The surveillance period was from September 29 , 2009 through December 18 , 2009 . All patients presenting for medical care at the ED of Saint Luke's Episcopal Hospital who met the case definition above were enrolled in the surveillance system . At the time of enrollment , study personnel explained the purpose of the surveillance project and obtained verbal consent for participation . Participants were interviewed to collect demographic data included on the standard PDSS data collection form , including age , sex , place of residence , and days from symptom onset to specimen collection . Clinical data that were recorded included the duration of fever , the presence or absence of headache , eye pain , myalgia , arthralgia , rash , nausea and vomiting , and hemorrhagic manifestations , including positive TT . Laboratory data recorded on the form include white blood cell count , platelet count , and highest and lowest hematocrit values . Performance of a white blood cell count , blood and urine cultures and other laboratory tests were at the discretion of the attending physician in the course of routine patient care but were not part of the study protocol . For all patients , a blood sample and two nasopharyngeal samples were collected for diagnostic testing . A TT was performed by trained study personnel . The standard TT was performed by inflating a blood pressure ( BP ) cuff on the upper arm of the patient to a point mid way between their systolic and diastolic pressure for 5 minutes , and then counting the number of petechiae in a 2 . 5 cm2 area on the volar aspect of the forearm just distal to the antecubital fossa 2 minutes after releasing the BP cuff . The TT was considered positive if 10 or more petechiae were identified . Leucopenia was defined as a total white blood cell count of <5 , 000/mm3 , in keeping with previously published criteria from Asia and the Americas [15] , [17] . Two nasopharyngeal samples were obtained . The first sample was tested onsite by using the QuikVue Influenza A + B rapid influenza test ( Quidel Corporation , San Diego , CA ) , and the second sample placed in viral transport media and refrigerated until transport to the CDC's Dengue Branch for confirmatory influenza testing by polymerase chain reaction ( PCR ) testing . A 5–10 ml venous blood sample was collected , immediately refrigerated at 4°C , centrifuged on site , and transported on ice within 3 days to the Dengue Branch for further testing . Samples were initially tested for the presence of DENV via serotype-specific reverse-transcriptase PCR ( RT-PCR ) [18]–[19] , DENV-specific non-structural protein-1 assay ( NS-1 ) [20] , and an anti-DENV IgM enzyme-linked immunosorbent assay ( MAC-ELISA ) [21] . Samples with sufficient quantity of serum remaining were subsequently transported on ice to the Bacterial Zoonosis Branch of the CDC in Atlanta for testing for leptospirosis . Specimens were screened for IgM antibodies to leptospirosis by using the rapid dipstick ELISA ImmunoDOT kit ( GenBio , Inc . , San Diego , CA ) . Specimens with positive or borderline results with the ImmunoDOT kit were further tested by using the microscopic agglutination test ( MAT ) [22] . Patient sera were serially diluted in the MAT and mixed with a panel of 20 Leptospira reference antigens that represented 17 serogroups . Resulting agglutination titers were read by using darkfield microscopy , and the final titer was expressed as the reciprocal of the last well that agglutinates 50% of the antigen . Samples from patients with illness of <3 days duration and with sufficient sera remaining were shipped at −70°C to the Picornavirus Laboratory at CDC and tested for enteroviruses by a pan-enterovirus real-time RT-PCR by using primers and probe targeting the 5′ non-translated region [23] . In PCR-positive specimens , the enterovirus type was identified by semi-nested PCR amplification and sequencing of a portion of the region encoding the VP1 capsid protein , as previously described [24] . All personnel performing laboratory testing were unaware of the clinical condition of the patient including TT and WBC count results . Data was initially entered into a Microsoft Access database ( Access 2007; Redmond WA ) and exported to SAS version 9 . 2 ( SAS Institute , Cary , NC ) . We calculated the sensitivity , specificity , positive predictive value ( PPV ) , and negative predictive value ( NPV ) of leucopenia and a positive TT , both in isolation and in combination , to differentiate dengue from other AFI . Proportions were compared using test or Fisher's exact test as appropriate . For each test , statistical significance was considered to be a P value≤0 . 05 .
A total of 284 patients with AFI were enrolled . Thirty-one ( 11% ) patients were laboratory confirmed as having dengue , 136 ( 48% ) influenza and 3 ( 1% ) were diagnosed with enterovirus . Dual infections were confirmed in two additional patients; one patient had influenza and dengue and the other had influenza and leptospirosis . These two patients were excluded from data analysis . Urinary tract infections were found in two patients ( Escherichia coli and Staphlococcus saprophyticus ) . None of the patients had a positive blood culture . No etiology was identified among the remaining 110 enrolled patients . Serotype information was available by PCR from 20/31 laboratory-confirmed pure dengue infections , 18 were DENV-4 and 1 each were DENV-1 and DENV-2 . A TT was performed on 247 ( 88% ) patients , of whom 54 ( 22% ) had a positive result ( Table 1 ) . No patient had an adverse event from performance of the TT or was unable to tolerate the procedure . Half ( 52% ) of the patients with laboratory-confirmed dengue had a positive TT while 18% of patients with influenza and 17% of patients with other AFI had a positive TT . WBC results were available for 276 ( 98% ) patients . Dengue patients were also significantly more likely to have leucopenia ( 87% ) than influenza patients ( 44% , P<0 . 001 ) and patients with other AFI ( 12% , P<0 . 001 ) . Forty five percent of dengue patients had a positive TT and leucopenia , whereas <10% of patients with either influenza ( 7% , P<0 . 0001 ) or another AFI ( 5% , P<0 . 0001 ) had both a positive TT and leucopenia . Almost all ( 94% ) dengue patients had either a positive TT or leucopenia compared with 57% of influenza patients and 26% of patients with other AFI . When compared by age group , a positive TT was more common among laboratory-confirmed dengue patients aged ≤15 years than dengue patients >15 years old ( 67% vs . 46% ) , but this difference was not statistically significant ( P = 0 . 43 ) ( Table 2 ) . Among patients aged ≤15 years , dengue patients were more likely to have a positive TT ( 67% ) than patients with influenza ( 18% ) or other AFI ( 17% ) . However , dengue patients were not more likely to have leucopenia when compared with influenza patients in this age group ( 67% vs . 54% , respectively , P = 0 . 98 ) . Among patients >15 years old , a higher proportion of patients with dengue had leucopenia ( 96% ) when compared with influenza patients ( 33% ) and patients with other AFI ( 11% ) ( P<0 . 001 ) . The proportion of dengue patients who presented with both a positive TT and leucopenia was similar for both age groups ( 44% vs . 46% ) . A higher proportion of dengue patients were positive for both tests either alone or in combination when compared with patients with influenza and patients with other AFI regardless of the patient's age . However , the difference was only statistically significant among patients in the older age group . The TT used alone correctly identified half ( sensitivity 52% ) of the patients who had dengue ( Table 3 ) . The TT performed better in identifying patients who did not have dengue ( specificity 82% ) , and a negative TT was highly associated with the absence of disease ( NPV 92% ) . Neither the specificity nor the NPV of the TT changed significantly when influenza patients were excluded . In contrast , the presence of leucopenia alone identified most laboratory-confirmed dengue cases ( sensitivity 87% ) . However , the specificity of leucopenia was less than that of the TT ( specificity 72% vs . 82% , respectively ) , and it was highly affected by the presence of influenza patients . The combination of a negative TT and normal white cell count correctly identified most patients who did not have dengue ( 94% specificity ) , whereas having either a positive TT or leucopenia correctly identified a similar proportion of patients with dengue ( 94% sensitivity ) . The relationship between TT results and platelet count , a marker of disease severity in dengue , is shown in Table 4 . A positive TT was associated with laboratory-confirmed dengue in patients with a platelet count greater than 100 , 000 ( P = 0 . 0001 ) but not in patients with a platelet count below this level . For patients without dengue , no association was seen between the platelet level and a positive TT .
Our study sought to evaluate the usefulness of a positive TT and leucopenia alone and in combination in identifying patients with dengue among children and adults with AFI in an ED of a dengue endemic area . Previous studies examining the use of one or both of these tests have primarily been performed in Asia among children suspected to have dengue and who required hospital admission [13]–[15] , [25]–[27] . While direct comparisons with the results of Asian studies are difficult because of differences in disease epidemiology and study design , our study supports the findings of previous studies conducted in Thailand [14] , [15] , which found that the presence of either leucopenia , a positive TT , or both is helpful in distinguishing between patients with and without dengue . In our study and others [17] , [28] , [29] , leucopenia alone was found to be an especially good indicator of dengue among adults . Future studies examining the predictive value of the tourniquet test for the diagnosis of dengue in adults should evaluate a positive TT in combination with leucopenia . As described in previous studies , we found that a positive TT alone was specific but not sensitive in distinguishing dengue from other AFI [25] , [27]–[29] . This is especially true when the WHO cut-off of 20 or more petechiae per 2 . 5 cm2 is used . As in other studies , we chose to maximize detection of positive dengue cases ( or the sensitivity of the tourniquet test ) by using the cut-off of 10 or more petechiae per 2 . 5 cm2 [13] , [14] , [30] . The presence of either a positive TT or leucopenia correctly identified 94% of patients who had dengue , and the absence of a positive TT or leucopenia was highly predictive of the absence of disease with a NPV >98% . That is , less than 2% of enrolled patients with neither a positive TT nor leucopenia had dengue . Compared with the data reported from Thailand on the combined performance of the TT and leucopenia in identifying dengue patients , our results showed a lower sensitivity ( 45% vs . 74% ) and PPV ( 52% vs . 73%–83% ) but similar specificity ( 94% vs . 86% ) [14]–[15] . Our finding of lower sensitivity most likely reflects differences in study design . These previous studies primarily enrolled hospitalized patients who then received daily TTs until the day of defervescence . Our study participants had a single TT performed at initial presentation to the ED . Previous research has demonstrated that the sensitivity of the TT depends on repeated testing and the timing of the test with respect to the day of illness with sensitivity increasing as a patient nears defervescence [14] , [26] . Most ( 68% ) of our participants were enrolled within three days of symptom onset so that the sensitivity of the TT found in our study corresponds to the day −4/day −3 values found in the study from Thailand ( 52% versus 46%–56% ) [14] . We feel that using values solely from the time of initial patient contact is more useful , as it uses only information that is available to physicians at the time of initial triage . Given that the study was performed in the setting of a concomitant influenza pandemic ( CDC , unpublished data ) and we did not restrict enrollment to suspected dengue cases that required hospitalization , our study also provides data on the performance of these triage criteria for dengue during periods of high influenza transmission . The lower PPV in our study largely reflects the effect of the study being conducted among patients with AFI ( versus only those suspected of having dengue ) at the time of a major outbreak of another AFI . While the specificity and PPV of the combination of tests was lower when cases with pandemic influenza A H1N1 were included , overall performance was still good . Reanalyzing the data after excluding influenza cases resulted in a PPV for the combination of leucopenia and positive TT of 74% , similar to previously published estimates [14] , [15] . As only 12% of patients without dengue or influenza had leucopenia in our study , there was only a minor decrease in the NPV upon excluding influenza patients for the combination of leucopenia and positive TT ( 92% vs 84% ) . Our study has some important limitations . It was performed at an ED in a tertiary-level referral hospital with a large catchment area , and patients seeking care at this facility are more likely to have severe disease than patients who seek care at lower-level facilities . Thus , dengue cases in our study were likely not representative of the whole spectrum of dengue occurring in Puerto Rico . Data were missing on TT and white blood cell count results for approximately 12% and 2% percent of patients , respectively . The 35 participants who did not have a TT performed did not differ significantly from those who did have a TT performed in terms of sex or median age . No cases of dengue were diagnosed among the patients who did not have a TT performed . In most cases , only one serum specimen was available for dengue testing , leading to the possibility that some true dengue cases were not detected and misclassified as dengue negative . A comprehensive testing strategy , including a highly sensitive , single-plex anti-DENV RT-PCR , a NS-1 assay , and a MAC ELISA , was used to try to limit the amount of this misclassification bias . The effects of dengue serotype and immunological status ( primary versus secondary dengue infection ) on the incidence of a positive TT or leucopenia have not yet been sufficiently investigated . Few studies have examined the role of these variables on the performance of these tests and these studies have usually been relatively small . Even fewer studies have had an adequate number of cases with both virological confirmation and immunological data to look at both variables simultaneously and stratify serotype specific results by immunological status , an important potential confounder . A positive TT was more frequently seen in children with dengue in Nicaragua when DENV-1 was the predominant serotype compared to an era when DENV-2 was circulating widely [31] . However this difference was not evident in the sub-group of virologically confirmed cases and the study did not stratify between primary and secondary infections . No difference in white blood cell count was seen between 385 adults with DENV-2 or DENV-3 infection in Taiwan [32] , but significantly more DENV-3 patients had secondary infections making accurate comparisons between groups difficult . No differences were seen in the proportion of European travelers with a positive TT between patients with a primary or secondary immune response [33] , but the small number of patients with a secondary immune response limited the power of the study . No significant differences for TT positivity or leukocyte count in primary and secondary infections were seen among 89 hospitalized DENV-1 patients in Fiji [34] . Overall our study indicates that a combination of two rapid , widely available tests can assist clinicians in distinguishing dengue from other AFIs that have similar clinical signs and symptoms . Previous investigators have reported that the combination of the TT and leucopenia is more accurate in identifying patients with dengue than the World Health Organization's 1997 clinical case definition [15] , and our data supports the contention that few patients with dengue are likely to be missed when these criteria are used . Twenty-nine of the 31 laboratory-positive dengue patients had either a positive TT or leucopenia . The two dengue patients that were not detected by these two tests were thrombocytopenic . Our study and others [31] suggest that the TT may be useful in identifying dengue patients before a major decrease in platelet count , a group for whom dengue is often overlooked as a diagnostic possibility in Puerto Rico . Patients in our study population with AFI who had a negative TT and normal WBC appear to be at low risk of having dengue , and patients with both leucopenia and a positive TT have a high likelihood of having dengue , even in the setting of an outbreak of another clinically similar illness . Increased emphasis should be placed on determining the usefulness of the TT in combination with white blood cell count in identifying patients with dengue in Puerto Rico and elsewhere in the Americas . Further exploration of the sensitivity , specificity and predictive value of these tests by day of illness would be of particular benefit for clinical decision making . Additional larger studies should also be conducted to explore the effect of dengue serotype and immune status on the diagnostic performance of these tests .
|
In the Americas , the incidence and severity of dengue cases has increased dramatically in the past 30 years . Early diagnosis and initiation of appropriate therapy can substantially reduce dengue morbidity and mortality . However the absence of a point-of-care diagnostic test and the non-specific clinical signs and symptoms in early disease make differentiating dengue from other acute febrile illnesses challenging . Identifying dengue during an outbreak of another disease is especially difficult . The combination of a simple bedside test , the tourniquet test ( TT ) , and a readily available laboratory test , the white blood cell count , has been reported to be a useful triage tool for identifying children with dengue in Asia , but little information exists on the performance of these tests in the Americas or among adults . We evaluated the utility of these tests in the setting of a concurrent influenza epidemic in Puerto Rico in 2009 . A positive TT or leucopenia ( white blood cell count <5000 ) was present in 94% of patients with laboratory proven dengue . Patients without either of these findings rarely had dengue . Our study indicates that a combination of two rapid , widely available tests can assist clinicians in distinguishing dengue from other illnesses with similar signs and symptoms .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"test",
"evaluation",
"public",
"health",
"and",
"epidemiology",
"diagnostic",
"medicine",
"epidemiology",
"infectious",
"disease",
"epidemiology",
"global",
"health",
"neglected",
"tropical",
"diseases",
"dengue",
"fever",
"viral",
"diseases",
"dengue"
] |
2011
|
Utility of the Tourniquet Test and the White Blood Cell Count to Differentiate Dengue among Acute Febrile Illnesses in the Emergency Room
|
Chlamydia is an obligate intracellular pathogen that causes a wide range of diseases in humans . Attachment and entry are key processes in infectivity and subsequent pathogenesis of Chlamydia , yet the mechanisms governing these interactions are unknown . It was recently shown that a cell line , CHO6 , that is resistant to attachment , and thus infectivity , of multiple Chlamydia species has a defect in protein disulfide isomerase ( PDI ) N–terminal signal sequence processing . Ectopic expression of PDI in CHO6 cells led to restoration of Chlamydia attachment and infectivity; however , the mechanism leading to this recovery was not ascertained . To advance our understanding of the role of PDI in Chlamydia infection , we used RNA interference to establish that cellular PDI is essential for bacterial attachment to cells , making PDI the only host protein identified as necessary for attachment of multiple species of Chlamydia . Genetic complementation and PDI-specific inhibitors were used to determine that cell surface PDI enzymatic activity is required for bacterial entry into cells , but enzymatic function was not required for bacterial attachment . We further determined that it is a PDI-mediated reduction at the cell surface that triggers bacterial uptake . While PDI is necessary for Chlamydia attachment to cells , the bacteria do not appear to utilize plasma membrane–associated PDI as a receptor , suggesting that Chlamydia binds a cell surface protein that requires structural association with PDI . Our findings demonstrate that PDI has two essential and independent roles in the process of chlamydial infectivity: it is structurally required for chlamydial attachment , and the thiol-mediated oxido-reductive function of PDI is necessary for entry .
Fundamental to understanding of intracellular bacterial pathogenesis is knowledge of the mechanism of bacterial attachment and subsequent entry into cells . There are two main processes by which bacteria stimulate their entry into nonphagocytic cells: by bacterial contact mediated activation of a cell surface receptor ( the “zipper” mechanism ) or by injecting bacterial proteins into the cell cytosol ( the “trigger” mechanism ) [1] , [2] . Once the bacterial and host factors involved in the invasion process are identified this knowledge can be employed to devise antimicrobial strategies that target cellular infection . Blockade of this first step of bacterial infection is ideal for intracellular bacteria as these pathogens are able to avoid a number of host defenses by residing within cells . Chlamydia is an obligate intracellular bacteria that can infect a number of different eukaryotic cells . Human chlamydial infection causes a wide range of pathologies . Chlamydia is the most common bacterial sexually transmitted disease [3] , the leading cause of infectious blindness [4] , and a community acquired respiratory pathogen [5] . Chlamydia infects cells as a metabolically inactive elementary body ( EB ) and then once within cells differentiates into the metabolically active but noninfectious form known as the reticulate body ( RB ) . The EB are small ( 0 . 3-µm ) and have a rigid outer membrane consisting of a mesh of disulfide cross-linked cysteine-rich proteins [6] . This membrane structure causes the EB to be osmotically stable and thus resistant to the stresses of the extracellular environment [7] . The RB , which is much larger ( 1-µm ) , is not osmotically stable owing to a decrease in disulfide cross-linked envelope proteins . Following replication the RB condense back into EB in a process that involves the expression of EB-specific disulfide-rich proteins and oxido-reductive processing . These EB can then infect neighboring cells or new hosts . Attachment and entry into cells are key steps in chlamydial development and pathogenesis , yet the mechanism governing these interactions is still unknown . A number of bacterial ligands , including the major outer membrane protein [8] , glycosaminoglycans [9] , [10] , heat shock protein 70 [11] , and OmcB [12] , [13] have been implicated in the process . It is likely that a host proteinacious factor ( s ) is involved in Chlamydia attachment as infectivity is lost following mild trypsin treatment of cells [14] . Several host proteins including: epithelial membrane protein 2 [15] , mannose 6-phosphate receptor [16] , the estrogen receptor complex [17] , platelet-derived growth factor receptor [18] , and protein disulfide isomerase ( PDI ) [17] , [19] influence Chlamydia attachment . However , only one mammalian protein , PDI , has been demonstrated to be involved in attachment of multiple species and serovars of Chlamydia [19] . The role for PDI in Chlamydia infection was originally elucidated by proteomic analysis of CHO6 cells . CHO6 cells were generated by chemical mutagenesis of Chlamydia-susceptible CHOK1 cells and the mutant cells are resistant to attachment of C . trachomatis , C . pneumoniae , and C . psittaci [20] . CHO6 have a defect in PDI processing and express two forms of the protein . CHO6 express full length PDI , which is found in the parental cell line CHOK1 , as well as a truncated form lacking the N-terminal signal sequence that is unique to CHO6 cells [19] . This truncated form of PDI present in CHO6 is not the result of a mutation in PDI itself but is instead likely due to an unidentified processing defect in CHO6 cells [19] . It was determined that ectopic expression of full length PDI in CHO6 rescues chlamydial attachment and consequently infectivity [19] . PDI is a multi-functional protein: it catalyzes the reduction , oxidation , and isomerization of disulfide bonds , it can act as a chaperone or anti-chaperone [21] , and PDI is a subunit of collagen prolyl 4-hydroxylase and microsomal triglyceride transferase [22] . PDI consists of five domains . It contains two thioredoxin-like catalytic domains ( a and a′ ) separated by two non-catalytic domains ( b and b′ ) and has a small C-terminal domain ( c ) . The two catalytic domains contain characteristic CGHC active-site motifs , the cysteines in these sites are essential for PDI enzymatic activity [23] . An ER retention signal ( KDEL ) is located at the C-terminal of PDI , PDI also has an N-terminal signal sequence . PDI is highly enriched in the endoplasmic reticulum but is also found in the cytosol , nucleus , and on the cell surface [24] . PDI reductive function at the cell surface has been shown to be essential for entry of a number of different viruses [25]–[28] , as well as being necessary for activating diphtheria toxin [29] . It is known that PDI is involved in chlamydial infection [19] , but the precise role of PDI is not clear as PDI has a number of diverse functions in cells . In this study we evaluated the role of PDI enzymatic activity in Chlamydia infectivity . We have determined that although cellular PDI is required for both Chlamydia attachment and entry the requirement is mechanistically different in the two processes . PDI cell surface enzymatic activity was necessary for entry of bacteria into cells . In contrast Chlamydia attachment to host cells required PDI but was independent of cell surface PDI enzymatic activity .
CHO6 cells , which have a mutation that affects PDI processing [19] , are resistant to attachment of multiple species of Chlamydia [20] . It has previously been shown that chlamydial infectivity of CHO6 can be rescued by expression of recombinant PDI [19] , suggesting that PDI is involved in bacterial attachment . The mutation in CHO6 leading to differential processing of PDI and consequent lack of Chlamydia attachment is unknown . PDI is essential for cell viability , thus gene disruption approaches cannot be used to test if PDI is necessary for Chlamydia attachment or if additional mutations play a role in the phenotype of the CHO6 cell line . In contrast , siRNA has been used to transiently knockdown expression of PDI in mammalian cells [30] . siRNA-mediated downregulation of PDI was used to test whether PDI alone is required for Chlamydia infectivity . HeLa cells were transfected with PDI-targeting siRNA and PDI knockdown was assessed by immunoblot . Approximately 80–90% reduction in PDI expression was achieved ( Figure 1A ) . When PDI was depleted from HeLa cells bacteria were no longer able to attach to and thus infect cells ( Figure 1B ) . The degree of chlamydial attachment to siRNA treated cells was quantified by counting the number of bacteria attached to cells . Following PDI depletion there were only 0 . 69±1 . 04 C . psittaci and 1 . 27±1 . 39 C . trachomatis per cell , a 97 . 8% and 97 . 7% reduction in attachment respectively ( Figure 1C ) . These results demonstrate that cellular PDI is essential for attachment of Chlamydia to the surface of host cells , making PDI the only host protein identified that is required for Chlamydia attachment . To further characterize the nature of the PDI requirement in Chlamydia infectivity , we tested the hypothesis that the oxido-reductive or cysteine isomerase roles of PDI were required for restoration of Chlamydia attachment and entry to CHO6 cells . The catalytic domain of PDI is well characterized and specific amino acids involved in PDI isomerase activity have been identified [31] . The role of PDI enzymatic activity in chlamydial infectivity was analyzed by generating recombinant PDI with disabled catalytic domains . This was accomplished by converting the four key cysteine residues within the catalytic domains to serine ( PDI-4CS ) . Although PDI-4CS is no longer able to reduce , oxidize , or rearrange disulfide bonds it can still fulfill PDI chaperone , anti-chaperone , and structural roles [31] . Prior to examination of Chlamydia infectivity restoration in cells expressing PDI-4CS , cell surface PDI enzymatic activity was evaluated . Plasma membrane PDI activity can be tested by cellular sensitivity to diphtheria toxin ( DT ) as cell surface PDI enzymatic activity is required for DT-mediated killing of cells . PDI reduces interchain disulfide bonds in DT triggering chain separation that then allows for translocation of DT chain A across the membrane and subsequent toxin-induced cell death [29] . CHO6 cells have aberrant cell surface PDI activity and are resistant to DT . If native PDI is expressed in CHO6 cells toxin sensitivity is restored [19] . DT sensitivity of CHO6 expressing native PDI ( CHO6+PDI ) or CHO6 expressing enzymatically disabled PDI ( CHO6+PDI-4CS ) were tested . As previously reported , CHO6+PDI became sensitive to toxin ( Figure 2 ) , whereas CHO6 cells expressing PDI lacking enzymatic function ( CHO6+PDI-4CS ) were largely resistant to DT ( Figure 2 ) . These results demonstrated that cell surface PDI enzymatic activity is rescued in CHO6 following expression of native PDI but not in cells expressing PDI lacking enzymatic function . These data establish a model that can be used to specifically address the role of plasma membrane PDI enzymatic activity in Chlamydia infectivity . CHO6 cells were transfected with vectors expressing native PDI or PDI-4CS and both attachment and entry were separately evaluated . Despite lacking PDI enzymatic activity at the cell surface , CHO6 cells transfected with PDI-4CS showed equivalent Chlamydia attachment as cells transfected with native PDI ( Figure 3A ) . The level of attachment recovery was quantified by counting the number of bacteria associated with cells ( Figure 3B ) . These data show that while PDI is necessary for chlamydial attachment , the function of PDI in attachment is independent of the protein's enzymatic activity . Because PDI that is enzymatically nonfunctional could restore bacterial attachment to the cell , the ability of Chlamydia to establish a productive infection in cells expressing the parental or enzymatic mutant PDI protein was evaluated ( Figure 3C ) . The development of Chlamydia laden vacuoles was only observed in the parental CHOK1 and CHO6 cells expressing PDI ( Figure 3C ) . Infection rescue was quantified by determining the number of inclusion per field of view ( Figure 3D ) , no inclusion were seen in CHO6 cells or CHO6 cells expressing PDI-4CS . It was noted that despite the lack of productive infection of CHO6+PDI-4CS , many of the bacteria remained persistently attached to the cells throughout the 24 h course of the experiment ( Figure 3C ) . Enzymatically nonfunctional PDI was capable of rescuing attachment but not infectivity . This outcome could be the result of an enzymatic role for PDI in either cellular entry or in other chlamydial developmental processes soon after entry that may prevent growth . Chlamydial entry was evaluated by allowing attached Chlamydia to enter cells by shifting them to 37° C for 2 h . Entry was analyzed by comparing the number of surface bound bacteria by immunofluorescence staining of unpermeabilized cells between CHO6+PDI versus CHO6+PDI-4CS . After 2 h at 37°C , entry was restored in CHO6+PDI with 95 . 7% of bacteria internalized following the 2 h incubation ( Figure 3E and 3F ) . In contrast , there was no significant internalization of Chlamydia that had attached to CHO6+PDI-4CS ( Figure 3E and 3F ) . Although PDI , independent of its enzymatic function was sufficient to restore chlamydial attachment , enzymatically functional PDI was required for uptake of bacteria into host cells . From these results it can be concluded that PDI serves two distinct yet essential functions for Chlamydia attachment and entry . The role of PDI in attachment is not enzymatic but perhaps limited to a structural or chaperone function , whereas subsequent bacterial entry requires PDI enzymatic activity . Evaluation of Chlamydia infectivity in CHO6 cells expressing enzymatically nonfunctional PDI indicate that cell surface PDI activity is required for entry of bacteria but not for attachment . However , from those experiments it is equivocal if the PDI enzymatic activity necessary for bacterial entry is occurring at the cell surface or intracellularly . To specifically test if the required PDI enzymatic activity was occurring at the cell surface , Chlamydia attachment and entry were evaluated in CHOK1 cells in the presence of bacitracin . Bacitracin is a membrane impermeable PDI-specific inhibitor that has been used to test the role of plasma-membrane-specific function of PDI in mammalian cells [32]–[34] . Bacitracin is considered to be a PDI inhibitor because it does not inhibit thioredoxin mediated reduction [32]–[34] . The precise mechanism of bacitracin inhibition of PDI function is not known . Bacitracin has previously been used with Chlamydia . Davis et al . [17] found that addition of the inhibitor during infection led to a 16 to 36% decrease in bacterial infectivity . Given the dual role of PDI in Chlamydia infection illuminated by genetic approaches , we revisited these early findings by examining the effect of bacitracin-mediated inhibition of cell surface PDI on not only Chlamydia infection but also attachment . CHOK1 cells were inoculated in the presence of bacitracin and attachment was then evaluated . In the presence of bacitracin Chlamydia attached to cells at similar levels as untreated cells ( Figure 4A ) . This demonstrates that extracellular PDI enzymatic function is not required for chlamydial attachment and supports the results of our experiments with CHO6 expressing enzymatically nonfunctional PDI ( Figure 3A and 3B ) . The role of plasma membrane PDI function in Chlamydia entry was examined by inoculating cells and cultivating them in media containing bacitracin for 2 h . When bacitracin was present during initial infection and for the 2 h following attachment Chlamydia were unable to enter cells and EB remained persistently attached to cells ( Figure 4B ) . This confirmed that there is a requirement for plasma membrane PDI activity for chlamydial entry and supports our previous analyses . The effect of bacitracin on bacterial attachment and infection was quantified using a quantitative real-time PCR assay as described previously [35] . When bacitracin was present the level of attachment was 92 . 9% that of untreated cells , indicating that cell surface PDI enzymatic activity is not necessary for Chlamydia attachment ( Figure 4C ) . In contrast when the level of infection was quantified 24 h after bacterial attachment , infection of bacitracin treated cells was reduced by 73 . 4% as compared to untreated cells ( Figure 4C ) . Along with being a PDI-specific inhibitor , bacitracin is also a bactericidal antibiotic that targets gram-positive bacteria . The mechanism of bacitracin-mediated bacterial death is that it inhibits bacterial cell wall synthesis by inhibiting dephosphorylation of lipid pyrophosphate . Chlamydia is a gram negative-like bacteria , but it remained important to ensure that the block of bacterial entry by bacitracin was not simply due to damage to chlamydial organisms by the inhibitor treatment . We determined that treating cells with the inhibitor only prior to infection and followed by washing had no effect on Chlamydia attachment or entry ( Figure 4D ) . Normal bacterial development was also seen if the inhibitor was removed or added after the first 8 h of the infection ( Figure 4D ) . The reversibility of the inhibition demonstrates that the observed effects were not due to damage to Chlamydia or the cell by the inhibitor treatment . To ensure that the requirement for PDI enzymatic activity was not unique to CHOK1 cells attachment and infectivity analysis with bacitracin was also conducted in HeLa cells . As in CHOK1 cells Chlamydia attached to HeLa cells in the presence of bacitracin but were unable to enter ( Figure S1 ) . Using genetics we determined that PDI lacking enzymatic function was able to complement attachment but not bacterial entry and subsequent development . DT analysis demonstrated that following complementation with enzymatically nonfunctional PDI there was a defect in cell surface PDI activity . The ability of bacitracin to inhibit bacterial entry , in addition to our results following PDI expression in CHO6 , define an essential role at the cell surface for PDI enzymatic function in Chlamydia entry . Having established that cell surface PDI enzymatic activity was required for Chlamydia entry into cells we next sought to determine the molecular mechanism of that activity . PDI is able to reduce , oxidize , and rearrange disulfide bonds and all three of these activities are arrested in the presence of bacitracin . Because of the role of PDI-mediated reduction in viral entry [25]–[28] , we evaluated if it was a reductive function that was necessary for Chlamydia entry into cells . CHOK1 cells were infected with Chlamydia in the presence of bacitracin , arresting infectivity of the bacteria at the cell surface . The membrane impermeant disulfide reducing agent TCEP ( Tris ( 2-carboxyethyl ) phosphine hydrochloride ) was then added and the cells were incubated at 37° C for 2 h . The addition of TCEP was able to overcome bacitracin inhibition of Chlamydia entry , confirming that the stimulus necessary for bacterial entry into cells is a PDI-mediated reduction occurring at the cell surface ( Figure 5A ) . Following entry and TCEP removal Chlamydia were able to establish a productive infection in cells demonstrating that the bacteria were not damaged by the 2 h incubation with TCEP ( Figure 5B ) . The mechanism of bacitracin mediated inhibition of PDI enzymatic activity is not known , making it possible that TCEP was simply inhibiting the interaction between PDI and bacitracin and in that manner restoring bacterial entry . To control for this possibility the effect of TCEP on entry of Chlamydia into CHO6 cells expressing enzymatically nonfunctional PDI ( CHO6+PDI-4CS ) was analyzed . Chlamydia are able to attach to CHO6 cells expressing PDI-4CS but the bacteria are unable to enter ( Figure 3 ) . When TCEP was added the bacteria were then able to enter the CHO6 cells expressing non-enzymatically functional PDI ( Figure 5C ) . From these experiments we can conclude that it is cell surface PDI mediated reduction that is required for Chlamydia uptake into cells . The functional role of PDI in the bacterial attachment stage was next examined . The most direct hypothesis is that Chlamydia bind PDI as a receptor . PDI is not an integral membrane protein . PDI is secreted from cells and then maintained at the surface through electrostatic interactions with other cell surface proteins [24] . To test for Chlamydia binding directly to cell surface PDI we generated a PDI protein that was tethered to the plasma membrane by a C-terminal gpi anchor . A similar strategy has been used to study the role of PDI in HIV entry [28] . This plasma membrane anchored PDI ( PDI-gpi ) was expressed in CHO6 cells and resulted in high-level cell surface PDI expression that was readily detectable by PDI-specific antibody ( Figure 6A ) . Prior to analysis of bacterial attachment to CHO6 expressing PDI-gpi , it was first determined if the gpi-anchored PDI was able to interact with other cell surface proteins in the same manner as unanchored PDI . This was tested by evaluating DT sensitivity of cell expressing PDI-gpi . PDI interaction with DT bound to its cell surface receptor ( heparin-binding epidermal growth factor ) is necessary for DT mediated cell death [36] . Whereas toxin sensitivity is recovered by expression of unanchored PDI ( Figure 2 ) , the gpi-anchored PDI was unable to normally interact with other cell surface proteins and could not restore DT sensitivity to CHO6 cells ( Figure 6B ) . The high-level of PDI-gpi expression did not interfere with normal cell surface interaction , as it did not reduce CHOK1 toxin sensitivity ( Figure 6B ) . From these results we could conclude that CHO6 expressing PDI-gpi functioned as a model for testing if Chlamydia was able to directly attach to PDI independent of PDI interactions with other cell surface proteins . Chlamydia attachment to CHO6 cells expressing PDI-gpi was evaluated , no recovery of bacterial attachment was observed ( Figure 6C ) . These results suggest there is a lack of Chlamydia binding directly to PDI . The experiments with the gpi-anchored PDI are indicative of a lack of direct attachment on Chlamydia to PDI , but the possibility that the gpi-anchor was causing a structural change in PDI that led to the lack of bacterial attachment could not be ruled out . To address this possibility the effect of polyclonal PDI-specific antibody of Chlamydia attachment was analyzed . The effect of antibody to PDI on Chlamydia infection has previously been evaluated by Davis et al . [17] , they observed a temperature dependent decrease in Chlamydia infection . We further developed those findings by specifically addressing inhibition of bacterial attachment as well as entry . When PDI antibody was present prior to and during bacterial attachment no significant change in the number of bacteria attached to cells was observed ( Figure 7A ) , suggesting that Chlamydia was not directly binding PDI . Bacterial entry in the presence of PDI antibody was also evaluated . The antibody significantly reduced bacterial entry ( Figure 7B ) and led to a persistent attachment phenotype similar to what was seen in the presence of bacitracin ( Figure 4 ) . The inhibition of Chlamydia entry by PDI antibody corroborated our previous determination that cell surface PDI reductive function was required for bacterial uptake . A polyclonal PDI antibody was used in these experiments , and it is likely that this antibody would inhibit PDI enzymatic activity by steric interference . To directly assess the level of PDI enzymatic activity a turbidimetric assay of insulin disulfide reduction was performed ( Protocol S1 ) . We determined that addition of PDI antibody to the reaction significantly reduced the rate and level of insulin reduction , indicating that the antibody had an inhibitory effect of PDI enzymatic activity ( Figure S2 ) . We conclude that while cellular PDI is necessary for Chlamydia attachment to cells the bacteria does not initiate attachment through a direct interaction with PDI . Much like HIV and diphtheria toxin , Chlamydia likely binds to a cell surface protein ( s ) that is associated with PDI .
Cell surface PDI-mediated disulfide bond reduction is involved in the infectious entry of a number of viruses . Upon binding of human immunodeficiency virus ( HIV ) envelope protein to CD4 receptor and co-receptor ( CCR5 or CXCR4 ) PDI reduces disulfide bonds in HIV gp120 , exposing the gp41 fusion peptide [28] . The fusion peptide then inserts into the target cell surface triggering viral and cell membrane fusion . For Newcastle disease virus , cell surface PDI enzymatic activity is required for the conformational changes in the viral fusion protein that are necessary for cell-viral membrane fusion [27] . Similarly PDI-mediated reduction of Sindbis virus envelope is required for membrane fusion and release of the viral genome into cells [25] . We have experimentally established that the initial stages of chlamydial infectivity , attachment and entry , each require host cell surface PDI; however , the functional participation of PDI was mechanistically unique at each stage . Cell surface PDI enzymatic activity was required for Chlamydia entry . Independent of PDI enzymatic function , PDI was additionally essential for bacterial attachment to host cells . The role for protein disulfide exchange in chlamydial infection has been previously explored using inhibitors . Davis et al . [17] and Raulston et al . [11] showed that inhibition of cell surface reductive function by addition of bacitracin or dithio-bis-2-nitrobenzoic acid ( DTNB ) adversely affected infectivity of C . trachomatis serovar E . These results can now be understood in terms of the direct enzymatic role for PDI in bacterial entry . Raulston et al . [11] found no inhibition of bacterial attachment using DTNB similar to our observations with bacitracin . When Davis et al . [17] and Raulston et al . [11] evaluated the effect of bacitracin or DTNB they reported a slight decrease in infectivity , whereas we observed a near complete loss of bacterial entry and subsequent infection . These differences are likely due to the fact that PDI is constitutively trafficked to the cell surface [24] , making bacitracin inhibition functionally reversible . Both Davis et al . [17] and Raulston et al . [11] removed the inhibitor after initial bacterial attachment allowing newly exported PDI to stimulate uptake . In our experimental design , the inhibitor was present throughout the course of the infection and if the inhibitor was removed following attachment , our results were similar to those of Davis et al . [17] and Raulston et al . [11] . Our data also significantly expand the previous observations as this now extends the requirement for disulfide exchange in the process of bacterial infection from one strain to multiple species of Chlamydia . A requirement for PDI in chlamydial attachment was implicated by complementation of the PDI gene in attachment-deficient mutant CHO6 cells [19] . We anticipated that the oxido-reductive enzymatic activity of PDI would be required for bacterial attachment , similar to what is characterized for viral attachment [25] , [27] , [28] . It was a surprise that chlamydial attachment to CHO6 cells was rescued by complementation with the enzymatically inactive PDI-4CS . Further evaluation using PDI inhibitors confirmed that Chlamydia attachment is independent of cell surface PDI enzymatic activity . Testing of gpi-anchored PDI tethered to the cell surface suggested that chlamydiae do not directly bind PDI as the sole target for attachment and this implicates an interaction with other proteins that require PDI . Consistent with characterized functions of PDI , it may be PDI's function as a chaperone or structural component of a host protein or protein complex [37] that is required for chlamydial attachment . PDI chaperone activity is independent of the protein's two catalytic domains ( a , a′ ) and PDI functions as a chaperone outside of the ER [38] , [39] . PDI could serve to stabilize a host receptor protein in the correct orientation or context for bacterial binding . There are also examples of PDI functioning as a structural subunit of proteins . PDI is the β-subunit of tetrameric enzyme collagen prolyl 4-hydroxylase , and PDI also makes up half of the heterodimeric protein complex microsomal triglyceride transfer protein [40] , [41] . The receptor that Chlamydia binds may be a multiprotein complex that includes PDI . In addition to a requirement of PDI for chlamydial attachment , it was shown that PDI enzymatic activity was necessary and sufficient to stimulate chlamydial entry into the host cell . Given the ability of simple chemical reduction to replace the enzymatic function of PDI for cell-adherent chlamydiae , it appears that reduction rather than disulfide exchange is minimally required for chlamydial entry . It is not known whether PDI functions to reduce a host or bacterial component to initiate entry . It has been previously shown that reduction of C . trachomatis L2 EB prior to infection leads to a decrease in inclusion forming units [42] . This experiment illustrates the problem of differentiating the effect of reduction on the infectious process . One can pre-reduce host cells or bacteria prior to infection , but due to the rapid rate of reoxidation following the reducing agent removal step and unknown detrimental effects , especially following pan-blocking of disulfides by chemical agents , make interpretation of outcomes confounded and uncertain . One can speculate that following initial bacterial attachment , PDI enzymatic activity mediates the establishment of a functional contact between the chlamydial organism and the host cell leading to bacterial uptake . It is known that plasma membrane PDI interacts with a number of host surface proteins . For example , PDI modifies the adhesion receptors integrin α2β1 and L-selectin resulting in receptor activation and ligand binding [33] . As well as modifying cellular surface proteins , PDI binds and modifies various viral proteins and toxins [25]–[29] . Thus , there is biological precedent supporting PDI modification of either host or microbial factors required for pathogenesis . Multiple cell surface proteins have been previously implicated in Chlamydia infectivity , these include epithelial membrane protein 2 [15] , mannose 6-phosphate receptor [16] , estrogen receptor complex [17] , and platelet-derived growth factor receptor [18] . Unlike PDI , these proteins appear to be involved in the infectivity of only a subset of chlamydial species or biovars . PDI interacts with a broad array of host proteins within the endoplasmic reticulum and at the cell surface allowing for the possibility that PDI may be an underlying requirement for several independent Chlamydia-host cell receptor interactions . PDI is involved in protein folding in the endoplasmic reticulum and PDI expression can be correlated to the level of secretion of a number of proteins [43]–[45] . We established using PDI-specific chemical inhibitors , anti-PDI antibodies , and chemical reduction of bound organisms that the function of PDI in chamydial infectivity requires surface accessible PDI . Proteomic analysis using 2-dimensional gel electrophoresis of biotin-labeled surface proteins failed to show any detectible and consistent difference between CHO6 and CHOK1 cell lines other than for PDI ( 19 and data not shown ) . This suggests that the lack of bacterial attachment to CHO6 cells is not due to a general defect in plasma membrane protein composition . If the enzymatic role of PDI is targeted to a host protein it seems likely that this occurs at the cell surface and not in a secretion pathway . Alternative to acting on a host protein , PDI could be targeting the chlamydial organism . The highly disulfide cross-linked structure of the chlamydial EB surface proteins that are only present in the infectious form of the chlamydial organism [7] , suggest that these could be the target for PDI . It has been shown that reduction of the EB surface is necessary for surface display of bacterial Hsp70 [11] and it is possible that PDI activates changes in chlamydial surface proteins that are required to initiate cellular entry . The chlamydial protein Tarp is translocated into host cells following bacterial attachment [46] . This translocation occurs via type III secretion and triggers actin recruitment to the site of bacterial attachment [46] . Tarp is localized within EB and is not surface exposed until the commencement of bacterial entry [46] . PDI-mediated reduction of EB surface proteins may be essential to activate Chlamydia type III secretion into the cell membrane and subsequent Tarp translocation . The type III secretion system needle protein of C . trachomatis , CdsF , was recently identified [47] . CdsF is conserved among the chlamydial species , and it is one of the few bacterial needle proteins that contains cysteine residues . One can speculate that PDI-mediated reduction of the EB surface is involved in CdsF function . While the full intricacies of the Chlamydia-host cell interaction remain enigmatic these findings illuminate important new details of the molecular mechanisms involved . Chlamydia attachment and entry into cells are separable processes that both have a unique requirement for PDI . Determination of the target of PDI enzymatic activity that leads to chlamydial entry may be expected to provide targets for generation of new anti-Chlamydia therapies and illuminate fundamental cell biological processes exploited by Chlamydia to mediate pathogen attachment and infectivity .
CHOK1 and CHO6 cells were maintained in RPMI 1640 ( Invitrogen , Carlsbad , CA ) supplemented with 10% heat-inactivated fetal bovine serum ( FBS ) ( Hyclone , Logan , UT ) , 2 mM glutamine ( Invitrogen ) , and 1 mM HEPES ( Invitrogen ) . HeLa and L929 cells were maintained in RPMI 1640 supplemented with 10% FBS . Cells were grown at 37°C in an atmosphere containing 5% CO2 . C . trachomatis L2/434/Bu EB and C . psittaci PF6 BC were purified from L929 cells on 30% and 30–44% discontinuous Renografin gradients ( E . R . Squibb and Sons , Cranbury , NJ ) [48] and stored at −80°C until use . On the night prior to analysis 1×105 cells were plated on 12 mm glass coverslips in 24 well plates . Cells were washed with Hanks buffered saline solution ( HBSS ) ( Invitrogen ) and then incubated with C . trachomatis L2/434/Bu or C . psittaci PF6 BC in RPMI 1640 with 10% FBS for 1 h at 24°C [19] , [49]–[52] . Following infection cells were washed 4 times with HBSS and methanol fixed for attachment analysis . For entry and infectivity analysis fresh cell culture media was added and cells were incubated 2 h for entry and 24 h for infectivity at 37°C in an atmosphere containing 5% CO2 . For entry analysis cells were fixed with methanol or with 4% paraformaldehyde , for infectivity assessment cells were fixed with methanol . Following fixation coverslips were incubated 15 min in blocking solution ( HBSS+2 . 5% bovine serum albumin ( BSA ) ) ( Fisher , Fair Lawn , NJ ) . Blocking solution was removed and coverslips were incubated 1 h with mouse anti-Chlamydia MOMP antibody for C . trachomatis or anti-Chlamydia LPS antibody ( Santa Cruz Biotechnology , Santa Cruz , CA ) for C . psittaci and then washed with HBSS . The wash was followed with a 30 min incubation with goat anti-mouse AlexaFluor 488 ( Invitrogen ) and Evans Blue . Coverslips were mounted in VectaShield Hard Set mounting media ( Fisher ) and evaluated on a Nikon Eclipse E800 microscope . When used bacitracin ( Sigma , St . Louis , MO ) was added 20 min prior to infection at 3 mM , infection was performed in media containing 3 mM bacitracin and following infection cells were maintained in culture media with 3 mM bacitracin . When used 50 mM TCEP ( Pierce , Rockford , IL ) in cell culture media was added following bacterial attachment . For attachment and entry inhibition analysis cells were incubated for 20 min with polyclonal rabbit anti-PDI antibody ( Stressgen Victoria , BC ) or goat anti-bovine IgG antibody ( Pierce , Rockford , IL ) , the bacterial inoculum was then added to the antibody containing media and cell were incubated for 1 h at 24°C . When Stressgen rabbit anti-PDI antibody was used for PDI visualization , PDI staining was performed similarly to Chlamydia staining and goat anti-rabbit AlexaFluor 594 ( Invitrogen ) was used as a secondary . For quantification the number of cell-associated apple-green fluorescing particles of size and shape consistent with 300 nM organisms were counted in two planes in 8 separate fields of view containing at least 10 cells . Fluorescent particles that appeared to be Chlamydia but were larger due to aggregation were enumerated separately when separate organisms could be discerned or when distinction was not possible were counted as one to provide a conservative estimate of bound organisms per cell . The gene encoding protein disulfide isomerase was cloned from CHOK1 cDNA and inserted into the expression vector pBICEP-CMV-1 ( Sigma ) yielding PDI with a N-terminal FLAG tag . PDI protein lacking enzymatic function ( PDI-4CS ) was generated by mutating 4 essential cysteine residues to serine with the Stratagene ( La Jolla , CA ) QuikChange Kit and primers Cys-55 , 58-Ser F ( 5′ - TGC CCC GTG GTC TGG CCA CTC CAA AGC TCT GG - 3′ ) , Cys-55 , 58-Ser R ( 5′ - CCA GAG CTT TGG AGT GGC CAG ACC ACG GGG CA - 3′ ) , Cys-399 , 402-Ser F ( 5′ - TAT GCC CCC TGG TCT GGC CAC TCC AAG CAG CT - 3′ ) , and Cys-399 , 402-Ser R ( 5′ - AGC TGC TTG GAG TGG CCA GAC CAG GGG GCA TA - 3′ ) . The gpi-anchor for PDI was amplified from human folate receptor 1 using primers GPI F ( 5′ - ATT GCC CGG GGC TGC AGC CAT GAG TGG - 3′ ) and GPI R ( 5′ – CAA TGT CGA CTC AGC TGA GCA GCC ACA GCA – 3′ ) . The gpi-anchor was than ligated to PDI lacking a stop codon and the PDI-gpi construct was ligated into the pBICEP-CMV-1 vector . PDI vectors were transfected into CHO cells with Effectene ( Qiagen , Valencia , CA ) following manufacturer's instructions . Chlamydia attachment and entry were evaluated 48 h after transfection . PDI was silenced in HeLa cells with PDI-specific siRNA ( 5′- GAC CTC CCC TTC AAA GTT GTT –3′ ) ( Dharmacon , Layafette , CO ) [30] . Dharmacon siControl non-targeting siRNA #1 was used at as a negative control . Cells were transfected as described in Hybiske and Stephens , 2007 with slight modification [35] . Cells were first transfected in 6 well plates with 10 µl of 40 µM siRNA duplexes with 5 µl of Olgiofectamine ( Invitrogen ) in OptiMem ( Invitrogen ) . 24 h after transfection cells were transfected again . 90 h after the first transfection cells were replated on fibronectin ( Sigma ) coated coverslips in 24 well plates , and attachment , infectivity , or protein expression was evaluated 6 h after replating . 24 h prior to toxin treatment cells were plated in 96-well plates at 1×104 cells per well . Cells were washed once with HBSS and than incubated for 4 h at 37°C with diphtheria toxin ( Biomol , Plymouth Meeting , PA ) ( 0 or 100 ng ml−1 ) in HBSS containing 15% FBS . Following incubation cells were washed three times with HBSS and were than maintain in RPMI 1640 without phenol red supplemented with 10% FBS , 2 mM glutamine , and 1 mM Hepes at 37°C in an atmosphere containing 5% CO2 . 72 h after toxin treatment cell viability was determined by Promega ( Madison , WI ) CellTiter 96 Aqueous One Cell Proliferation Assay as directed by the manufacturer's instructions . Cell viability is linearly proportional to absorbance . The number of Chlamydia per cell was determined using a quantitative PCR based method previously described by Hybiske and Stephens , [35] . Genomic DNA from CHOK1 cells infected with C . trachomatis was isolated using the High-Pure PCR template preparation kit ( Roche , Indianapolis , IN ) . Purified genomic DNA was used as a template in quantitative PCR to determine the relative levels of chlamydial ( 16S ) and CHOK1 ( ß-globin ) genomic equivalents . The NCBI Entrez ( http://www . ncbi . nlm . nih . gov/sites/entrez ? db=protein ) accession number for genes discussed in this paper are Homo sapiens PDI ( NP_006840 ) and Cricetulus griseus PDI ( AAM_00284 ) .
|
Chlamydia is a large burden on global health . It is the most common cause of infectious blindness , and the CDC ( Centers for Disease Control and Prevention ) estimates that in the United States alone there are more than 2 million people with sexually transmitted Chlamydia infections . Chlamydia is an obligate intracellular bacteria; thus , attachment and subsequent invasion of cells are key steps in Chlamydia pathogenesis . While strides have been made in understanding the molecular mechanism of Chlamydia infection , fundamental aspects of this process still remain elusive . We have identified a host protein , protein disulfide isomerase ( PDI ) , that is essential for Chlamydia attachment as well as for entry into cells . Cell-surface PDI-mediated disulfide reduction is required for Chlamydia entry into cells , whereas bacterial attachment is independent of PDI enzymatic activity . Although PDI is necessary for Chlamydia attachment , the bacteria apparently does not bind directly to cell-associated PDI , suggesting that Chlamydia attaches to a host protein ( s ) associated with PDI . This study advances our understanding of Chlamydia pathogenesis by the characterization of a host factor essential for independent stages of bacterial attachment and entry .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases/bacterial",
"infections",
"microbiology/microbial",
"growth",
"and",
"development"
] |
2009
|
Attachment and Entry of Chlamydia Have Distinct Requirements for Host Protein Disulfide Isomerase
|
Enteric infections are common where public health infrastructure is lacking . This study assesses risk factors for a range of enteric infections among children living in low-income , unplanned communities of urban Maputo , Mozambique . We conducted a cross-sectional survey in 17 neighborhoods of Maputo to assess the prevalence of reported diarrheal illness and laboratory-confirmed enteric infections in children . We collected stool from children aged 1–48 months , independent of reported symptoms , for molecular detection of 15 common enteric pathogens by multiplex RT-PCR . We also collected survey and observational data related to water , sanitation , and hygiene ( WASH ) characteristics; other environmental factors; and social , economic , and demographic covariates . We analyzed stool from 759 children living in 425 household clusters ( compounds ) representing a range of environmental conditions . We detected ≥1 enteric pathogens in stool from most children ( 86% , 95% confidence interval ( CI ) : 84–89% ) though diarrheal symptoms were only reported for 16% ( 95% CI: 13–19% ) of children with enteric infections and 13% ( 95% CI: 11–15% ) of all children . Prevalence of any enteric infection was positively associated with age and ranged from 71% ( 95% CI: 64–77% ) in children 1–11 months to 96% ( 95% CI: 93–98% ) in children 24–48 months . We found poor sanitary conditions , such as presence of feces or soiled diapers around the compound , to be associated with higher risk of protozoan infections . Certain latrine features , including drop-hole covers and latrine walls , and presence of a water tap on the compound grounds were associated with a lower risk of bacterial and protozoan infections . Any breastfeeding was also associated with reduced risk of infection . We found a high prevalence of enteric infections , primarily among children without diarrhea , and weak associations between bacterial and protozoan infections and environmental risk factors including WASH . Findings suggest that environmental health interventions to limit infections would need to be transformative given the high prevalence of enteric pathogen shedding and poor sanitary conditions observed . ClinicalTrials . gov NCT02362932
Diarrheal illness is estimated to cause approximately 1 . 7 million deaths annually and result in over 74 million disability-adjusted life years lost [1] , primarily among children in low-and middle-income countries where fecal contamination of the living environment is common . Diarrheal diseases are mostly caused by enteric pathogens , including bacteria , viruses , and protozoa , shed in human and animal feces . These pathogens can be shed in high numbers by both symptomatic and asymptomatic individuals [2] . Although the immediate and longer term health and productivity effects for asymptomatic individuals are unclear [3] , persistent asymptomatic infections are associated with environmental enteric dysfunction [4–6] and other conditions , including undernutrition , poor linear growth [7–13] , reduced immunogenicity of oral vaccines [14 , 15] , and cognitive deficits [16–18] . Enteric pathogens are transmitted via several fecal-oral pathways historically defined by the F-diagram [19] . Consumption of contaminated food and water and interaction with fecally contaminated environments have been implicated as dominant transmission pathways for bacterial and protozoan enteric pathogens [20 , 21] . While enteric viruses can be transmitted via similar routes , it is posited that person-to-person transmission is also important [22] . Improvements in WASH conditions can reduce risk of diarrheal disease by interrupting transmission pathways . A recent meta-analysis observed reductions in diarrheal disease risk by an average of 67% , 25% , and 30% for water , sanitation , and hygiene interventions , respectively [23] . Sanitation interventions may be more likely to interrupt transmission of protozoa , bacteria , and helminths which are primarily spread via indirect , environmentally mediated pathways than viruses which are often spread via person-to-person transmission [24] . Densely populated , urban , unplanned communities with inadequate sanitary infrastructure represent high-risk settings for exposure to enteric pathogens , though the great majority of sanitation-related exposure and health outcome research has been focused on rural communities where sanitation coverage is lowest and open defecation is common . In the context of the Maputo Sanitation ( MapSan ) trial [25] ( ClinicalTrials . gov Identifier: NCT02362932 ) , we conducted a baseline , cross-sectional survey of compounds ( defined as multi-household clusters with shared outdoor space ) served by shared latrines . The aim of our study was to estimate prevalence of selected enteric pathogens in stool samples of enrolled children from this cohort , and to identify WASH and other risk factors for enteric infections .
The head of the compound provided verbal assent for study activities before enrollment of any children within the compound . As children were ≤4 years old at the time of visitation , field enumerators obtained written informed consent from each child’s parent or guardian before enrollment . The study protocol was approved by the Comité Nacional de Bioética para a Saúde ( CNBS ) , Ministério da Saúde ( 333/CNBS/14 ) , the Ethics Committee of the London School of Tropical Medicine and Hygiene ( reference # 8345 ) , and the Institutional Review Board of the Georgia Institute of Technology ( protocol # H15160 ) . The associated MapSan trial has been registered at ClinicalTrials . gov ( NCT02362932 ) . This cross-sectional study measures enteric infections and key socio-demographic and WASH-related risk factors among children in low-income neighborhoods of Maputo . We defined four outcomes , based on analysis of stool for 15 common enteric pathogens , for our risk factor assessment: ( 1 ) detection of any enteric infection , ( 2 ) detection of any bacterial infections , ( 3 ) detection of any protozoan infections , and ( 4 ) detection of any viral infections . We also measured caregiver-reported diarrhea with 7-day recall [26] as a secondary outcome . We defined diarrhea as ≥3 loose or liquid stools in a 24-hour period or any stool with blood [27] . The study sites are located in densely populated , low-income , unplanned neighborhoods of Maputo , Mozambique . Poor sanitary conditions , inadequate infrastructure , environmental conditions including seasonal flooding , and increasingly high population density in these areas has led to a high burden of enteric disease and child mortality [28 , 29] . In 2015 , an estimated 53% of the urban population in Mozambique ( ∼4 . 5 million people ) lacked access to basic ‘improved’ sanitation facilities , as defined by the UNICEF/WHO Joint Monitoring Program [30] . In Maputo , approximately 89% of households use onsite waste disposal ( 10% have access to sewerage; an estimated 1% practice open defecation ) , and only 26% of fecal waste is safely managed [31] . An estimated 8% of urban sanitation in Mozambique is shared , often among the poorest households in informal neighborhoods [32] . All households in the MapSan trial used shared sanitation facilities that were in poor condition at the time of enrollment . Field teams enrolled children and collected baseline data concurrently between February 2015 and February 2016 . We enrolled all children who met the following eligibility criteria: ( 1 ) the child’s parent or guardian provided written informed consent , ( 2 ) the child was 1–48 months of age at the time of enrollment , and ( 3 ) the child resided in compounds meeting certain inclusion criteria . Compounds were eligible for enrollment if they were located within a predefined geographic area , were in close proximity to a legal piped water supply , had a minimum number of households ( 2 ) , and residents shared sanitation in poor condition and had stated demand for improved sanitation . The larger MapSan trial involved additional criteria to select compounds for intervention and details are presented in the supplementary information ( S1 Supporting information ) . Our enrollment period overlapped with the September 2015 rollout of the rotavirus A vaccination program in Mozambique . Children six weeks or younger at the time of rollout and children born after rollout began were eligible for immunization; some children enrolled in our study after September 2015 may have received the vaccination . Following enrollment , field teams collected data on socio-demographics and WASH-related risk factors using questionnaires and direct observation . Enumerators administered three levels of surveys in each compound with an enrolled child: compound-level , household-level , and child-level . For compound-level surveys , the head of compound or the head of compound’s spouse was the target respondent . For household- and child-level surveys , the child’s mother was the target respondent , though another parent or guardian was eligible to complete the questionnaire . All questionnaires were communicated in either Portuguese or the local language , Changana , as requested by the respondent . Surveys included socioeconomic and demographic questions such as child age and sex , household assets , caregiver’s education level , and breastfeeding practices . We calculated household wealth using an asset-based wealth index developed for Mozambique [33] . At each level , surveys included direct observations and questions about risk factors of enteric infection , including characteristics of household and compound level water and sanitation , sanitary condition of living spaces , presence of animals within the compound grounds , environmental conditions including flooding patterns , and measures of population density and crowding . We created a composite ‘latrine improvement score’ ranging from 0–4 with one point awarded for the presence of each of the following latrine features: permanent superstructure , tile or masonry slab , drop-hole cover , and ventilation pipe . Similarly , we created a “compound sanitary score” ranging from 0–3 with higher scores indicating poorer sanitary conditions . One point was awarded for each of the following potential risk factors: ( 1 ) compound floods during rainy season , ( 2 ) leaking or standing wastewater observed by latrine , and ( 3 ) feces or soiled diapers observed around compound grounds . Compound-specific population density was defined as the number of people who live in a compound divided by the area of that compound . We measured the area of the compound using high resolution , orthorectified and geolocated satellite imagery . Enumerators equipped with GPS enabled tablets would work with compound residents to identify landmarks and define the shape of a compound on the satellite imagery . We calculated compound area from the shapes and divided the number of compound residents by the calculated compound area to obtain our measure of compound-specific population density . We used rainfall data from the National Oceanic and Atmospheric Administration’s National Centers for Environmental Information ( https://www . ncdc . noaa . gov/cdo-web/datatools/findstation ) to calculate cumulative rainfall during the 30 days before data collection . We provided stool collection supplies , including diapers , plastic potties ( for older children no longer wearing diapers ) , and pre-labeled sterile sample bags to the caregiver of each enrolled child . Samples were collected , irrespective of reported symptoms , the following day . If a specimen was not immediately available , caregivers alerted the field team by phone when available . Following collection , samples were stored on cold packs , and transported to the medical parasitology laboratory at the Mozambican Ministry of Health ( MISAU/INS ) within six hours of collection for storage at -80°C . If a child produced a liquid stool , lab technicians stored a piece of the soaked diaper material ( “diaper samples” ) at -80°C upon receipt . Stool samples were shipped on dry ice with temperature probes to the Georgia Institute of Technology where they were stored at -80°C until analysis . We used the Luminex MagPix xTAG Gastrointestinal Pathogen Panel ( GPP , Luminex Corp , Austin , TX ) to analyze stool samples for the presence of 15 enteric pathogens: Campylobacter; Clostridium difficile , Toxin A/B; Enterotoxigenic Escherichia coli ( ETEC ) LT/ST; Shiga-like toxin producing E . coli ( STEC ) stx1/stx2; E . coli O157 , a serotype of STEC; Salmonella; Shigella; Vibrio cholerae; Yersinia enterocolitica; adenovirus 40/41; norovirus GI/GII; rotavirus A; Giardia; Cryptosporidium; and Entamoeba histolytica . The GPP is a stool-based multiplex RT-PCR assay that has been extensively tested for direct detection of enteric infections in a range of countries [34–43] . Per GPP protocol , we pretreated bulk stool samples with 1 mL of ASL stool lysis buffer ( Qiagen , Hilden , Germany ) and performed nucleic acid extraction for DNA and RNA using the QIAcube HT platform and the QIAamp 96 Virus QIAcube HT Kit ( Qiagen , Hilden , Germany ) . We eluted diaper samples in 2 . 5 mL of ASL stool lysis buffer . A sterile 10-mL syringe was used to facilitate elution via agitation by taking in and expelling the buffer 5 times . We used 1 mL of the final eluate in the pretreatment step and then proceeded with extraction as previously described . Extracts were stored at 4°C and analyzed by GPP within 24 hours of extraction . Sample size for the present study is based on enrollment in the larger MapSan trial . Sample size calculations for the larger MapSan trial have been described previously [25] . To minimize potential bias , we specified the statistical model and variables of interest before beginning the analyses . Details for individual variables used in these analyses—including definitions , coding schemes and proportions of missing values—are available in the supporting information ( S1 Table ) . We calculated unadjusted and adjusted risk ratios ( RRs ) and 95% confidence intervals for outcome variables and potential risk factors using generalized estimating equations ( GEEs ) to fit Poisson regression models with robust standard errors [44] . We used GEEs to account for clustering at the compound level . Outcome variables , including any infection and infection with bacterial , protozoan , or viral pathogens , were defined to identify differences in exposure risks from pathogen groups with different dominant routes of transmission ( e . g . person to person versus environment to person ) . All multivariable models were adjusted for a set of five variables determined a priori as contextually important covariates . These variables included child age and sex , breastfeeding practices , caregiver’s education level , and an index of household wealth . We also calculated RRs and aRRs for enteric infections using child age ( stratified by age group: 1–11 , 12–23 , and 24–48 months ) , sex , breastfeeding practices , and caregiver’s education as the predictors of interest . We ran separate multivariable models for each combination of risk factor and outcome and assessed multicollinearity of multivariable models using the variance inflation factor . We assessed crude and adjusted associations between specific enteric pathogens and diarrheal symptoms as described for the main risk factor analysis . Our primary analysis focused on complete observations . The proportion of incomplete observations per variable are denoted in supporting information ( S1 Table ) . In parallel with the complete case analysis , we ran all univariable and multivariable models on completed data following multiple imputation ( MI ) of missing values [45–49] . Details of the MI process are presented in supporting information ( S2 Supporting information ) . Briefly , we performed MI using chained equations ( also known as fully conditional specification ) to handle missing data [47 , 50] . MI models were congenial with previously discussed analysis models and included a fixed effect to account for clustering at the compound level . Auxiliary variables were included in the MI model if they were a priori defined as related to either an outcome or predictor , if they were correlated with observed values of an outcome or predictor ( r≥0 . 2 ) , or if they were correlated with missingness of any outcome or predictor variable ( r≥0 . 2 ) [46] . All statistical analyses were performed with Stata version 14 . 1 ( StataCorp , College Station , TX ) .
Field workers enrolled 519 of the 601 compounds approached regarding participation in the MapSan study . Eighty-two ( 15 . 8% ) compounds were ineligible for enrollment because they did not have a child <48 months old at the time of visitation . From those 519 compounds , workers enrolled 993 children in 815 households . Field teams administered child-level surveys for 980 of the 993 ( 99% ) enrolled children and collected stool samples from 759 ( 76% ) ( S1 Fig: flow diagram of enrollment and data collection activities ) . The average age of enrolled children was 23 months ( Table 1 ) . Approximately 27% ( 258/944 ) were <12 months old , while an equal percentage ( 28% , 266/945 ) were 12–23 months old , and the remainder ( 45% , 421/944 ) were 24–48 months old . Breastfeeding was very common among children 1–11 months old ( 87% , 224/258 ) , though 31% ( 82/266 ) of children 12–24 months were also breastfed . A little over half of child caregivers had completed primary school ( 527/980 ) . About 17% ( 163/975 ) of households met an a priori definition of crowding ( >3 people per room of living space ) . Almost all study children lived in a household that had access to a latrine in the compound ( 98% , 956/973 ) and most had access to latrines ( 61% , 576/950 ) shared by 3–5 households ( median = 4 ) . About half of children had latrines with drop-hole covers ( 57% , 557/974 ) , 37% ( 361/971 ) had a masonry or ceramic slab or pedestal , while only 31% ( 305/974 ) had a formal superstructure ( made of bricks or cement blocks ) , and 14% ( 138/975 ) had a vent pipe . Sanitary conditions of compounds were poor: 62% ( 606/974 ) of study children lived in compounds with wastewater leaking from in or around a latrine and 47% ( 455/974 ) lived in compounds where feces or soiled diapers were visible around the grounds . Disposal of child feces into a latrine was common for children 24–48 months old ( 57% , 238/421 ) . Feces of children between the ages of 1–23 months , most of whom wore diapers , was less frequently disposed of in a latrine ( 6 . 4% , 34/528 ) . Most children lived in study compounds with animals ( 65% , 645/993 ) , with cats ( 55% , 550/993 ) most commonly observed . All study households used piped water as their primary drinking water source and 78% ( 757/976 ) of children lived in households with access to a drinking water tap on the compound grounds . One or more pathogens were identified in stool samples from 655 ( ~86% ) of the 759 children from whom a sample was collected; most ( 59% , 445/759 ) had coinfections ( Table 2 ) . Stool samples from 66 ( 8 . 7% ) children yielded four or more enteric pathogens . The prevalence of coinfection ( ≥2 infections ) increased with age from 33% ( 69/208 ) in the youngest age group to 73% ( 214/293 ) in the oldest . Most children ( 76% , 579/759 ) had a bacterial infection , about half ( 53% , 402/759 ) had a protozoan infection , and only 14% ( 107/759 ) of children had a viral infection . Giardia , Shigella , ETEC , Salmonella , and norovirus were the most frequently detected pathogens among all children , though prevalence varied with age . Prevalence of any infection , and of bacterial and protozoan infections by themselves , increased with age and were largely driven by the most common bacterial and protozoan infections: Shigella and Giardia . Prevalence of Shigella infection increased from 9% ( 19/208 ) in children 1–11 months old to 65% ( 189/293 ) of children aged 24–48 months . Giardia showed a similar pattern with prevalence increasing from 14% ( 29/208 ) among 1–11 month-olds to 75% ( 219/293 ) prevalence in 24–48 month-olds . Prevalence of viral infections , largely driven by norovirus GI/GII , was highest among the youngest children ( 17% , 36/208 ) and lowest among the oldest children ( 11% , 33/293 ) . Prevalence of rotavirus was low among all age groups ( 1–2% ) . Prevalence of enteric infections was similar among boys and girls with the exception of viral infections which tended to be more frequent in girls ( 17% , 64/370 ) than boys ( 11% , 41/370 ) . Only 13% ( 126/980 ) of children were reported to have had diarrhea in the previous week . Reported diarrhea was higher among boys ( 16% , 74/464 ) than girls ( 10% , 50/498 ) and peaked in children aged 12–23 months ( 20% , 52/266 ) . Norovirus was the only infection associated with higher risk of reported diarrhea ( adjusted RR ( aRR ) : 1 . 76 , 95% CI: 1 . 03–3 . 02 adjusted for child age and sex , caregiver education , breastfeeding practices , and household wealth ( S2 Table ) , aRR 1 . 75 , 95% CI: 1 . 00–3 . 1 when also adjusted for presence of all other measured pathogens ) . Risk factors for enteric infection were assessed using generalized estimating equations in unadjusted models and models adjusted for age and sex of child , socioeconomic status , caregiver’s education , and any breastfeeding . Among complete cases ( Table 3 ) , presence of a latrine superstructure was associated with 7% reduced risk of any enteric infection in the unadjusted model ( risk ratio ( RR ) : 0 . 93 , 95% CI: 0 . 86–1 . 00 ) , though the association was attenuated in adjusted models ( RR: 0 . 95 , 95% CI: 0 . 89–1 . 02 ) . Presence of visible feces or used diapers in the compound was a risk factor in both unadjusted and adjusted models ( aRR: 1 . 07 , 95% CI: 1 . 01–1 . 14 ) . Compound-specific population density was also associated with higher risk of ≥1 enteric infection; children living in the most densely populated quintile of compounds had a 10% higher risk ( aRR: 1 . 10 , 95% CI: 1 . 00–1 . 21 ) of any enteric infection compared with children in the least densely populated compounds . Among a priori covariates adjusted for in models , any breastfeeding was associated with a 13% reduced risk of any infection in adjusted models . Child age was positively associated with enteric infection; children in the oldest age group were 1 . 21 times more likely to have an enteric infection than children in the youngest age category . Risk factors for the any infection were also assessed by multiple imputation ( S3 Table ) and results were consistent with complete case analysis ( Table 3 ) . Risk factors for any bacterial infection were assessed as previously described . Among complete cases ( Table 3 ) , presence of a drop-hole cover in the latrine was associated with reduced risk of any bacterial infection ( aRR: 0 . 90 , 95% CI: 0 . 83–0 . 99 ) . Among a priori covariates , any breastfeeding was associated with 19% reduced risk of bacterial infection in the unadjusted model but was not associated with bacterial infection risk in the adjusted model . Despite increasing prevalence of any bacterial infection with age , we found no association between age and bacterial infection in adjusted models . Results from multiple imputation models were consistent with models limited to complete cases ( S3 Table ) . Among complete cases ( Table 3 ) , presence of a latrine superstructure was associated with 20% reduced risk of any protozoan infection in the unadjusted model but was only marginally associated with reduced risk in the adjusted model ( aRR: 0 . 86 , 95% CI: 0 . 74–1 . 01 ) . In adjusted models , presence of visible feces or used diapers was associated with higher risk of protozoan infection ( aRR: 1 . 16 , 95% CI: 1 . 01–1 . 32 ) . Household crowding , as well as presence of a drinking water tap on the compound grounds , were associated with reduced risk of protozoan infection in adjusted models only ( aRR: 0 . 85 , 95% CI: 0 . 73–0 . 98 and aRR: 0 . 82 , 0 . 68–0 . 99 ) . Among a priori covariates included in all models , any breastfeeding was associated with reduced risk of protozoan infection in both unadjusted and adjusted models ( aRR: 0 . 49 , 95% CI: 0 . 36–0 . 66 ) . Caregiver completion of primary school was associated with 17% reduced risk of protozoan infection in the unadjusted model but was only marginally associated in the adjusted model ( aRR: 0 . 89 , 95% CI: 0 . 79–1 . 01 ) . Age was a risk factor for protozoan infection; children in the 12–23 month and 24–48 month age groups had a 2 . 41 ( 1 . 64–3 . 57 ) and 3 . 20 ( 2 . 14–4 . 80 ) times higher risk of protozoan infection , respectively , than children aged 0–11 months ( Table 3 ) . Among multiple imputation models , most results were in agreement with those in models limited to only complete cases ( S3 Table ) . The presence of visible feces or used diapers around the compound grounds was not associated with increased risk of protozoan infection in unadjusted or adjusted multiple imputation models . Viral infections were not associated with any of the risk factors assessed in adjusted complete case analysis ( Table 3 ) . Household crowding ( presence of >3 persons per room ) was only marginally associated with risk of any viral infection in adjusted models ( aRR: 1 . 55 , 95% CI: 0 . 95–2 . 43 ) . Among a priori covariates , sex was a predictor of viral infection , with girls at higher risk of infection than boys ( aRR: 1 . 65 , 95% CI: 1 . 17–2 . 31 ) . Children in the oldest age group ( 24–48 months ) had 52% reduced risk of any viral infection compared with the youngest age group ( 1–11 months ) . Results from multiple imputation models were consistent with results from models limited to only complete cases ( S3 Table ) . Among risk factors in multiple imputation models , household crowding was a risk factor in the unadjusted model ( RR: 1 . 55 , 95% CI: 1 . 04–2 . 32 ) , but not in the adjusted model . Sex remained a risk factor for viral infection in both unadjusted and adjusted MI models .
We observed a high prevalence of enteric infection , including coinfections , among study children yet most children lacked diarrheal symptoms . The prevalence of enteric infection , but not reported diarrhea , increased with age though pathogen-specific age-related patterns varied . We found some independent WASH or environmental risk factors to be associated with enteric infection , though magnitudes of specific associations were often small . In this setting where burden of disease was high and sanitary conditions were poor , pathogen acquisition , symptomology , and the duration of carriage ( colonization ) , may be driven by multiple interdependent risk and protective factors , including acquired immunity . These results are consistent with findings from other studies of enteric infection in resource-constrained but predominantly rural settings in Africa and elsewhere . The Global Enteric Multicenter Study ( GEMS ) site in the rural district of Manhiça , Mozambique identified one or more enteric pathogens in 85% of stools from children with moderate-to-severe diarrhea ( MSD ) and 76% of stools from control children ( without diarrhea in the 7 days preceding enrollment ) [51] . Similar trends were observed in the Etiology , Risk Factors , and Interactions of Enteric Infections and Malnutrition and the Consequences for Child Health and Development Project ( MAL-ED ) study sites where 77% diarrheal and 65% of non-diarrheal stool samples were positive for ≥1 enteric pathogen [52] . Studies using the GPP for enteric pathogen detection in similar settings in Ghana and Côte d’Ivoire have also found high prevalence of enteric infection among both symptomatic and asymptomatic children [34 , 35] . Compared with enteric infection , the prevalence of caregiver-reported diarrhea was low . We observed a decrease in caregiver-reported diarrhea in children aged 24–48 months compared with the younger age strata , similar to the pattern observed for viral infections . Decreases in reported diarrhea follows a trend observed in historic data of hospital admissions for acute diarrheal episodes among young children in Mozambique [51] . Though we could not formally calculate attributable fractions for etiologic agents of reported diarrhea with these data , we note that norovirus GI/GII was the only enteric pathogen associated with reported diarrhea . This is consistent with findings from the MAL-ED study sites where norovirus GII had one of the highest attributable fractions of diarrhea in children <2 years old [52] . In contrast with reported diarrhea and viral infection , prevalence of bacterial and protozoan infections tended to increase with age , though patterns varied by pathogen . The high prevalence observed here , especially in older children , could be due to the poor clearance and accumulation of persistent enteric infections over time [53] or could be a result of a high rate of reinfection due to frequent pathogen exposure [54] . As children age and become increasingly mobile they interact with their environment more , potentially leading to high exposures to fecal contamination and increased enteric infection [55] . While the overall prevalence of enteric pathogens was similarly high among our study and sites in GEMS and MAL-ED , there were differences in the frequency of detection of specific enteric pathogens . Giardia ( 51% ) , Shigella ( 44% ) , ETEC ( 30% ) , Salmonella ( 21% ) and norovirus GI/GII ( 10% ) were the most frequently detected pathogens in this cohort of children . Giardia , rotavirus , Cryptosporidium , E . histolytica , and enteroaggregative E . coli ( EAEC ) were the most common pathogens detected among cases and controls at the GEMS-Manhiça site , just 80 kilometers north of our study sites [51] . Across all MAL-ED sites , the most frequently detected pathogens in diarrheal and non-diarrheal stools were Campylobacter , Giardia , EAEC , and norovirus GII [52] . Notably , even though our data collection occurred largely before the rollout of the rotavirus vaccine in Mozambique in September 2015 , we detected almost no rotavirus in our study population . This is in stark contrast to findings from the GEMS-Manhiça site where rotavirus was deemed one of the principal causative agents of MSD and was detected in up to 18% of controls [51] . To further interrogate this difference , we tested the 8 rotavirus GPP-positive specimens and 84 randomly selected rotavirus GPP-negative specimens for the presence of rotavirus using the Premier Rotaclone ( Meridian Bioscience , Cincinnati , OH , USA ) in-vitro diagnostic fecal antigen enzyme-linked immunosorbent assay ( ELISA ) [56] . Using the ELISA results as the reference , we calculated the GPP to have 100% sensitivity and 100% specificity for detection of rotavirus A antigen in our fecal specimens . The variations in detection frequencies of enteric pathogens across these studies could be due to differences in detection methods or may suggest that pathogen profiles vary across even limited geographical distances . Molecular reanalysis of the GEMS specimens yielded higher detection frequencies of many bacterial pathogens than the original culture-based methods [57] . However , the GEMS reanalysis did not substantially change detection of Cryptosporidium or rotavirus , highlighting potential geographic differences . Results from this risk factor analysis are consistent with previous studies identifying the build quality or physical characteristics of latrines as factors for increased risk of infection exposure [58]; we found presence of a superstructure or a drop-hole cover to be associated with decreased infection risk . We did not identify any association between enteric infection prevalence and the presence of a cleanable slab , however , consistent with previous work from Tanzania [59] . Associations between the physical characteristics of a latrine and enteric infections were observed only for risk of bacterial and protozoan infections . Household crowding was also associated with a reduced risk of protozoan infection , further evidence that transmission of enteric bacterial and protozoan pathogens is likely to be largely environmentally mediated [20 , 21 , 24] . We did not identify any WASH or environmental variables associated with risk of viral enteric infection . This is consistent with our prior assumption that person-to-person transmission is likely the predominant pathway for viral infection in this setting [25] as has been observed elsewhere under similar conditions [20] . Consistent with previous work , any breastfeeding appeared protective for enteric infection risk in our analysis [60–65] . Adjusted estimates of association show that this observation is primarily driven by protection from Giardia infection ( RR = 0 . 50 , 95% CI 0 . 37–0 . 67 ) ; a similar correlation was also observed in the MAL-ED study [63] . Any breastfeeding limits enteric pathogen transmission by eliminating exposure via direct consumption of contaminated food or water . Maputo , like many cities of sub-Sahara Africa , is rapidly urbanizing [66] . Urbanization may result in higher risk of direct ( person-to-person ) or indirect ( environmentally-mediated ) transmission of enteric infection , especially in low-income , unplanned neighborhoods where WASH infrastructure is lacking [67 , 68] . Recent studies of population density and enteric infection risk have found mixed results , though most were based in rural areas or less dense urban settings [69–71] . In our study , we observed an association between higher compound-level population density and higher risk of enteric infections . There are important limitations to this study that qualify our results . First , our a priori selection of specific pathogen targets and our methods for stool sample analysis present key constraints to interpretation . The GPP tests for 15 of the most common enteric pathogens including bacteria , viruses , and protozoa , but this is a sub-set of all enteric infections and therefore an incomplete accounting of current infections . For example , the GPP does not detect EAEC , a pathogen commonly detected in young children in both MAL-ED and the GEMS-Manhiça site [51 , 52] and associated with malnutrition [64] . Metagenomics or other primer-independent approaches may have yielded information on additional targets of public health significance . Although detection of pathogens in stool samples was observed to be closely associated with age–suggesting persistent infections or frequent reinfection–we cannot make conclusions about either duration of infections or shedding or about the potential for rapid clearance and reinfection based on a single stool specimen . Detection of an enteric pathogen in stool can represent symptomatic or asymptomatic infection , pathogen carriage due to colonization of the gut , or simply passage due to recent exposure . Further , certain pathogens may be shed for weeks after clinical symptoms of infection have abated , and the onset or absence of symptoms following infection can depend on factors related to the environment , host , or pathogen strain of interest [53] . The GPP was designed to aid in diagnosis of enteric infections and the relatively high limits of detection ( 2 . 2x102–3 . 75x106 CFU or copies/mL stool ) [72] largely exceed the known infectious doses for target pathogens . This suggests that enteric pathogen detections via the GPP may primarily represent active infection ( symptomatic or asymptomatic ) or long or short term colonization of the intestinal tract . Although detection of enteric pathogens in feces is an unambiguous indication of past exposure and a clear indication that fecal waste from such individuals represents downstream exposure risks , absence of a particular pathogen in stool by the methods we used does not indicate absence of previous exposure to that pathogen . Because the detection limit of the assay we used is relatively high , a negative assay may not necessarily mean that the pathogen is absent in stool . Cross-sectional , end-point RT-PCR analysis of stool samples alone cannot reveal information on time since exposure , etiology of symptomatic infections , intensity of infections , health implications of infections , or infectivity of pathogens shed in stool . Enteric infections are on the causal pathway between exposures and all downstream health impacts of WASH , including diarrheal disease and environmental enteric dysfunction , but they should be considered an intermediate outcome of uncertain clinical significance . Second , the study population and the study setting , though diverse across some variables , was characterized by a limited range of WASH conditions . All participating households had access to shared sanitation without safe excreta management–a key criterion used in determining eligibility for the MapSan trial–and so exposures were likely to be high across our study sites . This lack of heterogeneity of WASH conditions may have limited our ability to observe variation in risk attributable to specific exposures . Third , certain inclusion criteria may limit the generalizability of our findings . Because our study only included children living in households sharing sanitation in densely populated urban neighborhoods , our results may not represent risks for children in rural areas or in households using private sanitation . Fourth , our analysis is constrained by missing data for variables , including the outcome . A secondary analysis used multiple imputation ( S2 Supporting information and S3 Table ) to handle missing values , and these methods are accompanied by different assumptions and limitations . We note , however , that results from the complete case models and estimates from multiple imputation were largely consistent . Finally , our modeling strategy did not include adjustment for multiple comparisons . While it is possible that some of our findings are spurious and due to type I error [73 , 74] , all variables in this analysis have strong foundations in the literature or plausibility as risk factors for enteric infection . Overall , we found high prevalence of enteric infection and comparatively low prevalence of reported diarrhea among children <4 years old living in informal neighborhoods of Maputo , Mozambique . Most infections were observed in reportedly asymptomatic children . Prevalence of bacterial and protozoan infection increased with child age and is likely due to variations in exposure profiles as children become more mobile . Certain sanitation facility characteristics were associated with decreased risks of enteric infection , though the magnitude of these associations was small . The importance of effective sanitation increases where prevalence of enteric infections is high: fecal wastes in such settings present elevated exposure risks , potentially driving burdens of infection and disease higher . Strategies to interrupt this cycle of infection and exposure risk should limit the possibility of exposure to excreta , including through multiple pathways of transmission .
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the U . S . Centers for Disease Control and Prevention .
|
Enteric pathogens such as bacteria , protozoa , and viruses can cause diarrhea and other longer-term health problems . Poor sanitary conditions , including inadequate sanitation facilities , can lead to contamination of the living environment and higher risk of exposure to and transmission of enteric pathogens . Young children , who are vulnerable to both the short- and long-term health effects of enteric infections , interact with their environment in different ways than older children or adults . In order to limit enteric pathogen transmission among this vulnerable group , we must understand the infection burden and the environmental or sanitation-related factors that are associated with infection . Among a group of children younger than four years old living in low-income neighborhoods of Maputo , Mozambique , we found over 85% tested had ≥1 enteric infection . Children living in environments visibly contaminated with feces were more likely to have an infection than children whose living environments were not visibly contaminated . In contrast , children living in compounds with certain latrine features , including walls and pit covers ( potential indicators of build quality ) , had reduced infection risk . Understanding that these risk factors may play important roles in exposure and transmission in this setting is key to planning effective interventions .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion",
"Disclaimer"
] |
[
"children",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"viral",
"transmission",
"and",
"infection",
"pathogens",
"microbiology",
"parasitic",
"diseases",
"health",
"care",
"diarrhea",
"age",
"groups",
"signs",
"and",
"symptoms",
"gastroenterology",
"and",
"hepatology",
"sanitation",
"medical",
"risk",
"factors",
"bacterial",
"pathogens",
"families",
"public",
"and",
"occupational",
"health",
"medical",
"microbiology",
"epidemiology",
"microbial",
"pathogens",
"protozoan",
"infections",
"people",
"and",
"places",
"environmental",
"health",
"diagnostic",
"medicine",
"virology",
"biology",
"and",
"life",
"sciences",
"population",
"groupings"
] |
2018
|
Risk factors for childhood enteric infection in urban Maputo, Mozambique: A cross-sectional study
|
Jacob , the protein encoded by the Nsmf gene , is involved in synapto-nuclear signaling and docks an N-Methyl-D-Aspartate receptor ( NMDAR ) -derived signalosome to nuclear target sites like the transcription factor cAMP-response-element-binding protein ( CREB ) . Several reports indicate that mutations in NSMF are related to Kallmann syndrome ( KS ) , a neurodevelopmental disorder characterized by idiopathic hypogonadotropic hypogonadism ( IHH ) associated with anosmia or hyposmia . It has also been reported that a protein knockdown results in migration deficits of Gonadotropin-releasing hormone ( GnRH ) positive neurons from the olfactory bulb to the hypothalamus during early neuronal development . Here we show that mice that are constitutively deficient for the Nsmf gene do not present phenotypic characteristics related to KS . Instead , these mice exhibit hippocampal dysplasia with a reduced number of synapses and simplification of dendrites , reduced hippocampal long-term potentiation ( LTP ) at CA1 synapses and deficits in hippocampus-dependent learning . Brain-derived neurotrophic factor ( BDNF ) activation of CREB-activated gene expression plays a documented role in hippocampal CA1 synapse and dendrite formation . We found that BDNF induces the nuclear translocation of Jacob in an NMDAR-dependent manner in early development , which results in increased phosphorylation of CREB and enhanced CREB-dependent Bdnf gene transcription . Nsmf knockout ( ko ) mice show reduced hippocampal Bdnf mRNA and protein levels as well as reduced pCREB levels during dendritogenesis . Moreover , BDNF application can rescue the morphological deficits in hippocampal pyramidal neurons devoid of Jacob . Taken together , the data suggest that the absence of Jacob in early development interrupts a positive feedback loop between BDNF signaling , subsequent nuclear import of Jacob , activation of CREB and enhanced Bdnf gene transcription , ultimately leading to hippocampal dysplasia .
Jacob is a protein that shuttles to the nucleus in response to activation of synaptic and extrasynaptic GluN2B-containing N-methyl-D-aspartate receptors ( NMDARs ) [1] and that following long-distance transport and nuclear import , encodes and transduces the synaptic and extrasynaptic origin of NMDAR signals to the nucleus [2] . Extracellular-signal-regulated kinase ( ERK1/2 ) -binding and ERK-dependent phosphorylation of a crucial serine at position 180 in Jacob encodes synaptic but not extrasynaptic NMDAR activation . A stable trimeric complex with proteolytically cleaved fragments of the neurofilament α-internexin is formed which protects Jacob and active ERK against phosphatase activity during retrograde transport . In the nucleus this signalosome-like complex enhances 'plasticity-related' and 'CREB-dependent' gene expression as well as synaptic strength [2] . In stark contrast , following extrasynaptic NMDAR activation nuclear import of Jacob results in sustained dephosphorylation and transcriptional inactivation of the transcription factor CREB ( CREB shut-off ) , loss of synaptic contacts , retraction of dendrites and eventually cell death [1 , 2] . In addition Jacob couples pathological Amyloid-β signaling to the nucleus via extrasynaptic GluN2B-NMDAR activation [3 , 4] and several lines of evidence suggest that the protein is involved in both neuronal plasticity as well as neurodegeneration . However , some reports indicate that the mouse orthologue of Jacob called nasal embryonic luteinizing hormone-releasing hormone ( LHRH ) factor ( NELF ) is essential for the migration of GnRH-expressing cells from the olfactory bulb to the hypothalamus during neuronal development [5 , 6] . These studies suggest that the protein functions as an extracellular guidance molecule on growth cones , which is essential for axon growth and routing of GnRH positive cells along vomeronasal olfactory-derived axons and eventually migration to the hypothalamus [5 , 6] . Several subsequent publications have tried to establish a link between mutations in the NSMF gene that encodes for Jacob/NELF and Kallmann syndrome ( KS ) [7–14] . One of these reports even suggests a monogenic causation of KS by a point mutation in the NSMF gene [13] . KS is a rare neurodevelopmental disorder that is characterized by a migration deficit of GnRH neurons that fail to migrate into the hypothalamus during embryonic development [15] . Mutations in several genes have been postulated to cause hypogonadism , delayed or absent puberty and almost invariably infertility [15] . Besides hypogonadotropic hypogonadism , KS patients exhibit anosmia or severe hyposmia and variable other phenotypes that depend upon the type of gene mutation that might underlie the condition [14] . Unfortunately , a number of discrepancies between reports on Jacob and NELF ( see above and [16] ) exist that obscure the interpretation of previous results . In brief , apart from reported differences in the subcellular localization and function also the originally published open reading frame of Nelf did not match to those of Jacob [5 , 1] . Most studies on loss of function phenotypes in either development or adulthood have been performed with in vitro preparations and utilized antisense oligonucleotides [5 , 6] or shRNA protein knockdown of Jacob/NELF [1 , 2] . Thus , deletion of the gene in vivo could potentially clarify some of the controversial issues . We therefore first sought to determine whether inactivation of the Nsmf gene results in phenotypes related to KS . Second , since very little is known about the role of Jacob in hippocampal development we examined whether hippocampal circuitry and function is affected by the lack the activity-dependent nuclear import of the protein in Jacob/Nsmf ko mice .
In order to delete Jacob in mice a targeting construct was generated by flanking the first three exons of the Nsmf gene with loxP sites to allow for Cre-mediated deletion in all tissues ( S1A Fig ) . Knockout was verified by genomic PCR ( S1B Fig ) , RT-PCR ( S1C Fig ) and by western blotting ( S1D Fig ) . Western Blots showed that the Jacob protein is expressed at high levels in mouse hippocampus and cortex , but is also present in other brain regions such as the striatum , hypothalamus and olfactory bulb . In heterozygous animals , protein expression is already markedly reduced and in ko mice the Jacob protein is not detectable ( S1D Fig ) . Jacob/Nsmf ko mice are viable and fertile , homozygous ( -/- ) , heterozygous ( +/- ) and wild-type ( wt , +/+ ) breeding pairs show comparable litter size and a Mendelian ratio as expected ( S1E Fig ) . Body weights were registered upon weaning ( 20–22 days ) and at the age of 4 months . No differences could be observed among the genotypes ( S1E Fig ) . We next investigated whether a gonadal defect evoked by the lack of Jacob expression is present in male and female mice . We found that the morphology and weight of Jacob/Nsmf ko and wt testes were unaltered . Hematoxylin and eosin staining revealed a comparable testicular anatomy ( Fig 1A and 1B ) . Jacob/Nsmf ko and wt mice showed no differences in appearance of various stages of differentiation from spermatogonia to sperm cells . All stages of spermatogenesis were observed within the tubules . Furthermore , no differences between genotypes were observed in chromosome preparations ( S2 Fig ) . Individual chromosomes can be distinguished during mitosis of spermatogonia ( S2A and S2E Fig ) . In diakinesis-metaphase I , the tetrads of the homologous chromosomes can be seen ( S2B and S2F Fig ) . In metaphase II the separation of the chromatids of each chromosome are visible ( S2C , S2D , S2G and S2H Fig ) . Additionally , testes of both genotypes were similar in mean size and weight ( Fig 1C ) . To exclude impaired infertility due to hormonal abnormalities , we additionally measured testosterone serum levels in all three genotypes . In male ko mice testosterone serum levels were slightly but not statistically significant reduced ( Fig 1D ) . In the next set of experiments , Jacob/Nsmf heterozygous and homozygous females were examined in comparison to wild-type ( wt ) littermates regarding reproductive organs , estrous cycle phases and sex hormones . Analysis of each estrous cycle phase revealed that Jacob deficiency prolonged the duration of the metestrus phase ( Fig 2C ) while shortening the diestrus phase . Statistically significant differences were found in both phases between Jacob/Nsmf ko and wt female mice . By contrast , no differences could be observed in the duration of the proestrus and estrus phase among all groups ( Fig 2C ) . An adequate folliculogenesis is a prerequisite for ovulation . To study the influence of Jacob deficiency on follicle development and ovulation we determined the number of the different follicle stages and corpora lutea . Fig 2A shows representative pictures of ovarian tissue from all genotypes . We found statistically significant lower numbers of primary and tertiary follicles in ko females when compared to wt females ( Fig 2B ) . In addition , the number of secondary follicles but not the number of corpora lutea was slightly reduced in Jacob deficient females compared to wt animals ( Fig 2B ) . Hormonal fluctuations during the estrous cycle are indispensable to ensure proper folliculogenesis , ovulation and preparation of the endometrium for implantation . Since the lack of Jacob affects the number of primary and tertiary follicles we sought to study whether these effects are associated with alterations in hormonal levels . Plasma levels of Luteinizing hormone and Progesterone were comparable among all genotypes throughout the estrous cycle ( Fig 2D ) , while Estradiol levels were significantly reduced in the estrus and metestrus phase when Jacob was absent ( Fig 2D ) . Taken together the data suggest that a Jacob/Nsmf gene ko has a modest effect on reproductive parameters in female mice but does not lead to hypogonadotropic hypogonadism or subfertility ( see S1E Fig ) . Since anosmia or hyposmia is observed in about 60% of patients with IHH and a distinctive feature for KS [17] we next tested whether Jacob-deficient mice display deficits in olfactory behavior . Synthetic 2 , 5-Dihydro-2 , 4 , 5 trimethylthiazoline ( TMT ) , a predator cue isolated from red fox anal secretion , is frequently used to induce unconditioned fear in rodents . Diethylphthalate ( DEP ) was used as a solvent for dilution of TMT and served as blank control in odor exposure experiments . Jacob/Nsmf ko and wt mice were individually exposed for 15 min to TMT or DEP ( S3 Fig and S1 Table ) . The exposition to TMT or DEP induced similar behavioral effects in mice from both genotypes ( S1 Table ) . Freezing levels were significantly higher in mice exposed to TMT , indicating that the predator odor was smelt ( S3 Fig ) . In contrast , mice treated with DEP showed significantly higher levels in sniffing , grooming and scratching , more rearing in the center area and leaning on the wall and higher locomotor activity ( S1 Table ) . However , no differences between genotypes were found in any of the other analyzed parameters except that Jacob/Nsmf ko mice produced a higher number of fecal boli during the odor exposition ( S1 Table ) . Along these lines we found that the general organization of the bulbus olfactorius is normal in Jacob/Nsmf ko mice ( Fig 3A–3C ) . No gross abnormalities in cell distribution and layering as well as the number of GnRH-immunoreactive ( IR ) neurons and fiber density in stratum glomerulosum ( Gl ) and stratum plexiforme externum ( EPl ) of the olfactory bulb were apparent ( Fig 3A–3D ) . A hallmark of KS is a migration deficit of GnRH-expressing neurons from the olfactory bulb to the hypothalamus and a reduced number of GnRH-immunopositive fibers in the hypothalamus . In order to determine if Jacob/Nsmf ko and wt littermates show differences in markers of the hypothalamic-pituitary-gonadal axis we investigated both genotypes for the number of GnRH-IR neurons in Gl and EPl of olfactory bulb ( Fig 3A–3C ) , in the medial septum ( MS ) , ventral diagonal ( VDB ) , horizontal diagonal band ( HDB ) of Broca and preoptic area ( PA ) ( Fig 3F–3H and 3D for quantification ) . A two-way repeated-measures ANOVA showed no significant interaction of REGION and GENOTYPE ( F ( 1 . 90 , 25 . 54 ) = 1 . 49 , p>0 . 05 ) and no significant effect of the between-subject factor GENOTYPE ( F ( 1 , 14 ) = 0 . 41 , p>0 . 05 ) . Thus , Jacob/Nsmf ko and wt mice show a comparable number of GnRH-IR neurons and no difference in the number of GnRH-expressing neurons in the hypothalamus . In addition GnRH-IR fiber densities were estimated in Gl , EPl , MS , VDB , HDB , PA , anteroventral paraventricular hypothalamic nucleus ( AVPe ) and nucleus arcuatus hypothalami ( Arc ) ( Fig 3E ) . Again , a two-way repeated-measures ANOVA showed no significant interaction for the region and genotype examined ( F ( 1 . 80 , 25 . 17 ) = 0 . 16 , p>0 . 05 ) and also no significant effect of the genotype on these measures ( F ( 1 , 14 ) = 0 . 47 , p>0 . 05 ) . We next asked whether a Jacob/Nsmf gene knockout affects brain organization and behavior . Nissl staining of mice brain sections revealed no obvious differences and indications of gross abnormalities like for instance in the hippocampus between both genotypes ( S4 Fig ) . MRI of the intact brain , however , revealed a small but statistically significant increase in the volume of the striatum ( S5 Fig ) . The volume of other brain regions was not affected by the genotype ( S5 Fig ) . Jacob-deficient mice did not exhibit clear signs for visual and auditory sensory deficits and motor behavior in the rotarod was not significantly altered in ko animals as compared to wt independently of the sex of the mice . Ko animals seemed to stay longer on the rotarod at slow velocities ( S6A Fig ) . Subsequent behavioral testing in the open field revealed a hyperactive phenotype of ko mice compared to wt littermates independent of the sex . Within the 15 min test period in the open field Jacob/Nsmf ko mice spent significantly more time with locomotor activity ( distance and speed , S6B and S6C Fig ) than wt mice . A social interaction test showed no significant behavioral differences , presenting only a trend towards a more aggressive behavior and more frequent anogenital sniffing in Jacob/Nsmf ko mice compared to wt mice ( S2 Table ) . Taken together the data indicate that Jacob/Nsmf ko mice are hyperactive when tested in the open field but no apparent sensory-motor deficits were found . We therefore next addressed whether conditioned behavior and cognitive function might be impaired in Jacob-deficient mice . To this end , we tested mice of both genotypes in auditory-discrimination learning , a cortex-dependent learning task [18] , and contextual fear conditioning , a behavioral paradigm that is sensitive to alterations in hippocampal circuitry and function [19] . Interestingly , Jacob/Nsmf ko mice exhibited normal auditory discrimination learning with a normal learning curve and performance over the course of the experiment ( S7 Fig ) . In stark contrast , Jacob-deficient mice were clearly impaired in contextual fear conditioning , a learning task that has been associated with the induction of Hebbian plasticity at hippocampal CA1 synapses [19] ( Fig 4A ) . Cued fear conditioning , that is independent of hippocampal function [19] , was not affected in ko mice ( Fig 4B ) . Furthermore Jacob/Nsmf ko mice were also clearly impaired in object recognition memory ( Fig 4C ) , a behavioral paradigm that is also sensitive to hippocampal dysfunction . In previous work we found that Jacob rapidly translocate to the nucleus after induction of NMDAR-dependent LTP but not long-term depression ( LTD ) at hippocampal CA1 synapses [20] . We therefore tested here whether Jacob/Nsmf ko mice exhibit normal Schaffer collateral LTP . In these experiments , we found a clear decay of the field excitatory postsynaptic potential ( fEPSP ) slope already 30–40 minutes after the induction of LTP ( Fig 4D ) , although basal synaptic transmission ( S8A Fig ) , paired-pulse facilitation and the input-output curve were not shifted as compared to wt controls ( S8B and S8C Fig ) . Late phase LTP was almost absent in Jacob/Nsmf ko mice despite a similar induction of early LTP like in wt mice ( Fig 4D ) . Collectively , these data indicate that deletion of the Nsmf gene in development results in an impairment of hippocampal function and reduced synaptic plasticity of hippocampal CA1 synapses in adulthood . We next wondered whether the observed functional deficits in adult mice could be a result of structural alterations and therefore analyzed synapto-dendritic cytoarchitecture of CA1 pyramidal neurons using the Golgi-Cox method . A subsequent Sholl analysis of Golgi-Cox stained cells revealed a clear reduction of dendritic complexity in CA1 neurons of Jacob/Nsmf ko mice ( Fig 5 ) . Apical and more prominently basal dendrites were affected ( Fig 5A–5C ) . Jacob-deficient mice exhibit shorter dendrites and much lower number of branches for a given dendritic segment . This rather prominent phenotype prompted us to investigate the density of spine synapses . Interestingly , spine density was clearly reduced in basal and apical dendrites in secondary branches of CA1 neurons ( Fig 5D–5G ) . Further evidence for structural deficits in hippocampal development in the Jacob/Nsmf ko mice came from TIMM-staining that revealed an enlargement of the dentate gyrus ( DG ) projection as compared to wt mice ( S9A–S9E Fig ) . In addition , we found an altered catecholaminergic innervation of the hippocampus with a clear increase in the number of tyrosine-hydoxylase positive fibres in the CA3 but not CA1 region or DG in Jacob/Nsmf ko mice ( S9F–S9J and S10 Figs ) . In contrast , catecholaminergic innervation in the striatum and ventral tegmental area was normal in ko mice compared to wt littermates as shown by tyrosine-hydroxylase staining ( S11 Fig ) . Thus , the analyses suggest that structural abnormalities in the hippocampus might underlie the functional deficits observed in adult Jacob/Nsmf ko mice . The data outlined above suggest that neuronal development of hippocampal pyramidal neurons might be compromised in Jacob/Nsmf ko mice . To further confirm this phenotype we investigated in the next set of experiments dendrite development and synaptogenesis in mouse hippocampal primary cultures . A Sholl analysis revealed that dendrite development is also impaired in hippocampal primary neurons deriving from Jacob/Nsmf ko mice . In comparison to wt controls , Jacob-deficient neurons exhibited a shorter neurite length and less branching of Microtubule-associated protein 2 ( MAP2 ) -positive dendrites at day in vitro ( DIV ) 10 but not at DIV 5 ( Fig 6A and 6B ) . Moreover , the number of synaptic contacts was reduced in later development ( DIV15 ) in Jacob/Nsmf ko neurons as evidenced by staining with the postsynaptic marker Homer1 and the presynaptic marker Synaptophysin ( Fig 6D ) . Thus , synaptogenesis is impaired in hippocampal primary neurons devoid of Jacob . Interestingly , a similar analysis of cortical primary neurons from wt and Jacob/Nsmf ko mice revealed no significant differences in dendrito- and synaptogenesis between both genotypes ( S12 Fig ) . Collectively these data point out a profound hippocampal dysplasia and impaired synapto-dendritic development in Jacob/Nsmf ko mice . We therefore sought to unravel the underlying mechanisms . In previous work we have been able to show that Jacob regulates transcriptional activity of CREB and CREB-dependent Bdnf gene transcription [2] . BDNF has been implicated in spinogenesis and dendrite development [21] . We therefore proceeded to examine hippocampal mRNA and protein levels of BDNF in wt and Jacob/Nsmf ko mice during development . Quantitative PCR experiments revealed decreased Bdnf exon IV expression in CA1 and CA3 regions from samples of P10 Jacob/Nsmf ko mice ( Fig 7A ) . ELISA-assays confirmed significantly lower BDNF levels in Jacob-deficient mice at P10 in CA1 and a trend towards reduced levels in CA3 ( Fig 7B ) . CA1 BDNF levels appear to be slightly but not statistically significant reduced in adulthood ( Fig 7C ) . We therefore next asked whether Bdnf promoter activity might be differentially regulated in Jacob/Nsmf ko neurons at P10 . To this end , we performed a reporter gene assay where mouse primary hippocampal neurons were transfected with a GFP construct fused to the Bdnf- exon1/2 promoter which contains a Cre-site and is regulated by Jacob [2] and then applied BDNF stimulation . In these experiments we observed that the relative BDNF-induced increase in GFP expression was clearly attenuated in Jacob-deficient neurons as compared to wt neurons ( S13 Fig ) , indicating that a positive feedback loop involved in Bdnf gene expression might be interrupted following Jacob/Nsmf gene knockout . Published work suggests that BDNF might have a positive impact on dendrite development and spinogenesis of hippocampal primary neurons [21] . We next performed in vitro experiments in which hippocampal primary neurons derived from Jacob/Nsmf ko mice were cultured in the presence of BDNF . Bath application of 100 ng/ml BDNF at DIV2 and DIV6 increased the complexity of the dendritic tree of Jacob/Nsmf ko neurons at DIV 10 ( Fig 6C ) as compared to the ko controls grown without BDNF . No effect of this treatment was seen in control wt neurons ( Fig 6C ) . Similarly , we found an increase in the number of spine synapses following BDNF supplementation at DIV2 and 6 in Jacob/Nsmf ko but not in wt neurons at DIV15 ( Fig 6D ) . Control experiments with administration of NGF and boiled BDNF yielded no or very little effect on dendrito- and synaptogenesis ( S14 Fig ) . Of note , quantitative immunoblotting revealed that TrkB as well as pTrkB levels are not altered as compared to wt in hippocampal lysates from adult Jacob-deficient mice ( S15 Fig ) . BDNF application could also not rescue the LTP phenotype in adult Jacob/Nsmf ko mice ( S16 Fig ) . Collectively the data suggest that specifically BDNF-signaling pathways appear to be sensitized to BDNF administration in Jacob-deficient neurons in comparison to wt , and enhancing BDNF levels corrects , at least in part , the deficit in synapto-dendritic development in cultured Jacob/Nsmf ko neurons . Previous reports have claimed that Jacob/NELF is a secreted protein that associates with the cell surface of neurons and might function as a guidance molecule [5 , 6] ( but see [16] ) . To exclude the possibility that Jacob might have an additional function as a guidance molecule on top of synapse-to-nucleus communication that might account for the findings described above we performed live-staining of rat hippocampal primary neurons with a Jacob antibody . These experiments revealed no specific staining above the background signal resulting from the secondary antibody ( S17 Fig; see S18 Fig for specificity of the antibody ) . A positive live-staining control with a prion protein ( PrP ) -antibody showed extensive labeling of hippocampal primary neurons ( S17 Fig ) . Hence , prominent Jacob immunofluorescence was evident following permeabilization of cells with a conventional TritonX-100 buffer ( S17 Fig ) . Collectively , these data do not support an extracellular localization of Jacob in rat hippocampal primary neurons and rather point to a function of Jacob in NMDAR-to-nucleus signaling , as shown previously [2] . We next wondered whether the low BDNF-levels in Jacob/Nsmf ko mice in early development might relate to the nuclear import of the protein and whether exogenous BDNF itself can drive Jacob into the nucleus . In the first set of experiments we found that acute bath application of BDNF ( 100ng/ml ) indeed results in nuclear accumulation of Jacob in rat hippocampal primary neurons at DIV10 and that this increase was completely blocked when the NMDAR antagonist AP5 was co-applied ( Fig 8A ) . In accordance with previous observations , we found a prominent increase in nuclear pJacob immunofluorescence in response to acute BDNF application at DIV10 ( Fig 8B ) . Additionally , BDNF application at DIV 10 substantially increased nuclear pCREB immunoreactivity ( Fig 8D ) whereas CREB immunofluorescence remained unaltered ( Fig 8C ) . This increase in pCREB was partially blocked by AP5 ( Fig 8D ) . BDNF-induced CREB activation from NMDAR-independent pathways likely accounts for the remaining increase . Collectively these data suggest a scenario where BDNF drives pJacob during dendritogenesis into the nucleus which then results in docking of pERK to the CREB complex [2] , increased serine 133 phosphorylation of CREB and CREB-dependent enhanced Bdnf gene transcription . Newly synthesized BDNF would then positively feedback to dendrite growth and synaptogenesis and concomitantly increase nuclear import of Jacob . In the next set of experiments we analyzed whether exogenous BDNF application increases pERK and pCREB in the nucleus of hippocampal primary neurons of Jacob/Nsmf ko mice . We found that the nuclear expression of pERK 30 minutes after the onset of stimulation with BDNF was significantly blunted in Jacob-deficient neurons at DIV10 following BDNF application for 30 minutes ( Fig 9A–9C ) . Moreover , basal nuclear ERK levels were greatly reduced in Jacob-deficient neurons as compared to wt neurons , indicating a prominent role of Jacob for BDNF-induced nuclear import of ERK/pERK during the development of mouse hippocampal primary neurons ( Fig 9A–9C ) . Additionally , under basal conditions pCREB but not panCREB immunofluorescence levels were significantly reduced in hippocampal neurons devoid of Jacob ( Fig 10 and S19 Fig ) . Application of the NMDAR antagonist AP5 to these cultures normalized the difference in pCREB staining intensity ( Fig 10 ) , suggesting that NMDAR in wt neurons are more efficiently coupled to pCREB . Interestingly , however , acute application of BDNF results in an increase of pCREB immunofluorescence that is relative to basal levels even more pronounced in Jacob-deficient than in wt neurons ( Fig 10 ) , indicating sensitization of BDNF-triggered signaling pathways coupled to activation of CREB . These pathways appear to be NMDAR-independent since co-application of AP5 resulted in a statistically significant effect only in wt neurons ( Fig 10 ) .
In the present study we addressed whether Jacob/Nsmf null mutant mice exhibit phenotypes related to KS and investigated the impact of Jacob on neuronal development with emphasis on the hippocampus . We found no evidence for anosmia or a major migration deficit of GnRH neurons in these mice . Sexual development , fertility hallmarks and reproductive capacity appear to be to a large extent normal . The results of the present work are in contrast to a recent report that claims subfertility and impaired puberty in female ko mice [22] . Our detailed analysis of a broad number of reproductive parameters clearly excludes subfertility and hypogonadotropic hypogonadism in both males and females . The reason for this discrepancy is not apparent but the mild phenotypes observed by us in knockout animals ( i . e . irregularities in folliculogenesis and in the estrus cycle that might be estradiol driven , lower testosterone levels in male mice ) clearly do not result in subfertility . The findings herein are also incongruent with previous reports claiming that Jacob/NELF acts on growth cones as an extracellular guidance molecule that enables GnRH neurons to reach their destination in the hypothalamus [5 , 6] . On the ground of the present and previous work [16] it is also unlikely that Jacob/NELF is a secreted molecule . It is therefore tempting to speculate that , if Jacob has any function in neuronal cell migration it will be related to the regulation of gene expression . Interestingly , at later stages of migration , NMDAR activation slows down the migration of GnRH neurons [23] and Jacob has been shown to translocate to the nucleus following NMDAR stimulation [1 , 2] . However , deficits in cell migration were not apparent in any of the brain regions examined in Jacob/Nsmf ko mice and Jacob protein levels are relatively low at birth and prominently increase only in the second postnatal week [1 , 24] . Moreover , it is important to emphasize that , in recent years , there has been extensive documentation of a high degree of heterogeneity within the GnRH neuronal population and it is unlikely that a single genetic mutation could completely prevent these neurons from reaching their destination [11 , 25] . Thus , GnRH deficiency in humans might only occur when the majority of GnRH neurons are affected by mutations in more than one gene . Along these lines it has been proposed that mainly digenic mutations in genes that are important for migration of GnRH neurons will result in IHH [9 , 11 , 25] . Taken together , the studies performed in null mutant mice provided no evidence for an involvement of Jacob in KS , although the experiments conducted in the present study do not rule out subtle defects in GnRH migration . The most compelling developmental phenotype that we observed in Jacob/Nsmf null mutant mice was hippocampal dysplasia . We could show that Jacob deficiency during development results in shorter dendrites , less dendritic branching , fewer synapses , an increased catecholaminergic innervation , an altered CA3-DG projection as well as reduced pCREB and BDNF levels . In consequence , Jacob/Nsmf ko mice show behavioral deficits in contextual fear conditioning and object recognition memory , two hippocampus-dependent learning tasks and impairments of synaptic plasticity in classical NMDAR-dependent Schaffer-collateral CA1 LTP . We surprisingly found no alteration in fEPSP size despite the fact that the number of CA1 synapses is reduced in Jacob/Nsmf knockout mice . There are a number of possible explanations for this unexpected finding , which includes a shift in the excitation/inhibition balance , changes in intrinsic excitability and synaptic surface expression of NMDA/AMPA receptors as well as others . Interestingly , BDNF application during tetanization could not rescue late LTP . This might be due to the impaired nuclear import of pERK and reduced CREB-dependent gene expression following activation of synaptic NMDAR . Since this import is already impaired in development it will affect synaptic function permanently . Although no obvious differences in TrkB and pTyr550TrkB levels were found in hippocampi of knockout mice we can therefore not exclude the possibility that synaptic changes occur that are either directly or indirectly related to BDNF/TrkB signaling in Jacob deficient neurons . Thus , the preliminary analysis does not exclude alterations in downstream TrkB signaling and further research should address these issues . In addition , the present study suggests that loss-of-function mutations in the Jacob/Nsmf gene are likely to cause selective cognitive dysfunction in humans . Of interest in this regard is a report on hippocampal dysplasia in two cases of KS [26] . Cortex- and amygdala-dependent auditory learning , in contrast , was not significantly affected in the mutants . We therefore focused our mechanistic analysis mainly on hippocampal CA1 pyramidal neurons . Nonetheless it will be interesting to see whether similar structural and functional deficits are also present in other brain regions . Initial experiments with cortical primary neurons revealed no clear-cut differences in dendrito- and synaptogenesis between both genotypes at DIV10 and DIV15 . The cortex on the other hand is very heterogeneous in terms of cellular architecture in different cortices and one should be therefore cautious to conclude that the effects of Jacob deficiency are specific for CA1 pyramidal neurons . The main novel mechanistic insight that we gained from the present work is that the underlying mechanism for the defect in synapto- and dendritogenesis in Jacob/Nsmf null mutant mice is probably due to an interrupted positive feedback loop between BDNF-signaling , subsequent nuclear import of pJacob in a complex with pERK , activation of CREB and enhanced BDNF gene transcription . Thus , the present study shows that the previously documented nuclear import of pJacob and subsequent regulation of CREB-dependent BDNF gene expression [2] seems to be also relevant for neuronal development in the hippocampus . Jacob-deficient neurons are more responsive to BDNF application , a fact that is reflected by higher relative increase in pCREB and because BDNF administration normalizes the number of synapses and dendrite complexity at a concentration and treatment regime that has no effect on wt neurons . Interestingly , a previous report has shown that BDNF also regulates dendritic development via co-activation of NMDARs , nuclear import of the synapto-nuclear messenger CRTC1 and binding of CRTC1 to CREB in neuronal primary neurons [27] . Thus , long-distance transport of proteins from the plasma membrane to the nucleus and docking to CREB seems to be a more common mechanism . Three main intracellular signaling cascades are activated by tropomyosin-receptor kinase B ( TrkB ) : the Ras-ERK , the PI3K-Akt and the PLCγ-Ca2+ pathway [28] . BDNF-induced nuclear import of Jacob requires co-activation of NMDAR and crosstalk with TrkB receptors and one possibility is that co-activation might directly result in sustained ERK-activation and thereby enhanced nuclear trafficking of Jacob [2] . Jacob preferentially associates with GluN2B-containing NMDAR and nuclear import essentially requires activation of this receptor subtype [1 , 2] . On the postsynaptic side a functional interaction of BDNF/TrkB signaling with the GluN2B subunit of NMDAR is well documented and this link might depend on activation of Src-family tyrosine kinases by PI3K-Akt [29–31] . The C-terminus of GluN2B contains a clathrin adaptor AP-2-binding site and the internalization motif YEKL [32–34] . The tyrosine 1472 within the YEKL motif of GluN2B is phosphorylated by the Src family of kinases [35 , 36] , and phosphorylation of Y1472 inhibits the binding of AP-2 and promotes surface expression of GluN2B [32 , 36] . BDNF and GluN2B-containing NMDARs contribute to long-term synaptic potentiation [37 , 38] and the BDNF/TrkB enhancement of GluN2B signaling is likely dependent on CaMKII phosphorylation , CaMKII/GluN2B binding [39 , 40] and subsequent AMPA receptor modifications [41–46] as well as activation of ERK [47] . It is tempting to speculate that in Jacob/Nsmf mutant mice this signaling might be affected in adulthood . Finally , the role of BDNF in dendritogenesis is rather controversial also because different dose and administration regimens were used and different cell types and developmental stages were analyzed . Exogenous application of BDNF has been shown in most studies to promote dendrite outgrowth in development and spine density and morphology in mature primary neurons ( reviewed in [21 , 37] ) . However , while a complete gene knockout is lethal heterozygous , Bdnf +/- , mouse mutants with reduced brain BDNF levels show only a mild phenotype [37 , 38] [48 , 49] . The analysis of conditional gene targeted mouse lines has revealed that the morphological effects of a Bdnf gene ko are relatively mild if one compares these effects to those of exogenous BDNF application in vitro [50–58] . Rauskolb et al . [58] reported a modest dendritic phenotype of CA1 pyramidal neurons in conditional Bdnf -/- mice , that was much less prominent than those observed in the striatum in the same study . Apart from methodological differences it is plausible that the complete knockout of the Bdnf gene in neurons already very early in neuronal development might have a different effect as compared to the situation in the present work . The Jacob pathway only comes into play later in postnatal development when Jacob levels start to increase . Thus , it is possible that wiring of hippocampal circuitry in the absence of BDNF is different than in the presence of lower levels in a restricted time window . Compensatory mechanisms are plausibly different in both scenarios and it should also be emphasized that altered gene expression in Jacob/Nsmf ko mice might not only affect BDNF levels but also other factors involved in dendritogenesis . Another issue is a possible sensitization to BDNF signaling in the mice that might account for the compensation seen after BDNF application in primary neurons . Along these lines TrkB signaling might be altered in Jacob null mutant mice and it is very likely that Jacob will have a synaptic function ( see above ) . Moreover , interruption of the positive feedback loop that we describe in the paper might interfere with synaptogenesis independent of dendritogenesis and will likely not account for all deficits in the null mutant mice . In summary , we could show that Jacob regulates dendrite growth in hippocampal pyramidal neurons in CA1 and that protein transport of Jacob from TrkB/NMDAR to the nucleus in development is part of a positive feedback loop that promotes both dendrito- and synaptogenesis . Later in development Jacob's nuclear transport is probably under increasing control of synaptic NMDAR as compared to earlier time points when the number of spines is low and we assume that the functional deficits in adult mice at least in part reflect dysplasia and structural deficits in CA1 that are imposed by a gene knockout early in development .
All experiments were carried out in accordance with the European Communities Council Directive ( 2010/63/EU ) and approved by the local authorities of Sachsen-Anhalt/Germany / Regierungspräsidium Halle Sachsen-Anhalt/Germany ( reference number 42502-2-987IfN ) . All animals used in this study were bred and maintained in the animal facility of the Leibniz Institute for Neurobiology , Magdeburg , Germany . Animals were housed in groups of up to 5 in individually ventilated cages ( IVCs; Green line system , Tecniplast ) with free access to food and tap water under controlled environmental conditions ( 22° +/- 2°C , 55% +/- 10% humidity , 12h , light—dark cycle , with lights on at 06:00 a . m . ) . Jacob/Nsmf ko mice were generated with help from Ozgene Pty . For conditional targeting exons 1 to 3 of the Jacob/Nsmf locus ( GenBank ID: 56876; RefSeq NM_001039386 . 1; NP_001034475 . 1 ) were flanked with loxP sites for Cre-mediated deletion via the design of two loxP arms ( 0 , 7 and 0 , 9 kb ) . The 5´loxP site was inserted into the 5´UTR of exon 1 whereas the 3´loxP site was inserted downstream of exon 3 to ensure deletion of the first 3 exons . Three loxP sites were used to reduce the size of the regions being floxed on either side of the PGK-Neomycin selection cassette ( see S1 Fig ) . Jacob/Nsmf constitutive ko mice were generated using standard procedures from targeted C57BL/6-derived Bruce-4 embryonic stem ( ES ) cells . For removal of the Neo cassette flanked by two FRT sites mice were bred to a FLPe recombinase line ( Ozgene ) resulting in wt/Jac-loxPΔneo/Flp mice . Jacob/Nsmf constitutive ko mice were obtained after breeding wt/Jac-loxPΔneo/Flp mice to a CMV-Cre deleter line ( Ozgene ) for excision in all tissues . In further crossing steps to wt mice the FLPe and Cre transgenes were removed . All breeding steps were performed with mice on C57BL/6J background . For most of the experiments mice from heterozygous breeders were used . Further information can be found in S1 Methods . To compare the litter size of Jacob-deficient breeders with heterozygous and wt breeders four males of each genotype were mated with the corresponding females . The mean litter sizes ( ± SEM ) between genotypes were compared over a period of 15 weeks . For analysis of the Mendelian ratio the genotypes of 573 animals ( offspring of heterozygous breeding pairs , both sexes ) were compared . Body weights of Jacob/Nsmf ko , heterozygous and wt littermates were taken after weaning ( 20–22 days ) and at the age of four month . For behavioral experiments different cohorts of mice were used , as defined in the corresponding paragraphs . However , Rotarod and Open Field test were performed with the same group of mice of both sexes ( n = 12 ko , n = 12 wt , half of each sex ) at the age of 4 month . Secondly , in odor exposure and social interaction test one group of male Jacob/Nsmf ko mice ( n = 16 ) and wt littermates ( n = 15 ) at the age of 4–5 months was used . One week prior to behavior experiment animals were separated and further kept individually . Hippocampi from 8–11 weeks old male mice ( Jacob/Nsmf ko and wt littermates ) were cut using a vibratome ( LeicaVT1000S ) into 350 μm thick slices . Hippocampal slices were incubated for 2h in carbogenated ( 95% O2 , 5% CO2 ) artificial cerebrospinal fluid ( ACSF , containing in mM: 110 NaCl , 2 . 5 KCl , 2 . 5 CaCl2·2H2O , 1 . 5 MgSO4·7H2O , 1 . 24 KH2PO4 , 10 glucose , 27 . 4 NaHCO3 , pH 7 . 3 ) at 31±1°C . fEPSPs were evoked by stimulation of CA1 Schaffer-collateral fibers with biphasic rectangular current pulses ( 200 μs/polarity , frequency 0 . 033Hz ) in a range of 4-5V through ACSF filled glass capillary microelectrodes ( 3–5 MΩ ) . fESPSs were recorded using ACSF filled capillary microelectrodes and amplified by an Extracellular Amplifier ( EXT-023 , npi ) and digitized at a sample frequency of 5 kHz by Digidata 1401plus AD/DA converter ( CED ) . Stimulation strength was adjusted to 40%~50% of the maximum fEPSP-slope values . Late-Long-term potentiation ( L-LTP ) was induced by tetanization consisting of three 1s stimulus trains at 100 Hz with a 6 min inter-train interval . Paired-pulse facilitation was measured by different interpulse interval ( ms ) . Input-output curves relating the fEPSP slopes with the afferent volleys were generated from field recordings in an interface chamber . In our experimental conditions , we could more precisely quantify the afferent volley in this configuration . The general procedure was similar to what was described above . Data are represented as mean ± SEM . Dissociated hippocampal neurons were prepared from P0-P1 Jacob/Nsmf ko and wt mice . Neurons were plated on glass coverslips coated with poly-L-lysine ( Sigma-Aldrich ) at a density of 40 , 000 for immunocytochemistry or 60 , 000 cells per well for transfection , in DMEM medium ( Gibco , Thermo Fisher Scientific ) supplemented with 8% FBS , 1% penicillin/streptomycin . Following attachment , cells were kept in Neurobasal medium ( Gibco ) supplemented with 0 , 5 mM Glutamax , B27 , and 1% penicillin/streptomycin ( all from Gibco ) , at 37°C , 5% CO2 and 95% humidity . Cells were divided into three groups , chronic BDNF treatment ( 100 ng/ml , Tocris ) ( DIV2 and DIV6 ) , acute BDNF treatment for 30 min or no treatment . All cells were fixed at DIV10 or DIV15 in PBS containing 4% paraformaldehyde for 10 min at room temperature . Hippocampal neurons were dissected from E18 Sprague Dawley rats as previously described [2] . Neurons were plated on 18 mm glass coverslips coated with poly-D-lysine ( Sigma-Aldrich ) at a density of 30 , 000 per well ( 12 well plates ) in 1 ml of DMEM medium supplemented with 10% FBS , 0 , 5 mM Glutamine and 1% penicillin/streptomycin ( Gibco ) . After 1 DIV the medium was exchanged to 1 ml of Neurobasal including B27 and 0 , 2 mM Glutamine ( Gibco ) . Cultures were incubated at 37°C , 5% CO2 and 95% humidity . At DIV11 neurons were divided into 4 groups , untreated , treated for 30 min with BDNF ( 100 ng/ml , Tocris ) , AP5 ( 20 μM , Sigma-Aldrich ) or both . After treatment cultures were fixed with 4% paraformaldehyde for 10 min . To dissect the role of BDNF in Bdnf gene transcription in hippocampal neurons from Jacob ko and wt animals we employed a Bdnf-promoter ( exonI and exonII ) driven GFP reporter system ( Bdnf I and II-eGFP , [2 , 61] ) . Dissociated wt and Jacob/Nsmf ko DIV 8 hippocampal neurons were co-transfected with plasmids expressing Bdnf I and II-eGFP and plasmids expressing mRFP under the actin promoter as a volume marker and transfection control using Lipofectamine 2000 ( Invitrogen ) and kept for 48 hr . Following transfection , cells were divided into two groups , treated with BDNF ( 100 ng/ml , Tocris ) or non-treated . Activity of the promoter was estimated by GFP-fluorescence levels upon indicated treatments in neuronal somata by measuring the GFP pixel intensity in the same ROI of maximal projections of two focal planes . P10 control and Jacob/Nsmf ko mice were sacrificed and tissue from CA1 and CA3 regions of the hippocampus was collected in PCR clean tubes ( Eppendorf ) , freshly frozen in liquid N2 and stored at -80°C . Total RNA was isolated ( RNeasy plus mini kit , Qiagen ) . 50 nanograms of RNA were reverse transcribed using random nonamers ( Sigma-Aldrich ) according to the manufacturer’s instructions ( Sensiscript , Qiagen ) . Bdnf exon IV and glyceraldehyde 3-phosphate dehydrogenase ( Gapdh ) mRNA ( as a reference gene ) were amplified using the iScript RT-PCR iQ SYBR Green Supermix ( BIORAD ) in a real-time quantitative PCR ( qPCR ) detection system ( LC480 , Roche ) using the following primers: Bdnf exon IV forward 5’-GCAGCTGCCTTGATGTTTAC-3’ and reverse 5’-CCGTGGACGTTTACTTCTTTC-3’ , and forward Gapdh 5’-TGCTGAGTATGTCGTGGAG-3’ and reverse 5’-GTCTTCTGGGTGGCAGTGAT-3’ . Each sample reaction was run in duplicate and Ct values of the reference genes from the samples were subjected to Grubbs’ outlier test . The relative expression levels were analyzed using the 2-ΔΔCt-method with normalization relative to GAPDH . Data were expressed as mean ± SEM . Two-tailed unpaired Student’s t-test was performed for comparison between two groups . P10 and P45 mice were sacrificed and bilateral hippocampi regions of CA1 and CA3 , were dissected and frozen with liquid N2 followed by storage at -80°C . Tissue was defrosted , scaled and lysed with lysis buffer ( 100 mM PIPES pH = 7 , 500 mM NaCl , 0 . 2% Triton X-100 , 0 . 1% NaN3 2% BSA and Complete protease inhibitor cocktail with EDTA , Roche ) . Hippocampi were sonicated and centrifuged at 16 , 000 g for 30 min at 4°C . 100 μl of supernatant was diluted in 4 volumes of DPBS buffer and acid treated with 10 μl of 1N HCl to decrease pH below 3 . 0 . After 15 min samples were neutralized with 10 μl of 1N NaOH . The BDNF Elisa was performed with BDNF Emax ImmunoAssay System ( Promega ) according to manufacturer’s protocol . Briefly , 96-well plate was coated with 100 μl of anti-BDNF antibody ( 1:1000 ) over night at 4°C . Plates were blocked with 200 μl 1x Promega Block and Sample buffer for 1 hour at room temperature . Following blocking samples were incubated at room temperature together with standard dilutions for 2 hours with shaking ( 400 rpm ) . After washing with TBST , samples were incubated with 100 μl of anti human BDNF polyclonal antibody ( 1:500 ) for 2 hours at room temperature . Following washing , 100 μl of anti-IgY horseradish peroxidase conjugate was added and incubated for 1 hour at room temperature . Plates were emptied again , washed with TBST buffer and developed with 100 μl of TMB One Solution . Reaction was stopped with 100 μl of 1N HCl . Absorbance was measured at 450 nm . Data were expressed as mean ± SEM . Two-tailed unpaired Student’s t-test was performed . Information about histological techniques , measurement of neuron number and fiber density , Golgi-Cox staining , TIMM staining , volumetric analysis of mouse brain by means of manganese-enhanced MRI , confocal laser-scanning and immunocytochemistry can be found in S1 Methods .
|
Kallmann syndrome ( KS ) is a neurodevelopmental disorder that is characterized by a migration deficit of GnRH neurons that fail to migrate into the hypothalamus during embryonic development . The affected patients suffer from idiopathic hypogonadotropic hypogonadism associated with anosmia or hyposmia . Mutations in NSMF ( NMDA Receptor Synaptonuclear Signaling And Neuronal Migration Factor ) have been associated with KS . In previous work we could show that Jacob , the protein encoded by the Nsmf gene , is involved in NMDA receptor synaptonuclear signaling . Jacob operates as a mobile hub that docks NMDA receptor-derived signalosomes to nuclear target sites and thereby plays a role in activity-dependent gene transcription . We found that mice that are deficient for the Nsmf gene do not present phenotypic characteristics related to KS . Instead , these mice exhibit hippocampal dysplasia with a reduced number of synapses and simplification of dendrites , reduced plasticity at CA1 synapses and deficits in hippocampus dependent learning . Brain-derived neurotrophic factor ( BDNF ) induces the nuclear translocation of Jacob in an NMDAR-dependent manner in early development , which results in increased phosphorylation of CREB and enhanced CREB-dependent Bdnf gene transcription . Nsmf knockout mice show reduced hippocampal Bdnf mRNA and protein levels as well as reduced pCREB levels during dendritogenesis . Moreover , BDNF application can rescue the morphological deficits in hippocampal pyramidal neurons devoid of Jacob . Taken together , the data suggest that the absence of Jacob in early development interrupts a positive feedback loop between BDNF signaling , subsequent nuclear import of Jacob , activation of CREB and enhanced Bdnf gene transcription , ultimately leading to hippocampal dysplasia .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"nuclear",
"import",
"medicine",
"and",
"health",
"sciences",
"nervous",
"system",
"cell",
"processes",
"brain",
"vertebrates",
"electrophysiology",
"mice",
"neuroscience",
"animals",
"mammals",
"animal",
"behavior",
"zoology",
"neuronal",
"dendrites",
"animal",
"cells",
"behavior",
"rodents",
"hippocampus",
"olfactory",
"bulb",
"cell",
"biology",
"anatomy",
"synapses",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"neurophysiology",
"organisms"
] |
2016
|
A Jacob/Nsmf Gene Knockout Results in Hippocampal Dysplasia and Impaired BDNF Signaling in Dendritogenesis
|
The developing crossveins of the wing of Drosophila melanogaster are specified by long-range BMP signaling and are especially sensitive to loss of extracellular modulators of BMP signaling such as the Chordin homolog Short gastrulation ( Sog ) . However , the role of the extracellular matrix in BMP signaling and Sog activity in the crossveins has been poorly explored . Using a genetic mosaic screen for mutations that disrupt BMP signaling and posterior crossvein development , we identify Gyc76C , a member of the receptor guanylyl cyclase family that includes mammalian natriuretic peptide receptors . We show that Gyc76C and the soluble cGMP-dependent kinase Foraging , likely linked by cGMP , are necessary for normal refinement and maintenance of long-range BMP signaling in the posterior crossvein . This does not occur through cell-autonomous crosstalk between cGMP and BMP signal transduction , but likely through altered extracellular activity of Sog . We identify a novel pathway leading from Gyc76C to the organization of the wing extracellular matrix by matrix metalloproteinases , and show that both the extracellular matrix and BMP signaling effects are largely mediated by changes in the activity of matrix metalloproteinases . We discuss parallels and differences between this pathway and other examples of cGMP activity in both Drosophila melanogaster and mammalian cells and tissues .
The vein cells that develop from the ectodermal epithelia of the Drosophila melanogaster wing are positioned , elaborated and maintained by a series of well-characterized intercellular signaling pathways . The wing is easily visualized , and specific mutant vein phenotypes have been linked to changes in specific signals , making the wing an ideal tissue for examining signaling mechanisms , for identifying intracellular and extracellular crosstalk between different pathways , and for isolating new pathway components [1–3] . We and others have been using one venation defect , the loss of the posterior crossvein ( PCV ) , to identify and characterize participants in Bone Morphogenetic Protein ( BMP ) signaling . The PCV is formed during the end of the first day of pupal wing development , well after the formation of the longitudinal veins ( LVs , numbered L1-L6 ) ( Fig 1B ) , and requires localized BMP signaling in the PCV region between L4 and L5 [4] . Many of the homozygous viable crossveinless mutants identified in early genetic screens have now been shown to disrupt direct regulators of BMP signaling , especially those that bind BMPs and regulate BMP movement and activity in the extracellular space [5 , 6] . The PCV is especially sensitive to loss of these regulators because of the long range over which signaling must take place , and the role many of these BMP regulators play in the assembly or disassembly of a BMP-carrying “shuttle” . As summarized in Fig 1A , the BMP Decapentaplegic ( Dpp ) is secreted by the pupal LVs , possibly as a heterodimer with the BMP Glass bottom boat ( Gbb ) . This stimulates autocrine and short-range BMP signaling in the LVs that is relatively insensitive to extracellular BMP regulators . However , Dpp and Gbb also signal over a long range by moving into the intervein tissues where the PCV forms [7–9] . In order for this to occur , the secreted BMPs must bind the D . melanogaster Chordin homolog Short gastrulation ( Sog ) and the Twisted gastrulation family member Crossveinless ( Cv , termed here Cv-Tsg2 to avoid confusion with other “Cv” gene names ) . The Sog/Cv-Tsg2 complex facilitates the movement of BMPs from the LVs through the extracellular space , likely by protecting BMPs from binding to cell bound molecules such as their receptors [8–11] . In order to stimulate signaling in the PCV , BMPs must also be freed from the complex . The Tolloid-related protease ( Tlr , also known as Tolkin ) cleaves Sog , lowering its affinity for BMPs , and Tsg family proteins help stimulate this cleavage [12 , 13] . Signaling is further aided in the PCV region by a positive feedback loop , as BMP signaling increases localized expression of the BMP-binding protein Crossveinless 2 ( Cv-2 , recently renamed BMPER in vertebrates ) . Cv-2 also binds Sog [14] ( Olsen , Halbisen , Li and Blair , in preparation ) , cell surface glypicans and the BMP receptor complex , and likely acts as a co-receptor and a transfer protein that frees BMPs from Sog [8 , 15 , 16] . The lipoprotein Crossveinless-d ( Cv-d ) also binds BMPs and glypicans and helps signaling by an unknown mechanism [17] . PCV development takes place in a complex and changing extracellular environment , but while there is some evidence that PCV-specific BMP signaling can be influenced by changes in tissue morphology [18] or loss of the cell-bound glypican heparan sulfate proteoglycans [17] , other aspects of the environment have not been greatly investigated . During the initial stages of BMP signaling in the PCV , at 15–18 hours after pupariation ( AP ) , the dorsal and ventral wing epithelia form a sack that retains only a few dorsal to ventral connections from earlier stages; the inner , basal side of the sack is filled with extracellular matrix ( ECM ) proteins , both diffusely and in laminar aggregates ( Fig 1B ) [19–24] . As BMP signaling in the PCV is maintained and refined , from 18–30 hours AP , increasing numbers of dorsal and ventral epithelial cells adhere , basal to basal , flattening the sack . The veins form as ECM-filled channels between the two epithelia , while in intervein regions scattered pockets of ECM are retained basolaterally between the cells within each epithelium; a small amount of ECM is also retained at the sites of basal-to-basal contact . This changing ECM environment could potentially alter BMP movement , assembly of BMP-containing complexes , and signal reception , as has been demonstrated in other developmental contexts in Drosophila [25–30] . We will here demonstrate the strong influence of the pupal ECM on PCV-specific long-range BMP signaling , through the identification of a previously unknown ECM-regulating pathway in the wing . In a screen we conducted for novel crossveinless mutations on the third chromosome , we found a mutation in the guanylyl cyclase at 76C ( gyc76C ) locus , which encodes one of five transmembrane , receptor class guanylyl cyclases in D . melanogaster [31–33] . Gyc76C has been previously characterized for its role in Semaphorin-mediated axon guidance; Malpighian tubule physiology , and the development of embryonic muscles and salivary glands [34–39] . Like the similar mammalian natriuretic peptide receptors NPR1 and NPR2 [40] , the guanylyl cyclase activity of Gyc76C is likely regulated by secreted peptides [35] , and can act via a variety of downstream cGMP sensors . Our evidence suggests that Gyc76C influences BMP signaling in the pupal wing by changing the activity of the cGMP-dependent kinase Foraging ( For; also known as Dg2 or Pkg24A ) [41] , also a novel role for this kinase . But rather than controlling BMP signal transduction in a cell-autonomous manner , we will provide evidence that Gyc76C and Foraging regulate BMP signaling non-autonomously by dramatically altering the wing ECM during the period of BMP signaling in the PCV . This effect is largely mediated by changing the levels and activity of matrix metalloproteinases ( Mmps ) , especially Drosophila Mmp2 . Genetic interactions suggest that the ECM alterations affect the extracellular mobility and activity of the BMP-binding protein Sog . This provides the first demonstration of Gyc76C and For activity in the developing wing , and the first evidence these proteins can act by affecting Mmp activity . Moreover , our demonstration of in vivo link from a guanylyl cyclase to Mmps and the ECM , and from there to long-range BMP signaling , may have parallels with findings in mammalian cells and tissues . NPR and NO-mediated changes in cGMP activity can on the one hand change matrix metalloproteinase expression secretion and activity ( e . g . [42–47] ) , and on the other change in BMP and TGFβ signaling [48–52]; we will discuss these below .
As many critical regulators of BMP signaling are likely to be required at earlier stages of D . melanogaster development , we screened for novel BMP regulators in the PCV by creating large mitotic recombinant clones homozygous for mutagenized third chromosomes in the posterior , PCV territory of the developing wing . We utilized posteriorly-expressed engrailed ( en ) -Gal4 and UAS- FLPase to induce mitotic recombination between mutagenized FRT-bearing chromosome [53 , 54]; the non-mutagenized FRT chromosomes carried Minute ( M ) mutations that dominantly slow the rate of cell divisions ( RpS17 for the 3R chromosome arm and RpS3 for 3L ) , allowing the homozygous mutagenized clones to divide more quickly and outcompete their M+/M- neighbors [55] . In en-Gal4 / UAS-FLP; FRT82B RpS3Plac92 ubi-GFP / FRT82B wing discs , homozygous wild type ( M+ ) clones , identified by the absence of GFP , fill almost the entire posterior ( Fig 1E ) . Inducing clones homozygous for a recessive mutation in dystrophin ( dys ) using en-Gal4 / UAS-FLP; FRT82B RpS3Plac92 ubi-GFP / FRT82B dysEP3397 reliably generated adults with a partial , “detached” PCV phenotype ( Fig 1F ) , similar to that caused by loss of dys function in the entire wing [56] . We screened 14 , 500 F2 adults; the 9 independent mutant chromosomes we found that reliably disrupted the PCV in large posterior clones were recessive and homozygous lethal . One 3L mutant chromosome gave crossveinless phenotypes over vvlM638 and vvlsep and three 3R mutant chromosomes were lethal over dysEP3397 . The remaining 5 ( Fig 2 ) ( S1 Fig ) complemented these and other candidates known to be required for PCV or vein formation , and complemented each other . In addition , one viable mutant found originally in an F2 male was not caused by third chromosome recombinant clones , but mapped to the X chromosome , and is allelic to ade5 as will be discussed below . Large posterior 3L043 clones result in the complete loss of the PCV in adult wings ( Fig 2E ) . Deletion mapping placed the lethality under Df ( 3L ) Exel9061 , a molecularly-defined deletion [58] that removes part of CG14101 and the coding exons of gyc76C ( Fig 2A ) . 3L043 was lethal over gyc76CKG0373ex33 , a homozygous lethal 8kb genomic deletion that leaves CG14101 intact but removes part of the 5’ UTR of gyc76C [34] ( Fig 2A ) . Sequencing 3L043 DNA identified an A to T transversion within exon 15 of gyc76C , resulting in the missense mutation L635H ( Fig 2B and 2C ) . Loss of gyc76C function mimicked the 3L043 phenotype . Flies carrying small homozygous gyc76CKG0373ex33 clones , generated using heat shock promoter 70 ( hs ) -driven FLPase , often had wings with disrupted PCVs ( Fig 2F ) . Flies carrying larger posterior gyc76CKG0373ex33 clones , generated using en-Gal4 UAS-FLPase and the Minute technique , did not survive to adulthood , but driving expression of a UAS-RNAi constructs directed against gyc76C ( UAS-gyc76C-RNAi , VDRC stocks 3057 and 6552 ) , with either the general disc driver A9-Gal4 or the posterior wing driver hh-Gal4 , caused partial or complete loss of the PCV ( Fig 2G and 2H ) . gyc76CKG0373ex33 clones and hh-Gal4-driven expression of UAS-gyc76C-RNAi also caused occasional morphological defects and wing blistering not seen with large posterior gyc76C3L043 clones ( Fig 2G and 2H ) , suggesting that the gyc76C3L043 allele is hypomorphic . Gyc76C shares the intracellular domain structure of vertebrate NPRs ( Fig 2B ) , including an intracellular protein tyrosine-like kinase ( PTKc ) domain , a putative coiled-coil dimerization domain ( “dimer” in Fig 2B ) , and a guanylyl cyclase catalytic ( CYC ) domain [59] . The L635H missense mutation in gyc76C3L043 alters a residue within the PTKc domain that is conserved in vertebrate NPRs ( Fig 2C ) . However , the Gyc76C PTKc domain , like that of the NPRs , lacks a glutamate that is required to catalyze phosphate transfer and thus is likely kinase dead [60]; this domain is thought to regulate the activity of the guanylyl cyclase domain [39 , 59] . While receptor guanylyl cyclases can increase cGMP levels , vertebrate NPR-A can also act independently of cGMP by directly binding to and altering the activity of TRPC3/C6 Ca2+ channels [61] . Two lines of evidence strongly suggest , however , that Gyc76C is acting in the wing via the production of cGMP . First , we compared the effects of overexpressing wild type and cyclase-dead versions of the Gyc76C . Overexpression of wild type gyc76C in the wing induces ectopic venation ( Fig 2I and 2J ) . In other contexts expressing a form of Gyc76C carrying a D945A mutation within its CYC domain has a dominant negative effect on endogenous Gyc76C cyclase activity , likely through the formation of non-functional homodimers as occurs with a similar mutation in NPR [34 , 62] . Overexpression of UAS-myc-gyc76CD945A using hh-Gal4 or A9-Gal not only failed to induce ectopic venation , but caused crossveinless phenotypes ( Fig 2L ) . This is in marked contrast to the equivalent mutation in NPR-A , which retained its ability to alter TRPC3/C6 channel activity [61] . Second , cGMP can act by stimulating cGMP-dependent protein kinases ( PRKGs ) , and loss of one of these mimicked the gyc76C mutant phenotype . D . melanogaster has three PRKGs: Pkg21D ( also known as Dg1 ) , CG4839 , and For ( also known as Dg2 ) [41 , 63 , 64] . RNAi-mediated knockdown of Pkg21D mimics the loss of Gyc76C in Malpighian tubule function , axon guidance and embryonic muscle and salivary gland formation [35 , 37–39] . However , en-Gal4-driven or A9-Gal4-driven expression of UAS-pkg21D-RNAi ( VDRC 34594 or 34595 ) did not disrupt PCV formation . CG4839MB10509 flies have a Minos transposable element inserted into one of the gene’s coding exons but are homozygous viable with normal wing venation . By contrast , PCVs were lost from the wings of pupae homozygous for the adult lethal fork04703 and for02 alleles . PCVs can normally be visualized from 16 to 32 hours AP using antisera specific to the C-terminal phosphorylated form of the BMP receptor-activated Smad , Mothers against Dpp ( anti-pMad ) , and after 22–24 hours AP by reduced expression of D . melanogaster Serum Response Factor transcription factor ( DSRF , also known as Blistered ) ( Fig 3A and 3B ) [4] . 24 hour AP for homozygotes lacked anti-pMad staining and DSRF downregulation in the PCV ( Fig 3C and 3C’ ) . Removal of for also blocks the effects of Gyc76C overexpression: hh-gal4-induced overexpression of gyc76C induced ectopic anti-pMad staining in pupal wild type or for02/+ wings , especially in a central region of the wing near the normal crossveins ( Fig 3M and 3N ) , but did not do so in for02 homozygotes ( 8/8 cases ) ( Fig 3O ) . We will show below that mutation of for not only mimics the effects of gyc76C knockdown on BMP signaling , but also its effects on the ECM , strongly supporting the involvement of Gyc76C and the PRKG For in a common pathway linked by cGMP . A role for cGMP in crossvein-specific BMP signaling provides an explanation for the crossveinless wings produced by defects in purine synthesis , a part of the “purine syndrome” [65–67] . As noted above , a homozygous viable X chromosome crossveinless mutation found in our screen ( X1 ) maps to and is allelic to the ade5 gene ( S3 Fig ) , which encodes an enzyme with 5-aminoimidazole ribonucleotide carboxylase and 4-[ ( N-succinylamino ) carbonyl]-5-aminoimidazole ribonucleotide synthetase activities , the sixth and seventh steps in purine synthesis [68] . The pupal crossvein defects in ade51 wings were similar to , although milder than , those caused by for mutations ( S3E and S3F Fig ) . We also attempted to manipulate cGMP in the wing using the D . melanogaster cGMP phosphodiesterase PDE6 , which can reduce levels of cGMP ( and cAMP in one assay ) after extraction from S2 cells [69]; overexpression of PDE6 causes a 25% reduction of cGMP levels in Malpighian tubules [70] . However , UAS-PDE6-RNAi driven with hh-Gal4 or en-Gal4 UAS-dcr2 UAS-PDE6-RNAi only rarely produced the ectopic venation expected from increased cGMP , and overexpression of wild type UAS-PDE6 or a mutated form lacking a prenylation site that alters its subcellular localization [71] , did not produce crossveinless wings with either hh-Gal4 , en-Gal4 , or MS1096-Gal4 . While the cGMP reductions caused by PDE might be expected to block PRKG activity , even ubiquitous PDE5/6 overexpression with actin promoter-driven Gal4 failed to reproduce the lethality of for or Pkg21D mutants , and adults appeared normal . PDE activity can be regulated at several levels , and cGMP , PDE and PRKG activities can depend greatly on subcellular localization [72 , 73] . Given our other experimental support for cGMP’s role in the gyc76C and for phenotypes , we think it likely that PDE6 does not cause a large enough cGMP change , in the correct subcellular compartment , to greatly affect For activity . We next investigated the role of the only known ligand for Gyc76C , but found it plays only a weak role in the wing . The VQQ neuropeptide , one of several produced from the Nplp1 peptide precursor protein , can stimulate Gyc76C-dependent cGMP production in S2 cells and Malpighian tubules , although the effects of its removal have not been tested [35] . Nplp1EY11089 is a P element insertion that introduces stop codons into the first coding exon of Nplp1 , 3’ to the signal peptide-coding region needed for secretion , but 5’ to the peptide coding region ( S4A Fig ) . But while this mutation blocked Nplp1 peptide production in the CNS ( S4B and S4C Fig ) , it failed to reproduce the lethality of strong gyc76C mutants , and caused only occasional ectopic branching from the PCV rather than its loss ( S4D and S4E Fig ) . Thus , either there are redundant Gyc76C-stimulating peptides , or Gyc76C has significant activity in the absence of peptide binding . Plexin A-mediated Semaphorin signaling can affect Gyc76C activity in embryonic axons and in vitro [34 , 39] , but we have reduced Plexin A signaling in the wing and found no effects on PCV development ( hh-Gal4 UAS-Plexin A-RNAi ) . The LVs are specified early in wing development by localized EGF-receptor-mediated MAPK activity , well prior to the appearance of the crossveins , but begin to express the BMP Dpp during early pupal stages [1 , 8 , 74] . Anti-pMad provides a measure of BMP signaling immediately downstream of receptor activation; anti-pMad staining appears around both the LVs and the PCV at 15 hours AP; by 18–20 hours the PCV always forms a continuous , gap-less line of pMad staining between L4 and L5 , despite the PCV not expressing Dpp or requiring EGF receptor-mediated MAPK activity until after 24 hours AP [4 , 8] . As in for mutants , knockdown of gyc76C using hh-Gal4-driven or en-Gal4-driven expression of UAS-gyc76C-RNAi , always blocked or created large gaps in anti-pMad staining in the 23–24 hour AP PCV , as shown by comparing pMad levels with those in the adjacent LVs ( Fig 3D and S5D Fig ) . This was accompanied by loss of DSRF downregulation in the PCV between 24 and 28 hours AP , slightly later than the equivalent effect in for mutants ( Fig 3C’ , 3E and 3F , S5D’ Fig ) . However , BMP signaling was still initiated in the PCVs of for mutant or gyc76C knockdown wings , and visible at 20–22 hours AP ( Fig 3H and 3I , S5C Fig ) . In for mutants signaling was often reduced at stages prior to formation of a vein lumen in the PCV region ( S6 Fig ) . The early BMP signaling was more robust after gyc76C knockdown than in for mutants; in fact , en-Gal4-driven knockdown of gyc76C often led to broader anti-pMad staining than in wild type wings at 20 hours AP , in both the PCV and the LVs ( Fig 3G and 3H ) . The broadening of the PCV and adjacent LVs was also apparent in the dorsal epithelium after driving dorsal-specific knockdown using ap-Gal4 UAS-gyc76C-RNAi ( S5A Fig ) and the width of L5 in adult hh-Gal4 UAS-gyc76C-RNAi wings was also significantly greater than in control hh-Gal4 wings ( S7A–S7E Fig ) . Thus , the effects on BMP signaling were complex: Gyc76C suppressed and refined BMP signaling around the LVs and the early PCV , but Gyc76C and For maintained BMP signaling in the older PCV . This is difficult to reconcile with an intracellular , “cell-autonomous” effect of Gyc76C and For on BMP signal transduction , which would be expected to lower pMad levels in all the vein cells . Instead , the effect is quite reminiscent of extracellular alterations in long-range BMP signaling: reducing the BMP shuttling mediated by extracellular BMP-binding proteins like Sog and Cv-Tsg2 can increase short-range signaling near the Dpp-expressing LVs , but decrease long-range signaling from the LVs into the PCV region [10] . As a more rigorous test of cell autonomy , we generated large homozygous gyc76C3L043 or gyc76CKG0373ex33 clones using the Minute technique and hs-FLPase , examining these at 28 hours AP which , because of the slowed development of M-/+ flies , corresponds to approximately 24 hours AP in wild type flies . Clones that encompassed the region of PCV formation on both the dorsal and ventral epithelia could result in the complete or near-complete loss of pMad from the PCV ( S8A , S8A’ , S8C and S8C’ Fig . Effects of additional gyc76C and for mutant clones on PCV development ) . However , individual PCV cells within smaller clones often had pMad levels identical to those in neighboring heterozygotic cells ( Fig 3J and 3K’ , S8B and S8B’ Fig ) . We observed similar non-autonomy in for02 mutant clones in 24 hour AP or older wings ( Fig 3L and 3L’ ) ; for02 clones could even occasionally disrupt PCV formation in neighboring for/+ or +/+ cells ( S8D and S8E Fig ) . These non-autonomous effects are quite similar to those caused by clones lacking the extracellular BMP-binding regulators Sog , Cv-Tsg2 and Cv-2 [10 , 15] . The overexpression of Gyc76C also induced ectopic venation and anti-pMad staining in a non-autonomous fashion . hh-Gal4 expression is limited to the posterior of the wing ( Fig 2K ) , but hh-Gal4-driven expression of UAS-myc-gyc76C resulted in ectopic venation in the anterior compartment of adult wings ( Fig 2J ) , and ectopic pMad anterior to the region of Gyc76C overexpression in pupal wings ( Fig 3M and 3N ) . The clonal analyses above strongly suggest that Gyc76C and For do not regulate BMP signaling in the pupal wing at the level of cell-autonomous signal transduction , but rather influence the extracellular regulation of BMP secretion , movement or reception . We therefore next examined the roles of BMPs and BMP-binding proteins in Gyc76C activity using genetic interactions . In the results that follow , at least 10 wings of each genotype were compared , and results were identical in all of them . First , Gyc76C can act downstream of Dpp expression . Overexpression of UAS-dpp-GFP using an L5-specific Gal4 driver [75] expanded the width of the adult L5 ( S7H Fig ) , but co-expression of UAS-Gyc76C-RNAi in L5 significantly reduced this expansion ( S7I and S7J Fig ) . Gyc76C’s vein-promoting activity also depended on the presence of the secreted BMP binding protein Cv-Tsg2 . Loss of Cv-Tsg2 prevents BMP signaling in the pupal PCV and thus PCV formation in adults [9–11] ( Fig 4A ) . The ectopic venation normally caused by en-Gal4 UAS-gyc76C was blocked in a cv hemizygous background , and the overexpression of gyc76C did not rescue crossvein formation ( Fig 4B and 4C ) . However , altering the levels of any single BMP , BMP regulator or effector did not rescue the defects caused by moderate gyc76C knockdown . The PCV disruption caused by A9-Gal4-driven expression of UAS-gyc76C-RNAi ( Fig 4D ) was not improved by individual co-expression of UAS-gbb-Flag , UAS-tlr , UAS-cv-His , cvEP1349 , UAS-myc-cv-2-V5 , cv-2EP1103 , UAS-sog , UAS-tkv-HA , UAS-punt , UAS-mad-Flag or UAS-medea ( the sole D . melanogaster co-Smad ) . Nonetheless , we found that greatly increasing Sog cleavage could counteract the PCV-disrupting effects . en-Gal4-driven overexpression of the Tlr protease rescued the crossveinless disruption of ade5X1 mutants ( S6G Fig ) . And while Tlr overexpression did not rescue PCV disruption in A9-Gal4 , UAS-gyc76CRNAi wings , expressing an activated form of Tld ( TldA53 ) did ( Fig 4E ) . This suggests that gyc76C knockdown decreases BMP signaling by increasing the affinity of the Sog/Cv-Tsg2 complex for BMPs . The effects of gyc76C overexpression are consistent with this hypothesis . Strong overexpression of UAS-sog with en-Gal4 always blocks PCV formation ( Fig 4G ) ; the BMPs can likely still move as part of the Sog/Cv-Tsg2 complex , but the excess Sog overwhelms the available Tlr and Cv-2 so that BMPs remain sequestered in the complex [8 , 74] . Increasing Sog cleavage with Tlr overexpression can rescue the crossvein defects caused by Sog overexpression [13] , as can overexpression of Cv-2 [8] . en-Gal4-driven expression of UAS-gyc76C also rescued the PCV loss normally caused by UAS-sog expression ( Fig 4H ) . Since Tsgs and Cv-2 can also decrease the BMP-sequestering activity of the Sog/Cv complex , we tested whether overexpression of Cv-Tsg2 or Cv-2 could rescue gyc76C knockdown in combination with each other or with Sog . The PCV disruption caused by A9-Gal4-driven expression of UAS-gyc76C-RNAi ( Fig 4D ) was not improved by co-expression of cvEP1349 , cv-2EP1103 or UAS-sog in any single or pair-wise combination . It was rescued , however , by triple co-expression of UAS-sog , cvEP1349 , and cv-2EP1103 ( Fig 4F ) . This result cannot be explained if gyc76C knockdown simply increased the affinity of the Sog/Cv-Tsg2 complex for BMPs , since adding excess Sog should increase BMP sequestration , not reduce it . Rather , we hypothesize that gyc76C knockdown also reduces the movement of the Sog/Cv-Tsg2 complex into the PCV region . Excess Sog can overcome this defect in diffusion , but only increases BMP signaling in a genetic background ( excess Cv-Tsg2 and Cv-2 ) that frees BMPs from the excess Sog . In summary , our results suggest that Gyc76C knockdown has complex effects on Sog function , both increasing Sog’s affinity for BMPs , but also decreasing the range of Sog movement ( see Discussion ) . The non-autonomous , complex effects of Gyc76C and For on BMP signaling and Sog function are reminiscent of similar effects caused by altering the ECM in different developmental contexts ( see Discussion ) . Moreover , the adult wing blistering caused by very strong loss of Gyc76G activity ( Fig 2H detail ) suggests a failure to properly adhere the two wing epithelia , an effect that can also be caused by altering the wing ECM and its receptors . Gal4-driver overexpression mediated by a UAS-containing EP insertion near the for locus was also reported to induce blistering [76] . We therefore examined the effects of cGMP activity on the levels and distribution of the ECM components Collagen IV using the 6G7 monoclonal antibody , LamininB2 ( LanB2 , also called Lanγ1 ) using anti-LanB2 , and the secreted perlecan heparan sulfate proteoglycan ( HSPG ) Terribly reduced optic lobes ( Trol , previously named l ( 1 ) zw1 ) using anti-Trol and a trol-GFP protein trap . In the normal pupal wing all three of these formed a diffuse ECM with scattered laminar aggregates; the aggregates were especially prominent with the 6G7 anti-Collagen IV . ECM proteins also concentrate in the hemocytes that circulate between the wing epithelia , and anti-LanB2 also stained the apical surfaces of the epithelia . We did not detect gross histological changes in the ECM after gyc76C knockdown prior to 24 hours AP , although the more open , pocket-like architecture of younger pupal wings makes it more difficult to detect ECM organization at this stage . Profound defects appeared , however , around 24–28 hours AP and strengthened from 28–34 hours AP , appearing slightly earlier in for mutants ( Fig 5 and S9G–S9J’ Fig ) . By 24 hours AP the ECM normally fills both the large vein channels and smaller basolateral pockets between cells in intervein regions ( Figs 1B and 5A–5C ) . Since the ECM is prominent in the normal veins , the PCV loss and LV expansion in for mutant wings , or in the posterior hh-Gal4 UAS-gyc76C-RNAi wings , caused parallel losses or gains of vein ECM ( Fig 5D–5E”‘ , 5I and 5I’ ) . However , we also observed histological changes in the organization of the ECM that were not simply reflections of altered venation . First , the basolateral pockets of intervein ECM , although initially normal , were progressively depleted in the posterior after posterior gyc76C knockdown ( Fig 5D–5E”‘ ) , or throughout the wing in for homozygotes ( Fig 5I and 5I’ ) ( for time course in S9G–S9J’ Fig ) . As this occurred the vein ECM became broader and more diffuse , and was often retained in abnormal vein-like blobs near the site of the PCV . The broadened ECM accumulation near veins did not strictly correlate with altered vein specification: the extremely broad L5 ECM caused by posterior gyc76C knockdown extended into regions lacking vein markers such as heightened pMad or reduced DSRF , and the vein-like blobs near the normal PCV site were retained after molecular markers of PCV development vanished ( S9A and S9A’ Fig ) . In hh-Gal4 UAS-gyc76C-RNAi wings diffuse Trol-GFP was especially strong in L5 and the PCV-like blobs ( Fig 5D”‘ ) . 6G7 anti-Collagen IV staining showed an abnormally high accumulation of laminar aggregates in the posterior of hh-Gal4 UAS-gyc76C-RNAi wings ( Fig 5E”‘ and S9J Fig ) . These gross organizational defects were preceded by a more subtle change: at 24 hours AP abnormally large intracellular vesicles appeared in the interveins ( Fig 5D”‘ and 5F–5H’ , S9B–S9F Fig ) . These containing Trol-GFP and were likely endocytotic , since many co-localized with the late endocytic marker Rab9 . YFP ( Fig 5H and 5H’ ) , although Trol-GFP only rarely overlapped the late endocytic marker Rab7 , and did not significantly overlap the early endocytic marker Rab5 , or the recycling vesicle marker Rab11 ( S9C–S9E’ Fig ) . Intriguingly , anti-Trol staining did not accumulate in the GFP-containing vesicles ( Fig 5G and 5G’ ) or co-localize with Rab9 . YFP ( S9B and S9B’ Fig ) . Since the GFP tag in Trol-GFP is inserted into the N terminal domain II , while the anti-Trol was produced against the C-terminal domain V [77] , it is possible that the GFP represents uptake of an abnormal cleavage product of Trol-GFP . We will present results below suggesting that the vesicles are a cellular reaction to breakdown of the ECM . The ECM phenotypes are not a general result of changes in BMP signaling or a crossveinless condition . cv null mutations block most or all PCV BMP signaling during the initial stages of PCV formation and broaden signaling in the LVs [9–11] , but did not obviously alter wing ECM outside the missing PCV ( Fig 5K and S9L Fig ) . Nor did we detect ECM defects outside the missing PCV in crossveinless wings mutant for the Rho-Rac GAP Crossveinless c ( cv-c1 ) or expressing en-Gal4 UAS-dys-RNAi ( Fig 5J , S9M and S9N Fig ) . The effects of cGMP activity on the ECM were not strictly cell autonomous . Moderate-sized gyc76C3L043 or for homozygous clones did not deplete the intervein pockets or increase accumulation of Collagen IV aggregates , even where dorsal and ventral clones overlapped ( S9K and S9K’ Fig ) . Posterior overexpression of gyc76C with hh-Gal4 also altered wing ECM in a non-autonomous fashion . While ECM in the LVs appeared normal , ECM in the intervein pockets was fainter and more diffuse in the posterior; this effect extended up to L3 , well anterior of the region of hh-Gal4 expression ( Fig 5L and 5L’ ) . Since reductions in Gyc76C or For activity disrupt several components of the wing ECM , we next searched for effects on components known to organize or modify the ECM . Posterior gyc76C knockdown did not cause posterior-wide changes in the levels of ECM receptors such as the glypican Dlp , Dystroglycan , or the integrins Mys , Mew and If , nor alter expression of the mys expression regulator Delilah [78] . Nor could we detect posterior-wide changes in the BMP receptor Thickveins or the vein-width regulator Notch . Changes were limited to those caused by altered venation , and only for those proteins whose levels are normally different in vein and intervein ( S10 Fig ) . The depletion of ECM from the interveins and the diffuse ECM appearance the veins , next suggested the involvement of the extracellular matrix metalloproteinases ( Mmps ) . This is also consistent with the possible appearance of a Trol-GFP cleavage product noted above , as vertebrate perlecan can be cleaved by Mmps [79] . There are two D . melanogaster Mmps: Mmp2 , which is predicted to be GPI-linked to the cell surface , and Mmp1 , which is diffusible [80–82] . An engineered Mmp2::GFP produced by the endogenous Mmp2 locus [83] is normally expressed in a slightly patchy pattern in intracellular structures and the cell cortex , more weakly in vein cells and more strongly near the wing hinge; the cytoplasmic GFP is stronger apically ( Fig 6C and 6D’ ) , without any strong anterior-posterior bias ( 11 of 11 23–25 hour AP wings ) . After posterior hh-Gal4-driven knockdown of Gyc76C , posterior Mmp-2 levels were higher and more uniform , an effect especially noticeable in more apical focal planes , and after a summing projection of all cross-sections along the proximo-distal ( X ) ( Fig 6A and 6B’ ) . Posterior apical Mmp2::GFP was increased by 25% or greater over anterior in 13 out of 16 23–25 hour AP wings , and the increase was significant in a comparison of all experimental and control wings ( Fig 6E ) . Those wings lacking the effect may reflect variation in its timing . Anti-Mmp1 staining in normal pupal wings was largely extracellular and concentrated in the diffuse ECM of veins and intervein pockets . hh-Gal4 UAS-gyc76C always increased the posterior levels of Mmp1 in the ECM of both veins , especially L5 and PCV-like blobs , and intervein pockets beginning at 24 hours AP ( details in Fig 6F–6I’; low magnification in S9S–S9U Fig; anti-Mmp1 in control wings in S9W and S9X Fig ) This increased staining was retained in the veins but was transient in the intervein pockets: as the intervein pockets became depleted of ECM beginning at 27–28 hours AP , Mmp1 levels in the pockets also decreased , although increased Mmp1 was sometime still observed in pockets partially depleted of anti-Trol ( Fig 6H and 6H’ ) . The increased Mmp1 is likely due to changes in Mmp1 secretion or accumulation rather than transcription , as an Mmp1-lacZ enhancer reporter that reproduces Mmp1 expression in other contexts [84] was not obviously altered in hh-Gal4 UAS-gyc76C-RNAi wings ( S9Q and S9R Fig ) . Given the strong association of Mmp1 with wing ECM , Mmp1’s abnormal accumulation in the abnormal ECM of gyc76C knockdown wings could be both a result and a cause . To test the role of Mmp activity in the ECM and BMP signaling changes caused by reduced Gyc76C activity , we first overexpressed the D . melanogaster member of the diffusible Tissue inhibitor of metalloproteases ( Timp ) protein family , which can inhibit both Mmp1 and Mmp2 activities [85 , 86] . hh-Gal4 UAS-gyc76C-RNAi UAS-Timp pupae did not produce adults and pupal wings were foreshortened , likely due to Timp’s effects on wing disc eversion [87] . Nonetheless , Timp greatly improved PCV development in hh-Gal4 UAS-gyc76C-RNAi pupal wings , forming normal or nearly normal PCVs as assessed by heightened pMad or reduced DSRF pMad in 9/9 28 hour AP or older wings ( Fig 6J ) . hh-Gal4-driven expression of UAS-Timp in a gyc76C knockdown background ( Fig 6K ) also increased the ECM , but largely in an in an abnormal proximal clump likely caused by excess inhibition of Mmp activity . Intriguingly , BMP signaling also expanded in the proximal wing adjacent to the excess ECM ( Fig 6J ) . We next tested the roles of the Mmps individually using RNAi lines with proven efficacy [84] . UAS-Mmp1-RNAi did not improve the defects of hh-Gal4 UAS-gyc76C-RNAi wings , but UAS-Mmp2-RNAi did . hh-Gal4 UAS-gyc76C-RNAi UAS-Mmp2-RNAi larvae and pupae were occasionally unhealthy-appearing , generating fragile wings with poor morphology and development , but the ECM in those that were healthy appeared almost normal: the width and strength of the staining around posterior veins and the strength and number of intervein ECM pockets were nearly normal ( Fig 6O and 6P ) . The abnormally large intracellular vesicles normally found in the interveins of gyc76C knockdown wings were also greatly reduced ( S9Y and S9Z Fig ) , suggesting that these are a cellular reaction to Mmp-induced breakdown of the ECM . While 6G7 anti-Collagen IV staining still showed the abnormally high numbers of Collagen IV aggregates observed in hh-Gal4 UAS-gyc76C-RNAi wings ( Fig 6P’ ) , it should be noted that hh-Gal4 UAS-Mmp2-RNAi in otherwise wild type wings also increases abnormal anti-Collagen IV aggregates in the posterior ( S9O and S9P Fig ) . BMP signaling in the PCVs was also largely rescued , as assessed by heightened pMad or reduced DSRF in 15/17 24–32 hour AP wings ( Fig 6L–6N ) . These results strongly suggest that most of the ECM and BMP signaling defects caused by reduced Gyc76C activity are caused by increased Mmp activity . We next asked whether the ECM and BMP signaling defects were linked , or were independent effects of Mmp activity , by looking at the effects of manipulated the ECM directly . It is difficult to remove most ECM components from the wing: null mutants are lethal or cause morphological abnormalities at earlier stages , and mosaic techniques cannot be used for those ECM components , like Trol and Collagen IV , that are likely transported into the wing via hemolymph and hemocytes [88 , 89] . Instead , we tested the ability of overexpressed ECM to rescue the effects of gyc76C knockdown , choosing Trol because hh-Gal4-driven expression of UAS-trol in a wild type background led to few abnormalities beyond a slight broadening of the PCV ( Fig 6Q ) . hh-Gal4-driven expression of both UAS-trol and UAS-gyc76C-RNAi increased the blistering sometimes observed in adult wings after gyc76C knockdown , but partially or wholly rescued PCV formation in adult ( Fig 6R ) and pupal wings ( 6/7 28 hours AP or older; Fig 6S ) . Unlike UAS-Mmp2-RNAi , the very strong overexpression caused by UAS-trol did not obviously improve the other ECM components in the posterior of hh-Gal4 UAS-gyc76C-RNAi pupal wings , but did cause an abnormal accumulation of diffuse 6G7 anti-CgIV staining between L4 and L5 proximal to the PCV ( Fig 6T and 6T’ ) , leaving open the possibility that the rescue was mediated by reorganization of the ECM , rather than by Trol directly .
It is noteworthy that the cGMP activity mediated by NPR or nitric oxide signaling can change also Mmp gene expression , secretion or activation in many different mammalian cells and tissues ( e . g . [42–47] ) . Both positive and negative effects have been noted , depending on the cells , the context , and the specific Mmp . Given the strong role of the ECM in cell-cell signaling , the contribution of cGMP-mediated changes in Mmp activity to extracellular signaling may be significant . There is also precedent for cGMP activity specifically affecting BMP and TGFβ signaling in mammals . cGMP-dependent kinase activity increases BMP signaling in C2C12 cells , and this effect has been suggested to underlie some of the effects of nitric oxide-induced cGMP on BMP-dependent pulmonary arterial hypertension [50 , 52] . Conversely , atrial natriuretic peptide stimulates the guanylyl cyclase activities of NPR1 and NPR2 and can inhibit TGFβ activity in myofibroblasts; this inhibition has been suggested to underlie the opposing roles of atrial natriuretic peptide and TGFβ during hypoxia-induced remodeling of the pulmonary vasculature [48 , 49 , 51] . However , unlike the pathway we observed in the fly wing , these mammalian effects are thought to be mediated by the intracellular modulation of signal transduction , with cGMP-dependent kinases altering BMP receptor activity or the phosphorylation and nuclear accumulation of receptor-activated Smads [50 , 51] . Nonetheless , it remains possible that there are additional layers of regulation mediated through extracellular effects , underscoring the importance of testing cell autonomy . Aside from its role in adult Malpighian tubule physiology , Gyc76C was previously shown to have three developmental effects: in the embryo it regulates the repulsive axon guidance mediated by Semaphorin 1A and Plexin A [34 , 39] , the proper formation and arrangement of somatic muscles [37] , and lumen formation in the salivary gland [38] . All these may have links to the ECM . Loss of gyc76C from embryonic muscles affects the distribution and vesicular accumulation of the βintegrin Mys [37] , and reduces laminins and the integrin regulator Talin in the salivary gland [38] . The axon defects likely involve a physical interaction between Gyc76C and semaphorin receptors that affects cGMP levels [39]; nonetheless , gyc76C mutant axon defects are very similar to those caused by loss of the perlecan Trol [90] . The parallels between the different contexts of Gyc76C action are not exact , however . First , only the wing phenotype has been linked to a change in Mmp activity . Second , unlike the muscle phenotype , the wing phenotype is not accompanied by any obvious changes in integrin levels or distribution , beyond those caused by altered venation ( S10 Fig ) . Finally , most gyc76C mutant phenotypes are reproduced by loss of the Pkg21D ( Dg1 ) cytoplasmic cGMP-dependent kinase [35 , 37 , 38 , 91] , instead of For ( Dg2 , Pkg24A ) as found in the wing , and thus may be mediated by different kinase targets . For has been largely analyzed for behavioral mutant phenotypes [92] , and the overlap between Pkg21D and For targets is unknown . While many targets have been identified for the two mammalian cGMP-dependent kinases , PRKG1 ( which exists in alpha and beta isoforms ) and PRKG2 , it is not clear if either of these is functionally equivalent to For . One of the protein isoforms generated by the for locus has a putative protein interaction/dimerization motif with slight similarity to the N-terminal binding/dimerization domains of alpha and beta PRKG1 , but all three For isoforms have long N-terminal regions that are lacking from PRKG1 and PRKG2 . In fact , a recent study suggested that For is instead functionally equivalent to PRKG2: Like PRKG2 , For can stimulate phosphorylation of FOXO , and is localized to cell membranes in vitro [93] . But For apparently lacks the canonical myristoylation site that is thought to account for the membrane localization and thus much of the target specificity of PRKG2 . FOXO remains the only identified For target , and foxo null mutants are viable with normal wings [94] . The loss of long range BMP signaling in the PCV region caused by knockdown of gyc76C can , like the ECM , be largely rescued by knockdown of Mmp2 . Two results suggest that it is the alteration to the ECM that affects long-range BMP signaling , rather than some independent effect of Mmp2 . First , the BMP signaling defects caused by gyc76C knockdown were rescued by directly manipulating the ECM through the overexpression of the perlecan Trol . Second , when Mmp activity is inhibited by overexpression of the diffusible Mmp inhibitor TIMP , this not only rescued the PCV BMP signaling defects caused by gyc76C knockdown , but also led to ectopic BMP signaling , not throughout the region of TIMP expression , but only in those regions with abnormal accumulation of ECM . The Mmp2-mediated changes in the ECM likely affect long-range BMP signaling by altering the activity of extracellular BMP-binding proteins , particularly Sog . The BMPs Dpp and Gbb produced in the LVs bind Sog and Cv-Tsg2 , shuttle into the PCV region , and are released there by Tlr-mediated cleavage of Sog and transfer to Cv-2 and the receptors ( see Introduction and Fig 1B ) . Our genetic interaction experiments suggest that knockdown of gyc76C both increases Sog’s affinity for BMPs and reduces the movement of the Sog/Cv-Tsg2/BMP complex into the crossvein region . Collagen IV provides the best-studied example for how the ECM might affect Sog activity . The two D . melanogaster collagen IV chains regulate BMP signaling in other contexts , and they bind both Sog and the BMP Dpp [25–27] . Results suggest that collagen IV helps assemble and release a Dpp/Sog/Tsg shuttling complex , and also recruits the Tld protease that cleaves Sog cleavage and releases Dpp for signaling [25 , 27 , 28] . D . melanogaster Mmp1 can cleave vertebrate Collagen IV [80] . Since reduced Gyc76C and For activity increases abnormal Collagen IV aggregates throughout the wing and diffuse Collagen IV in the veins , we hypothesize that these Collagen IV changes both foster the assembly or stability of Sog/Cv-Tsg2/BMP complexes and tether them to the ECM , favoring the sequestration of BMPs in the complex and reducing thelong-range movement of the complex into the region of the PCV ( Fig 7 ) . While few other D . melanogaster Mmp targets have been identified , it is likely that Mmp1 and Mmp2 share the broad specificity of their mammalian counterparts [80 , 81] , so other ECM components , known or unknown , might be involved . For instance , vertebrate Perlecan and can be cleaved by Mmps [79] . Trol regulates BMP signaling in other D . melanogaster contexts [29 , 30] , and Trol overexpression rescue gyc76C knockdown’s effects on BMP signaling . But while null trol alleles are lethal before pupal stages , normal PCVs were formed in viable and even adult lethal alleles like trolG0023 , and actin-Gal 4-driven expression of trol-RNAi using any of four different trol-RNAi lines did not alter adult wing venation . Loss of the D . melanogaster laminin B chain shared by all laminin trimers strongly disrupts wing venation [24] , and a zebrafish laminin mutation can reduce BMP signaling [95] . Finally , it was recently shown that Dlp , one of the two D . melanogaster glypicans , can be removed from the cell surface by Mmp2 [96] . While gyc76C knockdown did not detectably alter anti-Dlp staining in the pupal wing ( S10G and S10H Fig ) , it is noteworthy that Dlp and the second glypican Dally are required non-autonomously for BMP signaling in the PCV and that they bind BMPs and other BMP-binding proteins [15 , 17] .
The following were generated from Bloomington Drosophila Stock Center stocks , unless otherwise indicated . A9-Gal4 w y w; ap-Gal4 UAS-GFP/CyO y w; en-Gal4 y w; en-Gal4 UAS-FLP UAS-GFP; hh-Gal4/TM6 , Tb hh-Gal4 UAS-GFP/TM6 , Tb ( recombinant generated in lab ) L5-Gal4 ( 3 . 7KX-lacZ/UAS ) kindly provided by J . de Celis [75] . L5-Gal4 UAS-dpp-GFP recombinant generated in lab . y w; UAS-FLP y w hs-FLP; ubi-mRFP . nls FRT40A /CyO y w; FRT2A y w hs-Flp; hs-GFP RpS174 [also known as M ( 3 ) i55] FRT2A/TM3 , Sb ( kindly provided by G . Struhl ) y w; FRT82B y w , FRT82B , RpS3Plac92[also known as M ( 3 ) w or M ( 3 ) 95A] ubi-GFP/TM6B , Tb FRT82B dysEP3397/TM6 , Tb ( FRT recombinant kindly provided by D . Olson ) gyc76CKG03723ex33/TM3 , Sb , UAS-myc-gyc76C and UAS-myc-gyc76D945A , kindly provided by A . Kolodkin [34] . gyc76CKG03723ex33 FRT2A , en-Gal4 UAS-myc-gyc76C and hh-Gal4 UAS-myc-gyc76C recombinants were lab-generated . Gyc76C overexpression experiments used a second chromosome UAS-myc-gyc76C , except for those of Fig 3M and 3N which used a lab-generated for02; hh-Gal4 UAS-myc-gyc76C /CyO-TM6 , Tb stock . y cv1 v ade51 fl/FM6 , kindly provided by D . Clark [68] . Crossveinless males were crossed to Df ( 1 ) ED7165/FM7h; the deficiency covers ade5 but not cv . y w; Nplp1EY11089 for02/CyO , Tb , forK04703/CyO , Tb for02 FRT40A/CyO , recombinant generated in lab . UAS-PDE6-RNAi , UAS-PDE6 and UAS-PDE6C1128S kindly provided by Dr . S . Davies [70 , 71] . cv43 [11] UAS-dpp-GFP , UAS-tlr-HA , UAS-tldA53 , UAS-sog-HA , cv70 , kindly provided by M . O’Connor [10 , 13] . cv-2P ( EP ) 1103 ( Szeged Stock Center ) [4] cvP ( EP ) 1349; cv-2P ( EP ) 1103 UAS-sog-HA stock generated in lab . cv-c1 cvP ( EP ) 1349 y w; P ( UASp-YFP . Rab9 ) [97] CG4839MB10509 UAS-gyc76C-RNAi , UAS-dys-RNAi and UAS-Pkg24-RNAi lines were from the VDRC . UAS-gyc76C RNAi; hh-Gal4/CyO-TM6 , Tb and en-Gal4 UAS-dys-RNAi stocks generated in lab . trol-GFP ( P{PTT-un1}trolG00022 ) [98] from Kyoto DGRC . Mmp2::GFP/CyO , kindly provided by J . Sun [83] UAS-TIMP , UAS-Mmp1-RNAi , UAS-Mmp2-RNAi and Mmp1-lacZ kindly provided by D . Bohmann [84] . P ( GSV2 ) trolGS7407 ( UAS-trol ) kindly provided by J . Pastor-Pareja [88] . Bloomington deficiency kits and molecularly defined deletions and P element w+ insertions for mapping . 50 or more males were transferred to empty vials , allowed to dehydrate for 30 minutes , and then transferred overnight to new bottles contained filter paper soaked in a solution of 24mM EMS , 10Mm Tris pH 7 . 5 and 1% sucrose . The males were then transferred to dry tubes containing damp filter paper for 30 minutes before being crossed to 50 females for two days . Mapping was as described in the Results , except that 3L043 lethality was initially mapped using Bloomington deficiency kits DK3 , 3L and DK3 , 3R ( Bloomington ) , while ade5X1 was initially mapped using meiotic recombination relative to molecularly mapped w+ P element insertions [99] . The 3L044 missense mutation gyc76CL635H was identified by Sanger sequencing ( University of Wisconsin-Madison Biotechnology Center ) using primers 5’ ATGGATTGTTTGCCACCAACAG Fwd , and 5’ TCAAACAATCGGAATGAAGCTG Rev . Wing disc and pupal wing dissection , fixation and staining were as described previously [4 , 8]; identical methods were used for larval CNS staining . Images were captured on BioRad and Olympus FV1000 confocal microscopes . Projections , cross-sections and quantifications were generated from Z-series images using ImageJ . Concentrations and sources of primary antibodies were: 1:2000 rabbit anti-phosphoSmad3 ( Epitomics ) ; 1:500–1000 mouse anti-DSRF ( Cold Spring Harbor Laboratory Antibody Facility ) ; 1:50 mouse anti-Engrailed 4D9 , mouse anti-FasIII or mouse anti-Mmp1 ( Developmental Studies Hybridoma Bank ) ; 1:500 rabbit anti-MTYamide [100] kindly provided by L . Schoofs; 1:500 rabbit anti-Trol [77] kindly provided by S . Baumgartner; 1:1000 rabbit anti-Drosophila Lamininγ1 ( LanB2 ) ( ABCAM ) ; 1:50 mouse 6G7 anti-Collagen IV [22] kindly provided by J . Palka; 1:500 rabbit anti-GFP ( MBL ) ; 1:500 rabbit anti-Rab5 ( ABCAM ) ; 1:3000 rabbit anti-Rab7 [101] kindly provided by A . Nakamura; 1:1000 rabbit anti-Rab11 [102] kindly provided by D . Ready; 1:50 rabbit anti-Dei [78] kindly provided by A . Salzberg . Secondary staining used Jackson ImmunoResearch fluorescently-tagged ( FITC , RITC , Cy2 , Cy3 , or Cy5 ) Min X anti-mouse IgG ( H+L ) or anti-rabbit IgG ( H+L ) antisera .
|
Signaling between cells regulates many processes , including the choices cells make between different fates during development and regeneration , and misregulation of such signaling underlies many human pathologies . To understand how such signals control developmental decisions , it is necessary to elucidate both how cells regulate and respond to different levels of signaling , and how different types of signals combine and regulate each other . We have used genetic screening in the fruitfly Drosophila melanogaster to identify mutations that reduce or eliminate signals carried by Bone Morphogenetic Proteins ( BMPs ) , and show that BMP signaling is sensitive Gyc76C , a peptide receptor that stimulates the production of cGMP in cells . We identify downstream intracellular effectors of this cGMP activity , but provide evidence that the effects on the BMP pathway are not mediated at the intracellular level , but rather through cGMP’s effects upon the extracellular matrix and matrix-remodeling proteinases , which in turn affects the activity of extracellular BMP-binding proteins . We discuss differences and parallels with other examples of cGMP activity in Drosophila melanogaster and mammals .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
The Gyc76C Receptor Guanylyl Cyclase and the Foraging cGMP-Dependent Kinase Regulate Extracellular Matrix Organization and BMP Signaling in the Developing Wing of Drosophila melanogaster
|
Marker frequency analysis of the Escherichia coli recB mutant chromosome has revealed a deficit of DNA in a specific zone of the terminus , centred on the dif/TerC region . Using fluorescence microscopy of a marked chromosomal site , we show that the dif region is lost after replication completion , at the time of cell division , in one daughter cell only , and that the phenomenon is transmitted to progeny . Analysis by marker frequency and microscopy shows that the position of DNA loss is not defined by the replication fork merging point since it still occurs in the dif/TerC region when the replication fork trap is displaced in strains harbouring ectopic Ter sites . Terminus DNA loss in the recB mutant is also independent of dimer resolution by XerCD at dif and of Topo IV action close to dif . It occurs in the terminus region , at the point of inversion of the GC skew , which is also the point of convergence of specific sequence motifs like KOPS and Chi sites , regardless of whether the convergence of GC skew is at dif ( wild-type ) or a newly created sequence . In the absence of FtsK-driven DNA translocation , terminus DNA loss is less precisely targeted to the KOPS convergence sequence , but occurs at a similar frequency and follows the same pattern as in FtsK+ cells . Importantly , using ftsIts , ftsAts division mutants and cephalexin treated cells , we show that DNA loss of the dif region in the recB mutant is decreased by the inactivation of cell division . We propose that it results from septum-induced chromosome breakage , and largely contributes to the low viability of the recB mutant .
Most bacteria have a circular chromosome on which replication is initiated at single origin oriC and proceeds bi-directionally on the two replichores until forks meet in the terminus , opposite to oriC . The chromosome terminus is a particularly active region where several important processes take place: replication termination , chromosome dimer resolution and last steps of chromosome segregation ( Fig 1A ) . In E . coli , replication is arrested by the presence of sites called Ter that are bound by a specific protein Tus ( reviewed in [1 , 2] ) . Ter-Tus complexes allow replication to proceed in one direction only and thus create a replication fork trap in which replication forks enter but from which they infrequently exit . This system allows right and left replichores to be replicated principally in a clockwise and anti-clockwise direction , respectively , ensuring that replication is mainly co-directional with transcription [3–6] . 2D-gel analyses allowed the visualisation of replication forks arrested at TerC , and to a lesser extent at TerA and TerB [7] . In most bacterial species a site called dif , which is located opposite to oriC on the circular chromosome , is acted upon by the XerC/XerD site-specific recombination complex to resolve chromosome dimers ( reviewed in [8 , 9] ) . dif is also the site of inversion of the GC skew on the chromosome [2 , 10] and the site of orientation inversion of biologically active motifs such as Chi ( crossover hotspot instigator ) and KOPS ( FtsK-orienting polar sequences [11 , 12] reviewed in [13] ) . KOPS ( GGGNAGGG ) are used by the septum protein FtsK to orient the translocation of chromosomes to daughter cells , and the convergence of these sequences at dif makes it the last segregated chromosomal sequence in slow growing cells [14] . Although KOPS are present all around the chromosome , FtsK is particularly active in the 400 kb region , centred on dif [15] . Translocation of chromosomes by FtsK is arrested by encounter with the XerCD-dif complex [16 , 17]; FtsK then activates this complex to trigger chromosome dimer resolution at dif [18–20] . Finally , FtsK was proposed to displace the terminus-specific DNA-bound protein MatP [14] , a protein that organizes and condenses the 780 kb Ter macrodomain by binding specifically to short DNA sequences called matS [21–23] . In this manuscript we call “terminus” the chromosome region opposite to oriC , centred on the point of inversion of the GC skew , regardless of the position of replication forks merging . The septum forms at mid-cell by the assembly of several proteins in a defined order ( reviewed in [24 , 25] ) . Early proteins are FtsZ and its regulators , which include FtsA . Formation of the Z-ring is essential for the recruitment of late proteins , including FtsK and FtsI . FtsK is a bi-functional protein , its N-terminal domain is essential for cell division and is anchored in the septum; it is separated by a linker from the C-terminal domain , a cytoplasmic ATPase non-essential for viability and responsible for ( i ) DNA translocation in a direction imposed by KOPS , and ( ii ) activation of chromosome dimer resolution by interaction with XerCD ( reviewed in [26 , 27] ) . FtsI is a septum peptidase essential for constriction ( reviewed in [28 , 29] ) . The terminus region was reported to be a preferential region of genetic instability [30–33] . However , hyper-recombination in the terminus region was dependent on replication termination at Ter sites , or perturbation of the dimer resolution system XerCD/dif , or perturbation of FtsK-mediated chromosome segregation , and it occurred in a small subpopulation of cells ( at most 1% ) [30–33] . More recently , a limited region of the terminus was reported to be amplified in certain mutant contexts ( notably in a recG mutant ) and a nearly identical region was lost in recB mutants [34 , 35] . RecBCD is the enzyme that initiates recombinational repair of DNA double-strand breaks ( DSB ) in E . coli . This enzyme specifically recognizes DNA double-strand ends , unwinds and cleaves dsDNA via its coupled DNA helicase and exonuclease ( exo V ) activities , and when it encounters a Chi site it loads RecA on single-stranded DNA ( ssDNA ) [36–38] . Loss of terminus DNA in recB mutants was proposed to result from the formation of dsDNA ends by erroneous merging of replication forks leading to a transient over-replicated intermediate [35] . According to the proposed model , in the RecBCD+ context over-replicated dsDNA ends would be appropriately degraded by RecBCD , restoring intact chromosomes , while in the recB mutant extensive DNA degradation from these dsDNA ends by various single-stranded DNA exonucleases would cause DNA loss [35] . The amplification of terminal DNA in a recG mutant has been examined in detail [6 , 34 , 39 , 40] , while the DNA loss in the recB mutant has been less extensively explored , and the model of merging forks has not been directly tested experimentally [35 , 41] . Here we use Marker Frequency Analysis ( MFA ) and live-cell fluorescence microscopy to further characterize this phenomenon . As previously proposed [34 , 35] , we consider that loss of terminus DNA in the recB mutant results from the degradation of unrepaired DNA double-stranded ends , but we show that it is independent of the position of replication termination , which argues against the model of merging forks . In search for an alternative source of chromosome breakage , we show that terminus DNA loss in the recB mutant occurs during cell division and requires septum formation . We propose that chromosome breakage in the E . coli terminus region is septum-induced damage . In addition , we observed weak Tus-dependent DNA loss at Ter sites , which was only detected when division was prevented by mutation or when replication terminates at an ectopic Ter .
Marker frequency analysis ( MFA ) of the chromosome of wild-type and recB mutant cells in exponential growth confirmed a deficit of DNA reads in the terminus region of the chromosome in the absence of RecB ( Fig 2A , S1 Fig ) . To allow a direct comparison of MFA and microscopy results , all experiments were done in minimal medium ( M9 ) . This DNA loss is centred on the dif-TerC region when cells are grown exponentially in M9 glucose ( Fig 2A ) , as previously reported for cells grown in LB [34 , 35] . We have developed a live-cell microscopy approach to confirm that this phenomenon , which is observed by MFA in a population of growing cells , in fact occurs in a sub-population of individual cells and we have quantified this sub-population . Strains constitutively expressing the yGFP-ParBpMT1 protein from a chromosomal-borne gene and carrying parSpMT1 sites inserted at three different locations were used . Binding of the yGFP-ParB protein to its parS target site allows the visualisation of each chromosome parS sequence as a fluorescent focus . Three strain backgrounds carrying each a different parS sites were used . These parS sites were ydeV::parSpMT1 located between dif and TerC , 10 kb away from each site , yoaC::parSpMT1 located about 300 kb away from dif on the left replichore and ycdN::parSpMT1 located about 500 kb away from dif on the right replichore ( [42] , S1 Table , Fig 1B ) . Cells grown in exponential phase in M9 glucose medium were observed by fluorescence microscopy . In a wild-type context , nearly all cells showed foci and the proportion of cells with two foci increased with the distance from dif , as previously reported ( S2 Table ) [14 , 43] . In the recB mutant , the proportion of cells without any focus was much higher than in RecB+ cells: the recB mutant with ydeV::parSpMT1 ( the locus between dif and TerC ) showed 32% of cells with no focus and the control parS sites located 300 or 500 kb from dif showed 7–8% of cells with no focus ( Table 1 , S2 Table ) . These results argue that in a recB context about one third of cells have lost the dif-TerC region specifically , in agreement with the results of MFA experiments [34 , 35] ( Fig 2A ) . To better characterize this chromosomal DNA loss , the dynamic behaviour of foci was tracked by time lapse microscopy of recB cells growing on M9 glucose agarose pads , as described in Materials and Methods . As shown in Fig 2F and S1 and S2 Videos , ydeV::parSpMT1 foci were lost with the following characteristics: 1 ) the foci disappeared concomitantly with cell division , most often at the site of septum formation , and in one of the two daughter cells only ( yellow stars ) , no loss at any other time point was observed , 2 ) the loss occurred after duplication of this region , since most of the time two foci were clearly visible at earlier time points ( white arrows ) , 3 ) after cell division , the daughter cell that had lost a focus stopped growing and did not divide , whereas the cell that retained a focus divided again , and produced a focus-less cell at each generation after the first event ( yellow stars ) . Although the production of focus-less cells was asymmetrical and hereditary , we did not observe any bias toward the old or new pole . We called the first division that produces one focus-containing cell and one focus-free cell “the initial event” and calculated that these represented 17 . 7% of cell divisions ( not counting the “secondary events” that follow , Table 1 ) . Because divisions that produced focus-less cells also produced a focus-containing cell , the proportion of focus-less cells in the population was expected to be one half of the proportion of divisions that produced them in the absence of transmission of the phenomenon to progeny . In contrast to this , the proportion of focus-less cells ( 32% ) was higher than the proportion of divisions that produced them ( 17 . 7% ) , in agreement with the transmission of the phenomenon to following generations ( Table 1 ) . Nearly 75% of the initial events were transmitted to the progeny for as many generations as we could see ( up to five ) . In addition , in about 17% of the cases transmission was interrupted for one generation ( i . e . the cell that retained a focus produced two focus-containing cells , one propagated normally and the other one resumed the production of one focus-less cell at each generation ) . Although there is clearly some transmission in spite of the interruption , this second category of events was observed in all recB mutants and is not counted as transmitted in Table 1 . RecBCD has two activities , a recombinase activity that requires RecB and RecC but not RecD ( helicase and RecA loading activities ) , and an exonuclease activity , called exoV , which degrades linear dsDNA and is catalyzed by the entire RecBCD complex [36 , 38] . DNA degradation by exoV occurs in the absence of RecA or in the absence of Chi sites . recD mutant cells are recombination proficient but do not degrade dsDNA , and it was previously reported that terminus DNA loss is not detected by MFA in the recD mutant [35] . Accordingly , analysis of ydeV-parSpMT1 foci showed that the recD single mutant behaved like wild-type ( Table 1 ) , therefore we confirm that the presence of RecBC prevents DNA loss regardless of exoV activity . In contrast , the recC mutant that lacks both RecBC and RecBCD complexes and thus both recombination and exoV activities , behaved as the recB mutant ( Table 1 , S1 Fig ) . DNA loss was previously shown by MFA to occur in the dif-TerC region of recB mutants in two different E . coli genetic backgrounds: MG1655 [34] , in which replication terminates primarily at TerC and 4 to 5 times less often at TerA [7] , and in W3110 [35] . W3110 carries a large inversion between rrnD and rrnE around the replication origin , which enlarges the right replichore and shortens the left one by about 220 kb [44 , 45] ( Fig 1A ) . As a consequence of this inversion , the closest replication terminator from oriC in this context is TerA and not TerC , and therefore replication is expected to terminate at TerA more often than at TerC . Nonetheless , the position of the peak of DNA loss was the same in W3110 as previously reported in MG1655 [35] . This surprising observation prompted us to directly measure the influence of the position of replication termination on DNA loss in the dif/TerC region . For that purpose , we first compared recB and recB tus mutants ( in which Ter sites are non-functional ) by MFA ( Fig 2B , S1 Fig ) and by snap-shot fluorescence microscopy of ydeV::parSpMT1/yGFP-ParBpMT1 foci ( Table 1 ) . Inactivation of tus did not prevent DNA loss in the dif-TerC region detected by MFA ( Fig 2B ) and did not modify the percentage of focus-less cells for the ydeV::parSpMT1 locus close to dif , and for the yoaC::parSpMT1control locus ( Table 1 ) . Note that the ratio of reads in tus recB over the tus mutant increased in a large Ter region compared to the rest of the chromosome ( Fig 2B ) . We do not know the reasons for this phenomenon , but the existence of a mixed population partially masks DNA loss at dif/TerC in the MFA experiment . Time lapse experiments showed that the loss of ydeV::parSpMT1 foci followed the same scheme in tus recB as previously observed in Tus+ recB cells , i . e . loss of focus in one daughter cell , at the septum , at the time of division , and transmission of this defect to the progeny ( S3 Video ) , with a similar percentage of initial events ( Table 1 , 15 . 8% vs 17 . 7% ) , and a high level of events transmitted to progeny ( 87% ) . Replication was previously shown by 2D gels to terminate mainly at TerC in wild-type cells [7] , but our result in the tus mutant suggests that most of the loss of DNA in the dif-TerC region occurs independently of forks merging at TerC . Because in the tus mutant forks may still merge in the dif region that is opposite to the origin , we constructed a strain in which the clockwise replication forks are prevented from reaching TerC by the introduction of an additional TerB site that arrests replication prematurely , 29 kb downstream of TerA ( pspE::TerB , TerB* in Fig 1A and 1B ) . In the TerB* strain a new replication fork trap is created between TerA and TerB* and the dif site is mainly replicated by the counter-clockwise fork , instead of the clockwise fork in the wild-type strain . Fluorescence microscopy showed that loss of ydeV::parSpMT1 foci ( close to dif ) was increased to 48% in the strain containing TerB* , while loss of control yoaC::parSpMT1 and ydvN::parSpMT1 loci was unchanged ( Table 1 , S2 Table ) . As expected , the increase from 32% in recB to 48% in TerB* recB cells was Tus-dependent , as we counted 35% of ydeV::parSpMT1 focus-less cells in the TerB* recB tus mutant ( Table 1 , S2 Table ) . Time lapse experiments showed that in TerB* recB cells ydeV::parSpMT1 foci were still lost in one daughter cell , at the septum , at the time of division , with a transmission of this defect to the progeny ( Fig 2G , Table 1 ) , and the proportion of original divisions that yielded the first focus-less cells in an inheritable manner was increased from 17 . 7% to 25% ( Table 1 ) . Finally , measuring loss of control yoaC::parSpMT1 foci ( Table 1 ) and MFA analysis showed that the position of the peak of DNA loss at dif was not affected by the creation of this new replication fork trap , away from TerC ( Fig 2C , S2 Fig ) . Nevertheless , the MFA experiment also revealed a new Tus-dependent peak of DNA loss at TerB* , weaker than the DNA loss at dif ( Fig 2C and 2D , S2 Fig ) . It was previously proposed that DNA double-strand ends , target for RecBCD , were formed in the terminus region of the recB mutant by erroneous merging of replication forks , and that degradation of these unrepaired DNA double-stranded ends by the combined action of helicases and single-stranded exonucleases was the cause of terminus DNA loss [34 , 35] . Since in the presence of the additional TerB* site , replication forks are expected to merge at this site or at TerA , and are therefore unlikely to merge at TerC , the strong DNA loss that we observed in the dif/TerC region of the TerB* recB mutant ( Fig 2C , Table 1 ) cannot result from replication fork merging . We also propose that DNA loss results from DNA degradation by the combined action of helicases and single-stranded exonucleases , but propose that the dsDNA ends on which these enzymes act are produced by a DNA DSB occurring in dif/TerC region of the recB mutant chromosome , regardless of the position of replication termination . In addition , our results also show that displacing replication termination to TerB* creates a new hotspot of DNA loss , weaker than the dif/TerC hotspot , at the new replication termination site . The observation that the position of DNA loss in the dif-TerC region is independent of the position of replication fork merging raises the possibility that it might be determined by dif rather than TerC . dif is the site of chromosome dimers resolution , a XerCD- and FtsK- dependent reaction [18–20] . We tested a possible role of dimer resolution in DNA loss by inactivating xerC or removing the dif site . In RecB+ cells these mutations increased the proportion of ydeV::parSpMT1 focus-less cells from less than 1% to about 15% ( Table 1 , S2 Table ) and , accordingly , a weak DNA loss of the dif region could be detected by MFA ( S3 Fig ) . Time lapse experiments showed that focus-loss in the dif or xerC single mutants results from breakage of both chromosomes at the time of cell division ( Fig 3E , S4 Video ) ; Abnormal pattern of cell division in microcolonies of dif and xer mutants was previously observed and was proposed to result from breakage of chromosome dimers by septum closure , which was called guillotining [46 , 47] . 40–42% of cells lacked ydeV::parSpMT1 in xerC recB or dif recB mutants ( Table 1 ) . We propose that this higher level of focus-less cells compared to the recB single mutant results from a combination of broken dimers and septum-induced breaks ( about 50% of dimers are RecB-independent , [48] ) . Accordingly , time lapse microscopy confirmed that some focus-less cells result from the concomitant loss of both ydeV::parSpMT1 foci in the two daughter cells at the time of division ( presumably dimer breakage ) , whereas most of them result from the transmitted , asymmetric loss of one focus in one daughter cell at the time of division ( Fig 3E ) . Interestingly , in dif or xer cells that contain a dimer , cell division was delayed for more than an hour before we observed cell separation and focus-loss ( S4 Video ) , while cell division of the recB mutant was not delayed upon focus-loss in one daughter cell compared to cells that do not lose foci . The proportion of initial events in the dif recB mutant was 17 . 4% , similar to the recB single mutant , confirming that division-dependent loss of ydeV::parSpMT1 foci is independent from dimer resolution , and most of these events were transmitted to progeny ( Table 1 ) . Furthermore , the weaker loss of the control yoaC::parSpMT1 and ycdN::parSpMT1 foci compared to loss of the dif proximal ydeV::parSpMT1 site ( Table 1 , S2 Table ) , and MFA analyses of xerC and xerC recB mutant ( Fig 3A , S3 Fig ) confirmed that the loss of DNA remains centred on dif in the absence of dimer resolution . Two dif deletions were tested: one that lacks only the dif site , and one that also inactivates the adjacent hipA locus . HipA is a toxin that blocks growth by inactivating translation and is counteracted by the short-lived anti-toxin HipB ( [49] and ref therein ) . In the absence of both dif and HipA , the peak of DNA loss observed in MFA experiments in a recB mutant context was deeper and larger than in the recB single mutant ( Fig 3B , S3 Fig ) . Accordingly , in microscopy snapshot experiments the proportion of cells that lack the ydeV::parSpMT1 or the control yoaC::parSpMT1 focus was much increased , to 65% and 40% respectively ( Table 1 , S2 Table ) . Time lapse experiments showed that in the absence of HipA , focus-less cells grew and divided for several generations , which increased the proportion of these cells in the population , and presumably allowed the degradation of more and more chromosomal DNA with generations ( Fig 3F ) . This result showed that in recB single mutants the growth arrest of focus-less cells results from the degradation of the hipAB locus after chromosome breakage , which causes the accumulation of active toxin HipA . This phenomenon was previously described after breakage of chromosome dimers at the septum in a dif mutant [46] . It confirms genetically that chromosomes lacking the ydeV::parSpMT1 site originally conserve all genes required for growth and cell division , and that chromosome degradation is a slow process ( essential genes are absent from the terminus region , [47 , 50] ) . If DNA loss results from DSBs occurring at the peak of loss followed by nearly symmetrical DNA degradation by single-stranded nucleases , why are these DSBs introduced in the dif region even in the absence of this site ? The dif site is the last chromosome locus segregated in daughter cells because it is the site of convergence of the KOPS sequences , which are used by the FtsK protein to segregate replicated chromosomes to daughter cells ( [14 , 27] , Fig 1A ) . KOPS sites are present all around the chromosome but FtsK is mainly active in a 400 kb region approximating the one of decreased reads in the recB mutant [15] . We tested a putative role of FtsK in the localisation of terminus breaks with the use of a ftsKATPase mutant , in which ftsK carries a nucleotide substitution that specifically inactivates the ATPase activity and thus prevents DNA translocation without affecting DNA binding . We first analyzed ftsKATPase RecB+ cells by microscopy . Quantification of focus loss showed that the proportion of focus-less cells was increased compared to wild-type cells , particularly for the ydeV::parSpMT1 site located next to the dif locus ( from 0 . 6% to about 20% , Table 1 , S2 Table ) . This DNA loss presumably resulted mainly from a lack of dimer resolution in the absence of FtsK translocation activity . Inactivation of recB in the ftsKATPase mutant led to a large increase in focus-less cells ( nearly 55% of cells contain no ydeV-parSpMT1 focus and 14% contain no yoaC-parSpMT1 focus ) . Similar results were observed when FtsK translocation was inactivated by the deletion the entire protein C-terminal domain ( Table 1 , S2 Table ) . Time lapse microscopy showed two kinds of events leading to focus loss in the ftsK recB mutant context . In 15 . 5% of the divisions , one ydeV-parSpMT1 focus was lost at the septum , at the division time , in one daughter cell only , with a transmission of this defect to the progeny ( Table 1 , S5 Video ) . This result shows that the events occurring in the recB single mutant also occur in ftsKATPase recB cells , with a similar frequency ( Table 1 ) . In addition , the two daughter cells lost the ydeV-parSpMT1 foci during 12% of the divisions , presumably owing to dimer breakage ( S5 Video ) , and other types of focus loss could be observed , which presumably result from the segregation defect and account for the high percentage of focus-less cells in the ftsKΔCTer recB mutant population ( S6 and S7 Videos ) . In contrast with the dif and xerC mutants , DNA loss around dif was not detected by MFA in the ftsKATPase single mutant , and rather a weak DNA amplification was visible in the terminus region ( S3 Fig ) . Since microscopy results show a loss of the dif region in 20% of ftsKATPase and ftsKΔCTer mutants , this amplification reflects the existence of a mixed population of cells , some that lose the dif/TerC region as observed by microscopy , and some that amplify it and mask the loss in the MFA experiments . DNA degradation around the dif locus was observed by MFA in the ftsKATPase recB mutant ( Fig 3C , S3 Fig ) , but a larger DNA region was degraded than in the recB single mutant ( compare Fig 3C and Fig 2A ) . Furthermore , the MFA did not show the deep loss expected from the microscopy results . This loss could be masked if DNA amplification occurs in a subset of cells , as detected by MFA in single ftsKATPase mutant . We conclude that the DNA translocation activity of FtsK plays an important role in the sharp targeting of DNA loss to the dif region in the recB mutant , leading to a wider distribution of DNA loss in the absence of the FtsK C-terminal domain or ATPase activity . Nevertheless , DNA loss and therefore DNA breaks still occur specifically in the dif/TerC chromosome region when DNA translocation by FtsK is inactivated . To know whether in wild-type cells RecBCD acts in the dif/TerC region , we investigated RecA binding by ChIP followed by qPCR of sequences upstream and downstream of the first Chi site on each side of the dif/TerC region ( Fig 4 , RecBCD loaded at a DSB in the dif/TerC region unwinds DNA toward the origin , until it encounters properly oriented Chi sites at which it loads RecA ) . As previously reported [51] we detected a weak increase of RecA ChIP downstream of Chi when cells were grown in LB , but we did not detect any increase in cells grown in M9 glucose ( Fig 4 ) . Similar results were obtained in cells that over-express RecA owing to a mutation in the recA gene SOS operator ( Fig 4 ) . We conclude that in minimal medium , DNA breakage in the dif/TerC region does not occur in RecB+ cells , and thus only occurs in recB mutants . To address the question of the origin of chromosome breaks in the terminus region of a recB mutant two enzymes that cleave DNA were tested , Topoisomerase IV ( Topo IV ) and endonuclease I ( Endo I ) . Topo IV , encoded by the parC and parE genes , catalyzes the decatenation of daughter chromosomes after replication . Topo IV interacts with both XerC and FtsK , its decatenation activity is stimulated by its interaction with FtsK in vitro and a hotspot of activity was detected in vivo close to dif , which is dependent on its interaction with XerCD [52–55] . We hypothesized that if catenated chromosomes persist in the path of septum closure , an erroneous action of Topo IV during decatenation could be responsible for chromosome breakage . Because the inactivation of Topo IV by a ts mutation prevents cell division , and cell division is required for the DSBs studied here ( see below ) , the effects of a parE10ts mutation were tested by MFA at 37°C , where Topo IV is impaired but cell division is not affected enough to prevent cell growth [56] . As shown in S4 Fig , the parE10ts mutation did not affect DNA loss in the dif region at this semi-permissive temperature . Furthermore , interaction of Topo IV with XerCD is required to target its action close to dif [55] , therefore the lack of effect of xerC or dif inactivation described above argues against a direct action of Topo IV to break DNA next to dif . This was confirmed by using the observation that the Topo IV hotspot next to dif is abolished by over-production of the C-terminal region of ParC from a plasmid , ( parC-CTD plasmid , [55] ) : the presence of this plasmid did not affect focus loss ( Table 1 , S4 Fig ) . Therefore , the division-dependent DSBs studied here do not result from an erroneous action of Topo IV at dif . The most abundant endonuclease in E . coli is the periplasmic enzyme Endo 1 . We hypothesized that a leak of Endo 1 from the periplasm to the cytoplasm during division might cause cleavage of one chromosome in the terminus region . However , the inactivation of the endA gene encoding Endo 1 did not affect the proportion of focus-less cells in a recB context ( Table 1 , S2 Table ) . The enzyme ( s ) that introduces the DSBs responsible for DNA loss remain unidentified . Because time lapse experiments showed that the ydeV::parSpMT1 focus was often lost at the septum and always concomitantly with cell division , we tested whether septum formation plays a role in the loss of the dif region . We used three different conditions that affect cell division: ftsAts and ftsIts thermosensitive mutants , which block an early and a late step of divisome assembly at 42°C , respectively , and cephalexin , a drug that prevents the action of FtsI ( reviewed in [24 , 25] ) . Because blocking division produced cells that were highly elongated and difficult to analyse by microscopy , the loss of terminus DNA was examined by MFA . The ftsAts recB mutant was compared at 30°C and after 2 hours of incubation at 42°C , to block division ( Fig 5A , S5 Fig ) . Loss of DNA centred on dif was observed at 30°C where FtsA is active , but was much weaker at 42°C . A similar result was obtained with the ftsIts recB mutant ( Fig 5B , S5 Fig ) . Comparison of a cephalexin-treated with an untreated recB mutant showed that cephalexin also prevented DNA loss at dif ( compare S5 with S1 Fig ) . Therefore , results in all three cases indicated that loss of DNA in the dif region of a recB mutant is decreased when septum assembly is prevented . Control experiments showed that DNA loss was similar in a recB single mutant at 37°C and at 42°C ( S5 Fig ) . Ratios of recB over fts recB mutants grown at 42°C revealed two phenomena ( Fig 5C and 5D ) : ( i ) they confirmed that the loss of reads centred on dif/TerC is specific to dividing cells , ( ii ) they revealed 5–10% more reads in the TerA/TerD region in dividing versus non-dividing cells . The latter observation suggests that in a recB context blocking cell division causes a slight loss of TerA/TerD sequences . When results for the fts recB cells grown at 30°C and 42°C were compared this increase was not observed ( ftsAts ) or weak at TerD ( ftsIts ) , presumably because cell division is partially affected in these mutants at the permissive temperature ( compare Fig 5A and 5B with 5C and 5D ) . Comparisons of RecB+ and recB mutants in ftsIts ( or ftsAts ) contexts at 42°C , revealed two regions of DNA loss caused by recB inactivation , weaker than in dividing cells and centred on TerA and on the dif/TerC region ( Fig 5E and 5F ) . The effects of recB inactivation in ftsIts tus and ftsAts tus mutants were analyzed ( Fig 5G and 5H ) . The small peak in the TerA region disappeared in the tus context , showing that it results from replication arrest at TerA ( compare Fig 5G and 5H with Fig 5E and 5F , S6 Fig ) . Accordingly , in the ftsIts context the weak recB-dependent DNA loss in the TerB/TerC region was displaced to the dif region when tus was inactivated . The peak of DNA loss at dif in fts tus recB mutants is weaker than in dividing cells ( compare Fig 5G and 5H with Fig 2B and 2D ) , accounting for the difference between dividing and non-dividing cells shown above ( Fig 5A , 5B , 5C and 5D ) . We conclude from these experiments that ( i ) DNA loss around dif in the recB mutant is decreased by inactivating cell division ( Fig 5A–5D ) , and ( ii ) weaker peaks of DNA loss that require the Tus protein can be observed at TerA in non-dividing recB mutants ( Fig 5E–5H ) . To further test whether the site of convergence of GC skew determines the localisation of DNA loss , we constructed mutants in which a new GC skew convergence zone was created in the terminus region . First , we used a strain where the entire terminus region from 1 379 810 to 1 617 226 was deleted ( ΔLC3-R111 strain , Fig 1C ) . This 237 kb deletion removes half of the DNA region degraded in the recB single mutant including dif , hipA and TerC . It defines a new 102 kb replication fork trap between TerA and TerB and creates a new GC skew converging zone at the junction , next to which we inserted a parS site ( pspE::parSpMT1 ) . Because the ΔLC3-R111 mutant lacks the dif site , it showed 18% of focus-less cells . In the ΔLC3-R111 recB mutants 49% of cells were devoid of a pspE::parSpMT1 focus ( Table 2 ) . Time lapse analyses showed that the loss of a pspE::parSpMT1 focus in the ΔLC3-R111 recB resulted mostly from loss of one focus in one daughter cell at the time of division as in the original recB mutant ( Fig 6A ) ; we counted 19% initial events and 77% of them were transmitted to progeny ( Table 2 ) . Focus loss also occurred at a lower frequency in both daughter cells at the time of division , presumably resulting from dimer breakage ( Fig 6A ) . The ΔLC3-R111 deletion removes dif and hipA , but shortens the region devoid of essential genes that can be degraded without preventing cell propagation . Accordingly , we observed one or two divisions of focus-less cells owing to the absence of hipAB . For unknown reasons the MFA analysis of the ΔLC3-R111 chromosome ( S7 Fig ) , showed a breakpoint in the read copy number around TerA , which was not detected in the recB mutant ( Fig 7A ) . The ratio of reads in recB mutant over RecB+ cells is affected by this breakpoint , and in Fig 7A we present directly the MFA result of the ΔLC3-R111 recB mutant . The peak of DNA loss measured by MFA was located at the new junction of the chromosome arms , about 65 kb from TerB ( Fig 7A ) . Similarly to the lack of effect of tus inactivation on DNA loss in the recB mutant ( Table 1 ) , inactivating tus in the ΔLC3-R111 recB mutant did not prevent DNA loss at the GC skew convergence point ( Table 2 ) . For unknown reasons , inactivation of the DNA translocation activity of FtsK in the ΔLC3-R111 mutant led to an amplification centred on the midpoint between TerA/D and TerF and the breakpoint between TerA and TerB was not detectable ( ΔLC3-R111 ftsKΔCTer S7 Fig ) . However , DNA loss occurred in ΔLC3-R111 ftsKΔCTer recB as in ΔLC3-R111 recB , and the inactivation of FtsK translocase slightly widened the maximum point of DNA loss , to the entire 105 kb region between TerB and TerA ( Fig 7B ) , These results show that the position of division-induced DSBs is determined by the point of GC skew convergence , in a way that is independent of the sequence of this junction , and is more precisely targeted to the KOPS convergence point in the presence of the FtsK translocation activity . We then created a new GC skew convergence zone by inverting a region of the terminus . In the InvT3 mutant , a ~175 kb sequence is inverted on the right chromosome arm , which does not contain any Ter site and starts about 34 kb from dif ( Fig 1B ) . In this strain the main GC skew convergence zone is moved 209 kb to the left of dif , and the dif position becomes a minor convergence zone with on its left only 34 kb of DNA in the original orientation . InvT3 and InvT3 recB strains were compared by MFA ( Fig 7C , S8 Fig ) . Inactivation of recB in InvT3 created a new degraded region corresponding to the entire inverted sequence , but no peak of DNA loss at the new convergence zone . Importantly , the main DNA degradation peak in the dif region was still present ( Fig 7C , S8 Fig ) . Microscopy analysis confirmed a specific loss of the dif region by showing that the proportion of ydeV::parSpMT1 focus-less cells was similar in InvT3 recB and recB mutants ( 38% and 32% respectively , Table 1 , S2 Table; the additional focus-less cells observed in InvT3 RecB+ could result from a perturbation of segregation because of KOPS inversion , causing irreparable damage ) . The inactivation of the ATPase function of ftsK enlarged the degraded region but the maximum of DNA loss was still in the dif region ( Fig 7D ) . Comparison of FtsK+ and ftsK mutant MFA in the recB context suggests that FtsK-mediated translocation slightly protects the inverted region from degradation . It is also interesting to note that in the absence of the FtsK translocation activity , loss of the reads in the 1380–1554 kb region in the recB mutant was higher when this sequence was inverted than when it is in the original orientation ( compare Fig 7D with Fig 3C ) . This observation suggests the existence of a system other than FtsK able to detect the sequence orientation . Nevertheless , these results also indicate that a 175 kb GC skew convergence zone is not sufficient to create a division-induced DSB . In the InvT2 mutant a 150 kb region encompassing the TerA and TerD sites and located 209 kb from dif is inverted ( Fig 1B ) . Clockwise replication forks are expected to be arrested in the TerA-TerD region in this mutant , and the replication fork trap is moved between the inverted TerA and TerE . Accordingly , the MFA profile of the InvT2 mutant shows that replication forks meet in the TerA-TerD region , ~250–300 kb away from dif ( S9 Fig ) . As for the LC3-R111 mutant , direct results of the InvT2 MFA are shown in Fig 7E because the breakpoint of read copy number around TerA affects the ratio of RecB+ versus recB mutant reads ( Fig 7 , S9 Fig ) . Two peaks of DNA degradation were clearly detected by MFA in the InvT2 recB mutant: the one centred at dif observed in all recB mutants and a new one coincident with the inverted TerA site ( Fig 7E , S9 Fig ) . The inactivation of tus suppressed the TerA-associated degradation but not DNA loss at dif , and allowed the detection of some DNA degradation associated with the inversion region , as in InvT3 ( Fig 7F , S9 Fig ) . The observation that the InvT2 inversion did not affect inheritable division-dependent focus loss was confirmed by microscopy , as the proportion of ydeV::parSpMT1 focus-less cells increased in InvT2 from 8% to 42% upon recB inactivation ( Table 1 , Fig 6B ) . The observation of Tus-dependent DNA loss at TerA confirms that DNA breakage occurs at an artificially introduced Ter site that creates a new replication fork trap , as observed with TerB* ( Fig 2C and 2D ) . In addition , these results confirm that division-induced DSBs in the dif region are not affected by the creation of a new GC skew convergence zone , as observed with InvT3 . Furthermore , in InvT2 as in InvT3 the new DNA convergence zone does not show a peak of DNA loss , in contrast with ΔLC3-R111 , but the number of reads in the whole inverted region is lower than when this sequence is not inverted ( compare Fig 7C with Fig 2A and Fig 7F with Fig 2B ) .
Septum-induced breakage was previously reported in xer and dif mutants , in which chromosome dimers are not resolved to monomers and remain in the path of the closing septum [46]; as expected we observed dimer breakage in our experiments , which occurs specifically in mutants affected for dimer resolution ( xer , dif , ftsK ) and is characterized by a loss of ydeV-parSpMT1 foci in both daughter cells along with a significant delay in cell division . Dimer breakage during septum formation was called guillotining , a term that does not describe precisely the molecular events leading to DNA DSBs . If we assume that chromosome dimer breakage results from physical tension associated with the pulling of two linked chromosomes during segregation , then the breakage of one chromosome observed in recB cells implies that this chromosome is broken as a consequence of being attached in the terminus region while the origin is gradually pulled towards the daughter cell . This attachment could be a covalent link with the other daughter chromosome after replication completion , or a strong binding to a septum protein . It is unlikely that this link is topological , since DNA loss is unaffected in conditions that perturb Topo IV action . It is noticeable that breakage occurs without any delay in cell division , in contrast with dimer breakage . Alternatively , breakage could be enzymatic , but the nature of the nuclease remains mysterious . Importantly , we did not detect any focus loss in the recB mutant at any other time point in the cell cycle than cell division , and our measures of replication speed based on the MFA results show that the recB mutation does not affect replication progression ( the ratio of recB versus wild-type reads is constant all along the chromosome except at the terminus and equal to 1 in Fig 2A ) . This is in agreement with the recently published results , where authors using flow-cytometry analysis concluded that absence of RecB does not affect chromosome replication speed [57] . Hence , we propose that the main source of chromosome breakage in the recB mutant grown in M9 is not replication fork impediments but rather division-induced breaks in the terminus region of the chromosome . Following chromosome breakage , degradation of the DNA double-stranded ends by exonucleases is responsible for DNA loss . This step was postulated but never demonstrated [34 , 35] , and so far the formal possibility of under-replication of the dif region being responsible for the low number of reads observed in MFA experiments could not be excluded . In time lapse experiments , we most often see two ydeV-parSpMT1 foci before one of them disappears , which ascertains for the first time the presumed assumption that DNA loss does not result from a lack of replication but from DNA degradation of a replicated chromosome . Furthermore , the dif hipA and the ΔLC3-R111 mutants behaved as expected i . e . , in the absence of HipA-induced cell death , the broken chromosomes are slowly degraded and cells with a broken chromosome propagate until degradation by exonucleases reaches essential genes . DNA loss occurs at the site of GC skew convergence , and is observed at the new GC skew convergence zone in the strain that carries a large terminus deletion , confirming that the phenomenon is not DNA sequence specific . In addition , DNA loss is not affected by replication orientation , which progresses across dif in the clockwise direction in the majority of wild-type cells [7] , but not when replication is arrested prior to dif by an ectopic TerB* site ( pspE::TerB ) or by inverting TerA and TerD ( InvT2 ) . We have shown that division-induced chromosome breakage is independent of any specific DNA sequence . These observations support a model in which the chromosome terminus region is somehow specifically and precisely positioned in the path of the division machinery . This positioning is more centred on the KOPS convergence zone when FtsK translocation is active , but remains centred on dif in FtsK mutants , and therefore relies on a so far unknown process . Division-induced chromosome breakage occurs in a sub-population of dividing recB cells when the positioning is inappropriately controlled . In addition to causing breakage of one chromosome , the improper processing of the terminus leaves a mark on the intact chromosome , which is responsible for the transmission of the defect to the next generation . In addition to the septum-induced DNA DSBs described above , we observed a Tus-dependent loss of reads , suggesting DNA breakage in specific recB mutant conditions: ( i ) at ectopic or inverted Ter sites ( pspE::TerB Fig 2C and 2D , TerA in InvT2 Fig 7E and 7F ) and ( ii ) in the TerA-TerD region of cell division mutants ( ftsAts and ftsIts at 42°C , cephalexin treated cells , Fig 5 ) . The loss of reads around Ter sites is symmetrical , indicating that DNA breakage does not occur after blockage of the first fork that reaches Ter ( only the replicated side of Ter would then be degraded ) . Previous studies showed that forks blocked at ectopic Ter sites are stable , and that DNA double-strand ends are formed at such Ter sites upon arrival of a second round of replication behind the first blocked one , by rear-ending , but in these previous mutant strains fork-merging was prevented [58 , 59] . The Tus-Ter specific DNA breaks observed in the present work could therefore result from abnormal replication forks merging at Ter sites . However , although our MFA experiments are only semi-quantitative , Tus-dependent DNA loss at Ter sites seems to occur in a lower proportion of cells than division-induced breaks , with a weaker peak of DNA loss than the peak of division-induced DNA loss . Furthermore , the absence of DNA loss at TerA in wild-type cells shows that Tus-Ter induced breaks do not occur at a natural Ter site in cells that divide normally . The observation that blocking cell division triggers Tus-dependent DNA loss at TerA suggests an unknown link between replication fork merging and cell division , a link which would also be perturbed in dividing cells by arresting forks at an ectopic Ter site . Our observation suggests that replication termination at TerC , or forks merging at other sequences than Ter sites , is the most favourable condition during normal cell division and any change in this arrangement leads to loss of DNA at the new active Ter site . Further work will be needed to understand how Ter/Tus dependent DNA DSBs are made , and whether a common mechanism is involved upon division blockage at the natural TerA site and in dividing cells at ectopic Ter sites . In conclusion , we have shown that DNA degradation in the GC skew convergence region occurs in a subpopulation of recB growing cells . The reaction is transmitted to progeny and is strongly decreased when cell division is prevented . It is targeted to the KOPS merging zone by the translocase activity of FtsK , and occurs in a broader chromosome terminus region in mutants that lack this activity . Since our time-lapse experiments did not show any growth defect or loss of focus in the recB mutant at any other time than cell division , we propose that division-induced DNA breakage could be responsible for the decreased viability of recB cells under normal laboratory growth conditions . These findings open new fields of investigation in search for the molecular mechanism responsible for this reaction .
All E . coli strains are derivatives of MG1655 . Strains and plasmids are described in S1 Table . MM is M9 [60] supplemented with 0 . 4% glucose . Standard transformation and P1 transduction procedures were as described [60] . pspE::TerB-CmR , endA::KanR , araC::parBpMT1-CmR , araC::parBpMT1-ApraR , ydeV::parSpMT1-ApraR mutations were constructed by gene replacement ( recombineering ) as described in [61] , using DY330 [62] . All other strains were constructed by P1 transduction . All mutations introduced by P1 transduction were checked by PCR and all new mutations constructed by recombineering were checked by PCR and sequencing . recB mutations were tested by measuring UV sensitivity . Deletion ΔR111-LC3 and inversions InvT2 and InvT3 were made as described [63] . The araC open reading frame was replaced by the yGFP-parBpMT1-CmR sequence , to express yGFP-ParB protein under the control of the constitutively expressed araC promoter . For this construction , the Cm gene was amplified from pKD3 plasmid using primers harboring an HindIII site ( S3 Table ) . Amplified fragments were digested with HindIII and cloned into the HindIII site of pFHC2973 to make pFHC2973- yGFP-parBpMT1-CmR plasmid ( Nielsen et al . , 2006 ) . Clones were confirmed by PCR and sequencing using flanking primers . The yGFP-parBpMT1-Cm fragment was then amplified from pFHC2973- yGFP-parBpMT1-CmR plasmid using primers 582 and 583 ( S3 Table ) and inserted downstream of the araC promoter by the gene replacement method ( recombineering ) as described in [61] , using DY330 [62] . Cells were grown in M9 minimal medium supplemented with 0 . 4% glucose to exponential phase ( 0 . 2 OD 650 nm ) . Chromosomal DNA was extracted using the Sigma GenElute bacterial genomic DNA kit . 5 μg of DNA were used to generate a genomic library according to Illumina's protocol . The libraries and the sequencing were performed by the High-throughput Sequencing facility of the I2BC ( http://www . i2bc . paris-saclay . fr/spip . php ? article399&lang=en , CNRS , Gif-sur-Yvette , France ) . Genomic DNA libraries were made with the ‘Nextera DNA library preparation kit’ ( Illumina ) following the manufacturer’s recommendations . Library quality was assessed on an Agilent Bioanalyzer 2100 , using an Agilent High Sensitivity DNA Kit ( Agilent technologies ) . Libraries were pooled in equimolar proportions . 75 bp single reads were generated on an Illumina MiSeq instrument , using a MiSeq Reagent kit V2 ( 500 cycles ) ( Illumina ) , with an expected depth of 217X . An in-lab written MATLAB-based script was used to perform marker frequency analysis . Reads were aligned on the Escherichia coli K12 MG1655 genome using BWA software . Data were normalized by dividing uniquely mapping sequence reads by the total number of reads . Enrichment of uniquely mapping sequence reads in 1 kb non-overlapping windows were calculated and plotted against the chromosomal coordinates . Cells were grown in M9 minimal medium supplemented with 0 . 4% glucose to exponential phase ( 0 . 2 OD 650 nm ) and spread on a 1% ( wt/vol ) agarose pad for analysis . For snap-shot analyses , cell images were acquired using a DM6000-B ( Leica ) microscope with MetaMorph software ( Version 7 . 8 . 8 . 0 , Molecular Devices ) and analyzed using ImageJ . Images were taken from 5–10 different fields in each experiment . Two to three independent experiments were carried out to calculate mean and standard deviation for distributions of foci for each strain . For time-lapse analyses , 0 . 4% glucose agarose pads were used , the slides were incubated at 30°C and images were acquired every 10 minute by an Evolve 512 electron-multiplying charge-coupled device ( EMCCD ) camera ( Roper Scientific ) attached to an Axio Observe spinning disk ( Zeiss ) . Image acquisition was done using MetaMorph software ( Version 7 . 8 . 11 . 0 , Molecular Devices ) . At each time point , we took a stack of 32 bright-field images covering positions 1 . 6 μm below and above the focal plane . Image acquisition was performed on five selected different fields corresponding to different cell populations in each experiment . Final images were reconstructed from image stacks using an in-lab written MATLAB-based script . Image analysis was done manually using ImageJ software . For each mutant strain analyzed , two independent time-lapse experiments were realized , each providing five images with 5–10 bacteria per image at the start . The number of divisions that provided two foci-containing cells and the number of first divisions that provided one focus-containing and one focus-free cells were manually counted . Only cells that started with a normal division were taken into account ( few cells produced a focus containing-cell and a focus-less cell from the start and were not counted as initial events , as they did not show any normal division preceding the initial event ) . The percentage of initial events ( between parentheses in Table 1 ) corresponds to the ratio of cell divisions where a focus is lost in one daughter cell for the first time to the total number of cell divisions . For example , in the scheme shown in Fig 2E , we counted 2 initial events ( #2 and 7 ) out of 9 total cell divisions , and 100% heredity . Cells were grown in either M9 minimal medium supplemented with 0 . 4% glucose , 5 μM CaCl2 and 1 mM MgSO4 or LB medium supplemented with 0 . 5% glucose at 37°C as described in Cockram et al , 2015 [51] .
|
RecBCD protein complex is an important player of DSB repair in bacteria and bacteria that cannot repair DNA double-stranded breaks ( DSB ) have a low viability . Whole genome sequencing analyses showed a deficit in specific sequences of the chromosome terminus region in recB mutant cells , suggesting terminus DNA degradation during growth . We studied here the phenomenon of terminus DNA loss by whole genome sequencing and microscopy analyses of exponentially growing bacteria . We tested all processes known to take place in the chromosome terminus region for a putative role in DNA loss: replication fork termination , dimer resolution , resolution of catenated chromosomes , and translocation of the chromosome arms in daughter cells during septum formation . None of the mutations that affect these processes prevents the phenomenon . However , we observed that terminus DNA loss is abolished in cells that cannot divide . We propose that in cells defective for RecBCD-mediated DSB repair the terminus region of the chromosome remains in the way of the growing septum during cell division , then septum closure triggers chromosome breakage and , in turn , DNA degradation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"chemical",
"compounds",
"cell",
"division",
"analysis",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"carbohydrates",
"organic",
"compounds",
"glucose",
"dna",
"replication",
"bioassays",
"and",
"physiological",
"analysis",
"dna",
"sequence",
"motif",
"analysis",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"cell",
"analysis",
"chromosome",
"biology",
"chromosomal",
"aberrations",
"bioinformatics",
"chemistry",
"genetic",
"loci",
"biochemistry",
"chromosomal",
"translocations",
"dna",
"sequence",
"analysis",
"cell",
"biology",
"nucleic",
"acids",
"organic",
"chemistry",
"database",
"and",
"informatics",
"methods",
"genetics",
"monosaccharides",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences"
] |
2017
|
Division-induced DNA double strand breaks in the chromosome terminus region of Escherichia coli lacking RecBCD DNA repair enzyme
|
The interaction between fungal pathogens with the host frequently results in morphological changes , such as hyphae formation . The encapsulated pathogenic fungus Cryptococcus neoformans is not considered a dimorphic fungus , and is predominantly found in host tissues as round yeast cells . However , there is a specific morphological change associated with cryptococcal infection that involves an increase in capsule volume . We now report another morphological change whereby gigantic cells are formed in tissue . The paper reports the phenotypic characterization of giant cells isolated from infected mice and the cellular changes associated with giant cell formation . C . neoformans infection in mice resulted in the appearance of giant cells with cell bodies up to 30 µm in diameter and capsules resistant to stripping with γ-radiation and organic solvents . The proportion of giant cells ranged from 10 to 80% of the total lung fungal burden , depending on infection time , individual mice , and correlated with the type of immune response . When placed on agar , giant cells budded to produce small daughter cells that traversed the capsule of the mother cell at the speed of 20–50 m/h . Giant cells with dimensions that approximated those in vivo were observed in vitro after prolonged culture in minimal media , and were the oldest in the culture , suggesting that giant cell formation is an aging-dependent phenomenon . Giant cells recovered from mice displayed polyploidy , suggesting a mechanism by which gigantism results from cell cycle progression without cell fission . Giant cell formation was dependent on cAMP , but not on Ras1 . Real-time imaging showed that giant cells were engaged , but not engulfed by phagocytic cells . We describe a remarkable new strategy for C . neoformans to evade the immune response by enlarging cell size , and suggest that gigantism results from replication without fission , a phenomenon that may also occur with other fungal pathogens .
The interaction between a microbe and a host involves a complex response by both the pathogen and the infected individual . The host has multiple defence mechanisms to avoid infection , damage and disease . Microbial pathogens adapt to survive in a host through multiple changes that include signalling pathways that confer the capacity to survive immune-mediated stresses . Both entities , the host and the microbe , interact and each contributes to the outcome of infection [1] . In the case of fungal pathogens , the interaction with the host frequently results in morphological changes . For example , Candida albicans forms pseudohyphae and true hyphae during infection , phenomena associated with virulence [2] , [3] , [4] . Other examples of fungal pathogens that form filaments during infection are Aspergillus species and the agents of zygomycosis . In contrast , Histoplasma capsulatum and Blastomyces dermatitidis manifest a temperature regulated dimorphism , such that at ambient temperatures they form filaments and at 37°C transform into yeast cells [5] , [6] , [7] . Although the role of these morphological transitions is not completely understood , it is believed that the phenomenon of fungal dimorphism plays an important function during the interaction of each of these microbes with their host . The fungus Cryptococcus neoformans is the causative agent of cryptococcosis , a disease responsible for over 600 , 000 deaths per year , which makes this pathogen a major global threat . Cryptococcosis is currently the fourth leading cause of death from infectious diseases in Sub-Saharan Africa [8] . C . neoformans is unique among the major fungal pathogens in that it possesses a polysaccharide capsule surrounding a yeast cell body [9] . Capsular polysaccharides are also released into host tissues [10] , [11] , [12] , where they mediate numerous deleterious effects on host immune function [13] , [14] , [15] . In fact , the polysaccharide capsule is the factor that makes the greatest contribution to the virulence of C . neoformans [16] . Although C . neoformans can form pseudohyphae during mating [9] , this pathogen is mainly found in host tissues as round yeast cells . However , there is a specific morphological change associated with cryptococcal infection that involves a significant increase in capsule volume . Capsule size in C . neoformans depends on the growth condition ( reviewed in [17] ) . While capsule size is relatively small in standard laboratory media and in the environment , it undergoes a large increase in capsule size during pulmonary infection [18] , such that it can comprise more than 90% of the total volume of the cell [19] . Capsular enlargement is believed to confer an advantage to the microorganism during its interaction with the host . For example , capsule growth interferes with complement-mediated phagocytosis [20] and protects the yeast cell against free radicals and antimicrobial agents [21] . Furthermore , increased capsule size makes the yeast more difficult to phagocytose by a variety of phagocytic cells , including amoebas that can prey upon C . neoformans in the environment [21] . We now report another morphological change whereby gigantic fungal cells are formed in tissue . This change is achieved , not only by a significant increase in capsule size , but also by an enlargement of the cell body . During pulmonary infection , we observed that a significant proportion of yeasts in the lung had cell volumes 900-fold larger than cells grown in standard laboratory conditions . In retrospect , giant cells have been noted in prior studies [18] , [22] , [23] , [24] , [25] , but were never isolated or studied . The emergence of fungal giant cells poses a formidable problem for the immune system . In this study we present experimental evidence suggesting that C . neoformans gigantism may be a strategy that confers upon the organism the ability to survive within the host for long time periods .
While the typical size of C . neoformans cells ranges between 4–8 microns ( Figure 1A ) , we confirm the existence and report the recovery of C . neoformans cells of enormous size formed during infection ( Figure 1B ) . Although these cells manifested a large increase in capsule size , there was also a concomitant increase in cell body size . The cell body reached 25–30 µm in diameter , which was almost 7-fold greater than the 4 . 5 µm average size observed in vitro . This effect was more dramatic if the size of the capsule was included , with the giant cell size typically ranging from 40 to 60 µm in diameter , although extremely large cells with diameters around 70–100 µm were occasionally observed . If one considers volume applying the formula for a sphere ( V = 4/3×π× ( r3 ) ) , then giant cell formation involved an increase of 900-fold in cellular volume , compared to cells grown in Sabouraud medium . We investigated whether the phenomenon was found in different cryptococcal strains . Consequently , we infected different individual mice with ten different C . neoformans strains ( both serotype A and D , including standard strains , such as H99 , 24067 or B3501 , and clinical isolates from the Yeast Collection of the Spanish Mycology Reference Laboratory ) . For all strains , we found giant cells after three weeks of infection , indicating that this phenomenon applied to diverse strains ( results not shown ) . Hence , we focused our efforts on the model serotype A strain H99 , and we arbitrarily defined giant cells as those with a cell diameter greater than 30 µm ( capsule included ) , a size that is 5–6 times the usual size observed in vitro , and is virtually never encountered during in vitro experimental conditions . Using this strain , we observed giant cell formation in four different mouse strains ( CD1 , BALB/c , C57BL/6J and CBA/J ) , indicating that the emergence of giant cells was also not mouse strain specific . Giant cells had different cellular features than cells of regular size ( Figure 1 ) . Giant cells frequently contained multiple vesicles of unknown function that could reach more than 50 per cell ( Figure 1C ) . In addition , there was usually a single enlarged vesicle that occupied a significant proportion of the cell body volume . To better identify these intracellular structures , we stained the cells with the vacuole specific marker MDY-64 . In regular cells , we normally observed the presence of a single vacuole ( Figure 1D ) . In giant cells , we observed two patterns of staining with this specific marker ( Figure 1E , F ) . Multiple vesicles which stained with the vacuole marker were identified in approximately 50% of giant cells , whereas the remainder displayed staining mainly in a single large intracellular vesicle . These results suggest that in some giant cells , the vacuole fragmented into multiple vesicles or the smaller vesicles failed to coalesce . A peculiarity of the giant cells was the abnormally large width of their cell wall . This feature was most apparent when the cells were observed by transmission electron microscopy . Using this technique , we could determine that the cell wall of regular cells had a width between 50–100 nm ( Figure 1G ) . In contrast , the width in giant cells was 20–30 larger , ranging from 2 to 3 µm ( Figure 1H , I ) . In these pictures , it was also apparent that the density of the capsule differed between regular and giant cells . In the case of yeast obtained in vitro , the cells displayed a low density capsule with individual polysaccharide fibers attached to the cell wall ( Figure 1G ) . In giant cells , the capsule was significantly denser in the regions close to the cell wall ( Figure 1H , I ) . Fungal cell suspensions recovered from the lungs of infected mice had a dark brownish colour . We hypothesized that this phenomenon could be due to in vivo pigment accumulation at the cell wall level , in particular melanin [26] . To investigate this hypothesis , we stained giant cells with specific mAbs to melanin [27] . Giant C . neoformans cells bound mAb to melanin at the cell wall level ( Figure 1J , K ) , suggesting that this structure was melanized . In addition , we observed that giant cells showed a high degree of autofluorescence ( result not shown ) , which has been reported in cryptococcal cells grown in certain media [28] . Scanning electron microscopy images suggested that the capsule of giant cells was different from that of cells grown in vitro . For cells grown in standard Sabouraud medium we noted that the dehydration and fixing process resulted in polysaccharide shrinkage and aggregation of fibers such that regions of the cell wall became exposed ( Figure 2A ) . In contrast , the architecture of the capsule of giant cells appeared intact and well preserved , revealing a highly cross-linked polysaccharide net ( Figure 2B ) that accumulated around the cell body as a very compacted layer ( Figure 2C ) . In addition , we frequently observed the presence of “holes” in the capsule ( Figure 2 D , E and F ) , which we interpreted as pathways formed during recent budding . Transmission electron microscopy images confirmed that the capsule of giant cells was denser than the capsule of in vitro cultivated cells ( Figure 1 H , I ) . The higher degree of cross-linking in the capsule of giant cells relative to in vitro grown cells was confirmed by treating the cells with DMSO or γ-radiation , procedures known to strip the capsule of cells of regular size [29] , [30] , [31] . γ-radiation removed the majority of capsule of giant cells , but the inner region of the capsule remained attached to the cell ( Figure 3 ) . In contrast , DMSO treatment did not affect capsular size of giant cells while it routinely strips the capsule of cells grown in vitro ( results not shown ) . The increased resistance of the inner capsule to radiation and the overall capsule to organic solvents is consistent with a higher degree of capsular cross-linking by the giant cells . In C . neoformans , complement deposition is affected by the porosity and blocking capacity of the capsule [32] . Consequently , we characterized complement localization in the capsule of giant cells as a measure of capsule penetrability . As shown in Figure 4A , complement is known to deposit in the inner location of the capsule near the cell wall of typical yeast cells [20] . When giant cells were incubated in mouse serum , we observed that complement was not detected in the inner regions of the capsule ( Figure 4B ) . The exclusion of complement from the inner capsule is consistent with reduced permeability resulting from increased fibril cross-linking . To ascertain whether giant cells manifested antigenic differences from cells grown in vitro we used indirect immunofluorescence with mAb 18B7 . We compared cells of different size obtained from the lungs of infected mice , to avoid the possibility that factors of the immune system influenced the antigenic properties of the capsule . When stained with mAbs 18B7 , cells of small size exhibited a uniform annular binding pattern ( Figure 4C ) , which was identical to the binding of this mAb to cells grown in vitro [32] . In contrast , most of the fluorescence localized to the edge of the giant cell capsule , and this binding was diffuse and punctate ( Figure 4D–F ) . Moreover , many cells showed a double ring , punctate pattern , with a more uniform inner ring and a rougher , more diffuse outer ring . Chitin-like structures in the capsule were recently demonstrated by the specific binding of fluorescent wheat germ agglutinin , which binds to sialic acids and β-1 , 4-N-acetylglucosamine ( GlcNAc ) oligomers [33] . We used WGA to ascertain whether these structures were also present in giant cells . Cells grown in vitro bound WGA , especially at the neck between the mother cell and the bud ( Figure 4G ) . In giant cells , these structures were particularly prominent . Protrusions into the capsule were longer , reaching several microns ( Figure 4H ) . Giant cells were viable , since they replicated when placed on fresh agar plates . Daughter cells emerging from giant cells were not trapped inside the thick polysaccharide capsule , but rather traversed it in less than 0 . 08 seconds ( Figure 5A and B , supporting Videos S1 and S2 ) . In some cells , movement through the capsule was much faster , taking less than 0 . 01 seconds ( Figure 5C , supporting Video S3 ) . By measuring the distance travelled through the capsule by emerging buds and the transit time we estimated that the daughter cells traversed the capsule at 20–50 m/h , which is a remarkably high velocity for a microscopic unicellular particle in a gelatinous environment . This data suggested the existence of a motive force propelling and separating the buds from the mother cell's capsule . We observed that giant cells could produce several daughter cells over brief periods of time ( 2–3 hours ) , with buds always emerging from the same cell site ( data not shown ) . Despite ejection , the daughter cells remained close to the capsule of the mother cells where they replicated , producing abundant progeny around the giant cells . Even after the giant cells were surrounded by daughter cells , new buds were still ejected with significant force , since they were able to displace and move the surrounding cells upon impact ( see supporting Video S4 ) . When placed on agar , we observed that not all giant cells produced daughter cells after 24 h , suggesting that some of these cells were metabolically arrested , or had died prior to or during the isolation procedure . To measure the percentage of replicating cells we obtained giant cells from two mice and counted the proportion of giant cells producing colonies after 24 h on agar with a microscope . The percentage of giant cells reproducing was 60% and 73% for each mouse , respectively , indicating that the majority of giant cells were viable . As a secondary technique for testing giant cell viability we used the method based on the reduction of 2 , 3-bis ( 2-methoxy-4-nitro-5-sulfophenyl ) -2H-tetrazolium-5-carboxanilide inner salt ( XTT ) by alive cells . Giant cells manifested a strong capacity to reduce XTT , which was approximately 100-fold greater than the activity shown by the same number of cells grown in vitro ( data not shown ) . This result indicates that giant cells are metabolically active . We tried to induce giant cell formation in vitro by incubating the cells in different media . When we incubated the cells in minimal media , around 4–5% of the cells showed a marked increased in cell size over 4 days . These cells reached up to 25–30 µm in diameter ( capsule included ) , approximating , but not quite reaching the size of the giant cells recovered from mouse lungs . In addition , these in vitro giant cells showed other phenotypic differences with the giant cells obtained in vitro , such a smaller capsular size and a lack of enlargement of the cell wall ( result not shown ) . Although this in vitro medium only partially reproduced the gigantism phenomenon , we used it to study if there was any relationship between cellular enlargement and the age of the cells . We hypothesized that massive cellular growth required a prolonged period of time , so the majority of the giant cells obtained would be originated from the initial inocula . To explore this hypothesis , we labelled the cells with complement . Complement proteins , especially C3 , bind to the capsule covalently without inhibiting cell growth and do not segregate to buds after replication [34] . Consequently , we labelled cells grown in Sabouraud medium with mouse C3 and then incubated them in minimal medium ( which induces a small population of giant-like cells ) and Sabouraud medium . At time zero , all the cells were labelled with complement ( Figure 6A ) , but after 4 days of incubation in minimal medium , only a few cells remained labelled ( Figure 6B ) . When we measured the average size of the cells with complement bound after four days of incubation in minimal medium , we found that these cells had a significant larger size than the cells incubated in Sabouraud medium ( Figure 6B ) . We repeated this experiment , placing the cells in parallel in minimal medium , which induces cell enlargement in some cells , and in Sabouraud medium , in which cell enlargement is not expected . Then , we measured cell size of complement-labelled and non-labelled cells . In minimal medium , we observed that the complement-stained cells were significantly larger than the unlabelled population ( Figure 6C , D ) . Large cells were not found in Sabouraud medium after C3 labelling and there was no difference in the size of cells with and without C3 labelling ( Figure 6C , D ) . This result indicates that cellular enlargement and giant cell formation is correlated with the age of the cells , such that the giant cells are the older cells in the culture . We hypothesized that giant cell formation was a consequence of continued cell cycle progression without cellular fission . In other fungi , cell size can be related to DNA content [35] , [36] . Using flow cytometry to measure DNA content , we found that giant cells had low permeability to propidium iodide by regular staining protocols ( results not shown ) , but could be permeabilized by heating at 60°C for 45 minutes . As mentioned above , cells from the lungs of mice infected for 3–4 weeks showed strong autofluorescence , so we measured the intensity of the signal in the presence or absence of propidium iodide . We first analysed the difference in the forward scatter ( FSC , cell size ) and side scatter ( SSC , cell complexity ) . When we compared these parameters , we observed a population of larger yeasts in the cells isolated from the lung that was not present in the yeast cells obtained in vitro , as was expected with the presence of giant cells in vivo ( Figure 7A , B ) . This population was defined as region 1 ( R1 ) , which contained the giant cells present in the population . To estimate the DNA content , we added propidium iodide to these samples . When we measured the propidium iodide staining , we found that there was a high variation in the DNA content in the cell population obtained from lungs ( Figure 7C ) , which implied a high variation in the DNA content of cells in vivo . Fungal cells isolated from the lungs of infected mice also displayed significant autofluorescence in the absence of propidium iodide staining . To assess the staining of giant cells , we subtracted the autofluorescence of the cells present in region 1 from the fluorescence value determined in the presence of propidium iodide . The mean fluorescence intensity for the giant cells was almost 103 fold higher than the signal measured in cells grown in vitro ( Figure 7D ) . This result suggested that giant cells contain multiple copies of DNA . To quantify the ploidy level of the giant cells further , we performed real time PCR to amplify the ITS1 region from ribosomal DNA . We included controls of purified genomic DNA of known concentrations , which yielded a lineal relationship between crossing point values and DNA concentration . When we compared 300 giant and regular cells there was a difference in the Ct values of more than 3 cycles ( 36 . 95 in giant cells versus 40 . 00 in regular in vitro grown cells ) . When we estimated the amount of DNA present in each condition according to a standard curve generated using different concentrations of genomic DNA , we calculated that the amount of DNA in each giant cell was 1 . 3×10−7 ng . In contrast , in vitro cultivated cells contained 8 . 4×10−9 ng . This result indicated that giant cells contained 16× DNA than regular cells . This experiment was repeated using different cell concentrations and consistent results were obtained ( result not shown ) . DAPI was used to directly observe the nucleus of the giant cells . This staining revealed that both regular ( Figure 7E ) and giant cells ( Figure 7F ) contained a single nucleus . We investigated the potential involvement of two of the major signal transduction pathways in C . neoformans ( cAMP and Ras1 [37] , [38] ) in giant cell formation . Ras1-deficient cells produced giant cells in the lungs of infected mice ( Figure 8A ) . In contrast , mutants unable to accumulate cAMP ( lacking adenylate cyclase encoded by the CAC1 gene ) did not produce giant cells during murine infection ( Figure 8A ) , suggesting that this pathway was required for giant cell formation . The absence of giant cells in mice infected with the cAMP mutant was associated with a reduced fungal burden . Hence , the lack of giant cells with this mutant in vivo might be related to the ability of the host to rapidly clear the fungus . For this reason , we examined whether a cac1 mutant formed giant cells in minimal media . The cac1 mutant failed to produce giant cells , whereas a significant proportion of cells of the wild type ( H99 ) and reconstituted ( cac1/CAC1 ) strains manifested cellular enlargement ( Figure 8B ) , confirming that cAMP pathway is involved in giant cell formation . To further characterize this phenotype , we analysed the forward and side scatter profile of the cells from the three strains by flow cytometry , since this type of plot allows for clear differentiations in cell size . As shown in Figure 8C , the cac1 mutant yielded a very homogenous population of relatively small FSC and SSC values . In contrast , the wild type and reconstituted strains produced more heterogeneous populations , in which cells with larger FSC and SSC values were measured , indicating the appearance of cells of larger size . The distribution of cell sizes in vivo was extremely variable depending on the experimental conditions . Under our standard conditions ( infection of 6–8 weeks old mice with 105 yeast cells ) we consistently found that the proportion of giant cells in the lung was between 1–10% , with variation between individual mice and between experiments . However , in several experiments , we occasionally found that the proportion of giant cells was higher than 90% of the lung fungal cell population . Curiously , in those experiments where the proportion of giant cells was very high , there were no obvious signs of disease and the mice looked healthy . We decided to investigate this observation in more detail by studying the relationship between inoculum and giant cell formation . We hypothesized that infections with low inocula could reproduce chronic or latent asymptomatic infections . For this purpose , we performed infections with high ( 106/mouse ) or a low dose ( 104 cells/mouse ) . Mice infected with high inocula consistently developed typical cryptococcal disease , as indirectly shown by progressive weight loss ( Figure 9A ) . However , we found a high variation in outcome when the mice were infected with a low inoculum . Most of these mice ( 2 of 3 ) developed disease comparable with that seen after infection with a high dose . Severe disease was characterized by dense inflammation in the lungs , increasing the size and weight of these organs ( 1 . 3–1 . 8 grams ) such that they accounted for 7–8% of the total body weight ( Figure 9B ) . In contrast , the lung mass of asymptomatic mice ( control mice and one of the mice infected with low inocula ) was approximately 0 . 45 grams and ∼1% of total body weight . As expected , mice developing severe disease had a significantly higher number of CFUs ( >106/lung ) than the mice that did not manifest obvious signs of disease , where the number of CFUs remained very low ( <104 ) during the experiment ( Figure 9C ) . When we recovered the fungal cells from the lungs of these mice , we found profound differences in the size of the yeast cells . The average cryptococcal cell size in mice receiving a high inoculum was around 15–20 µm , which was significantly larger than the size reached when grown in vitro in rich medium ( Figure 9D , Sabouraud medium ) . Although approximately 5–20% of the yeast cells met criteria for giant cells , the enlargement of the majority of cells isolated was mainly due to increase in the capsule size , so the size of these populations was only slightly different from the size reached when capsule size is induced overnight in vitro ( see Figure 9D , in vitro enlarged capsule size ) . We were able to isolate around 1-3×103 C . neoformans cells from the lungs of asymptomatic , low dose infected mice and the average yeast cell size was around 40 µm ( Figure 9D ) . Notably , approximately 70% of the isolated yeast cells were giant forms . We also analysed the proportion of giant cells in the lungs using Classification and Regression Trees ( CART ) analysis . Using this approach , we found that there was a strong association between the total fungal cell size in the lungs and the degree of inflammation , such that high inflammation was predicted when the average fungal cell size was below 36 µm ( Figure 9E ) . This prediction is in accordance with our initial criteria of defining giant cells as those with a cell diameter greater than 30 µm . When we plotted the corresponding ROC curve , we found that the region under the curve was 0 . 80 , which provides strong support for the prediction . On the other hand , the model could not efficiently predict the proportion of giant cells according to inflammation , due to the low number of yeast cells found in lungs without significant inflammation . The relationship between the proportion of giant cells and inflammation was confirmed histologically . In control mice , the lungs revealed the typical structure in which alveolar spaces were present throughout the lungs ( Figure 10 ) . In the infected mice , we only observed this benign histology in the asymptomatic mouse ( mouse 1 of the group infected with 104 yeast cells , Figure 10 ) . In the rest of the low dose ( Figure 10 ) and all of the high dose infected mice ( result not shown ) , dense inflammation was observed and alveolar spaces contained numerous yeast cells and inflammatory cells . We performed another experiment using older mice ( 14–16 weeks old ) , which are more resistant to infection , and a lower infective dose than in the experiment previously described . We infected with either a low ( 103 cells/mouse , 10 mice ) or a high dose ( 105 cells/mouse , 3 mice ) , and measured fungal cell size after a month of infection . In this new model , the mice did not develop any visible sign of disease . When the mice were sacrificed after one month , only one of the mice infected with the high dose showed inflammation in the lungs ( table 1 , 105 , mouse 2 ) . In the group infected with low dose , we did not find yeast cells in 3 mice , indicating that the infection had been cleared . In the other seven mice , we found a low number of yeast cells , suggesting a chronic asymptomatic infection . In six of the seven mice , the average size of the fungal cells was above 30 µm , and the proportion of giant cells was between 50–90% . When the mice were infected with a higher dose , the two mice in which no inflammation was detected showed fungal cell sizes above 30 µm , with a proportion of giant cells around 70–80% . In the mouse with inflammation ( mouse number 2 ) , the average fungal cell size was smaller ( 14 µm ) , and the proportion of giant cells was less than 5% . This data supports the notion that the highest proportion of giant cells is found in hosts with chronic and longstanding infection . We measured the susceptibility of giant cells to oxidative stress produced by incubation in H2O2 . Giant cells were significantly more resistant to killing by oxidative stress than cryptococcal cells grown in vitro , with survival rates of 19%±10 and 46%±14 for cells grown in vitro and giant cells , respectively ( p = 0 . 014 ) . To further characterize the interaction of giant cells with host effector cells , we incubated macrophage-like cells with giant cells and observed the outcome of the interaction using live-imaging microscopy . When small-sized cryptococci were exposed to macrophages , we observed rapid and avid phagocytosis , yeast cell transfer between macrophages , fusion of infected macrophages after division , and intracellular replication of the C . neoformans cells , as described previously [39] , [40] , [41] , [42] , [43] and shown in supporting Videos S5 , S6 and S7 . None of these phenomena were observed when macrophage-like cells were exposed to C . neoformans giant cells , indicating that the interaction between these fungal cells and macrophages was different and had different outcomes . Although in some cases the macrophages seemed to adhere to the giant fungal cells , there was no phagocytosis or macrophage fusion after division ( supporting Videos S8 and S9 ) , indicating that macrophages could not cope with the giant cells .
Cryptococcus neoformans giant cells have been occasionally described in the literature primarily as curiosities in histological tissue sections [22] , [23] , [24] , but their importance in pathogenesis has remained obscure . Apart from the fundamental problems in cell biology posed by the mechanisms responsible for the transition to gigantism , we considered that the presence of fungal giant cells would pose a major problem for the immune system simply by virtue of their size . Giant cell formation was associated with several changes to the capsule relative to the typical cells observed in vitro . These changes represented an exaggerated response in capsule , cell body , and cell wall size during infection . Moreover , the resistance to capsule shedding after γ-radiation exposure suggests a more compact and dense structure . Such an increase in capsular compactness could confer a survival advantage in vivo since the capsule is known to protect against oxidative fluxes of the types produced by immune effector cells [21] . Consistent with this idea , we have observed that giant cells are more resistant to oxidative stress . Giant cells maintained their enormous size ex vivo , although they produced smaller cells in agar at replication rates similar to those observed in vitro . Another remarkable aspect of the budding process is the rapidity with which the buds traversed the capsule , especially considering the denseness and compactness of the polysaccharide noted by scanning and transmission electron microscopy . However , we did observe holes in the capsules of giant cells with dimensions that approximated the size needed for daughter cells to emerge . Similar capsule holes have been occasionally described in encapsulated cells grown in vitro [44] . The strong binding of WGA to these cells also suggests the presence of chitin-like structures , which have been proposed to be involved in the movement of the bud through the capsule of the mother cell [33] . Given prior work noting tunnel-like structures formed around buds [33] , [34] , it is possible that the rapid egress of buds from the mother cells represent movement along such structures that provide a non-obstructed conduit through the capsule . The cellular mechanisms by which cryptococcal cells enlarge to gigantic sizes are not known and a complete understanding of this phenomenon is beyond the scope of the current work . Nevertheless , we explored the potential mechanism of cell division without fission as a way for progressively increasing mass . Cell growth is intimately dependant on the cell cycle . The cells need to reach a critical size for cell cycle to progress , and there is a constant ratio between the mass of the cell and its DNA content ( reviewed in [45] , [46] , [47] , [48] ) . In plants , the phenomenon is striking , because their cells can enlarge in size by 100- or even 1000-fold , and this is achieved by endoreduplication , which is the process in which the cell increases the ploidy of the cells through several rounds of DNA replication [49] . Recently , it has been shown that bacteria from the genus Epulopiscium , which grow to lengths of 200–300 µm and widths of 40–50 µm , have extreme polyploidy , generating tens of thousands of copies of their genome [50] . In a process that may be relevant to cryptococcal gigantism , there are some symbiotic bacteria that undergo an important differentiation process achieved by genome amplification by endoreduplication in plant nodules resulting in significant cell enlargement [51] and provide beautiful examples of how some factors of symbiotic plants regulate the cell cycle of their symbiotic microorganisms . We hypothesized that giant C . neoformans cells achieved their size by repeatedly entering G1 cycles without dividing . To investigate this possibility we stained cells for DNA . Cells recovered from infected animals produced a noisy FACS profile that was interpreted as being consistent with cell-to-cell variation in DNA content . By analyzing cell size and fluorescence intensity we showed that cell size correlated with DNA content , thus establishing that larger cells have more DNA , a result confirmed by real-time PCR . These findings are consistent with a mechanism for DNA replication without cell fission . The proportion of C . neoformans giant cells in infected mouse lung was a function of total microbial burden and pulmonary inflammation . Okagaki et al also found differences in giant/titan cell proportions during infection experiments with MATa and MATα strains , with the proportion of titan cells being higher when mice were co-infected with both mating types ( Okagaki et al , see related article in the current PLoS Pathogens issue ) . These authors have concluded that titan cell formation is induced by the pheromone signalling pathway . Taking our and their results together , we can conclude that gigantism is a morphological response to host environments that impact cAMP and pheromone signalling pathways , which could regulate the cell cycle with the final purpose of generating giant cells during infection . The fact that the survival of the host is not compromised when the proportion of giant cells is high suggests that giant cells can survive in a local environment in the host for protracted periods of time without disseminating in the setting of intact host immunity , a finding in agreement with Okagaki's report . This notion is consistent with reports that a moderate increase in cell size due to capsule enlargement interferes with C . neoformans dissemination from the lung [52] , [53] . Various studies have suggested that C . neoformans dissemination is associated with intracellular survival inside macrophages [54] , [55] , [56] , [57] , but this model cannot be applied to giant cells since they exceed the size of macrophages . Hence , the increased size of the giant cells is likely to be an impediment for their dissemination as they are simply too large to cross biological barriers and/or transverse capillary diameters . Nevertheless , such cells are viable and capable of producing small sized variants when placed in suitable conditions , such as rich agar . Taken together , our findings suggest that giant cell formation could provide the fungus with a strategy for prolonged survival in a host . Cells could conceivably survive through the life of the host and then return to soils when an animal dies . Alternatively , the giant cells could await permissive host conditions such as in the setting of advanced HIV infection , immunosuppression after organ transplant , or other conditions impairing host immune responses that lead the fungus to proliferate to produce abundant progeny . In this context , it is noteworthy that micro-yeast forms have been described in the lung of infected mice [18] , [58] , suggesting another form of size polymorphism at the other end of the scale . All these findings indicate that during infection , C . neoformans can display a wide variation in cell sizes , ranging from micro-forms to giant cells , and suggest that each of these morphotypes have different roles in the pathogenesis of persistence and dissemination . We are aware of occasional reports of gigantic cells in other fungal species . For example , giant Candida albicans cells with diameter up to 30 µm have been described [59] , [60] , and similar large cells have been described for other pathogenic fungi during infection . The arthroconidia of Coccidioides immitis and Coccidioides posadasii swell to form giant spherules ( typically 30–150 µm in diameter ) during mammalian infection and the spherules produce a large number of endospores derived from the cell membrane , each with a single nucleus [35] . When the spherule is mature , the cell membrane is dissolved and the endospores are released . Another example of fungal giant cells occurs in species from the genus Emmonsia , responsible for adiaspiromycosis in humans . Emmonsia crescent cells reach up to 200–700 µm during infection and these forms are multinucleate . Emmonsia parva forms cells up to 40 µm , and they also contain several nuclei [36] . It is possible that gigantism is a general property of unicellular fungi that is expressed under certain conditions . If this is the case , the reproducibility of giant cell formation during cryptococcal infection provides an excellent experimental system for the study of this phenomenon . In summary , C . neoformans cells can achieve gigantic dimensions during infection and the phenomenon suggests that gigantism may be considered a new form of fungal dimorphism . The occurrence of extraordinarily large cells may enable an adaptation for persistence in certain hosts . The findings for C . neoformans together with the similar reports in other fungi suggest that this may be a general mechanism for fungal survival under certain environments and possibly contribute to persistence during host-pathogen interactions .
For most experiments , serotype A H99 strain was used [61] . In some experiments the following strains were also used: 24067 ( serotype D , ATCC ) ; B3501 ( serotype D , [62] ) ; RPC3 ( cac1::URA5 , [37] ) ; RPC7 ( cac1::URA5/CAC1 , [37] ) ; LCC1 ( ras1::ADE2 , [38] ) and different clinical isolates from the Yeast Collection of the Spanish Mycology Reference Laboratory ( CL2132 , CL4860 , CL5154 , CL5632 , CL5707 , CL5066 and CL4979 ) . Yeasts were grown in Sabouraud liquid medium at 30°C with moderate shaking ( 150 r . p . m . ) . In some cases , the yeast cells were grown in minimal media ( 29 . 4 mM KH2PO4 , 10 mM MgSO4 , 13 mM Glycine , 3 µM thiamine , 15 mM glucose , pH 5 . 5 ) . For melanization , L-DOPA containing medium was prepared as in [63] . In other experiments , the cells were transfer from the original Sabouraud culture to 10% Sabouraud medium pH 7 . 3 with 50 mM MOPS buffer , as described in [17] , to induce capsule enlargement in vitro . Six to eight weeks old female BALB/c , C57BL/6J ( Jackson Laboratories , Bethesda , MD ) and CD1 mice ( Charles River Laboratories ) were used in this study . In some experiments , older CD1 mice ( 16 weeks old ) were also used . C . neoformans strains were grown at 30°C , washed with sterile PBS , and suspended at specific cell densities . Fifty microliters of the selected yeast cell suspension were injected intratracheally into mice previously anesthetized with a xylazine/ketamine mixture , as described [64] . Lungs were excised from mice at different infection times and fixed in formalin for 48 h at room temperature . The tissues were then dehydrated and embedded in paraffin using an STP120 Tissue Processor ( Microm International , Walldorf , Germany ) . Then , 5 µm tissue sections were obtained using a Leica RM2245 microtome and placed on glass slides . Hematoxylin/eosin staining of the tissue sections was performed using standard protocols . Mice were euthanized at different times after infection and the lungs were removed . Lung tissue was then homogenized in 10 mL of PBS with 1 mg/ml collagenase ( Roche , Mannheim , Germany ) . The cell suspension was incubated for 1 h at 37°C with occasional vortex agitation , and washed several times with sterile distilled water . The cells were suspended in sterile distilled water , and immediately placed in fixative for microscopy , in fresh medium for microscopy observation , or in Sabouraud agar at 30°C overnight to observe in vitro budding . Cells were viewed with different microscopes . In some experiments , an Olympus AX70 microscope was used and pictures were taken with a digital camera using QCapture Suite V2 . 46 software for Windows . Alternatively , a Leica DMI3000B connected to a DFC300 digital camera with LAS 3 . 3 . 1 software , or a Leica DMI 4000B or a Leica DMRD microscope connected to a Leica DC200 digital camera with IM1000 software were used . To visualize the size of the capsule , the cells were mixed with an India ink suspension . Digital Images were processed with Adobe Photoshop 7 . 0 software ( San Jose , CA ) . For confocal microscopy , a SP5 confocal microscope ( Leica Microsystems ) was use . The macrophage-like cell line RAW264 . 7 was maintained in DMEM medium supplemented with 10% heat-inactivated fetal bovine serum , 10% NTCT , and 1% of non-essential amino acids at 37°C in the presence of a 5% CO2 atmosphere . For phagocytosis experiments , 5×104 macrophages were placed on 96-well plates and incubated overnight at 37°C in the presence of 5% CO2 , so that a total number of 105 macrophages was expected after this incubation given a phagocytic cell replication time of approximately 12 h . Fungal cells were added at a 1∶2 ( macrophage:yeast cells ) ratio in 200 µL of medium . Yeast cells of regular size were obtained by growing in Sabouraud medium overnight . To isolate giant cells , fungal cells were isolated from the lungs of infected mice as described above . Giant cells were separated from the rest of the fungal population by passing the sample through 11 µm filters . Although we defined giant cells as those larger than 30 µm ( capsule included ) , we observed that the capsule did not contribute to retention on the filters , since the size of the cell delimited by the cell wall was the main factor associated with retention or passage . The filters containing the giant cells were incubated in PBS with shaking for 20 minutes , and the cells were concentrated by centrifugation . After filtration , we observed that the population was significantly enriched in giant cells , being more than 90% of the sample . Finally , the cells that transited or were retained by the filter were counted using a haemocytometer . Once regular and giant cells were obtained and exposed to the macrophages , the 96-wells plate was placed under a Leica DMI 4000B microscope using a 20× objective with a 5% CO2 environment and 37°C . Pictures were taken at different time intervals ( see figure legend of the corresponding supporting videos ) . The videos generated by the Leica software were exported as . avi documents and processed with ImageJ ( National Institutes of Health , USA , http://rsb . info . nih . gov/ij/index . html ) and VidCrop 2 . 1 . 0 . 0 ( GeoVid ) softwares . The final videos were generated by merging 5 frames per second . Suspensions of cells isolated from mouse lung were air-dried on poly-L-lysine-coated slides ( Sigma ) . The slides coated with the cells were washed in PBS , incubated in blocking buffer ( Pierce , Rockford , IL ) for 1 h at 37°C followed by incubation with 10 µg/ml of the IgM melanin-binding monoclonal antibody ( mAb ) 6D2 for 1 h at 37°C . MAb 6D2 was generated against melanin derived from C . neoformans [27] . After washing , the slides were incubated with a 1∶1000 dilution of tetramethyl rhodamine isothiocyanate ( TRITC ) -conjugated goat anti-mouse ( GAM ) IgM ( Southern Biotechnologies Associates , Inc; Birmingham , AL ) for 1 h at 37°C . The slides were washed , mounted using a 50% glycerol , 50% PBS , and 0 . 1 M N-propyl gallate solution , and viewed with an Olympus AX70 microscope equipped with fluorescent filters . Negative controls consisted of cells incubated with mAb 5C11 , which binds mycobacterial lipoarabinomannan [65] , as the primary Ab or with TRITC-labeled Ab alone . Yeast cells were washed in PBS and suspended in fixing solution ( 2% p-formaldahyde , 2 . 5% glutaraldehyde , 0 . 1 M sodium cacolydate ) . Cells were then serially dehydrated with ethanol , coated with gold palladium and visualized using a JEOL ( Tokyo , Japan ) JAM 6400 microscope . Cells grown in vitro or isolated from the lungs of infected mice ( see above ) were fixed with 2 . 5% glutaraldehyde in 0 . 1 M sodium cacodylate buffer . The cells were treated with osmium tetraoxide and serially dehydrated . The samples were embedded in epoxy resin and ultrathin sections were obtained , stained with uranyl acetate and lead citrate , and observed in a CM12-Phillips transmission electron microscope . Yeast cells with enlarged capsule were exposed to varying amounts of γ-radiation from 137Cs to remove layers of the polysaccharide capsule as described [29] , [66] . Briefly , giant and in vitro-grown cells were washed three times in PBS to remove shed capsular polysaccharides , suspended in 1 mL of distilled H2O , and irradiated using a Shepherd Mark I Irradiator at the dose rate of 1388 rads/min . For all experiments , cells were irradiated for 40 minutes . Irradiated cells were collected by centrifugation . In other experiments , the fungal cells were suspended in DMSO as described in [29] . The presence of capsule after the treatments ( γ-irradiation or DMSO ) was visually observed by suspending the cells in India Ink and regular microscopy . Complement ( C3; complement protein 3 ) deposition on the cryptococcal capsule was performed as in [20] . Briefly , C57BL/6J mice were bled from the retro-orbital cavity and serum was obtained by centrifugation . Approximately 2×107 cryptococcal cells were suspended in 700 µL freshly-obtained serum , and incubated at 37°C for 1 h . Cells were extensively washed and suspended in PBS . C3 was then detected using a fluorescein-isothiocyanate ( FITC ) conjugated GAM C3 antibody ( 4 µg/mL , Cappel , ICN , Aurora , OH ) . Yeast washed and not suspended in serum were used as controls . To delineate the capsular edge , mAb 18B7 ( 10 µg/mL ) specific for GXM [67] was added , and detected using a TRITC conjugated GAM IgG1 antibody ( 10 µg/ml , Southern Biotechnology Associates , Inc ) . The cells were observed under fluorescent filters with the Olympus AX70 microscope , QCapture Suite V2 . 46 software for Windows , and Adobe Photoshop 7 . 0 for Macintosh . Yeast cells were isolated from infected mice as described above , and placed on Sabouraud agar plates for 18 h 30°C . Initially , we tried recording the budding of giant cells by basic microscopy techniques , such as taking pictures every few minutes or seconds . However , the separation of the bud through the capsule of the giant cell was too fast , so we developed a new approach to record the phenomenon . The surface agar plate was observed with an Olympus AX70 microscope to visualize and continuously record giant cells . To record real-time daughter cell emergence , the cells were observed in the computer screen with the “Preview” option , and the image of the screen was recorded with a Digital Handycam Sony Camcorder affixed to a tripod . The videos were converted into digital files using Windows Movie Maker software provided by Microsoft Windows and processed with the Quick Media Converter ( V . 3 . 6 . 5 ) software . Although this method provided lower resolution than the regular CCD used in microscopy , it permitted a precise measurement of the phenomenon . Giant cells were obtained by filtering the yeasts obtained from the lung of infected mice through 22 µm filters . Then , the yeast cells were separated from the filter by gently shaking the filters in 20 mL of water in 50 mL centrifuge tubes . After 20 minutes , the filters were removed , and the tubes centrifuged at 2000 r . p . m . Then , the cells were suspended in 2 mL of sterile water and the cell concentration was estimated using a haemocytometer . Approximately 105 giant cells were placed on 96 wells plates . In parallel , regular cells were obtained by overnight incubation in Sabouraud , washed with sterile water and counted with a haemocytometer . Then , the same number of cells ( 105 ) was placed in 96-wells plates . As negative controls , equal numbers of giant and regular cells were heat-inactivated ( 45 minutes at 60°C ) and placed in different wells of the 96-wells plates . Viability measurement based on the reduction of 2 , 3-bis ( 2-methoxy-4-nitro-5-sulfophenyl ) -2H-tetrazolium-5-carboxanilide inner salt ( XTT ) by living cells was performed as described in [21] with minor modifications , which involved the use of 1 mg/mL of XTT and 25 µM menadione . Optical density at 450 nm was recorded every 30 minutes for 18 hours in a iEMS Spectrophotometer ( Thermofisher ) . Differences in metabolic activities were calculated by fold differences in the optical densities of the different wells . To detect capsular features , we observed the immunofluorescence pattern after incubating the cells with the mAb 18B7 to the capsular polysaccharide as described above ( see complement labelling and detection section ) , but using goat anti-mouse IgG1-FITC conjugated as the detection Ab . Yeast washed and incubated with the IgG1-FITC alone were used as controls . In some experiments , calcofluor ( 10 µg/mL ) was included to visualize the cell wall . To observe the presence of chitin-like structures , fungal cells with enlarged capsule ( incubated in 10% Sabouradud in 50 mM MOPS buffer pH 7 . 3 ) were treated as in [33] . Briefly , the cells were washed with PBS and suspended in 4% p-formaldehyde cacodylate buffer ( 0 . 1 M , pH 7 . 2 ) and incubated for 30 min at room temperature . The fixed cells were washed in PBS and suspended in 100 µl of a 5 µg/mL of WGA conjugated to Alexa 594 ( Molecular Probes , Invitrogen ) for 1 hour at 37°C . Cell suspensions were mounted over glass slides and photographed with a Leica DMI 3000B fluorescence microscope . To identify the vacuole in the yeast cells , we used the specific dye MDY-64 ( Molecular Probes , Invitrogen , Eugene , Oregon ) following the manufacturer's recommendations . Briefly , the cells were suspended in 10 mM HEPES buffer ( pH 7 . 4 ) supplemented with 5% glucose . MDY-64 was dissolved in DMSO , and added to 106 cells at a final concentration of 10 µM . The cells were incubated for 5 minutes at room temperature , and washed twice with the same buffer . The cells were observed with a SP5 confocal microscope ( Leica Microsystems ) . Cells were isolated as described above and fixed by heating the cells at 60°C for 45 minutes in PBS buffer . Then , the cells were separated in two parallel samples , and propidium iodide was added to one of them at a final concentration of 10 µg/mL . DNA content was analyzed using a FACSCalibur cytometer ( Becton Dickinson ) . As a control , cells grown in vitro in Sabouraud medium were also analyzed . To visualize the nucleus , the cells were treated with 3 . 7% formaldehyde for 30 minutes . Then , the cells were washed with PBS and DAPI was added at 0 . 3 µg/mL . The cells were incubated for 10 minutes at 37°C , and then washed twice with PBS . Finally , fluorescence was visualized in a Leica DM3000 microscope . Giant cells were obtained by filtering the lung extracts through 22 µm filters as above . The filters were then placed in 50 mL tubes containing 20 mL of sterile water with moderate shaking , and after 20 minutes , the filters were removed . The tubes were centrifuged , and the pellet suspended in 0 . 5 mL of sterile water . Then , the cell concentration was determined using a haemocytometer . In parallel , cells obtained from a fresh liquid culture in Sabouraud were counted , and a cell suspension was prepared at the same concentration as that calculated for the giant cells . A real-time PCR using whole cells was then designed using equivalent numbers of the different cell types in the well . The reaction ( final volume of 20 µL ) contained 2 . 8×103 or 2 . 8×102 cells , 1 . 5 mM MgCl2 , and 0 . 8 µM of ITS1 ( 5′TCCGTAGGTGAACCTGCGG3′ ) and ITS2 ( 5′GCTGCGTTCTTCATCGATGC3′ ) oligonucleotides , which amplify the ITS1 region from the ribosomal DNA . The real time was performed using the SensiMix Kit ( Quantance ) using the enzymes and SYBR green concentrations recommended by the manufacturer . The reaction mix was placed in a 96-wells plate and PCR was performed in a LC480 real-time PCR machine ( Roche ) . We included wells with a known concentration of C . neoformans genomic DNA ( 20 , 2 , 0 . 2 and 0 . 02 ng ) to quantify the results . The PCR was performed according to the following protocol: initial step of 10 minutes at 95°C and 45 amplification cycles ( 10 seconds at 95°C , 5 seconds at 54°C and 30 seconds at 72°C ) . Once the PCR was finished , a standard curve was calculated using the wells of the known genomic DNA concentration , and this curve was used to calculate the estimate of DNA present in each of the samples . The CART system was proposed by Breiman et al . [68] , and is characterized by binary-split searches , automatic self-validation procedures and surrogate splitters . This analysis is used to find associations between events with statistical support . CART analysis ( CART 6 . 0 Salford Systems , Ca . , USA ) was used to find associations between giant cell formation and inflammation in the lungs . This analysis was performed with the following methodological conditions , Gini method , minimum cost tree regardless of the size for selecting the best tree , 10 v-fold-cross-validation , equal priors , no costs , and no penalties . Relative error of 0 means no error or perfect fit , whereas 1 represents the performance of random guessing . The statistical support for this association is given by the ROC curve . In this graph , specificity ( false positive rate ) vs . sensitivity ( true positives rate ) is calculated , and the area under the curve is analysed . When this area is 1 ( 100% of sensitivity and 0% false positives ) , a total agreement for the prediction is obtained . An area of 0 . 5 or below is indicative of random guess . Yeast cells ( regular and giant ) were obtained as described above . The cells were incubated in PBS with or without 1 mM H2O2 at a cell density of 104 cells/mL . After two hours of incubation at 37°C , 100 µL of each sample was plated on Sabouraud agar medium . In addition , a 1/10 dilution was done in PBS , and 100 µL of this dilution was also plated . The plates were incubated at 30°C for 48 hours and the colonies were enumerated . The survival was expressed as the percentage of colonies counted in the samples incubated with H2O2 compared to colonies of control samples not exposed to the oxidative agent . Normal distribution in group samples were assessed using the Shapiro-Wilk and Kolmogorov-Smirnov tests using Unistat 5 . 0 ( Unistat Ltd , London , England ) and Analyse-it ( Analyse-it Ltd , Leeds , England ) softwares for Excel . Statistical differences between groups were tested using Student's t-Test ( normal distributions ) or Kruskal-Wallis test ( non-parametric test for non-normally distributed samples ) . Differences were considered significant when p value was below 0 . 05 . All the experiments involving the use of animals have been performed following the guidelines of the Bioethical and Animal Welfare Committee of the Instituto de Salud Carlos III ( approved protocol PA-349 , to be performed at the National Centre for Microbiology ) .
|
In this article we describe the formation of giant cells by the human fungal pathogen Cryptococcus neoformans during infection , involving an approximately 900-fold increase in volume compared to that of yeast cells grown in vitro . This switch to gigantism is a dramatic transition that is posited to have important consequences during infection . The paper reports the phenotypic characterization of these cells and the relationship between giant cell formation and polyploidy which suggests that gigantism is achieved by continued cell growth and DNA replication without fission . During infection , we observed an inverse correlation between the proportion of giant cells in the lung of infected mice and the inflammatory response elicited by the animals . In conclusion , our results indicate that during infection , C . neoformans forms giant cells , which might be implicated in fungal survival in the host during long time periods , especially during chronic and asymptomatic infection . The capacity for gigantism is an important new facet in fungal pathogenesis that provides the pathogen with the ability to escape host defences . We propose that the transition to gigantism can have profound consequences for the host-pathogen interaction including promoting fungal persistence in the host that can translate into latency and disease relapses .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases/fungal",
"infections"
] |
2010
|
Fungal Cell Gigantism during Mammalian Infection
|
Adverse drug effects ( ADEs ) are one of the leading causes of death in developed countries and are the main reason for drug recalls from the market , whereas the ADEs that are associated with action on the cardiovascular system are the most dangerous and widespread . The treatment of human diseases often requires the intake of several drugs , which can lead to undesirable drug-drug interactions ( DDIs ) , thus causing an increase in the frequency and severity of ADEs . An evaluation of DDI-induced ADEs is a nontrivial task and requires numerous experimental and clinical studies . Therefore , we developed a computational approach to assess the cardiovascular ADEs of DDIs . This approach is based on the combined analysis of spontaneous reports ( SRs ) and predicted drug-target interactions to estimate the five cardiovascular ADEs that are induced by DDIs , namely , myocardial infarction , ischemic stroke , ventricular tachycardia , cardiac failure , and arterial hypertension . We applied a method based on least absolute shrinkage and selection operator ( LASSO ) logistic regression to SRs for the identification of interacting pairs of drugs causing corresponding ADEs , as well as noninteracting pairs of drugs . As a result , five datasets containing , on average , 3100 potentially ADE-causing and non-ADE-causing drug pairs were created . The obtained data , along with information on the interaction of drugs with 1553 human targets predicted by PASS Targets software , were used to create five classification models using the Random Forest method . The average area under the ROC curve of the obtained models , sensitivity , specificity and balanced accuracy were 0 . 837 , 0 . 764 , 0 . 754 and 0 . 759 , respectively . The predicted drug targets were also used to hypothesize the potential mechanisms of DDI-induced ventricular tachycardia for the top-scoring drug pairs . The created five classification models can be used for the identification of drug combinations that are potentially the most or least dangerous for the cardiovascular system .
Adverse drug effects ( ADEs ) are one of the top 10 causes of death in developed countries , are one of the main reasons for stopping the development of new drug-candidates and are the main reason for drug recalls from the market [1 , 2] . Cardiovascular effects are some of the most serious ADEs that may lead to hospitalization or death , and , at the same time , are widespread [1] . The ADE profile of a particular drug-candidate is usually investigated during standard preclinical animal tests and clinical trials according to the GLP and GCP requirements . However , many rare , but serious , ADEs cannot be revealed by these studies , because of interspecies differences , the limited number of patients or animals and the duration of studies; thus , additional in vitro and in silico methods for the detection of serious ADEs are currently being developed [3–8] . These methods are based on the determination of the relationships between several chemical and biological features of drugs and their ADEs . Among these features are molecular descriptors , known and predicted drug targets , gene expression changes induced by drugs , phenotypic features such as perturbed pathways , or known ADEs . The relationships between these features and ADEs are usually established using various machine learning methods and network-based approaches . It is accepted that the interaction with human proteins is the most common cause of ADEs; therefore , known and predicted human targets are the most common type of drug features that are used in corresponding studies . Many of the developed methods require knowledge of only the structural formula of a drug-candidate to predict its potential ADEs; therefore , they can be used at the earliest stages of drug development , which may sufficiently increase their effectiveness [3 , 4 , 8] . In real clinical practice , the treatment of human diseases often requires the administration of several drugs , which can lead to drug-drug interactions ( DDIs ) , thus causing an increase in the frequency and severity of ADEs [9] . An evaluation of the effect of DDIs on the manifestation of ADEs is a nontrivial task and requires numerous preclinical and clinical studies . To solve this problem various computational approaches for the prediction of DDIs were developed [10–22] . Most of these approaches are based on the calculation of similarities between the profiles of various chemical and biological features of two drugs . These similarities can be calculated based on molecular fingerprints , drug targets , their amino acid sequences , pathways and Gene Ontology ( http://www . geneontology . org/ ) annotations , the Anatomical Therapeutic Chemical ( ATC ) Classification terms ( https://www . whocc . no/atc_ddd_index/ ) , as well as known ADEs of individual drugs [10 , 12 , 13 , 15–17 , 18 , 20 , 22] . The Tanimoto coefficient is the most common similarity that is measure in these studies; however , more complicated measures can be used , e . g . , several approaches were developed to calculate the proximity of the protein targets of two drugs in a protein-protein interaction network [12 , 17] . Similarity measures based on the profiles of different features can be integrated into single interaction scores that allow drug pairs to be ranked according to their potential ability to interact with each other . To estimate the parameters of such integration and validation of obtained results , information about known DDIs was used . Such data can be obtained from various public databases , including DrugBank ( https://www . drugbank . ca/ ) . For example , Cheng F . with colleagues [13] used several machine learning methods with drug phenotypic , therapeutic , chemical and genomic similarities used as features to predict DDIs . The classifiers were trained on the set of known DDIs from the DrugBank database and the same number of randomly chosen drug pairs as the negative examples . The best result with the area under the ROC-curve ( AUC ) 0 . 67 was achieved using a support vector machine with a Gaussian radial basis function kernel . In addition to approaches that are based on similarities , some other methods were developed [14 , 19] . Zakharov A . V . with colleagues [19] used separate training sets of pairwise drug combinations for each of four isoforms of cytochromes P450 , which are examples of known DDIs . The corresponding information was obtained from the literature . Drug pairs were represented as mixtures of compounds in ratio 1:1 , and several types of molecular descriptors were generated for them . The prediction models were generated by using the radial basis function self-consistent regression and a Random Forest . The balanced accuracies that were obtained from the cross-validation procedure varied from 0 . 72 to 0 . 79 , depending on the dataset [19] . Luo H . , with colleagues , used the sums and differences of the docking scores for 611 human proteins to describe 6328 drug pairs , which represented known DDIs from the DrugBank database , and the same number of drug pairs was randomly chosen as a negative example . A predictive model was created based on l2-regularized logistic regressions to obtain their values . The obtained accuracy , sensitivity and specificity that were calculated based on the 10-fold cross-validation procedure were 0 . 804 , 0 . 847 and 0 . 772 , respectively [14] . Despite the significant progress in predicting DDIs , all of these methods allow for estimating only the fact of interaction , but not the resulting ADEs , whereas such information is important to assess the clinical significance of DDIs . The main problem is the absence of known data for most of the DDI-induced ADEs . The major source of data on ADEs of individual drugs is drug labels [23]; however , they usually contain very few data on ADEs that are induced by DDIs . Nevertheless , the corresponding information can be obtained through the analysis of spontaneous reports ( SRs ) which are received by regulatory agencies from healthcare professionals and patients . Each SR contains information about all drugs that are prescribed to a patient , as well as information about developed ADEs . An analysis of large sets of SRs allows for relationships between certain ADEs and individual drugs [24–29] , or drug combinations [30–35] , to be revealed . The datasets of individual drugs with information about ADEs obtained by an analysis of SRs were earlier successfully used for the creation of predictive models that are based on structure-activity relationships [27 , 29] . The corresponding information on ADEs that is induced by pairwise drug combinations may also potentially be used for this purpose . We developed a computational approach for the assessment of cardiovascular ADEs of DDIs . The approach is based on a combined analysis of SRs and predicted drug-target interactions ( DTIs ) and allows for the prediction of five cardiovascular ADEs of DDIs: myocardial infarction , ischemic stroke , ventricular tachycardia , arterial hypertension and cardiac failure , with balanced accuracies from 0 . 73 to 0 . 81 . Unlike most of the other methods , our approach requires only structural formulas to predict cardiovascular adverse effects for any pair of drugs , and , therefore , may be applied for new , drug-like compounds that have not yet been studied . The developed approach can be used for the identification of pairwise drug combinations that are potentially the most or least dangerous for the cardiovascular system .
We developed a new computational approach for the assessment of cardiovascular ADEs of DDIs through a combined analysis of SRs and predicted DTIs ( Fig 1 ) . The approach is based on two main steps: creation of datasets on cardiovascular DDI-induced ADEs containing drug pairs that potentially cause or do not cause ADEs , and the creation of classification models for each dataset based on predicted drug targets as descriptors . The creation of datasets is based on the analysis of SRs from the standardized version of publicly available parts of the FDA database [36] . The analysis was performed using least absolute shrinkage and selection operator ( LASSO ) logistic regression with the addition of propensity scores as independent variables [35] ( see Materials and Methods for details ) , which allows for the identification of drug pairs that potentially cause or do not cause cardiovascular ADEs–positive and negative examples . Each “positive” drug pair represents a potential synergistic or additive effect of DDI on the development of ADEs . This method takes into account the confounding effects of other drugs and risk factors on the manifestation of ADEs and , thus , allows for datasets with lower numbers of false positives to be obtained . To further improve the quality of datasets , information about the ADEs of individual drugs [37] was used to filter out potentially false positive and false negative examples ( see Materials and Methods ) . Since the created datasets may still contain non-causal drug pair-ADE associations , we used an approach based on inference scores ( ISs ) [38] derived from Comparative Toxicogenomics Database ( http://ctdbase . org/ ) to validate them and estimate their quality ( see Materials and Methods ) . At the second step of the approach , a PASS Targets software [39] was used to predict interactions of individual drugs that were from obtained datasets with 1553 human protein targets . The sums and absolute values of the differences in the probability estimates of interaction with targets were used as descriptors for drug pairs . The classification models were built using Random Forest along with a method that allows for the applicability domain to be determined . The accuracy of prediction is estimated using a 5-fold cross-validation procedure ( see Materials and Methods ) . To demonstrate the practical benefit of the obtained models , predictions of ADEs for a large amount of drug pairs were performed . The analysis of the biological role of predicted protein targets for the top predicted drug pairs that potentially cause ADEs allows for proposing the potential mechanisms of corresponding DDIs . At the first step of the proposed approach , we created five datasets of drug pairs that potentially cause and do not cause five cardiovascular ADEs through the analysis of SRs ( see Materials and Methods ) , namely , ventricular tachycardia , myocardial infarction , ischemic stroke , arterial hypertension and cardiac failure ( see S1 Table ) . Each positive drug pair represents an example of a potential synergistic or additive DDI that causes a corresponding ADE . The datasets contain , on average , more than 3100 drug pairs belonging to 335 individual drugs and 166 ATC terms ( https://www . whocc . no/atc_ddd_index/ ) of the fourth level ( Table 1 ) , reflecting the chemical/therapeutic/pharmacological subgroup of drugs , which indicates that the created datasets are representative . Since the datasets were created by analysis of SRs and were not confirmed experimentally , they may still contain non-causal associations between drug pairs and ADEs . To validate them , we used a method based on inference scores ( ISs ) [38] from Comparative Toxicogenomics Database ( http://ctdbase . org/ ) . ISs are calculated from known drug-gene-disease relationships and reflect the influence of drugs on disease manifestation ( therapeutic or adverse effect ) ( see Materials and Methods ) . We compared ISs for corresponding diseases between drug pairs from created datasets , which potentially cause and do not cause cardiovascular ADEs . We calculated AUC values for each dataset and p-values based on the Wilcoxon test to estimate their statistical significance ( Table 2 ) . The corresponding values range from 0 . 901 to 0 . 615 and reflect the quality of the datasets . According to AUC values , the dataset for arterial hypertension has the best quality , whereas the dataset for ischemic stroke has the worst quality . It is important to note that AUC values reflect both errors in datasets , caused by disadvantages of the analysis of SRs , and errors of approach , which was used for the calculation of corresponding ISs . Thus , the AUC values reflecting the quality of datasets must really be higher . According to the obtained results , we can conclude that the created datasets have from good to moderate quality and can be used for further analysis . We used Random Forest to create classification models based on five datasets and the local ( Tree ) approach to determine their applicability domain [40] . The models were created based on sums and absolute values of differences of probability estimates of interaction with 1553 human protein targets that had been calculated for individual drugs by PASS Targets software [39] . The accuracy estimates were obtained by a 5-fold cross-validation procedure with use of the “compound out” approach [41] ( see Materials and Methods for details ) . The obtained average values of AUC , sensitivity , specificity and balanced accuracy were 0 . 837 , 0 . 764 , 0 . 754 and 0 . 759 , respectively , whereas 95 . 7% of the drug pairs were in the applicability domain of the models ( Table 3 ) . The accuracy values generally correlate with the AUC values obtained using ISs ( Table 2 ) . We also estimated the prediction accuracy of ventricular tachycardia and arterial hypertension on two external test sets , which are based on the data from the DrugBank database ( see Materials and Methods ) ( Table 4 ) . The obtained relatively high accuracies ( Tables 3 and 4 ) allow for the application of the created models to solve practical tasks , e . g . , to perform a search of new pairwise combinations of drugs that potentially interact and cause cardiovascular ADEs . The created datasets contain from hundreds to thousands of drug pairs that potentially cause cardiovascular ADEs depending on the effect; however , the number of possible pairwise drug combinations is much higher . To investigate the practical benefit of the created classification models , we performed a prediction of the DDIs-induced ADEs for all of the possible drug pairs that were generated from individual drugs with known data on five cardiovascular ADEs ( see Materials and Methods ) [37] . Five large datasets were generated with more than 230000 drug pairs on average , and 190000 pairs ( 84% ) of them were in the applicability domain of the models ( see Table 5 ) . Surprisingly , nearly half of the drug pairs in the datasets were predicted to cause corresponding DDI-induced ADEs . A large number of predicted drug pairs can be explained by a prediction probability distribution ( Fig 2 ) . Most of the predicted drug pairs have probability estimates are near the threshold P > 0 . 5 , and they are unlikely to cause ADEs , whereas there are near 2 . 6% of drug pairs potentially cause ADEs at probability threshold P > 0 . 8 ( Table 5 ) . To roughly estimate the accuracy of predictions for large datasets , we calculated AUC values based on ISs from Comparative Toxicogenomics Database at different thresholds of probabilities ( Fig 3 ) . Fig 3 demonstrates that the AUC values for most of ADEs increase with increasing the probability threshold . The obtained AUC values at high probability thresholds are near the corresponding values obtained on training sets ( see Table 2 ) . Thus , high probability thresholds should be chosen for the selection of drug pairs potentially causing ADEs . The results of these analyses and the results of 5-fold cross-validation ( the average area under the ROC curve , sensitivity , specificity and balanced accuracy were 0 . 837 , 0 . 764 , 0 . 754 and 0 . 759 , respectively; see Table 3 ) indicate that the accuracy of the prediction of the most of DDI-induced cardiovascular ADEs is relatively high and that the created models can be applied in the search for new pairwise combinations of drugs that are the most or the least dangerous for the cardiovascular system . Because DTIs are needed for the creation of models that were predicted by PASS Targets software based on structures of drugs , the developed models can be used for any drug-like compounds , including those for which only structural formulas are known . For example , they can be used to predict DDI-induced ADEs for drug candidates on the stage of clinical trials . Since DDI-induced ADEs are effectively estimated by using data on predicted DTIs , the corresponding information on drug targets may also be used to reveal the potential mechanisms of cardiovascular ADEs and influence of DDIs on their manifestation . We performed a corresponding analysis for the top 10 drug pairs from the large dataset with the highest probability scores for ventricular tachycardia ( Table 6 ) . We selected only those pairs where corresponding drugs do not cause ventricular tachycardia when administrated separately . According to prediction results , the drugs possibly cause ventricular tachycardia when they are administered together . We found that the DDIs for these drug pairs may occur at both levels of pharmacokinetics and pharmacodynamics . First , the drugs from five of ten pairs are metabolized by the same cytochromes P450 . Second , corresponding drugs potentially interact with protein targets to influence the action potential of cardiac cells . These targets , either known or predicted , are shown in Table 6 . It is important that only chlorphenamine and alfentanil were predicted to interact with the HERG ( KCNH2 ) potassium channel , which is a well-known protein that is associated with ventricular tachycardia [5] . However , this and other drugs from selected pairs that are known to or are predicted to interact with human proteins form compact fragments of the regulatory network ( Fig 4 ) and indirectly change the action potential . Such changes may form a basis for the induction of ventricular tachycardia in predisposed patients .
The data on cardiovascular ADEs of individual drugs were obtained from our previous study [37] . Briefly , we created five datasets of individual drugs which cause and do not cause the following cardiovascular ADEs: ventricular tachycardia , myocardial infarction , ischemic stroke , arterial hypertension , and cardiac failure . The primary source of information for the creation of datasets was SIDER 4 . 1 ( http://sideeffects . embl . de/ ) , which contains data on ADEs of drugs obtained from drug labels [23] . For each drug-ADE pair , we manually checked the section of the drug label where the ADE was described . If it was described in “Boxed Warning” or “Warnings and Precautions” sections , we considered that drug causes ADE . If ADE was described in section “Adverse reactions , ” which may contain effects unrelated to drug intake , it had to be verified . To do this , additional information on ADEs was obtained using the following sources and approaches: We considered drug-ADE association from “Adverse reactions” section to be verified if it was confirmed from at least one additional source . If ADE was not indicated in the drug labels and publications although the compound had been used clinically for > 5 years and had > 50 SRs about other effects , then it was considered not to cause the corresponding effect . We proposed that integration of information from various sources allow filtering out most of false positive and false negative drug-ADE associations from created datasets . In our current study , we used the AEOLUS database [36] as a source of SRs . AEOLUS is a curated version of publicly available parts of the FDA database of SRs ( https://www . fda . gov/Drugs/GuidanceComplianceRegulatoryInformation/Surveillance/AdverseDrugEffects/default . htm ) , where the names of ADEs , drugs and indications are standardized . We selected only those SRs that contain description of drugs , ADEs and drug indications , because all of these types of data are required for further analysis . A total of 4028051 SRs were selected . The ADEs and indications in the database were described by the preferred terms ( PTs ) of the MedDRA dictionary ( https://www . meddra . org/ ) . Since some PTs may describe pathologies that are related to the same or similar ADEs , we selected the main PTs , which exactly match the investigated ADEs and supporting PTs , which are conditions that are similar to or are indirectly related to ADEs . The main and supporting PTs for five investigated cardiovascular ADEs are presented in Table 7 . At the next step , we selected those drugs in the AEOLUS database that have annotations on five investigated cardiovascular ADEs: ventricular tachycardia , myocardial infarction , ischemic stroke , arterial hypertension and cardiac failure . The data on drugs that caused and did not cause five ADEs was obtained from our previous study [37] ( see above ) . The following numbers of drugs were selected: 496 drugs for ventricular tachycardia , 460 drugs for myocardial infarction , 447 drugs for ischemic stroke , 398 drugs for arterial hypertension , and 467 drugs for cardiac failure . The data on the five ADEs of these individual drugs are represented in S2 Table . We selected drug pairs that were formed by these drugs with at least 100 SRs wherein both drugs are mentioned . For each pair of drugs and each PT from Table 7 , we performed an analysis which is based on three steps . At the first step , we found which of the drug pairs are associated with selected PTs . At the second step we used LASSO logistical regression [35] to estimate the potential synergistic and additive DDIs that are associated with the drug pairs that were selected in step 1 . At this step , noninteracting drug pairs were also determined . At the third step , we integrated the obtained data on different PTs into single ADEs to create datasets with positive and negative examples of DDI-induced ADEs ( see Table 1 ) . Since the datasets on five cardiovascular ADEs were created using analysis of SRs , they may still contain false positive and false negative associations between drug pairs and corresponding effects , thus , datasets have to be validated before performing further analysis . For this purpose we used inference scores ( ISs ) from Comparative Toxicogenomics Database [38] . ISs were calculated based on known interactions of drugs with human genes which have links to corresponding diseases in literature . ISs reflect the degree of similarity between drug–gene–disease networks and a similar scale-free random network . The higher the score , the more likely the inference network has atypical connectivity ( see original publication [38] for details ) and the higher the probability of possible relationship between drug and disease . If drug is not known to cause ADE according to data from our previous study ( see above ) [37] we took IS from Comparative Toxicogenomics Database for corresponding disease; however if the drug is known to cause ADE we took the maximal value of ISs among all drugs . It was done because many drugs , which have description of cardiovascular ADEs in “Boxed Warning” and “Warnings and Precautions” sections of drug labels , demonstrate low ISs due to insufficient information on target genes in literature . To describe pairs of drugs with corresponding ISs we used sums of the scores of individual drugs . We calculated AUC values for each of the five datasets based on ISs for corresponding diseases ( ventricular tachycardia , myocardial infarction , ischemic stroke , arterial hypertension and cardiac failure ) . We proposed that the values of AUC reflect the quality of datasets . Interactions of individual drugs with human proteins were predicted by the PASS Targets software [39] . PASS ( Prediction of Activity Spectra for Substances ) [43–45] can be used for the prediction of various types of biological activities and is associated with several hundred success stories of its practical application , with experimental confirmation of the prediction results [45 , 46] . It uses Multilevel Neighborhoods of Atoms ( MNA ) descriptors and the Bayesian approach and is available as a desktop program as well as a freely available web service on the Way2Drug platform ( http://www . way2drug . com/PASSOnline/ ) [47] . PASS Targets is a special version of PASS that is based on training data from the ChEMBL database ( https://www . ebi . ac . uk/chembl/ ) and allows for predicting interactions with 1553 human protein targets with an average AUC 0 . 97 and a minimal AUC 0 . 85 [39] . The full list of human targets is presented in S3 Table . PASS provides two estimates of probabilities for each target of a chemical compound: The Pa probability to interact with a target , and the Pi probability to not interact with a target . If a compound has Pa > Pi , it can be considered as interacting with the target . The larger the Pa and Pa−Pi values , the greater the probability of obtaining an activity against a target in the experiment . In this study , we used a threshold Pa>0 . 3 for the estimation of protein targets of drugs from the top 10 scored drug pairs potentially causing ventricular tachycardia ( see the last section of the Results and Discussion ) . We used sums and absolute values of differences of Pa/ ( Pa+Pi ) values , calculated by PASS for individual drugs , to obtain corresponding values for pairs of drugs . Thus , each drug pair was described by a vector of 3106 values , which were further used as descriptors for the creation of classification models ( see below ) . Classification models for the prediction of five DDI-induced cardiovascular ADEs were created by the Random Forest method . We used the RandomForest function from “RandomForest” R package ( https://cran . r-project . org/web/packages/randomForest/ ) for this purpose . All arguments of this function were set to default . Since the training sets were imbalanced ( see Table 1 ) which is a problem for the creation of accurate classification models we used multiple under-sampling procedure when majority class of the training set was randomly sampled up to the size of the minority class . This process was repeated multiple times , and prediction probabilities from multiple models were averaged . The applicability domain of the obtained models was determined by the local ( Tree ) approach , which was described earlier [40] . The accuracy of created models was determined by a 5-fold cross validation procedure according to the “compound out” approach , wherein each drug pair in the test set must contain at least one drug that is absent in all drug pairs of the training set [41] . The accuracies of the models for ventricular tachycardia and arterial hypertension were also estimated on two external test sets generated based on information from DrugBank ( https://www . drugbank . ca/ ) database . The database contains some data on known DDIs that lead to ventricular tachycardia ( or prolongation of the QT interval on an electrocardiogram ) and arterial hypertension . These DDIs were extracted from drug labels and scientific publications by DrugBank team . We used this data as positive examples to create external tests sets . To create negative examples , we randomly generated drug pairs in the amounts equal to positive examples . We did not include as negative examples those pairs , where both individual drugs cause corresponding ADE according to data from our previous study [37] ( see above ) , as potentially false negatives .
|
Assessment of adverse drug effects as well as the influence of drug-drug interactions on their manifestation is a nontrivial task that requires numerous experimental and clinical studies . We developed a computational approach for the prediction of adverse effects that are induced by drug-drug interactions , which are based on a combined analysis of spontaneous reports and predicted drug-target interactions . Importantly , the approach requires only structural formulas to predict adverse effects , and , therefore , may be applied for new , insufficiently studied drugs . We applied the approach to predict five of the most important cardiovascular adverse effects , because they are the most dangerous and widespread . These effects are myocardial infarction , ischemic stroke , ventricular tachycardia , arterial hypertension and cardiac failure . The accuracies of predictive models were relatively high , in the range of 73–81%; therefore , as example , we performed a prediction of the five cardiovascular adverse effects for the large number of drug pairs and revealed the combinations that may potentially cause ventricular tachycardia along with potential molecular mechanisms . We consider that the developed approach can be used for the identification of pairwise drug combinations that are potentially the most or least dangerous for the cardiovascular system .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Materials",
"and",
"methods"
] |
[
"heart",
"rate",
"myocardial",
"infarction",
"stroke",
"medicine",
"and",
"health",
"sciences",
"cerebrovascular",
"diseases",
"drug-drug",
"interactions",
"drug",
"research",
"and",
"development",
"vascular",
"medicine",
"ischemic",
"stroke",
"neurology",
"tachycardia",
"pharmacology",
"hypertension",
"blood",
"pressure",
"drug",
"interactions",
"drug",
"information",
"cardiology"
] |
2019
|
Assessment of the cardiovascular adverse effects of drug-drug interactions through a combined analysis of spontaneous reports and predicted drug-target interactions
|
Vibrio cholerae and a subset of other Gram-negative bacteria , including Acinetobacter baumannii , express proteins with a C-terminal tripartite domain called GlyGly-CTERM , which consists of a motif rich in glycines and serines , followed by a hydrophobic region and positively charged residues . Here we show that VesB , a V . cholerae serine protease , requires the GlyGly-CTERM domain , the intramembrane rhomboid-like protease rhombosortase , and the type II secretion system ( T2SS ) for localization at the cell surface . VesB is cleaved by rhombosortase to expose the second glycine residue of the GlyGly-CTERM motif , which is then conjugated to a glycerophosphoethanolamine-containing moiety prior to engagement with the T2SS and outer membrane translocation . In support of this , VesB accumulates intracellularly in the absence of the T2SS , and surface-associated VesB activity is no longer detected when the rhombosortase gene is inactivated . In turn , when VesB is expressed without an intact GlyGly-CTERM domain , VesB is released to the extracellular milieu by the T2SS and does not accumulate on the cell surface . Collectively , our findings suggest that the posttranslational modification of the GlyGly-CTERM domain is essential for cell surface localization of VesB and other proteins expressed with this tripartite extension .
The type II secretion system ( T2SS ) is a multi-protein complex used by many Gram-negative bacteria to secrete extracellular proteins [1–4] . Most notably , Vibrio cholerae , the causative agent of cholera , uses the T2SS to secrete cholera toxin [5 , 6] . Cholera toxin and other T2SS substrates are secreted in a two-step process . First , proteins translocate across the inner membrane via recognition of their signal peptide by the Sec or Tat systems . Then , the secretion intermediates fold , engage with the T2SS and traverse the channel formed by the outer membrane embedded secretin [7–9] . These substrates are then free to diffuse away from the cell . In addition to cholera toxin , V . cholerae transports a number of other proteins , including three serine proteases , VesA , VesB and VesC , across the outer membrane via the T2SS [6 , 10] . VesB , the focus of this study , is a trypsin-like serine protease that contains an N-terminal signal peptide , a protease domain , and an immunoglobulin-like domain [11] ( Fig 1A ) . The protease domain is 30% identical to trypsin and includes a typical His-Asp-Ser catalytic triad and activation site . VesB is made as a zymogen and cleavage at the activation site ( Arg32-Ile33 ) results in the removal of the N-terminal propeptide to generate active VesB [11] . VesB , along with five other V . cholerae proteins , VesC , VesA , Xds , VCA0065 , and VC1485 also contain a C-terminal domain called GlyGly-CTERM [12] ( Fig 1B ) . Most , if not all , of these so-called GlyGly-CTERM proteins are hydrolytic enzymes . Proteolytic activities of VesA , VesB and VesC have been verified using low molecular weight fluorogenic peptides [10 , 11] . VesA and VesB are expressed and active in intestinal cecal fluid during V . cholerae colonization of rabbits as determined by an activity-based protein profiling technique [13] . While the range of their natural substrates has yet to be determined , VesA and VesB are capable of cleaving the A subunit of cholera toxin , thus separating the A1 and A2 domains required for toxin activation [10] , and VesC induces a hemorrhagic fluid response when injected into rabbit and mouse ileal loops [14 , 15] . Xds is the most well characterized of the GlyGly-CTERM containing proteins . Through its nuclease activity Xds generates nutrients , combats neutrophil extracellular traps , and modulates biofilm [16–18] . Lastly , VCA0065 is a RpoH-regulated putative Zn-metalloprotease , while VC1485 is a protein with unknown function that may be essential for growth [19–22] . The novel GlyGly-CTERM domain has a consensus tripartite motif that contains two prominent glycines surrounded by serines , followed by a hydrophobic helix and positively charged residues ( Fig 1B ) . It was originally identified bioinformatically in Gram-negative bacteria of the Vibrio , Shewanella , Acinetobacter and Ralstonia genera [12] . Through in silico partial phylogenetic profiling , proteins that have a GlyGly-CTERM extension were identified to co-exist with a putative intramembrane protease , rhombosortase , while GlyGly-CTERM proteins are absent in bacteria lacking rhombosortase [12] . In addition , in species encoding only one GlyGly-CTERM protein , the gene is linked to the gene coding for rhombosortase . This co-distribution led the authors to speculate that rhombosortase , if expressed as an active enzyme , may target the C-terminal extension of the GlyGly-CTERM proteins [12] . However , besides this bioinformatics analysis , the relationship between GlyGly-CTERM and rhombosortase has not been experimentally validated . Rhombosortase is homologous to rhomboid proteases , a class of membrane-embedded serine proteases that cleave single pass transmembrane proteins at or within the plane of the lipid bilayer [23–29] . X-ray crystallography studies of the rhomboid protease GlpG indicate that it is comprised of six transmembrane domains and a cavity within its folded form that contains the catalytic residues Ser and His [23–26] . Rhomboid proteases are nearly ubiquitous and function in a wide range of processes [30–32] . They have been implicated in human and parasitic diseases; however , only one rhomboid protease substrate has been identified thus far in bacteria . The Providencia stuartii rhomboid protease AarA cleaves off a short N-terminal peptide and thereby activates the channel-forming TatA of the twin-arginine translocation system [33 , 34] . In this study , we address the role of the T2SS , rhombosortase and the GlyGly-CTERM domain in VesB biogenesis . Analysis of the localization and activity of VesB in different mutant backgrounds reveals that rhombosortase cleaves the GlyGly-CTERM domain and promotes T2SS-dependent surface localization of active VesB .
Previously , we have shown that while VesB is a type II secreted protease with its N-terminal signal peptide and propeptide removed [10 , 11] , immunoblotting of culture supernatants and cells indicated that a large fraction of VesB remains cell-associated . To determine whether the cell-associated form is active , we expressed plasmid-encoded VesB in a strain lacking the genes for the three related serine proteases VesA , VesB and VesC ( ΔvesABC [10] ) . We included two additional strains , one containing an empty vector and one expressing an active site mutant protein , VesB-S221A , to serve as negative controls . Cultures were grown in the presence of 50 μM IPTG in LB for 4 hours , after which culture supernatants and cells were separated by centrifugation and the VesB activity was determined . When using the fluorogenic peptide Boc-Gln-Ala-Arg-AMC as a substrate , VesB activity was found in the culture supernatant and in association with intact cells ( Fig 1C ) , suggesting that both the secreted and cell-associated forms of VesB are proteolytically active . While VesB-S221A was inactive in this assay , it could be detected in culture supernatant and in association with cells with anti-VesB antibodies ( Fig 1D ) . Like other trypsin-like proteases , VesB is produced as a zymogen and removal of the propeptide through processing at Arg32-Ile33 results in activation of VesB [11] . The finding that VesB-S221A migrated slower compared to VesB when analyzed by SDS-PAGE and immunoblotting indicated that it still contains the propeptide , suggesting that VesB is an autoactivating enzyme ( Fig 1D ) . Further support for this came from comparing the size of plasmid-expressed VesB-S221A in the absence and presence of native VesB . When co-expressed with native VesB in a wild type ( WT ) strain background , VesB-S221A migrated on SDS-PAGE indistinguishably from that of WT VesB ( Fig 1E ) . Because we detected cell-associated VesB activity , we hypothesized that VesB may be surface-localized in T2SS competent cells , but remains internally trapped in T2SS mutant cells ( Δeps ) . To test this , we expressed plasmid-encoded VesB in the ΔvesB and ΔvesBΔeps strains and subjected the cells to surface labeling with anti-VesB antibodies and Alexa Fluor 488 F ( ab′ ) 2 goat anti-rabbit IgG . The cultures were grown in M9 medium supplemented with glucose and casamino acids to avoid autofluorescence observed with LB grown cells . Under this growth condition , the majority of VesB protein and activity was cell-associated in the ΔvesB/pVesB strain and very little VesB was found in the culture supernatant ( Fig 2A & 2B ) . In contrast , in the ΔvesBΔeps/pVesB strain , while the majority of VesB remained cell-associated , it was not active ( Fig 2A ) . Furthermore , the plasmid-encoded VesB migrated slightly slower in the ΔvesBΔeps strain than in the ΔvesB strain ( Fig 2B ) , suggesting that it may retain the N-terminal propeptide and that zymogen activation of VesB by cleavage of the Arg32Ile33 bond likely occurs after outer membrane translocation . For the VesB surface labeling experiment , the ΔvesB and ΔvesBΔeps strains with empty vector served as negative controls . After thorough washing , the labeling efficiency was quantitated by fluorimetry of cells in suspension and the fluorescent units were normalized to the OD600 of the original cultures . The fluorescence measured for ΔvesB/pVesB cells was 4 . 8-fold higher when compared to ΔvesBΔeps/pVesB cells , suggesting that VesB is surface-localized in WT cells , while it primarily remains intracellular in the Δeps cells ( Fig 2C ) . The cells were also visualized by fluorescence microscopy ( Fig 2D ) . Quantitative analysis indicated that 94% of ΔvesB/pVesB cells were fluorescent , whereas there was no detectable fluorescence observed with ΔvesBΔeps/pVesB mutant cells . Similarly , the catalytically inactive VesB-S221A was also localized to the surface of cells with an intact T2SS but remained intracellular in Δeps cells ( Fig 2C and 2D ) . This suggests that activation of VesB is not required for surface localization . Taken together , our results suggest that outer membrane translocation by the T2SS is required for VesB localization to the surface of WT cells , and that following translocation VesB autoactivates . Through partial phylogenetic profiling , a previous study identified a putative rhomboid protease , rhombosortase , to be co-distributed with proteins containing a GlyGly-CTERM sequence , leading to the speculation that rhombosortase may cleave these proteins [12] . To determine the relationship between VesB and the rhombosortase protease RssP of V . cholerae , we inactivated the rssP gene and subjected culture supernatants and cells to SDS-PAGE and immunoblotting with VesB antibodies . To simplify the analysis and keep the number of processing events to a minimum , we initially used VesB-S221A , which does not undergo autoactivation ( Fig 1C and 1D ) . Culture supernatants and cells were separated following growth of the various mutants in the presence of 50 μM IPTG in M9 media supplemented with glucose and casamino acids , and putative size differences of VesB-S221A were analyzed by SDS-PAGE and immunoblotting ( Fig 3A ) . Because VesB is less stable in rssP::kan backgrounds it was necessary to load differing amounts of sample from the various strains in order to clearly visualize potential size differences of VesB-S221A . Accordingly , we loaded ten times less of the samples in lanes 1 , 2 , 10 and 12 ( Fig 3A ) . When the rhombosortase gene was inactivated ( rssP::kan ) , the total amount of VesB-S221A was reduced and a portion migrated slower than VesB-S221A in cells expressing a functional rhombosortase ( Fig 3A , compare lanes 2 and 4 ) . This unprocessed form of VesB-S221A was exclusively detected in the cell fraction ( Fig 3A , lane 4 ) . Interestingly , a small amount of VesB-S221A was still cleaved and released extracellularly ( Fig 3A , lane 3 ) ; however , this cleaved form was no longer released when the T2SS was inactivated in the rssP::kanΔeps double mutant ( Fig 3A , lane 5 ) . While these results suggested that rhombosortase cleaves VesB-S221A , we speculated that the residual processing of VesB-S221A in the rssP mutant was due to the activity of the ubiquitous rhomboid protease GlpG , which shares 32% amino acid sequence identity with RssP . To test this , we made a double mutant , where both glpG and rssP were inactivated ( rssP::kanΔglpG ) and found that in this background VesB-S221A remained associated with the cells in an un-processed form ( Fig 3A , lane 8 ) , suggesting that GlpG is capable of cleaving VesB-S221A when RssP is absent , albeit inefficiently . As VesB-S221A was completely cleaved in the ΔepsΔglpG double mutant ( Fig 3A , lanes 9 and 10 ) , it suggests that RssP is the primary protease that cleaves VesB . The results shown in lanes 9–12 also indicate that RssP-mediated processing of VesB does not require the presence of a functional T2SS and suggests that cleavage likely occurs before VesB engages with the T2SS . To establish the functional consequences of VesB processing , we next analyzed WT VesB expressed from its native promoter . Specifically , we wanted to determine whether generation of active VesB is dependent on cleavage by rhombosortase . Cultures were grown in LB , and culture supernatants and intact cells were analyzed for VesB activity . In both WT and ΔglpG strains , VesB was active in the culture supernatants and in association with cells ( Fig 3B ) , further confirming that GlpG does not play a significant role in VesB biogenesis when rhombosortase is present . In the experiment shown in Fig 3A , VesB-S221A was found to be unstable and detected in two forms when overexpressed in the rssP::kan mutant and only the GlpG-cleaved form was found in the culture supernatant . Here , when WT VesB was expressed from its native promoter in the rssP::kan mutant grown in LB no VesB activity was detected ( Fig 3B ) . This is consistent with the earlier finding that VesB is unstable in the absence of rhombosortase , but also suggests that the GlpG-cleaved form of VesB is inactive , and therefore , that the C-terminal cleavage of VesB by rhombosortase is required for the correct maturation of active VesB . Since VesC and VesA collectively contribute to approximately 20% of the proteolytic activity towards Boc-Gln-Ala-Arg-AMC [10] , the lack of activity for the rssP::kan mutant suggests that VesC and VesA are also substrates of rhombosortase . When the rssP::kan mutant was complemented with the plasmid pRssP , we observed restoration of protease activity in both the culture supernatant and cells ( Fig 3B ) . In contrast , no complementation of the rssP::kan mutant was apparent in the presence of pRssP-S102A and pRssP-H160A , which express mutant forms of rhombosortase in which the predicted catalytic residues serine and histidine , respectively , are substituted with alanine ( Fig 3B ) . To further analyze the effect of inactivating the rssP gene , we took a proteomics approach that involved isolating and concentrating the culture supernatants of the WT and rssP::kan strains and subjecting them to off-line SDS-PAGE , gel segmentation , in-gel digestion with trypsin , and LC/MS/MS . Many proteins were detected using this method; however , we focused on VesB and the other five GlyGly-CTERM containing proteins . The proteomic data were used to compare the relative amount of the GlyGly-CTERM containing proteins in the culture supernatant of the WT and rssP::kan strains following growth in LB , a growth condition that supports the expression of all six GlyGly-CTERM proteins . We hypothesized that the amount of the GlyGly-CTERM containing proteins would be reduced in the culture supernatant of the rssP::kan mutant unless they are cleaved by GlpG , similarly to VesB , in the absence of rhombosortase . Spectral counts for each protein were normalized to total number of spectra in each sample and protein size to obtain Normalized Spectral Abundance Factor ( NSAF ) values [35] for each protein in the two samples ( Table 1 ) . The results showed that the amount of VesA , VesC , Xds , and VCA0065 was significantly reduced in the absence of RssP by four-fold or more ( Table 1 ) . Consistent with results from our immunoblotting experiments , VesB was only reduced by a factor of 1 . 5 in the rssP::kan mutant , while the level of VC1485 was increased by a factor of 2 . Neither of these latter results was statistically significant and is likely due to GlpG’s ability to cleave and thus release VesB , and perhaps also VC1485 , to the culture supernatant when rhombosortase is absent . To identify the rhombosortase cleavage site in VesB and to begin to address the mechanism by which VesB is associated with the surface of V . cholerae , we subjected both the secreted and surface associated forms of VesB to affinity purification and mass spectrometry analysis . First , we grew V . cholerae into stationary phase in LB to maximize the yield of overexpressed VesB . Following removal of cells , VesB was purified from the culture supernatant on benzamidine sepharose as described previously [11] . N-terminal sequencing confirmed that the signal peptide and propeptide were removed . Purified VesB was then subjected to reversed phase chromatography and electrospray mass spectrometry ( ESI-LC/MS ) to obtain its intact molecular mass . Two major peaks representing masses of 37 , 583 and 37 , 429 Da were obtained along with two minor peaks representing species of 37 , 185 Da and 37 , 027 Da , respectively ( Fig 4A; left panel ) . An additional species , 38 , 086 Da ( Fig 4A; right panel ) eluted at a higher percentage of the organic solvent acetonitrile during HPLC , suggesting that it is more hydrophobic than the others . As the predicted mass of active VesB with its signal peptide and propeptide removed , but with an intact C-terminus would be 39 . 7 kDa , these masses are consistent with rhombosortase cleaving off part of the C-terminus of VesB . The two smallest species of 37 , 185 Da and 37 , 027 Da ( marked with asterisks in Fig 4A; left panel ) matched within +/- 0 . 01% mass accuracy to molecular species that have either 23 or 25 aa removed from the C-terminus . This represents cleavage of VesB at either Ser378Ala379 or Ser380Ser381 immediately upstream of the GlyGly-CTERM domain , which starts at Gly382 ( Fig 4A; left panel ) . The two larger species with masses 37 , 583 and 37 , 429 Da did not match with theoretical cleavage products and may represent VesB with additional posttranslational modifications . Purified VesB was also subjected to SDS-PAGE , in-gel trypsin digestion , and LC-MS/MS analysis . Peptide mapping gave a 100% sequence coverage between residues Arg32 and Ser382 . MS/MS analysis revealed a ragged C-terminus of VesB ( Fig 4B ) , consistent with the results from the intact mass analysis described above and with VesB being subject to C-terminal proteolysis . As expected , no peptides containing the intact GlyGly-CTERM were observed . Interestingly , spectra representing modified forms of the C-terminal peptide , IQLDTSPFAPSASSGG , were present . The fragmentation pattern for this peptide showed modification of the C-terminal glycine with either a 43 Da or 197 Da moiety ( exemplified in Fig 4C and 4D ) . The presence of the 43 Da and 197 Da moieties on Gly383 generates theoretical intact VesB masses of 37 , 430 Da and 37 , 583 Da , respectively; consistent with the major peaks observed in Fig 4A ( left panel ) . Taken together , the two mass spectrometry approaches suggest that VesB primarily undergoes posttranslational modification at Gly383 . The additional sites of cleavage upstream of the GlyGly-CTERM domain may be the result of proteolysis by extracellular proteases during prolonged growth in LB and/or purification . Support for this suggestion was obtained when VesB was purified at ambient temperature ( instead of 4° C ) and in the absence of proteinase inhibitors , which yielded the 37 , 027 Da form of VesB as the dominating species ( S1 Fig ) . Next , we extracted and purified active cell-associated VesB following growth in M9 medium , which maximizes the yield of the membrane bound form of VesB . While treatment of cells with high salt ( 2M NaCl ) , glycine buffer ( pH 2 . 5 ) or carbonate buffer ( pH 11 ) did not result in appreciable amount of VesB extraction , the non-ionic detergent Triton X-100 was capable of efficiently removing VesB from V . cholerae cells suggesting that VesB may be associated with cells via protein-lipid or lipid-lipid interaction . We extracted the membrane bound form of VesB with Triton X-100 , purified it by benzamidine affinity chromatography and subjected it to SDS-PAGE , in-gel trypsin digestion and LC-MC/MS analysis . Again , we detected two species of the C-terminal peptide IQLDTSPFAPSASSGG with either a 43 Da or 197 Da modification of the terminal Gly383 ( Fig 4E and 4F ) . The 43 and 197 Da modifications of the terminal Gly383 are consistent with ethanolamine ( 61 Da minus the loss of a water molecule ) and glycerophosphoethanolamine ( 215 Da minus the loss of H2O ) , suggesting that VesB is attached to a phosphoethanolamine moiety ( Fig 4G ) . A peptide with an intact phospholipid such as phosphatidylethanolamine was not identified possibly due to unwanted fragmentation in the mass spectrometer . Nevertheless , the 38 , 086 Da species detected by intact mass analysis of extracellular VesB ( Fig 4A; right panel ) is consistent with VesB being modified with phosphatidylethanolamine containing C16 and C18 fatty acids at Gly383 . The presence of a modified VesB species containing an intact phosphatidylethanolamine among extracellularly released VesB may be due to VesB’s association with outer membrane vesicles ( OMVs ) , a finding recently reported by Mekalanos and colleagues [36] . To test this , we subjected filtered culture supernatant from overnight cultures of V . cholerae overexpressing VesB to high speed centrifugation , SDS-PAGE and immunoblot analysis . In addition to the major outer membrane protein OmpU , VesB was detected in the pellet ( Fig 5 , lane 3 ) , suggesting that a fraction of extracellular VesB is pelletable and likely associated with OMVs or OM fragments . To further address the importance of the GlyGly-CTERM extension , we determined the functional consequences of expressing VesB without an intact GlyGly-CTERM domain . Two VesB variants , VesBΔ5 and VesBΔ30 , were constructed and their subcellular location and activity were examined . VesBΔ5 is lacking the C-terminal five residues including the positively charged residues and VesBΔ30 has the additional Gly/Ser rich motif and hydrophobic region removed . These constructs and WT VesB were expressed in ΔvesABC cells and at various time points , culture supernatants and cells were collected and analyzed for localization and activity . While VesBΔ5 displayed a slightly higher activity than VesB in the culture supernatant ( Fig 6A ) , no cell-associated activity could be detected ( Fig 6B ) . VesBΔ30 activity was only observed in the 8 . 5h culture supernatant sample ( Fig 6A ) . The immunoblots correlated well with the activity assays , in that VesB was present in the culture supernatant and cells , while VesBΔ5 was found mostly in the supernatant with a small amount cell-associated . The total yield of VesBΔ30 was lower and this truncated form of VesB was detected exclusively in the supernatant ( Fig 6C ) . The change in subcellular distribution of VesBΔ5 andVesBΔ30 suggests that the presence and the subsequent removal of the GlyGly-CTERM domain by rhombosortase are essential for the cell surface retention of VesB . Consistent with the finding that VesBΔ5 and VesBΔ30 are primarily released from the cell was the finding that little to none of these truncated proteins were associated with OMVs ( Fig 5; lanes 6 and 9 ) . In summary , the GlyGly-CTERM domain and processing by rhombosortase are prerequisites for surface localization of VesB . To further address the requirement for processing by rhombosortase , we fused the GlyGly-CTERM domain from VesB to the C-terminus of the E . coli heat-labile enterotoxin B subunit , EtxB . While EtxB , a homolog of the cholera toxin B subunit , was secreted by V . cholerae as has been shown previously [6 , 37] , EtxB-GlyGly-CTERM remained associated with the cells ( Fig 7 ) . When expressed in the rssP::kan mutant , EtxB-GlyGly-CTERM was barely detected , but appeared to be larger in size , suggesting that it may retain the GlyGly-CTERM domain in the absence of rhombosortase . This result suggests that EtxB-GlyGly-CTERM is recognized and processed by rhombosortase; however , whether fusing a GlyGly-CTERM domain to any protein with an N-terminal signal peptide will result in a hybrid protein that is targeted by rhombosortase needs to be addressed in much greater detail .
In this study , we used the trypsin-like serine protease VesB as a model protein to determine the relationship between the GlyGly-CTERM domain , rhombosortase , and the T2SS in V . cholerae . T2S substrates translocate through the outer membrane and are either released into the extracellular milieu or retained on the cell surface through a variety of mechanisms . We show here that the majority of VesB is surface-associated; however , depending on growth conditions , various amounts of the surface-localized VesB are released to the extracellular space . Rhombosortase cleaves off the GlyGly-CTERM domain and the newly generated C-terminus is further modified to retain VesB on the cell surface once transported through the T2SS . When rhombosortase is absent , full-length VesB is largely subjected to degradation , but a small fraction is cleaved by GlpG . However , there is likely no additional posttranslational modification of GlpG-cleaved VesB and , therefore , GlpG-cleaved VesB is released to the extracellular milieu . Additionally , rhombosortase-mediated C-terminal processing leads to subsequent VesB auto-activation , while cleavage by GlpG in the absence of rhombosortase results in an inactive VesB , indicating that either correct C-terminal processing of VesB and/or its localization to the outer membrane are required for its auto-activation . Taking all of the data together , we have assembled a model for successful delivery of VesB to the cell surface of V . cholerae that involves its GlyGly-CTERM domain , rhombosortase and the T2SS ( Fig 8 ) . VesB is synthesized in the cytoplasm with an N-terminal signal peptide and a C-terminal GlyGly-CTERM domain . VesB enters the Sec system via the signal peptide and , as the protein is translocated through the Sec system , the GlyGly-CTERM domain is positioned in the inner membrane . The signal peptide is cleaved off and VesB folds . Rhombosortase cleaves the GlyGly-CTERM domain and possibly further modifies the newly generated VesB C-terminus by attaching it to a glycerophosphoethanolamine containing lipid ( possibly phosphatidylethanolamine ) via transamidation . It is also possible that a second enzyme is responsible for this posttranslational modification event; however , we have no evidence for a separate transamidase enzyme at this time . VesB is then translocated from the inner membrane by the T2SS and delivered to the cell surface . Once at the cell surface , we speculate that the propeptide is removed in trans by a nearby active VesB resulting in auto-activation . Under some growth conditions , surface-localized VesB may then be released through outer membrane blebbing and/or detachment by extracellular protease ( s ) . The rhombosortase/GlyGly-CTERM system may be compared to the sortase/LPXTG system of Gram-positive organisms and the archaeosortase/PGF-CTERM system of archaea . In all three systems , substrates contain a C-terminal tripartite domain consisting of a recognition motif ( GG , LPXTG or PGF ) followed by a hydrophobic region and positively charged residues [38 , 39] . For example , in Staphylococcus aureus , sortase cleaves staphylococcal protein A ( SPA ) , a LPXTG substrate , between the Thr and Gly residues , resulting in an acyl-enzyme intermediate that is resolved by lipid II , a precursor of peptidoglycan ( PG ) , instead of H2O [40] . This transamidation process results in PG-anchoring of SPA [40] . VesB may also undergo C-terminal transamidation rather than complete hydrolysis , where the terminal NH2 group of an ethanolamine-containing component serves as the attacking nucleophile instead of H2O , which is scarce in the membrane environment . The detection of a species containing a 197-Da modification of the C-terminal glycine in VesB is consistent with a glycerophosphoethanolamine modification . Because extraction of VesB from the surface of V . cholerae requires a detergent , it is possible that VesB is attached to a glycerophosphoethanolamine containing lipid , possibly phosphatidylethanolamine generating a size of VesB that is consistent with the protein peak in Fig 4A; right panel . While this is the first time that this type of posttranslational modification is described in prokaryotes , Atg8 , a ubiquitin-like protein required for autophagosome formation in yeast , is cleaved at the C-terminus by a cysteine protease to expose a glycine that is subsequently attached to phosphatidylethanolamine ( PE ) via amidation [41 , 42] . When Atg8-PE was analyzed by mass spectrometry a C-terminal 197 Da glycerophosphoethanolamine moiety was also detected at a C-terminal glycine . There is also some resemblance between surface localization of VesB and glycosylphosphatidylinositol ( GPI ) -anchored proteins in eukaryotes [43] , as both processes involve the removal of a trans-membrane domain and attachment of the cleaved protein to a lipid via a phosphoethanolamine linker . Further studies are planned to elucidate the mechanism of C-terminal modification of VesB and to determine whether rhombosortase has transamidase activity . Rhombosortase does form a distinct group with other rhombosortases within the greater rhomboid protease family . Rhombosortase is only 23% identical to E . coli GlpG and it is lacking the 90 amino acid N-terminal cytoplasmic domain , whose function has yet to be determined . This globular domain is present in most rhomboid proteases except Haemophilus influenzae GlpG and P . stuartii AarA , and in the mitochondrial PARL proteases it is replaced by a transmembrane domain [24 , 44] . It is possible that sequence differences and the lack of the cytoplasmic domain favor transamidation over hydrolysis by rhombosortase . For example , the active site of rhombosortase may be less exposed to water than that of rhomboid protease . As a consequence , the transient acyl-enzyme intermediate , formed when the serine hydroxy group of rhombosortase attaches to the acyl moiety of VesB , will react with the nucleophilic NH2 group of ethanolamine instead of water [45] . Under some conditions , we observe more VesB in the supernatant than others ( for example , in LB vs . M9 medium containing casamino acids and glucose ) . Similar findings have been reported for other surface-attached proteins . For example , it was recently shown that the cell wall-attached sortase substrate SPA is released into the culture supernatant by murein hydrolases [46] . Additionally , GPI-anchored proteins , like prostasin , can be released from the cell surface by lipid-targeting enzymes like phospholipase C [47] . Surface-localized VesB could similarly undergo processing by an enzyme such as an extracellular protease that removes it from the cell surface . Our results suggest that release of VesB may also occur via OM blebbing . The dual locations of VesB pose an interesting question of why a fraction of the protein is released to the extracellular space . While it is possible that the release of surface-localized VesB is a laboratory-induced artifact , dual locations have also been reported for the GlyGly-CTERM containing protein ExeM required for biofilm formation by Shewanella oneidensis . While ExeM has been found as an active nuclease in the culture supernatant of S . oneidensis [48] , other studies involving proteomic analysis of membranes have identified ExeM as a membrane-associated protein [49 , 50] . V . cholerae Xds , an ExeM homolog , displays a similar distribution , as Xds nuclease activity has been detected in both culture supernatant and in association with cells [17 , 51] . VesB and other GlyGly-CTERM proteins , which are mostly hydrolytic enzymes , may be of benefit to the individual bacterium when retained on the surface where they can break down macromolecules and generate nutrients such as short peptides for immediate cellular uptake . On the other hand , in the context of a biofilm , the released form of these proteins could benefit the community at large . In summary , VesB has a novel GlyGly-CTERM domain and utilizes rhombosortase and the T2SS to be correctly processed , translocated across the cell envelope , and retained on the cell surface of V . cholerae . The rhombosortase/GlyGly-CTERM system offers a new alternative method of surface association of T2S substrates that differs from previously observed mechanisms . PnlH from D . dadantii possesses a non-cleavable TAT specific signal peptide that is needed for its outer membrane retention [52] , heat-labile enterotoxin from ETEC is localized on the cell surface via an interaction with lipopolysaccharides [53] and the lipidated N-terminus of pullulanase keeps it surface-associated in K . oxytoca through an unknown mechanism [54] . Furthermore , the newly acquired knowledge of the GlyGly-CTERM domain of VesB provides insight to the subcellular distribution of ExeM and Xds from S . oneidensis and V . cholerae , respectively . Based on our findings , a large fraction of these proteins are also likely retained on the cell surface in a process that involves their GlyGly-CTERM domains and rhombosortase [17 , 48–51] . Our findings may also provide insight into the mechanism of C-terminal processing of the PGF-CTERM domain of archaeal S-layer glycoprotein by archaeosortase . While the S-layer glycoprotein is lipid modified and can be extracted with Triton X-100 , it is not known where archaeosortase cleaves and the site of lipid modification has not yet been identified [38] .
The V . cholerae El Tor O1 strain , N16961 , and the isogenic Δeps [37] , ΔepsD [55] , ΔvesB and ΔvesABC [10] mutants were used in this study . All plasmids and primers are listed in Table 2 . All polymerase chain reactions ( PCR ) , cloning and restriction enzyme digestions were done with Phusion Polymerase , T4 DNA ligase and restriction enzymes from New England Biolabs and primers that were synthesized at IDT Technologies . pCRScript ( Stratagene ) and pMMB67EH constructs were transformed into E . coli MC1061 and pCVD442 constructs into SY327λpir . Triparental conjugation was performed with a helper strain , MM294/pRK2013 to transfer plasmids into N16961 and its isogenic mutants [56] . A kanamycin insertional rhombosortase mutant ( rssP::kan ) was created by amplifying the rhombosortase ( rssP ) gene ( VC1981 ) from V . cholerae N16961 chromosomal DNA using rssP primers . The fragment was ligated into a high copy plasmid , pCRscript . A kanamycin cassette was amplified from pK18mobsacB using kan primers containing BclI restriction enzyme sites and cloned into the native BclI site in rssP . The rssP::kan fragment was moved into the suicide plasmid , pCVD442 . The rhomboid protease gene from V . cholerae , glpG ( VC0099 ) was deleted from the chromosome by amplifying 500 base pair regions upstream and downstream from glpG followed by overlap extension PCR of the two fragments to generate a 1 . 0 kbp fragment that was cloned into pCVD442 . For complementation , rssP was amplified and cloned into pMMB67EH , a low copy vector with an IPTG inducible promoter and ampicillin cassette to generate pRssP . Using primers with base pair changes and the pRssP plasmid as template , fragments with rssP-S102A and rssP-H160A were made and cloned into pMMB67EH . vesBΔ5 and vesBΔ30 were amplified from pCRscript carrying vesB using newly created reverse primers with the forward primer originally used to clone vesB [57] . PCR fragments were cloned into pMMB67EH . Plasmids pMMB68 [58] and pVesB [57] and primers annealing to the ends of etxB and vesB , were used to generate EtxB fused to the GlyGly-CTERM extension of VesB . Strains were either grown on Luria-Bertani ( LB ) agar/broth ( Fisher ) or M9 media containing 0 . 4% glucose and 0 . 4% casamino acids with 100 μg/mL of carbenicillin ( Sigma ) when plasmids were present . ΔvesABC/pVesB was grown in the presence of 50 μM IPTG in M9 media supplemented with glucose and casamino acids for 4 h . Following the removal of culture supernatant , cells were resuspended and sonicated . After removing unlysed cells , the cell envelope was pelleted at 170 , 000 x g for 1h . Membrane pellet was resuspended in 50 mM Tris-HCl pH 8 . 0/450 mM NaCl buffer containing 2% Triton X-100 . Following another high speed centrifugation step , Triton X-100 soluble VesB was purified by benzamidine affinity chromatography [11] in the presence of 2% Triton X-100 . Following washing with the above buffer containing Triton X-100 , VesB was eluted with 10 mM benzamidine in the presence of Triton X-100 . Fractions containing purified VesB were collected , pooled precipitated with pyrogallol red-molybdate-methanol as described [10] , and subjected to SDS-PAGE , in-gel trypsin digestion and LC-MS/MS analysis as described below . Samples were prepared and analyzed by SDS-PAGE and immunoblotting as described previously [37] . Polyclonal antiserum against VesB [11] was incubated with culture supernatant from the ΔvesABC mutant for 1 h to pre-absorb cross-reactive antibodies prior to incubating with the nitrocellulose membrane for 2 h ( 1:5 , 000 ) . OmpU antibodies ( gift from K . Klose ) were used at 1:20 , 000 and incubated for 2 h and monoclonal EtxB antibody was used at 1:30 , 000 [37] . Horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G ( Bio-Rad ) used at 1:20 , 000 was incubated with the membrane for 1 h . Membranes were developed using ECL 2 Western blotting reagent ( Thermo Fisher ) and visualized using a Typhoon Trio variable mode imager system and Image Quant software . V . cholerae supernatants and whole cells were measured for protease activity using N-tert-butoxycarbonyl-Gln-Ala-Arg-7-amido-4-methylcoumarin as described previously [11] . Change in fluorescence per minute was calculated and converted to moles of methlycoumarin ( AMC ) cleaved per minute via a standard curve with known concentrations of AMC . The rate of AMC generation was normalized by OD600 of the cultures . Cells were washed , blocked with 2% BSA and incubated with 1:1000 of VesB antiserum that was pre-incubated with ΔvesABC cells to remove cross-reactive antibodies . Following incubation with 1:1000 of Alexa Fluor 488 F ( ab′ ) 2 goat anti-rabbit immunoglobulin G ( Invitrogen ) and washing , fluorescence was measured ( Ex 495 nm/ Em 519 nm ) . The cells were also visualized by differential interference contrast and fluorescent microscopy using a Nikon Eclipse 90i fluorescence microscope equipped with a Nikon Plan Apo VC 100× ( 1 . 4 numerical aperture ) oil immersion objective and a CoolSNAPHQ2 digital camera as previously described [57] . Intact mass analysis of VesB ( 10 μl of a 20-μM solution in 10 mM Tris-HCl ) , purified from V . cholerae culture supernatant as described [11] , was conducted on an Agilent 6224 ESI-TOF mass spectrometer in conjunction with an Agilent 1260 Infinity binary pump HPLC system as follows: mobile phase A , 0 . 1% formic acid in water; mobile phase B , 0 . 1% formic acid in 90% acetonitrile/10% water; gradient , 20% B to 90% B in 15 min; flow rate , 200 μl/min; positive ion detection; profile mode; mass range , m/z 100– m/z 3 , 200; fragmentor , 225 V; skimmer , 65 V; capillary voltage , 4 , 000 V; gas temperature , 325°C . The maximum entropy deconvolution algorithm in MassHunter BioConfirm ( Agilent ) was used for molecular mass determination of components in the sample . Culture supernatants of WT and rssP::kan strains were prepared as described with the following modifications [59] . Culture supernatants from three independent experiments were combined and 20 μg of protein was processed by SDS-PAGE . The gel was stained with InstantBlue ( Expedeon ) and excised into ten equal sized segments . Gel segments were digested with a ProGest robot ( DigiLab ) with the following protocol: washed with 25mM ammonium bicarbonate followed by acetonitrile , reduced with 10mM dithiothreitol at 60°C followed by alkylation with 50mM iodoacetamide at RT , digested with trypsin ( Promega ) at 37°C for 4h , quenched with formic acid , and the supernatant was analyzed directly without further processing . Each gel digest was analyzed by nano LC/MS/MS with a Waters NanoAcquity HPLC system interfaced to a ThermoFisher LTQ Orbitrap Velos Pro . Peptides were loaded on a trapping column and eluted over a 75μm analytical column at 350nL/min; both columns were packed with Jupiter Proteo resin ( Phenomenex ) . The mass spectrometer was operated in data-dependent mode , with MS performed in the Orbitrap at 60 , 000 FWHM resolution and MS/MS performed in the LTQ . The fifteen most abundant ions were selected for MS/MS . Data were searched using a local copy of Mascot ( Matrix Science , UK ) with the following parameters: Enzyme: Trypsin; Database: Swissprot V . cholerae El Tor N16961 ( concatenated forward and reverse plus common contaminants ) ; Fixed modifications: Carbamidomethyl ( C ) ; Variable modifications: Oxidation ( M ) , Acetyl ( N-term ) , Pyro-Glu ( N-term Q ) , Deamidation ( N , Q ) ; ethanolamine ( C-term ) , glycerophosphoethanolamine ( C-term ) and carbamyl ( C-term ) . Mascot DAT files were processed in Scaffold ( Proteome Software Inc . ) for determination of protein and peptide probabilities . Data were filtered using protein and peptide thresholds of 99% and 95% , respectively , and requiring at least two unique peptides per protein . The complete mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium ( http://proteomecentral . proteomexchange . org ) via the PRIDE partner repository [60] with the dataset identifier PXD000896 . Data for the six GlyGly-CTERM proteins are shown in Table 1 . In-gel trypsin digested VesB , purified from culture supernatant or Triton X-100 extracted membranes of V . cholerae , was analyzed by nano LC/MS/MS with a Proxeon EASY-nLC 1000 HPLC system interfaced to a ThermoFisher Q Exactive mass spectrometer . Data were searched using a local copy of Mascot and subsequently processed in Scaffold with the parameters described above . Additional Mascot searches were performed with the enzyme specified as “no enzyme” instead of trypsin , in order to search for truncated C-terminal peptides The data were deposited to the ProteomeXchange Consortium with the dataset identifier PXD003261 . Statistical significance between WT and mutant samples was assessed by Student’s T-test . Results yielding a P value of <0 . 05 were considered statistically significant .
|
The type II secretion system ( T2SS ) critically contributes to pathogenesis of many Gram-negative bacteria by transporting toxins and a variety of enzymes including proteases , lipases , nucleases and carbohydrases to the extracellular environment . Cholera toxin , the principal virulence factor of Vibrio cholerae , is an example of a T2SS substrate that employs this multi-protein complex for translocation across the outer membrane . While cholera toxin freely diffuses away from the bacterial cell following outer membrane translocation , other T2SS substrates remain surface associated or may reattach to the bacterial cell surface following extracellular release . We show here that a large fraction of VesB , another V . cholerae T2SS substrate , stays associated with the cells and that this depends on the posttranslational modification of its C-terminal domain , GlyGly-CTERM , by rhombosortase . Rhombosortase is a member of the nearly ubiquitous rhomboid family of intramembrane serine proteases . As the Providencia stuartii TatA is the only rhomboid protease substrate identified in bacteria to date , this study expands the list of known bacterial rhomboid protease substrates and likely includes all proteins expressed with GlyGly-CTERM extensions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"bacteriology",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"chemical",
"compounds",
"aliphatic",
"amino",
"acids",
"molecular",
"probe",
"techniques",
"enzymes",
"pathogens",
"vibrio",
"microbiology",
"enzymology",
"immunoblotting",
"organic",
"compounds",
"serine",
"proteases",
"membrane",
"proteins",
"vibrio",
"cholerae",
"secretion",
"systems",
"outer",
"membrane",
"proteins",
"amino",
"acids",
"molecular",
"biology",
"techniques",
"cellular",
"structures",
"and",
"organelles",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"glycine",
"microbial",
"physiology",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"chemistry",
"molecular",
"biology",
"cell",
"membranes",
"biochemistry",
"bacterial",
"physiology",
"cell",
"biology",
"post-translational",
"modification",
"virulence",
"factors",
"organic",
"chemistry",
"biology",
"and",
"life",
"sciences",
"proteases",
"physical",
"sciences",
"signal",
"peptides",
"organisms"
] |
2018
|
C-terminal processing of GlyGly-CTERM containing proteins by rhombosortase in Vibrio cholerae
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.