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 |
|---|---|---|---|---|---|
Zoonotic visceral leishmaniasis is a vector-borne disease caused by Leishmania infantum in the Mediterranean Basin , where domestic dogs and wild canids are the main reservoirs . The promastigote stage replicates and develops within the gut of blood-sucking phlebotomine sand flies . Mature promastigotes are injected in the dermis of the mammalian host and differentiate into the amastigote stage within parasitophorous vacuoles of phagocytic cells . The major vector of L . infantum in Spain is Phlebotomus perniciosus . Promastigotes are routinely axenized and cultured to mimic in vitro the conditions inside the insect gut , which allows for most molecular , cellular , immunological and therapeutical studies otherwise inviable . Culture passages are known to decrease infectivity , which is restored by passage through laboratory animals . The most appropriate source of promastigotes is the gut of the vector host but isolation of the parasite is technically challenging . In fact , this option is not viable unless small samples are sufficient for downstream applications like promastigote cultures and nucleic acid amplification . In this study , in vitro infectivity and differential gene expression have been studied in cultured promastigotes at the stationary phase and in promastigotes isolated from the stomodeal valve of the sand fly P . perniciosus . About 20 ng RNA per sample could be isolated . Each sample contained L . infantum promastigotes from 20 sand flies . RNA was successfully amplified and processed for shotgun genome microarray hybridization analysis . Most differentially regulated genes are involved in regulation of gene expression , intracellular signaling , amino acid metabolism and biosynthesis of surface molecules . Interestingly , meta-analysis by hierarchical clustering supports that up-regulation of 22 . 4% of the differentially regulated genes is specifically enhanced by the microenvironment ( i . e . sand fly gut or culture ) . The correlation between cultured and naturally developed promastigotes is strong but not very high ( Pearson coefficient R2 = 0 . 727 ) . Therefore , the influence of promastigote culturing should be evaluated case-by-case in experimentation .
The genus Leishmania ( Kinetoplastida: Trypanosomatidae ) is responsible for leishmaniasis , a vector-borne parasitic disease with an estimated prevalence of 12 million people worldwide . Visceral leishmaniasis is the most severe form . It is fatal if left untreated . In fact , it causes about 60 , 000 deaths annually [1 , 2] . L . infantum is the etiological agent of zoonotic visceral leishmaniasis in the Mediterranean Basin , where co-infection with the HIV has been reported [3 , 4] . The main reservoirs of L . infantum are domestic dogs and wild canids . However , hares and rabbits have been involved as reservoirs in an outbreak in humans reported in the southwest of the Autonomous Community of Madrid [5 , 6 , 7] . The life cycle of the parasite is digenetic and involves an insect stage ( promastigote ) and a vertebrate stage ( amastigote ) . Promastigotes replicate and differentiate inside the gut of hematophagous female sand fly vectors ( Diptera: Psychodidae , Phlebotominae ) that innoculate metacyclic promastigotes into the mammalian host´s dermis when feeding . Promastigotes engulfed by phagocytes are able to develop into amastigotes , which multiply inside parasitophorous vacuoles . The proven vectors of L . infantum in Spain are Phlebotomus perniciosus and P . ariasi [8] . The former is the major vector in the central and the western Mediterranean Basin [9] . In the 1960s and 70s , axenic culture of promastigotes was developed in undefined media . To some extent , this procedure allows obtaining biomass and reproducing in vitro the conditions inside the gut of the sand fly [10 , 11 , 12 , 13] . In fact , promastigote cultures are grown at 26–27°C ( reviewed in [14 , 15] ) . The axenic culture model is more stable and reproducible in promastigotes than in amastigotes [16 , 17] . Conversely , obtaining promastigotes from the gut of the sand fly is technically challenging . Established laboratory colonies are required and the promastigote biomass isolated is usually insufficient for subsequent procedures . For these reasons , promastigotes are generally cultured . The original features , infectivity and virulence of the parasite become attenuated after numerous culture passages , which is often remedied by passages through laboratory animals ( reviewed in [14] ) . Axenic culture does not influence genome content analysis and other types of studies . However , others are affected , such as parasite-host cell interaction and immunological studies . RNA polymerase II transcription is performed without a canonical promoter for each gene in Leishmania spp . Protein coding genes are arranged in long polycistronic gene clusters ( PGCs ) [18 , 19 , 20] that are constitutively transcribed . The steady-state transcript levels are post-transcriptionally regulated [21] . Relatively low differential gene expression rates have been described in these organisms [22] . Differential gene expression profiling studies provided data about relative transcript abundance of hundreds of genes , as well as valuable information about the biology of Leishmania spp . ( e . g . , a succession of transient and permanent changes in gene expression during differentiation of promastigotes to amastigotes [23] and the relevance of temperature increase and acidification in this process [17] ) . The purpose of this study is comparing in vitro infectivity and differential transcript abundance of promastigotes isolated from two different environments: the anterior thoracic midgut of the experimentally infected P . perniciosus sand flies ( Pro-Pper ) and stationary phase of axenic culture ( Pro-Stat ) . Small amounts of RNA could be isolated and mRNA was amplified for subsequent microarray hybridization analysis . For obvious reasons , proteome analysis is not suitable so far [24] . The study contributes to explain the reliance of promastigote axenic cultures .
Blood samples from a New Zealand White rabbit were required for infecting P . perniciosus with L . infantum . Breeding , handling and sampling was performed following the EU ( 2010/63 ) and Spain ( RD1201/2005 ) regulations . The ISCIII Ethics Committee for Research in Animal Welfare approved the blood extraction protocol included in the CBA PA73-2011 license . The Leishmania infantum isolate MCAN/ES/98/10445 ( zymodeme MON-1 ) was axenically cultured . The inocula were aliquots of mid logarithmic phase promastigote cultures at the fourth passage after extraction from the gut of experimentally infected sand flies ( see below ) . These aliquots were cryopreserved at -196°C in heat inactivated fetal bovine serum ( HIFBS ) containing 10% DMSO . Three independent biological replicate cultures were performed at 27°C in complete medium , which contained RPMI 1640 supplemented with L-glutamine ( Cambrex , Karlskoga , Sweden ) , 10% HIFBS ( Cambrex ) and 100 μg/ml streptomycin– 100 IU/ml penicillin ( Cambrex ) . After 72 h , promastigote samples were harvested at 2000 x g for 10 min and processed daily ( see below ) . Only stationary phase promastigote samples obtained the day before the beginning of the death phase ( Pro-Stat ) were subsequently analyzed . In vitro infection of the human histiocytic leukemia U937 cell line [25] ( ATCC CRL1593 . 2 ) and high-throughput differential gene expression analysis were performed with Pro-Pper and Pro-Stat . The U937 cells were cultured at 37°C in complete medium in 175 cm2 flasks in an atmosphere of 5% CO2 for 72 h . Once the cells had been harvested at 250 x g for 10 min , differentiation was triggered with 20 ng/ml phorbol 12-myristate 13-acetate ( Sigma , Saint Louis , MO ) in complete medium for 72 h [26] . Then , the cells were mildly rinsed with RPMI 1640 supplemented with L-glutamine ( Cambrex ) and detached by vigorous shaking in the presence of 0 . 5 g/l trypsin , 0 . 2 g/l EDTA ( Cambrex ) . Trypsin inactivation was carried out by adding one volume of complete medium . Next , phagocytes were centrifuged and mixed with stationary phase promastigotes at a proportion of 1:20 . Infection was allowed at 37°C in complete medium in a water bath for 2 h . During incubation , the mixture was mildly shaken every 15 min . Then , the cells were centrifuged and incubated in complete medium at 37°C in an atmosphere of 5% CO2 for 72 h . The cultures were rinsed with complete medium after 2 and 16 h post-infection in order to eliminate remaining promastigotes . This procedure was used to obtain infected phagocytes for subsequent infection of sand flies ( see below ) . The in vitro infection assays were performed following a similar procedure but on 8-well cell chamber slides ( LabTek , New York , NY ) . In this case , Pro-Pper and Pro-Stat were added directly onto the stimulated U937 cells attached to the surface ( 1:5 ) . Therefore , detachment of phagocytes with trypsin was not required . Fixation and staining were performed 48 h post-infection . First , the wells were rinsed with fresh complete medium by thorough pipetting . Then , the cells were treated with hypotonic solution ( complete medium:water 9:11 ) for 5 min . Next , four washes with 150 μl ethanol-acetic acid 3:1 were carried out . The cells were fixed with the same solution for 10 min three times . The preparations were allowed to air dry and the wells were removed from the slide . Modified Giemsa staining was performed with Diff-Quick Stain Solution I and II ( Dade Behring , Marburg , Germany ) . The preparations were rinsed with distilled water , air dried and mounted with Entellan Neu ( Merck , Darmstadt , Germany ) . Finally , the number of amastigotes ( Amas ) per infected cell was estimated by counting 100 cells per biological replicate randomly . The experiment was performed in triplicate . The differences in infectivity between Pro-Pper and Pro-Stat were statistically assessed by the paired Student's t-test . An established colony of P . perniciosus was maintained at 27–28°C , 90–100% relative humidity , 17 h light– 7 h darkness photoperiod and 30% fructose solution in a climatic chamber . Two million U937 cells from an infected culture ( see above ) were resuspended in 2 ml defibrinated rabbit blood . The mixture was used to feed 150–200 female sand flies over a 3-day chicken skin membrane [27] . Then , the sand flies were maintained at 27–28°C , 90–100% relative humidity , 17 h light—7 h darkness photoperiod in a climatic chamber , obviously in the absence of the fructose solution . The course of infection was followed every day . For this purpose , a sample of sand flies was dissected using entomological needles and the guts were removed and studied at the light microscope ( 40X ) . Once promastigotes reached the stomodeal valve in the monitoring samples , twenty sand flies per experimental replicate sample were dissected in a PBS drop . Then , the entire guts were isolated and the anterior thoracic midgut was separated in a new PBS drop on the same slide . Only this portion was slightly pressed with a coverslip , and the drop containing promastigotes ( Pro-Pper ) in suspension was recovered with a Pasteur pipette . Hence , carryover of gut tissue was minimized as much as possible . Finally , the sample was directly added onto stimulated U937 cells attached to the well in the case of the infection experiment , or centrifuged at 4°C and washed in PBS for subsequent RNA isolation ( see below ) . After 72 h of development , the RNA samples were prepared daily until sand flies began to die as a consequence of infection . Therefore , Pro-Pper is defined as the promastigotes samples isolated from the sand fly anterior thoracic midgut behind the stomodeal valve one day before the beginning of the dead phase . This criterion was also applied for preparation of Pro-Stat , thus ensuring that both populations are comparable . Immediately after isolation of Pro-Pper and Pro-Stat , RNA was purified by extraction with 0 . 5 ml TRizol reagent ( Life Technologies , Carlsbad , CA ) following the manufacturer’s instructions . One μg/ml glycogen ( Life Technologies ) was added as carrier prior to RNA precipitation with isopropanol . At this point , RNA was stored at -80°C until use . Two mRNA amplification rounds were performed with MessageAmp II aRNA Amplification Kit ( Life Technologies ) as described [28] . RNA quality was assessed by conventional agarose gel electrophoresis and Experion RNA HighSens Analysis Kit ( Bio-Rad Laboratories , Hercules , CA ) according to the manufacturer's instructions . In the first case , the electrophoresis cell , tray and comb were rinsed with hydrogen peroxide and the aaRNA samples were run at 5 V/cm in a 1 . 5% agarose gel prepared with RNase-free water . The gel was pre-stained with GelRed Nucleic Acid Gel Stain ( Biotium , Hayward , CA ) diluted 1:10 , 000 . The first strand aminoallyl-cDNA was synthesized . First , 10 μg of aaRNA were mixed with 6 μg of random hexamer primers ( Life Technologies ) . The mixture was denatured at 70°C for 10 min and immediately cooled on ice . Thereafter , samples were incubated at 46°C for 3 h with 230 μM dTTP , 340 μM aminoallyl-dUTP , 570 μM ( each ) dATP , dCTP and dGTP , 10 μM DTT and 600 U SuperScript Reverse Transcriptase ( Life Technologies ) in a final reaction volume of 30 μl . Next , RNA was degraded at 70°C for 30 min in 100 mM NaOH/10 mM EDTA . The solution was neutralized with 3 μl of 3 M sodium acetate pH 5 . 2 . Then , cDNA was purified with QiaQuick PCR Purification Kit ( Qiagen , Hilden , Germany ) . The manufacturer’s instructions were followed except for using phosphate wash buffer ( 5 mM KPO4 , 80% ethanol , pH 8 . 0 ) and phosphate elution buffer ( 4 mM KPO4 ) instead of the respective commercial buffers . Then , the purified aminoallyl-cDNA samples were completely dried in a vacuum centrifuge and resuspended in 10 μl of water . Five μl of Cy3 or Cy5 ( respectively for Pro-Stat and Pro-Pper ) monofunctional dye ( GE Healthcare , Chalfont Saint Giles , UK ) dissolved in DMSO at 12 ng/μl were added . Coupling was allowed at room temperature in darkness for 1 h . Finally , the labeled cDNA samples were purified with QiaQuick PCR Purification Kit ( Qiagen ) according to the manufacturer's instructions . Whole genome shotgun DNA microarrays of L . infantum ( GEO Accession number GPL6781 ) were soaked with 0 . 1% N-lauroylsarcosine in 2XSSC; then in 2XSSC . After denaturation at 95°C for 3 min , they were fixed in chilled 100% ethanol and spin-dried in a slide mini centrifuge . A 60 μl drop containing 2XSSC , 0 . 3% N-lauroylsarcosine , 60 mM Tris-HCl pH8 . 0 , 83 ng/ml denatured herring sperm DNA and 1% BSA was deposited over a Hybri-Slip coverslip ( Sigma ) . The slide was attached on the coverslip . Blocking was allowed at 42°C in a hybridization chamber submerged in a water bath for 30 min . Thereafter , labeled cDNA samples were mixed in equimolar amounts of each dye ( 50 pmol ) and incubated at 40°C with blocked microarrays for 16 h ( same as blocking solution except for 0 . 1% BSA , 25 ng/ml poly ( T ) , 50% deionized formamide ) . Finally , the slides were soaked with 2XSSC , 0 . 2% SDS at 40°C and consecutively in 1XSSC and 0 . 2XSSC at room temperature . Genomic DNA was isolated from non-infected sand flies by phenolic extraction as described [29] and directly labeled with Cy5 using GenomiPhi DNA Amplification Kit ( GE Healthcare ) . For this purpose , 350 μM each dATP , dCTP , dGTP and [1/3 Cy5-dUTP , 2/3 dTTP] mix were used . Next , it was hybridized with the microarrays as a cross-hybridization control . Hybridization data were acquired with a GenePix 4100A scanner ( Axon , Foster City , CA ) . Raw data of local feature background medians were subtracted with GenePix Pro 7 . 0 software . The LOWESS per pin algorithm was applied for data normalization and differential expression was contrasted by the Student’s t-test . The AlmaZen software ( BioAlma , Tres Cantos , Spain ) was used for both purposes . The cutoff values for differential gene expression were the following: ( i ) fold change F ≥ 2 ( Cy5/Cy3 ratio if Cy5 > Cy3 ) or F ≤ -2 ( -Cy3/Cy5 ratio if Cy3 > Cy5 ) , ( ii ) total relative fluorescence intensity value > 5000 arbitrary fluorescence units and ( iii ) p < 0 . 05 . Three biological replicates were considered in the experiment . The clones that complied with the cutoff values mentioned above were sequenced with the M13-pUC18 primers and assembled as described [29] . Correctly assembled clones fulfilled the following: ( i ) e-value < 1e-10 for both ends , ( ii ) convergent orientation in the genome sequence and ( iii ) length ≤ 11 kbp , according to the features of the genome library used for microarray construction [29] . The analyzed clones were classified in three categories: a clones ( only a pair of alignments complies with all three conditions ) , b clones ( more than a pair does due to adjoining sequence repeats; only the best sequence identity is considered ) and c clones ( not complete fulfilment of the requirements mainly due to the presence of two or more inserts in the clone ) . Then , clones were associated to genes annotated in the genome by using a Perl script that excludes 5% of the ORF end sequence that overlaps with the boundaries of the clone . Clones that do not fulfill this criterion but align with less than 5% of the length of a given annotated ORF were identified using the genome browser [29] . Clones that do not map with any ORF were aligned with complete transcript sequences including UTRs that were obtained by RNAseq in L . major [30] . Genes were classified in functional categories according to the Gene Ontology database ( GO ) and associated to EC identifiers and KEGG pathways [31] with BLAST2GO software [32] . In addition , the GeneDB [33] and TriTrypDB [34] databases were useful to retrieve information about gene functions , as well as literature . CLUSTALW2 alignments allowed distinguishing between gene copies . Unlabeled single stranded cDNA was synthesized following the same procedure as for microarray hybridization but using a mixture stock of 10 mM each dNTP . Custom TaqMan MGB Assay-by-Design ( specifically primers and FAM-NFQ MGB probes , Life Technologies ) were mixed with 1:5 serial dilutions of cDNA samples ( 10 , 2 and 0 . 4 ng cDNA per reaction ) and with TaqMan Universal Master Mix 2X ( Life Technologies ) in a final reaction volume of 10 μl . Primer and probe sequences are listed in S1 Table . The qRT-PCR reactions were run in a 7900HT Fast Real Time PCR system using the SDS 4 . 1 . software ( Life Technologies ) following the procedure specified by the manufacturer . The thermal cycling conditions were: 95°C for 5 min; 40 x [95°C for 30”; 60°C for 1 min , data acquisition] . After checking coefficients of variation , PCR efficiencies were calculated by the standard curve best fit method using the data obtained in the triplicate dilution series experiment for each gene and cDNA sample ( Pro-Pper/Pro-Stat ) . The normalized quantities were calculated by dividing the efficiency-corrected raw quantities ( efficiency to the power of–Ct ) for the gene of interest by those for the reference gene ( L . infantum gGAPDH ) . Fold changes were obtained by dividing the normalized quantities ( Pro-Pper/Pro-Stat ) . This procedure is based on specifications provided by Bookout et al . [35] . The relative expression profiles of Pro-Pper/Pro-Stat were compared with the Pro-Pper/Amas [24] , Pro-Stat/Pro-Log and Pro-Stat/Amas [28] ones . Pro-Log are defined as cultured promastigotes at early logarithmic phase ( about 48 h in culture , hence undifferentiated ) . For this purpose , the TIGR's Experiment Viewer 4 . 9 ( MEV ) software was used . The normalized fold-change values of each microarray hybridization experiment were loaded . The hierarchical clustering-support tree ( HCL-ST ) algorithm was run using Euclidean metrics and setting jackknifing resampling , 100 iterations , the complete linkage method and the Pearson correlation coefficient as the distance metrics for the ST .
Infected sand flies were dissected to extract their digestive tracts . Next , the anterior part of the thoracic midgut , containing the stomodeal valve and Pro-Pper promastigotes , was isolated ( Fig 1A ) . After slightly pressing with a coverslip , the PBS drop containing Pro-Pper promastigotes in suspension was recovered with a Pasteur pipette . Following this procedure , carryover of gut tissue was minimized as much as possible . Each Pro-Pper sample contained material from 20 sand flies . Three samples per condition were prepared for the in vitro infection and gene expression profiling experiments . Pro-Pper has been defined above as the promastigote samples obtained from the sand fly anterior thoracic midgut behind the stomodeal valve the day before the dead phase began ( day 6 in this case ) , whereas Pro-Stat are equivalent populations in stationary phase of axenic culture ( day 7 in this case ) . Despite both are intrinsically heterogeneous , they are enriched in metacyclic promastigotes , especially Pro-Pper . The in vitro infection experiment of U937 cells has confirmed that Pro-Pper are significantly more infective than Pro-Stat ( Student's t-test , p < 0 . 0001 ) , as the mean ± SD of the number of amastigotes per infected cell at 48 h post-infection is 4 . 8 ± 0 . 9 and 2 . 7 ± 0 . 4 , respectively ( Fig 1B ) . All samples were immediately washed with PBS once and lysed with TRIzol reagent . The yield of total RNA isolation was about 20 ng per biological replicate sample in the case of Pro-Pper . For this reason , two rounds of mRNA amplification were performed to obtain enough starting material for microarray analysis . The Pro-Stat samples were prepared following the same procedure . The results of mRNA double amplification are shown in Fig 1C . Microarray hybridization data of control spots are included in S2 Table . As expected , the amastigote-specific A2 gene , spotted in the microarrays as a control gene [29] , is not differentially expressed between Pro-Pper and Pro-Stat . The number of differentially regulated genes between Pro-Pper and Pro-Stat is 286: 148 are up-regulated in Pro-Pper and 138 in Pro-Stat ( Fig 2 , Table 1 and S3 Table ) . All clones that represent differentially regulated genes of known function are provided in S4 and S5 Tables , whereas Tables 2 and 3 contain only those discussed below . A complete explanation of Tables S4 and S5 is provided in S1 Text . Certain clones selected in the microarray hybridization analysis are undetermined because they overlap with more than one gene annotation . Most were solved by TaqMan Probe-based qRT-PCR analyses , thus determining the actual differentially regulated gene . This approach also validated 13 . 8% of the microarray results ( Tables 2 and 3 , S4–S7 Tables ) , together with the internal controls mentioned ( S2 Table ) . The microarray and qRT-PCR results are quite consistent qualitatively . In some cases , remarkable quantitative differences are observed , which is expected due to the inherent wide dynamic range and sensitivity of qRT-PCR compared to the high-throughput microarray hybridization analysis . Constant expression values have been obtained by qRT-PCR in certain clones that overlap with more than one gene . In these cases , at least one of the remaining is presumably differentially regulated . Consequently , inconsistencies between both approaches have not been detected so far . The GO terms were assigned with BLAST2GO . This software is based on the NCBI database , where the second version of the L . infantum genome was deposited . In order to ensure that the changes introduced in the last version of the genome sequence released ( TriTrypDB ) did not affect the GO analysis , we aligned all genes contained in Tables 2 and 3 , S4–S7 Tables against the NCBI database . As a result , at least 98% identity was found in all cases . According to the GO analysis , gene expression regulation is affected between Pro-Pper and Pro-Stat . In fact , the GO terms ncRNA metabolic process , RNA processing , translation , post-translational modification and proteolysis are more represented in genes up-regulated in Pro-Pper . Conversely , a considerable number of ribosomal proteins are over-expressed in Pro-Stat ( Fig 3 ) . Changes related with signal transduction are also expected because certain differentially expressed protein kinase ( PK ) and phosphatase ( PP ) genes are associated to the GO term protein phosphorylation . Some are up-regulated in Pro-Pper and some others in Pro-Stat . All changes in the transcript levels of genes related with gene expression regulation and intracellular signaling might be associated to the GO term response to stimulus ( associated to Pro-Pper ) and response to chemical stimulus ( associated to Pro-Stat ) ( Fig 3 ) . Certain genes related with vesicle-mediated transport and metabolic processes of carbohydrates , lipids , amino acids and nucleotides are also differentially regulated between Pro-Pper and Pro-Stat . The most relevant differentially regulated genes of known function ( Tables 2 and 3 ) are discussed in detail in this section . A more detailed explanation can be found in S1 Text ( S4 and S5 Tables ) . Important differences in abundance of transcripts related with most major cellular processes have been found between promastigotes developed in both microenvironments studied ( Fig 4 ) . Unless indicated , the expression “up-regulation in Pro-Pper” is relative to Pro-Stat and vice versa . The findings described have raised interesting hypotheses that may be tested in the future . The differential expression profile of Pro-Pper/Pro-Stat described in this study has been compared to the Pro-Pper/Amas [24] , the Pro-Stat/Pro-Log and the Pro-Stat/Amas [28] expression profiles by HCL-ST ( Fig 5 ) . All microarray hybridization experiments were performed by the same procedure and clone nomenclature is equivalent . Fig 5A consists of clusters obtained by HCL-ST of two sets of differentially regulated genes: Pro-Pper/Pro-Stat ( 286 genes; Tables 2 and 3 , S4–S8 Tables ) and Pro-Pper/Amas ( 213 genes [24] ) . The intersection of both sets is 64 genes ( 30 . 0% and 22 . 4% , respectively ) . Two thirds of these 64 common changes ( 20 . 2% and 15 . 0% , respectively ) correspond to up-regulated genes in Pro-Pper and the remaining 21 ( 9 . 9% and 7 . 3% , respectively ) to down-regulated . In principle , this suggests that common over-expression of the 43 genes in Pro-Pper vs . Pro-Stat and vs . Amas is important for life cycle progression , i . e . for development of promastigotes from the sand fly anterior thoracic midgut into amastigotes . In other words , up-regulation of these genes is required at the beginning of differentiation into amastigotes only whether promastigotes have been developed within the sand fly gut but not in axenic culture . In fact , none of these genes is up-regulated in Pro-Stat vs . Amas ( Fig 5B ) [28] . Conversely , the 21 commonly down-regulated genes in Pro-Pper vs . Pro-Stat and vs . Amas indicate that Pro-Stat require higher steady-state levels of the corresponding transcripts upon differentiation to amastigotes . Differentiation of promastigotes to an infective stage is successful in culture , as well as within the sand fly gut . However , the efficiency of infection is higher in the natural life cycle than in experimental infections of cell lines and animals . For example , a single P . dubosqi sand fly is able to transmit as low as 600–1 , 000 L . major promastigotes capable of establishing infection in mice successfully [52] , whereas experimental infection with L . major stationary phase promastigotes is usually performed with 106 promastigotes ( e . g . [53] ) . This may be explained by the in vitro infection procedure , the reduction of infectivity of cultured parasites [14] and the absence of molecules from the saliva of the sand fly . In fact , L . infantum Pro-Pper promastigotes are significantly and considerably more infective than Pro-Stat ( Fig 1B ) . Therefore , future studies on the 43 commonly up-regulated genes in Pro-Pper vs . Pro-Stat and vs . Amas may contribute to explain the increased infection rate during progression of the natural life cycle . Lahav et al . [54] described that the correlation between the transcript and protein levels is about 25% quantitatively in L . donovani . According to clustering analysis , qualitative coincidence ( up-regulation , down-regulation , constitutive expression ) is about 65% [54] . Proteome analysis is not possible in Pro-Pper due to sample amount limitations . Therefore , amplification of the transcriptome is the only alternative so far . For this reason , our estimate of genes that are actually related with progression of the life cycle is at least 29 in sand fly foregut promastigotes and 13 in stationary phase promastigotes . Future studies on these genes may reveal new vaccine candidates and drug targets . For example , we described that the L . infantum tyrosine aminotransferase is up-regulated in metacyclic promastigotes in culture at the transcript level [29] and it has been confirmed at the protein level by Western blot . This protein is a drug target candidate and the pharmacophore has been predicted [55] after structural study [56] . GO terms were associated to differentially regulated genes in Pro-Pper/Pro-Stat ( Fig 3 ) and Pro-Pper/Amas [24] . According to this analysis , genes involved in carbohydrate and amino acid metabolic processes are up-regulated in Pro-Stat and Amas with respect to Pro-Pper , whereas cellular macromolecule catabolic processes , fatty acid metabolic processes and vesicle mediated transport are linked to genes up-regulated in Pro-Pper . In both comparisons , changes in steady-state transcript levels involved in response to internal and external stimuli were found to be higher in Pro-Pper , except for response to chemical stimuli . In the case of the Pro-Pper/Amas comparison , the systematic analysis of GO terms provided insight into the nature of the stimuli ( abiotic stimuli , DNA damage and inorganic substances ) [24] , unlike in Pro-Pper/Pro-Stat . In both cases , the considerable number of differences in genes related to intracellular signaling and regulation of gene expression is remarkable , including the post-transcriptional , translational and post-translational levels and protein degradation via the ubiquitin-proteasome system . In fact , the 43 commonly up-regulated genes in Pro-Pper vs . Pro-Stat and vs . Amas are involved in most major cellular processes like DNA repair ( MSH6 ) , gene expression regulation ( H4 , HDAC , U2-snRNP , Lsm5p , eIF4E , 60S acidic ribosomal protein ) , proteolysis ( UBC , ubq , A22B ) , cytoskeleton remodeling ( PFN ) , intracellular signaling ( PP2B-A2 , PK , β-prop , PI4K ) , metabolism ( GNAT , DHAK , coxV , oxidoreductase ) , transport ( PT , ABCE1 , AP3δ1 ) and biosynthesis of surface molecules ( PI4K , PIGA ) . Nineteen hypothetical proteins are also included . Among these genes , H4 , GNAT , AP3δ1 and four hypothetical protein genes are also up-regulated in Pro-Log and the expression levels of GNAT and PP2B-A2 are higher in Pro-Stat than in Amas ( Fig 5B ) [28] . The 21 up-regulated genes in Pro-Stat and Amas with respect to Pro-Pper are ESAG5 , L21 , gPEPCK , NSDHL , SbGRP , AIP , three amastins and 12 hypothetical protein genes . Three of them ( gPEPCK and amastin-like proteins LinJ . 08 . 0680/90 ) are also over-expressed in Pro-Stat with respect to Pro-Log ( Fig 5B ) [28] . Future studies on these genes that are common to both datasets are of great interest . The first set may contribute to explain the relationship between signal transduction and effector mechanisms of gene expression regulation in the differentiation process of Pro-Pper to Amas , whereas the second set would be applicable for the equivalent process from Pro-Stat to Amas . In summary , most differential gene expression profiles are distinct between Pro-Pper/Pro-Stat and Pro-Pper/Amas as expected , except for 13 . 2% genes specifically up-regulated by the effect of the microenvironment ( i . e . sand fly gut or culture ) at the beginning of differentiation of promastigotes to amastigotes . Therefore , these genes are essential for life cycle progression in the respective microenvironments . In fact , they participate in processes affecting key regulatory biological processes . The data presented above contribute to answer this question . Axenic cultures of promastigotes mimic the conditions inside the gut of the sand fly to some extent [10 , 11 , 12 , 13] and they are relatively stable and reproducible when compared with amastigote cultures [16 , 17] . However , this study confirms that L . infantum promastigotes obtained from the anterior thoracic midgut of P . perniciosus are considerably more infective than promastigotes in stationary phase of axenic culture ( Fig 1B ) . Evidence about important differences in their transcriptome is also provided , i . e . 286 differentially regulated genes . As described above , the correlation coefficient between both expression datasets is R2 = 0 . 727 ( Fig 2 ) . The meaning of this finding is that the Pro-Pper and Pro-Stat populations are strongly correlated , although important differences are still observed . Indeed , the shape of the M/A scatter plot ( Fig 2 ) is a non-dispersed ( rank -4 < M < 4 ) dot-cloud symmetric about the M = 0 line ( i . e . , lack of differential expression ) . In light of these findings , we contend that the axenic culture model of promastigotes is generally valid , but it should be cautiously questioned case by case for every particular experimental design . After all , Pro-Pper promastigotes are the result of development in their natural microenvironment and consequently , their infectivity is higher . Pro-Pper populations are more infective than Pro-Stat ones . Their transcriptome profiles are substantially different . In fact , certain genes involved in DNA repair , gene expression regulation , metabolism , transport including vesicle trafficking , intracellular signaling , cytoskeleton remodeling and biosynthesis of surface molecules are differentially regulated . However , the Pearson correlation coefficient between the normalized fluorescence intensity values of both populations is R2 = 0 . 727 . This indicates strong correlation but also remarkable differences . Consequently , the adequacy of the axenic culture model should be studied a priori in each particular experimental design . The HCL-ST analysis has revealed that the degree of similarity in differentially expressed genes between Pro-Pper/Pro-Stat and Pro-Pper/Amas is 13 . 2% ( 64 genes ) . Life cycle progression ( differentiation to amastigote ) in the natural microenvironment of promastigotes would require up-regulation of 43 out of the 64 common genes . The information obtained in this high-throughput study is useful for understanding better the differences between promastigotes from culture and the sand fly . Specific information about relative expression is also a criterion for selecting possible vaccine candidates and/or drug targets . | The protozoan parasite Leishmania infantum causes visceral leishmaniasis in humans and is responsible for a recent outbreak reported in central Spain . Domestic dogs and wild canids are the main reservoirs . The life cycle of the parasite involves two stages and two hosts . The motile promastigote stage differentiates within the gut of the sand fly vector host and develops into non-motile amastigotes within phagocytes of the mammalian host . Promastigotes are routinely cultured in liquid media because it is assumed that they mimic the conditions within the gut of the insect . Therefore , the culture model is used in most studies about the biology of the parasite , pathogenesis and development of vaccines and new compounds for treatment . Isolating promastigotes from the natural microenvironment ( i . e . the vector host ) is desirable but technically challenging . We were able to perform a high-throughput analysis of gene expression thanks to mRNA amplification . The over-expressed genes detected may influence life cycle progression depending on the promastigote microenvironment ( i . e . culture or vector host ) . Upcoming studies based on these results may reveal new therapeutic targets or vaccine candidates . Our results suggest that evaluating the influence of cultures in experimentation is convenient . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"gene",
"regulation",
"microbiology",
"sand",
"flies",
"cloning",
"protozoan",
"life",
"cycles",
"parasitic",
"protozoans",
"developmental",
"biology",
"protozoans",
"leishmania",
"bioassays",
"and",
"physiological",
"analysis",
... | 2016 | Influence of the Microenvironment in the Transcriptome of Leishmania infantum Promastigotes: Sand Fly versus Culture |
Wide and improper application of pyrethroid insecticides for mosquito control has resulted in widespread resistance in Aedes albopictus mosquitoes , an important dengue vector . Therefore , understanding the molecular regulation of insecticide resistance is urgently needed to provide a basis for developing novel resistance diagnostic methods and vector control approaches . We investigated the transcriptional profiles of deltamethrin-resistant and -susceptible Ae . albopictus by performing paired-end sequencing for RNA expression analysis . The analysis used 24 independent libraries constructed from 12 wild-caught resistant and 12 susceptible Ae . albopictus female adults . A total of 674 , 503 , 592 and 612 , 512 , 034 reads were obtained , mapped to the Ae . albopictus genome and assembled into 20 , 091 Ae . albopictus transcripts . A total of 1 , 130 significantly differentially expressed genes included 874 up-regulated genes and 256 down-regulated genes in the deltamethrin-resistant individuals . These differentially expressed genes code for cytochrome P450s , cuticle proteins , glutathione S-transferase , serine proteases , heat shock proteins , esterase , and others . We selected three highly differentially expressed candidate genes , CYP6A8 and two genes of unknown function ( CCG013931 and CCG000656 ) , to test the association between these 3 genes and deltamethrin resistance using RNAi through microinjection in adult mosquitoes and oral feeding in larval mosquitoes . We found that expression knockdown of these three genes caused significant changes in resistance . Further , we detected 1 , 162 single nucleotide polymorphisms ( SNPs ) with a frequency difference of more than 50% . Among them , 5 SNPs in 4 cytochrome P450 gene families were found to be significantly associated with resistance in a genotype-phenotype association study using independent field-collected mosquitoes of known resistance phenotypes . Altogether , a combination of novel individually based transcriptome profiling , RNAi , and genetic association study identified both differentially expressed genes and SNPs associated with pyrethroid resistance in Ae . albopictus mosquitoes , and laid a useful foundation for further studies on insecticide resistance mechanisms .
Arboviruses such as Zika , dengue , and Chikungunya by Aedes mosquitoes represent an expanding threat to global health [1–3] . Chemical control through the use of insecticides is one of the essential tools in the global strategy for mosquito-borne infectious disease control [4] . Pyrethroid is often the insecticide of choice due to its low mammalian toxicity and high insecticidal activity [4 , 5] . Permethrin , deltamethrin , cypermethrin , and cyfluthrin are currently the most commonly used pyrethroids for adult mosquito control in indoor residual spraying ( IRS ) , insecticide-treated nets ( ITN ) , and space spray treatment [6 , 7] . For instance , ultra-low-volume ( ULV ) spray of pyrethroids has been a major measure for controlling Aedes adults in dengue-endemic areas in China [8] . However , excessive and improper application of pyrethroids for mosquito control , together with widespread agricultural usage , has resulted in a high selection pressure for pyrethroid resistance [9 , 10] . Insecticide resistance has become a major impediment to controlling mosquito-borne diseases worldwide . As a response to this dilemma , the World Health Organization ( WHO ) launched the Global Plan for Insecticide Resistance Management ( GPIRM ) to guide rational insecticide usage [11] . Insecticide resistance in the Asian tiger mosquito , Aedes albopictus , the major dengue vector , came to public attention in China with the dengue outbreak in 2014 , when over 45 , 000 dengue cases were reported in Guangdong province [12] . Insecticide was applied extensively under the government-directed mosquito control program , and resistance has been reported in multiple Ae . albopictus populations in Guangdong [13] . Current resistance monitoring relies on bioassays with live mosquitoes . Understanding resistance mechanisms is essential to developing sensitive molecular monitoring tools of resistance and resistance management . Resistance to DDT and pyrethroids has been documented in different mosquito species [10 , 14 , 15] . Two major resistance mechanisms have been previously identified , including target site insensitivity resulting from mutations in the insecticides’ target protein genes and increased metabolic detoxifications for insecticides [4] . Non-synonymous mutations in the voltage-gated sodium channel ( VGSC ) gene causing target-site insensitivity have been found in multiple mosquito species of public health importance , including Anopheles gambiae [16 , 17] , Culex quinquefasciatus [15 , 18] , Ae . aegypti [10 , 14] , and Ae . albopictus [13] . In Ae . aegypti , V1016G , V1016I , and F1534C mutations were found to be correlated with pyrethroid resistance [19 , 20] , and the F1534S mutation was correlated with deltamethrin resistance in Ae albopictus [13] . Metabolic resistance involves the insecticide molecule’s bio-transformation via higher expression or presence of more efficient detoxification enzymes . Detoxification enzymes belong to four gene families: cytochrome P450 monooxygenases ( P450s ) , carboxylesterases ( COEs ) , glutathione S-transferases ( GSTs ) , and UDP-glucuronosyltransferases , but a number of species-specific genes within each gene family have been found to be involved in metabolic resistance in various mosquito species , such as Anopheles sinensis , Cx . quinquefasciatus , and Ae . aegypti [10 , 14 , 15 , 17 , 21–24] . Other resistance mechanisms include mosquito physiological changes , such as cuticle thickening and digestive tract modification , which may lead to changes in insecticide absorption and penetration , and in behavioral avoidance [25 , 26] . We have previously identified one mutation ( F1534S ) in the VGSC gene in Ae . albopictus that is significantly associated with pyrethroid resistance [13] . However , little is known about metabolic resistance in Ae . albopictus , which may result from both altered expression and mutations in the detoxification genes . For example , SNPs in GSTe2 were found to confer resistance to DDT in the malaria mosquito Anopheles funestus [27] . With this in mind , this study aims to characterize novel resistance mechanisms in Ae . albopictus . mRNA sequencing ( RNA-seq ) is a powerful technique for comprehensive profiling of transcripts . To date , several studies have used RNA-seq to investigate resistance mechanisms in mosquito species , including An . gambiae [28] , An . sinensis [29] , Ae . aegypti [30] , and Cx . quinquefasciatus [31] and Cx . pipiens pallens [32] , and these studies have demonstrated that the RNA-seq technique can successfully yield useful resistance targets . In southern China , Ae . albopictus is the most important dengue vector with strong pyrethroid resistance [13] . In this context , the present study used RNA-seq on single mosquitoes and identified genes differentially expressed between resistant and susceptible Ae . albopictus mosquitoes and SNPs associated with resistance . In addition , the function of 3 candidate genes was further functionally verified by RNA interference ( RNAi ) . Using a genotype-phenotype association study , we further identified 5 new SNPs significantly associated with resistance in natural populations .
This study did not involve endangered or protected species , and was not conducted on protected lands . All collections were performed on public land , and no specific permits were required for mosquito field specimen collection . RNAi analysis used a laboratory-selected deltamethrin-resistant population ( Lab-DR strain ) . Briefly , larvae from a laboratory Foshan strain of Ae . albopictus were bioassayed to determine the 50% lethal concentration ( LC50 ) , following the WHO standard method [34] . This bioassay measured larval mortality after 24 hr of exposure to 5 concentrations of deltamethrin , each concentration in three replicates . The Lab-DR strain was selected from the larvae of the Foshan strain , using 0 . 001 mg/L of deltamethrin ) as the starting concentration , and gradual increments of deltamethrin concentrations for 10 generations . The 10th generation selection used a 0 . 015 mg/L concentration of deltamethrin . The WHO insecticide susceptibility tube test on Lab-DR adults found an 89% mortality rate , indicating modest resistance . RNAi analysis was conducted in both adult mosquitoes and larval mosquitoes with 3 highly differentially expressed genes . For adult mosquito RNAi , 3-day post-emergence Lab-DR mosquitoes was randomly selected and divided into control group or RNAi groups . For each of the three RNAi groups , 500ng siRNA duplex ( siRNACYP6A8 , siRNA01931 , or siRNA00656 ) was microinjected into a female mosquito thorax ( S2 Table ) . For the control group , a siRNA duplex lacking significant sequence homology to any genes in the Ae . aegypti genome , siRNA_C , was injected ( S2 Table ) . All siRNAs were synthesized by IDT Technologies ( Coralville , IA , USA ) . After the injection , mosquitoes were allowed to recover in a cup and maintained with 8% sucrose . An overall 16% mortality was observed 24 hr after the injection . Forty-eight hours after injection , gene knockdown efficiency was examined in 12 randomly selected individuals in each injection group using qRT-PCR , and its impact on mosquito resistance was determined using a standard WHO insecticide susceptibility tube test with 0 . 05% deltamethrin . For each gene , RNAi-treated female mosquitoes 48 hr after injection underwent the WHO insecticide susceptibility bioassay , with 20–25 mosquitoes per replicate for a total of 5 replicates per group [36] . Control group and RNAi groups were tested simultaneously . For larval mosquito RNAi , we conducted oral RNAi following the method of Zhang et al . [37] . Briefly , each target gene siRNA and the siRNA_C described above ( S2 Table ) was mixed with Chitosan to form a Chitosan/siRNA nanoparticle . The nanoparticle was then mixed with cat food and agarose with green food dye , before feeding to 20 second-instar larvae in a petri dish with 100 ml water on a daily basis until the adult mosquitoes emerged . A total of 20 replicates with a total of 400 mosquito larvae were used for each target gene group and the control group . All larvae were successfully fed with the nanoparticles as evidenced by the green food dye in their guts . Three days after emergence , for each group 12 female adult mosquitoes were randomly selected to detect gene expression by qRT-PCR . Deltamethrin resistance bioassay was conducted using the standard WHO insecticide susceptibility tube test , with 20–25 mosquitoes per replicate for a total of 5 replicates per group . All RNA-seq data generated in this study are available under NCBI BioProject number PRJNA475859 .
Twenty-four cDNA libraries were constructed individually for deep sequencing on an Illumina-Solexa HiSeq 2500 platform . Raw read quality control was carried out via base composition analysis and base sequence quality analysis . A total of 674 , 503 , 592 and 612 , 512 , 034 paired-ends reads from resistant and susceptible conditions , respectively , were obtained . A summary of filtered paired-end reads is shown in Table 1 . An average of 56 , 208 , 633 reads in resistant individuals and 51 , 042 , 670 in susceptible individuals were mapped to the Ae . albopictus reference genome . Gene coverage statistics for resistant and susceptible individuals showed that most of the genes ( 84% and 77% , respectively ) reached coverage of 90–100% . All unique matched reads from resistant and susceptible individuals were assembled into 20 , 091 transcripts in a total of 11 , 090 genes . From these genes , 1 , 130 genes were significantly differentially expressed genes ( DEGs ) when using the criteria of FPKM ratio of each phenotype >2 ( in either direction ) , P < 0 . 0001 , and FDR < 0 . 05 . Among them , 874 genes ( 77 . 3% ) were up-regulated and 256 genes ( 22 . 7% ) were down-regulated in the resistant mosquitoes . These DEGs included P450s , cuticle proteins , UDP-glucuronosyltransferases , lipases , serine proteases , heat shock proteins , esterase , peptidases , ATP-binding cassette transporters , and others ( Fig 1A ) . All DEGs were classified into three main GO categories: biological processes , cellular components , and molecular functions ( Fig 1 ) . For cellular components , genes involved in cells ( GO: 0005623 , 249 out of 705 DEGs , or 35 . 3% ) and cell parts ( GO: 0044464 , 249 out of 705 DEGs , 35 . 3% ) were the most abundant ( Fig 1B ) . As for molecular functions , binding ( GO: 0005488 , 435 out of 838 DEGs , 51 . 9% ) and catalytic activity ( GO: 0003824 , 238 out of 838 DEGs , 28 . 4% ) were the most highly represented categories ( Fig 1C ) . For biological processes , the categories most represented were cellular processes ( GO: 0009987 , 294 out of 1 , 013 DEGs , or 29 . 0% ) and metabolic processes ( GO: 0008152 , 256 out of 1 , 013 DEGs , 25 . 3% ) ( Fig 1D ) . A total of 694 DEGs were mapped into 309 KEGG pathways . The largest category was metabolic pathways , containing 92 annotated DEGs ( 20 . 6% ) , followed by biosynthesis of secondary metabolites , with 36 DEGs ( 8 . 1% ) . In general , the most represented GO and KEGG categories were “metabolic process” and “metabolic pathway , ” which were widely acknowledged to be involved in insecticide metabolism [40 , 41] . A total of 11 DEGs belonging to the P450 family or GST which were known to be related to insecticide detoxification are shown in Table 2 . We did not find any significant DE genes related to metabolic inhibitors or cuticle protein genes . We used qRT-PCR to validate the expression level obtained from RNA-seq . As shown in Fig 2 , a highly significant correlation was found between the two methods ( R2 = 0 . 67 ) , demonstrating the validity of the RNA-seq method for expression quantitation . CYP6A8 ( CCG018948 ) , CCG013931 . 2 , and CCG000656 . 1 were selected for RNAi analysis based on expression differences between resistant and susceptible mosquitoes . The RNA-seq analysis indicated that the expression of CYP6A8 was 3 . 1-fold higher in the resistant samples than in the susceptible ones , and CCG013931 . 2 and CCG000656 . 1 were 13 . 7 and 5 . 5-fold higher as evidenced from the ( Fig 2 ) . For microinjection RNAi in adult mosquitoes , the CYP6A8 expression level was reduced by 52 . 8% by siRNACYP6A8 , compared with the control ( P < 0 . 01 ) ( Table 4 ) . Similarly , CCG013931 . 2 expression level was reduced by 66 . 3% ( P < 0 . 01 ) , and by 71 . 4% for CCG000656 . 1 ( P < 0 . 01 ) . The resistance bioassay found 100% mortality in RNAi treated groups for the three genes whereas a 90 . 1% mortality rate was found in the control group ( P < 0 . 01 ) ( S3 Table ) . The effect of adult mosquito RNAi using the microinjection method on mosquito knockdown time is shown in S2 Fig . Mosquitoes injected with siRNA01931 exhibited significantly shorter KDT50 time ( S3 Table ) . Oral feeding of siRNA nanoparticles in larvae indicated a similar gene expression knockdown effect compared to the microinjection group . The CYP6A8 expression level was reduced by 51 . 9% compared with the control ( P < 0 . 01 ) , by 53 . 5% for CCG013931 . 2 ( P < 0 . 01 ) , and by 48 . 0% for CCG000656 . 1 ( P < 0 . 01 ) ( Table 4 ) . Gene expression knockdown corresponded to increased mortality rate in the RNAi-treated groups . Mortality rate was 90 . 5% in the control group whereas the RNAi-treated groups showed 100% mortality rate ( P < 0 . 01 ) ( S4 Table ) . The mosquito knockdown time in insecticide resistance bioassay found a significantly shorter average knockdown time in RNAi treated groups than the control group ( S2 Fig ) . Non-synonymous mutation was detected at codon 1534 in the selected Lab-DR resistant strain adults but not in the susceptible Foshan strain . At codon 1534 , a change from wildtype codon TTC ( Phe ) to TCC ( Ser ) was detected in 14 out of the 30 samples , of which all were heterozygotes . The L1534S mutation frequency was 23 . 3% in the selected Lab-DR resistance strain . We examined 9 SNPs within the 7 P450 genes identified from the RNA-seq analysis to determine SNP genotypes’ association with deltamethrin resistance using 70 randomly selected resistant mosquitoes and 70 susceptible mosquitoes ( Table 5 ) . Among the 9 SNPs examined , 5 ( Arg226Ser in CYP6A8 , Pro175Gln and His877Tyr in CYP9B2 , Met120Ile in CYP6A8 , and Cys212Ser in CYP1A1 ) showed significant associations with resistance .
In the present study , we used the RNA-seq technique to study the transcriptional profile of deltamethrin-resistant and -susceptible Ae . albopictus mosquitoes . The individual mosquito-based sampling strategy used in this study is distinct from the commonly used pooled sampling strategy , and this new strategy enables us to simultaneously identify differentially expressed genes and differential SNPs that may be associated with pyrethroid resistance . We identified a total of 1 , 130 differentially expressed genes , including 874 up-regulated and 256 down-regulated genes . A total of 1 , 162 SNPs with large frequency differences between resistant and susceptible mosquitoes was identified . Using genotype-phenotype association analysis of natural mosquito populations , we identified 5 SNPs in the P450 gene family that are significantly associated with resistance . RNAi in adult mosquitoes through microinjection and larval mosquitoes through oral feeding confirmed three highly differentially expressed genes’ roles in resistance . In our data analysis , most of the DEGs are related to metabolic pathways involving insecticide absorption , especially genes coding for cytochrome P450 monooxygenase ( P450 ) and cytochrome C oxidase . The highly expressed DEGs in resistant individuals included specific P450 genes , such as CYP6A8 , CYP9Z4 , and CYP6D4 , that were identified in other insects , such as Musca domestica , An . gambiae , Cx . quinquefasciatus , and An . sinensis [42–45] . We also detected one glutathione S-transferase ( GST ) gene ( GSTE1 , AGAP009195-PA ) that showed increased expression in resistant mosquitoes , which was consistent with previous studies [15 , 34 , 46] . In addition to the up-regulated CYP genes , it was interesting to notice that there were several down-regulated CYP genes , including CYP301A1 , CYP4AC2 , and CYP4V2 . Several GST genes were found to be down-regulated ( GSTE1 and GSTE5 ) , which apparently contradicts the current knowledge on GST . GST enzymes have been shown to metabolize insecticides by facilitating reductive dehydrochlorination or by conjugation reactions with reduced glutathione , to produce water-soluble metabolites that are more readily excreted . The increased GST expression in resistant mosquitoes contributes to the removal of toxic oxygen-free radical species produced through the action of pesticides [47] . Thus , the current study’s reverse finding indicates that the GST enzyme may confer insecticide resistance in a more complex way . One explanation for decreased expression in the resistant individuals is that the decrease in gene expression may be due to reactions to various endogenous and exogenous compounds , or alternatively may represent a pathophysiological signal [48 , 49] . Such a GST gene down-regulation observed in resistant Ae . albopictus individuals was also found in other mosquito species: Cx . quinquefasciatus and An . sinensis [50 , 51] . These findings suggest complex associations between metabolic genes’ expression changes and insecticide resistance . An RNA interference assay was conducted to verify the function of candidate genes identified from RNA-seq analysis . We used two RNAi methods: the classic microinjection-based and oral feeding RNAi . We demonstrated that both methods led to >50% reduction in the expression of 3 candidate genes , and that RNAi treatment altered the resistance phenotype . Our results indicated that these 3 genes each may played a role in deltamethrin resistance regulation , however the impact of these genes on insecticide resistance at the organismal level needs further determination . One major disadvantage of the microinjection-based RNAi method is that siRNA injection led to substantial mortality in adult mosquitoes . In the control , we observed 16% mosquito mortality , likely due to mechanical injury associated with microinjection . The oral RNAi showed less than 1% mortality from the oral delivery process , consistent with oral delivery RNAi methods reported by other studies [52] . Another advantage of oral delivery RNAi is the relative ease of laboratory operation , without the need for microinjection , and subsequently a large number of RNAi-treated mosquitoes can be used for the resistance bioassay . Therefore , oral RNAi may be considered in future studies . Using a genotype-phenotype association study with field mosquito samples from southern China , we identified 5 SNPs within the CYP gene family that were significantly associated with pyrethroid resistance . In the SNP analysis , we also found higher F1534S mutations in the kdr gene among the 12 resistant individuals than in the 12 susceptible individuals ( 29 . 1% vs . 16 . 7% ) . The L1534S mutation frequency was 23 . 3% in the selected resistant strain but 0 in the laboratory susceptible colony . Therefore , the impact of mutations in the kdr gene on pyrethroid resistance is contingent on the mosquitoes’ genetic background . Mutations in other genes and expression of detoxification genes need to be considered in developing predictive biomarkers for resistance . Our study has several limitations . First , only 12 resistant and 12 susceptible individuals were subjected to RNA-seq analysis . RNA-seq using a large sample size would detect more genetic variants with small effects on resistance . Secondly , the laboratory resistant population used for RNAi functional studies was not highly resistant to deltamethrin , which lowered the ability to detect candidate genes’ phenotypic impact . Nonetheless , the present study reveals the major genetic variants associated with pyrethroid resistance in Ae . albopictus .
By comparing the expression profile of deltamethrin-resistant and -susceptible wild mosquitoes from China , we identified 3 candidate genes with increased expression in resistant Ae . albopictus . We determined that if interfered with , the mosquito would change phenotype status from resistant to susceptible , suggesting that these genes play a role in insecticide resistance . We also identified 5 SNPs in 4 P450 genes that were significantly associated with insecticide resistance . Overall , this study demonstrates the power of individually based transcriptome profiling , with the combination of RNAi and genotype-phenotype association analysis , in research on resistance mechanisms . That is , both differentially expressed genes and SNPs associated with pyrethroid resistance were identified in Ae . albopictus mosquitoes . Findings from this study provided candidate resistance genes and SNPs for future functional verification using the new gene editing techniques such as CRISPR/Cas9 . | Aedes albopictus represents a serious threat to public health as a vector for a number of arboviruses , such as dengue , Zika , and chikungunya viruses . Its ability to resist insecticides has recently been detected in several countries . However , little is known regarding molecular mechanisms that confer pyrethroid resistance . Determining such mechanisms is essential for early detection , monitoring , and management of insecticide resistance . Here , we use high-throughput RNA sequencing techniques , RNAi , and genotype-phenotype association study to characterize qualitative and quantitative differences in gene expression as well as single nucleotide polymorphisms between resistant and susceptible Ae . albopictus individuals collected from a dengue-endemic area in southern China . We identified 3 candidate genes and 5 candidate SNPs associated with deltamethrin resistance . These results enhance the understanding of insecticide resistance mechanisms among Aedes albopictus . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion",
"Conclusions"
] | [
"death",
"rates",
"sequencing",
"techniques",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"rna",
"interference",
"gene",
"regulation",
"animals",
"genome",
"analysis",
"epigenetics",
"molecular",
"biology",
"techniques",
"rna",
"sequencing",
"population",
... | 2018 | Comparative transcriptome analysis and RNA interference reveal CYP6A8 and SNPs related to pyrethroid resistance in Aedes albopictus |
In South America , the Protist parasite that causes visceral leishmaniasis , a potentially fatal human disease , is transmitted by blood-feeding female Lutzomyia longipalpis sand flies . A synthetic copy of the male produced sex-aggregation pheromone offers new opportunities for vector control applications . We have previously shown that the pheromone placed in plastic sachets ( lures ) can attract both females and males to insecticide treated sites for up to 3 months . To use the pheromone lure in a control program we need to understand how the application of lures in the field can be optimised . In this study we investigated the effect of increasing the number of lures and their proximity to each other on their ability to attract Lu . longipalpis . Also for the first time we applied a Bayesian log-linear model rather than a classic simple ( deterministic ) log-linear model to fully exploit the field-collected data . We found that sand fly response to pheromone is significantly related to the quantity of pheromone and is not influenced by the proximity of other pheromone sources . Thus sand flies are attracted to the pheromone source at a non-linear rate determined by the amount of pheromone being released . This rate is independent of the proximity of other pheromone releasing traps and indicates the role of the pheromone in aggregation formation . These results have important implications for optimisation of the pheromone as a vector control tool and indicate that multiple lures placed in relatively close proximity to each other ( 5 m apart ) are unlikely to interfere with one another .
The sand fly Lutzomyia longipalpis ( Diptera: Psychododae ) is the major vector of Leishmania infantum , a Protist parasite and the causative agent of Zoonotic Visceral Leishmaniasis ( ZVL ) in Latin America . Approximately 90% of the cases of ZVL that occur in the Americas are recorded in Brazil where the greatest number of cases were found in the North East of the country [1] . The range of the vector has been gradually expanding and consequently human and canine cases of the disease are now found throughout the central and southern states where it was previously absent [1 , 2] . In Minas Gerais State , Brazil , ZVL transmission is intense and over the last few decades has expanded from rural regions into cities [3 , 4] . Incidence rates within the state were 1 . 6 per 100 , 000 inhabitants in 2012 and 1 . 4 per 100 , 000 in 2013 , almost equal with incidence rates in North-eastern Brazil [5] . The causes of ZVL urbanisation are unclear but it is likely to be related to the movement of people and their animals from rural to urban settings as well as the ability of the vector to adapt to an urban environment [1 , 6] . In Governador Valadares ( GV ) ( a municipality in eastern Minas Gerais State ) and the surrounding areas , ZVL transmission was believed to have been reduced after intervention with DDT spraying and extensive dog culling in the 1960’s [7] . However , it has re-emerged as a public health concern after the control campaign was interrupted in the 1990’s [8 , 9] . From 2008 until 2013 , 127 human cases were recorded with a fatality rate of 16% . In 2015–2017 the Centre for Control of Zoonoses ( CCZ ) reported 53 cases of human VL with 9 deaths [10] . Domestic dogs infected with Le . infantum are the proven reservoir host of human infection [1 , 11] . Studies in GV in 2007 found that an average of 13% of dogs from 175 samples obtained across 2 districts; one urban and one rural , were seropositive [8] . That average had risen to 30 . 2% during 2008–2011 from 16 , 529 dog samples taken from 35 urban and rural districts of GV [9] . Leishmaniasis control relies on control of the sand fly vector via 1 . / reactive spraying of insecticide as recommended by the Ministry of Health ( MoH ) 2 . / reservoir control which is focused on the proactive diagnosis and removal of infected dogs and 3 . / the use of therapeutic drugs [12] . However , despite these intensive efforts the vector and disease continue to affect new areas of the country [1] . Lu . longipalpis is a species complex and there are divergent views on how to define the members of the complex in Brazil . However , they can be distinguished from each other by the sex-aggregation pheromone that is produced by the males of each of the members of the complex [13] . Males of the most widespread member of the complex in Brazil , and the rest of South and Central America , produce ( S ) -9-methylgermacrene-B ( 9MGB ) [13] . The Lu . longipalpis sex-aggregation pheromone attracts both males and females to mating aggregations ( leks ) which can become very large and well established in a particular location , with males returning night after night to compete with each other for access to females [14] . A synthetic copy of the 9MGB pheromone has been developed and experiments with a prototype lure , which allowed sustained controlled release of the pheromone for 24 hours [15] , showed that increasing the quantity of prototype lures ( which is equivalent to increasing the amount of pheromone being released and therefore equivalent to a greater number of males ) from 1 to 10 increased the numbers of female Lu . longipalpis attracted by 278% [16] . Subsequently a long-lasting version of the lure which was attractive to females and males for up to 3 months in the field was developed . The long-lasting lure releases synthetic pheromone at an average rate ( which is likely to be influenced by the ambient temperature ) of approximately 4–7μg h-1 . This is equivalent to the estimated range of pheromone produced by a natural lek ( 1μg h-1 to over 10μg h-1 ) [15 , 18] . Located inside or next to pyrethroid insecticide-sprayed chicken sheds , the pheromone could be used as the attractive element of a ‘lure-and-kill’ ( sometimes referred to as “attract-and-kill” ) strategy for vector control . In the current study we investigated the potential to improve sand fly capture rates by increasing the numbers of long-lasting lures at the trapping site ( equivalent to increasing the release rate of pheromone ) . We also wished to establish the size of area around the long-lasting lure that might be under the influence of the pheromone as this has i ) direct implications for the spatial deployment of an “attract-and-kill” intervention , and ii ) may enhance our understanding of sand fly dynamics under experimental conditions . For example , in previous experiments with synthetic pheromone we established an experimental protocol whereby sand fly catches in pheromone baited traps were compared with those of un-baited control traps in a variety of situations [15–18] . However , the extent to which the catches of the control traps are influenced by the proximity of the pheromone baited traps is unknown . Based on limited laboratory-based wind-tunnel experiments pheromone ( with host odour present ) appears to attract females over 2 . 4m [19] , thus control traps have typically been placed 3m from test traps . From a control perspective , it is important to understand the likely influence of adjacent “lure-and-kill” focal points on each other , in order to optimise their placement . Therefore , we tested the potential interaction between adjacent pheromone baited traps by increasing the distance between them .
The study was carried out in Governador Valadares ( Minas Gerais State , Brazil ) , a city of approximately 280 , 000 people 320 km northeast of Belo Horizonte the state capital . The city is located in the valley of the River Doce where , according to the Köppen—Geiger classification , the climate is temperate , characterised by dry winters and hot wet summers [20] . High densities of 9MGB sex-aggregation pheromone producing Lu . longipalpis have been found in several districts of Governador Valadares where Lu . intermedia and Lu . cortelezzii , vectors of American cutaneous leishmaniasis ( ACL ) , have also been found [8 , 9] . Experiments were conducted in 4 houses , based within four separate neighbourhoods of the city ( Fig 1 ) . House 1 was in Vila Mariana within the urban perimeter of the city , not far from the city centre , whilst the other 3 houses were on the opposite side of the Rio Doce and outside the main urban perimeter . House 3 in Vila Parque Ibituruna was considered semi urban and the two houses in Village da Serra ( House 2 ) and Chacaras Recanto da Cachoeria ( House 4 ) were considered to be rural properties . The inclusion criteria for households were; accessibility , the ownership of chickens ( >5 ) and Lu . longipalpis catch rate ( average >15 , males and females combined per trap per night ) and finally the home owners’ willingness to comply with the long-term study requirements . A pre-experimental investigation in 20 houses selected at random one week prior to the experiment determined that 4 of the houses surveyed met the inclusion criteria; mean number of Lu . longipalpis = 35 ( SE = 15 , SD = 30 ) . All experiments were undertaken within the ‘quintal’ ( yard ) area of the property where small bushes , some grass and fruit trees were characteristic of the vegetation . The chickens either roosted overnight in trees or a “shed” , constructed from wood , chicken wire , corrugated sheeting and any other construction materials available to the home owner . Some qualitative characteristics of the four study houses are summarised in Table 1 . The experiments took place between July and September 2016 which coincided with the end of winter and the beginning of summer , this period is characterised by dry and relatively cool weather ( average rainfall <20mm and average nightly temperature 22°C ) . All experiments were conducted using modified Hoover Pugedo ( HP ) Centers for Disease Control ( CDC ) style light traps , without a light and suspended inside experimental chicken sheds [17] . Sand flies attracted to the HP trap were collected in nylon netting Barraud cages suspended beneath the trap . Experimental chicken sheds were made from 4 plywood panels ( 105cm long , 55cm wide ) , arranged in a square plan ( 55cm x 55cm ) . The panels were held together by cable-ties passed through holes ( 10 mm diameter ) drilled in the top and bottom corners of each panel . A wooden dowel ( 20 mm diameter ) placed across the top of the experimental chicken shed was used to suspend the HP trap inside the shed . A chicken , chosen at random from the householder’s chicken roost , was placed on the ground inside the experimental shed overnight . Pheromone lures , each containing 10mg of synthetic sex pheromone , were suspended from the underside of the lid of the HP trap [17] . Traps were set out at dusk between 15:00 and 17:00 hours local time . The netting cages and pheromone lures were removed from the HP traps the following morning between 07:00 and 09:00 hours and returned to the laboratory for examination . Sand flies were removed from the cage using a battery powered aspirator , placed in a -20°C freezer and numbers of both male and female Lu . longipalpis sand flies were determined by examination under a stereo-microscope ( x20 ) ( Quimis Q744S , SP , Brasil ) . Lures were placed in a freezer between sampling points to prevent loss of pheromone . The aim of the first experiment was to measure the effect of increasing the number of lures on Lu . longipalpis recruitment to pheromone treated sites . As such the experimental set-up used 1 pair of experimental chicken sheds per house . Each pair included a test shed and a control shed set 30 metres apart and each experimental shed contained a HP trap . Within a pair of sheds the position of the test and control were alternated each night to control for any potential positional bias . The pairs of traps were placed at 2 out of the 4 available household sites ( e . g . the combination with 2-lures in the test shed vs . 1-lure in the control was performed in houses 1 and 2 , the combination with 5-lures in the test shed vs . 1-lure in the control was carried out in houses 2 and 3 , the combination with 10-lures in the test shed vs . 1-lure in the control was carried out in houses 3 and 4 , the combination with 20-lures in the test shed vs . 1-lure in the control was carried out in houses 1 and 2 , finally the combination with 50-lures in the test shed vs . 1-lure in the control was carried out in houses 3 and 4 ) . The two houses used for each of these combinations of lures were chosen in a semi-recurring sampling design . All experimental sheds ( control and test ) were placed so that they were equidistant from the normal chicken roost ( i . e . where the chickens normally roosted overnight ) . This resulted in the chickens’ normal roost being positioned mid-way between the test and control experimental chicken sheds . Chicken roosts were located ca . 5m from the house . Each combination was tested for 6 nights at each household location for a total of 60 trap nights ( raw data provided in S1 Table ) . The aim of the second experiment was to determine over what spatial scale competing experimental sheds interact . Using paired experimental chicken sheds , the test trap was baited with 5 lures and the control trap with 1 lure . Pairs of experimental chicken sheds were set out either 5 , 10 , 20 or 30 m apart and nightly numbers of male and female Lu . longipalpis captured determined as previously described . The combination with the test shed vs . control shed 5 m apart was performed in houses 1 and 3; 10 m apart was performed in houses 2 and 4; 20 m apart was performed in houses 1 and 3; 30m apart was performed in houses 2 and 3 . Each distance between control and test traps ( 5 , 10 , 20 , 30 m ) was tested for 6 nights at each of two separate households ( raw data provided in S2 Table ) for a total of 48 trap nights . As in Experiment 1 , the two houses used for each of these combinations of lures were chosen in a semi-recurring sampling design . In order to estimate 1 ) the attractiveness of different numbers of pheromone lures and 2 ) the interference between the test and control trap , we compared the number of sand flies caught by the test traps with those caught in the control traps for each respective experiment and employed a Bayesian log-linear model for analysis of contingency tables [21] . Bayesian analysis allows the calculation of posterior probabilities of the full model and its sub-models ( allowing for covariate selection ) , independent of submodel size and structure , and therefore allows the inclusion of uncertainty into the inferential process [22] . Tables 2 and 3 in the Results section show the contingency tables prepared for the 2 experiments , summarising the sand fly catches for all possible combinations of house , trap type ( test vs control ) and experimental test condition ( lure number or distance ) . As such , the total number of entries per contingency table ( yi ) is 40 for Experiment 1 , and 32 for Experiment 2 , including 0s for non-experimental nights ( experimentally untested combinations ) . If we exclude experimentally untested combinations , data are reduced to 20 non-zero , usable entries for Experiment 1 and 16 for Experiment 2 . In our Bayesian model , the contingency table entries , yi , are observations of independent Poisson random variables with mean μi , yi ~ P ( μi ) , and likelihood: l ( μ|y ) =∏i=1n1μi ! μiμie-μi with mean parameter μ modelled as log ( μi ) =βxiT=gi where β is the vector of ( regression ) coefficients for the dummy variables or indicators xi of a given variable X ( e . g . in the case of lures , the dummy variable for the 5-lure experimental condition will contain 1s for those entries associated with 5 lures and 0s for the others ) . Since μ has the same length of y , the model is saturated i . e . the number of parameters equals or is larger than the number of data entries ( but see below for the solution ) . The focus of this analysis was on the β coefficients , which are the measure of the association between the factors ( household , trap type and test condition ) , their combinations ( interaction ) , and the number of sand flies caught in traps . The meaning of the β coefficients in the log-linear model is identical to log odds ratios . The first step was , therefore , to assign a prior to the β coefficients that can allow identifiability of these parameters . Therefore , we assumed a noninformative Zellner’s G-prior ( a multivariate normal distribution ) [23] for the β coefficients of the indicator variables contained in X: β~Ni ( 0i , σ2 ( XTX ) -1 ) where σ2 is the scale parameter . The β posterior is: p ( β|y ) ∝|XTX|12Γ ( m2 ) ( βT ( XTX ) β ) -m2π-m2exp{ ( ∑i=1nyixi ) Tβ-∑i=1nexp ( xiTβ ) } where m is the number of β parameters; Γ is the gamma function; and π is the Pi greco constant . The constant terms |XTX|12 , Γ ( m2 ) andπ-m2 were applied in the equation because they change according to the sub-model dimension ( considered here in order to compare the importance of factor combinations in the full model ) . The posterior distribution of β cannot be derived analytically . Samples from the posterior distribution of β were obtained by using a Markov Chain Monte Carlo ( MCMC ) , an algorithm that allows the exploration of all the important regions of parameter-space . The MCMC algorithm is based on a random walk Metropolis-Hastings sampler proposed by Marin & Robert [23] . That is , initial β coefficients values and covariance matrix were obtained from a maximum likelihood estimation method . New β coefficients values are proposed from the G-prior using initial ( later updated ) β coefficients , a scale parameter ( fixed at 0 . 5 ) and the fixed covariance matrix . Proposed β coefficients are accepted or rejected based on the log-likelihood ratio between p ( β│y ) with proposed β coefficients and p ( β│y ) with initial or updated β coefficients . To apply the above model to the contingency Tables 2 and 3 ( presented in the results section ) , we first converted the three factors ( test/control , house , lures or distance ) into indicators . We called the indicators u ( test and control ) , v ( house number ) and z ( number of lures in Experiment 1 , or distance between experimental boxes in Experiment 2 ) . u takes L values ( i . e . 2 ) , v takes J values ( i . e . 4 ) and z takes K values ( i . e . 5 in Experiment 1 or 4 in Experiment 2 ) , so that the log-model for the mean parameter can be rewritten as: log ( μi ( l , j , k ) ) =g+glu+gjv+gkz+gljuv+glkuz+gjkvz+gljkuvz for l in 1 , ‥ , L; j in 1 , ‥ , J; and k in 1 , ‥ , K . where g is the reference average effect identifying marginal discrepancy for terms like glu or interaction discrepancy for terms like gljuv . In Experiment 1 ( varying the number lures , with a constant distance between test and control ) , we assumed no three-factor interaction , i . e . gljkuvz=0 , and no interaction between house ( v ) and number of lures ( z ) , i . e . gjkvz=0 . The latter is equivalent to considering v and z conditionally independent given u . Therefore , the full model is: log ( μi ( l , j , k ) ) =g+glu+gjv+gkz+gljuv+glkuz As seen in Tables 2 and 3 , half of the comparisons are not made , therefore we have 20 entries for 30 parameters ( saturated model , parameters coming from the dummy variables for test , house and number of lures , and permitted interactions between house x test , house x number lures , test x number of lures ) in experiment 1 . We have 16 entries for 27 parameters in experiment 2 ( dummy variables for test , house and distance , and permitted interactions between house x test , house x distances , test x distances ) . To ensure identifiability of the parameters , which makes the ANOVA comparison between the full model and its sub-model feasible , constraints are imposed . By setting to zero the parameters corresponding to the first category of each variable ( excluded category ) : g1u=g1v=g1z=g11uv=g12uv=g21uv=g13uv=g14uv=g11uz=g12uz=g21uz=g13uz=g14uz=g15uz=0 the saturated model becomes non-saturated ( 20 entries for 16 parameters ) . In Experiment 2 ( varying distance between test and control traps with a constant number of lures ) , we again assumed no three-factor interaction , gljkuvz=0 , and no interaction between house ( v ) and distance between chicken shed ( z ) , gjkvz=0 . Again , since half of the comparisons are not made the number of entries is 16 for 27 parameters . By setting to zero the parameters corresponding to the first category of each variable: g1u=g1v=g1z=g11uv=g12uv=g21uv=g13uv=g14uv=g11uz=g12uz=g21uz=g13uz=g14uz=0 the total number of parameters is reduced to 14 for 16 entries ( non-saturated model ) . The sub-models considered for Experiment 1 are: log ( μi ( l , j , k ) ) =g+glu+gjv+gkz+glkuz without interaction test/control and house number log ( μi ( l , j , k ) ) =g+glu+gjv+gkz+gljuv without interaction test/control and number of lures and for Experiment 2: log ( μi ( l , j , k ) ) =g+glu+gjv+gkz+glkuz without interaction test/control and house number log ( μi ( l , j , k ) ) =g+glu+gjv+gkz+gljuv without interaction test/control and distance between chicken sheds . Comparisons between full model and submodels ( full model in which an interaction between factors is removed ) are made by calculating the Bayes factor [24]: BF=likelihoodfullmodellikelihoodsubmodel We then take the log10 ( 1/BF ) [25] and interpret this value by using Jeffrey’s scale of evidence [26]: if log ( 1/BF ) is larger than 1 , we consider the removed interaction to be significant in the full model ( the extended scale of evidence is: removed interaction anecdotal for values <0 . 5; substantial for values between 0 . 5–1; strong for values between 1 and 2; and decisive for values > 2 ) . The statistical analysis was repeated for the data in both Tables 2 and 3 and calculated for the number of caught male sand flies , number of caught female sand flies and total ( male+female ) sand flies . The analysis was performed in R-cran software [27] , specifically using the Bayes package [28] for the MCMC algorithm calculations ( functions hmnoinfloglin and loglinnoinflpost ) . The project , including the involvement of householders , was reviewed and approved by the Faculty of Health and Medicine Ethical Review Committee ( FHMREC15125 ) at Lancaster University . Consent was obtained from the Governador Valadares district health authority Centro de Controle de Zoonoses ( CCZ ) to conduct the study within their administrative jurisdiction . This study was carried out in accordance with the guidelines of the Animals in Science Regulation Unit ( ASRU ) and in compliance with the Animals ( Scientific Procedures ) Act ( ASPA ) 1986 ( amended 2012 ) regulations and was consistent with UK Animal Welfare Act 2006 and The Welfare of Farmed Animals ( England ) Regulations 2007 and 2010 .
The models for experiment 1 and 2 were run for 100 , 000 iterations to evaluate if the MCMC engine converged ( in other words , if we reached a stable configuration of the posterior and its parameters ) . Thus , we investigated the MCMC traces and histograms generated in the Bayesian analysis . The MCMC traces for each β coefficient , and the histograms for each β posterior distribution for experiment 1 and 2 are provided in supplementary files S1 , S2 , S3 , S4 and S5 Figs . Convergence statistics for experiment 1 are shown in Table 4 ( and in supplementary file S3 Table for experiment 2 , where distance 5m and house 1 were the reference variables ) , where for the last 30 , 000 iterations divided in 3 blocks of 10 , 000 iterations , the variations in the posterior mean and posterior variance are presented . These values are small and guarantee stable posterior histograms for each coefficient . The overall conclusion is that the models for experiment 1 and experiment 2 converged to stable posteriors for all the parameters . This guarantees a good approximation of the credible intervals ( CR ) for all the parameters . Mean , variance and CR of the β coefficients for the last 30 , 000 iterations are shown in Table 5 . This table shows that the interaction between test and control traps and number of pheromone lures are all significantly different from zero ( ßcl5 , ßcl10 , ßcl20 , ßcl50 ) despite some non-interactive term not being significantly different from zero ( shaded rows ) ( house 2 and 3 and lure numbers 5 and 10 ) . Overall , differences in catches between test and control traps were positively associated with the different numbers of pheromone lures deployed in the trap . CR is the credible interval . Inter , is the intercept; test is the variable containing test and controls ( 0 for controls and 1 for tests ) ; ch is the interaction between test and house; cl is the interaction between test and pheromones; h is the house ( house number 2 , 3 and 4 ) ; and l is the number of pheromone lures . Shaded areas indicate non-significant interactions ( i . e . the credible interval crosses 0 ) . In the ANOVA comparison between the full model and its components , by adopting the Jeffrey’s scale of evidence ( see Methods ) the interaction between house number and test and control is substantial ( log10 ( 1/BF ) = 0 . 73 ) , while the interaction between number of lures and test and control is decisive ( log10 ( 1/BF ) = 14 . 65 ) . In addition , and to test our initial assumption that there was conditional independence between house number and level of pheromone ( i . e . house characteristics are not influenced by pheromones and vice versa ) we ran sub-models containing only a single interaction , and we found that only the interaction between test and control and number of lures is decisive ( log10 ( 1/BF ) = 12 . 01 ) , compared to test and control and house number ( log10 ( 1/BF ) = 0 . 05 ) and house number and number of lures ( log10 ( 1/BF ) = 0 . 0001 ) ( which for Jeffrey’s scale of evidence are considered anecdotal ) . Given the results shown in Tables 4 and 5 , we were able to restrict the full model to contain only the interaction between test and control and number of lures ( largest coefficients and significantly different from 0 for all the houses ) . The restricted ( parsimonious ) model , includes coefficients for the indicators for the number of lures and indicators of the interactions between test and control and number of lures are reported in Table 6 ( and histograms in supplementary information S5 Fig ) . Table 6 shows that all the coefficients for the pheromone model are significantly different from 0 . Table 7 shows that in all the comparison between lures , 20 lures has the largest ratio . A comparison between the coefficients of the interactions ( Table 6 and Fig 2 ) shows the greatest increase in capture rate ( relatively larger coefficient ) from 2 lures ( 2 . 6–6 . 2 , 95% CR , 2vs20 in Fig 2 ) to 20 lures ( indicated with βcl20 in Table 6 ) , which represents a 3 . 8-fold increment in total sand fly capture as lures are increased 10-fold . By comparison , the change in coefficient from 20 and 50 lures indicates the capture rates are not significantly increased . In the cross-comparison between numbers of lures ( Table 7 ) , statistically significant increases ( when both the CR limits are above one ) were found for all the comparisons apart from between 10 and 5 lures or between 20 and 50 lures . The interaction between test and control and house number has a log10 ( 1/BF ) = 0 . 92 ( substantial for the Jeffrey’s scale of evidence ) for male and 0 . 001 ( anecdotal ) for female sand flies; while the interaction between test and control and number of lures has a log10 ( 1/BF ) = 6 . 40 ( decisive ) for male and 5 . 56 ( decisive ) for female sand flies . This means that even when considering the sand flies by sex , the house has little effect on the number of catches , while the pheromone quantity influences both male and females sand fly trap catches , with a slight preference for male sand flies ( however this may also reflect the proportion of male/female in the sand fly populations ) . In fact , even if female sand flies show the largest fold increase from 2 lures to 20 lures ( 5 . 9 ) , this is statistically insignificant ( -33 . 1–41 . 2 , 95%CR ) . Table with numbers of male and female sand flies is shown in S4 Table . In this experiment , the numbers of male , female and total sand flies attracted to test ( 5 lures ) and control ( 1 lure ) experimental sheds , were investigated at different distances ( 5 , 10 , 20 and 30m ) between sheds . No significant relationship was found for either male , female or total sand flies when modelled with test and control , house and distance variables ( taking as reference variables the distance at 5m and house number 1 ) . In particular: Anecdotal interaction between test and control and house number , log10 ( 1/BF ) = 0 . 004 for total number of sand flies; 0 . 0009 for male and 0 . 103 for female sand flies; Anecdotal interaction between test and control and distance between experimental boxes , log10 ( 1/BF ) = 0 . 0002 for total number of sand flies; 0 . 001 for male and 0 . 001 for female sand flies . Experiment 2 results show that the distance between the test and control chicken sheds does not influence the number of sand flies caught in the test or control traps . Instead the numbers of sand flies caught is determined by the number of lures used in the trap , as the difference between test and control trap catches was significant at all the distances apart ( confirming results above ) . ß coefficient for the dummy variable test representing 5 lures is equal to 1 . 7 ( 1 . 39 , 2 . 02 , CR ) for the total number of sand flies , 1 . 3 ( 0 . 76 , 2 . 05 , CR ) for female sand flies , and finally 1 . 8 ( 1 . 43 , 2 . 22 , CR ) for male sand flies . These coefficients do not change significantly when considering the same model applied to each distance individually: median difference from 1 . 7 of -0 , 03 ( -0 . 21 , 0 . 16 CR ) for the total number of sand flies , median difference from 1 . 3 of 0 . 03 ( -0 . 34 , 0 . 41 , CR ) for female sand flies , and finally median difference from 1 . 8 of -0 . 04 ( -0 . 26 , 0 . 16 , CR ) . The results suggest that differences between test and control are similar at the different distances and that increasing the separation of the test and control traps does not favour trapping either males or females . This finding seems to indicate that the trapping is operating within a spatially homogeneous sand fly population in the peridomestic environment . If the sand fly population was spatially heterogeneous we might expect the proportion of flies caught in the 1-lure traps to 5-lure traps to change substantially as the distance between the traps changed .
This is the first time that a Bayesian log-linear model has been employed to quantify exogenous effects on sand fly catches . The model used allowed analysis of the contingency table obtained from multiple concurrent experiments containing categorical variables only , to identify the most important factors affecting the number of sand fly catches and to include model and data uncertainty in the model inference . Classically , analyses using simple ( deterministic ) log-linear models of sand fly count data are applied ( e . g . [29 , 30] ) , this can lead to limited interpretation of the β coefficients , and therefore of the effect of each factor , since a measure of uncertainty is missing . In addition , simple ( deterministic ) log-linear models do not allow for a comparison between the distributions ( the values ) of the β coefficients , which allows us to obtain credible interval data from differences between two β coefficients ( Fig 2 ) . To our knowledge , a similar approach has only been applied once before in a study that examined the differences between human landing catches and light trap catches for capture of Anopheles gambiae [31] . However , that publication did not account for the interaction between levels within and between each variable , leaving the method ( and analysis ) not fully exploited . Our approach considered the interactions between factors and therefore allowed us to dissect the effect of each factor on the full model and consequently view the outcomes with a high degree of certainty . The first experiment showed that increasing the number of lures increased the total number of Lu . longipalpis ( both males and females ) caught by the traps ( Table 2 ) . Overall , we collected more male than female sand flies confirming previous observations , using a variety of different trap types in GV and elsewhere , that suggest that there may be more male than female Lu . longipalpis in the population [6 , 15–17] . A group of 20 or 50 pheromone lures was significantly more attractive than groups of 2 , 5 or 10 lures ( Table 6 ) . Overall the increase in catch was not proportional to the increase in the number of lures i . e . the number of Lu . longipalpis caught increased in steps ( Fig 2 ) as the number of lures increased . Increasing the quantity of pheromone lures is equivalent to increasing the release rate of the sex pheromone and therefore the effective distance at which the insects can detect the pheromone , it follows therefore , that more sand flies should be caught with higher quantities of pheromone than with lower quantities , as the pheromone would be able to attract sand flies from further away [32] . However , in this study , the relative contribution of additional lures to total catch reached a plateau: 50 lures were not significantly more attractive than 20 lures ( after the effect of houses on trap catches is accounted for ) . It is also possible that at very high release rates the pheromone is repellent [33] . When the proximity between the test and control traps , ( but not the quantity of pheromone ) , varied , we found that the proximity of the test traps ( 5 lures ) had no effect on the control trap catches ( 1 lure ) even when the test and control traps were only 5 m apart . At all the distances tested , the traps with 5 lures were significantly more attractive than the traps with 1 lure ( confirming previous result of experiment 1 ) , but critically , that this difference in capture rate ( in both absolute and relative terms ) between 5 lures and 1 lure was consistent over the tested distances . Thus , at distances equal to or greater than 5 m neighbouring traps do not appear to affect each other’s attractiveness , at least when the number of lures used is relatively small . Indicating no contamination between test and control traps over distances used in the established protocol . The differences between the test and control traps were consistent and significant at all the distances tested , therefore it is unlikely that the two traps are combining their effect . Although the role of the male pheromone as an attractant and in forming aggregations is well established [14 , 15 , 16 , 34] , this result may indicate the important role of the pheromone in maintaining aggregations . The results show that the sand flies are attracted to the area of the pheromone release in proportion to the amount of pheromone present and once in the vicinity of the pheromone the HP trap samples that population . The results suggest that the sand flies are aggregated ( or arrested ) at that site because the proportion of flies collected in the 1-lure traps compared to the 5-lure traps is similar regardless of how far apart they are placed . However , in these experiments once the sand flies enter the trap they cannot migrate towards the area of greater pheromone concentration . Therefore , we cannot be sure that eventually all sand flies would not move towards the site of greater pheromone release . In the future we could test this hypothesis using a mark release recapture experiment . It would be interesting to determine the minimum amount of pheromone required to maintain aggregations and if interaction between pheromone sources occurs at distances less than 5m . There is also a possibility that males captured in the Barraud cages would attract other males and females thus adding a source of error . We discount this as a major source of error for either experiment because of the small numbers of males involved compared to the pheromone released by the lures . An implication of this result is that naturally established aggregations would compete with synthetic pheromone lure-and-kill sites . However , little is known about the stability , longevity or density of naturally established aggregation sites [14] however the lure-and-kill sites would be active for long periods of time and present in greater numbers than untreated sites where aggregations are known to form . It could also be the case that the attractiveness of sites treated with synthetic pheromone lures would be enhanced by the presence of real male L . longipalpis sex aggregation pheromone . The results also suggest the possible existence of a spatially homogenous Lu . longipalpis population in the peridomestic environment . This is an interesting possibility requiring further work to clarify the situation . Previous work using CDC miniature light traps has suggested that Lu . longipalpis is heterogeneously distributed and that adult males and females aggregate sporadically in chicken sheds and other animal shelters . However , it is not understood why , in an area with many potential aggregation sites , aggregations appear to develop at some sites but others . Our historical understanding of Lu . longipalpis distribution may be related to the use of miniature CDC light traps which possibly may not be adequate for sampling the Lu . longipalpis population . Taken together the results of experiment 1 and 2 suggest that the most effective way to use the pheromone would be to use as many lures as possible distributed widely in any given area rather than use the same number of lures grouped together in a small number of places within the same area . Increasing the number of lures and thereby increasing the release rate of pheromone would also increase trap catches but clearly there is a trade-off between the number of lures used , their cost and effectiveness . The effect of the house location was substantial but not decisive ( apart from house 4 , the only one with a significant coefficient when considering the full model ) , in other words , house conditions slightly improve the model and therefore additional studies should focus on which factors , related to the house , influence the overall catches in order to be taken into account in future analyses . It has been shown that sand flies , in particular Lu . longipalpis , favour humid environments with lots of vegetation [35] . Habitat-specific effects on pheromone attraction is also a potentially important factor playing a critical role in shaping the response to pheromones [36] . This study showed that increasing the number of pheromone lures increased the numbers of sand flies captured . This is not a linear relationship and increasing the number of lures by a given factor does not lead to a similar increase in the number of sand flies caught . However , in the context of a control programme , greater numbers of pheromone lures placed next to an insecticide sprayed wall , would result in more sand flies being killed and potentially a more effective control programme . The second experiment highlighted the attraction and possible aggregation behaviour by Lu . longipalpis once a pheromone source had been located . In the context of a control program this result suggests that the sand flies would remain strongly aggregated at the synthetic pheromone insecticide treated site and that “lure-and-kill” sites even when positioned in relatively close proximity would not compete with each other . | Lutzomyia longipalpis sand flies are the insect vectors of the Protist parasite Leishmania infantum which causes visceral leishmaniasis ( VL ) in Brazil . Control of VL has focussed on vector and infected reservoir control , but despite the sustained efforts of the Brazilian Health authorities the disease burden doubled between 1990 to 2016 . New approaches to VL control are urgently needed . We previously demonstrated that Lu . longipalpis synthetic sex-aggregation pheromone placed alongside insecticide sprayed surfaces can attract and kill female sand flies . However , before the synthetic pheromone can be effectively exploited in any VL control program it is essential to understand how it might be deployed . In this study we investigated the effect of different amounts of pheromone and the spatial relationship between different pheromone sources on Lu . longipalpis catches . We developed a robust Bayesian analysis to fully exploit the field data which showed that optimal use of the pheromone could be achieved by placing individual or small numbers of pheromone releasing devices ( lures ) within the peridomestic environment and these can be positioned relatively closely without competing with each other . The results also revealed the significance of the pheromone in maintaining aggregations of Lu . longipalpis and suggested that Lu . longipalpis may be more evenly distributed in the peridomestic environment than previously recognised . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"statistics",
"tropical",
"diseases",
"geographical",
"locations",
"vertebrates",
"sand",
"flies",
"contingency",
"tables",
"parasitic",
"diseases",
"animals",
"mathematics",
"neglected",
"tropical",
"diseases",
"infectious",
"dise... | 2018 | Attraction of Lutzomyia longipalpis to synthetic sex-aggregation pheromone: Effect of release rate and proximity of adjacent pheromone sources |
Lasting alterations in sensory input trigger massive structural and functional adaptations in cortical networks . The principles governing these experience-dependent changes are , however , poorly understood . Here , we examine whether a simple rule based on the neurons' need for homeostasis in electrical activity may serve as driving force for cortical reorganization . According to this rule , a neuron creates new spines and boutons when its level of electrical activity is below a homeostatic set-point and decreases the number of spines and boutons when its activity exceeds this set-point . In addition , neurons need a minimum level of activity to form spines and boutons . Spine and bouton formation depends solely on the neuron's own activity level , and synapses are formed by merging spines and boutons independently of activity . Using a novel computational model , we show that this simple growth rule produces neuron and network changes as observed in the visual cortex after focal retinal lesions . In the model , as in the cortex , the turnover of dendritic spines was increased strongest in the center of the lesion projection zone , while axonal boutons displayed a marked overshoot followed by pruning . Moreover , the decrease in external input was compensated for by the formation of new horizontal connections , which caused a retinotopic remapping . Homeostatic regulation may provide a unifying framework for understanding cortical reorganization , including network repair in degenerative diseases or following focal stroke .
The mature brain is not as hard-wired as traditionally thought . Long-term in vivo imaging has revealed that dendritic spines appear and disappear frequently , accompanied by synapse formation and elimination [1] . Spine and synapse formation and elimination are induced by learning [2]–[4] and are associated with long-term memory storage [5]–[7] . Similarly , peripheral lesions , which permanently alter input to cortical areas , trigger extensive spine formation and elimination [8]–[11] . Likewise , large-scale axonal sprouting and pruning in cortical areas are associated with focal retinal lesions [12] , [13] , whisker trimming [8] , and digit or limb amputation [14] , [15] . Axonal and dendritic arborizations are profusely intertwined [16] , so a neuron can already access a large pool of neurons by just extending its dendritic spines or slightly changing the length of its neurites ( axons or dendrites ) . Despite the relevance of structural changes for cortical adaptations after alterations in sensory input [1] , [9] , [10] , the driving forces behind structural plasticity remain elusive . Particularly , it is unknown whether neuronal structural changes may be induced primarily by the neuron's own activity level or critically depend on the activity level of other neurons as well , as in associative forms of synaptic plasticity such as STDP . We therefore investigated whether the need of neurons to maintain their average electrical activity at a particular level ( homeostatic regulation ) could account for the changes in neuronal morphology and the functional rewiring of cortical circuitry observed in the primary visual cortex after focal retinal lesions [10] , [13] . Electrical activity controls the growth of axons [17]–[21] and dendrites [22]–[26] , and the formation of spines [11] , [27]–[30] and boutons [13] . Alterations in dendritic membrane potential can induce very localized structural changes that seem to follow Hebbian synaptic plasticity rules [31] . After lesions , however , when the activity of the whole neuron is altered , the soma may innitiate compensatory structural changes of the entire neuronal morphology . The way in which the neuron's electrical activity influences neurite outgrowth and synapse numbers suggests that neurons try to maintain their level of electrical activity at a particular set-point ( homeostasis ) [32]–[36] . When electrical activity is below the homeostatic set-point and above a certain mimimal level , neurons extend their dendrites towards sources of activity [37] , [38] and enhance spine formation [1] . When electrical activity is too high or very low ( below a required minimal level ) , neurons arrest neurite outgrowth and eliminate spines [11] , [39]–[43] . What global network dynamics may arise from such local activity-dependent changes in neuronal morphology ? Changes in neuronal morphology affect synaptic connectivity , altering neuronal and network activity , which in turn changes neuronal morphology . The consequences of these reciprocal interactions are hard to predict from experimental studies alone . Computational modelling is required to help elucidate these complex dynamics . Most computational models [44] , [45] , however , consider network structure as fixed , with plasticity merely arising from changes in connection strengths ( synaptic plasticity ) . To explore what structural changes in neurons and network circuitry may emerge after a loss of input , we used a novel computational model , called the structural plasticity model [36] , [46] , [47] , in which a neuron creates new spines and boutons when its activity level is within an optimal range , and delete spines and boutons when activity is outside this range . We found that the dendritic and axonal dynamics in the model after a loss of external input were remarkably similar to the formation and deletion of dendritic spines in mice [10] and the changes in axonal boutons in monkeys [13] after focal retinal lesions . The model generated , as an emergent property of the local activity-dependent rules for structural plasticity , an axonal overshoot and subsequent pruning of horizontal connections that re-occupy deafferented neurons and so cause a retinotopic remapping . Our results demonstrate that functional reorganization in terms of cortical remapping can result from structural rewiring of cortical circuitry alone and does not necessarily require associative forms of synaptic plasticity such as STDP [48] . The model generates a variety of testable predictions with respect to cortical circuit rewiring following changes in sensory input .
As shown in mice [10] and monkeys [13] , a circumscribed loss of input to cortical networks triggers massive structural changes in neuronal morphology . We used our structural plasticity model ( Fig . 1 ) to explore the impact of a loss of input ( referred to as ‘lesion’ ) on network rewiring caused by activity-dependent changes in axonal and dendritic morphology . A circumscribed area in the model , corresponding to the lesion projection zone ( LPZ ) in the animal experiment , is deprived of external input . The LPZ can be further subdivided into a center and a border region . Whereas the border region still receives some input via horizontal connections from outside the LPZ , the neurons in the center region are almost completely deprived of input . In accordance with the experimental literature , the area with intact external input directly surrounding the LPZ is defined as the peri-LPZ . In order to be able to predict what changes occur at the network level as a result of activity-dependent changes at the cellular level , the model is detailed enough to include measurable physiological variables such as membrane potential , spiking , intracellular calcium concentration and numbers of dendritic spines and axonal boutons , yet computationally simple enough to characterize network changes such as cortical rewiring and functional remapping . From a theoretical point of view , the probability that two neurons form new synapses or break existing ones depends on , in addition to the distance between the neurons , the number of synaptic contact possibilities that each neuron has , i . e . the number of axonal boutons and dendritic spines . In principle , changes in spine and bouton numbers can be caused directly by changes in the numerical densities of spines and boutons or indirectly by neurite outgrowth and retraction . Since we are interested in the effective change in synaptic contact possibilities , we abstracted away from the detailed neuronal morphology by using one-compartment neurons that carry sets of synaptic elements representing axonal boutons ( axonal elements ) , excitatory dendritic spines ( excitatory dendritic elements ) and inhibitory postsynaptic densities ( inhibitory dendritic elements ) . Excitatory and inhibitory neurons form exclusively excitatory or inhibitory axonal elements , respectively , but can express both excitatory and inhibitory dendritic elements . Synaptic elements can develop independently of potential contact partners [49] , [50] and are highly selective in connecting to matching synaptic counterparts [51] , which motivated our idea of representing axonal and dendritic elements as ‘plugs’ and ‘sockets’ ( cf . Fig . 1 ) . Synapses are formed by merging corresponding synaptic elements or are deleted when synaptic elements are lost . Thus , activity-dependent changes in the number of synaptic elements alter synaptic connectivity . The execution of the model procedes as follows . We continuously determine for each neuron its electrical activity on a millisecond timescale . Electrical activity causes influx of calcium . Calcium concentration also decays , so the neuron's intracellular calcium concentration effectively represents the time-averaged level of its electrical activity . The intracellular calcium concentration then drives continuous changes in the number of synaptic elements per neuron [23] . Structural changes , i . e . changes in the number of synaptic elements , are much slower than changes in electrical activity , so reorganization of network connectivity and electrical activity dynamics take place on different time scales and do not directly interfere with each other . Because formation and deletion of synaptic elements occur continuously over time ( reflecting an implicit growth process ) but the formation of synapses ( by merging synaptic elements ) and breaking of synapses are singular events , we update connectivity at discrete time steps . Since changes in synaptic elements are slow and the algorithm for updating connectivity is computationally costly , we do the update in connectivity only every . The structural dynamics are slow enough to be comparable to the experimentally observed morphological changes in the visual cortex of rodents after focal retinal lesion over a time span of 72 days . However , we did not unnecessarily slow down structural plasticity in the model to the point that the entire simulation time would sum up to 72 days in milliseconds . In every connectivity update , we determine for each neuron how many vacant synaptic elements are available for synapse formation , and if the number of synaptic elements decreased , which of its synaptic elements , either vacant or bound in a synapse , should be removed . In every connectivity update , deletion of bound synaptic elements immediately causes the breaking of synapses . Vacant synaptic elements ( i . e . not bound in a synapse ) spontaneously decay with a certain time constant . For synapse formation , we distribute vacant synaptic elements on complementary elements in a random , distance-dependent manner , i . e . excitatory axonal elements bind to excitatory dendritic elements , and inhibitory axonal elements connect to inhibitory dendritic elements . Synapse formation is more likely between adjacent than between remote neurons . This scheme is repeated for every update in connectivity . An important feature of the structural plasticity model is that after the breaking of a synapse , the complementary synaptic element remains , i . e . the axonal element if the dendritic element was deleted or the dendritic element if the axonal element was deleted . The vacant synaptic element is available for synapse formation again , so synaptic connections can rewire . The structural plasticity model used to model cortical deafferentation resembles the canonical cortical microcircuit [52] consisting of ( excitatory ) pyramidal cells and inhibitory , GABA-ergic interneurons . The mircocircuit is represented by a two dimensional recurrent network of 400 neurons , of which 80% are excitatory and 20% inhibitory [53] . In the model , excitatory and inhibitory neurons differ only in the sign of synaptic transmission; all other parameters are the same . Excitatory neurons were placed with a spatial variance of on a 20×16 grid with a distance between two grid points of . More precisely , the x , y-coordinates of each neuron were derived from a normal distribution ( with the chosen spatial variance as standard deviation ) that was centered at an individual grid point . For the 80 inhibitory neurons we defined a second 10×8 grid positioned in such a way that the inhibitory neurons become equally distributed among the excitatory ones; the precise x , y coordinates were determined as was done for the excitatory neurons . Given that neurons in the adult cortex of rodents [54] are capable of rewiring their axonal branches over a couple of hundred micrometers , the expected distance between neurons of in the model is a plausible choice . Neurons generate electrical activity . To describe neuronal firing , we used the Izhikevich model [55] , a biophysical neuron model that provides a good trade-off between physiological realism and computational cost . In this model , the membrane potential of a neuron is computed by two differential equations: ( 1 ) where is in , time is in , , , , is a membrane recovery variable and represents the input to the neuron . Each time the membrane potential exceeds ( i . e . fires a spike ) , and are reset to their resting values: ( 2 ) Parameters were set to , , , . Each neuron receives input that comprises synaptic input from other neurons in the network ( horizontal input from adjacent and distant cortical neurons ) and external input ( vertical input from the eye via the thalamus ) . Neurons interchange electrical signals on a millisecond timescale without a synaptic delay . Synaptic input consists of the incoming action potentials from the presynaptic neuron low-pass filtered by an exponential filter function with decay constant . Network connectivity is defined as the number of synapses from neuron to . If a synapse exists , it has a fixed strength of . Neurons are either excitatory or inhibitory . Indices refer to excitatory neurons if or and to inhibitory neurons if or . As in the original version of the neuron model [55] , we defined the external input as a noisy current ( white noise ) with mean and standard deviation . In [56] , the total thalamocortical input to individual neurons in the primary visual cortex was estimated experimentally to be about 5 mV . However , this study did not report on the frequency of this input . The frequency may roughly be estimated by the number of thalamocortical synapses impinging on an individual neuron in the primary visual cortex and by the firing frequency of the thalamocortical projection neurons . Synapse numbers have been estimated to be about one hundred ( cf . chapter 8 , page 323 in [57] ) , while the firing frequency is about 10 Hz for each projection neuron [58] . If one hundred projection neurons fire with an expected firing rate of 10 Hz , it is perhaps reasonable to assume that with input amplitude 5 mV is delivered on average at every millisecond . Although it is not exactly known how neuronal morphology in the visual cortex changes after loss of input due to focal retinal lesions , existing experimental data , e . g . from tissue culture , strongly support that changes in neuronal morphology are activity-dependent and that a certain level of activity is optimal for axonal and dendritic outgrowth and spine and bouton formation [33] ( reviewed in [32]; cf . also Introduction ) . The intracellular calcium concentration can be used as an indicator of a neuron's average electrical activity [59]–[61] . In the model , increases by a fixed amount every time neuron fires , and otherwise decreases exponentially to zero [61] with decay time . With these values of and the intracellular calcium concentration represents the time-averaged electrical activity of a neuron and the concentration decay is of the same magnitude as measured experimentally [62] . ( 3 ) Through changes in the level of intracellular calcium , electrical activity modulates actin and tubulin polymerization and therefore neurite outgrowth [33] . If the average electrical activity of a neuron exceeds some maximum level , it will withdraw dendritic spines [63] , [64] and retract neurite branches [65] , which may reduce connectivity and hence activity . If activity becomes lower , the neuron will generate vacant synaptic elements [66] , increasing connectivity and activity . However , if activity comes below some minimum level , outgrowth and synapse formation will be halted [10] , [42] , [43] , [67] . These findings are further supported in a study by Richards et al . [30] showing that reducing electrical activity in hippocampal slice cultures increases the number of spine head protrusions , whereas this increase does not occur with a complete block of activity . Earlier studies [27] , [28] are in line with these results . Thus , we conclude that homeostatic regulation in combination with a minimal level of electrical activity required for outgrowth may serve as a guiding principle for the formation of synaptic elements , i . e . for neurite outgrowth [17] , [33] and spine and bouton formation [68] , [69] . Synapses consist of merged pairs of synaptic elements , i . e . axonal boutons and dendritic spines . Changes in the number of synaptic elements can be brought about by outgrowth or retraction of neurites , or by changes in the numerical density of these elements on neurites . In the model , we abstract over these two possibilities and consider merely the number of synaptic elements rather than the detailed morphology of the neuron . We define as the number of axonal elements per neuron , representing the overall number of axonal boutons . Likewise , we define and as the number of excitatory and inhibitory dendritic elements , respectively , representing excitatory dendritic spines and inhibitory postsynaptic densities . Dendritic elements can be excitatory or inhibitory regardless of the type of the hosting neuron , whereas all axonal elements of a neuron are either excitatory or inhibitory . In the numerical integration , , and are treated as continuous variables , but when synaptic elements are deleted or used for synapse formation , the values of , and are rounded off to their smallest integer values . We postulate Gaussian-shaped growth curves for the activity-dependent formation and deletion of every type of synaptic element , i . e . excitatory and inhibitory axonal elements , excitatory dendritic elements and inhibitory dendritic elements : ( 4 ) where is the continuous number of synaptic elements . By choosing , we obtain the growth curve of a particular type of synaptic element ( Fig . 2 ) . Eq . 4 applies to the total number of synaptic elements of the respective type regardless of whether synaptic elements are bound in a synapse or unbound . Growth curves are bounded to , where is the maximum speed of increase or decrease in the number of synaptic elements , with . We used , which is small enough to represent the slow time scale of growth yet large enough not to slow down the simulations unnecessarily . Growth curves are further determined by a low and a high set-point , at which points there are no activity-dependent changes in the number of the respective type of synaptic element . Parameter is a stable fixed point to which the system will converge when of all neurons is higher than . The value of depends on the type of synaptic element ( cf . Results section for the choice of and ) , whereas is the same for all types . The center and width of the Gaussian-shaped growth curve ( Fig . 2 ) are given by and , respectively . Synaptic elements that are not bound in a synapse ( vacant elements ) decay with time constant connectivity updates . ( 5 ) For reasons of simplicity and because of lack of detailed experimental constraints , we used the same type of growth rules for both axonal and dendritic elements . We also used identical growth rules for excitatory and inhibitory synaptic elements . Every we perform an update in connectivity based on the activity-dependent continuous change in , and . The continuous changes determine how many discrete synaptic elements on neuron have to be deleted . For example , if neuron had previously 100 axonal elements bound in 100 outgoing synapses stored in ( with ‘ . ’ indicating the entire column of the connectivity matrix ) and decreased to e . g . 95 . 32 due to Eq . 4 , is rounded off to 95 and consequently neuron has to delete outgoing synapses at the next update in connectivity . For the update in connectivity , the algorithmic procedure then determines which connections to postsynaptic neurons are to be reduced . This is done by listing all individual axonal elements per neuron with the index of its postsynaptic target neuron and randomly selecting synaptic elements from this list for deletion . Vacant elements are included in this list and can be chosen for deletion as well . All synaptic elements have an equal chance to be deleted , regardless of whether they are vacant or bound in a synapse . The removal of a bound synaptic element immediately causes the breaking of the synapse , i . e . reducing the respective by one ( or more if more than one element bound to a dendritic element on postsynaptic neuron was chosen for deletion ) . Note that with breaking of synapses , the total number of dendritic elements of the postsynaptic neuron remains unchanged , but the number of vacant dendritic elements of neuron increases . The same procedure is done for all types of elements , and in every update in connectivity . The above is the description of the algorithmic procedure implemented . Given that a number of synaptic elements , and needs to be deleted , the expected loss of synapses between any pairs of neurons can be predicted analytically . For excitatory and inhibitory dendritic elements the change can be predicted as follows: ( 6 ) with . Likewise , the expected loss of outgoing synapses caused by deletion of axonal elements of neuron with either or is ( 7 ) If , excitatory synapses will break , whereas if inhibitory synapses will break . In both cases , the postsynaptic neurons affected by synapse loss can be excitatory or inhibitory . A neuron can undergo synapse formation when it has gained vacant synaptic elements . For each type of element , the number of vacant synaptic elements , and on neuron or is the difference between the total number and the number of bound synaptic elements , with the continuous variables , , being rounded off to their smallest integer values ( as described in “Update in synaptic elements due to synapse deletion” ) . The number of bound synaptic elements is equal to the number of incoming synapses ( for bound dendritic elements ) or outgoing synapses ( for bound axonal elements ) of a neuron . Thus , ( 8 ) Initially , the numbers of bound and vacant synaptic elements are zero . Vacant synaptic elements may come from an increase in synaptic elements due to the activity-dependent growth rules ( Eq . 4 ) or from the previous breaking of synapses formed earlier ( Eq . 6–7 ) . For synapse formation , all vacant synaptic elements from all neurons are randomly and simultaneously assigned to a complementary synaptic element , i . e . excitatory axonal elements to excitatory dendritic elements and inhibitory axonal elements to inhibitory dendritic elements . If there are more synaptic elements of a particular type than there are matching counterparts , some elements will remain vacant . Whether or not assigned pairs of complementary synaptic elements actually form synapses depends on the Euclidean distance between the neurons . A two-dimensional Gaussian kernel with width ( where is the distance between two grid points ) determines the distance-dependent likelihood for synapse formation between any pair of neurons: ( 9 ) with the x-coordinate and the y-coordinate of the location of postsynaptic neuron , and and the coordinates of presynaptic neuron in . The expected number of synapses to be formed from neuron to can be approximated on the basis of the number of vacant synaptic elements and the distance-dependent likelihood : ( 10 ) Eq . 10 provides , rather than a description of the algorithmic procedure implemented , an analytical approximation of the expected increase in synapse numbers for each update in connectivity . The denominator of Eq . 10 gives the maximum of the total number of vacant axonal elements and the total number of vacant dendritic elements in the entire network . The numerator is the product of the number of vacant axonal elements of presynaptic neuron and the number of vacant dendritic elements of postsynaptic neuron . Eq . 10 can be interpreted as follows: Given that there are , for example , in total more vacant dendritic elements in the network than vacant axonal elements , the vacant axonal elements from neuron are proportionally distributed over all postsynaptic neurons offering dendritic elements . In this example , we would expect ( ignoring for the moment the distance-dependent part of Eq . 10 ) that postsynaptic neuron will connect proportionally to as many axonal elements of presynaptic neuron as the number of vacant dendritic elements that postsynaptic neuron offers , namely . Hence , the allocation of vacant elements , and thereby the formation of new synapses , is more likely when a neuron has more vacant elements . The distribution of vacant synaptic elements , in combination with the neurons' distance to other neurons , determines the chance that a neuron connects to other neurons . Note that for pairs of distant neurons , the expected number of synapses to be formed could be much below one because of a very low . The expected number considerably increases when the number of vacant synaptic elements becomes higher . Such an increase in expected synapses can be interpreted as resulting from neurite outgrowth ( not explicitly represented in the model ) . Initially , synapses between distant neurons are unlikely , but over time , when more vacant synaptic elements are produced , synapses between distant neurons will also become established . For every simulation run , networks start with zero connectivity and develop connectivity over time in a self-organizing process according to the procedure described above . Synapse formation and reorganization of network connectivity ceases when electrical activities of each neuron is close to the desired homeostatic set-point . A rectangular lesion of the retina caused by photo-coagulation as described in [10] leads to a circumscribed deafferentation of the primary visual cortex—the lesion projection zone ( LPZ , Fig . 1 ) . The LPZ is modelled by depriving a central square of about 49 excitatory and 9 inhibitory neurons ( numbers can slightly vary between simulations , since neurons are placed with some jitter on the two-dimensional plane , whereas the size and position of the square is fixed ) of external input , i . e . by permanently setting . Before the lesion , all model neurons are at the high set-point of electrical activity , at which they do not considerably change their synaptic connectivity ( ‘mature’ network; cf . section “Electrical activity triggers changes in neuronal morphology” ) . This equilibrium state is usually obtained after 5000 connectivity updates , but to be absolutely sure that the network in all simulations reached an equilibrium , we applied lesions only after 8000 connectivity updates . After the lesion , simulations are then continued for another 5000 connectivity updates . In total , each simulation thus undergoes 13000 connectivity updates . In the time period after the lesion , we distinguished an early phase ( the first 1000 connectivity updates after the lesion ) , a middle phase ( 2000–3000 connectivity updates ) and a late phase ( 4000–5000 connectivity updates ) . The structural dynamics occuring over 5000 connectivity updates are comparable to the morphological changes observed experimentally over 72 days [10] , although in the model the speed of structural dynamics is faster than in reality in order not to prolong the simulation unnecessarily . We compared all measurements to an unlesioned network ( ‘control’ ) after the same number of connectivity updates . Mann-Whitney U test or ANOVA was used with a Bonferroni post-hoc test , as described in the text . is the number of cells for each condition . All error bars denote standard deviation ( SD ) unless otherwise stated .
Comparison with the experimental data revealed that growth rules with low and high produced dynamics of axonal and dendritic elements that are remarkably similar to the axonal bouton [13] and dendritic spine changes [10] observed in the viusal cortex after a focal retinal lesion . In agreement with the experimentally observed dendritic spine dynamics , our model produced the highest turnover rate in dendritic elements in the center of the LPZ ( Fig . 6 ) , which was significantly higher than the turnover in the border region , in the peri-LPZ and in controls . Turnover rate is defined as the sum of lost and gained dendritic elements/spines over a certain time period divided by twice the initial amount at the beginning of this period . In both the experiment and the model , also the border region showed an increased turnover rate compared to that in controls ( ; ANOVA with Bonferroni correction for multiple tests ) . Furthermore , strong similarities were found in survival fraction rate and cumulative addition rate between dendritic elements in the model and dendritic spines in the experiment [10] . Whereas the survival fraction rate indicates how many of the dendritic elements/spines before lesion are still present after lesion , the cumulative addition rate measures how many new spines compared to the initial number have been formed meanwhile . Interestingly , the survival fraction rates in both the experiment and the model showed a significant difference between the center and border of the LPZ ( from post-lesion day 4 onwards in the model; Mann-Whitney U test ) , whereas the cumulative addition rates in the center and border region did not differ significantly ( from day 45 onwards in the model ) . On the basis of our model results , we suggest the following explanation: The low activity level in the center of the LPZ leads to a stronger decrease in dendritic elements in the center as compared to the border and consequently to a lower survival fraction . While activity is the determining factor for dendritic element survival in the model —and possibly also for dendritic spines in the experiment—it is not the limiting factor for the cumulative addition of dendritic elements . In the model , the number of vacant axonal elements available for synapse formation and in reach for postsynaptic neurons is decisive for the cumulative addition rate of dendritic elements . Dendritic elements that did not bind to an axonal element quickly decayed and therefore limited the cumulative addition rates . Consequently , additional axonal boutons—in the visual cortex most likely provided by sprouting axonal branches [12] , [13]—stabilize dendritic spines and thereby increase cumulative addition rates of dendritic spines . It remains unclear , however , what could determine the total gain in dendritic spines in the experiment . While there is a clear increase in spine density after a loss of input to the visual cortex caused by juvenile or adult monocular deprivation [9] , the gain of total spine density after focal retinal lesions seems limited [10] . It is likely that if plasticity in synaptic strength was included in the model , fewer synaptic connections would be sufficient to enable neurons to return to the high set-point of electrical activity . Similarities between model and experiment were also observed with respect to axonal dynamics . The model produces an overshoot and subsequent pruning in number of axonal elements after the loss of input ( Fig . 7 ) . In the late phase after the lesion , the number of axonal elements remains significantly elevated compared to controls ( , Mann-Whitney U test ) . In the model , the surplus of axonal elements is the source for additional horizontal connections from the peri-LPZ into the LPZ . The overshoot is caused by an early phase of decreased activity in the peri-LPZ followed by a middle phase of activity higher than the set-point ( cf . Fig . 5 ) . Unfortunately , there is only limited experimental data on axonal growth into the LPZ after focal retinal lesions . The first examples were provided by [12] , [70] . More recently , fine structure imaging of sprouting axonal branches and proliferating boutons indeed revealed an overproduction and subsequent pruning of axons and boutons [13] . The presence of axonal overshoot in the model strongly suggests that the dynamics in the model and the experiment are comparable . In conclusion , compensatory network rewiring as observed after lesions may be accounted for by local activity-dependent changes in individual neurons and the subsequent emerging cooperative effects between neurons inside and outside the LPZ . Many studies have shown that a loss of input due to a focal retinal lesion causes a remapping of retinotopic representations in the primary visual cortex [71]–[80] . Traditionally , cortical remapping is thought to be caused by changes in synaptic strengths , for instance due to STDP [48] . Recently it has been shown that , in the primary visual cortex , lesion-induced functional remapping coincides with massive structural changes in dendritic spines [10] and axonal branches and boutons [13] . To investigate whether retinotopic remapping could be accounted for by our structural plasticity model , we tested spatial input representations in the model at different time points before and after the lesion ( Fig . 8a ) . This was done in line with experimental procedures to assess retinotopic mappings [10] , as follows: The plane of neurons was subdivided into 6×6 non-overlapping rectangular areas . Subsequentially , each area received a noisy test current with mean and standard deviation for . During the stimulation , we measured the number of action potentials in all neurons of the network in order to determine which areas of the network responded to focussed input . At the time of stimulation , we froze the connectivity in the network so that the test stimulus did not change the retinotopic mapping , to test how focussed input spreads throughout the network via horizontal connections . Afterwards , we color-coded each neuron according to the input area it strongest responded to . Immediately after the loss of input ( no test inputs were given to the neurons in the LPZ ) , all neurons inside the LPZ are irresponsive to any inputs given to any neurons outside the LPZ . In the early phase after the lesion , neurons at the edge of the LPZ start to fire in response to input from adjacent areas in the peri-LPZ . In the middle phase , bordering representations begin to enlarge from the peri-LPZ into the LPZ and thereby ‘fill’ the LPZ from the border to the center . In the late phase , even neurons in the center of the LPZ become responsive again . The course of remapping from the border to the center coincides with the formation of additional horizontal connections ( Fig . 9 ) , in the early phase from the peri-LPZ to the border of the LPZ , and in the middle and late phases from the border to the center . In the late phase , representations of the peri-LPZ have almost completely ‘filled’ the entire LPZ . Only in the very center of the LPZ did some neurons remain irresponsive , depending on the size of the lesion and the width of the kernel function ( Eq . 9 ) . Importantly , in our model the remapping arises purely from structural changes , such as synapse formation and reorganization of ‘horizontal’ intra-cortical connections . We did not impose any form of synaptic plasticity such as LTP or STDP . Therefore , remapping in the model is not the result of a goal-directed associative process but emerges from the interactions between local activity-dependent changes in spine and bouton numbers . The course of remapping resembles the remapping in mice visual cortex ( Fig . 8b ) . So far we have shown that functional network reorganization similar to that observed experimentally can arise from local activity-dependent rules for axonal and dendritic element formation . Now we contrast these growth rules with different growth rules leading to aberrant network reorganization . Here we discuss two other cases with fundamentally different time courses of network reorganization . In the first aberrant case , the set-point is relatively high compared to while is rather low . In this scenario , network repair is impossible because neuronal activity in the LPZ does not exceed and therefore neurons reduce their dendritic elements instead of forming additional elements in a compensatory manner . Consequently , axonal elements from the peri-LPZ are not be able to find dendritic targets in the LPZ , so no new synapse are formed and neuronal activity cannot increase ( Fig . 10 ) . Thus , not only the high set-point [35] , [81] but also the minimum activities and for axonal and dendritic element formation , respectively , are crucial for whether or not neurons can restore their activity level after the lesion . In the second aberrant case , are equally low . In this scenario , neurons that are deprived of external input literally pull themselves up by their own bootstraps to restore their activity . Because the neurons simultaneously generate dendritic and axonal elements , massive recurrent connections are formed within the LPZ ( Fig . 11 ) . Recurrent connections enable the neurons to amplify fluctuations in activity and raise their average activity to the high set-point . Interestingly , the activity state of the network changes from irregular to more synchronized firing ( Fig . 12 ) , with clear oscillations in activity , when massive recurrent connections form . The time course of activity recovery differs fundamentally from the time course of normal network repair ( cf . Fig . 3 ) , in that neurons do not sequentially return to the high set-point . Neurons with the same initial activity level recover simultaneously . The difference between the different scenarios becomes especially clear for LPZs that cover almost the entire network . Networks with a high and a low ( physiological case ) do not allow for network repair when ( almost ) the complete external input is removed ( Fig . 13a ) but only show some restricted reorganization in the border region of the LPZ . By contrast , networks with equally low and do recover well ( Fig . 13b ) . In fact , LPZs of any size will recover as long as the remaining activity in the LPZ after the lesion exceeds and . Large LPZs most clearly show the simultaneous recovery of neurons . Furthermore , due to the formation of massive recurrent connections , neurons frequently become hyperactive after network rewiring . This structural dynamics is comparable to that described in [35] , [81] . As shown in Fig . 9 , functional reorganization in terms of the remapping of spatial input representations is the result of a sequential formation of synapses between axonal elements in the peri-LPZ and dendritic elements in the border of the LPZ , and thereafter between axonal and dendritic elements in the border and the center , respectively . This can be interpreted as a sequential ingrowth of axons from the peri-LPZ into the border and from the border into the center . Remapping therefore emerges from cooperative effects between the peri-LPZ and the LPZ , in the sense that new synaptic connections transmit additional activity into the LPZ . New synaptic connections help neurons in the LPZ to restore their activity to the high set-point , thereby enlarging the input representations from the peri-LPZ into the LPZ—the remapping observed . Both scenarios with aberrant network reorganization discussed here do not show this ‘ingrowth’ of new synaptic connections into the LPZ . For the case with , this is due to the lack of newly formed dendritic elements in the LPZ . As expected , functional reorganization in terms of remapping does not occur ( Fig . 14a ) . Interestingly , remapping is also absent in the case with ( Fig . 14b ) , in spite of robust network repair in the sense that all neurons returned to the high set-point . Network repair is in this case caused by recurrent synaptic connections within the LPZ and hardly by newly formed connections from the peri-LPZ into the LPZ . The lack of remapping in this scenario supports the notion that ingrowing connections are crucial for remapping . It further demonstrates that functional reorganization depends on growth rules with the right proportion of activity-dependent increase and decrease of axonal and dendritic elements . Thus , the values of , and are crucial in determining the principally different types of network reorganization . Network parameters such as neuron numbers or neuron densities or neuronal parameters for electrical activity and calcium dynamics are not critical , and the same results can be obtained with other choices of these parameters ( cf . Fig . S1 ) . We assumed biologically plausible growth rules inspired by what is known from in vitro studies about how neurons reshape their morphology in response to changes in electrical activity . The precise morphological response of neurons in the visual cortex to changes in activity has never been determined . Our modeling study revealed that growth rules adopted from in vitro studies can produce structural and functional network dynamics that show striking similarities to experimentally observed changes in the primary visual cortex after focal retinal lesion . Thus , we predict that activity-dependent changes in dendritic spines and axonal boutons of visual cortex neurons are governed by growth rules that purely depend on the cell's own level of activity . These growth rules can be described by simple Gaussian-shaped curves , and state that neurons have a narrow range of activity for which axonal and dendritic outgrowth , or at least axonal bouton and dendritic spine formation , is optimal , and that very high and very low activity cause retraction . It has never been considered in the context of retinotopic remapping that neurons may retract horizontal connections when their activity is too low . Furthermore , to obtain functional remapping , axonal and dendritic growth rules should not be identical but dendrites must start growing at lower activity levels than axons ( Table 1 ) . As a further consequence of the growth rules , we predict that the presence of different activity levels inside and outside the LPZ is what drives the rewiring of the network after the lesion . To support this prediction , we lowered the external input to the peri-LPZ in the model by either 20% or by a complete blockade . We could suppress compensatory network reorganization significantly by blocking external input to the peri-LPZ ( Fig . 15 , right column ) . This indicates that the recovery of the LPZ indeed crucially depends on the higher activity level in the peri-LPZ than in the LPZ . Interestingly , a slight reduction in the input to the peri-LPZ by 20% ( Fig . 15 , middle column ) had a promoting effect on axonal element formation ( for data points later than post-lesion day 7; Mann-Whitney U test ) . This is because the axonal growth rule reaches a maximum for activities slightly below the set-point . Since the ‘filling’ of the LPZ with adjacent representations depends on the size of the lesion relative to the width of the kernel function ( Eq . 9 ) , we further predict that larger LPZs may functionally recover at the rim but will not completely reorganize . Since all elements of the growth rules , such as electrical activity , intracellular calcium and dendritic and axonal morphology , can be measured experimentally , our predictions are amenable to experimental testing . Taken together , our work points to the need to quantify the activity-dependent growth of neurons in the visual and other cortices in order to understand better how alterations in sensory input cause the cortex to rewire its circuitry .
By means of our structural plasticity model , we have shown that specific , functional network reorganization can emerge from local homeostatic growth rules . The dynamic changes in axonal and dendritic elements produced by these growth rules were comparable to the dynamics of dendritic spine remodelling and axonal sprouting observed in the primary visual cortex after focal retinal lesions in mice [10] and in monkeys [13] , respectively . We therefore conclude that neurons in V1 may have a narrow range of electrical activity in which axonal and dendritic growth is optimal . Importantly , axons may need higher activitity to grow out than dendrites . Initial activity differences , with higher activities in the peri-LPZ and lower activities in the border region of the LPZ , may drive axonal outgrowth in the peri-LPZ and dendritic spinogenesis in the border of the LPZ , respectively . New synaptic connections from the peri-LPZ to the border of the LPZ restore the activity from the outside to inside of the LPZ . Indeed , experimental studies have shown that a circular wave front of electrical activity travels slowly from the edge of the LPZ inwards [80] . Our model indicated that if the area of deprived input is not too large , cortical reorganization is capable of restoring electrical activity of most neurons in the LPZ . Interestingly , quantitative changes in the long-term neuronal activity markers c-fos and zif268 indeed suggest that neurons inside the LPZ strive towards homeostasis in electrical activity [82] , [83] . Therefore , we conclude that homeostatic regulation may serve as a fundamental principle in reshaping cortical circuitry after focal loss of input . The growth rules postulated in our model try to capture the essential aspects of activity-dependent axonal and dendritic changes reported in experimental studies . Dendritic spines need a certain level of electrical activity to persist [28] , [30] , which corresponds to our model assumption of a minimum activity level required to maintain dendritic elements ( the first set-point ) . Incidentally , the minimum required level of activity measured experimentally is unexpectedly low [28] , in line with our model result that needs to be lower than for functional reorganization . Reducing electrical activity , but not below the minimal required level [28] , promotes the formation of spines [66] , [84] , [85] , enabling neurons to explore the neuropil for axonal contacts [86] . Our model assumption that lowering electrical activity increases the neuron's number of dendritic elements is thus supported by experimental findings . Finally , increasing electrical activities attenuates dendritic growth , while overstimulation may lead to dendritic pruning [65] . This is reflected by the model assumption that too high activity is not beneficial for the formation of dendritic elements . Axonal growth , too , is modulated by electrical activity , and the way axons respond to activity is comparable to how dendrites respond . Blockade of activity can promote axonal sprouting and synapse formation , whereas increasing the activity level can suppress axonal outgrowth [17] . The precise levels of , and the homeostatic set-point remain to be quantified experimentally . The model showed that functional reorganization can arise from local structural plasticity alone and does not require associative forms of plasticity such as LTP or STDP . Sprouting of axonal elements from the peri-LPZ into the LPZ gives rise to an enlargement of representations adjacent to the LPZ and consequently to a remapping of spatial input representations in the network . Although the number of neurons in the network is relatively low compared to the number of neurons in the primary visual cortex , we believe that the essential dynamics in the model , with network repair and retinotopic remapping following focal retinal lesions [71]–[80] , is not dependent on neuron number . We simulated a network with a higher neuron density ( neurons in a network of the same physical size as before with neurons ) and obtained the same structural and functional network reorganization ( Fig . S1 ) . Although axonal sprouting was found following focal retinal lesions [12] , [13] , [70] , it has as yet not been possible , because of methodical limitations , to show that new axons establish synapses with deafferented neurons . On the basis of our modeling results , we postulate that the formation of additional horizontal connections from the peri-LPZ into the LPZ may account for the retinotopic remapping as seen , for example , in mice [10] . Since axonal growth in the mature cortex is restricted to a few hundreds of micrometers [54] , this would further explain why the compensatory capacity of cortical networks is limited and larger LPZs do not recover [87] , [88] . Nevertheless , as our structural plasticity model predicts , functional recovery at the rim of the LPZ usually takes place even with larger lesions [80] . Our model stands in contrast to the Kohonen model for self-organizing maps [89] , [90] , traditionally applied to explain cortical remapping . In the Kohonen model , cortical neurons receive input from all cells in the retinal input layer and merely change their response properties towards intact retinal input . However , there is strong experimental support that horizontal connections within the cortex in fact are the source of functional reorganization [91] , [92] . Moreover , our model also better resembles the experimentally observed time course of the ‘filling’ of the LPZ with adjacent representations from the border to the center [80] , [83] . By contrast , the Kohonen model would predict a simultaneous change in the response properties of all neurons in the entire LPZ ( cf . Table 1 ) . Our conclusion that structural plasticity in the cortex after retinal lesions [9] , [10] , [12] , [13] , [70] may contribute to cortical remapping complements the current notion that retinotopic remapping could arise from synaptic plasticity [93] , in particular from spike timing-dependent plasticity [44] , [48] . Our modeling results are the first to indicate that cortical remapping may not necessarily require associative plasticity but can already emerge from structural reorganization of synapses ( structural plasticity ) alone . In the living brain , cortical remapping is most likely the result of both structural and synaptic plasticity , but the power of a modelling approach is that the impact of both types of plasticity can be explored separately . In our future models , we intend to include both types of plasticity and investigate whether synaptic plasticity can contribute to the experience-dependent ( fine ) tuning of structurally formed cell assemblies [9] , [94] and the experience-dependent formation of memory traces therein . An important question is how inhibitory neurons contribute to network repair . A recent study [95] has shown that inhibitory neurons respond to a loss of input with a rapid decrease in structural connectivity that is strongest in the center of the LPZ and fades with distance to the LPZ . It needs to be tested how the dramatic loss of inhibition influences the activity in the network . A straightforward interpretation is that disinhibition of neurons contributes to the difference in activity between the LPZ and peri-LPZ . Waves of spreading activity are a common phenomenon after brain lesions , for example after focal stroke [96] . Modeling work has revealed complex interactions between activity and structure , with various forms of oscillatory activity , when inhibition was substantially altered [81] . Previous work on cortical reorganization after loss of visual input has raised the question of whether synapse formation is specific or unspecific [9] . The observation that synapse formation takes place particularly at the edge between areas with intact and deprived input led to the notion that cortical organization may be highly specific [9] , [80] . Associative forms of synaptic plasticity such as STDP [48] may serve as such specific mechanisms . However , these mechanisms require that neurons keep track of their current firing and , in case of STDP , even of the firing history of other neurons . Here we propose an unspecific mechanism that can give rise to specific cortical reorganization . We dub this mechanism neuron-centric plasticity , since neurons change their morphology purely on the basis of their own level of electrical activity , i . e . form synaptic elements and offer them to the neuropil [49] , [50] . Remarkably , without knowing the activity levels of other neurons and randomly recombining their vacant synaptic elements , neurons can achieve homeostasis in electrical activity ( at both the cellular and the network level ) and even cause functional cortical remapping . This notion stands in sharp contrast to associative , synapse-centric forms of plasticity such as STDP , in which the timing relations between pre- and postsynaptic activity determine the strengthening or weakening of synapses , and which are traditionally regarded as the standard paradigm for cortical reorganization . Other forms of neuron-centric plasticity , such as synaptic scaling , also contribute to homeostasis in electrical activity [97] . Neurons in the cortex [98] , hippocampus [99] and spinal cord [100] can up- or down-regulate all their incoming synapses when their firing rates become too low or too high , respectively . The main difference with network reorganization by activity-dependent structural plasticity is that synaptic scaling has an immediate effect on postsynaptic firing rates , whereas activity-dependent structural plasticity requires the availability of complementary synaptic elements for synapse formation . Synaptic scaling and structural plasticity operate on different time scales , but may also cooperate . Neurons may form synapses de novo to explore new sources of activity when synaptic scaling has become insufficient to maintain activity homeostasis , for example because presynaptic neurons are also inactive . We propose that activity-dependent structural plasticity may provide a unifying framework for understanding cortical reorganization in a wide range of situations including network repair in degenerative diseases or following focal stroke . | The adult brain is less hard-wired than traditionally thought . About ten percent of synapses in the mature visual cortex is continually replaced by new ones ( structural plasticity ) . This percentage greatly increases after lasting changes in visual input . Due to the topographically organized nerve connections from the retina in the eye to the primary visual cortex in the brain , a small circumscribed lesion in the retina leads to a defined area in the cortex that is deprived of input . Recent experimental studies have revealed that axonal sprouting and dendritic spine turnover are massively increased in and around the cortical area that is deprived of input . However , the driving forces for this structural plasticity remain unclear . Using a novel computational model , we examine whether the need for activity homeostasis of individual neurons may drive cortical reorganization after lasting changes in input activity . We show that homeostatic growth rules indeed give rise to structural and functional reorganization of neuronal networks similar to the cortical reorganization observed experimentally . Understanding the principles of structural plasticity may eventually lead to novel treatment strategies for stimulating functional reorganization after brain damage and neurodegeneration . | [
"Abstract",
"Introduction",
"Model",
"Results",
"Discussion"
] | [] | 2013 | A Simple Rule for Dendritic Spine and Axonal Bouton Formation Can Account for Cortical Reorganization after Focal Retinal Lesions |
Sequence-specific binding by the human p53 master regulator is critical to its tumor suppressor activity in response to environmental stresses . p53 binds as a tetramer to two decameric half-sites separated by 0–13 nucleotides ( nt ) , originally defined by the consensus RRRCWWGYYY ( n = 0–13 ) RRRCWWGYYY . To better understand the role of sequence , organization , and level of p53 on transactivation at target response elements ( REs ) by wild type ( WT ) and mutant p53 , we deconstructed the functional p53 canonical consensus sequence using budding yeast and human cell systems . Contrary to early reports on binding in vitro , small increases in distance between decamer half-sites greatly reduces p53 transactivation , as demonstrated for the natural TIGER RE . This was confirmed with human cell extracts using a newly developed , semi–in vitro microsphere binding assay . These results contrast with the synergistic increase in transactivation from a pair of weak , full-site REs in the MDM2 promoter that are separated by an evolutionary conserved 17 bp spacer . Surprisingly , there can be substantial transactivation at noncanonical ½- ( a single decamer ) and ¾-sites , some of which were originally classified as biologically relevant canonical consensus sequences including PIDD and Apaf-1 . p53 family members p63 and p73 yielded similar results . Efficient transactivation from noncanonical elements requires tetrameric p53 , and the presence of the carboxy terminal , non-specific DNA binding domain enhanced transactivation from noncanonical sequences . Our findings demonstrate that RE sequence , organization , and level of p53 can strongly impact p53-mediated transactivation , thereby changing the view of what constitutes a functional p53 target . Importantly , inclusion of ½- and ¾-site REs greatly expands the p53 master regulatory network .
The tumor suppressor p53 ( OMIM no . 191170 ) is a sequence-specific master regulatory gene that controls an extensive transcriptional network providing for genome integrity in response to cellular and environmental stresses or damage [1]–[3] . p53 differentially regulates the expression of target genes , as well as microRNAs associated with cell cycle control , apoptosis , DNA repair , angiogenesis , senescence and carbon metabolism [4] , [5] . A variety of factors such as stress and cell-type dependent , post-translation modifications and transcriptional co-factors can influence p53-induced transcriptional changes [6]–[8] . Paramount to p53 transcriptional function is the direct interaction between p53 and its targeted DNA sequence . The nature of this interaction could per se determine transactivation capacity , as well as influence p53-mediated biological processes [9] . Such activities are often altered during human cancer development as highlighted by the frequent appearance of p53 missense mutations in its sequence-specific DNA binding domain [10] , [11] which can abrogate or alter p53 transactivational activity that result in changes in biological responses , such as the balance between apoptosis and survival in response to DNA damage . [3] , [10] , [12]–[15] . Through in vitro studies , a consensus p53 DNA binding sequence has been derived comprising a motif of two decamers ( half-sites ) RRRCWWGYYY ( n ) RRRCWWGYYY , ( where R = purine , W = A or T , Y = pyrimidine and n is 0–13 bases ) where each decamer is composed of two adjacent p53 monomer binding sites ( quarter-sites ) in inverted orientation [16]–[18] . p53 binds cooperatively to the consensus RE as a dimer of dimers , where a tetramer is the accepted functional unit required for full transcriptional activity [19]–[26] . Most functional response elements ( REs ) identified in association with p53 target genes depart from this consensus , where base changes are tolerated at each position with the exception of the C and G at positions 4 and 7 in each half-site [27] , [28] . Recent crystal structures show that the interactions between p53 and DNA are influenced by the base pairs present in the RE sequence [29] while binding assays in solution established a wide range of dissociation constants among natural p53 REs [23] . Functional studies in model systems demonstrate that the RE sequence and amount of p53 expression can dramatically influence the level to which p53 can transactivate from a specific RE [3] , [30] . However , results of those various approaches do not fully overlap and it is still not well-understood how the loose consensus , in terms of base variation and spacer length between half-sites affects p53 binding to DNA , or how differential binding relates to transactivation specificity [31]–[33] . More recently , several studies have refined the accepted , or canonical p53 consensus binding sequence based on presumed unbiased chromosome- or genome-wide in vivo binding assays using chromatin immunoprecipitation ( ChIP ) [34]–[37] . However , the findings may be skewed towards stronger p53 interacting sequences or influenced by the agent used to induce p53 protein , as well as the human cell lines examined . While ChIP provided a powerful tool for identifying sequences that p53 could bind and the results pointed to a refined p53 consensus with more restrictive sequence features [34] , these assays could not address the strength of p53 binding . Furthermore , the experimental approach failed to retrieve known p53 binding sites , indicating that the technique could miss interactions between p53 and weaker response elements which might include noncanonical sites that do not match the accepted consensus motif . Bioinformatics studies to identify p53 REs are guided by identified sequences . Based on the consensus sequence , computational algorithms have exploited the base pair composition of known p53 REs to determine binding probabilities of p53 towards sequences and generate position-weight matrices ( PWM ) or “logos” [38] , [39] . While such algorithms may guide the identification of p53 binding sites in the genome , they cannot assess the strength of p53 binding to these sequences . The ability of p53 to bind ( and presumably transactivate from ) a sequence is often related to compliance with the canonical consensus sequence where each position in the motif is assumed to be equivalent and mutually exclusive in terms of affecting p53 binding [40] . However , PWM values appear not to be a good predictor of p53 transactivation [30] . While the canonical consensus sequence has guided studies of the p53 network , an additional layer of complexity was identified recently suggesting that the canonical motif is not limited to two decamers in human cells . p53 dependent-transactivation was detected in association with a decamer half-site created by a single nucleotide polymorphism ( SNP ) in the promoter of the FLT-1 gene ( the vascular endothelial growth factor ( VEGF ) receptor-1 gene ) [41] . These results showing that p53 can function from a noncanonical consensus sequence imply that the number of potential p53 target sites within the genome may be much greater than anticipated . Current methods that identify and/or define putative p53 REs would overlook such noncanonical binding elements . Stimulated by the observation of a transcriptionally active p53 half-site , as well as by the finding that even a single nucleotide change can greatly impact transactivation potential of a RE [42] , [43] , we have systematically deconstructed the canonical p53 RE sequence to address the requirements for a functional p53 binding site . Utilizing in vivo systems we developed in the budding yeast Saccharomyces cerevisiae , the transactivation potential of p53 was first examined from p53 canonical consensus REs containing base changes and/or variations in organization of binding motifs ( i . e . , multiple REs and various spacer lengths ) and then assessed transactivation from ½- and ¾-site REs , which we refer to as noncanonical REs . For these noncanonical REs the surrounding sequence does not resemble a p53 binding sequence and the motif itself would not be recognized within standard descriptions of a p53 canonical consensus sequence . Within the yeast system , we have addressed not only the ability of p53 to function from specific target sequences ( i . e . , on/off ) , but also the extent of transactivation from these sites at variable levels of expression . These studies have been extended to transactivation capacity and p53 promoter occupancy in a human cell system , and the in vivo functionality evaluations were compared with results obtained in a recently developed semi-in vitro binding assay using human cell nuclear extracts . Several structural mutants were examined to assess the p53 structural requirements for transactivation from noncanonical REs . Overall , the findings expand our understanding , as well as the anticipated size of the human p53 master regulatory network .
Previously , we had developed a plasmid-based haploid yeast system to systematically evaluate the contribution of RE sequence and p53 expression level towards p53 differential transactivation [14] , [30] . While many factors may ultimately determine p53 transactivation of individual genes in human cells including stress stimuli , post-translational modifications , and transcriptional cofactors , the yeast system addresses the potential for wild type and mutant p53 to bind and transactivate from various REs derived from human genes when placed in a constant chromatin environment . We have expanded this experimental system to diploid yeast to further assess the transactivation capacities of p53 ( WT or mutant ) ( Figure 1A ) . Two panels of modified S . cerevisiae strains were generated . The first was a set of p53 host strains in which p53 ( WT or mutant ) is directed by a “rheostatable” GAL1 promoter that allows for controlled , over 200-fold , inducible expression of p53 in yeast depending on the carbon source in the media ( Figure 1B and Figure S1 ) . Importantly , similar to p53 protein accumulation in mammalian cells following stress [44] , expression from the GAL1 promoter in yeast displays a graded transcriptional response such that there is a range of activity from the promoter as opposed to a binary , or on/off response [45]–[47] . Biggar and Crabtree [45] demonstrated through fluorescence-activated cell sorting ( FACS ) experiments that expression of green fluorescent protein from a GAL1-GFP reporter within a population of cells generated a single fluorescent peak where the intensity of the peak was dependent upon the concentration of galactose supplemented in the media . Thus , increases in galactose will result in a homogeneous response where the vast majority of cells in the population respond ( within our system expressing an induced amount of p53 ) rather than merely increasing the percentage of cells within the population expressing the maximal level of protein . The second set of strains of opposite mating type contained promoter REs upstream of the minimal CYC1 promoter and either the ADE2 color reporter or the firefly luciferase reporter [28] . To facilitate the construction of a large number of p53 mutants and REs at chromosomally located target loci we employed the delitto perfetto system for in vivo mutagenesis . Delitto perfetto utilizes oligonucleotides and targeted homologous recombination to rapidly generate S . cerevisiae yeast strains with specific genetic alterations [48] , [49] . Mating of the reporter and p53 host strains results in isogenic , diploid yeast that enable the rapid assessment of the transactivation potential for WT or mutant p53 proteins towards many individual REs in the p53 transcriptional network [14] . Importantly , all the conditions in the cells are constant , i . e . , isogenomic , where the only variables between strains are the RE sequence , WT or mutant p53 and level of expression . The luciferase reporter provided a quantitative estimate of transactivation capacity of WT p53 from the various REs in cultures of logarithmically growing cells . The strength of transactivation ( relative light units/µg protein ) from each RE were compared to transactivation from the strongly transactivated p21-5′ RE at high p53 expression ( 0 . 024% galactose ) . An example of transactivation capacity of p53 over a range of expression is shown by the quantitative assessment of p53-induced transactivation from the p21-5′ RE in vivo ( Figure 1C ) . Functional assessment with the luciferase reporter assay provided an indication of the kinetics of transactivation with a basal level of transactivation at 0 . 00% galactose ( 2% raffinose ) , initial induction of transactivation ( between 0 . 004–0 . 008% galactose ) and maximal level of transactivation ( between 0 . 016–0 . 024% galactose ) . A common feature of many p53 target genes , including p21 , PUMA , BAX , and DDB2 ( p48 ) , is the presence of multiple p53 binding sites [50]–[56] . The distance between binding sites is variable and can even be overlapping . The Mdm2 gene , which targets p53 for degradation by the ubiquitination pathway through a negative feedback loop , is an example of a p53 target gene that contains two promoters . The upstream promoter is constitutively active and does not rely on p53 for transactivation , whereas the second is in the first intron and is p53-dependent [57] , [58] . The second promoter contains two full-site p53 REs separated by 17 nucleotides ( nt ) . Using our diploid yeast-based p53 rheostatable system , we investigated the interaction between these two REs and the impact of this spacer by changing the distance between the murine MDM2 REs . The induction from the individual MDM2 REs at high p53 expression was much weaker than observed with the strong p21-5′ RE ( Figure 2A ) : ∼33% ( RE1 ) and ∼18% ( RE2 ) relative to p21-5′ RE . However , transactivation from the natural MDM2 , containing the 17 base spacer , was much higher than the sum of the individual REs , reaching the p21-5′ RE levels . The synergy was apparent at both moderate and high p53 levels . To determine the impact that a spacer may play for full-site REs , the distance was reduced to either 10 or 5 nt . As shown in Figure 2B , decreasing the separation to 10 nt had little effect . However , a decrease to 5 nt resulted in a substantial reduction in transactivation suggesting that synergistic interactions are lost as full REs become closely spaced , although transactivation from the two sites remain additive . While weak full-site REs can interact synergistically when separated by 17 bases , previous studies have indicated that spacers between decamer half-sites of the canonical consensus RE can alter the ability of p53 to transactivate from a RE [18] , [59] , [60] . We systematically investigated the role that spacers may play on p53 transactivation using the yeast-based rheostatable promoter and a system based on expression following transfection into human cells . Addition of a one-base spacer between the p21-5′ half-sites resulted in a dramatic 60% decrease in p53 transactivation ( Figure 3A ) in yeast . Addition of a second nucleotide further decreased p53 transactivation to approximately 25% of transactivation at high p53 expression . Importantly , at lower p53 levels transactivation was essentially abolished , demonstrating that the impact of spacer is markedly affected by p53 expression level . Further increases in spacer resulted in decreased transactivation and at 5 nt there was almost no detectable luciferase activity at high p53 expression levels . Similar findings were observed in a plasmid-based haploid yeast system when p53 transactivation was measured from p21-5′ REs containing a spacer of 0 , 2 , 5 , or 10 nt ( Figure S2A ) [3] , [28] , [30] . Thus , the length of spacer sequence between decamer half-sites combined with level of p53 , greatly influences the ability of p53 to transactivate from the p21-5′ sequence . Interestingly , transactivation by the p53 family members p63β and p73β was also compromised when p21-5′ RE contained spacers ( Figure S2B ) ; a 2 nt spacer essentially abolished transactivation . To determine the effect of spacers upon weak target REs in the p53 transcriptional network , transactivation was assessed from the RE of the human apopotosis gene Noxa [61] with and without a 5 nt spacer between the decamer half-sites . In comparison to the p21-5′ RE , transactivation from the Noxa RE was ∼30% of the levels of p21-5′ at high p53 expression ( Figure 3A ) . This was comparable to p53 transactivation from the p21-5′ RE containing a spacer of 1 or 2 bases . A spacer of 5 bases between the Noxa half-sites abolished p53 responsiveness ( also found with the haploid yeast system; Figure S2A ) . Finally , we wanted to establish the impact that naturally occurring spacers in REs might have on transactivation . TIGAR ( TP53 induced glycolysis and apoptosis regulator ) contains a p53 target RE with a two-base spacer between decamers . This gene reduces levels of glycolysis , decreases free reactive oxygen species and attenuates the apoptotic response [62]–[64] . Interestingly , the natural TIGAR RE is one of the few examples in the genome where the binding element matches the canonical p53 consensus sequence precisely ( i . e . , no mismatches ) . However , p53 could only induce transactivation from this sequence to ∼20% of the levels with the p21-5′ RE , which does not precisely match the consensus p53 sequence ( Figure 3A ) . When the spacer was removed , p53 could transactivate from TIGAR to levels comparable to the p21-5′ RE . Similar results were obtained with the haploid yeast system ( Figure S2A ) . The impact of spacers on the ability of p53 to function from a RE was also assessed in human cells under conditions of high p53 expression . Luciferase reporter vectors containing the p21-5′ RE with spacers of increasing length were generated in the vector pGL3 promoter ( pGL3-P ) . The ability to transactivate from transfected p21-5′ REs was assessed in p53 null SaOS2 cells ( derived from a human osteosarcoma line ) that were transfected with a cytomegalovirus ( CMV ) based p53 expression plasmid [65] . Consistent with the results observed in yeast , p53-dependent transactivation decreased with increasing spacer length between decamer half-sites . As shown in Figure 3B , expression of WT p53 resulted in an ∼35-fold induction of transcription from the natural p21-5′ RE as compared to transfection with a p53 deficient plasmid , whereas transactivation from the p21-5′ RE containing a one- or two-base spacer resulted in a 45% and 67% reduction in relative luciferase activity , respectively . Similar to the situation in yeast , additional nucleotides resulted in >90% net reduction in transactivation . It is interesting that within the three systems–haploid and diploid yeast and human cells–transactivation from a RE with a spacer of 10 bases was slightly increased in comparison to a RE with a 5 base spacer . To determine if the difference in p53-dependent transactivation from REs containing spacers is simply due to a reduction in p53 promoter binding , we investigated in vivo occupancy using chromatin immunoprecipitation ( ChIP ) assays and in vitro binding using a newly developed microsphere binding assay ( Noureddine et al . , submitted ) . ChIP assays were performed on the luciferase reporter plasmids containing the p21-5′ RE with spacers of varying lengths which had been transfected into SaOS2 cells along with the p53-expressing plasmid . As shown in Figure 3C , a one nt spacer decreased occupancy at the p21-5′ RE by two-fold . Further increases in spacer length reduced p53 occupancy . No occupancy was observed in mock-transfected cells . Thus , the pattern of occupancy by p53 mirrored that for transactivation . Consistent with the transactivation results , p53 occupancy at the p21-5′ RE with a spacer of 10 nucleotides was slightly increased in comparison with a spacer of 5 nucleotides ( 0 . 6% vs . 0 . 4% ) . To further characterize the impact of spacer on p53 interactions with a RE , we utilized a fluorescent microsphere binding assay to evaluate sequence-specific p53-DNA binding interactions ( Noureddine et al . , submitted ) . Briefly , this assay addresses p53 binding to individual beads with specific RE test sequences , where each bead “type” ( i . e . , beads with specific REs ) is identified with a unique double stranded oligonucleotide 24-nt “tag” sequence . Several bead types are then multiplexed in a binding assay and analyzed using Luminex technology to determine the amount of p53 bound to each bead type . We generated a series of oligos with the p21-5′ RE sequence of interest that contained various spacer lengths flanked by non-specific DNA . Each of these RE sequences was conjugated to beads . The bead types were combined and incubated for 60 minutes with nuclear cell extracts obtained from human lymphoblastoid cell lines that were either not induced or induced for p53 expression with doxorubicin ( Dox ) . The level of p53 expression was approximately 15-fold greater in extracts from Dox-treated versus nontreated cells ( data not shown ) . The p53 interaction with each bead type ( i . e . , each p21-5′ RE variant ) was determined after incubation with p53 antibodies and secondary antibodies conjugated with phycoerythrin . As displayed in Figure 3D , the mean relative binding of p53 to the p21-5′ RE with a spacer of 1 nucleotide ( 0 . 83±0 . 09 ) was comparable to that for the p21-5′ RE ( 1 . 0±0 . 11 ) with no spacer . However , p53 binding to the p21-5′ REs was affected when the spacer between half-sites was ≥2 nucleotides , with a dramatic decrease at >4 nucleotides . The sequence of the spacer used to increase the distance between the decamer half-sites had no apparent effect on binding activity ( data not shown ) . Furthermore , p53 did not bind to the negative control ( NC ) sequence which had the p21-5′ RE replaced with a scrambled sequence . Interestingly , p53 displayed the same residual binding to the individual half-sites of the p21-5′ RE [p21-5′ left ( L ) and right ( R ) ] as it did towards REs containing a spacer of 4 nucleotides or more ( 0 . 06 and 0 . 08 , respectively ) . Since a low level of binding or transactivation was observed with widely separated decamers , we investigated p53 binding and transactivation from single decamer sequences in the yeast diploid and human cell systems . Transactivation was barely detectable from the left or right decamers of the p21-5′ RE ( Figure 4A ) . However , two additional complete consensus half-sites , designated Con G and Con D , were able to support transactivation at a level corresponding to 2 . 4% and ∼10% , respectively , of that from the full p21-5′ RE . Importantly , transactivation was only observed at high levels of p53 expression . By way of comparison , transactivation from the con D half-site was comparable to levels obtained from the low-responding 14-3-3σ RE . To assess the ability of p53 to transactivate from decamer half-sites in a mammalian system , luciferase reporter vectors containing the p21-5′ RE half-sites were transiently transfected with or without a vector containing p53 under the CMV promoter into p53 null SaOS2 cells [65] . As shown in Figure 4B ( also see Figure 3B ) , WT p53 induced transactivation from the full p21-5′ RE was 35-fold greater than with an empty pGL3-P vector . However , there was clear induction from the left and right half-sites , 4- and 6-fold , respectively . These results correlated well with the relative p53 occupancy assessed by ChIP: 4% for the full RE , 0 . 4% for the left decamer and 0 . 9% for the right decamer and agree with the previous findings in Menendez et al . [66] that p53 can function from a half-site RE . Based on these findings , the p53 transcriptional network may incorporate more downstream targets than previously predicted . The sequence requirements for p53-dependent transactivation were examined further , utilizing a plasmid-based haploid yeast system with rheostatable p53 expression ( previously used to assess transactivation capacity of canonical full-site REs; see Materials and Methods and [30] ) . Similar to results from mammalian cells , p53 was able to weakly transactivate from p21-5′ half-site REs at a level that was ∼1% of the levels of transactivation from the p21-5′ full-site RE ( Figure 4C ) . This result differed somewhat from that of the diploid yeast system where transactivation from the p21-5′ half-site REs was <0 . 5% of the transactivation from the full-length p21-5′ RE ( Figure 4A ) . Transactivation from Con D was greater than p53 transactivation from either of the p21-5′ half-sites and Con D was greater than Con G in both systems . We also examined the impact of changes in the WW of the core CWWG ( W = A or T ) . The 3 bases on either side consisted of GGG and TCC which had been shown to enhance binding at full-site REs [33] ( Figure 4C ) . Decamers required the central CATG motif at the junction of the monomer binding sites for modest levels of transactivation , ∼7% of the levels from the p21-5′ RE ( Figure 4C ) . Altering the motif from CATG to CAAG or CTTG decreased transactivation another 2- to 3-fold whereas changing this motif to CTAG nearly eliminated transactivation . Transactivation was also assessed with various combinations of sequences surrounding the CATG core domain . As shown in Figure 4C , transactivation with the GGG/CCC flanking sequence was comparable to the GGG/TCC sequence tested with the CWWG motifs . However , transactivation was reduced when other alterations were made . For example , a change in flanking sequence to GGG/CTC or GGA/CTC resulted in almost no transactivation . We also addressed the ability of p53 to transactivate from ¾-site REs . A consensus ( Con ) binding site was created for each of the two possible configurations of a ¾-site RE . The first , designated Con J , consists of a ¼-site directly adjacent to a ½-site , whereas the second , designated Con K , contains a 5 base spacer between the ¼-site and the ½-site . Transactivation in the diploid yeast system from the Con J and Con K ¾-site REs was 25% and 18% , respectively , of the full p21-5′ RE level ( Figure 5A ) , at high p53 expression ( 0 . 024% galactose ) . This was substantially more than observed for half-site REs . Similar findings were observed in the haploid yeast system ( Figure S3A ) . Further evaluation of the sequence requirements for p53 transactivation from ¾-sites with the haploid yeast system showed p53 could transactivate from a ¾-site RE containing a CTTG core ( Figure S3B ) . Transactivation was also effected by surrounding flanking sequences . In our examination of full-site RE sequences containing spacers , several REs contradicted our finding that large spacers between decamer half-sites abolish p53 transactivation . For example , the PIDD RE ( p53 induced protein with death domain ) [67] which promotes apoptosis has an 8 nt spacer in its RE . Contrary to our expectation , high expression of WT p53 resulted in ∼20% of the level of transactivation from the p21-5′ RE . As expected , removal of the 8 nt spacer between the decamer half-sites increased the levels of p53 transactivation ( Figure 5B ) . Examination of the PIDD RE sequence suggested that rather than functioning from a canonical full-site RE , p53 might transactivate from a noncanonical ¾-site RE . The canonical PIDD element was first separated into two noncanonical ¾-site REs ( designated PIDD ¾ A and PIDD ¾ B ) , both of which utilized the “spacer” sequence as part of the binding element . The noncanonical REs were comprised of a ¼-site directly adjacent to a ½-site , but differed in the central CWWG motif in the half-site and number of mismatches from the consensus binding sequence . Transactivation assays revealed p53 could function from the PIDD ¾ A sequence to levels equivalent to the canonical PIDD RE containing the 8 nt spacer at high expression ( Figure 5B ) . In contrast , WT p53 could not transactivate from the PIDD ¾ B RE . These findings showed that the PIDD RE was neither a true canonical consensus RE nor an exception to the “spacer” rule , but rather a noncanonical ¾-site RE which can support p53 transactivation . The results of the PIDD RE sequence analysis suggested that other known p53 REs previously identified as canonical full-site REs are actually ¾-site REs . Based on an empirically derived set of RE rules previously established in our lab to predict p53 transactivation capacity [14] , [30] , [68] , several established p53 target REs were re-examined to determine if their responsiveness to p53 was actually due to the presence of a noncanonical ¾-site RE instead of a full-site RE . As shown in Figure 5C , the REs of p21-3′ , PCNA , 14-3-3σ ( site 2 ) and Apaf-1 REs were actually comprised of a genuine ¾-site binding element ( containing a ½-site directly adjacent to a ¼-site ) followed by an additional sequence that vastly deviated from a consensus ¼-site . Luciferase assays in the diploid yeast strains revealed that p53 transactivation from these sites was clearly reduced in comparison to transactivation of p21-5′ RE ( Figure 5C ) , but comparable to the levels of transactivation from the consensus ¾-site REs , Con J and Con K ( see Figure 5A ) . These results are consistent with p53 transactivation capacities towards the p21-3′ , PCNA , and Apaf-1 REs assessed with the haploid yeast system in a recent study focused on examining conservation of RE sequence and conservation of RE functionality [68] . We examined several p53 mutations that could address the structural requirements for transactivation from noncanonical REs ( ½- and ¾-site REs ) in comparison to canonical REs ( weak and strong full-sites ) . The following REs were examined: ½-site REs ( con G , con B , p21-5′ left , and p21-5′ right ) , ¾-site REs ( con J and con K ) , as well as weak ( TIGAR and Gadd45 ) and strong ( p21-5′ and TIGAR-spacer ) full-site REs . Mutations in the oligomerization domain were analyzed for transactivation capacity to determine the extent to which the tetramerization of p53 facilitated functionality from noncanonical REs . Also examined were N-terminal transactivation domain and C-terminal regulatory domain mutants . The results are summarized in Table 1 and presented relative to transactivation from the p21-5′ RE . The variation in transactivation by the p53 mutants towards the REs was not due to differences in levels of expressed p53 protein ( Figure S4 ) . The Arg337 residue is located on the surface of the tetramerization domain and plays a role in tetramer stabilization through a salt bridge with Asp352 on the opposite monomer [69]–[73] . The mutations R337C and R337H , associated with Li-Fraumeni-like syndrome ( LFL ) and adrenal cortical carcinoma ( ACC ) , respectively , appear to compromise the ability to form tetramers . The impact of the R337C mutation ( previously described as a partial function mutation ) on p53 functionality has been postulated to arise by a change in thermodynamic stability , shifting the equilibrium from the tetramer to dimer and/or monomeric states [74] , [75] . The R337H mutation is considered to have a subtle functional effect causing a pH dependent effect on folding [69] , [74] , [76]–[79] ( Storici and Resnick , unpublished data ) . In agreement with previous findings by Lomax et al . [74] , R337C was found to be an altered function mutation in the diploid yeast and SaOS2 cell systems , significantly diminishing transactivation from both canonical and noncanonical REs . Transactivation from the strong , full-site REs was reduced to less than 25% of the levels of WT p53 ( “+” in Table 1 ) , whereas transactivation from the weaker full-sites and the ¾-site REs was barely detectable ( “+/−” in Table 1 ) with no transactivation from the ½-site REs . Similar to previous results [74] , the p53 R337H mutant could transactivate from the p21-5′ RE to levels comparable to WT p53 in the haploid yeast system . However , although R337H could function from the p21-5′ RE in the diploid yeast and SaOS2 cell transactivation assays , the level was reduced in comparison to WT p53 . Similar results were found for transactivation from the weaker full-site REs and ¾-site REs . The differences between WT and mutant p53 were not as clear for the ½-site REs . L344 is one of five residues that comprise the hydrophobic core of the α-helices which form the interface for p53 dimer-dimer interactions [70] , [72] , [73] , [80] , [81] . The mutation L344A prevents tetramer formation although stable dimers are formed [72] , [80] , [82] , [83] . The missense mutation L344P disrupts the helix [74] , [84]–[86] and is associated with Li-Fraumeni syndrome ( LFS ) . This alteration prevents dimer formation resulting in monomeric p53 protein [74] , [75] , [86]–[89] . Within the haploid and diploid yeast systems , the L344A protein had reduced transactivation activity towards the strong canonical full-site REs and noncanonical ¾-site REs ( Table 1 ) and no activity towards half-sites . However , the L344A protein was capable of transactivating to low levels from the p21-5′ half-site REs within mammalian cells . Similar results were seen in SaOS2 cells with an N345S dimer mutation , where transactivation was measurable , but significantly reduced in comparison to WT p53 transactivation ( data not shown ) . Consistent with previous findings showing L344P is a loss-of-binding mutation , the monomeric p53 L344P was not able to transactivate from the full-site , ¾-site , or ½-site REs tested within the 2n yeast or SaOS2 cells ( Table 1 ) [74] , [75] , [86]–[89] . Similarly , the germline mutation L330H , which is predicted to form only monomers through destabilizing the β-strand of the tetramerization domain [80] , [87] , [90] , was also inactive for transactivation from the p21-5′ full-site RE and half-site REs in the mammalian system ( data not shown ) . Finally , deletion of the tetramerization domain ( Δ325–357 ) or a terminal truncation ( 331 stop: similar to the p53β isoform ) resulted in complete loss of transactivation from any of the binding sequences examined , including full-site REs , in the yeast or SaOS2 cell systems ( data not shown ) . These results were in agreement with previous findings that the tetramerization region is necessary for p53 binding to and transactivation from a RE [83] , [91] . Previously , we reported that the terminal deletion Δ368 resulted in reduced transactivation capacity towards full-site REs when assessed with a yeast ADE2 reporter plate assay ( on plates ) in haploid yeast [30] . Those studies have been extended to address its ability to transactivate from full- , ¾- and ½-site REs . We found that removal of the C-terminal tail resulted in at most a modest transactivation from the noncanonical and canonical REs ( Table 1 ) . The inhibition varied with the strength of the RE sequence , as well as with level of p53 expression ( data not shown ) for all REs . Thus , in comparing WT to mutant , the C-terminal domain of p53 appears in several cases to actually enhance rather than repress transactivation from canonical , as well as noncanonical REs . Finally , we also examined a deletion in the N-terminal domain , Δ1–39 , to assess if transactivation from ¾- and ½-site REs was differentially affected . The mutation resulted in <25% residual transactivation from the full-site REs ( strong: p21-5′ and TIGAR - spacer; moderate: GADD45 ) at low expression in the haploid yeast system ( data not shown ) . At the high expression levels required to examine noncanonical REs ( 0 . 128% galactose ) , transactivation was also reduced from ¾-site REs ( con J and con K ) and a ½-site RE ( con G ) suggesting that the canonical and noncanonical REs are similarly affected by defects in the transactivation domain .
Yeast has been extensively used as an in vivo test tube to analyze p53 transactivation capacities towards p53 response elements ( REs ) derived from human genes and assess the potential role that target sequence plays in p53 functionality [3] , [14] . A distinct advantage of yeast is the opportunity to rapidly modify in vivo either the target REs or to create mutant p53 coding sequences utilizing a highly efficient recombination-based system , known as delitto-perfetto [Italian for “perfect murder”] that targets desired changes with oligonucleotides [49] , [92] , [93] . The newly developed diploid yeast approach extends our previous work with a plasmid-based haploid yeast system in order to capitalize upon the opportunity to conveniently merge a large number of integrated ( single copy ) p53 mutants with transactivation capabilities at many response elements simply through mating of strains . Not only can the rheostatable , diploid yeast system differentiate between functional and nonfunctional REs , it can also estimate weak to strong functionality at different levels of p53 expression . The results in yeast have proven useful in guiding studies with highly expressed WT and mutant p53 towards potential target REs in human cells and evaluations of direct DNA binding . Previous studies have shown that widely-separated full-size REs associated with the same p53 target gene , such as the muscle creatine kinase ( MCK ) , can interact to synergistically transactivate the associated target gene [94] . The mechanism has been suggested to involve looping out of the intervening DNA so that multiple p53 tetramers can “stack” and concentrate the basal transcription machinery [94] , [95] . While this mechanism of transactivation may hold for REs separated up to 3 kb , the synergy was proposed to be lost when the distance was less than 25 bases due to steric hindrance [94] , an arrangement that holds for several p53 target genes including MDM2 . We found that p53 could function synergistically from the two weak MDM2 REs separated by a 17 and 10 nt spacer , but further reduction to 5 nt resulted in additivity . Interestingly , the length of the spacer between the two REs within the promoter of Mdm2 has been conserved between mouse and human although the sequence of the spacer is 50% diverged . In the haploid yeast system , p53 was found to function similarly from the human and murine MDM2 REs in terms of synergistic transactivation [68] . Given the divergence in the sequence of the spacer , it is unlikely that an additional transcription factor binding site would be found in the 17 nucleotides that would allow a second transcription factor to interact cooperatively with p53 to induce the observed synergistic transactivation from the two REs . Instead , there may be a functional conservation that assures the level of p53-mediated transactivation . This is supported by sequence analysis across 14 species , corresponding to at least 70 million years of evolution , where 12 had maintained the 17 nucleotide spacer ( with varying degrees of sequence divergence ) between RE1 and RE2 ( including chimp , rat , rabbit , dog , elephant , armadillo and hedgehog ) , while opossum and bat had a one nucleotide indel [96] , [97] . From this preliminary search , it appears that the spacer sequence is under stabilizing selection such that variations in the length of the spacer which would affect the ability of p53 to synergistically function from these sites were not observed . Interestingly , a similar phenomenon , is observed in the DNA sequences within and between ( but not flanking ) cis-regulatory elements of the otx , delta , wnt8 , and brachyury genes where insertions or deletions of random sequence do not occur as assessed by a comparison between the orthologous cis-regulatory regions and flanking sequences of the sea urchins , Strongylocentrotus purpuratus and Lytechinus variegatus [98] . We are currently investigating whether functional conservation of spacer length as a mechanism for regulating p53 functionality holds with other closely-spaced full-sized p53 REs . The impact of the spacer between full-sites acts opposite to that of a spacer between half-sites . In previous studies , it was proposed that REs had a “rotational specificity” where spacers between half-sites could abrogate p53-dependent transactivation if they were on opposing faces of the DNA helix , but had little impact on transactivation when the half-sites were on the same face [18] , [59] , [60] . Recently , in electrophoretic mobility shift binding assays ( EMSA ) , a 2 nt spacer did not significantly alter p53 affinity ( Kd ) towards a consensus RE [29] . Furthermore , in a 3D model derived through cryoelectron microscopy of full length p53 , where p53 dimers are proposed to form through interactions of α-helices on the N- and C-termini rather than α-helices on the C-termini , the best fit RE sequence had a spacer length of 6 nucleotides [99] . These in vitro findings imply if p53 transactivation is primarily dependent on the ability of p53 to bind target REs , a small spacer would not affect p53-dependent transactivation . The effect of spacer upon transactivation from a RE at varying levels of p53 was first analyzed with the WAF1/Cip1 p21-5′ RE . The present results are in agreement with an earlier report showing that increases in spacer length of 4 or 14 bases between half-sites decreases p53 transactivation in yeast [18] . However , there is no evidence for rotational specificity influencing transactivation . This is the first demonstration that a spacer of 1 or 2 nt between half-sites can dramatically affect the ability of p53 to transactivate from a RE . Importantly , the impact of spacer on the ability of p53 to transactivate from a RE is greatly influenced by the intrinsic potential transactivation strength of the RE , as shown for NOXA , as well as level of p53 expression . We also established with the haploid yeast system that a spacer had a similar negative effect on the ability of p53 family members , p63β and p73β , to transactivate from a RE . The biological importance of spacer as a means for affecting response to p53 was clearly demonstrated for the TIGAR RE . We propose that unlike the implications from the established consensus RE , a spacer between half-sites in natural elements serve to modulate the levels to which p53 can regulate the associated target gene and may play an important mechanistic role in the evolution of RE responsiveness to p53 . In the case of TIGAR , the opportunity to address evolutionary implications is limited due to the apparent recent emergence of the putative p53 binding site in the primate lineage , however , the orthologous TIGAR RE in Pan troglodytes ( chimpanzee ) , and Macaca mulatta ( rhesus monkey ) also contain a 2 nt spacer [96] , [100] . Several modes of binding have been postulated for p53 to a RE containing a spacer that include shifting of the tetramer on the DNA to compensate for the reduced protein-protein interactions , bending and/or kinking of the DNA within the spacer and/or RE and rotational motions of p53 to accommodate supercoiling of the spacer sequence [24] , [27] . In any of these , binding would modify either the conformation of the p53 tetramer and/or the DNA sequence itself in a fashion that is not predictable from current structural studies . While functional studies demonstrate that the RE sequence can dramatically influence the level to which p53 can transactivate from a specific RE [3] , the relationship between in vitro binding and efficiency of in vivo transactivation had not been established prior to this study . We utilized a novel fluorescent microsphere , semi-in vitro binding assay to determine the effect of spacer upon p53 binding . Recently , we demonstrated a good correspondence between binding and in vivo transactivation for several human p53 target REs ( Noureddine et al . , submitted ) . While differences were observed between the in vivo and in vitro assays for p53 binding to the p21-5′ RE containing a single nucleotide spacer , increasing the length of a spacer between the decamer half-sites beyond one nucleotide had a large impact on the ability of p53 to bind in the semi-in vitro assays , contrary to reports with pure components [22] , [23] . Interestingly , p53 binding to a RE with a ≥3 nt spacer was nearly equivalent to binding to the individual p21-5′ half-site REs , suggesting that p53 may recognize each decamer sequence as a distinct binding motif once a spacer increases beyond a certain length . These findings strongly argue that the 0–13 nucleotide spacer in the established p53 consensus sequence , which is based on interactions between purified protein and REs , should be reduced to no more than a few bases . It appears that factors present in nuclear extracts or cells limit the ability of p53 to recognize REs with spacers between half sites . It will be interesting to identify these putative spacer-discriminating factor ( s ) . The observations that p53 can function , as defined by binding and transactivation , from consensus half-site REs in yeast and from the p21-5′ half-site REs in mammalian cells indicate the p53 transcriptional network is comprised of many more downstream targets than previously predicted . Transactivation from half-sites is strongly dependent upon the targeted sequence and level of p53 expression . Similar to full-site REs [30] , the central binding motif at the monomer junction ( i . e . , CATG ) had the greatest impact upon the ability of p53 to function from half-sites . The difference between the yeast and mammalian systems in functionality from the half-sites indicates that additional co-factors in human cells may assist in p53 interactions with weaker binding sequences . Tetrameric p53 can bind full-site and half-site binding elements in vitro , but the p53 bound to a half-site had a much higher disassociation rate in vitro as indicated through “trap” assays that had measured dissociation of p53 from a labeled RE sequence [22] . Given the higher probability of a decamer sequence occurring in the genome compared to a 20 base sequence , it will be interesting to determine the number of functional p53 half-site REs in the human genome which is composed of ∼3 billion base pairs . In a preliminary genomics screen , we have identified over 1 , 400 “perfect” p53 half-sites containing a CATG core motif in the genome within 2 kb of a transcriptional start site . However , the question remains as to which of these sites will have a relevant biological function in p53-dependent stress responses or whether the sequences are merely “noise” within the genome . Half-sites could serve many roles in the genome . For example , a collection of half-sites could affect chromatin accessibility to transcriptional machinery ( i . e . , loosening the chromatin or recruiting and sequestering chromatin modifiers ) . Such sites might function in the proposed opening of the genome to make it more available for repair [101] . On the other hand , half-sites might also serve to titrate p53 . Additionally , p53 half-sites may play a role in bringing together different transcriptional networks as described for a SNP in the FLT-1 promoter [41] , [66] . Subsequent studies of the FLT-1 half-site have shown that p53 bound to the T-SNP can cooperate with estrogen receptor ( ER ) , also bound to a half-site of its own consensus sequence , to synergistically transactivate from the FLT-1 T promoter [66] . Such co-regulation may be advantageous for precise fine-tuning of p53-regulated responses and may make activity from the noncanonical ½-site REs more dependent on availability of cooperating transcription factors or additional cofactors and on levels of nuclear p53 . Within yeast , p53 was able to transactivate from noncanonical consensus ¾-site REs containing a ¼-site adjacent to a ½-site or a ¼-site separated from a ½-site by a 5 nt spacer . The lack of a significant difference between the transactivation from both ¾-site REs imply that p53 encounters the ¾-site RE as a tetramer protein . Transactivation from ¾-site REs was moderate relative to full-sites . Furthermore , the hierarchy of transactivation , in terms of the level to which p53 could bind and transactivate in vivo , was comparable to earlier in vitro studies that measured p53 binding efficiencies towards ½-sites , ¾-sites and full-site sequences with competitive gel retardation assays [59] . The impact of a ¾-site was revealed in our study of the RE for the PIDD gene where p53 transactivation was substantial even though the RE contained a spacer greater than 3 nt . Transactivation assays revealed the responsiveness of the originally described PIDD RE was actually due to a noncanonical ¾-site RE . Transactivation from the PIDD ¾ A RE , but not the PIDD ¾ B reinforces the requirement for a strong core element for functionality and that a functional RE cannot contain greater than 3 mismatches within a half-site . There are likely to be many functional ¾-site REs in the genome , among which are sequences that were originally considered as canonical , full-site REs containing >2 nt spacers . Furthermore , as recently pointed out in our functional conservation analysis of p53 REs from several species , there are clear examples of weaker , apparently noncanonical ¾-site REs , such as p21-3′ and APAF1 being conserved in evolution [68] . While this report is the first to systematically evaluate p53 function from noncanonical ½- and ¾-site REs , there are other reports of a transcription factor binding to noncanonical binding sites within the genome . Johnson et al . [102] , recently mapped the in vivo interactome for the transcription factor neuron-restrictor silencing factor ( NRSF ) using a large scale ChIP analysis and found that NRSF bound to noncanonical half-site binding motifs . A variety of approaches such as ChIPSeq or FAIRE , formaldehyde-assisted isolation of regulatory elements , may be able to reveal interactions at noncanonical sites under different stress conditions and in various cell types [103] , [104] to better understand the dimensions of the p53 master regulatory universe . Having established WT p53 interacts with sequences that do not fully conform to the canonical p53 binding sequence , a panel of p53 mutants was analyzed to determine what structural features of p53 play a role in transactivation from ½- and ¾-site REs versus full-site REs . Together , the mutations suggest that common functional aspects of p53 are required for the ability of p53 to function from full , ¾- , and ½-site REs . In general , the observed transactivation of oligomerization mutations towards full-site REs was in agreement with previous reports of mutants able to form tetramers or dimers retaining transactivation towards strong REs , while monomeric proteins were inactive [89] . Having established WT p53 interacts with noncanonical sequences , a panel of p53 mutations was analyzed to determine what structural features of p53 play a role in transactivation from ½-site and ¾-site REs versus full-site REs . The observed transactivation of oligomerization mutations towards full-site REs has confirmed previous reports of mutants able to form tetramers or dimers retaining transactivation towards strong REs , while monomeric proteins were inactive [89] . Together , the mutations suggest that common functional aspects of p53 are required for the ability of p53 to function from full , ¾- , and ½-site REs . The DNA binding domain of p53 has been shown in vitro to bind DNA in the absence of the oligomerization domain and that dimeric p53 can bind to half-sites in a cooperative fashion independent of tetramerization [22] , [105] . While able to transactivate at a low level from strong full-site REs and ¾-site REs , the L344A dimerization mutation was unable to transactivate from the weak full-site REs or consensus half-site REs in yeast . In agreement with these findings , Waterman et al . previously showed that the L344A protein could bind in vitro to oligonucleotides containing an optimal consensus binding site and/or half-site , but not to an oligonucleotide containing a suboptimal full-site [83] . The impact of the dimer mutation L344A , as well as N345S , was less severe in mammalian cells . Similarly , a designed dimer mutant ( Met340Gln/Leu344Arg ) was capable of transactivating from a p21 full-size RE within SaOS2 cells to half the level of WT p53 suggesting a direct correlation between the oligomeric state and transactivation activity [75] . While L344A dimeric proteins may bind to the sites in a cooperative fashion , tetramerization which is a necessity for efficient transactivation may be required to stabilize the binding and reduce the off-rate of p53 from the DNA in yeast . It is possible that other factors in SaOS2 cells may contribute to this stabilization accounting for the difference between the systems . Since altered tetramer proteins and dimers exhibited a reduced ability to transactivate from noncanonical REs , transactivation is not simply an additive process determined by the number of available pentamer sequences . It is possible that binding and transactivation in vivo , but not binding alone , may require a conformational change of the p53 protein that is only obtainable with a tetrameric protein . Several transcription factors have displayed such a characteristic including the heat shock protein which can bind chromatin weakly as monomer , but is not sufficiently active to cause a biological response until after it is induced and forms a trimer [106] , [107] . Furthermore , in the case of a half-site , where there is not a sequence to which the second p53 dimer can bind , tetramerization may assure the second dimer is in the vicinity of the p53 binding element in order to interact with other factors required for transactivation . The finding that the p53 Δ368 mutation reduced transactivation from both canonical and noncanonical REs agrees with studies showing that p53 does not require modification of the C-terminal to engage its target binding sites [31] , [32] . Furthermore , these results indicate that the C-terminal is not absolutely essential for function from canonical or noncanonical sites , but may play a role in regulating the level of transactivation . The deletion of the first N-terminal transactivation domain ( TAD ) which mimics a naturally occurring alternatively spliced form of p53 that is differentially expressed in breast tumors , had a more severe impact upon transactivation than previously reported with the qualitative ADE2 color reporter ( data not shown ) [30] , [108] . The results may reflect the greater quantitative assessment of transactivation with the luciferase assay . Nevertheless , both assays revealed that the decreased opportunity to interact with the transcriptional machinery through the loss of the transactivation domain strongly affects p53 transactivation from both canonical and noncanonical REs . Through deconstructing the canonical consensus sequence and assessing functionality , as determined by binding and transactivation within three in vivo model systems and a semi-in vitro binding assay , it has been possible to refine the requirements for functional p53 binding elements . The organization and arrangement of binding motifs , as well as the level of p53 expression , have a large impact on the ability of p53 to transactivate from a RE sequence . It is interesting that p53 may not function from reported canonical consensus REs yet functions from noncanonical REs that are common to the human genome . Half-sites and spacers between full-sites may provide additional levels of regulation in p53 transcriptional network . Truly functional REs may be restricted to decamers separated by <3 nucleotides . When the spacer increases beyond 3 nucleotides , we suggest p53 recognizes the half-sites as separate binding entities . The ability of p53 to function from noncanonical decamer half-sites introduces a new realm of target sequences and genes into the p53 transcriptional network and expands the universe of p53 regulated genes . Noncanonical sequences might provide p53 responsiveness at high levels of expression or in combination with other transcription factors , as for the case of the FLT-1 gene [66] . As discussed in the “piano model” for transactivation from a broad range of REs [14] , variation in the level of p53 , as well as various mutants can markedly affect the spectrum of p53 responses . Determining the relationship between expression levels and responsiveness at canonical and noncanonical target sequences is important in understanding how p53 implements cellular fate in response to stress , such as cell cycle arrest or apoptosis . It is also important for addressing the consequences of p53 alterations , particularly those mutations that retain transactivation capabilities , as well tailoring individual therapies .
The site-directed mutagenesis system , delitto perfetto [48] , [49] was used to generate a panel of isogenic “p53-host” strains and a panel of response element ( RE ) reporter strains in the budding yeast , S . cerevisiae . Each “p53-host” strain , yAT-iGAL::p53 ( MATa leu2-3 , 112 trpl-1 his3-11 , 15 can 1-100 ura3-1 , trp5::pGAL1:p53:cyc1-Ter , lys2::HygroR ) , contains the wild type p53 cDNA controlled by the inducible , “rheostatable” GAL1 promoter [109] integrated at the TRP5 locus on chromosome VII . p53 mutations in the tetramerization and basic domains were constructed using a derivative of the p53 host strain containing a CORE cassette ( CO , counterselectable , KLURA3; RE , reporter , KanMX4 resistance gene ) integrated within the p53 cDNA at nucleotide position 1105 [110] . Modification of the p53 cDNA was performed using the delitto perfetto approach [48] , [49] where the CORE cassette was replaced with an oligonucleotide containing the mutation of interest to generate a full-length mutant p53 cDNA or the desired deletion of the C-terminus . Replacement of the CORE was confirmed by selection on 5-FOA and kanamycin sensitivity . Specific p53 alterations were confirmed by colony PCR and sequencing ( Big dye , Applied Biosystems , Foster City , CA ) . The second panel of isogenic strains containing p53 REs upstream of the CYCl minimal promoter and the firefly luciferase reporter were obtained starting from the yLFM-ICORE strain , as previously described ( Table S1 ) [42] . The reporter strains are also isogenic with the p53 host strains , but LYS2 and Hygros . Mating of the reporter and p53-host strains followed by selection for diploid cells on Lys− Hygro+ plates , results in isogenomic yeast that enable the assessment of the transactivation potential for WT or mutant p53 proteins towards individual REs in the p53 transcriptional network . Strains differ only by the mutation of interest and 4–5 nucleotide variation in the RE . Individual p53-inducible RE reporter colonies were inoculated into 5 ml rich media , YPDA plus adenine [200 mg/L] , and grown overnight at 30°C with shaking . The overnight culture was diluted 1∶50 in H2O . For each reaction , 1 ml of the diluted culture was spun down , washed of residual glucose with H2O and re-suspended in 2 mL synthetic complete - LYS media containing 2% raffinose or raffinose supplemented with either 0 . 008% or 0 . 024% galactose . These dilute raffinose cultures were grown overnight ( ∼18 hr ) at 30°C to ∼2–4×107 per ml . The 2 ml cultures were spun down and the supernatant was aspirated . The remaining pellet was resuspended in 100 µl reporter lysis buffer ( Promega , Madison , WI ) and an equivalent amount of 425–600 micron acid-washed , glass beads was added ( Sigma , St . Louis , MO ) . Samples were homogenized for 30 seconds in the Biospec Products , Inc . mini-bead beater ( Bartlesville , OK ) , briefly incubated on ice and spun for 20 minutes at 16k relative centrifugal force ( rcf ) in an Eppendorf 5415R centrifuge ( Batavia , IL ) to separate the soluble protein fraction . The standard protocol recommended by the manufacturer ( Promega; Madison , WI ) was performed for the luciferase assay system starting with 10 µl of protein extract . Luciferase activity was measured from 96-well , white optiplates ( Perkin Elmer , Waltham , MA ) in a Wallac Victor2 multilabel counter ( Perkin Elmer , Waltham , MA ) . Light units were standardized per µg protein as determined by a Bio-Rad protein assay ( Bio-Rad; Hercules , CA ) . Luciferase assays in haploid strains were performed as previously reported [14] . Human SaOS2 ( HTB-85 , ATCC ) osteosarcoma cells were grown in McCoy's A5 medium supplemented with 10% FBS and 1× penicillin/streptomycin ( Gibco , Carlsbad , CA ) . All cultures were incubated at 37°C with 5% CO2 . Human lymphoblastoid cells were grown in RPMI 1640 media supplemented with 15% heat-inactivated fetal bovine serum ( Invitrogen , Carlsbad , CA ) and incubated at 5% CO2 at 37°C with 1% penicillin-streptomycin antibiotics ( Invitrogen ) . The lymphoblast cell lines GM12824 and GM12825 used in the semi-in vitro binding assay were purchased from Coriell Cell Repositories ( Camden , NJ ) . Plasmids pC53-SN3 [111] coding for human p53 cDNA under the control of CMV promoter and the control vector pCMV-Neo-Bam were kindly provided by Dr . Bert Vogelstein . p53 mutations within the pC53-SN3 vector were generated by site-directed mutagenesis ( QuickChange site-directed mutagenesis kit , Stratagene ) and were confirmed by sequencing . Luciferase reporter constructs containing the p53 REs were constructed in pGL3-Promoter backbone ( Promega , Madison , WI ) . Partially complementary oligonucleotides containing the RE and 5 additional flanking nucleotides from both sides were annealed in vitro to yield double-stranded molecules that would be compatible for an in vitro ligation reaction with XhoI/BamHI double digested pGL3-promoter vector . Ligation products were transformed into XL1 blue E . coli cells , purified , amplified and sequenced . pRL-SV40 , a reporter plasmid coding for Renilla reniformis luciferase ( Promega , Madison , WI ) was used as a control of transfection efficiency in the luciferase reporter assay . For luciferase assays SaOS2 cells were seeded in 24-well plates 24 hours before transfection . Cells were transfected using Fugene-6 ( Roche , Indianapolis , IN ) according to manufacturer's instructions at ∼80% confluence ( with 200 ng of reporter constructs ) . When appropriate , 25 ng of the p53 was co-transfected . Total plasmid DNA per well was adjusted to an equal level by adding the empty vector pCMV-Neo-Bam . Forty-eight hours post transfection , extracts were prepared using the Dual Luciferase Assay System ( Promega ) following the manufacturer's protocol and luciferase activity was measured in a Victor Wallac multilabel plate reader ( PerkinElmer ) . For each construct , relative luciferase activity is defined as the mean value of the firefly luciferase/Renilla luciferase ratios obtained from at least three independent experiments . SaOS2 cells were seeded in 10-cm dishes and transfected at 80% confluence with WT p53 expression vector along with the pGL3-P reporter plasmid containing the p21-5′ RE or its derivatives . ChIP on plasmid assays were performed as described previously [66] using the ChIP kit ( Upstate Biotechnology ) following the manufacturer's instructions . A mouse monoclonal anti-p53 antibody DO7 ( Pharmigen , BD ) was used . PCR amplifications were performed on immunoprecipitated chromatin using a pair of primers to amplifya specific region in the pGL3-P backbone containing the p53 REs ( 5′-ATAGGCTGTCCCCAGTGCAA-3′ and 5′-TGGAATAGCTCAGAGGCCGA-3′ ) . The PCR cycles were as follows: an initial 10 min Taq Gold polymerase at 95°C followed by 40 cycles of 95°C for 15 s and 60°C for 1 min . The PCR products were then run on a 1 . 8% agarose gel and quantified with IMAGEQUANT V5 . 1 ( Molecular Dynamics-GE , Piscataway , NJ ) . To evaluate the sequence-specific p53-DNA binding interactions , we used the semi-in vitro fluorescent microsphere binding assay as previously described ( Noureddine et al . , submitted ) . Briefly , lymphoblast cells were grown to ∼8 . 5×105 cells/mL before exposing to 0 . 6 ug/mL ( 1 mM ) doxorubicin ( Sigma , St . Louis , MO ) for 18 hours at 37°C . Nuclear protein was extracted from non-treated and treated cells using a Nuclear Extraction Kit according to manufacturer's protocol ( Active Motif , Carlsbad , CA ) and protein concentration was measured in triplicate using the BCA Protein Assay Kit ( Pierce , Rockford , IL ) , followed by a plate read using a Perkin Elmer HTS 7000 BioAssay Reader . All the oligonucleotides reported in this study were synthesized by Invitrogen ( Carlsbad , CA ) . Fluorescent microspheres bearing double stranded DNA fragments containing a sequence of interest were multiplexed and incubated for 1 hour in the presence of 1 . 75 µg of nuclear protein extracts from either treated or non-treated cells . Following incubation in cell extracts , the beads were incubated with p53 antibodies ( DO-7 , BD Biosciences , San Jose , CA ) and secondary antibodies conjugated with phycoerythrin ( R-phycoerythrin-coated goat anti-mouse ) for 30 minutes . The p53 interaction with each bead was measured on a BioPlex Machine ( BioRad ) as raw fluorescence intensity signal generated from phycoerythrin-conjugated secondary antibody to mouse anti-p53 . For signal normalization , an aliquot of the DNA-conjugated beads was treated separately with phycoerythrin-conjugated streptavidin for 20 minutes in the absence of any nuclear extracts . All binding reactions were conducted in triplicate . For each bead type in every multiplex set of beads , relative binding signal was obtained by normalizing the absolute binding signals of extract-treated beads ( mean of 3 replicates ) to mean signals obtained from the same set of beads that were independently treated with phycoerythrin-conjugated streptavidin ( mean of 3 replicates ) . This normalization accounts for bead type-specific oligo content . Net binding for each oligo is the numerical difference between NT and DOX-treated signal obtained for this oligo . | Within human cells , the tumor suppressor p53 is the central node of regulation required to elicit multiple biological responses that include cell cycle arrest and death in response to stress or DNA damage , where mutations in p53 are a hallmark of cancer . As a master regulatory gene , p53 controls the action of target genes within its network by directly interacting with a widely accepted consensus DNA binding sequence , composed of two decamer ½-sites that can be separated by up to 13 bases . While mismatches from consensus sequence are frequent , the canonical consensus sequence places a limitation upon the organization and number of target genes within the p53 transcriptional network . Using yeast and human cell systems , our goal was to further understand how the DNA sequence , DNA organization , and level of p53 expression might influence the inclusion of genes within the p53 regulatory network . We found that increases in spacer beyond a few bases greatly reduce responsiveness to p53 . Importantly , we established that p53 can function from noncanonical sequences comprising only a decamer ½-site or a ¾-site . These findings further define and expand the universe of potential downstream target genes which may be regulated by p53 and bring further diversity into the p53 regulatory network . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry/molecular",
"evolution",
"molecular",
"biology/transcription",
"initiation",
"and",
"activation",
"molecular",
"biology/molecular",
"evolution",
"cell",
"biology",
"evolutionary",
"biology/genomics",
"biochemistry/transcription",
"and",
"translation",
"genetics",
"a... | 2008 | Noncanonical DNA Motifs as Transactivation Targets by Wild Type and Mutant p53 |
Herpesviruses encode a characteristic serine protease with a unique fold and an active site that comprises the unusual triad Ser-His-His . The protease is essential for viral replication and as such constitutes a promising drug target . In solution , a dynamic equilibrium exists between an inactive monomeric and an active dimeric form of the enzyme , which is believed to play a key regulatory role in the orchestration of proteolysis and capsid assembly . Currently available crystal structures of herpesvirus proteases correspond either to the dimeric state or to complexes with peptide mimetics that alter the dimerization interface . In contrast , the structure of the native monomeric state has remained elusive . Here , we present the three-dimensional structures of native monomeric , active dimeric , and diisopropyl fluorophosphate-inhibited dimeric protease derived from pseudorabies virus , an alphaherpesvirus of swine . These structures , solved by X-ray crystallography to respective resolutions of 2 . 05 , 2 . 10 and 2 . 03 Å , allow a direct comparison of the main conformational states of the protease . In the dimeric form , a functional oxyanion hole is formed by a loop of 10 amino-acid residues encompassing two consecutive arginine residues ( Arg136 and Arg137 ) ; both are strictly conserved throughout the herpesviruses . In the monomeric form , the top of the loop is shifted by approximately 11 Å , resulting in a complete disruption of the oxyanion hole and loss of activity . The dimerization-induced allosteric changes described here form the physical basis for the concentration-dependent activation of the protease , which is essential for proper virus replication . Small-angle X-ray scattering experiments confirmed a concentration-dependent equilibrium of monomeric and dimeric protease in solution .
The family Herpesviridae is divided into the three subfamilies alpha- , beta- and gammaherpesviruses . These infectious agents cause a variety of diseases in many different hosts including humans . Pseudorabies virus ( PrV ) is a neurotropic porcine alphaherpesvirus [1] and the causative agent of Aujeszky's disease . The pig is the only susceptible species that can survive a PrV infection depending on the age of the animal and virulence of the virus , while most other mammals die within a few days . Only higher primates including humans and equids are resistant to infection . Due to its broad host range PrV has become an important model system to study herpesvirus biology in cell culture and in the natural host . PrV genome organization and protein content exhibit significant homology to that of the human herpes simplex virus type 1 ( HSV-1 ) [2 , 3] , which is among the best-studied herpesviruses . Capsid assembly of HSV-1 has been intensively analyzed . However , since herpesvirus capsid proteins are well conserved , it is very likely that the process leading to mature , DNA-filled nucleocapsids is also similar . The proteolytic activity of the serine protease is essential for this process [4] . In HSV-1 and PrV , this protease is encoded by the UL26 gene [5] , which is the longest open reading frame in a family of in-frame overlapping genes [5–8] . UL26 overlaps in frame with UL26 . 5 [3] . UL26 and UL26 . 5 possess identical 3'-termini , which encode a scaffold protein while the unique 5'-terminus of UL26 contains the protease domain . There are at least two target sites for the protease in the full-length UL26 protein ( pUL26 ) [8 , 9] . Autoproteolytic activity at the release site ( R-site ) results in release of the N-terminal protease domain ( pUL26N , also called VP24 or generic: assemblin ) and the C-terminal part containing the scaffold protein ( pUL26C , also called VP21 or generic: assembly protein ) [10] . Due to the presence of a linker region pUL26C is 21 amino-acid residues longer than pUL26 . 5 ( Fig 1 ) . Near the C-terminus of the scaffold protein is the maturational site ( M-site ) where pUL26 . 5 and pUL26C are cleaved [10] . The scaffold protein binds to the major capsid protein pUL19 ( VP5 ) and directs it to the nucleus [11–17] . During capsid assembly , the scaffold protein forms a scaffold core with the major capsid protein bound to the C-termini of pUL26/pUL26C/pUL26 . 5 [17] . When the capsid is fully assembled the scaffold is cleaved at its M-site releasing the ring-like scaffold structure from the capsid , which is then expelled during DNA packaging . In contrast , the protease remains in the nucleocapsid [18] . Without protease activity , the scaffold remains in the capsid resulting in capsids without viral DNA designated as B-capsids . Upon activation of the protease , the B-capsids mature and subsequent steps of viral replication occur as shown with a temperature-sensitive HSV-1 pUL26 mutant [19] . The function of pUL26 . 5 can be taken over by pUL26C but with reduced efficiency as well as loss of an apparent core structure in the resulting capsids [22] . Proteases of several herpesviruses , such as human cytomegalovirus ( HCMV ) and Kaposi's sarcoma-associated herpesvirus ( KSHV ) , have one or more internal cleavage sites to regulate activity or promote destabilization of the protease [23–25] . For proteases of PrV , HSV-1 and herpes simplex virus type 2 ( HSV-2 ) no internal cleavage sites have been reported . Herpesvirus maturational proteases exist in a monomer-dimer equilibrium . Dimers are weakly associated with dissociation constants ( KD ) in the micromolar range [26] and are active , while monomers are almost inactive [27] . It was shown that dimerization of the HCMV assemblin is dependent on protein concentration . The fraction of dimeric protease at 0 . 2 μM was demonstrated to be ~0 . 3 increasing to ~0 . 7 at 4 . 5 μM [27] . Additionally , dimerization is favored by high concentrations of kosmotropic compounds like glycerol [26 , 27] . Activity of the herpesvirus protease has to be strictly regulated . The enzyme is expressed as full-length pUL26 in the cytosol of infected cells where its concentration is low and thus the inactive , monomeric form is predominant [28] . The major capsid protein is bound by pUL26 . 5 and translocated to the nucleus via the nuclear localization sequence within pUL26 . 5 [29] . Since pUL26 encompasses pUL26 . 5 , it also contains this nuclear localization sequence . Therefore , autoproteolytic activity of the protease in the cytosol would prevent its localization to the nucleus . When capsid assembly is completed , the protease has to become active to release the scaffold protein from the capsid and to allow DNA packaging . In the capsids , the concentration of protease is much higher than in the cytosol , thus promoting dimerization [28] . Additionally , it was proposed that the capsid environment itself might enhance proteolytic activity of the protease [30] . Several structures of homologous assemblins of other herpesviruses have been published , revealing the overall fold , active site and biological assembly [31–38] . The sequence identities of these structures to the PrV assemblin range from 60% ( alphaherpesviruses ) to 30% ( beta- and gammaherpesviruses ) [39 , 40] . All assemblins consist of 6–9 α-helices surrounding a β-barrel formed by two β-sheets . The catalytic triad is unique among the serine proteases and consists of one serine and two histidine residues . The active site is solvent accessible and distal to the dimer interface . Nevertheless , dimerization drastically influences the activity [27] . There is evidence , that upon dimerization the oxyanion hole is formed by structural changes of a loop containing two strictly conserved arginine residues [41 , 42] . Currently , crystal structures are available for native dimeric and covalently inhibited dimeric assemblins . Recently , three structures of the truncated KSHV assemblin in complex with helical-peptide mimetics were published [43 , 44] . These compounds bind to the dimerization area and disrupt the dimerization interface of full-length KSHV assemblin [45] . Here , we report crystal structure analyses of the active dimer , the diisopropyl phosphate-inhibited dimer as well as the non-inhibited monomer of pUL26N from PrV ( 224 amino-acid residues ) . The latter is the first-ever structure of an assemblin in its native monomeric state . Comparison of the monomeric and dimeric forms provides insight into the regulation of protease activity by dimerization and a structural basis for rational design of therapeutic substances that trap the protein in the inactive monomeric state . Additional details and information about herpesvirus proteases and its involvement in capsid maturation are reviewed elsewhere [21 , 46–52] .
The active site serine is solvent accessible and the catalytic residues are part of β-strands β5 ( Ser109 ) and β6 ( His128 ) and the β2-β3 loop ( His43 ) , like in earlier structures [31–37] . In our inhibitor complex , covalent binding of diisopropyl fluorophosphate has formed a phosphate ester with the active-site serine in a manner also observed in crystal structures of homologous assemblins [34] . In the native structure , a chloride ion occupies the oxyanion hole , which is formed by the β6-β7 loop ( residues 133–142 , further referred to as oxyanion-hole loop , OHL , Fig 3 ) . Two consecutive arginine residues in this loop ( Arg136 and Arg137 ) are strictly conserved throughout all herpesvirus maturational proteases ( S3 Fig ) . The backbone N-H of Arg136 provides the first hydrogen-bond donor of the oxyanion hole . A water molecule as second hydrogen-bond donor supports anion stabilization , as does the positive local environment established by these two arginine residues . For HCMV assemblin it was shown that this water molecule plays a role in catalysis [56] . It is kept in place by a second water molecule that is positioned by hydrogen bonds of the peptide backbone oxygen of Leu10 ( loop β1-α1 ) and the peptide N-H of Leu110 ( β5 ) . Both water molecules are present in the dimeric structures of assemblins from PrV ( this report ) , HSV-2 ( pdb entry 1at3 ) , KSHV ( pdb entries 1fl1 , and 2pbk ) and HCMV ( pdb entries 1cmv , 1wpo , 1id4 , 1iec , 1ied , 1ief , and 1ieg ) . The positions of these water molecules seem to be conserved in all active assemblins . Our findings for the composition of the oxyanion hole are consistent with those reported for HSV-2 [34] . An extensive network of hydrogen bonds stabilizes the OHL ( S4 Fig ) . This network includes conserved water molecules and parts of the α1 region , α8 and β5 . There are five strictly conserved residues in the OHL and each of these is involved in the network of hydrogen bonds that positions Arg136 and stabilizes the oxyanion hole . The side chain of Arg137 forms hydrogen bonds to the peptide backbone of Leu20 ( α1 ) and Leu110 ( β5 ) , which are also conserved in assemblins ( S3 Fig ) . The OH-group of the conserved Thr140 forms a hydrogen bond to the backbone oxygen of Arg137 whereas the backbone oxygen of Thr140 accepts a hydrogen bond from the backbone N-H of Ala108 ( β5 ) . Inspection of other assemblin structures showed that these hydrogen bonds are present in all dimeric structures . In the PrV assemblin , the peptide oxygen of Val138 is connected to the peptide N-H of Leu12 via hydrogen bonds mediated by a conserved water molecule ( found in HSV-2 and KSHV assemblins ) . Most likely this water molecule and hydrogen bonding pattern are present in all herpesvirus maturational proteases . Additionally , the OHL of PrV pUL26N is maintained by hydrogen bonds of the side chains of Asp16 ( α1 ) , Glu214 and Arg211 ( both from α8 ) with the peptide backbone of Val138 , Gly139 and Gly135 , respectively . These hydrogen bonds are also present in HSV-2 assemblin with Glu214 replaced by Gln . Identical residues are present in the VZV assemblin , but the side chains of Glu and Asp point in different directions and do not form the corresponding hydrogen bonds . This ambiguity may result from the limited resolution of that structural model ( pdb entry 1vzv with a resolution of 3 Å , no structure factors were deposited ) . Most likely , the loop is stabilized comparable to HSV-2 and PrV assemblins . Diffraction datasets of dimeric pUL26N were collected from needle-shaped crystals and phasing by molecular replacement was successful either by using a monomer or the complete dimer as search models . For datasets derived from morphologically different , plate-shaped crystals , molecular replacement was only successful when using a monomer as a search model . In the resulting structure , two chains are present in the asymmetric unit . These chains and their adjacent symmetry mates are not in a proper position to form the known dimer . PDBePISA [53] suggests one assembly . This putative dimer has no local dyad , which is unusual for biologically relevant dimers [57 , 58] . The interface area ( 978 Å2 ) is much smaller than average for homodimers of this molecular weight ( ~1 , 500 Å2 ) [57] . Furthermore , the helices forming the interface have B-factors ( 60–100 Å2 ) above average ( 52 Å2 ) . Thus , the suggested dimeric interaction is a mere crystal contact and the crystal structure actually corresponds to monomeric pUL26N . The region corresponding to helix α1 in dimeric PrV pUL26N is disordered in the monomeric form . For ease of comparison , in this report numbering of the helices in the monomeric protease is adapted to that of the dimeric protease . The β-barrel and distal side of the dimerization area of chain A align very well with chain B of the asymmetric unit . The OHL and helices α3 , α7 and α8 of the dimerization area , on the other hand , differ in chain A and B ( Fig 4 ) . The OHL has two alternative conformations in chain A with equal occupancies . One conformation correlates to the OHL of chain B and the other one is shifted towards α8 due to crystal contacts . Compared to chain B the helices α3 and α7 of chain A are bent and helix α8 is slightly tilted . Accordingly , the overall r . m . s . deviation of chain A and B is 1 . 03 Å ( 207 aligned Cα atoms ) . Furthermore , there is no electron density observed for residues 16–19 of chain A ( β1-α2 loop ) and very weak electron density for residues 14–18 and 194–196 of chain B ( β1-α2 loop and the short α7-α8 loop , respectively ) . The core of the monomer is rigid as indicated by low B-factors of Cα atoms in the β-barrel , helix α4 and most of helix α8 of chain A ( Fig 4 ) . In contrast , the periphery of both monomers is flexible as evidenced by increased B-factors of Cα atoms at the distal side of the dimerization area , the OHL , and the dimerization helices α3 and α7 as well as in helix α8 of chain B . B-factors around 100 Å2 are also observed at the N-termini of helices α8 in both chains . Taken together , these observations show that the dimerization area and the parts necessary for formation of the oxyanion hole in dimeric pUL26N are flexible and not strictly ordered in monomeric pUL26N . The tertiary structures of monomeric and dimeric PrV pUL26N are partially similar . The r . m . s . deviations of inhibited dimeric pUL26N chain B with chain A and chain B of monomeric protease are 0 . 93 Å ( 192 aligned Cα atoms ) and 1 . 01 Å ( 190 aligned Cα atoms ) , respectively . The β-barrel , the distal side of the dimer interface , and helix α4 of the dimer interface are almost identical in monomeric and dimeric pUL26N ( Fig 5 ) . The r . m . s . deviations of these parts of inhibited dimeric pUL26N chain B with chain A and chain B of monomeric protease are 0 . 54 Å ( 151 aligned Cα atoms ) and 0 . 63 Å ( 149 aligned Cα atoms ) , respectively . In contrast , significant differences are observed for the OHL , N- and C-termini of helices α3 , α7 and α8 of the dimer interface , and the α7-α8 loop ( Fig 5 ) . In the monomer , the N-terminus of helix α8 is one and a half turn longer , but the C-terminus of helix α7 is one turn shorter . Thus , the α7-α8 loop conformation is shifted towards α8 upon dimer formation . In dimeric pUL26N helices α3 and α7 form a hydrophobic cleft that is occupied by helix α7 of the second monomer within the dimer . This cleft is closed in the monomer by bending the C-termini of these helices towards each other ( Fig 5 ) . In the monomeric pUL26N , the OHL is positioned near the N-terminus of α8 . Thus , the activation of pUL26N relies on dimerization-induced allosteric changes , which shift the OHL towards the active site ( Fig 6A ) . The top of the OHL ( Arg136Cα ) moves approximately 11 Å . In the dimer , the N-terminus of α8 is unwound , so some residues which point towards the protein core in the monomeric pUL26N are buried by the adjacent monomer of the dimer ( Fig 6B ) . However , the side chain positions of the catalytic triad are unchanged as suspected by Batra et al . [41] . Comparison of the active and inactive conformations of the OHL reveals a key location in the structure that is occupied by alternative hydrophobic residues . In dimeric PrV pUL26N , Ile134 is present at this position , whereas in the monomeric form Val138 takes its place ( Fig 6A ) . The hydrophobic character of the Ile134 position is highly conserved in all sequences and three-dimensional structures of assemblins ( S3 Fig ) . The hydrophobic character of Val138 is type-conserved in alphaherpesvirus sequences . Both residues are kept in place by hydrophobic interactions with two side chains of helix α8 ( Leu207 , Val208 ) or the corresponding helices in related structures . The hydrophobic character of these residues is conserved in herpesvirus proteases , although Thr and long , partially aliphatic side chains appear to be tolerated ( e . g . Thr in Epstein-Barr virus ( EBV ) assemblin , or Lys in KSHV assemblin , S3 Fig ) . Another consequence of the different conformation of the OHL in the monomeric form is that the Asp16-containing region and α1 are disordered . This loop , the OHL and major parts of the dimer interface area are flexible in monomeric pUL26N as indicated by partially disordered segments and high B-factors . Thus , the crystallographically observed increase of order upon dimerization of PrV assemblin is in line with the disorder-to-order mechanism of dimerization previously proposed for KSHV assemblin [28 , 43] . In our crystallization assays we used a construct of PrV pUL26N that was shortened by one C-terminal amino-acid residue ( Ala225 ) compared to the in vivo form . The resulting overall fold , conformation of the OHL , and position and orientation of the residues of the catalytic triad are identical to those in previously determined dimeric full-length assemblin structures of related herpesviruses , strongly suggesting that the C-terminal deletion does not affect the activity or structure of the protease . Moreover , native and inhibited dimers of PrV pUL26N crystallized isomorphously and the diisopropyl phosphate-ester of the active-site serine is quantitatively observed in the inhibited structure . The inhibitor diisopropyl fluorophosphate reacts specifically with the active-site serine , but hydrolyzes in aqueous solutions with a half-life of one hour at pH 7 . 5 and even faster at higher pH [59] . After incubation of PrV pUL26N with a sevenfold excess of diisopropyl fluorophosphate ( 5 mM ) in a buffer at pH 7 . 5 at room temperature for one hour , the protein was crystallized at pH 8 . Based on these constraints quantitative inhibition of PrV pUL26N ( minimum 90% in the crystal ) prior to inhibitor hydrolysis can be calculated to occur at a second order reaction rate of at least 0 . 09 s−1 M−1 . This is consistent with the known reaction rate of full-length HSV-1 assemblin with diisopropyl fluorophosphate of 1 s−1 M−1 at higher pH ( pH 8 . 0 ) and higher temperature ( 30°C ) [60] . Thus , the catalytic site of PrV pUL26N remains fully reactive with respect to this substrate-like inhibitor . Furthermore , earlier work has shown that proteolytic activity with respect to substrates containing the M-site is not significantly reduced by 3- or 8-residue C-terminal truncations [61] . C-terminally extended assemblins on the other hand do show a considerable decrease in activity [37 , 61 , 62] , presumably because such an extension sterically interferes with the proper positioning of helix α8 , which is required to establish the correct conformation of the OHL . The absence of Ala225 , however , does not interfere in any way with helix α8 , the conformation of the active site , and the OHL . Indeed , in the crystal structure of VZV assemblin the C-termini are disordered [35] , further confirming that the C-terminal region does not have a substantial impact on the overall protein fold . Structural properties and self-association behavior of PrV pUL26N in solution were characterized by small-angle X-ray scattering ( SAXS ) . Data were recorded at protein concentrations between 0 . 5 mg/ml and 10 mg/ml . Normalized SAXS intensities vary with protein concentration , suggesting the presence of more than one oligomerization state . The experimental SAXS curves can be accurately described assuming a monomer-dimer equilibrium and using form factors for the individual states derived from the monomeric and dimeric crystallographic models ( a representative example is shown in S5 Fig ) . Monomer volume fractions resulting from curve fitting with the program OLIGOMER [63] are shown in S5 Fig and S1 Table . At the highest protein concentration investigated ( 10 mg/ml , which approximately corresponds to the starting concentration in our crystallization experiments ) , a considerable volume fraction ( 30% ) of pUL26N is present in the monomeric form . This observation is consistent with the fact that crystals of the monomer could be obtained under these conditions . In contrast , inclusion of 0 . 2 M MgCl2 in the buffer used for SAXS measurements markedly shifts the monomer-dimer equilibrium , resulting in nearly complete dimerization at a protein concentration of 10 mg/ml . This result may explain why crystallization of the dimeric form of pUL26N required the presence of MgCl2 . Particle shape under conditions that result in virtually complete dimerization according to the analysis with OLIGOMER ( i . e . 10 mg/ml protein in the presence of MgCl2 ) was also determined independently by means of an ab initio approach . The size and oblong shape of the final bead model obtained with DAMMIN [64] are in good agreement with the crystallographic structure of the pUL26N dimer ( S5 Fig ) , further confirming the nature of the concentration-dependent self-association that is observed here . The dissociation constants with and without MgCl2 can be estimated to 200 μM and 50 μM , respectively ( S5 Fig ) . Higher oligomers than dimers are not observed . Three structures of KSHV assemblin ( KA ) with helical-peptide mimetics ( HPMs ) have been published to date with pdb entries 3njq , 4p2t , and 4p3h [43 , 44] . These HPMs disrupt dimer formation in full-length KA as shown by size-exclusion chromatography [45] . In solution , both C-terminal helices of the monomeric KA are unfolded according to NMR- and CD-spectroscopic data [28 , 43] . Therefore , 34 C-terminal residues of KA were truncated for crystallization of these HPM complexes . One of these truncated helices is the major dimer-interface helix , so this truncated KA is obligate monomeric . Consequently , the models of these HPM complexes were stated as monomeric [43 , 44] . Inspection of these models led us to the conclusion , that the HPMs function as an artificial dimer interface for this truncated KA . PDBePISA suggests dimers or higher association states for HPM complexes ( buried surface area of ~1 , 200 Å2 per monomer ) . The monomer is termed as unstable in solution , because the hydrophobic HPMs would be heavily solvent exposed . Compared to the native dimeric structure of full-length KA , one monomer is rotated approximately 80° around an axis roughly perpendicular to the dimer interface in the artificial , inactive HPM complexes of truncated KA ( S6 Fig ) . Although artificial , the HPM-interface underlines the importance of hydrophobic interfaces as suitable drug targets . The used HPMs were reported to bind to assemblin dimer interfaces of all herpesvirus classes . The IC50-values for these HPMs against the representative alphaherpesvirus assemblin ( HSV-2 assemblin ) were significantly higher than those against HCMV , EBV and KSHV assemblins , indicating much weaker binding to HSV-2 assemblin [44] . The C-terminal helices of alphaherpesvirus assemblins are likely ordered due to the hydrophobic key position being occupied by alternative conserved hydrophobic residues of the OHL . Thus , the HPMs have to compete against the C-terminal helices for binding which explains the observed higher IC50-values . Additionally , the most important residues for dimerization in KA , the so-called “hot spot” residues [65] , are Trp109 ( α4 ) [43] and maybe Phe76 ( α3 ) . These correspond to Tyr ( α4 ) and Leu ( α3 ) , respectively , in PrV , VZV , HSV-1 and HSV-2 assemblins ( S3 Fig ) . Neither Leu nor Tyr are considered as “hot spot” residues [65] and thus , binding of these HPMs is likely weakened . Screening for suitable mimetics is necessary to achieve specific and efficient inactivation of alphaherpesvirus proteases . As mentioned above the HPMs force an artificial assembly of truncated KA in contrast to our native monomeric PrV assemblin . Therefore , a detailed comparison of monomeric PrV assemblin and HPM complexes of truncated KA is discussed in the supplement ( S1 Text and S7–S10 Figs ) . A putative cation was found in both dimeric structures from PrV . A distorted octahedral arrangement of the coordinated water molecules with mean distances of 2 . 1 Å strongly suggests the presence of a cation rather than water . Since the crystallization buffers contain MgCl2 , it is highly probable that these cations are Mg2+ ions . For verification , divalent cations with higher electron density were tested in the crystallization procedure . Mn2+ was able to substitute for Mg2+ but the resulting crystals diffracted considerably less well . The best dataset that we managed to collect at a wavelength near the Mn absorption edge had a resolution of 3 Å , but no significant anomalous signal was detected and electron density at the putative Mg2+/Mn2+ position was too weak to unambiguously confirm the presence of a Mn2+ ion . Presumably , the presence of putative Mg2+ ions in the structures is a direct result of the crystallization conditions ( containing 400 mM MgCl2 ) and does not reflect functional Mg2+ binding by the native protein since the metal ions compensate negative charges from two adjacent dimers , rather than within one dimer ( S11 Fig ) . Thus , Mg2+ ions or divalent cations with related properties are presumably necessary for the observed crystal packing of dimeric pUL26N from PrV . For HSV-1 protease a temperature-sensitive ( ts ) phenotype was described by mutating Y30F and A48V [66] . Mutating the corresponding residues in PrV protease ( Y13F/A30V ) , however , did not result in the desired phenotype . Hence , these mutations may not sufficiently destabilize PrV pUL26N . In HSV-1 protease , it is highly probable that the result of these mutations is a destabilization of the region around helix α1 . This helix and surrounding loops stabilize the OHL in the active conformation . In the PrV assemblin Arg24 is located at the N-terminus of α2 . This residue stabilizes the α1 region , because its side chain forms a hydrogen bond to the peptide oxygen of Asp59 ( Fig 7 ) . This hydrogen bond is missing in HSV-1 assemblin , since the corresponding residue is Pro42 . Consequently , we propose a Y13F/R24P/A30V mutant for achieving the desired ts phenotype in PrV assemblin . If the additional R24P exchange completely inactivates the protease , an R24K variant could also be taken into account . The slightly shorter side chain of Lys may cause a varied hydrogen-bond network and a weaker hydrogen bond to the peptide oxygen of Asp59 . During capsid assembly , pUL26 accumulates in the nascent capsids because of its C-terminal scaffold-protein part . Thus , the local high concentration of the protease is promoting dimerization and autoproteolysis occurs to release the scaffold from capsids for DNA packaging . In comparison with its dimeric structure , the monomer of pUL26N reveals changes at the dimerization area , in line with allosteric changes of a loop forming the oxyanion hole . As previously anticipated , the core of the protease including the positions and orientation of the active-site residues remains unchanged , but the oxyanion hole is disrupted in the monomeric form [41 , 42] . The monomeric structure presented here is not truncated and does not contain any inhibitors . Thus , it constitutes the first reliable model for native monomeric structures of other assemblins . Dimerization induces the following allosteric events: helix α7 of a monomer interleaves between helices α3 and α7 of a second monomer moving these helices farther apart . At the same time , helix α7 is elongated by one turn at its C-terminus to the cost of one N-terminal turn of helix α8 . This allosteric process forces the OHL to shift to the vicinity of the active-site serine and builds a far-reaching network of hydrogen bonds with side chains of helix α8 and the polypeptide of strand β5 . Furthermore , residues 13–20 of the β1-α2 loop become ordered and form helix α1 in the dimer by getting involved in that network of hydrogen bonds . In this position , the OHL forms the oxyanion hole and activity of the protease is established . The extent of disorder at the dimerization area will vary in assemblins of different herpesviruses , but the general disorder-to-order mechanism of dimerization [28 , 43] will very likely hold for all assemblins . The molecular structure of dimeric pUL26N will help to engineer a temperature-sensitive phenotype of the PrV protease . A temperature-sensitive variant will be a powerful tool to observe subsequent steps of viral replication in a synchronous wave [19] . This will provide valuable data on kinetics for cleavage , packaging of the DNA , nuclear egress and intracellular trafficking of the virions . The structure of monomeric PrV assemblin is the paradigm for monomeric states of other assemblins , primarily from alphaherpesviruses . Detailed knowledge of this structure , conformational changes and sequence specific contacts upon dimerization are a rational basis for the development of drugs binding to the dimerization area and , thus , trapping the inactive monomeric state [45] . Inhibition of dimerization suppresses protease activity and therefore prevents the assembly of fully functional virus capsids . Small-angle X-ray scattering ( SAXS ) of PrV assemblin in solution revealed that dimerization increases with protein concentration and in the presence of MgCl2 . Dissociation constants in the micromolar range are comparable to those observed for other assemblins [26 , 27] . Divalent cations like Mg2+ or Mn2+ are required for crystallization of the dimeric PrV assemblin , because these cations support crystal packing by compensating negative charges of neighboring dimers . Since MgCl2 shifts the condition of equilibrium towards dimeric pUL26N , the concentration of the monomeric form is likely below the critical nucleation concentration resulting in crystals of the dimeric form only . Accordingly , crystallization of the monomeric assemblin requires absence of divalent cations . The monomeric fraction of ~0 . 3 is sufficient for nucleation and the monomer-dimer equilibrium provides a steady supply of monomeric pUL26N . The monomeric form is increasingly favored since crystallization of monomeric pUL26N decreases the concentration of pUL26N in solution . Similar cases with a minor monomeric fraction crystallizing from a monomer-dimer equilibrium were reported earlier [67] .
Full-length pUL26 cleaves itself at two positions , and therefore expression of the full-length protein leads to an inhomogeneous product that is unlikely to crystallize . Accordingly , cleavage was prevented by cloning a stop codon behind Gln224 . The resulting coding sequence contains the protease fraction of pUL26 ( pUL26N ) only . N-terminally ( His ) 6-tagged pUL26N was expressed using a pET28a+ vector in E . coli strain BL-21 ( DE3 ) . Cells were grown in LB medium to an OD600 of 0 . 5–0 . 8 at 37°C and then induced by addition of isopropyl β-D-1-thiogalactopyranoside to a final concentration of 1 mM . Cells were lysed by sonication . The protein was purified by performing immobilized metal-ion affinity chromatography using a Poros MC 20 column loaded with Ni2+ ions ( 0 . 5 M NaCl , 50 mM Tris/HCl pH 7 . 5 , 5% glycerol , eluted with a gradient of 0–250 mM imidazole ) . The protein was checked for its purity by SDS-PAGE and then concentrated to ~20 mg/ml . Aliquots were stored at −80°C . The purified protein was not tested for enzymatic activity . Crystals were obtained using the hanging-drop vapor diffusion method at 22°C . First crystals grew in drops containing 1 μl pUL26N concentrate and 1 μl precipitant solution ( 0 . 1 M Hepes pH 7 . 5 , 20% PEG 8 , 000 ) within several days . The quality and size of the crystals could be increased by optimization of the composition of the precipitant solution and the ratio of protein solution to precipitant solution . The morphology of the crystals changed from plate-shaped to needle-shaped when MgCl2 was used as an additive in the crystallization procedure . Plate-shaped crystals turned out to be monomeric pUL26N , whereas the dimer formed needle-shaped crystals . For crystallization with inhibitor , the concentrated protein solution was incubated with a final concentration of 5 mM diisopropyl fluorophosphate for 1 hour prior to crystallization . Best crystals of native dimeric , inhibited dimeric and monomeric pUL26N grew in drops containing 2 μl pUL26N concentrate and 1 μl precipitant solution . Precipitant solution for monomeric pUL26N consisted of 0 . 1 M Tris/HCl pH 8 , 8% PEG 8 , 000 and crystals grew within one week . Precipitant solutions for native dimeric and inhibited pUL26N consisted of 0 . 1 M Tris/HCl pH 8 , 14% PEG 8 , 000 , 0 . 4 M MgCl2 and 0 . 1 M Tris/HCl pH 8 , 20% PEG 8 , 000 , 0 . 2 M MgCl2 , respectively . All crystals were cryo-protected by soaking for 15 s in drops of precipitant solution with increasing amounts of PEG 400 . Final concentrations of PEG 400 were 25% ( monomer ) , 18% ( native dimer ) and 17% ( inhibited dimer ) . Cryo-protected crystals were flash-frozen in liquid nitrogen . Datasets were measured at 100 K with a PILATUS-6M detector at beamline BL14 . 1 , operated by the Helmholtz-Zentrum Berlin ( HZB ) at the BESSY II electron storage ring ( Berlin-Adlershof , Germany ) [68] . 1 , 800 images were collected at X-ray wavelength 0 . 91841 Å with an exposure time of 0 . 5 s and an oscillation range of 0 . 1° . All datasets were processed using XDS and Aimless [69–72] . Further details on data collection and processing are listed in Table 1 . For determination of the Matthews coefficient and solvent content of the unit cell the CCP4 Program suite ( version 6 . 4 . 0 ) was used [73–76] . The structures of monomeric and native dimeric pUL26N were solved via molecular replacement using Phaser [77] . The starting model for native dimeric pUL26N was the A-chain of pdb entry 1at3 ( the homologous protein from HSV-2 ) edited with Chainsaw [78 , 79] . The A-chain of native dimeric pUL26N ( pdb entry 4v07 ) was used as the starting model for monomeric pUL26N . Since both dimeric forms were isomorphous , the inhibited structure could be solved by refining the native structure against the dataset of the inhibited form . Cycles of model building and refinement were carried out using Coot ( version 0 . 7 . 1 ) and Refmac5 ( version 5 . 8 . 0073 ) , respectively [80–88] . Further details on refinement are listed in Table 1 . The oligomerization states were confirmed by PDBePISA [53] . All figures representing structural models were prepared using PyMOL version 1 . 7 . 1 . 3 [89] . The diffraction data and refined models of monomeric , native dimeric and inhibited dimeric pUL26N were deposited with the Protein Data Bank under entry codes 4v0t , 4v07 , and 4v08 , respectively . A preliminary dataset of the monomeric pUL26N at 2 . 5 Å resolution was deposited with entry code 4cx8 . SAXS data were recorded at beamline P12 of the EMBL outstation at PETRA III , DESY , Hamburg [90] , using a PILATUS 2M pixel detector , a sample-to-detector distance of 3 . 1 m and a wavelength of 1 . 24 Å . Solutions contained 0 . 5 M NaCl , 50 mM Tris pH 7 . 5 , 0 . 25 M imidazole , 5% glycerol , 50 mM urea and pUL26N as indicated . In all experiments the sample temperature was 283 K . Measurements covered the momentum transfer range 0 . 008 < s < 0 . 47 Å-1 ( s = 4π sin ( θ ) / λ , where 2θ is the scattering angle and λ is the X-ray wavelength ) . To monitor radiation damage , 20 successive 50 ms exposures of protein solutions were compared , revealing no significant change . The data were normalized to the intensity of the transmitted beam and radially averaged . Scattering of the buffer was subtracted and the difference curves were scaled to unity protein concentration ( 1 mg/ml ) . For further data analysis , version 2 . 6 . 0 of the ATSAS package was used [91] . Form factors were generated from the monomeric and dimeric crystallographic models by means of the program FFMAKER . For subsequent curve fitting , the program OLIGOMER [63] was used . An ab initio model corresponding to the highest protein concentration in the presence of MgCl2 was generated using the programs DAMMIF [92] , DAMAVER [93] and DAMMIN [64] via the PRIMUS interface [63] , in "slow" mode and without imposing particle symmetry . The scattering data , structural models and curve fittings of dimeric PrV pUL26N were deposited with the small-angle scattering biological data bank ( SASBDB ) with entry code SASDA58 [94] . PrV UL26: Gene ID 2952508 HSV-1 UL26: Gene ID 2703453 PrV UL26 . 5: Gene ID 2952525 HSV-1 UL26 . 5: Gene ID 2703454 PrV pUL26: Q83417 in UniProtKB , S21 . 001 in MEROPS [95] PrV pUL26N: Amino-acid residues 1–224 of Q83417 in UniProtKB , S21 . 001 in MEROPS [95] PrV pUL26C: Amino-acid residues 226–524 of Q83417 in UniProtKB PrV pUL26 . 5: Q83418 in UniProtKB KA: Q2HRB6 in UniProtKB , S21 . 006 in MEROPS [95] | Herpesviruses encode a unique serine protease , which is essential for herpesvirus capsid maturation and is therefore an interesting target for drug development . In solution , this protease exists in an equilibrium of an inactive monomeric and an active dimeric form . All currently available crystal structures of herpesvirus proteases represent complexes , particularly dimers . Here we show the first three-dimensional structure of the native monomeric form in addition to the native and the chemically inactivated dimeric form of the protease derived from the porcine herpesvirus pseudorabies virus . Comparison of the monomeric and dimeric form allows predictions on the structural changes that occur during dimerization and shed light onto the process of protease activation . These new crystal structures provide a rational base to develop drugs preventing dimerization and therefore impeding herpesvirus capsid maturation . Furthermore , it is likely that this mechanism is conserved throughout the herpesviruses . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Dimerization-Induced Allosteric Changes of the Oxyanion-Hole Loop Activate the Pseudorabies Virus Assemblin pUL26N, a Herpesvirus Serine Protease |
The search for an HIV-1 cure has been greatly hindered by the presence of a viral reservoir that persists despite antiretroviral therapy ( ART ) . Studies of HIV-1 latency in vivo are also complicated by the low proportion of latently infected cells in HIV-1 infected individuals . A number of models of HIV-1 latency have been developed to examine the signaling pathways and viral determinants of latency and reactivation . A primary cell model of HIV-1 latency , which incorporates the generation of primary central memory CD4 T cells ( TCM ) , full-length virus infection ( HIVNL4-3 ) and ART to suppress virus replication , was used to investigate the establishment of HIV latency using RNA-Seq . Initially , an investigation of host and viral gene expression in the resting and activated states of this model indicated that the resting condition was reflective of a latent state . Then , a comparison of the host transcriptome between the uninfected and latently infected conditions of this model identified 826 differentially expressed genes , many of which were related to p53 signaling . Inhibition of the transcriptional activity of p53 by pifithrin-α during HIV-1 infection reduced the ability of HIV-1 to be reactivated from its latent state by an unknown mechanism . In conclusion , this model may be used to screen latency reversing agents utilized in shock and kill approaches to cure HIV , to search for cellular markers of latency , and to understand the mechanisms by which HIV-1 establishes latency .
A major obstacle to the eradication of HIV-1 is the persistence of the latent viral reservoir . While antiretroviral therapy ( ART ) has been extremely effective at suppressing viral replication , it has not eradicated this reservoir [1] . Upon the removal of ART , HIV-1 emerges from the latent state and replicates to levels akin to an acute infection that leads to disease progression [2 , 3] . The low frequency of latently infected cells within the HIV-1 patient ( between 1 and 60 in 106 resting CD4 T cells ) complicates the study of this viral reservoir in vivo [4 , 5 , 6] . This has prompted the development of models of HIV-1 latency based on chronically infected cell lines and primary human CD4 T cells [7] . To obtain an accurate representation of HIV-1 latency in vivo , it is essential to fully characterize these different models . Transcriptome profiling by microarrays or RNA-Seq allows the simultaneous evaluation of transcriptional activity of the entire genome within a sample , thus providing a comprehensive analysis of the biological condition of the cell population at a given time . These technologies are becoming important for the study of HIV-1 latency , particularly in the search for biomarkers of HIV-1 latency [8 , 9] and for evaluating the effects of latency reversing agents [10 , 11 , 12 , 13] . Krishnan and Zeichner [14] , utilized cDNA spotted microarrays to compare gene expression in cell lines chronically infected with HIV-1 ( i . e . , ACH-2 , J1 . 1 , U1 ) and their parental uninfected lines to identify 32 genes that were consistently differentially regulated . The laboratory of Fabio Romerio used Agilent microarrays to profile latently infected and uninfected conditions from four donors using their primary CD4 T cell latency model , and a gene encoding a surface receptor , CD2 , was identified to be enriched in latently infected cells [9] . RNA-Seq is the current state-of-the-art technology with respect to transcriptomics and is thought to exhibit greater specificity and dynamic range than microarrays [15] . The first RNA-Seq study of a primary CD4 T cell model of latency incorporated a GFP expressing virus [16] . When samples from a single donor were profiled over time , a large number of genes were identified as dysregulated during the latent phase ( N = 227 ) and were associated with chemokine receptors , cytokine signaling , and general immune responses . We previously used RNA-Seq to profile latently infected and uninfected samples from 4 donors [17] from the first iteration of a cultured primary central memory CD4 T cell ( TCM ) model [18 , 19] . This study demonstrated that the defective vectors used to seed the latent reservoir were recombining to reconstitute actively replicating HIV-1 . This observation led to the revision of the cultured TCM model of HIV-1 latency by incorporating wild type virus ( HIVNL4-3 ) and ART to suppress virus replication [20] . The purpose of the present RNA-Seq study was primarily to identify mechanisms involved in the establishment of HIV-1 latency but also to characterize this modified cultured TCM model of latency [20] to confirm that the model reflects a latent state . Comparison of latently infected to uninfected cells identified differential expression of genes in the p53 signaling pathway . Treatment with pifithrin-α , an inhibitor of the transcriptional activity of p53 [21] , reduced the ability of latently infected cells to be reactivated in the cultured TCM model . The main finding from these results suggests a direct effect of p53 on the establishment and ability to reactivate the latent reservoir .
To dissect the viral and cellular status of the cultured TCM model of latency [20] , gene expression profiles were generated by RNA-Seq for a total of 16 samples from 4 different donors representing the following 4 conditions: uninfected ( UI ) , latently infected ( LI ) , uninfected and activated ( UIA ) , and latently infected and activated ( LIA ) . Briefly , naïve CD4 T cells from four HIV-1 negative individuals were isolated and activated with αCD3/αCD28 beads for Three days in conditions that generate central memory CD4 T cells ( Fig 1A ) [18] . After activation , cells were allowed to expand in the presence of IL-2 and infected by spinoculation at day 7 with HIV-1NL4-3 at a low MOI that rendered 3–7% of cells infected at day 10 ( Fig 1B , day 10 ) . Cells were then seeded for an additional three-day period in a 96 well ( round bottom ) plate to increase the efficiency of virus transmission ( Fig 1B , day 13 ) . At that time point , cells were diluted and cultured for 4 extra days in the presence of IL-2 and ART ( Raltegravir plus Nelfinavir ) . At day 17 , CD4 positive T cells were isolated using magnetic bead sorting . This strategy was chosen because productively infected cells downregulate CD4 expression on the cell surface due to the expression of the accessory genes Nef and Vpu ( Fig 1B , day 17 ) [22 , 23] . This procedure largely eliminates productively infected ( p24 positive ) cells , as well as CD4 negative cells present in the culture . Therefore , cells in the UI and LI conditions were CD4 positive cells that were not expressing detectable levels of viral antigens . UI and LI cells were further activated with αCD3/αCD28 beads in the presence of ART for 48 hours to generate cells for the UIA and LIA conditions . Initially , the resting and activated conditions were compared ( i . e . , UI vs . UIA and LI vs . LIA ) to validate the phenotype of the resting cells . When comparing RNA content between these conditions , transcriptional amplification had occurred ( S1 Fig ) . Traditional normalization procedures for transcriptomic data do not account for transcriptional amplification [24 , 25] . Therefore , Biological Scaling Normalization ( BSN ) using ERCC spike-in control RNAs was used to normalize RNA-Seq data to allow comparison of resting and activated conditions [26] ( See Materials and Methods ) . A number of gene expression markers of CD4 T cell activation were modulated following activation in both the LIA and UIA conditions ( Fig 2A ) . Notably , IL2 and components of its receptor ( IL2RA , IL2RB , IL2RG ) were upregulated upon activation , as were members of the NFκB complex ( NFKB1 , NFKB2 , REL , RELA , RELB ) , and CD28 itself . The KLF2 gene , which is highly upregulated in quiescent CD4 T cell lymphocytes , but repressed during activation [27 , 28] , was significantly downregulated as expected . The modulation of IL2 and KLF2 upon T cell activation was further confirmed by RT-qPCR ( Fig 2B ) . In summary , resting cultured TCM cells have the phenotypic characteristics of a quiescent T cell and stimulation with αCD3/αCD28 beads modulates known markers of CD4 T cell activation . Next , the effect of antigen stimulation mimicked by αCD3/αCD28 beads on HIV-1 transcription was evaluated . HIV-1 transcription in the resting ( LI ) and activated ( LIA ) states was compared after BSN . Treatment with αCD3/αCD28 beads induced global upregulation of total HIV-1 reads from the resting to the activated conditions ( average 6 . 6 fold change , s . d . ±3 . 6 , t-test p-value = 0 . 04 ) with an increase in all major splicing groups: unspliced ( US ) , singly spliced ( SS ) , and a significant increase ( p = 0 . 015 ) in multiply spliced ( MS ) ( Fig 3A ) . A significant increase of US , MS and total polyadenylated HIV-1 transcripts upon activation was confirmed by RT-qPCR ( Fig 3B ) with a concomitant increase in HIV-1 p24 protein ( Fig 3C ) . The fold change increase in polyadenylated transcripts appears much more variable than US or MS transcripts ( Fig 3B ) . It is unclear what is driving this variation but measurements of polyadenylated transcripts reflect fully elongated and correctly processed HIV-1 mRNA , which relies on the host transcriptional machinery . It is possible that the efficiency of polyadenylation varies across donors , whereas US and MS are less variable because they measure HIV transcripts , whether polyadenylated or not . In support of this , single nucleotide polymorphisms that vary between donors and effect post-transcriptional processing and subsequent gene expression have previously been identified [29] . In summary , examination of HIV-1 US , MS and polyadenylated transcripts further confirmed that the LIA condition of the cultured TCM model reflected activation of transcription from an HIV-1 latent state ( LI ) . The UI and LI samples were compared to identify 826 differentially expressed genes ( DEGs , S1 Table ) , 275 of which were downregulated and 551 upregulated ( false discovery rate [FDR] corrected p-value < 0 . 05 ) . The top 100 DEGs presented in the heatmap ( S2 Fig ) demonstrated the consistency of gene expression across donors . Although a large signal of differentially expressed genes was identified between the UI and LI conditions , it should be noted that this signal may be confounded by low proportions of latently infected cells and bystander effects . Specifically , a small proportion of latently infected cells in a background of uninfected cells is being compared to a population of 100% uninfected cells . The possible impact of this potential limitation is fully expanded upon in the Discussion section of this manuscript . The DEGs identified in the analysis were compared to other primary CD4 T cell [9 , 16] and cell line models of latency [14] . Although no up- or downregulated DEGs were identified in common across all models , ( Fig 4A & 4B ) , greater overlap was identified when comparing primary cell models in a pairwise fashion . In particular , there were 65 genes upregulated during latency in common between our cultured TCM model [20] and the model used by Iglesias-Ussel et al . [9] ( S2 Table ) . Unfortunately , the upregulation of CD2 during latent infection previously identified by Iglesias-Ussel et al . [9] was not confirmed in our TCM model . One explanation to this result is that in the model used by Iglesias-Ussel et al . [9] , the cells isolated to perform the analysis were expressing intracellular p24 . Our cultured TCM model of HIV-1 latency largely eliminates such cells . In order to develop a better understanding of genes perturbed during latency in the cultured TCM model , the 826 genes differentially expressed between the LI and UI conditions were subjected to pathway and protein interaction ( PIN ) analysis . The only Kyoto Encyclopedia of Genes and Genomes ( KEGG ) pathway that attained significance for over-representation of DEGs was the "p53 signaling pathway" ( FDR corrected p-value = 5 . 5E-06 ) . The results of differential gene expression were overlaid on the p53 signaling pathway and revealed that multiple threads related to apoptosis and DNA repair and damage prevention were upregulated in latently infected cells [30] ( Fig 5 ) . Of note , a number of p53 related genes ( ACTA2 , BBC3 , DDB2 , DRAM1 , FDXR , GADD45A , and TNFRSF10B ) were present in the 65 upregulated genes ( S2 Table ) in common between our study and that of Iglesias-Ussel et al . [9] . A PIN constructed using genes differentially expressed during latency complemented KEGG pathway analysis by confirming the importance of genes related to p53 activity ( Fig 6A and Table 1 ) . This PIN contained two major hubs ( AR and MDM2 ) and one minor hub ( TNFRSF10B a . k . a . DR5 or TRAIL-R2 ) , which are all related to p53 activity . For example , MDM2 facilitates negative feedback to the p53 signaling pathway by degrading p53 and the upregulation of this gene in the LI condition may be indicative of prior p53 activity [31] . AR is negatively regulated by p53 signaling [32] , and correspondingly , is downregulated in the LI condition . TNFRSF10B , a gene involved in programmed cell death , is upregulated and known to be induced by p53 in response to DNA damaging agents [33] . Therefore , the PIN extended KEGG pathway analysis by identifying p53 related hub genes ( e . g . AR ) and their targets that were not curated into the KEGG p53 signaling pathway . Several genes ( BBC3 , FAS , GADD45A , HEXIM1 , MDM2 , TNFRSF10B and TP53I3 ) selected from the p53 signaling pathway and the PIN were subjected to RT-qPCR analysis ( Fig 6B ) . The upregulation of transcripts of BBC3 , GADD45A , MDM2 , TP53I3 , and downregulation of HEXIM1 during latent infection was confirmed as significant by RT-qPCR . The direction of fold change was validated by RT-qPCR for the cell surface markers FAS and TNFRSF10B ( Fig 6B ) , which were previously noted as significantly upregulated in the RNA-Seq data ( Fig 6C ) . The cell surface expression of these markers was interrogated in an independent donor set by flow cytometry ( Fig 6D ) but only FAS ( CD95 ) was confirmed as significantly upregulated on the surface of T cells ( fold change of 1 . 47 ± 0 . 20 , Fig 6E ) . In summary , functional analysis of genes perturbed during latency using pathway and PIN analyses identified dysregulation of genes associated with p53 activity . Pifithrin-α , an inhibitor of p53 transcriptional activity [21] , was added to CD4 T cells from five additional donors after the initial infection at day 10 and again at day 13 in the cultured TCM model to investigate the possible role of p53 signaling in HIV-1 latency ( Fig 7A ) . Inhibition of p53 had no effect on HIV-1 replication since no difference was observed in the levels of p24 between treated and untreated samples at days 13 and 17 ( Fig 7B and 7C ) . However , inhibition of p53 transcriptional activity using pifithrin-α resulted in a reduction of the number of cells producing p24 ( average reduction of 32% , s . d . ±8% ) after reactivation with αCD3/αCD28 beads ( Fig 7B and 7D ) . Interestingly , when pifithrin-α was added only during the reactivation step ( day 17 ) , there was no effect on viral reactivation ( Fig 7E ) . These results suggest that inhibition of p53 during the active replication phase of HIV-1 may have an effect on the establishment or maintenance of HIV-1 latency . To test this hypothesis , integrated HIV-1 was analyzed by Alu-PCR in LI cells treated or not treated with pifithrin-α . There was a trend towards the reduction of integrated HIV-1 in pifithrin-α treated samples ( Fig 7F , p = 0 . 056 ) . However , the reduction on integration ( average reduction of 9% , s . d . ±5% ) does not fully explain the reduction in p24 observed after reactivation with αCD3/αCD28 beads ( Fig 7D , average reduction of 32% , s . d . ±8% ) . Therefore , it is possible that in addition to reducing the number of integrated copies of HIV that pifithrin-α may be further inactivating the integrated provirus , albeit through an unknown mechanism . To support this , integrated copies of the provirus were correlated with the percentage of p24 producing cells following reactivation independently for pifithrin-α treated and untreated cells ( Fig 7G ) . The correlation lines in this plot are parallel and shifted to the left for pifithrin-α treatment suggesting that the integrated virus in cells treated with pifithrin-α may be less prone to reactivation with αCD3/αCD28 beads . To compare both populations , we calculated the reactivation index measured as the percentage of cells expressing p24 after αCD3/αCD28 reactivation divided by the number of integrated copies before reactivation . This index compares the ability of an integrated copy to be reactivated by αCD3/αCD28 beads . Interestingly , cells that were treated with pifithrin-α have a lower ability to reactivate latent HIV-1 ( Fig 7H ) . In summary , these studies confirmed that p53 signaling may play an important role in the establishment and maintenance of HIV-1 latency in this cultured TCM model .
In this study , RNA-Seq was utilized to characterize a cultured TCM model of HIV-1 latency [20] , which incorporates a replication-competent virus ( HIV-1NL4-3 ) and ART to suppress HIV-1 replication . RNA-Seq analysis demonstrated that the resting condition in this model ( LI ) reflects a quiescent and a latent state when compared to the activated state ( LIA ) both at the level of host and virus transcription . Notably , an increase in HIV-1 transcription was observed in all donors after activation ( Fig 3A ) , indicative of a departure from a latent state [49] . The result of subtracting MS from US reads in the LI condition ( mean difference 4 , 134 , s . d . ±5448 ) was significantly different ( p = 0 . 03 , paired t-test ) than this subtraction in the LIA condition ( mean difference -10 , 928 , s . d . ±3 , 408 ) . This demonstrates a significant shift from US to MS reads upon activation suggesting an increase in early HIV-1 transcripts ( Vpr , Tat , Rev , and Nef ) after activation . The greater numbers of US versus MS reads in the LI condition of the TCM model is supported by previous reports that consistently detect a greater signal for US over MS transcripts in the resting CD4 T cells isolated from HIV-1 infected patients on ART [50 , 51] . The increase in HIV-1 transcription was not solely due to abortive transcripts since polyadenylated viral transcripts , which reflect fully elongated and correctly processed HIV-1 mRNA , were significantly upregulated after activation . By comparing expression profiles from UI to those from LI cells , a total of 826 differentially expressed genes were identified ( S2 Table ) . While there was only minimal overlap in genes dysregulated during latency when comparing this cultured TCM model to published data from three additional models of HIV-1 latency ( Fig 4A & 4B ) , there appeared to be greater overlap between primary CD4 T cell models compared to models based on cell lines ( S2 Table ) . Several reasons may account for these differences . First , this overlap might have been greater if the same transcriptomic technologies had been utilized , e . g . Iglesias-Ussel et al . [9] used the Agilent-012391 Whole Human Genome Oligo Microarray ( G4112A ) to profile gene expression compared to RNA-Seq in this study . Second , the RNA-Seq study of Mohammadi et al . [16] analyzed samples from only a single donor . Greater overlap between primary HIV-1 latency models may occur when better powered transcriptomic studies ( i . e . , multiple donors ) are performed using the state-of-the-art technology ( i . e . , RNA-Seq ) . Functional analysis of DEGs identified between the LI and UI conditions clearly implicated the transcriptional activation of genes by p53 ( Figs 5 and 6 , and Table 1 ) . For example , the genes ACTA2 , BBC3 ( a . k . a . Puma ) , DDB2 , DRAM1 , FAS , FDXR , GADD45A , PRDM1 , RHOC , TNFRSF10B , and TP53I3 are known to be induced by p53 [33 , 36 , 39 , 43 , 45 , 48 , 52 , 53 , 54] , and were upregulated in the LI condition , which is in agreement with previous studies showing the activation of the p53 pathway mediated by HIV-1 through type I IFN signaling [55 , 56 , 57 , 58] . Of these genes , FAS ( CD95 ) was further confirmed at the protein level ( Fig 6E ) . In addition to upregulated p53 related genes , activation of p53 by genotoxic stress has been shown to result in downregulation of AR expression [32] , also downregulated in this TCM model ( Fig 6A ) . The p53 protein itself is highly regulated and its activity must be tightly controlled to allow normal cellular functioning [59] . To maintain homeostasis , p53 will not only activate genes that enhance and stabilize its activity but also genes that repress its activity through negative feedback loops . Further evidence of prior p53 signaling activity was demonstrated by the upregulation of genes which act to degrade p53 following activation ( Table 1 ) . For instance , the protein product of MDM2 mediates ubiquitination and breakdown of p53 resulting in the inhibition of p53-mediated apoptosis [31] , and was upregulated in the LI condition ( fold change = 1 . 503 ) . Several other genes also involved in the degradation or inhibition of p53 [31 , 34 , 35 , 44 , 60 , 61] , were upregulated in the LI condition: BCL3 , CCNG1 , HEXIM1 , LIF , NR4A1 , and PTK2 . In summary , the latently infected ( LI ) condition in the TCM model of HIV-1 latency exhibited evidence of prior p53 activation and subsequent negative feedback of this signaling pathway . The identification of p53 related genes modulated during HIV-1 infection led to the hypothesis that this pathway may be important for the establishment of latency . Experiments with the p53 inhibitor pifithrin-α demonstrated that inhibition of this pathway did not affect viral replication or cell death , but did limit the number of cells that could be reactivated from latency ( S3 Fig , Fig 7C , 7D and 7E ) . A number of explanations can account for these results . First , Alu-PCR analysis suggested that p53 may be required for successful integration of HIV-1 ( Fig 7F ) . The p53 protein is not only involved in apoptosis and cell cycle arrest , but also in the activation of DNA repair mechanisms [62] . A number of p53-responsive genes identified in our study ( BBC3 , DDB2 , GADD45A , FDXR , PCNA and XPC ) are related to radiation-induced DNA damage [63] and point towards involvement of the nucleotide excision repair ( NER ) pathway , a process that recognizes and removes helix-distorting DNA lesions from the genome [64 , 65] . For example , GADD45A binds to UV-damaged DNA where it is believed to modify DNA accessibility within chromatin [66] . Although prior studies have primarily implicated other DNA repair pathways such as non-homologous end joining and base excision repair in HIV-1 DNA integration [67 , 68] , the results of the present study suggest the p53-responsive genes that are components of the NER pathway may also play a role in the establishment of latency . Interestingly , a previous siRNA screen to characterize DNA repair factors involved in HIV-1 integration demonstrated that siRNA knockdown of DDB2 , a component of the NER pathway , resulted in a large reduction ( 71 . 3% ) of HIV-1 integration [67] . A second hypothesis to explain the effects observed by pifithrin-α could be the silencing of the HIV-1 provirus . It should be noted , that despite approaching significance and in the same direction for each donor , the reduction in HIV integration by pifithrin-α is modest ( Fig 7F ) , and does not correspond entirely to the magnitude of the difference in p24 following reactivation ( Fig 7D ) . Therefore , in addition to reducing the establishment of latency , pifithrin-α may be inactivating the integrated provirus , albeit through unknown mechanisms . Such mechanisms might include a shift in integration sites , inactivated provirus , and/or silencing through alterations in DNA methylation or histone modification at the HIV promoter . To support this , correlating the number of integrated copies of the provirus with the percentage of p24 producing cells following reactivation independently for pifithrin-α treated and untreated cells demonstrates parallel lines but shifted to the left for pifithrin-α treated cells ( Fig 7G ) . Furthermore , there was a significant drop in the reactivation index between pifithrin-α treated and untreated cells ( Fig 7H ) . These data suggest that pifithrin-α may induce changes to the integrated provirus , either directly or at the level of the epigenome , resulting in less efficient reactivation with αCD3/αCD28 beads . This is not without precedent since it has been demonstrated that HIV DNA synthesis by the virus reverse transcriptase is more accurate in the presence of p53 , which has exonucleolytic proofreading capabilities [69] . Therefore , inhibition of pifithrin-α in the TCM model of HIV-1 latency may lead to more error prone HIV DNA synthesis during the expansion phase resulting in greater numbers of dysfunctional provirus . Further studies of this pathway will be needed to completely understand the role of p53 in the establishment of HIV-1 latency in cultured TCM . The importance of p53 signaling may not be confined to only the TCM model analyzed here as several genes related to the p53 pathway were also identified in the overlap with the primary CD4 T cell model examined by Iglesias-Ussel and colleagues [9] . Specifically , the p53 related genes ACTA2 , BBC3 , DDB2 , DRAM1 , FDXR , GADD45A , and TNFRSF10B were significantly upregulated in both models ( S2 Table ) . It will be interesting to determine if inhibiting the p53 pathway in other models of HIV latency also affects the establishment of latency . This may suggest common mechanisms involved across primary CD4 T cell models . Finally , the contribution of p53 to the establishment of latency in vivo needs to be evaluated . Interestingly , Castedo and colleagues [70] have previously shown that activation of p53 can be detected in HIV-1 infected patients in both peripheral blood mononuclear cells ( PBMCs ) as well lymph nodes . Moreover , the authors demonstrated that p53 activation correlates with viral load . These results suggest that p53 may also play a role in the establishment of latency in vivo . The present study allowed the comparison of gene expression between the LI and UI conditions of this TCM model of HIV-1 latency , which may represent potential biomarkers of latency . However , the search for these biomarkers in this study was somewhat limited by the relatively low proportion of latently infected cells in the LI condition ( mean 2 . 92% , s . d . ±0 . 71% ) that were in a large background of uninfected cells . Therefore , a gene must be upregulated greater than 30-fold in individual latently infected cells in order for it to be identified as being upregulated by 2-fold when comparing the LI and UI conditions . Furthermore , bystander effects could be contributing to the signal of differential gene expression between the LI and UI conditions , whereby signals emanating from previously infected cells ( e . g . , cytokines and chemokines ) may be perturbing gene expression in uninfected cells in the LI condition . Similarly , it is plausible that these results may be confounded by the triggering of innate immune pathways in TCM cells exposed but not latently infected by HIV particles . However , robust triggering of innate immune responses is unlikely , since only 11 interferon stimulated genes with known antiviral properties [71] were found to be differentially expressed between the LI and UI conditions , out of a total of 826 genes , and only 2 of these 11 genes were in the top 100 DEGs ( S2 Fig ) . Therefore , it appears that TCM cells in this model were cultured for a sufficient period of time following virus exposure to allow innate immune responses to recede . Regardless of these limitations , gene expression markers of HIV-1 latency undoubtedly exist within the 826 DEGs identified between the UI and LI conditions ( S1 Table ) and will need to be verified in future work . In summary , the primary finding from this RNA-Seq analysis of the cultured TCM model [20] was that p53 signaling may be involved in the establishment of HIV latency . Furthermore , the RNA-Seq data was used to demonstrate that this model was truly reflective of a latent state and thus suitable for screening latency reversing agents for shock and kill approaches to an HIV-1 cure [72 , 73] , further investigating mechanisms associated with establishing a latent state , and identifying gene expression biomarkers of HIV-1 latency . It would be beneficial to subject all primary CD4 T cell models of HIV-1 latency [7] to RNA-Seq analysis in statistically powered studies ( i . e . >3 donors ) so that similarities and differences across models may be dissected . Finding similar genes across models will lead to the identification of gene expression biomarkers of HIV-1 latency that may be used to isolate latently infected cells from HIV-1 infected subjects and utilized in innovative cure strategies ( e . g . , radioimmunotherapy [74] ) or killing latently infected cells by means of immunotoxins [75] . The profiling of other HIV-1 latency models on the omics scale will lead to the validation or modification of these models , which will undoubtedly result in a better understanding of both in vivo latency and the therapies that can be used to facilitate a cure for HIV .
Nelfinavir was obtained through the AIDS Research and Reference Reagent Program , Division of AIDS , NIAID , NIH; raltegravir from Merck & Company , Inc . ; human IL-2 from Dr . Maurice Gately , Hoffmann-La Roche Inc . [76]; HIV-1NL4-3 from Dr . Malcolm Martin [77] . Pifithrin-α , p-Nitro , Cyclic [78 , 79] , a more potent analogue of pifithrin-α with a longer half-life , was obtained from Santa Cruz Biotechnology . Sample preparation and generation of infected TCM cells was fully described previously [20] . Briefly , PBMCs were isolated from HIV-1 negative individuals or obtained from the Gulf Coast Regional Blood Center ( Houston , TX ) . Naïve CD4 T cells were isolated and then activated using human αCD3/αCD28-coated magnetic beads ( one bead per cell , Thermo Fisher Scientific , Cat . #11131D ) in the presence of αIL-4 , ( 1 μg/mL; Peprotech; Cat . No . 500-P24 ) αIL-12 ( 2 μg/mL; Peprotech , Cat No . 500-P154G ) , and tumor growth factor ( TGF ) -β1 ( 10 μg/mL; Peprotech; Cat . No . 100–21 ) for 3 days at a density of 500 , 000 cells/ml in 96 well round bottom plates . Cells were expanded in medium containing human IL-2 ( 30 IU/ml ) for additional 4 days . IL-2 and media were replaced at day 4 and day 5 . At day 7 , cells were infected ( or mock-infected ) with HIV-1NL4-3 by spinoculation at 2900 rpm at 37C for 2 hours at a multiplicity of infection of 0 . 1 . After infection , cells were further cultured in medium containing IL-2 for 3 days , subjected to crowding in round bottom plates in the presence of IL-2 for another 3 days , and then cultured for a further 4 days in the presence of IL-2 and ART in a cultured flask ( nelfinavir , 0 . 5 μM; raltegravir , 1 . 0 μM ) . Every time that media and IL-2 were replaced , cells were kept at a density of 106 cell/ml . At day 17 , any remaining productively infected cells were removed by magnetic isolation of CD4+ cells using the Dynabeads CD4 Positive Isolation Kit ( Thermo Fisher Scientific , Cat . No . #11551D ) following the manufacturer’s instructions with a minor change , i . e . 75 μl per 107 cells was used instead of 25 μl to increase the recovery of CD4 positive cells . CD4 beads were removed from the cells following the manufacture instructions . At this stage , samples were taken for the latently infected ( LI ) condition . Uninfected ( UI ) cells were cultured under the same conditions and collected at the same time as LI cells . Additional cell aliquots were subjected to reactivation with αCD3/αCD28-coated magnetic beads in the presence of ART for 2 days for the uninfected activated ( UIA ) and latently infected activated ( LIA ) conditions . To address whether p53 transcriptional activation plays a role in the establishment of latency , 7 . 5 μM pifithrin-α was added 3 days post infection during the crowding stage ( day 10 ) , and then washed out with replenishment at day 13 and thus maintained in culture until just prior to reactivation ( day 17 ) when it was washed out . Production of p24 was then assessed at day 19 after two days of reactivation with αCD3/αCD28-coated magnetic beads . In a separate experiment , to confirm that pifithrin-α was not affecting the reactivation process , pifithrin-α was added only during the reactivation step ( day 17 to 19 ) and p24 production similarly assessed at day 19 . Total RNA was extracted from 16 TCM cell samples from 4 donors for the 4 conditions ( UI , LI , UIA and LIA ) using the RNeasy Plus Mini Kit ( QIAGEN , Cat . No . 74134 ) according to manufacturer’s instructions , with the addition of an on-column DNase treatment ( RNase-free DNase Set , QIAGEN , Cat . No . 79254 ) . RNA integrity ( RIN ) values of samples were on average 9 . 9 ( s . d . ±0 . 1 ) as determined using a Bioanalyzer 2100 ( Agilent Technologies , CA , USA ) and RNA concentration was measured by Nanodrop 2000 ( Thermo Fisher Scientific ) . To account for transcriptional amplification , synthetic RNA standards from the External RNA Controls Consortium ( ERCC RNA Spike-In Mix 1 , Ambion , CA , USA ) were spiked into total RNA isolated from each sample . TCM cells in each sample were counted in quadruplicate and ERCC spike-ins were added at 1 μl of 1:100 dilution per million cells . RNA-Seq libraries were prepared from 100ng of total RNA using the TruSeq Stranded Total RNA Library Prep kit ( Illumina , CA , USA ) after depletion of cytoplasmic and mitochondrial ribosomal RNA with Ribo-Zero Gold ( Epicentre , WI , USA ) . All libraries were sequenced to a read depth of >75 million reads using the Illumina HiSeq2000 to generate 50bp paired-end reads ( 100 bp total read length ) . FASTQ files for each sample were mapped to the human genome ( hg38 ) using Tophat ( version 2 . 0 . 13 ) [80] and counted against the human GENCODE [81] annotation ( v21 ) with HTSeq [82] . FASTQ files are available through GEO accession number ( GSE81810 ) . All reads were then aligned to the HIVNL4-3 genome ( Genbank accession number AF324493 ) using TopHat and counted using HTSeq [82] . Levels of HIV-1 US , SS , and MS transcripts were estimated by the method of Mohammadi and colleagues [16] which counts reads that pass through the two major HIV-1 splice sites D1 and D4 . Finally , all reads were mapped to the 92 ERCCs , using Bowtie ( version 1 . 1 . 1 ) [83] and then counted against individual ERCCs using HTSeq . When identifying differences in host and HIV-1 gene expression between resting ( UI and LI ) and activated ( UIA and LIA ) conditions , BSN was utilized to account for transcriptional activation [26] . Briefly , the expression of ERCC spike-in controls were used to estimate a scaling factor between resting and activated conditions which was then used to adjust the expression levels of host genes and HIV-1 reads . When identifying differences in host gene expression between the UI and LI conditions , RUVSeq [84] was used to normalize the data since transcriptional amplification was not an issue in this comparison . Following normalization , differentially expressed genes were identified with EdgeR [85] ( FDR corrected p-value < 0 . 05 ) . Please refer to ( S1 Supplemental Methods ) for more details . Pathway analysis was performed with ToppGene [86] using the functional analysis enrichment tool , ToppFun , with the KEGG pathways selected . Pathway images were generated from the KEGG Pathway Database [30] . Fold change data were log2 transformed , colored , and overlaid upon the p53 signaling pathway . A protein interaction network ( PIN ) was generated with MetaCore and visualized through Cytoscape v2 . 8 . 3 [87] . MetaCore draws connections between the protein products of differentially expressed genes if they have protein-protein or protein-DNA interaction confirmed from the literature record . The advantage of this approach is that the PIN often reveals biological associations that have not been curated in KEGG pathways . For PIN construction , genes were filtered using a log2 fold change of 0 . 5 between latently infected and uninfected cells . Read pileup figures were generated with the Integrated Genome Browser [88] . Venn diagrams were constructed using Venny 2 . 1 . 0 [89] . RT-qPCR validation of the expression of host and virus genes identified by RNA-Seq was performed using TaqMan Gene Expression Assays ( Thermo Fisher Scientific ) as previously described [90 , 91 , 92] . Changes in host and virus gene expression were calculated using the 2-ΔΔCT method with the spike-in ERCC control ERCC_00130 , as the normalizer . Please refer to ( S1 Supplemental Methods ) for more details on RT-qPCR analysis . For the detection of surface FAS ( CD95 ) expression , cells were stained with FAS/CD95 Antibody ( DX2 ) , FITC conjugate ( Molecular Probes ) . For the detection of surface TNFRS10B ( DR5 ) expression , cells were stained with CD262 ( DR5 , TRAIL-R2 ) Antibody ( DJR2-4 ( 7–8 ) ) , APC conjugate ( Biolegend ) . Cells were also stained with a viability dye ( Fixable Viability Dye eFluor 450 , Affymetrix , eBioscience , San Diego , CA ) . For the dual detection of CD4 and HIV-1 p24 Gag , cells were first stained with the viability dye ( Fixable Viability Dye eFluor 450 ) , followed by staining with CD4 antibody ( S3 . 5 ) , APC conjugate ( Molecular Probes ) . After staining , cells were fixed , permeabilized , and stained for HIV-1 p24 Gag as previously described [20] . In all experiments , CD4 positive HIV-1 p24 Gag negative staining regions were set with uninfected cells treated in parallel . Flow cytometry was performed with a BD FacsCanto II flow cytometer using FACSDiva acquisition software ( Becton Dickinson , Mountain View , CA ) . Data were analyzed with Flow Jo ( TreeStar Inc , Ashland , OR ) . DNA from 2x106 cells was isolated using DNeasy Blood and Tissue Kit ( Qiagen ) . DNA was quantified using NanoDrop 1000 ( Thermo Fisher Scientific ) . Genomic DNA was subjected to nested quantitative Alu-LTR PCR for integrated provirus as previously described [93] , with modifications . For the first reaction , 250 ng of total DNA was amplified using Platinum Taq DNA polymerase ( Invitrogen ) . Reactions were carried with 1 . 5 mM of MgCl2 , 200 μM dNTPs , 400 nM of Alu164 primer ( 5’-TCCCAGCTACTCGGGGAGGCTGAGG-3’ ) and 400 nM of PBS primer ( 5’-TTTCAAGTCCCTGTTCGGGCGCCA-3’ ) . Amplifications were performed in a MultiGene Optimax ( Labnet International , Inc ) with the following parameters: 1 ) 94C 5 min; 2 ) 18x 94C 30 sec , 66C 30 sec , 72C 5 min; 3 ) 72C 10 min . PCR samples were subject to a 1/10 dilution in water , then 2 μl of the diluted sample was subject to qPCR reactions in a LightCycler 480 ( Roche ) using PCR Master Mix ( 2X ) ( Thermo Fisher Scientific ) . Final concentration of primers ( AE989-2 5’-CTCTGGCTAACTAGGGAACCCAC-3’; AE990-2 5’-CTGACTAAAAGGGTCTGAGGGATCTC-3’ ) and probe ( 5’-FAM-TTAAGCCTCAATAAAGCTTGCCTTGAGTGC-BHQ1-3’ ) were 400 nM and 200 nM respectively . A serial dilution of pcDNA3 . 1-LTR was used for a molecular standard curve . pcDNA3 . 1-LTR was generated by cloning the 3’ LTR from NL43 into pcDNA3 . 1 . PBMCs were isolated from HIV-1 negative individuals following IRB-approved protocol no . #67637 ( University of Utah ) or unidentified source leukocytes designed “for research only” were purchased from the Gulf Coast Regional Blood Center ( Houston , Texas ) . All HIV-1 negative individuals provided written informed consent . | The major hindrance to an HIV cure is the ability of the virus to persist in a latent state despite antiretroviral therapy . It is difficult to study this latent state in the HIV-infected patient because only a small proportion of cells in the body are affected and current technologies are not able to identify these cells . Therefore , models in the laboratory have been developed to study HIV latency . However , these models have not been adequately characterized with the latest genomic technologies . We have characterized our model of HIV latency using global gene expression analysis ( i . e . , RNA-Seq ) . Our model aims to reflect HIV latency in patients by using primary central memory CD4 T cells , wild type virus , and antiretroviral therapy . Our main finding was that signaling through the p53 protein characterized the latent state , and may be important in its establishment . This has implications for a better understanding of HIV latency which may lead to new therapies . In a broader context , we validated the latent state of our model of HIV latency , which can now be used with confidence to evaluate compounds used in strategies to cure HIV , search for markers of HIV latency , and further investigate the mechanisms leading to the establishment of latency . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"blood",
"cells",
"t",
"helper",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"antiviral",
"therapy",
"pathogens",
"immunology",
"microbiology",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses... | 2016 | Transcriptomic Analysis Implicates the p53 Signaling Pathway in the Establishment of HIV-1 Latency in Central Memory CD4 T Cells in an In Vitro Model |
Schistosomiasis is among the most neglected tropical diseases , since its mode of spreading tends to limit the contamination to people who are in contact with contaminated waters in endemic countries . Here we report the in vitro and in vivo anti-schistosomal activities of trioxaquines . These hybrid molecules are highly active on the larval forms of the worms and exhibit different modes of action , not only the alkylation of heme . The synergy observed with praziquantel on infected mice is in favor of the development of these trioxaquines as potential anti-schistosomal agents .
Malaria and schistosomiasis are the two most important parasitic diseases in tropical and sub-tropical areas . The parasite species responsible of these diseases are quite different: Plasmodium is an intracellular protozoa while Schistosoma is a metazoan worm . However these parasites share a common feature , they are both hematophagous . During their development in human blood stream , they digest a large quantity of host hemoglobin . As a consequence of the proteolytic digestion of this heme-containing protein , free heme is released and constitutes a major threat for both parasites due to its easy reduction by endogenous electron sources . The iron chelated by the protoporphyrin-IX ligand of heme is particularly efficient in oxygen reduction by monoelectronic transfer . This catalytic dioxygen reduction is at the origin of highly toxic reactive oxygen species ( ROS ) . Despite their high phylogenetic divergence , convergent evolution has conducted these two parasite species to use a similar heme detoxification pathway . The hemozoin pigment , known as malaria pigment in Plasmodium , is an aggregation of heme dimers turning the iron inactive . Hemozoin is a dark-black inert crystalline pigment , which is structurally identical in Plasmodium and Schistosoma [1] , [2] . The treatment and control of schistosomiasis currently rely on the use of a single drug , the praziquantel ( PZQ , Figure 1 ) . Praziquantel , a safe and effective drug , has been used for the last fourty years . However , several schistosome strains with lower sensitivity to praziquantel with possibility of resistance have been identified in African countries [3] , [4] . Having a single drug to treat a disease that affects hundred millions of people is a real concern , due to the possible resistance of the parasite to this drug . As a consequence , in the last ten years important efforts have been made in either developing new drug series [5]–[7a , b] , or testing existing drugs originally used on non-related diseases [8]–[13] . Among the existing medications , the antimalarial drugs targeting heme ( i . e . before the hemozoin formation ) are particularly interesting since free heme is not present in non-infected persons . In this field , two major series of molecules can be considered according their mechanism of action: trioxane-based molecules , that are heme-alkylating agents [14]–[17] , and aminoquinoline-based molecules , that are heme-stacking agents [18] , [19] . Trioxane-based molecules , whether naturally extracted or chemically synthesized , have shown moderate anti-schistosomal activities [10]–[11] . Similarly , aminoquinoline derivatives have been shown to be efficient against schistosomes experimentally infected animals , e . g . the treatment with mefloquine ( MFQ ) significantly reduced the number of eggs [20] . Because schistosomiasis and malaria are co-endemic in several countries , using an anti-malarial molecule against schistosomiasis might select drug-resistance in malaria parasites [8] , [9] . However , many malaria patients treated with artemisinin-based combination therapy ( ACT ) are indeed co-infected with schistosomes . In fact , a study carried out in Côte-d'Ivoire evidenced that children infected with S . haematobium , treated with mefloquine-artesunate administered in accordance with the currently recommended malaria treatment schedule , showed significantly higher egg reduction rates compared to children treated with artesunate ( ARTS ) or mefloquine alone [21] . The future challenge in the treatment of schistosomiasis may be not to use native anti-malarial drugs but to develop new drugs considering the mechanism of action of the anti-malarial molecules targeting free heme . In these conditions , it might be useful to develop an anti-schistosomal peroxide-based drug that will be also active against malaria parasites , with the requirement that the drug should not easily induce the selection of drug-resistant strains of Plasmodium . This strategy should provide new molecules active on both parasites with limited side effects . Trioxaquines ( TXQ ) are hybrid drugs containing two pharmacophores within a single molecule: a 1 , 2 , 4-trioxane and a 4-aminoquinoline [22] . Initially developed against malaria , they exhibit a dual mode of action: alkylation of heme with the trioxane entity , and stacking with heme due to the aminoquinoline moiety , leading to inhibition of hemozoin formation in vitro [23]–[25] . As reported for artemisinin derivatives [14] , [15] , owing to their trioxane entity , trioxaquines are indeed efficiently activated by iron ( II ) -heme , leading to the formation of covalent heme-drug adducts detected in the spleen of malaria infected mice [16] , [24] . Because of the relationship of 1 , 2 , 4-trioxane-containing drugs with heme metabolism , we have decided to evaluate the in vitro activity of trioxaquines on Schistosoma mansoni . Several of these molecules were found highly active on both larval and mature stages of S . mansoni [12] . A better understanding of the molecular bases of the antischistosomal activity of trioxaquines is requested to design new active drugs , and to optimize the existing drug candidates . As a confirmation that heme is a general target of drugs active against blood-feeding parasites , we recently reported that trioxaquine PA1259 alkylates heme in female adult S . mansoni , and heme-drug adducts were identified from treated worms [26] ( for the structure of PA1259 , see Figure 1 ) . Although important , this feature is probably not the only mode of action of trioxaquines in schistosomes . So , we decided to further evaluate their reactivity toward hemozoin , and also to attempt to have a general picture of damages induced by trioxaquines in S . mansoni . For this purpose , the action of a trioxaquine “prototype” , PA1259 , was compared to that of three other drugs: the reference drug praziquantel ( PZQ ) , artemether ( ARTM ) and mefloquine ( MFQ ) ( Figure 1 ) .
The laboratory has received the permit N° A 66040 for experiments on animals from both French “Ministère de l'Agriculture et de la Pêche” and French “Ministère de l'Enseignement supérieur et de la Recherche” . Housing , breeding and animal care of the mice followed the ethical requirements of our country . The experimenter possesses the official certificate for animal experimentation delivered by both ministeries ( décret n° 87–848 of October 19th 1987; authorization no 007083 ) . Animal experimentation follows the guidelines of the French CNRS . The different protocols used in this study have been validated by the French veterinary agency . Before parasite infection , mice were anaesthetized by injection of 0 . 1 mL/10 g of body weight of a mixture of Rompun ( 0 . 5 mL , 20 mg/mL; Bayer ) and Imalgène ( 1 . 0 mL , 100 mg/mL; Rhône Mérieux ) in 8 . 5 mL of autoclaved NaCl 8 . 5 ( o/oo ) . The host-parasite system used was an albino variety of Biomphalaria glabrata from Brazil and a strain of Schistosoma mansoni , also from Brazil , maintained in Swiss CD1 mice ( Depré , Bourges , France ) . Detailed methods for mollusc and mouse infections and for parasite recovery were previously described [27] . In vitro tests were performed on both free larval ( cercariae ) and parasitic stages ( schistosomules and adult worms ) . Cercariae were recovered in spring water under binocular microscope . Parasitic stages were recovered after percutaneous infection of mice using either 120 or 400 parasite cercariae . Mice exposed to 400 cercariae were sacrificed at 21 days after infection for schistosomule recovery , while mice exposed to 120 cercariae were sacrificed at 49 days after infection for adult recovery . Schistosomules or adult worms , freshly recovered , were washed and placed in RPMI 1640 medium ( supplemented with L-glutamine and Hepes 25 mM ) and store in incubator chamber at 37°C . Ten to 20 freshly recovered 21day-schistosomules or 49day-adults were placed in 24-well or in 6-well Falcon plate containing 1 mL or 3 mL of RPMI 1640 medium ( supplemented with L-glutamine and Hepes 25 mM ) , respectively . Fifty cercariae were placed in 24-well Falcon plate containing 1 mL of spring water . The drugs PA1259 , PZQ , ART or MFQ were first dissolved in DMSO to give mother solutions at 100 mg/mL . All further dilutions were done in DMSO except the last one that was realized in RPMI 1640 or spring water , for parasitic stage or free larval stage , respectively . The dilution was complemented with Tween 80 in order to obtain this final ratio dilution: culture medium/Tween80/DMSO , 1000/0 . 95/3 . 8 , v/v/v . These drug solutions were added to Falcon plate that contained worms . The S . mansoni cultures were then incubated with each drug at final concentration of 5 or 50 µg/mL for larval stages ( cercariae or schistosomules ) or adult stage , respectively ( a concentration of 5 µg/mL corresponds to 10 , 16 , 17 , and 13 µM for PA1259 , PZQ , ARTM , and MFQ , respectively ) . Control worms were treated with the same culture medium/Tween 80/DMSO , but without drug . Each test was performed in duplicate . Every 30 minutes , moving worms were counted , in order to define the percentage of survivors . Observation was extended to 8 h . Kaplan-Meier survival analyses followed by pairwise log-rank tests were used to compare survival data . Parasites showing no body contractions during a 30-s observation may be considered dead . Mice were infected percutaneously with 120 cercariae each . Twenty-one days ( schistosomule stage ) or 49 days ( adult worm stage ) post-infection , groups of five mice were treated orally . In monotherapy , PZQ or PA1259 oral treatment was performed at 100 mg/kg/d for five consecutive days , or at four doses of 50 mg/kg each , given every three hours ( overall treatment period: 9 h ) . For bitherapy evaluation , treatments were made at the following PZQ/PA1647 wt/wt ratio: 100/0 , 75/25 , 50/50 , 25/75 , and 0/100 . Administration consisted in 4 oral doses of 50 mg/kg each , given every three hours ( total drug dose: 200 mg/kg ) . In all cases , control mice were treated with solvent but without drug . Fifteen days after treatment , mice were killed and worms were recovered by retrograde perfusion . The viscera were observed to count the worms in each mouse . As a preliminary , the in vitro concentration of PZQ or PA1259 that kills half of the parasites after one hour of incubation ( LC50 ) was determined ( 30-secondes immobilized worms were considered as killed , since no worms recovered an activity after that 30-sec period of immobilization ) . For this purpose , the following ranges of concentrations were tested: 0 , 0 . 01 , 0 . 1 , 0 . 5 , 1 , 2 . 5 , 5 µg/mL for PZQ , and 0 , 5 , 10 , 20 , 30 , 40 , 50 µg/mL for PA1259 . The incubation medium and drug solvents were as described above for in vitro treatment of schistosomes . LC50 values found were 50 µg/mL and 0 . 075 µg/mL for PA1259 and PZQ respectively . No difference was observed between male and female worms . PA1259 and PA1019 , synthetized as reported in the patent application WO/2007/144487 , and PA1647 were provided by Palumed . PA1647 is the diphosphate salt of PA1259 ( stoichiometry checked by CHN-elemental analysis ) . Artemether ( ARTM ) was a gift from Rhône-Poulenc Rohrer Doma ( Antony , France ) . Praziquantel ( PZQ ) and mefloquine ( MFQ ) were purchased from Sigma-Aldrich , as well as sodium dithionite ( Na2S2O4 ) , dimethylsulfoxyde ( DMSO , ACS spectrophotometric grade ≥99 . 9% ) , pyridine ( ≥99% ) , and methanol ( Chromasolv ≥99 . 9% ) . Formic acid ( 99+% ) was from ACROS . Hemin ( ferriprotoporphyrin IX chloride , 98 . 0% ) from bovine blood was purchased from Fluka ( Switzerland ) . All chemicals and solvents were used as purchased without further purification . Milli-Q water ( resistivity ≥18 . 2 MΩ ) was used for LC-MS eluent preparation . Adult S . mansoni females were recovered in mice seven weeks after infection , and maintained in culture in RPMI 1640 medium , supplemented with L-glutamine and Hepes 25 mM , at 37°C . Groups of twenty-five schistosomes , freshly recovered , were carefully washed with RPMI 1640 medium , then treated with PA1259 or ARTM at 50 µg/mL for 3 hours . Then , worms were crushed with sand , and the obtained powder was extracted with pyridine ( 500 µL ) . The mixture was vigorously stirred ( vortex ) , placed in an ultrasonic bath for 30 min and , finally , magnetically stirred at 37°C overnight . After centrifugation at 4000 rpm for 30 min , the supernatant was withdrawn and filtered through ptfe 0 . 45 µm syringe filters before analysis by HPLC or LC-MS .
The in vitro activities of PA1259 ( ▪ ) , praziquantel ( ★ ) , artemether ( ⧫ ) , and mefloquine ( • ) on cercaria- , 21 day-old schistosomules , and 49 day-old adult worms are reported in Figure 2A , 2B , and 2C , respectively , as Kaplan-Meier plots . When treated with PZQ at 5 µg/mL , after 4 h of contact with the drug , all cercaria were immobilized for a 30 s-observation time . To obtain such immobility after treatment with PA1259 or MFQ required only 60 or 90 min , respectively . In the presence of ARTM , 8 h after treatment , more than 80% of cercariae were still moving ( Figure 2A ) . The treatment of 21-day schistosomules with PZQ at 5 µg/mL resulted in a complete immobilization of all larvae after 3 h of contact . With PA1259 or MFQ , the same effect was obtained after 5 h and 8 h , respectively . When ARTM was used in the same conditions , no significant effect was observed after 8 h ( Figure 2B ) . The treatment of 49-day adult schistosomes with drugs at 50 µg/mL resulted in complete immobilization of all worms in 2 h in the presence of PZQ , and 3 h in the presence of MFQ or PA1259 . With ARTM at the same concentration , more than 60% of worms were still moving after 8 h ( Figure 2C ) . In addition , the activity of PA1019 , which is the 4-aminoquinoline residue contained in PA1259 , was evaluated on schistosomules and adult worms . Seven hours of contact of PA1019 was required to immobilize only 27% of schistosomules and 79% of adult worms . In mice infected by S . mansoni , the reduction of worm burden upon oral treatment by PA1259 or PZQ is reported in Table 1 . In mice treated with PA1259 at five daily doses of 100 mg/kg starting from day 21 post-infection , 26 . 6±5 . 4 worms were collected 15 days later . When treatment was done at day 49 post-infection , 21 . 0±7 . 0 worms were collected . These values correspond to a reduction of 31% and 42% on larval and adult stage , respectively , with respect to mice treated by excipient alone ( control mice ) . With the same treatment schedule , PZQ induced a worm burden reduction of 20% and 79% on schistosomules and adult worms , respectively . With four doses of 50 mg/kg of PA1259 given every three hours , the reduction of the worm burden was 53% ( 16 . 8±7 . 2 worms ) or 40% ( 20 . 0±5 . 1 worms ) on larval stage or adult stages , respectively . The same treatment schedule but with PZQ induced a worm burden reduction of 41% and 86% on schistosomules and adult worms , respectively . These treatments did not induce any visible adverse effect in mice . The effect of a combination of PZQ and trioxaquine PA1647 on the reduction of the worm burden in S . mansoni infected mice is reported in Table 2 . The first course consisted of 4 oral treatments with 50 mg/kg of drug given every three hours . The drug combinations were 100% of PZQ ( line 2 ) , 75 wt% PZQ/25 wt% PA1647 ( line 3 ) , 50 wt% PZQ/50 wt% PA1647 ( line 4 ) , 25 wt% PZQ/75 wt% PA1647 ( line 5 ) , and 100% of PA1647 ( line 6 ) , respectively . In these conditions , the reduction of the schistosomules burden with respect to control mice was 24% with 100% PZQ ( line 2 ) , 73% with 75 wt% PZQ/25 wt% PA1647 , ( line 4 ) , and 18% with 100% of PA1647 ( lines 2 , 4 , and 6 , respectively ) . These treatments did not induce any visible adverse effect in mice . Adult S . mansoni females were treated in vitro with PA1259 , PZQ , ARTM or MFQ at 50 µg/mL for three hours , and compared with untreated worms ( excipient only ) .
The in vitro activities of PA1259 ( ▪ ) on cercariae , schistosomules , and adult worms S . mansoni are reported in Figure 2A , 2B , and 2C , respectively , along with activities of praziquantel ( ★ ) , artemether ( ⧫ ) , and mefloquine ( • ) given as comparison . Results depicted in Figure 2 show that PA1259 exhibits a significant anti-schistosomal activity on all parasite stages . Concerning the free cercarial stages , PA1259 and MFQ have similar efficacy ( all cercariae were dead within 60–90 min , Figure 2A ) . These two drugs are significantly more efficacious than PZQ ( after 90 min of treatment with PZQ , 80% of cercariae were still moving , and 4 hours was needed to immobilize all cercariae ) . On the schistosomules stage , the time required to kill all schistosomules was 3 h , 5 h , or 8 h , with PZQ , PA1259 , or MFQ , respectively ( Figure 2B ) . So , PZQ and PA1259 are more active than MFQ . On adult parasites , the activities of these three last molecules are not significantly different ( p>0 . 05 ) , with all worms killed at 2–3 h ( Figure 2C ) . The in vitro activities obtained for both MFQ and PZQ are consistent with previous reports [30] , [31] . Compared to MFQ , PA1259 is more potent on schistosomule stage , and has the same activity on adult stage . It is noteworthy that MFQ , based on a quinoline moiety , is active on the cercarial stage , whereas ARTM , containing a trioxane , is not . This feature suggests that the quinoline part of PA1259 may play a role in its activity against schistosomes ( especially the cercarial stage ) , and that the parasite heme is not the only target of this drug ( cercariae do not contain heme ) . In fact , a non-heme target has recently been proposed for MFQ [30] . ARTM is inactive on cercariae and schistosomules , and only poorly active on adult worms . In fact , this latter drug was reported to be active only when hemin was added in the culture medium [32] . In addition , the activity of PA1259 on schistosomules and adult worms was compared with that of PA1019 which is the 4-aminoquinoline residue contained in PA1259 ( see Figure 1 for the structure of PA1019 ) . PA1019 was only poorly active , and 7 h were required to immobilize only 27% of schistosomules and 79% of adult worms . By contrast , treatment with PA1259 immobilized 100% of schistosomules and adult worms after 5 h or 3 h , respectively . This result support an additive or synergistic effect of the quinoline and trioxane moieties of PA1259 . In S . mansoni mice , the reduction of worm burden upon oral treatment by PA1259 or PZQ is reported in Table 1 . The administration of trioxaquine PA1259 at five daily doses of 100 mg/kg resulted in a reduction of 31% and 42% on larval and adult stage , respectively . With four doses of 50 mg/kg every three hours , the reduction of worm burden was 53% or 40% on larval or adult stages , respectively . Whatever the used protocol , the activity of PA1259 on larval stage was very close ( slightly higher ) to that of PZQ . On adult stage , PA1259 exhibited a significant activity , with a reduction of the worm burden being half of that obtained with PZQ . It is noteworthy that the efficacy of PA1259 was very close on schistosomules and adult worms , whereas PZQ exhibited a significantly higher efficacy on adult schistosomes compared to schistosomules . This feature suggests a different mode of action for these two drugs . For comparison purpose , quinoline- or artemisinin-based molecules present variable efficacy . For instance , among quinoline-based molecules , MFQ exhibits a significant efficacy against schistosome infections , but chloroquine is inactive against schistosome infections [8] . For comparison , artemisinin derivatives or synthethic trioxolanes mainly possess activity against schistosomules [33] , [34] . In a mouse model , ARTM was reported to be inactive at 800 mg/kg on adult S . mansoni infection [33] . In addition , on S . mansoni schistosomules , 4 doses of 50 mg/kg of PA1259 every three hours ( total dose of 200 mg/kg given over a period 9 hours ) was found to be a more efficacious protocol than a total dose of 500 mg/kg given over a period of 5 days ( 100 mg/kg administered daily ) . Owing to the complementarity of PZQ and trioxane based drugs against schistosomes , we investigated the reduction of the worm burden in mice infected by 21-day S . mansoni schistosomules when orally treated with an association of PZQ and trioxaquine PA1647 , the diphosphate salt of PA1259 ( Figure 1 ) . Five courses were done with different proportions of PZQ and PA1647 . Each course consisted of four oral doses of ( PZQ + PA1647 ) , at a total amount of 50 mg/kg for each dose ( Table 2 ) . The reduction of the schistosomule burden with respect to control mice was 73% with 50 wt% PZQ/50 wt% PA1647 , ( corresponding to 32 mol% of PA1647 , line 4 ) . For comparison , it was only 24% or 18% when PZQ or PA1647 , respectively , were used as monotherapy ( lines 2 and 6 , respectively ) . A simple additive effect between PZQ ( 24% worm burden reduction at 50 mg/kg ) , and PA1647 ( 18% worm burden reduction at 50 mg/kg ) would have provided at best a ( 24+18 ) /2 = 21% worm reduction with the combination PZQ 25 mg/kg + PA1647 25 mg/kg . In fact , such a combination resulted in a significantly higher reduction of 73% ( line 4 ) . So , the results are consistent with an additive or synergistic effect against schistosomules . Then , due to this promising effect of PZQ and PA1647 on schistosomules , the PZQ/PA1647 association should be considered as a drug-association candidate for future clinical tests , after additional optimization on the association drug ratio , in order to target all parasite stages . Adult S . mansoni females were treated in vitro with PA1259 , praziquantel , artemether or mefloquine were examined by photon- and electron microscopy ( SEM and TEM ) , with special focus on hemozoin , vitelline cells , musculature and tegument . In all cases , morphological alterations were apparent , but in highly variable extent and nature . The main observations reported in the Result Section were summarized in Table 3 where the most specific results have been indicated in bold . The following points are worth being emphasized . A multidisciplinary approach allowed us to propose the following mode of action of the antischistosomal trioxaquine PA1259 . First , the easy absorption due to the quinoline moiety of PA1259 should allow the drug to interact with the schistosome process of polymerization of heme to hemozoin . The reaction of the peroxide function of trioxaquine with free heme resulting from the degradation of hemoglobin was assessed by the production of heme-drug adducts characterized in worm extracts . As heme-drug adducts are probably able to inhibit the heme polymerization , and are themselves unable to polymerize ( this feature has been demonstrated for heme-artemisinin adducts [51] ) , this reaction led to accumulation in the worm of soluble redox-active heme derivatives . Then , the iron chelated by the protoporphyrin-IX ligand of heme or heme-drug adducts induced the production of reactive oxygen species , which should readily destroy hemozoin by radical chain reactions , resulting in ( i ) drastic discoloration of worms from black to brown , and ( ii ) much lower density of hemozoin in worms treated with PA1259 than in control worms , as observed in microscopy ( Figures 3 and 6 ) . The high quantity of nitric oxide detected in the parasite gut is certainly an important element of this oxidative stress ( Figure 8 ) . Third , this oxidative stress would cause gastrodermis perforation and the passage of the gut content through the gastrodermis , as observed by electronic microscopy ( Figure 6 ) . The oxidative cleavage of heme also releases free iron that should increase the production of radical oxygen species . Finally , the substructure of the tegument was altered by the invasion of oxygen reactive species ( Figures 4 and 7 ) . However , several questions remain to be adressed . Indeed , the fact that cercariae are heavily affected by treatment , the external tegumental alteration , and the effect on vitelline cells observed with PA1259 are not directly explained by reaction of the drug with heme . But the sequence hemoglobin/heme/hemozoin is not the single source of iron in schistosomes and several other iron proteins may play a role . Schistosomes have high demand for iron and are dependent on host iron for early development within the mammalian host [52] . In each stage , from miracidia to adult , schistosomes express tegumental divalent metal transporters involved in iron uptake [53] . Moreover , schistosomes are known to bind host transferrin at their surface [52] . The parasite also possesses two ferritin isoforms involved in storage and release of iron [54] , [55] . One isoform , called yolk-ferritin or ferritin 1 , is predominantly found in mature egg laying female , where it has been localized in the vitelline cells and in the ovary [56] . Vitelline stores of iron are implicated in eggshell formation [57] . Interestingly , the two isoforms of ferritin are also expressed in the parasite egg suggesting these molecules are also synthesized by embryo [57] . These observations emphasize the importance of iron in schistosome metabolism and egg formation , even on worm stages that do not ingest hemoglobin yet . Ferritin has not yet been characterized in cercariae , however the expressed sequence tags of ferritin 2 are available in sequence database suggesting that cercariae express this enzyme . Several recent articles suggest that iron metabolism should be a valuable target for either chemotherapy or vaccine development against schistosomes [58] , [59] . Using an anti-schistosomal drug that is also active on malaria parasites is a real matter of discussion [PA1259 is curative on P . vinckei petteri infected mice orally treated at 25 mg/kg/day during four days ( 5/5 mice cured without recrudescence at Day-30 , unpublished data from Palumed ) ] . Such dual activity might be considered as an advantage since patients are suffering from both diseases in many endemic areas . One the other side , one can argue about the possibility to generate trioxane-resistant strains of P . falciparum . But we should keep in mind that no trioxaquine-resistant strain has been selected after two years of drug pressure ( unpublished results ) . In conclusion , trioxaquine PA1259 is the most active against schistosomes among the trioxane-containing drugs that have been tested up to now . Its phosphate salt , PA1647 acts in synergy with praziquantel , specially against schistosomules , when infected mice are treated by oral administration . This opens the route to an efficient bitherapy of a highly neglected disease . | Schistosomiasis is a tropical disease affecting more than 200 million people throughout the sub-tropical and tropical world . The treatment and control of schistosomiasis rely on the use of a single drug , the praziquantel and no vaccine is available . However , schistosome species with low sensitivity or resistance to praziquantel have been identified in several countries . It is an urgent need to develop new drugs against this parasite . In this context , our study reports the activity the trioxaquine PA1259 . PA1259 is an hybrid drug containing two pharmacophores within a single molecule: a trioxane and an aminoquinoline . Initially developed against malaria , the trioxaquines target the heme a disposal product resulting from the digestion of the hemoglobin . The first action of the trioxaquine is an alkylation of the heme with the trioxane entity , and the second action is stacking with the heme due to the aminoquinoline moiety . In this study we show that this new drug is active in vitro against all schistosome stages ( cercariae , schistosomule and adult ) . The PA1259 is also active in vivo and shows synergistic action in association with praziquantel . This opens the route to an efficient bitherapy of a highly neglected disease . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"infectious",
"diseases",
"drug",
"research",
"and",
"development",
"drugs",
"and",
"devices",
"neglected",
"tropical",
"diseases",
"pharmacology",
"biology",
"zoology"
] | 2012 | Antischistosomal Activity of Trioxaquines: In Vivo Efficacy and Mechanism of Action on Schistosoma mansoni |
To better understand the tissue-specific regulation of chromatin state in cell-fate determination and animal development , we defined the tissue-specific expression of all 36 C . elegans presumptive lysine methyltransferase ( KMT ) genes using single-molecule fluorescence in situ hybridization ( smFISH ) . Most KMTs were expressed in only one or two tissues . The germline was the tissue with the broadest KMT expression . We found that the germline-expressed C . elegans protein SET-17 , which has a SET domain similar to that of the PRDM9 and PRDM7 SET-domain proteins , promotes fertility by regulating gene expression in primary spermatocytes . SET-17 drives the transcription of spermatocyte-specific genes from four genomic clusters to promote spermatid development . SET-17 is concentrated in stable chromatin-associated nuclear foci at actively transcribed msp ( major sperm protein ) gene clusters , which we term msp locus bodies . Our results reveal the function of a PRDM9/7-family SET-domain protein in spermatocyte transcription . We propose that the spatial intranuclear organization of chromatin factors might be a conserved mechanism in tissue-specific control of transcription .
Chromatin state can be regulated at the tissue-specific and cell-type-specific levels . Both active transcription and gene silencing require chromatin regulatory mechanisms to render DNA accessible or inaccessible , respectively . The post-translational modification of specific lysine residues in the tails of histone proteins by methylation or acetylation is an important mechanism of chromatin regulation [1 , 2] . Histone lysine methylation can determine cell-type-specific gene expression states [3] , and broad histone methylation domains at promoters are necessary for cell-fate maintenance; misregulation of such domains can drive oncogenic states [4 , 5] . Lysine methylation of histone tails is predominantly catalyzed by the SET domains of lysine methyltransferases ( KMTs ) , and the SET domain can be used to identify putative lysine methyltransferases across eukaryotic genomes [6–9] . Fifty presumptive KMTs are encoded by the human genome , most of which remain uncharacterized biochemically or functionally . While extensive analysis of histone methylation profiles in specific cell-types has revealed tissue-specific chromatin regulation , it has been difficult to determine the tissue-specificity of KMT function . In mammals , PRDM-type KMTs have emerged as tissue-specific chromatin regulators . The founding member , PRDM1 , was identified as a master transcription factor in hematopoietic differentiation [10] . PRDM9 functions in the determination of meiotic recombination sites in the mammalian germline [11] . PRDM3 and PRDM16 are required for heterochromatin formation in adipose tissues [12] . The SET-domain families of PRDM-type KMTs are named according to their mammalian representatives and are broadly conserved across metazoans . However , domains outside the SET-domain can vary by organism . For example , whereas C . elegans BLMP-1 and its mammalian ortholog PRDM1 share both SET and Zn-finger domains , SET-17 ( which is the only other PRDM-type KMT in C . elegans ) does not share the KRAB or Zn-finger domains of its mammalian SET-domain family counterparts PRDM9 and PRDM7 ( S1A and S2B Figs ) . The organization of the genome into spatial domains is emerging as an important factor in the specification of cell-fate [13] . These domains can facilitate the physical interaction of enhancers with target promoters [14] and can define heterochromatin domain boundaries [15] . Domain organization of the genome is important in development and disease [16] . The relationship between tissue-specific chromatin signatures , spatial domains and transcriptional regulation remains poorly understood [17 , 18] . The C . elegans genome encodes 36 presumptive KMTs , most of which are conserved in mammals based on their SET domain sequences ( S1A Fig ) . Many of these KMTs remain uncharacterized . Previous analysis of C . elegans KMT mutants showed that loss of most individual KMTs does not affect embryonic development or post-embryonic vulval development [9] . Our goal was to define the tissue-specific expression patterns of all C . elegans KMTs and focus functional analyses on a specific tissue .
To analyze the tissue-specificity of KMTs in C . elegans , we determined the endogenous KMT mRNA expression patterns for all 36 presumptive KMT genes using single-molecule fluorescence in situ hybridization ( smFISH ) [19] . smFISH allows the detection of endogenous mRNA in all C . elegans tissues , including the germline , in which multi-copy transgenes are silenced [20] . The KMT gene expression patterns in the L1 larval stage fell into two categories , broad and tissue-specific: 12 KMTs were expressed broadly , while 22 were detectable in only one or two tissues ( Figs 1 and 2 ) . No expression was detected for two KMTs . The 12 broadly expressed KMTs included the C . elegans orthologs of general chromatin modifiers set-16 ( MLL; Figs 1A and 2 ) , met-1 ( SETD2 ) , met-2 ( SETDB1 ) , set-25 ( SUV39 ) and set-4 ( SUV420 ) . The SET-domain orthologs of some KMTs that function tissue-specifically in mammals were also broadly expressed , such as set-17 , the SET domain of which is similar to the SET domains of the PRDM9/7 family ( Figs 1B , S1A and S3B ) . mes-2 , the C . elegans ortholog of the Polycomb KMT EZH , was expressed in only hypoderm and germline ( Fig 1C ) . We verified the expression pattern of mes-2 , and that of mes-4 , with two independently designed smFISH probe-sets . Consistent with our findings , recent RNAseq studies of single postembryonic C . elegans cells also have reported somatic expression of mes-2 and mes-4 [21] . Twelve KMTs were detectable in only a single tissue: five were germline-specific ( e . g . Fig 1D ) , three muscle-specific ( e . g . Fig 1E ) , three hypoderm-specific and set-11 was neuron-specific ( Fig 1F ) . Our data show that at least 11 KMTs were expressed in any given tissue , and the germline was the tissue with the broadest KMT expression . We conclude that most KMTs are expressed tissue-specifically , although expression of 12 KMTs occurs in most cells . To assess the functions of individual KMTs , we focused on germline-expressed KMTs . The mRNAs of all KMTs that were detectable in primordial germ cells in the L1 were also detectable in the adult germline ( S1B Fig , S2 Fig ) . The primary measure of germline function is fertility . We therefore determined the number of progeny of all available and viable mutants of germline-expressed KMTs ( Fig 2 , Fig 3A ) . All of the mutations we tested were out-of-frame deletions and therefore likely to cause complete loss of function . It had been previously reported that the fertility of met-1 and met-2 mutants declines over several generations at elevated temperatures [9] . To avoid such effects , we maintained all strains at lower temperatures and raised the temperature only for fertility measurements . Our assay confirmed the previously reported fertility defects of met-1 and met-2 mutants [9] , while animals with individual mutations of most germline-expressed KMTs exhibited normal fertility . set-17 ( n5017 ) mutants showed a 50% reduction in fertility ( Fig 3A and 3B ) . set-17 ( n5017 ) deletes the first 135 amino acids of set-17 , including part of the SET-domain ( S3A Fig ) . A second allele , set-17 ( gk417488 ) , an ochre termination codon at amino acid 84 , caused a fertility defect similar to that of set-17 ( n5017 ) ( hereafter set-17 , S3A Fig ) . Expression of wild-type set-17 partially rescued the fertility defect of set-17 mutants when expressed as a single-copy insertion [22 , 23] from its endogenous promoter ( Pset-17 ) or expressed in the germline from a germline-specific promoter ( Pmex-5 , Fig 3B ) . We found that the expression of set-17 was restored only partially by either the set-17 endogenous promoter ( 58% of the wild-type level ) or the germline-specific mex-5 promoter ( 47% of the wild-type level ) ( S6B and S6F Fig ) . This partial restoration of set-17 expression could account for the partial rescue of fertility . Mutations in the other germline-expressed KMTs we tested did not affect the fertility defect of set-17 mutants ( S3C Fig ) . We conclude that loss of set-17 function reduces fertility and that set-17 function in the germline is sufficient to promote fertility . set-17 could function in either the sperm- or the oocyte-producing germline , or both , to promote fertility . To distinguish among these possibilities , we conducted mating experiments with set-17 males and females . We feminized set-17 hermaphrodites using the fog-2 ( q71 ) mutation , which suppresses sperm production in hermaphrodites [24] . If and only if males were mutant for set-17 was there a reduction in the number of cross-progeny ( Fig 3C ) : when mated to wild-type males , set-17 females produced the same number of progeny as wild-type females; by contrast , both wild-type and set-17 females displayed a 50% reduction in the number of progeny when mated to set-17 males . We conclude that SET-17 functions in only the male germline to promote fertility . In C . elegans spermatogenesis , germ cells differentiate into primary spermatocytes while they are undergoing meiosis I . One primary spermatocyte then differentiates into a secondary spermatocyte , which generates four haploid spermatids with the completion of meiosis II . Spermatids are then stored in the spermatheca of the hermaphrodite or the male . Before fertilization spermatids are activated by external cues from oocytes to become actively crawling spermatozoa . To determine if set-17 functions in spermatid production , we counted the number of spermatids in the spermatheca of young adult hermaphrodites after spermatid production had concluded . set-17 mutants exhibited a 50% reduction in the number of spermatids per spermatheca ( Fig 3D ) . This defect was rescued by expressing wild-type set-17 either from its endogenous promoter or germline-specifically . We did not detect functional abnormalities in set-17 spermatozoa in an in vitro activation assay or when monitoring set-17 spermatozoa crawling from the uterus to the spermatheca after mating , suggesting that set-17 spermatids function normally once produced . We conclude that set-17 functions in spermatid production to promote fertility . To identify morphological abnormalities that might be related to the decreased spermatid production and fertility of set-17 mutants , we used electron microscopy to examine the stages of spermatocyte development and spermatid formation ( Fig 4A ) in set-17 males . Although the gross morphologies of both spermatocytes and spermatids appeared normal ( Fig 4B and 4C ) , the cytoplasm of mature set-17 primary spermatocytes displayed a striking ultrastructural abnormality . Fibrous-body membranous organelles ( FB-MOs , Fig 4D–4F ) , which store paracrystals of the major sperm proteins ( MSPs ) [25] , defined relatively little of the cytoplasmic area of set-17 primary spermatocytes . MSPs are produced in spermatocytes and function in spermatozoa , in which they form the pseudopod that drives motility . We found that FB-MOs covered about 20% of the cross-sectional cytoplasmic area of wild-type primary spermatocytes and only about 8% of the cytoplasmic area in set-17 spermatocytes ( Fig 4E ) . This abnormality was rescued in set-17 males that expressed wild-type set-17 as a transgene driven by the endogenous set-17 promoter ( Fig 4E ) . We found that the reduction in overall FB-MO cross-sectional area in set-17 spermatocytes was caused by a reduction of individual FB-MO cross-sectional area to half of that in wild-type spermatocytes ( Fig 4F ) . set-17 males were normal in their number of FB-MOs ( S4A and S4B Fig , Methods ) . The reduction of individual FB-MO area in set-17 spermatocytes leads us to conclude that set-17 promotes FB-MO production in primary spermatocytes . We suggest that an absence of set-17 function causes a reduction of FB-MO production in spermatocytes . We speculate that reduced FB-MO production contributes to reduced spermatid number that in turn leads to reduced fertility ( Fig 4G ) . To determine the expression pattern and subcellular localization of SET-17 protein , we used a single-copy transgene that expresses SET-17::GFP under the set-17 endogenous promoter and can rescue the fertility , sperm-production and FB-MO-production defects of set-17 mutants ( Figs 3C , 3E and 4E ) . Using confocal microscopy , we examined immobilized live male germlines ( Fig 5A ) and fixed male germlines that were extruded from adult males by microsurgery and labeled by immunofluorescence ( Fig 5B ) . We found that in the male germline SET-17::GFP was expressed predominantly in primary spermatocytes ( Fig 5A and 5B ) . SET-17::GFP was first detectable at the initiation of spermatocyte differentiation in L4 and adult males . The SET-17::GFP level increased with the stage of spermatocyte progression and peaked in late-stage primary spermatocytes ( Fig 5A and 5B ) . Strikingly , SET-17::GFP localized to foci in the nuclei of primary spermatocytes ( Fig 5B i-iii ) . The SET-17 foci were associated with DNA , ranged from 4–14 per primary spermatocyte and increased in number with overall SET-17::GFP expression ( S5A Fig ) . In hermaphrodites , SET-17::GFP was similarly expressed in primary spermatocytes and localized to nuclear foci ( S5B Fig ) . In hermaphrodites , spermatid production occurs during the L4 stage . We also detected SET-17::GFP expression during oocyte production ( S5C Fig ) , even though SET-17 does not obviously function in oocytes for fertility ( Fig 3C ) . In hypodermal nuclei , the distribution of SET-17::GFP was pan-nuclear ( S5D Fig ) , suggesting that SET-17 localization to foci is related to the function of SET-17 in spermatocytes . The SET-17::GFP expression pattern is consistent with our conclusion that SET-17 functions in primary spermatocytes and suggests that SET-17 functions in nuclei in chromatin-associated foci . To investigate if SET-17 localization affects the chromatin state of primary spermatocytes , we determined the distribution of the presumptive products of the lysine methyltransferase activity of SET-17 . SET-17 purified from E . coli can methylate lysine 4 of histone 3 to the mono- and di-methyl states ( H3K4me1 and H3K4me2 ) but not the tri-methyl state in vitro [26] . By immunostaining dissected male germlines , we found that H3K4me1 and H3K4me2 were distributed throughout the nuclei of primary spermatocytes . While both H3K4me1 and H3K4me2 were detectable at SET-17 foci , they were not enriched at SET-17 foci ( Figs 5C , 5D , S5E and S5F ) . These results show that H3K4me1 and H3K4me2 are both present at SET-17 foci and suggest that SET-17 is not the only KMT that can catalyze the generation of H3K4me1 and H3K4me2 in primary spermatocytes . The localization of SET-17 to chromatin in primary spermatocytes coupled with the reduction in set-17 mutants of FB-MOs , spermatids and fertility suggested that SET-17 functions to promote transcription in primary spermatocytes . To determine the role of set-17 in spermatocyte gene expression , we analyzed the transcriptome of hermaphrodites at the sperm-producing L4 stage . Loss of set-17 affected the expression of 123 transcripts , down-regulating them 2-fold on average; this transcriptional misregulation was rescued by wild-type set-17 expression using the set-17 promoter ( S6A and S6B Fig ) . We asked if set-17 affects the expression of genes enriched in spermatocytes . Based on known gene-expression data from individual dissected germlines in the oogenic or spermatogenic state [27] , we categorized transcripts into four categories: spermatogenic , oogenic , spermatogenic and oogenic or non-germline . Of the 123 transcripts misregulated in set-17 mutants ( defined as “all set-17” ) , 60 were spermatogenic ( “set-17 spermatogenic”; Fig 6A ) . The 60 set-17 spermatogenic transcripts were on average 2-fold downregulated in set-17 mutants ( Fig 6B ) . Wild-type set-17 expression in the germline rescued the expression of set-17 spermatogenic transcripts ( Fig 6B ) . Loss of set-17 did not affect spermatogenic gene expression globally , as on average the 2306 previously identified spermatogenic transcripts were unchanged in set-17 ( Fig 6B ) . A rank-correlation analysis of all set-17 misregulated genes indicated that the most highly misregulated transcripts were enriched for spermatogenic genes ( S6C Fig ) . We conclude that SET-17 functions in the germline to promote the expression of a small subset of spermatogenic genes . Since FB-MO production was reduced in set-17 spermatocytes and FB-MOs store MSP paracrystals , we asked if the genes encoding MSPs were among the down-regulated spermatogenic transcripts in set-17 mutants . The msp gene family comprises 28 distinct genes that encode identical MSP peptides . While the coding mRNA sequences of the 28 msp genes are 95% identical , it is possible to assay the expression level of each msp gene individually using RNAseq because of their completely divergent 3’ and 5’ UTRs as well as the randomly distributed single nucleotide polymorphisms that occur throughout the coding sequences . We designed a set of 16 smFISH probes that covered the full length of the msp coding sequence . Each individual msp gene can be recognized by at least 10 probes , so collectively these probes detect all 28 msp transcripts . Collectively the 28 msp genes were down-regulated in set-17 mutants . 26 of 28 msp genes were among the 60 set-17 spermatogenic transcripts ( Fig 6C ) . msp expression was reduced to 50% of wild-type levels in set-17 mutants , and this defect was rescued by the expression of wild-type set-17 specifically in the germline ( Fig 6C ) . We conclude that set-17 promotes the expression of msp genes . To explore how set-17 affects the expression of msp and other spermatogenic genes , we investigated the genomic positions of the genes encoding the set-17 spermatogenic transcripts . msp genes cluster in three distinct regions in the genome: one on chromosome II and two on chromosome IV [28] . Non-msp spermatocyte-specifically expressed genes are also enriched in these clusters [29] . First , we analyzed the distribution of all known spermatogenic genes on chromosomes II and IV to confirm the clustering of spermatogenic genes at the msp gene loci ( blue regions , Fig 6D and 6E ) . We then plotted the fold-change in expression levels in set-17 vs . wild-type as a function of genomic position for the set-17 spermatogenic genes ( shown in red; yellow indicates msp gene ) . We found that 53 of the 60 set-17 spermatogenic genes were located in the spermatogenic gene clusters on chromosomes II and IV ( 26 msp genes , 27 non-msp genes; Fig 6D and 6E ) . We conclude that set-17 preferentially promotes expression of genes in spermatogenic gene clusters . To investigate whether set-17 might also regulate spermatogenic genes in the clusters on chromosomes II and IV in addition to the ones already identified in our global analysis of gene expression in set-17 mutants , we asked if the spermatogenic genes in the clusters were more likely to be down-regulated in set-17 mutants when compared to spermatogenic genes outside of the clusters . This analysis showed that loss of set-17 preferentially caused the down-regulation of spermatogenic genes located in the clusters on chromosomes II and IV and that expression of wild-type set-17 in the germline rescued this effect ( Figs 6F , 6G , S6G and S6H ) . This preferential down-regulation of spermatogenic genes in clusters remained statistically significant after removing the msp genes from the analysis ( S6I and S6J Fig ) . Many additional spermatogenic genes have been identified since the spermatogenic gene clusters on chromosomes II and IV were identified . We therefore hypothesized that previously unidentified non-msp spermatogenic gene clusters might exist in the C . elegans genome and that set-17 might affect the expression of the genes in such a cluster . We therefore analyzed the distribution of spermatocyte-specifically expressed genes defined by Ortiz et al . on all C . elegans chromosomes [27] . We identified a previously unknown cluster on chromosome V ( S6K Fig ) . We analyzed the effect of loss of set-17 on spermatogenic genes in the chromosome V cluster compared to the spermatogenic genes outside the cluster on chromosome V . We found that loss of set-17 specifically reduced expression of the chromosome V cluster genes and that this reduction was rescued by expression of wild-type set-17 specifically in the germline ( Fig 6H ) . These results show that SET-17 acts in the germline to promote expression from four spermatogenic gene clusters on chromosomes II , IV and V . That SET-17 promotes the expression of spermatogenic genes suggests that SET-17 promotes their transcription in spermatocytes . Using the set of msp smFISH probes described above , we found that in mature primary spermatocytes of L4 hermaphrodites loss of set-17 caused a 50% reduction of total msp mRNA ( Fig 7A and 7B ) . We conclude that set-17 is needed for full msp expression in spermatocytes . To determine the transcriptional activity of msp genes , we quantified msp RNA foci at transcription sites . RNA polymerase II transcription occurs in bursts of several transcripts , causing the transient accumulation of pre-mRNAs at transcription sites [30] . The transcriptional activity of a gene transcribed by RNA polymerase II can be determined by quantifying nuclear RNA foci by smFISH [31 , 32] . We examined primary spermatocytes in L4 hermaphrodites and counted msp transcription sites in the nuclei of primary spermatocytes . We detected transcription sites in most wild-type spermatocyte nuclei and found that msp transcription sites were reduced in number in the spermatocytes of set-17 mutants ( Fig 7C and 7D ) . We scored primary spermatocyte nuclei as positive or negative for msp transcriptional activity , depending on whether at least one msp transcription site was detectable . The percentage of positive nuclei was reduced in set-17 primary spermatocytes to about 60% of that of wild-type ( Fig 7C ) , suggesting that msp gene transcription is reduced in set-17 spermatocytes . To determine the transcriptional activity of single msp gene clusters , we counted the number of transcription sites per nucleus in wild-type and set-17 L4 hermaphrodite spermatocytes ( Fig 7D ) . The median number of transcription sites per primary spermatocyte nucleus was two in wild-type and one in set-17 animals . We computed the probability of individual msp gene cluster transcription from the observed distribution of transcription sites ( S7A and S7B Fig , Methods ) . Strikingly , we found that loss of set-17 caused a 50% reduction in the probability of transcriptional activity of individual msp clusters ( Fig 7E ) . These results demonstrate that set-17 regulates the transcriptional activity of msp genes in primary spermatocytes . SET-17 is required for the transcription of msp gene clusters , and SET-17 localizes to foci in spermatocyte nuclei ( Fig 5B ) . We hypothesized that SET-17 foci might localize to msp gene clusters to promote their transcription . Strikingly , we found that msp transcription sites colocalized with SET-17::GFP foci in primary spermatocytes ( Fig 7F ) . 90% of msp transcription sites occurred at SET-17::GFP foci . msp transcription sites occurred in the absence of SET-17::GFP foci only in very early primary spermatocytes with either one or no SET-17::GFP foci . The colocalization of SET-17::GFP foci with msp transcription sites is consistent with the hypothesis that SET-17 functions to promote but might not be absolutely essential for msp gene expression . To characterize structural and dynamic features of SET-17 foci , we used time-lapse imaging and fluorescence recovery after photo-bleaching ( FRAP ) studies of individual SET-17::GFP foci in spermatocytes in vivo . We found that individual SET-17::GFP foci persisted in immobilized 1-day adult males for the duration of our experiments ( ~10 min ) and that SET-17::GFP fluorescence recovered after photo-bleaching in individual foci ( Fig 7G ) . By quantifying the SET-17::GFP fluorescence recovery for individual foci we determined the rate of exchange of SET-17::GFP in the foci ( Fig 7H ) . SET-17::GFP recovered fluorescence in spermatocyte foci on a time-scale of tens of seconds ( Fig 7H ) , while recovery occurred on a time scale of seconds in hypodermal nuclei , where the protein is distributed throughout the nucleus ( Fig 7G and 7H ) . The time of half-maximal fluorescence recovery is a measure of the equilibrium constant of the exchange of SET-17 from the foci with the environment . The time of half-maximal recovery of SET-17::GFP in spermatocyte foci was significantly slower than that in hypodermal nuclei ( Fig 7I and 7J ) . These data revealed a barrier to SET-17 exchange from the foci with the nucleoplasm in spermatocytes . Foci with diffusion barriers can be the consequence of liquid-liquid phase separation [33] . In addition to slow diffusion , liquid-liquid phase-separated droplets exhibit fusion and fission , which can distinguish them from other macro-molecular assemblies with slow diffusion rates [33] . Although SET-17 foci were largely stationary in immobilized animals , we did observe some fusion and fission of SET-17 foci ( S7C and S7D Fig ) . We speculate that SET-17 foci might be novel phase-separated stable nuclear structures associated with msp gene clusters that facilitate msp transcription . Our data suggest that the reduction of fertility in set-17 mutants is caused by a reduction of transcription of msp genes . The transcriptional regulator ELT-1 can bind in vitro to a promoter motif found upstream of 44 spermatocyte-specifically expressed genes , including 19 msp genes [34] . We showed that these 44 predicted elt-1 target genes were enriched in the three known spermatogenic clusters on chromosomes II and IV ( 33 out of 44 , S8A and S8B Fig ) and that 26 are also among the set-17 spermatogenic genes ( Fig 8A ) . All 44 predicted elt-1 targets were downregulated in set-17 mutants , and their expression was restored by expression of wild-type set-17 in the germline ( Fig 8B ) . Thus SET-17 and ELT-1 might function together in msp transcription . Does ELT-1 function in endogenous msp regulation and fertility ? Because elt-1 is required for viability , we examined a viable partial loss-of-function mutant of elt-1 . elt-1 ( ku419 ) carries a P298S missense mutation in one of the two DNA-binding domains of ELT-1 [35] . Using smFISH , we found that this elt-1 mutation caused a ~50% reduction in endogenous msp transcript levels in primary spermatocytes , similar to the reduction in set-17 mutants ( Fig 8C and 8D ) . msp mRNA levels of set-17; elt-1 double mutants were indistinguishable from those of set-17 single mutants ( Fig 8D ) . These results indicate that ELT-1 regulates msp expression together with rather than independently of SET-17 . elt-1 mutants were reduced in fertility to ~50% of that of the wild-type , similar to set-17 mutants ( Fig 8E ) . The elt-1 mutation did not enhance the fertility defect caused by a set-17 mutation , indicating that elt-1 and set-17 act together to control fertility . We conclude that ELT-1 and SET-17 function together to promote msp expression and fertility . To ask how fertility and msp expression might be related , we plotted fertility as a function of msp expression levels for wild-type , set-17 , elt-1 , set-17; elt-1 . msp mRNA levels were tightly correlated with fertility ( Fig 8F , r = 0 . 9981 , P < 0 . 0001 ) , suggesting that msp expression levels might determine fertility . We explored this relationship further by plotting msp expression and fertility for all strains , including the msp expression data from the RNAseq experiments ( S8C Fig , R2 = 0 . 97 ) . msp expression and fertility corresponded for set-17 and elt-1 mutants , as well as for the set-17 ( + ) rescued strains , independent of experimental measurement , further demonstrating a correlation between msp expression and fertility . We conclude that SET-17 functions in fertility by promoting the transcription of msp gene clusters .
set-17 is one of only two PRDM-type KMTs in C . elegans . The SET-17 protein is a member of the PRDM9/7 SET-domain family but lacks any additional domains , such as the KRAB domain of PRDM9 and PRDM7 or the Zn fingers of PRDM9 ( S2B Fig ) [11] . In the mammalian germline , PRDM9 is involved in the positioning of DNA double-strand breaks for recombination [11 , 36 , 37] . Loss of PRDM9 function causes infertility in male mice as sperm arrest in meiotic prophase [38] . It is unknown how PRDM9 recruits to its target loci the recombination machinery for double-strand break induction . We speculate that PRDM9 might function in SET-17-foci-like nuclear bodies for the initiation of double-strand breaks . Indeed , recent analysis of the induction of double-strand breaks and meiotic recombination have revealed protein accumulation in distinct foci and led to the proposal that such structures represent “recombinosomes” [39] . Molecular analysis of PRDM9 transcripts in whole mouse testes identified a short isoform that lacks the Zn-Finger array and the function of which is unknown [38] . Perhaps this short SET-domain-only isoform of mammalian PRDM9 functions in transcriptional regulation in mammalian spermatocyte development , analogous to SET-17 function in C . elegans spermatocytes . The function of PRDM7 remains unexplored . Loss of the human H3K4me2 demethylase LSD1 in the paternal germline can cause developmental defects in offspring [40 , 41] . The C . elegans ortholog of LSD1 , SPR-5 , functions in transgenerational germline immortality [42] . Mutation of set-17 can suppress the progressive sterility and accumulation of H3K4me2 in spr-5 mutants [26] . Intriguingly , transcriptional changes associated with loss of fertility in spr-5 mutants are enriched for set-17 spermatogenic genes , including a subset of the msp genes ( S8E Fig , P < 1E-20 ) . Many of the spermatocyte-specific genes that are down-regulated in set-17 are up-regulated in late-generation adult spr-5 mutant hermaphrodites . Based on these observations , we suggest that set-17 promotes spermatogenic gene expression from clusters and spr-5 restricts this activity and that PRMD9 or another SET-17-like chromatin regulator functions in mammalian sperm production to deposit chromatin marks that must be appropriately erased by LSD1 for proper development . Histone genes cluster in the genome , are highly expressed and associate with distinct nuclear structures , known as histone locus bodies ( HLBs ) , to facilitate their expression [43] . Replication-dependent histone genes are transcribed during S-phase to produce histones for the replicated genome . Multiple histone genes in each cluster encode the same histone protein , analogous to msp genes in C . elegans . Like histone genes , msp genes are short ( ~400 bp ) and lack introns . We propose to call the SET-17 foci at msp gene clusters msp locus bodies ( MLBs ) by analogy with HLBs . Like HLBs , MLBs are stable over long periods of time , exhibit slow exchange of protein components and localize to transcriptionally active genomic loci . We also observed instances of MLB fusion and fission ( S7 Fig ) . Dynamic spatial clustering of RNA polymerase II at loci of active transcription has been observed in mouse embryonic fibroblasts [44] . MLBs might function to facilitate the clustering of RNA polymerase II at msp gene clusters in C . elegans spermatocytes . We suggest that spatially restricted nuclear structures concentrate chromatin factors and transcription machinery to support the controlled activation of transcription . That the expression of most putative KMTs is restricted to only one or two postembryonic tissues suggests that KMTs function broadly in tissue-specific chromatin regulation . Further analysis of KMT specialization , localization and redundant function should enable the identification of additional aspects of chromatin regulation and have important implications for the understanding of chromatin biology in the context of development and disease . For example , a view of oncogenesis is emerging in which tumors originate as a consequence of cell-fate transformations in the cell-type of tumor origin [18] . One focus has been identifying master transcription factors or chromatin drivers of tissue-specific gene expression that might be involved in these cell-fate transformations . Our data show that SET-17 is a novel type of chromatin-modifying driver of tissue-specific transcription states . We suggest that SET-17-like tissue-specific chromatin modifiers are important for transcriptional regulation in development and disease .
Probe design and smFISH experiments were performed essentially as described elsewhere [19] with 10% formamide in the hybridization buffer . Samples were incubated overnight at 30˚C and washed for 30 min twice . Imaging was performed using an inverted microscope with a 100x oil objective ( Nikon , NA 1 . 4 ) . smFISH signal was detected using a Pixis 1024 ( Princeton Instruments ) with exposure times of 2 sec . For msp FISH the high fluorescence intensity required exposure times of 500 msec . and a scMos Orca Flash 4 . 0 camera ( Hamamatsu ) was used for fluorescence detection . For broodsize determination , young L4 animals were individually placed on plates with food and transferred to fresh plates every 24 hr for 3 days . The number of progeny was determined by counting the number of adult animals produced over 6 days . All experiments were performed at 25˚C . Strains were maintained at 20˚C and transferred to 25˚C . | During animal development , the generation of diverse cell types , such as nerve cells or muscle cells , requires differential gene expression . Lysine methyltransferases ( KMTs ) help determine which regions of the genome and hence which genes are expressed in each cell type . To enable the study of KMT function in development , Engert et al . defined the spatial expression patterns of all 36 presumptive KMT genes in the nematode Caenorhabditis elegans . While one-third of KMT genes were expressed broadly , the expression of most KMT genes was restricted to only certain cell types . The KMT gene set-17 was expressed in sperm-producing cells , where it proved to be necessary for normal fertility . The SET domain of the SET-17 protein is similar to the SET domain of the mammalian KMT PRMD9 , which functions in fertility in mammals; SET-17 lacks the other domains of PRDM9 . The authors showed that set-17 promotes fertility in sperm cell precursors by enabling the expression of clusters of genes that are required for the generation of sperm . Unexpectedly , the SET-17 protein localized to distinct spatial regions in the nuclei of sperm cell precursors . The authors propose that the spatial organization of KMTs and other factors that control gene expression is a broadly used mechanism in cell-type-specific gene regulation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"invertebrates",
"spermatocytes",
"caenorhabditis",
"dna-binding",
"proteins",
"animals",
"dna",
"transcription",
"germ",
"cells",
"animal",
"models",
"caenorhabditis",
"elegans",
"model",
"organisms",
"experimental",
"organism",
"systems",
"epigenetics",
"chromatin",
"sper... | 2018 | A Caenorhabditis elegans protein with a PRDM9-like SET domain localizes to chromatin-associated foci and promotes spermatocyte gene expression, sperm production and fertility |
The formation of metastases is driven by the ability of cancer cells to disseminate from the site of the primary tumour to target organs . The process of dissemination is constrained by anatomical features such as the flow of blood and lymph in the circulatory system . We exploit this fact in a stochastic network model of metastasis formation , in which only anatomically feasible routes of dissemination are considered . By fitting this model to two different clinical datasets ( tongue & ovarian cancer ) we show that incidence data can be modelled using a small number of biologically meaningful parameters . The fitted models reveal site specific relative rates of dissemination and also allow for patient-specific predictions of metastatic involvement based on primary tumour location and stage . Applied to other data sets this type of model could yield insight about seed-soil effects , and could also be used in a clinical setting to provide personalised predictions about the extent of metastatic spread .
For most forms of cancer , occurrence of distant metastases equals incurable disease [1 , 2] . Regional metastases , i . e . positive lymph nodes , also implies an inferior prognosis [3 , 4] . In order to diminish the risk of dissemination of the primary tumor , patients commonly receive adjuvant treatment with radiotherapy and/or some kind of medical oncological treatment . Yet , in many cases , the patient is later on faced with residual or recurrent disease [5] . On the other hand , many patients receive adjuvant treatment with subsequent side effects , even though their illness would never have disseminated [6 , 7] . Hence , increased knowledge about the extent of metastasis at diagnosis would improve the care of many cancer patients . Metastatic spread is known to follow certain disease specific patterns [8 , 9] , but despite the current state of knowledge , a quantitative understanding of the process of metastatic spread would improve our ability to optimise therapy and reduce over-treatment . Here will we make use of stochastic modelling to quantitatively assess the importance of different routes of dissemination . In addition , this method also makes it possible to estimate the extent of metastatic spread based on tumour stage and location . Our approach is versatile and we highlight this by applying it to two different types of tumours: carcinoma of the oral tongue , which disseminates primarily through the lymphatic system and ovarian carcinoma , which spreads both intraperitoneally , lymphatically and through the blood circulatory system . The tendency of a primary tumour to form a metastasis is the hallmark of malignant cancer [10] , and the process by which this occurs is a multi-step process [11 , 12] . Typically tumour cells detach from the primary tumour , invade the surrounding stroma and find their way to local lymph nodes or blood vessels [13] . Although a late stage tumour can release very large number of circulating tumour cells ( CTCs ) into the blood stream ( up to 4 × 106 cells shed per gram of tumour per day [14] ) , the low probability of forming metastatic foci [15] coupled with the low probability of passing through capillary beds [16] lead to the conclusion that micrometastases in for example the liver ( for gut malignancies ) and the lung ( for all malignancies ) are necessary for further hematogenic dissemination [17–19] . These , often microscopic [20] , lesions release CTCs into the arterial side of the circulation ( in the case of the lung ) [21] , and hence amplify CTC numbers in arterial blood , which otherwise would be low due to the filtration occurring in the lung capillary bed . A similar process is at work during lymphatic spread [22] , where CTCs get trapped in lymph nodes , where they form micrometastases , which shed CTCs that travel further downstream in the lymphatic system . This process is known as secondary seeding , and although it is of paramount importance for the metasatic process we still have a limited understanding of the steps involved . Here mathematical modelling can provide a helping hand , since it allows for a quantitative understanding of biological processes that are difficult to measure directly . Mathematical modelling of metastasis dates back to a series of seminal papers by Liotta and coworkers [23–25] . They considered the release of CTCs from an implanted tumour in mice and the subsequent formation of lung metastases . Using both deterministic and probabilistic methods they could derive predictions of how the number of metastatic foci changes over time and how the probability of being free of metastasis changed [24] . These predictions agreed well with experimental data and highlights the stochastic yet predictable nature of metastatic spread . Another important contribution was made by Iwata et al . [26] who formulated and analysed a model which accounts for secondary seeding . That model predicts how the size distribution of metastases changes as the disease progresses . Predictions of the model were tested by Baratchart et al . [27] in a murine model of renal carcinoma , and while the model could predict the total metastatic burden , it was unable to describe the size distribution of metastases . However , the model prediction could be improved by assuming interactions between metastatic foci . The Iwata model has also been used for connecting presurgical primary tumor volume and postsurgical metastatic burden and survival [28] . The model was applied to two datasets from mouse models and one clinical dataset and the analysis revealed a highly nonlinear relationship between resected primary tumor size and metastatic recurrence . A similar approach has been used by Hanin and coworkers in a model which accounts for the growth rate of the primary tumour , shedding of metastases , their selection , latency and growth in a given secondary site [29] . They proved that the parameters of the model are identifiable in the case of Gompertzian growth of the primary tumour , and apply a maximum likelihood method to identify the model parameters in the case of a single patient with a large number of lung metastases . Other mathematical models have successfully been applied in order to describe the dynamics of metastatic spread in pancreatic cancer [30] and breast cancer [31] . The importance of secondary seeding was investigated by Scott et al . [19] in the context of self-seeding [32] , the process whereby a primary tumour can accelerate its own growth by releasing CTCs that return to the site of origin . A careful mathematical treatment of this hypothesis showed that secondary seeding is indeed required for this pathway to contribute to primary growth . Secondary seeding also has an impact on estimates of metastatic efficiency , since undetected micrometastatic lesions render the apparent spread directly from the primary to target sites more efficient than it actually is [18] . The idea of metastatic spread occurring on a network , where the nodes represent organs and the links correspond to routes of spread was first described by Scott et al . [33] , and later modelled quantitatively by Newton et al . [34 , 35] . They considered a stochastic model where the dissemination of cancer cells is modeled as an ensemble of random walkers on the network . The dynamics of the model is determined by a transition matrix , which was obtained by fitting the model to a large autopsy data set [36] . The entries of this transition matrix give information about rates of dissemination between different organs . Here we build on the work of Newton et al . , but with secondary seeding in mind , make modifications which alleviate the problem underdetermination , which plagued their work . In order to parametrise their model they had to assume that the observed patterns of metastasis correspond to a steady-state distribution of the model . We instead make use of primary tumour stage to create a model which is temporal and considerably easier to parametrise . Focusing on two primary tumours we show that our model is able to estimate the dissemination rates with high confidence , and crucially allows us to estimate rates between sites , which are inaccessible if one simply analyses incidence data .
The drainage of the lymphatic system in the head and neck area can be described by a network where the nodes correspond to lymph node stations and the links represent flow between the stations . The location of the primary tumour determines how much it drains into the different stations , Tumours of the tongue mainly drain into station I and II , and to some extent to station III . Station I drains into II , which in turn drains into III , which drains into IV [38] . The dissemination network is shown in Fig 1 , where the directed links show how CTCs flow in the system . Given a primary tumour metastases will eventually form in downstream stations , but the dynamics of the model crucially depend on the parameters ( λI , λII , λIII , ϕI , ϕII , ϕIII ) . The aim is therefore to estimate these from clinical data . Our data set contains 141 patients with carcinoma of the oral tongue diagnosed between 2004 and 2014 and treated at the department of Oncology at Sahlgrenska University Hospital in Gothenburg ( see Methods and S1 Data ) . For each patient we have information about the stage of the primary tumour and the presence/absence of metastasis in LN station I-IV . In our model we assume that secondary seeding is responsible for all metastases that occur in sites not directly connected to the primary tumour . This implies that we expect all patients with station IV positive to also be positive for station III . This is true for all but 2% of the cases ( 3 patients ) , and we therefore exclude these cases from further analysis . Another option would be correct for potentially undetected micrometastases in station III , but we decided for the more cautious option of exclusion . Given the assumption outlined above about a constant downstream flow of cancer cells and the necessity of secondary seeding the state of the system evolves according to a continuous-time Markov chain with state space that corresponds to all possible states of metastatic spread ( see Methods for details ) . The flow rates in the network dictate transition rates for the Markov chain and the probability distribution over the states changes according to a master equation . Since we know the initial state of the system ( no metastasis at tumour initiation ) we can , given a set of parameter values , numerically solve the master equation and obtain the probabilities of all the metastatic states for all future times . This allows us to estimate the model parameters by comparing the metastatic state of each patient with solutions to the master equation and computing the likelihood of the data given certain parameter values ( see Methods ) . In order to visually compare the parametrised model outcome to the clinical data we transform the data in the following way . Let us first focus on station I and let n t I denote the number of patients with stage t primary tumours that are positive in station I . Here t can take the values 1 , 2 , 3 , 4 corresponding to primary tumour stage T1 , T2 , T3 and T4 . Let Nt denote the total number of patients with stage t disease . The fraction of patients with stage t tumour with positive lymph node at station I is then given by p t I = n t I N t . ( 1 ) The same procedure is applied to stations II-IV yielding stage-dependent fractions . The fraction p t i can be interpreted as the probability of finding a patient with stage t disease with a metastasis in LN station i . This quantity can readily be calculated from the model ( see Methods for details ) . A comparison of the clinical data and the model fit is shown in Fig 2 , where each panel correspond to a LN station ( I-IV ) . The estimated dissemination rates are shown in Table 1 together with 95% confidence intervals obtained using parametric bootstrap ( see Methods ) . It is worth noting that the confidence intervals for ϕI and ϕIII obtained from bootstrapping shows that the variability in flow from the station I to station II is large , ranging from practically zero to 0 . 42 , the largest of all rates . Also the flow from station III to station IV exhibits a large variability . With the current amount of the data we can conclude that flow from station II to III has a large impact on the metastatic process , whereas we are unable to ascertain the importance of flow from station I to II and from III to IV . Ovarian cancer has the highest mortality rate of the gynecological cancers and a majority of patients are diagnosed in an advanced stage [42] . Ovarian cancer predominantly metastasises within the peritoneal cavity and through the pelvic lymph nodes [43] . In the peritoneal cavity cancer cells metastasize through a process commonly described as transcoelomic dissemination , where the cancer cells loose cell-cell contact and exfoliate into the peritoneal cavity . They float in the peritoneal fluid and are spread across the peritoneal cavity , where they attach to the peritoneal organs and form a metastatic tumour [44] . Ascites produced in the peritoneal cavity is drained through lymph vessels in the diaphragm [45] , enabling cancer cells to enter into the blood circulation . Historically , hematogenous metastasis has been regarded as occurring only in late stages of ovarian cancer . Recent work however , suggest that this mode of dissemination may be more common than previously thought [46] . In the case of distant metastases , the most common sites are liver , lung , brain and skin [47] . The data set on ovarian cancer was obtained from the SEER-database ( see Methods ) . As of 2010 SEER contains information about the presence or absence of metastases at diagnosis in liver , lung , brain and bone . The status of regional LNs ( including the pelvis and diaphragm ) is also available . Given the available data we did not try to model transcoelomic dissemination , and instead focused on dissemination to local LNs and hematogenic spread to the organs represented in the SEER database . From a primary tumour located in the ovaries cancer cells can thus either spread to regional LNs or via the venous blood vessels to the lungs ( see Fig 3 ) . Metastases in regional LNs also allow for dissemination to the lungs , and from there metastatic lesions shed CTCs to all organs of the body , including the liver , bone and brain ( for which we have data ) . Again our aim is to estimate the dissemination parameters ( λ1 , λ2 , ϕ1 , ϕ2 , ϕ3 , ϕ4 ) from clinical data . In this case the data set is considerably larger containing 16 055 patients diagnosed with ovarian cancer . Again we exclude patients that exhibit skip metastases , leaving us with 15 536 cases ( 3% of patients are excluded ) . The primary tumour stage for this data is more refined and each T-stage is divided into three substages a , b and c , giving us in total 9 different stages ( T1a-T3c ) . We estimate the parameters of this model in precisely the same way as for the tongue cancer model ( see Methods ) . Again we compare the parametrised model to the clinical data by calculating the probability of finding a metastasis in each of the sites considered ( regional LN , lung , liver , brain and bone ) as a function of the primary tumour stage , using Eq ( 1 ) . A comparison between the data and the model is shown in Fig 4 , which shows that the model is able to recapitulate the overall behaviour of the data . Due to the low incidence for bone and brain ( in total 22 and 5 cases respectively ) the data is quite noisy and the model represents a poor fit for those sites . The parameter values are collected in Table 2 together with 95% confidence intervals obtained using parametric bootstrap ( see Methods ) . For this data set the uncertainty in the inferred parameter values is considerably smaller , which can be attributed to the almost hundred-fold larger data set .
We have presented a novel method for inferring the rates of metastatic dissemination , and shown that one can obtain reliable estimates ( with small CIs ) for both lymphatic and hematogeneous spread . Our work builds on previous mathematical models [34] , but with additional biological knowledge we simplify the model structure and make parameter estimates more reliable . Firstly , we make use of the fact that high filtration rates imply that secondary seeding is responsible for metastatic spread beyond the first capillary bed/lymph node . This implies that we do not need to consider all possible links between the sites ( e . g . no direct link between ovary and liver ) . Secondly , we make use of known anatomical structures and flow directions to further prune the network ( e . g . the flow in the lymphatic system dictates the topology of the tongue cancer network , see Fig 1 ) . Lastly , we make use of primary tumour stage as a proxy for time , which means that we can resolve the data temporally . This implies that we do not have to rely on assumptions about stationarity of the underlying metastatic process , and instead fit the parameters for the time-dependent problem . We assumed that tumour stage maps linearly to an arbitrary time scale , which implies that the inferred dissemination rates are estimated with arbitrary and unknown units . In order to investigate how this assumption affected the inferred parameter values we also considered a model with exponential growth of the primary tumour ( see Methods ) . That model resulted in similar parameter values , which suggest that our approach is robust to assumptions about tumour growth dynamics . Our method makes it possible to infer dissemination rates between sites and also their confidence intervals . The model does not fit the data perfectly , and although this may have to do with the scarcity of the data set , we cannot exclude that this may be due to some limitations of the assumptions used in the model . Traditionally this type of data is analysed by looking at incidence rates [48] . Our analysis goes beyond this by disentangling incidence rates into dissemination rates between different lymph node stations/organs . This means that we can quantify processes that at a first glance seem inaccessible , but which appropriate assumptions and modelling techniques can help us reveal . It could be argued that we have used an excessively complex model to fit a data set , which could be described with four ( and five ) straight lines ( see Figs 2 and 4 ) , and hence four ( and five ) parameters corresponding to the slopes of the lines . However , such a model would have no connection to the underlying biology and would be unable to say anything about the dynamics that generated the data . In our model on the other hand the parameters have an immediate biological interpretation . They correspond to the ( relative ) flow rates of cancer cells between different anatomical sites , numbers that could be of interest to clinicians when deciding on the extent of surgical resection or radiotherapy . The magnitudes of the flow rates also reveal the importance of different routes of spread . For example for ovarian cancer we can conclude that the rate of metastasis formation in the regional lymph nodes is ten times higher than the direct spread to the lungs . The flow rates in our model factor in not only physical flow , but also the ability of cancer cells to survive during transport to the target site , and their ability to form metastases in the target site . Assuming that the survival in the circulatory system is independent of target site , knowledge of physical flow would make it possible to estimate the relative rate of metastasis formation in different target sites . This would correspond to effect of the “soil” in the well-established seed-soil hypothesis [49 , 50] . Unfortunately , lymphatic flow is difficult to measure [51] , and there are currently no estimates of lymphatic flow in the head and neck region . Values of relative blood flow are however readily available [52] , and we can use these to calculate soil-effect for ovarian cancer with respect to liver , bone and brain . The relative blood flow to these organs is given by 6 . 5% , 5% and 1 . 2% and by dividing these flow rates with ϕ2 , ϕ3 and ϕ4 , we get relative metastasis formation rates of 1 . 42 , 0 . 3 and 0 . 33 . From this we can conclude that the rate of metastasis formation in the liver is roughly five fold higher compared to bone and brain . This suggests that ovarian cancer cells are considerably better at colonising the liver compared to bone and brain , which is in agreement with previous data [53] . We did encounter a couple of issues when it comes to parameter identifiability . In the case of tongue cancer our model was unable to accurately identify the rates ϕI and ϕIII . In the case of ϕI the problem arises because we are dealing with a site ( station II ) which has flow from both the primary tumour and an upstream site ( i . e . station I ) . With current size of the dataset ( n = 141 patients ) we are unable to obtain accurate estimates of the flow into this station . With a larger dataset we would most likely find more patients who are only positive for station II making it possible to obtain better estimates for λII and consequently ϕI . The large confidence interval obtained for ϕIII is due to the low number of cases with metastases in station IV . In the bootstrap procedure we generate synthetic data based on the estimated parameter values , and since the actual number of patients with station IV positive is small ( only two cases ) we sometimes generate data without any prevalence of metastasis in station IV . This results in estimating the flow rate to ϕIII = 0 , which in turn leads to a large confidence interval . This problem does not appear for the ovarian model since , although the inferred parameter values are small , the large number of patients in the data set imply that we almost surely generate synthetic data with some patients positive for metastases in the brain or bone . From a clinical point of view , this work could be of importance , by contributing to an increased possibility to predict the risk of future regional and/or distant metastases . Especially so in the current era , with new treatment modalities emerging and a current development towards more individualised treatment programs . Although the models are parametrised with population level data they might still be used in order to make predictions on the individual level . For example if a patient is diagnosed with tongue cancer of a specific stage the parametrised model could provide probabilities of different metastatic states , which could be factored in with other clinical data to guide treatment . It would also be possible to include the effect of treatment in the model ( e . g . radiation ) , which would reveal potential benefits of treating metastatic sites and making the model better suited to describe clinical procedure . In conclusion we believe that this framework for analysing metastatic spread , which incorporates known anatomical constraints and a temporal dimension , allows for novel insights and will hopefully be of assistance to both cancer biologists and clinicians in the future .
For the tongue cancer data approval has been granted by Regional ethical review board in Gothenburg . The model consists of N nodes each representing a specific site/organ ( excluding the primary site ) . A node takes the value 0 at time t if the site is void of metastases and 1 if the site contains one or more metastases . We assume that once a site has become positive it will remain so for all future times . Since we have N nodes and each can be in two states we have a state space for the entire network that contains 2N states . Each state is denoted by a binary string ( e . g . 0100 correspond to site 1 , 3 and 4 being negative and site 2 being positive ) , but for notational simplicity we enumerate them with integers from 1 to 2N . The flow rates in the anatomical network dictate transition rates between states . We assume that a site becomes positive at a rate which equals the sum of flow rates from all positive upstream sites . Since the transition rates only depend on the current state ( and not the history ) the system can be described as a continuous-time Markov chain with state space {0 , 1}N . If we let Pi ( t ) , where i = 1 , … , 2N , denote the probability of being in state i at time t , then these probabilities evolve according to the master equation [54]: d P i ( t ) d t = Q i i P i ( t ) + ∑ j = 1 2 N Q i j P j ( t ) where Qii ≤ 0 is the rate of leaving state i and Qij is the rate of moving from state j to i . Since we know that the patient is free of metastasis at tumour intitiation ( t = 0 ) , we know that the initial condition for the master equation is given by P1 ( t ) = 1 and Pi ( t ) = 0 for 1 < i ≤ 2N . Numerical solutions of this equation will be used in order to parametrise the two models . Since we only have access to the primary tumour stage of the patients in the two data sets we make the following assumption about the mapping from tumour stage at diagnosis to time from tumour intitiation to diagnosis . For simplicity we assume that the tumour radius grows linearly with time , which is a simplification , since it is known that volume and hence radius grows non-linearily with time [37] . Tumour stage is informed by the linear size of the lesion [55] , e . g . for tongue cancer T1 corresponds to a primary tumour less than 1 cm in diameter and T2 is a tumour larger than 2 cm , but less than 4 cm . This means that there is a strong correlation between tumour stage and thickness [56] . We therefore assume that time from tumour initiation depends linearly on the stage . Since we have no exact information about the details of this mapping we will for simplicity consider an arbitrary time scale and simply let the stage correspond directly to time since initiation . Below we challenge this assumption by considering a model in which the primary tumour volume grows exponentially . | For most cancer patients the occurrence of metastases equals incurable disease . Despite this fact our quantitative knowledge about the process of metastatic dissemination is limited . In this manuscript we improve on a previously published mathematical model by incorporating known biological facts about metastatic spread and also consider the temporal dimension of dissemination . The model is fit to two different cancer types with very different patterns of spread , which highlights the versatility of our framework . Properly parametrised this type of model can be used for making personalised predictions about metastatic burden . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"cancer",
"detection",
"and",
"diagnosis",
"medicine",
"and",
"health",
"sciences",
"classical",
"mechanics",
"fluid",
"mechanics",
"cancer",
"treatment",
"cancers",
"and",
"neoplasms",
"basic",
"cancer",
"research",
"oncology",
"neurological",
"tumors",
"lymph",
"nod... | 2019 | Inferring rates of metastatic dissemination using stochastic network models |
Many disordered proteins function via binding to a structured partner and undergo a disorder-to-order transition . The coupled folding and binding can confer several functional advantages such as the precise control of binding specificity without increased affinity . Additionally , the inherent flexibility allows the binding site to adopt various conformations and to bind to multiple partners . These features explain the prevalence of such binding elements in signaling and regulatory processes . In this work , we report ANCHOR , a method for the prediction of disordered binding regions . ANCHOR relies on the pairwise energy estimation approach that is the basis of IUPred , a previous general disorder prediction method . In order to predict disordered binding regions , we seek to identify segments that are in disordered regions , cannot form enough favorable intrachain interactions to fold on their own , and are likely to gain stabilizing energy by interacting with a globular protein partner . The performance of ANCHOR was found to be largely independent from the amino acid composition and adopted secondary structure . Longer binding sites generally were predicted to be segmented , in agreement with available experimentally characterized examples . Scanning several hundred proteomes showed that the occurrence of disordered binding sites increased with the complexity of the organisms even compared to disordered regions in general . Furthermore , the length distribution of binding sites was different from disordered protein regions in general and was dominated by shorter segments . These results underline the importance of disordered proteins and protein segments in establishing new binding regions . Due to their specific biophysical properties , disordered binding sites generally carry a robust sequence signal , and this signal is efficiently captured by our method . Through its generality , ANCHOR opens new ways to study the essential functional sites of disordered proteins .
The classical point of view on protein function claims that the functionality of a protein requires the presence of a well-defined three dimensional structure . However , as the amount of experimental evidence against the generality of this concept grew , this paradigm had to be reassessed [1] . It has become evident that there is a large number of proteins that do not require a stable structure even under physiological conditions in order to fulfill their biological role [2]–[4] . These intrinsically unstructured/disordered proteins ( IUPs/IDPs ) lack a well defined tertiary structure and exhibit a multitude of conformations that dynamically change over time and population . The importance of protein disorder is underlined by the abundance of partially or fully disordered proteins encoded in higher eukaryotic genomes [5] , [6] . Disordered proteins are involved in many important biological functions [2] , [7] , which complement the functional repertoire of globular proteins [7] . Recent characterization of IUPs based on their functions shows that disorder can help these proteins to fulfill their functions in various ways [8] , [9] . In the case of entropic chains , the biological function is directly mediated by disorder ( e . g . MAP2 projection domain [10] , titin's PEVK domain [11] , NF-M and NF-H between neurofilaments [12] , [13] , nucleoporin complex [14] ) . Furthermore , disordered segments often act as flexible linkers between folded domains in multidomain proteins [2] , [15] . Alternatively , many disordered proteins function by binding specifically to other proteins , DNA or RNA . This process , termed coupled folding and binding involves a transition from disordered state to a more ordered state with stable secondary and tertiary structural elements [16] , [17] . The coupled folding and binding confers several functional advantages in certain types of molecular interactions . Since – at least partial – folding happens together with binding , the entropic penalty counterbalances the enthalpy gain coming from the binding [18] , [19] . This way disorder uncouples specificity from binding strength allowing for weak transient , still specific interactions that are essential for signaling processes . These properties enable disordered proteins to play an important role in molecular recognition including gene regulation , cell cycle control and other key cellular processes [20]–[23] . The kinetic and thermodynamic details of the binding are influenced by conformational preferences present prior to binding [24] . Although disordered proteins in general lack secondary and tertiary structure , some exhibit partial secondary structure at closer inspection . For example , CD analysis indicated that p21 and p27 possess α-helical segments [19] , [25] , [26] . Detailed NMR characterization of p27 and other proteins showed that several segments can have a pronounced tendency to adopt α-helical , or even β strand conformations [9] . Upon binding , these inherent structural preferences can either be solidified or overwritten by the partner molecule [27] . Some regions can preserve flexibility even within the complex , mitigating the unfavorable entropy term [28] . This allows the fine-tuning of the affinity of interactions over a wide range . As a general rule , however , these interactions are driven largely enthalpically by the favorable interactions formed with the partner molecule [18] , [19] , [29] . The inherent flexibility of disordered proteins offers further advantages in binding . It results in a malleable interface that can allow binding to several partners or to adopt different conformations , manifested in increased binding capability [8] , [20] . In accordance , several analyses of protein interaction networks revealed that disordered proteins are abundant among hub proteins , proteins with a large number of interacting partners [30] , [31] . In a different scenario , the binding partners of an ordered protein are disordered , as shown for binding of 14-3-3 proteins , thus allowing a single protein to bind multiple partners [32] . Beside their involvement in protein-protein interactions , these proteins are also subjects of various post-translational modifications that control their functions , localization and turnover [33] . In this way , these proteins can integrate and mediate multiple signals of various sources , and act as the central elements in signaling or regulatory networks . The centrality of these proteins , however , is also their weakness . It has been suggested that the targeted attack of hubs can cause serious disruption in protein interaction networks [34] . Furthermore , disordered proteins are often associated with various diseases [35] . For example , the primary importance of p53 originates from its involvement of 50% of cancers [36] . In general , 79% of human cancer associated proteins have been classified as IUPs , compared to 47% of all eukaryotic proteins in SwissProt database [22] . Disordered proteins were also suggested to be common in diabetes and cardiovascular diseases [35] , [37] . Several disordered proteins - such as Aβ , τ , α synuclein , and prion protein - are involved in neurodegenerative diseases and are also prone to amyloid formation [38]–[40] . On the other hand , due to their specific way of interactions , disordered proteins can also be attractive targets for drug discovery . A novel strategy for drug discovery exploiting binding sites within disordered regions has already been suggested [41] . This adds further support to the importance of finding specific functional sites in proteins that undergo disorder-to-order transition upon binding or disordered binding regions in short . Despite their importance , the number of well characterized examples of disordered proteins undergoing disorder-to-order transition is very small . The PDB also offers only a limited sample of proteins adopting a well defined conformation as part of a complex . However , recent comparisons of these structures with complexes formed between ordered proteins pointed out several differences [42]–[44] . In general , disordered proteins adopted a largely extended conformation in the complex exposing the majority of their residues for interacting with their partner . The interface of disordered proteins was enriched in hydrophobic residues compared to the interface of ordered proteins , but also to disordered regions in general . The higher number of interchain contacts was suggested to be a sign of better adaptation of disordered proteins to the surface of their partner . In general , the regions that become ordered were shorter as compared to globular domains , usually less than 30–40 residues . While the interface of globular proteins was most often formed by distant segments of the amino acid sequence brought together by folding , disordered binding sites were much more localized in the primary structure . These features demonstrate that the underlying principles of molecular recognition of disordered binding regions are different from the complex formation of globular proteins [43] . Disordered binding sites are also expected to be distinguishable from general disordered sites that are not directly involved in binding . A common notion is that protein disorder comes in many flavors , and these should be targeted by specific prediction methods [45] , [46] . However , training specific methods would require significantly larger datasets than those that are available today . Nevertheless , existing general protein disorder prediction methods might already be equipped for this problem . It has been suggested that specific patterns of disorder prediction profiles can be associated with regions undergoing disorder-to-order transitions [47] . Since these regions can be ordered as well as disordered , there is no clear recipe whether these regions should be predicted ordered , disordered , or as borderline cases . A recent analysis compared several methods to recognize short protein-protein interaction motifs containing α-helical elements in their bound state , the so-called α-MoRFs [48] . As expected , the various methods showed large variations in predicted order/disorder tendency corresponding to binding regions . One of the earliest prediction method PONDR VL-XT [49]–[51] was quite consistent in predicting these regions as ordered within a broader disordered region , giving them the characteristic appearance of dips in the prediction output . Based on this specific prediction output , a method was developed to recognize α-MoRFs from the amino acid sequence [48] , [52] . First , regions predicted with dips in the output of VL-XT were selected and were filtered further by a neural network using several additional properties . This prediction method is restricted to recognize short , α-helical binding regions within disordered proteins . Here we present a general method to identify specific binding regions undergoing disorder-to-order transition . Our method relies on the general disorder prediction method IUPred [53] , [54] . IUPred is based on the assumption that disordered proteins have a specific amino acid composition that does not allow the formation of a stable well-defined structure . The method utilizes statistical potentials that can be used to calculate the pairwise interaction energy from known coordinates . Using a dataset of globular proteins only , a method was developed to estimate the pairwise interaction energy of proteins directly from the amino acid sequence . By virtue of this algorithm , disordered residues can be predicted by having unfavorable estimated pairwise energies . The estimation of the energy for each residue is based on its amino acid type , and the amino acid composition of its sequential neighborhood . Through the amino acid composition of the sequential environment , IUPred can take into account that the disorder tendency of residues can be modulated by their environment [53] . This property of IUPred is exploited in order to recognize regions that are most likely to undergo a disorder-to-order transition based on their estimated pairwise energies in different contexts . The prediction of binding sites is based on estimating the energy content in free and in the bound states , and identifying segments that are potentially sensitive to these changes . In a previous work , the ability to predict specific contacts was emphasized in order to recognize disordered regions that are involved in binding externally rather than internally [46] . In our model , however , there was no attempt made to model specific interactions . Instead , the environment is taken into account simply at the level of amino acid composition . Here we show that this simple model captures the essential property of disordered binding regions and allows their robust prediction . We termed our disordered binding site prediction method ANCHOR , to reflect the primary importance of short segments driving the complex formation between a disordered protein and its partner .
The goal of the present work was to recognize a special class of disordered segments from the amino acid sequence , namely those that are capable of undergoing a disorder-to-order transition upon binding to a globular protein partner . The essential feature of such binding regions is that they behave in a characteristically different manner in isolation than bound to their partner protein . In their free state , they behave as disordered proteins , existing as a highly flexible structural ensemble . In their bound state they usually adopt a rigid conformation , similar to regions within globular structures . This capability to behave in drastically different ways in different environments is targeted by our approach . We seek to identify segments in a generally disordered region that cannot form enough favorable intrachain interactions , however they have the capability to energetically gain by interacting with a globular partner protein . Our prediction is based on three properties . These properties are estimated individually and are combined into a single predictor via optimized weights . In more detail , the prediction of these three properties relies on the energy estimation framework implemented in IUPred , a general disorder prediction method . The core element of IUPred is the energy predictor matrix P . The parameters in Pij were trained on globular proteins with known structures only , without relying on any kind of disordered dataset . These parameters were determined to minimize the difference between the estimated energies and the energies calculated from the known structures on the dataset of globular proteins . Using the energy predictor matrix IUPred predicts the E interaction energy for each residue based on the following formula in default: ( 1 ) where i denotes the type of the k-th amino acid , Pij is the element of the energy predictor matrix that estimates the pairwise energy of residue of type i in the presence of residue type j , is the fraction of residue type j in the sequential environment within w0 residues from residue k . The size of neighborhood considered ( w0 ) equals 100 residues in both directions and the result is smoothed over a window size of 10 ( also in both directions from the k-th residue so in fact 21 residues are considered in total ) . For the final prediction output , the energies are transformed into probability values , denoted as sk . For more details see Dosztányi et al . [53] . The disordered binding site prediction is based on three different scores that are calculated with a slight modification of the original energy estimation scheme . The parameters of Pij were taken directly from IUPred . The following three scores are assigned to each residue in a protein according to the above described criteria ( 1–3 ) : 1 , To measure the tendency of the neighborhood of an amino acid for being disordered we use the IUPred algorithm and assign an Sk score to the k-th residue of the chain by averaging the IUPred scores in the w1 neighborhood of the residue in question: ( 2 ) where sj is the IUPred score of the j-th residue of the chain , N is the number of amino acids in the averaging and blower and bupper are the lower and upper boundaries of the neighborhood of the i-th residue , that is blower = max ( k−w1;1 ) and bupper = min ( k+w1;l ) , where l is the chain length . 2 , We estimate the pairwise interaction energy the given residue may gain by forming intrachain contacts . This is done the exact same way as in IUPred using ( 1 ) , only here the size of the considered neighborhood ( w2 ) is left as a parameter and is set during the training of the predictor: ( 3 ) The smaller window size corresponds to more local behavior . 3 , The pairwise energy that the residue may gain by interacting with a globular protein is approximated using the average amino acid composition of globular proteins: ( 4 ) where is the fraction of residue type j in the averaged reference amino acid composition of globular proteins shown in Table 1 . By subtracting this energy from one can estimate the energy that the residue may gain by interacting with a hypothetical globular protein compared to forming intrachain contacts ( ) . The final prediction score of the residue is given by the linear combination of the above three terms: ( 5 ) where the p1 , p2 and p3 coefficients are determined during the training of the predictor together with the optimal values of w1 and w2 window sizes . Ik is then converted into a p value that expresses the probability of that residue being in a disordered binding site . For a binary classification residues with scores above 0 . 5 are predicted to be in a disordered binding site . Since the second and third terms of ( 5 ) may vary heavily between neighboring residues , the final score is smoothed in a window of 4 residues . The optimal values for the three weights ( p1 , p2 and p3 ) and the two window sizes ( w1 and w2 ) are determined using a dataset of disordered protein complexes and ordered monomeric proteins by three-fold cross validation ( See Methods and Figure S1 for a schematic representation and outline of this procedure ) . The small dataset of known disordered proteins bound to ordered proteins represent a serious bottleneck during optimization . Therefore , it is a clear advantage of our approach that it greatly reduces the dependence on the existing dataset of disordered complexes , and leaves us with only 5 parameters to be optimized on this small dataset . The behavior of various scores is shown for an example , the N terminal domain ( residues 1–100 ) of human p53 tumor supressor protein that plays an important regulatory role [55] . Its N terminal region is completely disordered [56] and is known to be able to bind to ( at least ) three different globular proteins as shown in Figure 1 . The segment between residues 17–27 binds to MDM2 [57] , the other two binding sites overlap with residues 33–56 binding to RPA 70N [58] and residues 45–58 binding to the B subunit of RNA polymerase II [59] . The three calculated quantities for this domain are also shown in Figure 1 . It is worth noting that the MDM2 binding site in the N-terminal region of p53 appears to be on the border of being disordered . Although the disordered prediction is part of ANCHOR , the output of this prediction ( Eint , described in Theory ) is linearly combined with two other quantities meaning that predicted disorder is not strictly a prerequisite of a successful disordered binding site prediction . Testing of the predictor was done by dividing both our negative and positive datasets ( Globular proteins and Short disordered complexes ) into three subsets , training the predictor on two of these and evaluating it on the remaining third one . This was done in all three possible combinations yielding three optimal parameter sets . The parameters calculated on the training sets are shown in Table 2 together with the respective True Positive Rates ( TPR ) and the fraction of the amino acids in disordered regions of the Disprot dataset predicted to be in disordered binding sites ( F values ) . The optimal parameters were chosen to maximize the amount of correctly predicted disordered binding sites ( TPR ) while minimizing predicted binding sites in globular proteins ( FPR ) and also restricting predicted binding sites within disordered regions in general ( F ) . The fact that the three parameter sets do not differ significantly implies that our method is robust . The output of the predictor with all three parameter sets and the combined final predictor ( the average of these three ) are shown for the example of the N terminal region of p53 in Figure 1 . A few additional well characterized examples are shown in the Supporting Information ( Figure S2 , Figure S3 , Figure S4 , Figure S5 , and Figure S6 ) . The results obtained on the three independent testing subsets as well as their average are given in Table 3 . Since the cutoffs are given by the training process such that we achieve exactly 5% False Positive Rate ( FPR ) on the respective training sets ( ie . the part of the original Globular proteins dataset that was used in the training of the respective subpredictor ) , the FPR's are also quoted ( they can differ slightly from 5% ) . Besides the overall TPR calculated on a residue basis ( marked TPRAA ) , we also calculated the percentage of binding sites identified , termed TPRSEG . A binding site was considered to be found if at least five of its amino acids are correctly classified . The results show that ANCHOR performs at 62% TPRAA with a slightly higher TPRSEG of 68% on average , while maintaining a 5% FPR . ANCHOR is also specific to disordered binding sites as opposed to disorder to general . If all disordered proteins had approximately equal capability of binding then the fraction of correctly identified disordered binding sites ( TPR ) could not be significantly different from the fraction of disordered regions predicted to be binding sites ( F value ) . As this is not the case ( TPR = 62% vs . F = 42% ) we can conclude that common features of known disordered binding sites that distinguish them from general disordered protein regions are successfully recognized . Another standard way of describing prediction algorithms is by Receiver Operating Characteristic ( ROC ) curves [60] , that is the TPR versus the FPR of the algorithm . This relationship is mapped by scanning the interval between 0 and 1 with the score cutoff . The three ROC curves of the predictor with the three different parameter sets evaluated on the respective testing sets are shown in Figure 2 . A single number measure to characterize the performance is the area under the curve ( AUC ) with random predictors scoring AUC = 0 . 5 and perfect predictors scoring AUC = 1 . The AUC values of the predictors trained and tested on the respective subsets are 0 . 8675 , 0 . 8781 and 0 . 8993 . Since the interacting regions of a disordered and an ordered protein are inherently different we expect that the predictor will only recognize binding sites in disordered proteins that interact with globular proteins but are not part of globular proteins themselves . In order to verify this hypothesis we tested the combined final predictor on a dataset of complexes containing only ordered chains ( that is three-state complexes – see Methods ) . The prediction was done on the short interacting chain of the complexes . This gave a false positive rate of only 3 . 7% that is even lower than the value obtained on our testing set , although this might be only a consequence of the relatively small size of our ordered complex set ( 72 complexes ) . Overall , we could ensure that our predictor makes very few mistakes on both globular proteins and complexes of globular proteins , while it can still recognize the majority of disordered binding regions . This implies that our algorithm is specific to disordered binding sites as opposed to globular proteins , the interface between globular proteins or disordered proteins in general . Our predictor was also tested on a completely independent dataset of α-MoRFs , short disordered complexes that was assembled by Cheng et al . [48] and composed of 40 proteins containing binding regions that adopt mostly α-helical structure upon binding . The results of the prediction on this dataset can be seen in Table 4 . Although the residue based TPR is somewhat lower than that calculated on our testing set ( 57 . 0% instead of 61 . 8% ) , the segment based TPR is almost the same for the two sets ( 67 . 5% and 68 . 3% ) . Overall these results are comparable to the ones calculated on our training set . The specific construction of the algorithm for the prediction of interaction energy implies that the method will be sensitive to amino acid compositions . The differences between the composition of disordered binding sites and the amino acid composition of any of the negative sets ( globular proteins , ordered interfaces and disordered proteins in general ) are shown in Figure 3A , 3B , and 3C , respectively . The amino acid compositions of all three datasets are significantly different from that of disordered binding segments ( data not shown ) . The final prediction is based on three different scores that combine local and global disorder tendency with sensitivity to the structural environment . Although the individual quantities that are combined for the final score can work selectively better or worse for various types of residues , the effect of these differences on the efficiency of the final prediction is not trivial . This effect was tested by comparing the amount of the different amino acids in the short disordered binding sites to the amount recovered from these by the predictor . These data are shown in Table 5 together with the calculated p values quantifying their differences . As all of the p values are fairly large , these differences are likely to occur by chance alone . For example , proline rich binding sites are found with similar accuracy as binding sites enriched in hydrophobic amino acids . Therefore , one may conclude that there is no statistical evidence based on the available dataset that the efficiency of the predictor depends significantly on the amino acid composition of the disordered binding site in question . The relationship between the efficiency of the prediction and the secondary structure types was also assessed , by considering the three types of secondary structural elements: helix ( H , including α- and 310 helices ) , extended ( E ) and coil ( C , including everything else ) as defined by DSSP [61] . The number of amino acids in different conformations that can be found in the PDB structures of our positive training set ( short disordered complexes ) , in the interacting residues of these structures and the interacting residues that are correctly identified by the predictor are shown in Table 6 . These data are represented graphically as distributions in Figure 4 . The secondary structure content in this type of interactions is heavily biased towards coil conformation . It can also be seen on Figure 4 that the predictor seems to work slightly better for H and E conformations . However assessing the difference of the distributions of secondary structures in interacting residues and in the subset identified correctly by ANCHOR shows that this difference is not statistically significant at a 5% level ( χ2 = 5 . 32 , p = 0 . 070 ) . Furthermore , a similar result holds true if binding sites are categorized based on their dominant secondary structure type - that is there is no significant correlation between the secondary structure type the binding regions adopt upon binding and the efficiency of the predictor . ( Dataset S1 shows the secondary structure types determined for the short disordered chains in the disordered complexes as described in Protocol S1 . ) Overall , this means that there is no significant difference in the efficiency of the prediction on different secondary structural elements . Since the predictor was trained on the short disordered dataset it is informative to see how it performs on long disordered binding sites . There is experimental evidence that at least some long disordered chains are not uniform concerning binding strength but contain short stretches of strongly interacting residues separated by segments that interact with the partner only weakly if at all [19] . In these cases , it is expected that the predictor will be unable to identify the weakly interacting parts since – though these parts may also form interchain contacts – they would not be able to bind to the partner in the absence of their sequential neighbors . The distribution of predicted binding regions for the short and long disordered chains in Figure 5A shows a strong preference for predicting multiple interacting regions for longer chains . This inevitably yields lower residue based TPR but the segment based TPR is not expected to drop . Testing the predictor on the long disordered data confirms this assumption with a decreased residue based TPR of 47 . 7% ( as opposed to 65 . 8% obtained on running the final predictor on the whole set of short disordered complexes ) but with a basically unchanged segment based TPR of 78 . 6% ( compared to the 76 . 1% calculated on short disordered complexes ) . These data suggest that the method either finds short disordered binding sites as a whole or completely misses it . However , this may not be true for long binding regions . Figure 5B shows the distribution of the fraction of amino acids successfully identified during prediction in the two types of binding sites . The effect can clearly be seen as about 59% of short binding regions are either fully recovered or are completely missed ( the sum of the rightmost and leftmost columns ) whereas this ratio is only about 29% for long binding sites . This type of behavior is illustrated on the disordered human p27 . This protein is involved in controlling eukaryotic cell division through interactions with cyclin-dependent kinases . Its kinase inhibitory domain binds both subunits of the CDK2-cyclin A complex in an extended conformation ( PDB ID: 1jsu [62] ) . It is known from kinetic measurements that the binding of p27 is hierarchical through its three domains: first , the D1 domain ( residues 25–36 ) binds to cyclinA which anchors the neighboring LH domain ( residues 38–60 ) that exhibits transient helical structure in monomer state as well [63] . After the binding of D1 this transient structure is stabilized and positions the rest of the chain ( D2 domain , residues 62–90 ) in the correct position to bind to CDK2 . Figure 6 shows the prediction output for p27 . Four interacting regions are identified with the first one ( 27–37 ) clearly corresponding to D1 . The gap between the first two regions ( 38–58 ) coincides with the weakly interacting LH domain . The last three regions ( 59–67 , 74–77 and 79–90 ) cover the strongly interacting D2 . Figure 6 also shows the number of atomic contacts/residue for p27 ( averaged in a window of size 3 ) . This contact number profile exhibits well pronounced peaks that line up with the regions that are predicted by our algorithm . The figure also shows the four predicted regions mapped to the crystal structure of the complex . The examples discussed so far represent various fragments of proteins . Here we present an additional case showing the prediction output for a complete protein sequence . The human Wiskott-Aldrich Syndrome protein ( WASp ) is a 502 residue long protein that is expressed in the cells of the hematopoietic system [64] . Its mutations can be linked to the Wiskott-Aldrich Syndrome ( WAS ) , a disease characterized by actin cytoskeleton defects leading to deficiencies in blood clotting and immune response . The protein is composed of various functional domains . It contains the WH1 domain near the N terminus ( residues 39–148 ) , the GTPase-binding domain ( GBD , 230–310 ) , a polyproline-rich region and a C-terminal verpolin homology/central region/acidic region ( VCA , 430–502 ) domain [65] that also contains the WH2 domain ( 430–447 ) . Apart from the structured WH1 domain , it is predicted to be largely disordered and contains several low complexity regions ( enriched in P , G and acidic amino acids ) . There is experimental evidence that the activated WASp hubs a number of interactions with partners including CDC42 , RAC , NCK , FYN , SRC kinase FGR , BTK , ABL , PSTPIP1 , WIP , and the p85 subunit of PLC-gamma as well as the Arp2/3 complex . However , the location of many of these binding regions is not known . The domain structure of WASp is shown in Figure 7 together with the known binding regions . In its inactive state WASp exists in an autoinhibited form with the GBD domain bound to the VCA domain . When WASp is activated , the GBD domain is bound to CDC42 and this interaction disrupts the GBD-VCA interaction . This initiates a conformational change where WASp opens up and becomes able to bind to the Arp2/3 complex leading to its activation and actin nucleation . Both GBD and VCA regions were shown to be disordered in their free state [65] , [66] , with GBD adopting a loosely packed , compact conformation . However , the structure of both complexes could be determined using NMR , by covalently linking GBD to CDC42 or the VCA region , respectively [65] , [67] . In these two structures WASp GBD adopts related but distinct folds . The plasticity that can be seen by comparing these two complexes is enabled by the absence of discrete tertiary structure in isolation . As it can be seen on Figure 7 , ANCHOR captures these disordered binding sites correctly . It is known that WASp is able to bind to SRC Homology 3 ( SH3 ) domains through one of its proline rich regions although the exact binding site is not known . The interaction with SH3 domains is usually mediated by a short , linear sequence motif that is present in the interaction partner . In the collection of Eukaryotic Linear Motifs ( ELM ) database ( http://elm . eu . org/ [68] ) there are five different motifs annotated as SH3 recognition sites . Multiple instances of the following three can be found in human WASp: LIG_SH3_1 , LIG_SH3_2 and LIG_SH3_3 represented by the following consensus sequences: [RKY] . . P . . P , P . . P . [KR] and …[PV] . . P , for interaction with Class I/ClassII SH3 domains and those SH3 domains with a non-canonical Class I recognition specificity , respectively . The found motifs are clustered in two separate regions mainly falling into the proline-rich regions of WASp ( Figure 7 ) . Although there is no direct evidence for the location of interaction with SH3 domains on human WASp , the interaction sites have been identified for Las17 [69] , the yeast homologue of this protein . In total , four distinct regions containing multiple binding sites were identified experimentally in Las17 that interact with various SH3 domains . These sites correspond to the proline rich regions in WASp ( 155–194 and 306–427 ) that also match with several SH3 binding motifs . As linear motifs were shown to have a preference to reside in disordered regions [70] , it is plausible to expect ANCHOR to be able to recognize the SH3 binding region of WASp . In accordance with this , both regions containing putative SH3 binding sites contain binding sites predicted by ANCHOR . This prediction can restrict the candidate sequence regions for SH3 binding and can guide experimental studies to localize true binding sites . In order to gain some evolutionary insight concerning disordered binding sites , the predictor was run on the 736 complete proteomes ( 53 archaea , 639 bacteria and 44 eukaryota , see Dataset S5 , Dataset S6 , and Dataset S7 , respectively ) that are currently available from the SwissProt database ( ftp://ftp . expasy . org/ ) . In agreement of previous analyses [5] , [6] there is a clear trend of increasing amount of protein disorder as the complexity of the organism increases ( see Figure 8 ) . However , Figure 8 also shows that the fraction of disordered amino acids predicted to be in disordered binding sites increases even compared to fraction of disordered residues , as the complexity of organisms grows . Generally , archaea have the least amount of both disorder and binding sites . On the other hand , eukaryota have generally the largest ratio of disordered and binding amino acids with bacteria being between these two groups on average . However there are a few exceptions to these general trends , marked separately on Figure 8 . Considering archaea , mesophiles generally contain a larger amount of disorder and a larger fraction of disordered binding sites than most extremophiles ( thermophiles , cryophiles and acidiphiles ) . However the group of halophile archaea ( archaea that favor high saline concentration ) is a distinct exception with fraction of disordered amino acids ranging from 0 . 2 to 0 . 25 as opposed to other extremophiles' values not exceeding 0 . 07 . This group includes all the halophile archaea in our study , namely Natronomonas pharaonis , Haloarcula marismortui , Haloquadratum walsbyi and two types of Halobacterium salinarum . Cenarchaeum symbiosum , the only example of obligate endosymbiont among archaea also has an unusually large amount of disordered protein segments in its proteome ( 0 . 12 ) . While Cenarchaeum symbiosum is closely related to thermophile archaeas , it is adopted to the much lower living temperature of its host [71] . This adaptation could explain the relatively large amount protein disorder and disordered binding sites . In general , these clear differences in the predicted disorder between various archaea organisms points to different strategies to adapt to various extreme environmental conditions resulting in biased amino acid compositions . However , we cannot rule out the possibility that under such extreme conditions , as high salt concentration or high temperature , the amount of disorder can be over- or underpredicted depending how these conditions affect the presence of protein disorder . Among bacterial proteomes , there are a few examples of organisms that seem to utilize a surprisingly large fraction of their disordered amino acids in binding . The three most extreme cases ( Carsonella ruddii , Sulcia muelleri and Buchnera aphidicola subsp . Cinara cedri ) are marked separately on Figure 8 . These are the three smallest complete bacterial proteomes , none of them reaching the size of the smallest archaea proteome . These organisms present extreme cases of streamlined genomes as a result of endosymbiosis [72]–[74] . As these proteomes are very small , the predicted amount of disorder and disordered binding sites are within the false positive range , and should be treated more cautiously . Eukaryotes tend to appear more consistent both in using larger amount of disordered residues and larger fraction of disordered residues for binding compared to the other two kingdoms ( Figure 8 ) . The only notable outlier both in terms of extremely low amount disordered proteins and disordered binding sites is Encephalitozoon cuniculi . This organism is the only microsporidian parasite in our dataset and has an extremely small proteome . This lack of complexity and dependence on a eukaryotic host to function might explain the lack of disordered proteins . The length distributions of the predicted disordered regions and binding sites in the three kingdoms of life was also analyzed and are shown in Figure 9A and 9B , respectively . As complexity increases , longer disordered segments are preferred , and the difference between eukaryota and lower complexity organisms becomes even more apparent for longer regions ( over 30 residues ) . A similar trend can be observed in the length distribution of disordered binding sites . While in archaea and bacteria predicted binding regions are generally below 30 residues , longer binding sites in eukaryota organisms are much more common . There are at least three different effects that can contribute to this phenomenon . First , as the number of binding sites rise there is also an increasing possibility of these binding sites becoming very close to each other or even overlapping with each other . This scenario was demonstrated in the case of the N-terminal domain of p53 , as shown in Figure 1 . Second , extremely large disordered binding regions may be needed for special functions . Some members of the mucin protein family provide an example for this . Human MUC1 contains a large repeat region ( 20–120 repeats , one repeat being 20 amino acids long ) that enables it to aggregate and to perform its function [75] . As each repeat is correctly identified as a disordered binding site , the whole repeat region is predicted as one large binding region . This mechanism can create binding sites up to the length of several hundreds of residues in extreme cases . Third , we cannot exclude the possibility that longer binding sites are not always segmented by weakly interacting regions like in the case of p27 , thus forming long , continuous binding regions . Nevertheless , the majority of predicted binding sites is shorter than 30 residues , although such restriction on the length of disordered binding sites was not enforced .
Regions undergoing disorder-to-order transitions upon binding are essential elements in the molecular recognition process involving disordered proteins . The main property of these binding regions is that they can exist in a disordered state as well as in bound state , adopting at least partially a well-defined conformation . The presence of these two separate states discriminates them from monomeric globular proteins as well as from complexes formed between globular proteins and from disordered proteins in general . They are also expected to differ from dual personality fragments [76] , which occur within globular proteins , however , mostly as a result of perturbations of environmental conditions . In this work we aimed to recognize such disordered binding regions from the amino acid sequence . So far , the limited number of well characterized examples hindered the development of general prediction methods . Nevertheless , biophysical considerations suggest that in most cases there is a strong signal in the amino acid sequence highlighting regions involved in coupled folding and binding . These regions are linear in sequence , unlike in the case of globular proteins , where distinct sites in the amino acid sequence are brought together to form the interface for interaction [43] . An additional difference is that binding of disordered proteins is driven by a large enthalpic component to compensate for the entropy penalty due to the loss of conformational freedom [9] . These features result in a relatively short sequence segment containing residues with a pronounced tendency to make interactions , leading to a characteristic sequence signal . Our approach relies on a basic physical model of disordered binding sites and it is based on modeling the interaction capacity in the free disordered state and in the bound ordered state . Previously , it was shown that ordered proteins can be discriminated from disordered proteins based on estimated pairwise energy content and this approach was implemented in IUPred , a general disorder prediction method [53] . This method takes into account that disorder/order tendency can be modulated by the sequential neighborhood simply at the level of amino acid composition , without attempting to model the specific interactions . Taking it one step further , the same energy estimation calculations were used to identify disordered binding regions in proteins . Our model assumes that the specific properties of disordered binding sites are dictated by the combination of preferences to bind to an ordered protein on the one hand , and the ability to remain in a disordered state in isolation , on the other . Based on this simple model , ANCHOR achieved approximately 67% accuracy at predicting 5% false positive rate ( Tables 2–4 ) . Furthermore , this approach was validated by the ability to reproduce the specific amino acid composition of disordered binding sites , that is distinct from that of ordered proteins as well as disordered proteins in general ( Table 5 ) . During binding , the formation of intermolecular contacts is accompanied by the formation or the stabilization of secondary structure elements . The secondary structure composition of the binding sites is highly unequal ( Table 6 and Figure 4 ) . The most dominant secondary structure element adopted in the bound conformation is coil , while β strand conformation is rare . Helical conformations are observed as frequently in disordered complexes as in globular proteins [27] . It was found that the adopted secondary structure can be predicted from the amino acid sequence with similar accuracy as in the case of globular proteins , suggesting that the adopted secondary structure can be imprinted into the sequence of the binding motif [27] . The secondary structure observed in the complex can also be dictated by the template structure . An extreme example of this is the C-terminal region of p53 ( see Supporting Information ) , observed in all three secondary structure classes [32] . It is clear that not all of these conformations can be the result of inherent preferences . Interestingly , our prediction method does not seem to be sensitive to the adopted secondary structure conformation and it works with the same accuracy for all secondary structure conformations ( Table 6 and Figure 4 ) . This independence of secondary structure elements underlines the generality of ANCHOR . These results also suggest that disordered binding sites can be recognized without taking into account of the adopted secondary structure in the majority of cases . Nevertheless , the details of conformational preferences can be still crucial in selecting the specific binding partner , or determining the kinetic and thermodynamic properties of the associations . Beside our algorithm , a previously published method called α-MoRF predictor also exploited a general disorder prediction method to recognize short binding elements [48] , [52] . Although the direct comparison between the two methods was not possible , because the α-MoRF predictor is not yet publicly available , some basic differences between the two methods should be noted . First , the α-MoRF predictor directly relies on the prediction output of PONDR VXLT , which essentially predicts binding regions as ordered structural elements , and a subsequent neural network is applied to filter out valid disordered binding sites . Although very high accuracies were reported for the performance of the neural network based filtering , the complete method is limited by finding dips based on PONDR VLXT [49]–[51] . Therefore it should be taken into account that this program is a first generation prediction method that was trained on only 15 proteins . In the case of IUPred , dips corresponding to certain binding sites were also observed , although to a smaller extent [48] , [53] . This observation , however , is not directly exploited in our prediction method . Instead , the core parameters of the energy prediction of IUPred are used to create three separate scores characterizing three important attributes of disordered binding regions . The second main difference is that ANCHOR is not restricted to a single secondary structure class like the α-MoRF predictor that was trained to recognize only α-helical segments . The example of the C-terminal region of p53 ( Figure S2 ) , where four short overlapping regions were shown to bind in different conformations representing all three secondary structure classes , indicates that such restriction can be a serious disadvantage for recognizing some extremely adaptable disordered binding motifs . An alternative approach for binding site identification is based on the observation that protein-protein interactions are often mediated through short linear motifs ( approximately three to eight residues ) [77] . Such motifs are defined by a consensus pattern , which captures the key residues involved in function or binding . Prominent examples include the nuclear receptor box motif , MDM2 binding sites , SH2/SH3 domain recognition patterns or 14-3-3 domain binding sites [68] . Although there are known examples of motifs that reside within globular domains , many of them are required to be in a disordered region to function properly and it was suggested that such motifs share many similarities with disordered binding regions [70] . Our preliminary results support previous observations of the partial overlap between short linear motifs and disordered binding segments . Nevertheless , short disordered binding sites and sequence specific linear motifs capture different aspects of certain binding regions . Linear motifs are defined on the basis of a per residue binding strength , and they are specific to a certain partner or to a group of partner molecules . However , such short linear motifs can also occur purely by chance , with no biological significance . Also , sequence patterns alone cannot ensure the accessibility of the site and the potential flexibility of the binding region that could be necessary for the complex formation . Complementary to sequence motifs , ANCHOR aims to capture a broader structural context . Based on their specific structural properties , it can recognize disordered binding regions that are capable of undergoing disorder-to-order transition . The predictions are made without taking into account the partner molecules and are expected to be less sensitive to sequence details . For certain motifs , this molecular environment can be a prerequisite of functionality and could help to identify biologically significant binding motifs . In our work we assumed , that short binding regions undergoing disorder-to-order transition can be viewed as elementary binding units that are necessary for the molecular recognition . Therefore , such examples were used for the optimization of our method . In accordance with their elementary unit picture , ANCHOR recognized them generally as a single continuous binding site ( Figure 5 ) . Regions undergoing disorder-to-order transition , however , are not limited to such short segments as there are several examples of longer disordered segment becoming ordered upon complex formation . Such segments can be as long as 100 residues . However , these longer regions can contain segments which bind only weakly or might not become ordered at all [63] , [78] , [79] . This segmentation of longer binding regions can occur for structural reasons . The segmentation can prevent the accumulation of the critical amount of residues that would lead to the formation a collapsed structure or non-specific aggregates . The possible functional advantages of the segmented nature of a binding site were demonstrated for the well characterized example of p27 . The kinase inhibitory domain of p27 can be divided into several subdomains which dock and fold in a stepwise manner on the surface of the Cdk2-cyclin A complex [19] . These segments can also evolve independently , increasing the repertoire for specificity for different cellular location or species . Intervening segments of higher flexibility are accessible for modifications such as phosphorylations and ubiquitinations . This way p27 can integrate and process various signals to regulate cell proliferation , in which the flexibility and modularity of p27 is essential [63] . The segmented nature of binding is reflected in the prediction output , with predicted binding sites corresponding to the strongly interacting regions ( Figure 6 for p27 , and Figure S4 for a similar example , calpastatin ) . In the dataset of longer disordered binding segments , we found this segmentation to be quite general . In these cases , the predicted sites generally give only partial coverage of the PDB structure , and multiple binding sites are predicted in the majority of cases ( Figure 5 ) . This suggests that our prediction method is likely to find those sites that interact more strongly , anchoring the disordered segments to their partner protein . While the segmented nature of binding is prominent in the case of long binding regions , to a smaller extent , it can also affect shorter binding regions . Indeed , around 20% of short disordered binding regions are predicted as 2 or 3 segments ( Figure 5 ) . This could also account for the significantly lower per residue efficiency compared to the segment based efficiency . By looking at further individual examples , one can already see remarkable variations in the details of disorder-to-order transitions even within the limited collection that is available today . The adopted conformation in these complexes can be quite different , both in terms of secondary or tertiary structure . Furthermore , the transition to an ordered structure might not be complete [28] . This could leave terminal residues or linker regions flexible and inaccessible to structure determination . It was also suggested that specific binding can be possible even without adopting a well-defined conformation as in the case of the ζ-chain of T-cell receptor [80] ( see Figure S6 ) . Differences are also present at the level of the sequence . Some binding regions rely largely on hydrophobic or aromatic residues ( MDM2 binding regions , Figure 1 ) , others use proline rich regions ( WASp SH3 binding regions , Figure 7 ) . Disordered binding regions can contain conserved linear motifs , while large divergence in sequence was noted in other cases ( C terminal domain of histones [81] ) . These examples represent multiple ways disordered regions can be utilized for binding . A single protein sequence can contain several distinct binding regions , however , a single region can be involved in binding to multiple partners , or use these regions in combination to hub several interactions ( p53 – see Figure 1 and Figure S2 , WASp – see Figure 7 ) . In an alternative scenario , disorder present in the partner molecules allows to bind a well-folded protein by a large number of proteins ( β-catenin [82] , Figure S3 ) . Even further variations are expected as the number of examples will grow in the future . Nevertheless , the success of ANCHOR confirms our hypothesis , that despite these differences disordered binding regions have a common property that predispose them for coupled folding and binding . The occurrence of disordered binding sites is clearly tied to the presence of disordered protein regions . Their relationship was further analyzed at the level of complete proteomes . Previous studies have shown that the amount of predicted disordered regions increases with the complexity of organisms throughout evolution and reaches a high level in multicellular organisms [5] , [6] . This increase can be mostly attributed to the appearance of long , domain-sized segments of protein disorder or fully disordered proteins ( Figure 9A ) . Our analysis showed that the amount of disordered binding segments increases in eukaryotes in a similar way , however , their fraction is elevated even compared to disordered regions in general ( Figure 8 ) . The observed trend is valid through a wide range of organisms , and occasional exceptions occur either due to adaptation to extreme habitat conditions , or as a result of endosymbiosis . These findings imply that the newly introduced disordered proteins and protein segments mainly serve as a carrier for new binding regions in eukaryotic organisms . The importance of disordered regions in protein-protein interactions is also supported by the increased ratio of disordered proteins among hub proteins [30] , [31] . Disordered segments are often involved for complex signaling and regulatory processes [20] such as cell cycle control , gene regulation or signal transduction in the intracellular region of transmembrane proteins [83] . These processes rely on interactions involving multiple partners and high specificity/low affinity interactions , that disordered binding segments can provide by their very nature . The disordered segments can harbor multiple binding sites which can act relatively independently . In other cases segmented binding sites can be involved in simultaneous binding to larger complexes . Overlapping binding sites ( such in the case of p53 N and C terminal regions ) suggest competition between binding partners . We are only beginning to comprehend how disordered binding regions are exploited to provide versatile interaction sites in proteins . In conclusion , disordered binding regions represent a specific subclass of disordered proteins that can undergo a disorder-to-order transition upon binding . These binding sites generally have distinct properties both structurally and functionally . Due to the inherent flexibility , these regions are difficult to study experimentally [84] , making specific prediction methods even more valuable . While there are several methods available for prediction of disordered regions [85] , [86] , recognizing disordered binding sites was regarded as a more challenging problem [9] due to the limited number of well-characterized examples . In this work we report a general method to recognize disordered binding sites based on a basic biophysical model . Our method relies on a simple energy estimation procedure that was developed earlier for the IUPred disorder prediction method . This way , the problem of small datasets can be largely avoided . We showed that these regions can be characterized by highly disordered sequential neighborhood , unfavorable intrachain energies and more favorable interaction energies with a globular partner . The combination of these properties allowed the recognition of disordered binding sites independent of their secondary structure or amino acid composition , underlining the generality of the method . As such binding sites are essential functional elements of disordered proteins , their prediction directly provides information about functionally important residues in these proteins . In this way , ANCHOR broadens the repertoire of prediction methods for functional sites in proteins aiming to decrease the large number of unannotated sequences [87] . Generally , the complete understanding of protein-protein interactions involving disordered binding sites requires the knowledge of their partners as well as possible post-translational modifications that can influence their binding . While predictions can be made even without taking the partner molecule into account , certain cases might require incorporating the specific feature of the partner . Nevertheless , our method can provide the starting point for such scientific explorations , by finding potential regions involved in such binding .
The primary source of data for the present analysis is a carefully assembled dataset of binding regions undergoing disorder-to-order transition . The strict requirement of the experimental verification of both the disordered status in isolation and the formation of an ordered structure in complex distinguishes our dataset from a previously collected dataset for disordered binding regions [88] . The length of disordered regions involved in the binding can vary on a large scale . In the case of longer regions it is not guaranteed that each residue is equally important for binding , therefore complexes of short disordered regions were treated separately , and only these were used for tuning the method . The optimal parameters were determined by a three fold cross-validation , by dividing both our negative and positive datasets ( Globular proteins and Short disordered complexes , respectively ) into three parts . In each turn we used two parts for training and the remaining part for testing . To avoid any bias , the different subsets were chosen such that the distribution of chain lengths in both the positive and negative sets and the distribution of secondary structure types in the positive set were approximately the same . Our approach relies on IUPred , a general disorder prediction method , and its energy predictor matrix . These parameters ( ie . the elements of the energy predictor matrix ) have been determined earlier , independently of disordered binding regions . Only five additional parameters , w1 , w2 , p1 , p2 and p3 were optimized for this specific problem and were selected by a grid search procedure . Specifically , w1 was varied in the range of 20 to 100 in steps of 10 ( giving 9 possible values ) , w2 was varied in the range of 5 to 35 in steps of 2 ( giving 16 possible values ) , and p1 , p2 and p3 was selected from 1000 sets of randomly generated values . Taking into account that the prediction performance is insensitive to the norm and the sign of the vector corresponding to the p1 , p2 and p3 values , the search was restricted to 1000 random sets that were evenly distributed on the surface of the upper half of the unit sphere . This means that p1 and p2 were randomly selected from the interval [−1;1] and p3 was selected from the interval [0;1] in a way that the sum of their squares is always equal to 1 . This yielded 1000 different ( p1 , p2 , p3 ) combinations . These , combined with all possible values of w1 and w2 gave 144 , 000 different parameter sets in total . These were considered in order to select the optimal one , containing the five optimal parameters for each round of the cross-validation . To quantify the performance of the predictor given a set of parameters we calculated the True Positive Rate ( TPR ) at False Positive Rates ( FPR ) fixed at 5% calculated on globular proteins as the negative set . However , a full characterization of the performance of the algorithm would also require a set of disordered proteins that are known not to bind to globular proteins . Unfortunately , such dataset cannot be constructed since there is hardly any way to give evidence for a protein that it does not contain binding sites . This problem was addressed by calculating the fraction of amino acids that are predicted as binding sites in general disordered regions of Disprot database that are correctly recognized as disordered by IUPred . This fraction was denoted as F . Optimal parameters should combine high TPR with low F at the expense of very low FPR . During optimization of the algorithm , the performance on three different datasets needed to be monitored at the same time ( set of globular proteins , set of disordered binding sites and Disprot ) . The best parameter set was chosen manually , by reducing the parameter set in a step-wise manner based on the following steps: 1 , Calculate TPR ( at fixed FPR = 5% ) and F for each of the 144 , 000 candidate sets of parameters 2 , Discard all for which F>50% 3 , Discard all for which TPR<60% 4 , From the remainder choose the 20 for which the difference between TPR and F is the largest 5 , Choose the one for which TPR is maximal ( the TPR-F difference among these 20 sets vary only within a range of less then 0 . 02 so that is not a good measure to choose the best one ) The negative and positive sets were divided into three parts , resulting in three different optimal parameter sets . The final predictor algorithm is constructed by averaging these three outputs . As the training sets only contained binding regions of at least 10 amino acids and we aim to identify at least 5 residues of each region , all predicted binding sites were removed that did not exceed 5 consecutive residues . A schematic figure of the training procedure is given in Figure S1 . ANCHOR is available upon request from the authors . | Intrinsically unstructured/disordered proteins ( IUPs/IDPs ) do not adopt a stable structure in isolation but exist as a highly flexible ensemble of conformations . Despite the lack of a well-defined structure these proteins carry out important functions . Many IUPs/IDPs function via binding specifically to other macromolecules that involves a disorder-to-order transition . The molecular recognition functions of IUPs/IDPs include regulatory and signaling interactions where binding to multiple partners and high-specificity/low-affinity interactions play a crucial role . Due to their specific functional and structural properties , these binding regions have distinct properties compared to both globular proteins and disordered regions in general . Here , we present a general method to identify disordered binding regions from the amino acid sequence . Our method targets the essential feature of these regions: they behave in a characteristically different manner in isolation than bound to their partner protein . This prediction method allows us to compare the binding properties of short and long binding sites . The evolutionary relationship between the amount of disordered binding regions and general disordered regions in various organisms was also analyzed . Our results suggest that disordered binding regions can be recognized even without taking into account their adopted secondary structure or their specific binding partner . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"molecular",
"biology/bioinformatics",
"biochemistry/bioinformatics",
"computational",
"biology/macromolecular",
"sequence",
"analysis",
"biochemistry/theory",
"and",
"simulation",
"evolutionary",
"biology/bioinformatics",
"computational",
"biology",
"computational",
"biology/systems",... | 2009 | Prediction of Protein Binding Regions in Disordered Proteins |
In metabolic networks , metabolites are usually present in great excess over the enzymes that catalyze their interconversion , and describing the rates of these reactions by using the Michaelis–Menten rate law is perfectly valid . This rate law assumes that the concentration of enzyme–substrate complex ( C ) is much less than the free substrate concentration ( S0 ) . However , in protein interaction networks , the enzymes and substrates are all proteins in comparable concentrations , and neglecting C with respect to S0 is not valid . Borghans , DeBoer , and Segel developed an alternative description of enzyme kinetics that is valid when C is comparable to S0 . We extend this description , which Borghans et al . call the total quasi-steady state approximation , to networks of coupled enzymatic reactions . First , we analyze an isolated Goldbeter–Koshland switch when enzymes and substrates are present in comparable concentrations . Then , on the basis of a real example of the molecular network governing cell cycle progression , we couple two and three Goldbeter–Koshland switches together to study the effects of feedback in networks of protein kinases and phosphatases . Our analysis shows that the total quasi-steady state approximation provides an excellent kinetic formalism for protein interaction networks , because ( 1 ) it unveils the modular structure of the enzymatic reactions , ( 2 ) it suggests a simple algorithm to formulate correct kinetic equations , and ( 3 ) contrary to classical Michaelis–Menten kinetics , it succeeds in faithfully reproducing the dynamics of the network both qualitatively and quantitatively .
Our approach to PINs relies on a modified QSSA introduced by Borghans , DeBoer , and Segel in 1996 [7] . They proposed that , for conditions when ET and S0 are comparable numbers , the proper intermediate timescale variable is Ŝ ( t ) = S ( t ) + C ( t ) . In terms of this variable , the governing equations are Borghans , DeBoer , and Segel called this the total QSSA ( tQSSA ) . A sufficient condition for the uniform validity of the tQSSA was derived by Tzafriri [8]: where ST = S ( 0 ) + C ( 0 ) , ɛ ( ET , ST ) = ( k2/2k1ST ) · f ( r ( ET , ST ) ) , f ( r ) = ( 1 – r ) −1/2 − 1 , and r ( ET , ST ) = 4ETST ( Km + ET + ST ) −2 . Tzafriri showed that Hence , Equation 3 is satisfied if k−1 ≫ k2; i . e . , the dissociation rate of the enzyme–substrate complex is much faster than the catalytic conversion of substrate into product . Notice that Equation 3 is not badly violated even if k–1 = 0; so the tQSSA is likely to be an excellent approximation for any ratio of enzyme to substrate and for any ratio of timescales . It is of course possible , using the quadratic formula , to solve Equation 1 exactly for C as a function of Ŝ and ET and to substitute this formula into Equation 2 for dŜ/dt . One may instead use the expression which is a good approximation so long as C2 ≪ ETŜ . Defining ρ = ( Ŝ + Km ) /ET , we can write the condition C2 ≪ ETŜ as which is certainly satisfied for ρ ≫ 1 and ρ ≪ 1 . At its minimum , the left-hand side of Inequality 6 is 4 , so the approximate Equation 5 would seem to be quite good for any values of ET and ST . Per Borghans , DeBoer , and Segel , we call Equation 5 the Padé approximant , and we use it whenever possible . Tzafriri [8] provides a careful discussion of the conditions for validity of the tQSSA and the more restrictive Padé approximant ( which he calls the first-order tQSSA ) . Recently , Pedersen et al . applied the tQSSA to the case of an enzyme converting two different substrates into products [9] . Continuing along this line , we apply the tQSSA to an isolated Goldbeter–Koshland ( GK ) switch [10] and then to PINs of increasing complexity .
In 1981 , Goldbeter and Koshland [10] introduced the notion of an ultrasensitive switch , composed of a substrate–product pair ( say , S and Sp ) that are interconverted by two enzymes ( say , E and D ) ; see Figure 1A . ( Think of Sp as a phosphorylated form of S and of E and D as a kinase and a phosphatase , respectively . ) Assuming the MM conditions , ET ≪ S ( 0 ) + Kme and DT ≪ Sp ( 0 ) + Kmd , Goldbeter and Koshland wrote a single dynamical equation for the time evolution of the switch: In Equation 7 , x = Sp ( t ) /ST , ST = S ( 0 ) + Sp ( 0 ) , Je = Kme/ST , Jd = Kmd/ST , Ve = k2eET/ST , Vd = k2dDT/ST , and the subscripts “e” and “d” refer to the kinase and phosphatase reactions , respectively . Goldbeter and Koshland showed that the steady state solution of Equation 7 is given by the GK function , The steady state response ( Sp* ) as a function of stimulus strength ( kinase level , ET ) is simply If Je and Jd ≪ 1 , then Sp* is a steeply sigmoidal function of ET . Goldbeter and Koshland called this signal-response curve zero-order ultrasensitivity . Goldbeter and Koshland's analysis of phosphorylation–dephosphorylation cycles is fine for metabolic control systems , where metabolite concentrations ST are orders of magnitude larger than enzyme concentrations , ET and DT . But for PINs , the condition for the classical MM rate law is not valid , and we must keep track of the enzyme–substrate complexes ( Figure 1B ) . A similar argument has recently been suggested for the mitogen-activated protein kinase pathway [11] . Re-deriving Equation 7 , using the tQSSA and the Padé approximation to the algebraic equations , we find Equation 7 subject to some new definitions: The signal–response curve , is ultrasensitive if Je and Jd ≪ 1 , but this requires that the total enzyme concentrations be small with respect to the total substrate concentration: the standard MM requirement . For PINs , we cannot expect this requirement to be satisfied , which suggests that protein phosphorylation–dephosphorylation cycles are unlikely to be ultrasensitive . In Figure 1C we plot the signal–response curve , Ŝp/ST , as a function of ET/DT given by Equation 8 , for three different values of DT . As expected , the response function is ultrasensitive for Je and Jd small , but ultrasensitivity is lost as Je and Jd increase . This conclusion about ultrasensitivity being lost in PINs , based as it is on the Padé approximant , is not reliable when enzymes and substrates are present in similar concentrations [12] . In fact , numerical calculations ( Figure 1D ) , based on the full tQSSA equations ( Table 1 ) , show that a GK switch may still be ultrasensitive for reasonable values of Je and Jd ( see Figure 1D ) . Therefore , the issue of ultrasensitivity for covalent modifications in PINs should be settled using exact steady state calculations , without making the Padé approximation . In the next section we show that the tQSSA , besides giving insights into the steady state behavior of the network , provides a good approximation of its temporal dynamics as well . In Table 2 ( columns D and E ) we assign specific values to the rate constants and total concentrations for the GK module discussed in the previous section . Our choices are not consistent with MM requirements and are only partly consistent with favorable conditions for the tQSSA . In Figure 2 we compare a numerical solution of the full set of governing differential equations with approximate solutions computed from the usual QSSA and the tQSSA ( the equations are provided in Table 1 and in machine-readable form in File Collection S1 ) . In addition to a time course ( Figure 2C ) , we plot ( Figure 2A and 2B ) one of the fast variables ( [E:S] or [Dp:Sp] ) as a function of the relevant slow variable ( S or Sp for QSSA and Ŝ or Ŝp for tQSSA ) , to highlight the presence of two different timescales . If the timescales are clearly separated , a sudden increase of the complex ( displacement along the y-axis ) will precede the slower conversion of substrate into product ( displacement along the x-axis ) ; see [7] . From the exact solution ( black curves in Figures 2A and 2B ) we see that , at the beginning of the reaction , Sp forms a complex with D , and before D:Sp reaches a maximum , Sp starts to be converted into S . As soon as S is produced , E:S is created and converted back into Sp , with most of E forming a stable complex with S ( [E:S]/ET = 0 . 935 at steady state ) . Although the fast and slow dynamics are not perfectly separated in either QSSA or tQSSA , the timescale separation is clearly more pronounced in tQSSA ( red curve in Figure 2B ) than in QSSA ( blue curve in Figure 2A ) . Time courses ( Figure 2C ) show that throughout the simulation tQSSA does a better job in reproducing the dynamics of the network than QSSA . Notice that at the beginning of the simulation , QSSA gives negative values to S and [E:S] to comply with both initial conditions and conservation relations . In this section we study the steady state behavior of a system of two coupled GK switches ( Figure 3A ) . Building upon the previous network , we consider the possibility that E exists in phosphorylated ( Ep ) and unphosphorylated ( E ) forms . Suppose that Ep is a less active form , and E → Ep is catalyzed by S . S and E are antagonists since they phosphorylate and inactivate each other . F is a phosphatase that converts Ep back to E . ( In our notation for complexes , A:B , the enzyme comes first and substrate follows; for example , for the reaction whereby S phosphorylates E , the enzyme–substrate complex is denoted S:E . ) This is no arbitrary example; it describes exactly the interactions between two regulators of the G2-to-mitosis ( G2/M ) transition in the eukaryotic cell cycle [13] . In that case , S = MPF ( M-phase promoting factor , a dimer of Cdc2 and cyclin B ) and E = Wee1 ( a kinase that phosphorylates and inactivates Cdc2 ) . The parameter values that we choose for these coupled enzymatic reactions ( Table 2 ) are taken , for the most part , from a careful study of the biochemical kinetics of these reactions in Xenopus egg extracts , done by Marlovits et al . [14] . Novak , Tyson , and collaborators implemented these values into a mathematical model for the G2/M transition during early embryonic cell cycles . The equations of the model were derived from an implicit application of the QSSA , but to our knowledge an explicit derivation has never been done: this observation prompted us to investigate the matters addressed in this paper . The antagonism between E and S ( Figure 3A ) constitutes a positive feedback loop ( sometimes called a double-negative feedback loop ) . Novak and Tyson proposed that this loop contributes to the bistability that characterizes the G2/M transition in the eukaryotic cell cycle [15–17] . In fact , according to their original model , the antagonism between Wee1 and MPF alone could sustain bistability . However , this conclusion was drawn from an improper application of the QSSA; hysteresis is lost once the network is unpacked to its elementary steps ( Figure 3B ) . Even more , according to Advanced Deficiency Theory ( performed with the software package CRNT [Chemical Reaction Network Toolbox] developed by Feinberg [18] ) , this network ( Figure 3B ) cannot have bistable behavior for any positive values of the kinetic rate constants . Is it possible to recover hysteresis from the antagonism between S and E ? In Novak and Tyson's original model , both Sp and Ep had some residual activity , a feature that in their model was not essential to generate hysteresis and that we have not taken into account so far . This residual activity becomes essential when the network is reduced to elementary steps: we find that bistability is recovered when we add the reactions E + Sp ↔ Sp:E → Ep + Sp , i . e . , when Sp retains some limited kinase activity . ( In Figure 3B the additional reactions are drawn in grey , and in Figure 3C , top panel , the signal–response curve is drawn with and without the additional reactions . ) Bifurcation analysis suggests that the new reactions create a steady state where S is inhibited and E is active , which is paradoxical because the additional reactions provide an alternative path for inactivating E . The paradox is resolved when we realize that a large fraction of S is sequestered in Sp:E complexes , thus helping E to inactivate S . Indeed , hysteresis is possible ( unpublished data ) with the association–dissociation reactions alone ( E + Sp ↔ Sp:E ) without the catalysis step ( Sp:E → Ep + Sp ) . Of course , bistability can also be restored by allowing Ep to phosphorylate S . To apply the tQSSA to such networks , we first define “hat” variables to include a single free molecular species plus all the complexes in which this species appears . Defined thus , the association–dissociation reactions and all the reactions where a chemical species serve as a catalyst cancel out . Here we define whose rates of change are given by Our definition of hat variables extends what was originally proposed for a single enzymatic reaction by Borghans , DeBoer , and Segel . In terms of the hat variables , we can describe each catalytic reaction in the network with equations similar to Equations 1 and 2 . For example , the rate of phosphorylation of S in Figure 3B is described by Equation 2 with k2C → e2 [E:S] and with the concentration of enzyme–substrate complex given by Equation 1 with C = [E:S] , , and ( see equations in Table 3 ) . Concerning this network: we perform no numerical simulations to compare tQSSA and QSSA; we will do that in the next section for a larger and biologically more significant network . Rather , we use the model reduced with tQSSA to perform phase plane analysis ( Figure 3C , bottom panel ) , which confirms that the additional reaction strengthens the negative effect that E exerts on S , thereby bending the Ŝ nullcline and creating three intersection points . In the G2/M network , the antagonism between Wee1 and MPF is aided by a second positive feedback loop , involving MPF and Cdc25 , a phosphatase that removes the inactivating phosphate group from Cdc2 [13] . ( In our notation , Wee1 is E , MPF is S , and Cdc25 is D . ) Because S can phosphorylate D , and Dp is a more active form , S and Dp activate each other . Finally , we have C , an unregulated phosphatase that converts Dp back to D . Altogether , the network consists of three GK modules: the first ( C/D/S ) controls D's phosphorylation , the second ( D/S/E ) affects the phosphorylation state of S , and the last ( S/E/F ) controls the activity of E ( Figure 4A and 4B ) . Again , we take most of the parameter values for the additional reactions from Marlovits et al . [14] . In Table 4 we provide the governing equations for these coupled GK switches in three versions: full ( no approximations ) , QSSA , and tQSSA . Bifurcation analysis and CRNT show that the network is bistable even with no residual activity for Sp or Ep ( unpublished data ) . Indeed , the positive feedback between Dp and S suffices , by itself , to generate hysteresis , as confirmed by CRNT . Interestingly , given the parameter values in Table 2 , the network is bistable ( unpublished data ) in a region very similar to that determined experimentally by Sha et al . [17] . Since the model performs satisfactorily concerning its steady state behavior , we move on to compare the dynamics of both QSSA and tQSSA to the exact solution ( Figure 4C ) . To apply tQSSA , we define the hat variables whose dynamics are described by the following equations: The concentrations of the enzyme–substrate intermediates are given by solving simultaneously a set of six coupled quadratic algebraic equations ( see Table 4 ) . Notice how the modularity of the network ( composed of six identical MM reactions ) is mirrored by the modularity of the tQSSA equations . Numerical simulations ( Figure 4C ) demonstrate the superiority of tQSSA ( red lines ) over QSSA ( blue lines ) in accurately capturing the exact dynamics ( black lines ) of this regulatory network . To compare further the QSSA and tQSSA with the exact solution , we plot in Figure 5 the complexes ( [Dp:Sp] and [E:S] ) as functions of the slow variables ( Sp and S for QSSA and Ŝp and Ŝ for tQSSA ) . The initial hump of [Dp:Sp]—due to phosphorylation of D by S followed by dephosphorylation of Dp by C— ( black curves in Figure 5A and 5B ) , is captured qualitatively by the tQSSA ( red curve ) but completely missed by the QSSA ( blue curve ) . Similarly , the initial accumulation of [E:S] is closely approximated by the tQSSA but badly overshot by the QSSA . Both approximations capture the latter part of the time course reasonably well . The QSSA completely misses the initial stages of the simulation because it assigns negative values to Ep , [F:Ep] , and D .
Coupled enzymatic reactions ( e . g . , interacting kinases and phosphatases ) are common features of PINs . We analyze one of these networks , composed of three phosphorylation and three dephosphorylation reactions , altogether comprising 14 chemical species . The network was originally proposed by Novak and Tyson [13] to model the activation of MPF in early embryos of the frog Xenopus . In their original work , Novak and Tyson neglected the contribution of enzyme–substrate complexes . Using a combination of mass-action and MM kinetics , they showed that the network may have multiple steady state solutions . We have relaxed the assumptions , writing down differential equations for each species , including the enzyme–substrate complexes . Using parameters obtained indirectly from published data , we confirmed that the model exhibits bistability in the same parameter range as predicted theoretically by Novak and Tyson [13] and measured experimentally [16 , 17] . As for the dynamics , our simulations show that the enzyme–substrate complexes , neglected in the original model , are present in non-negligible concentrations . The complexes play an important role because of the topology of the network . In the cell cycle model , some molecular species are at the same time enzymes and substrates of each other . If for one reaction enzyme concentration is negligible compared with substrate concentration , then the opposite must be true when the roles are exchanged , allowing for a significant fraction of the substrate to be sequestered in the complex . Similarly , when some molecules form complexes with several enzymes , even if each complex is not present in a large amount , their sum may not be negligible . The role played by enzyme–substrate complexes in PINs could be more important than currently appreciated . In the cell cycle network , Wee1 and MPF are antagonists that phosphorylate and inhibit each other . In our simulations , the concentrations of Wee1:MPF and MPF:Wee1 are not negligible . The presence of high concentrations of such complexes can be interpreted as a second way for Wee1 to inhibit MPF , by sequestration . In this sense , Wee1 behaves as both an inhibitory kinase and a stoichiometric cyclin dependent kinase ( CDK ) inhibitor . In our simulation we notice that MPF:Cdc25 is also present in high concentration ( 40% of total MPF , 10% of total Cdc25 ) . If confirmed , that would suggest an intrinsic way to attenuate the positive feedback loop between MPF and Cdc25 . Finally , trimeric complexes might form as well ( e . g . , Wee1:MPF:Wee1 ) , and such molecular species may have great effects on the qualitative behavior of PINs ( Sabouri-Ghomi , Ciliberto , Novak , and Tyson , unpublished data ) . Of course , one can follow the exact dynamics of a control system by solving the full system of ordinary differential equations ( ODEs ) . However , a full description of the system comes with many equations , making any qualitative analysis difficult . A typical way to reduce the number of equations is the QSSA originally formulated for an isolated enzyme-catalyzed reaction , when substrate is in great excess over enzyme . When applying the QSSA to a network of coupled catalytic reactions , with realistic values of rate constants and total protein concentrations , we found that the reduced system of ODEs obtained by the QSSA does not faithfully reproduce the dynamics of the full system of ODEs . The tQSSA works for a larger range of parameter values than does the QSSA; in particular , it is valid even when the enzyme is in excess compared with the substrate . Such an approximation is particularly appealing in PINs where , as mentioned above , enzymes and substrates often exchange roles . Given its appeal , we applied tQSSA to a full model of the PIN that regulates the G2/M transition in the cell cycle . We found by numerical simulations that the tQSSA does a good job describing coupled catalytic reactions . Moreover , applying tQSSA to PINs generates a set of differential–algebraic equations of standard format , Equations 1 and 2 , suggesting that a computer algorithm may be devised whereby tQSSA is used to reduce a full set of multi-timescale ODEs to a smaller set of slow equations . The reduced set of equations may be especially useful for stochastic simulations of PINs by Gillespie's algorithm . Summarizing , we propose that large networks of coupled enzymatic reactions should first be written in full and then reduced by applying the tQSSA . This way it will be possible to reduce the number of dynamic equations while maintaining the complexity of the network ( i . e . , including enzyme–substrate complexes ) and simultaneously to achieve reliable approximate solutions for the transient dynamics of the network .
All calculations have been made using XPPAUT , software developed by Ermentrout [19] and freely available online at http://www . math . pitt . edu/∼bard/xpp/xpp . html . XPPAUT solves the algebraic part of differential–algebraic equations by Newton's method , given an initial guess for the unknowns . In File Collection S1 we provide . ode files , readable by XPPAUT , for reproducing the results in this paper . | The physiological responses of a cell to its environment are controlled by gene–protein interaction networks of great complexity . To understand how information is processed in these networks requires accurate mathematical models of the dynamical behavior of large sets of coupled chemical reactions . To avoid producing large and hardly manageable models , such reaction networks are often simplified using phenomenological reaction rate laws , such as the Michaelis–Menten rate law for an enzyme-catalyzed reaction . We show that , in regulatory networks where proteins swap places as enzymes and substrates , such simplifications must be carried out with care , keeping track of enzyme–substrate complexes . The risk is to provide a simplified description of the molecular networks that at best is correct for the long-term behavior but fails to represent the short-term dynamics of the real network . To avoid such a possibility , we suggest using an alternative approach called the total quasi-steady state approximation . We apply this alternative formalism to a model of the network controlling the entry into mitosis in the eukaryotic cell cycle , composed of three coupled protein modification cycles . Whereas the classical Michaelis–Menten formalism fails to represent the dynamics of this network correctly , the one we propose captures the behavior with economy and accuracy . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"cell",
"biology",
"eukaryotes",
"computational",
"biology"
] | 2007 | Modeling Networks of Coupled Enzymatic Reactions Using the Total Quasi-Steady State Approximation |
Dengue virus ( DENV ) activity has been reported in Dhaka , Bangladesh since the early 1960s with the greatest burden of dengue fever and dengue hemorrhagic fever cases observed in 2000 . Since this time , the intensity of dengue activity has varied from year to year , and its determining factors remained relatively unknown . In light of such gaps in knowledge , the main objectives of this study were to determine the magnitude of seroprevalence and seroconversion among the surveyed population , and establish the individual/household level risk factors for the presence of DENV antibodies among all age groups of target populations in the city of Dhaka . Considering the lack of fine scale investigations on the factors driving dengue activity in Bangladesh , a prospective cohort study involving serological surveys was undertaken with participant interviews and blood donation across the city of Dhaka in 2012 . Study participants were recruited from 12 of 90 wards and blood samples were collected during both the pre-monsoon ( n = 1125 ) and post-monsoon ( n = 600 ) seasons of 2012 . The findings revealed that the seroprevalence in all pre-monsoon samples was 80 . 0% ( 900/1125 ) while the seropositivity in the pre-monsoon samples that had paired post-monsoon samples was 83 . 3% ( 503/600 ) . Of the 97 paired samples that were negative at the pre-monsoon time point , 56 were positive at the post-monsoon time point . This resulted in a seroprevalence of 93 . 2% ( 559/600 ) among individuals tested during the post-monsoon period . Seroprevalence trended higher with age with children exhibiting a lower seropositivity as compared to adults . Results from this study also indicated that DENV strains were the only flaviviruses circulating in Dhaka in 2012 . A multivariate analysis revealed that age , possession of indoor potted plants , and types of mosquito control measures were significant factors associated with DENV seroprevalence; while attendance in public/mass gatherings , and use of mosquito control measures were significantly associated with DENV seroconversion after adjusting for all other variables . Our study suggests that there is a high level of endemic dengue virus circulation in the city of Dhaka which has resulted in significant DENV seroprevalence among its residents . Seropositivity increased with age , however , a substantial proportion of children are at risk for DENV infections . Our serological analysis also documents considerable DENV seroconversion among study participants which indicates that a large proportion of the population in the city of Dhaka were newly exposed to DENV during the study period ( pre-and post-monsoon 2012 ) . High levels of seroconversion suggest that there was an intense circulation of DENV in 2012 and this may have resulted in a significant risk for viral associated illness . Findings of our study further indicated that home-based interventions , such as removing indoor potted plants and increased bed net use , in addition to vector control measures in public parks , would reduce exposure to DENV and further decrease risk of viral associated disease .
The spread of dengue virus ( DENV ) –a viral pathogen transmitted by mosquitoes , primarily Aedes aegypti and Ae . albopictus–has been unprecedented in recent decades . At the present time , 390 million people are exposed to DENV each year resulting in 96 million annual cases of viral associated disease globally [1] while approximately 3 billion people living in the tropical and subtropical regions are at risk of infection [2–6] . According to estimates from the World Health Organization ( WHO ) , approximately 500 , 000 people develop severe disease each year , and among them , about 1 , 250 ( 2 . 5% ) die [7] . The reasons for increased incidence and rapid geographic spread of dengue are not fully understood; however , numerous ecological , biological , social , economic , and cultural factors could be responsible . Due to the fact that the most advanced dengue vaccine candidate has hitherto exhibited only intermediate efficacy [8] despite considerable investment and research , the current dengue prevention and epidemic response primarily relies upon vector control measures designed to reduce vector abundance . Effective vector control requires a clear understanding of the drivers and/or determining factors which drive DENV transmission . It is therefore important to determine the epidemiological , socioeconomic and ecological factors and the dynamics of exposure to DENV infection , risk and epidemic potential . Among numerous factors , globalization of the world economy , increases in international travel , breakdowns in public health infrastructure , ineffective vector control programs and enhanced climate variability have been cited for their potential roles in resurgence of DENV [1 , 9–10] . The underlying drivers of DENV activity operate at various spatial scales , and among factors at the regional/local scale , uncontrolled urbanization [11] , human population density and clustering of households with DENV activity seem to play a prominent role . At the finer scale of neighborhoods and households , the absence of public utility services including reliable sources of potable water , effective sewer systems , and appropriate waste disposal appears to contribute significantly to the problem [12–15] . The complexity of the factors influencing the dynamics of patterns of DENV transmission are reflected in the heterogeneity of dengue incidence in time and space and this heterogeneity needs to be considered when attempting to understand or predict patterns of disease incidence . Although the first recognized outbreak of dengue in the city of Dhaka , the capital of Bangladesh , was recorded in 1964 [16] , followed by sporadic cases of dengue fever ( DF ) during 1977–78 and 1996–1997 [17] , the extent of dengue prevalence in Bangladesh is poorly documented . It is also possible that sporadic or ongoing transmission occurred during the 1964 to 1978 and 1978 to 1997 time frames , but went undetected , due to under-reporting . The first identified epidemic of DF and dengue hemorrhagic fever ( DHF ) in Bangladesh , took place during the monsoon season of 2000 , and resulted in 5 , 521 officially reported cases with 93 fatalities [17–18] . Based on molecular diagnostic testing of persons with acute disease , all four dengue serotypes have been found in circulation in recent years [19–21] . From 2000–2009 , 91 . 0% of all reported dengue cases were from Dhaka making it the most endemic urban area of the country [22] . Cases of DF were clustered during the time periods: 2000–2002 , 2003–2005 , and 2006–2009 , and Dhaka was the locality with ‘the most likely cluster for DF in all three periods’ [18] . Since 2009 , the number of reported dengue cases declined in Bangladesh [23]; however , as discussed by Sharmin et al . [22] , this was likely an artifact of changes in reporting criteria ( i . e . , laboratory confirmation was required in order to confirm cases instead of using a clinical diagnosis ) . As a result , the ‘tip of the iceberg’ analogy of dengue reporting is likely in play in Bangladesh , as it has been well-documented that passive surveillance involving case notifications does not accurately reflect the burden of dengue in such a geographical location [24] . The local drivers of dengue activity in Bangladesh are poorly defined . For example , Mahmood and Mahmood [25] suggested that several macro-level risk factors including overpopulation , uncontrolled urbanization , and poor waste management play prominent roles in the emergence of dengue in Bangladesh . Furthermore , piped water supply , drainage and adequate sewage disposal are unevenly distributed throughout the city , all of which can have an effect on dengue transmission [26] . At the present time , detailed studies on the factors operating at the household level that impact dengue transmission dynamics are generally lacking in Bangladesh . Given the absence of human surveillance system for dengue and a lack of fine scale investigations on the factors driving dengue activity in this country , a prospective cohort study involving serological surveys was undertaken with participant interviews and serological sampling across the city of Dhaka in 2012 . Because dengue incidence in Bangladesh is highly monsoon dependent [10] , two separate household serosurveys were conducted—one pre-monsoon ( to estimate seroprevalence ) and one post-monsoon ( to determine the extent of seroconversion ) . The target population included all household members including children less than 12 years of age . Potential risk factors for dengue exposure were solicited from data collected during structured participant interviews . To the best of our knowledge , this study is the first population serosurvey on dengue in Bangladesh , with the objective i ) to determine the magnitude of dengue seroprevalence and seroconversion , and ii ) simultaneously to determine the individual/household level risk factors among all age groups for dengue transmission in various socioeconomic status zones of the city of Dhaka . It was anticipated that the outcomes of this investigation would inform epidemic risk management authorities and assist in the development of effective intervention tools and strategies .
The city of Dhaka is ranked as the 9th largest city in the world [27]; it is also among the most densely populated cities . Considering that the city of Dhaka has experienced the most intense dengue activity in Bangladesh in recent years [18] and it has been cited as the most endemic area of the country [22] , Dhaka City Corporation [ ( DCC ) location: 23 . 77° N; longitude , 90 . 38° E ) ] was chosen as the study area for the present investigation . Dhaka experiences a hot , wet and humid tropical climate with a distinct monsoonal season . The annual average temperature is 28°C and annual average rainfall of 2002 mm [28] . The pre-monsoon season is very hot with an average maximum of 36 . 7°C , a very high rate of evaporation and erratic but occasional heavy rainfall from March to June . The post-monsoon season is short-lived and characterized by decreased rainfall and gradual lowering of night time minimum temperatures . More than one-third of the population of this city is poor and lives in dense squatter settlements , and their livelihood usually relies on daily wage earnings [29] . The city has the natural and social conditions which provide ample larval development sites for mosquito populations , particularly during the monsoon , rainy seasons . Aedes aegypti , as the primary vector , occupies the overwhelming majority of the wards ( administrative units ) in Dhaka [30] , while the secondary vector–Aedes albopictus–occupies only a few wards in and around the fringe areas [31–32] . The primary sampling units of this study were individual households ( HHs ) . A multi-stage , stratified sampling design was used for this investigation . The 90 administrative wards ( according to the previous administrative system ) of the city were categorized by socio-economic status ( SES ) ( i . e . high , medium , and low ) . The criteria for classifying the wards , following a Delphi method , into three SESs are described elsewhere [30] . A probability proportional sampling ( PPS ) method [33] was employed to select sample wards from each stratum , which resulted in 2 wards from high , 5 wards from medium , and 5 wards from low SES . Finally , a spatial randomization procedure was followed where 100m X 100m grid cells were superimposed on each ward , and a total of 100 grid cells were randomly selected . From each ward , 100 households were selected as sampling units for the surveys , resulting in a total of 1200 target households ( i . e . , 200 from high , 500 from medium , and 500 from low SES wards ) . Initially , the project was designed to examine i ) the seroprevalence of DENV antibodies in the target populations and ii ) the prevalence , abundance and distribution of DENV vector mosquitoes . As a result , no specific clinical information was used for recruitment . For the purpose of this study , seroprevalence was defined as the percentage of total individuals in a population who tested positive for the presence of IgG antibodies to DENV . Individuals identified as seroconverted were found to be negative for dengue antibody during the pre-monsoon period but who became positive for dengue IgM or IgG based on the serological analysis of their post–monsoon sample . The extent of seroconversion refers to the percentage of new cases as well as new exposures identified between the pre- and post-monsoon period . Two household serosurveys were conducted in 2012 in order to assess the seroprevalence and seroincidence of DENV . The first serosurvey was completed during the pre-monsoon season of 2012 ( June and July ) . At this time , information on demographic , socio-economic and other relevant characteristics was collected from the study participants using a structured socioeconomic questionnaire . The inclusion criteria for household members in our study were: all household members shared one roof , all meals , and common living space . Participants were a priori required to give informed consent . Special attention was given to age , sex , place of residence , socioeconomic status , water storage and supply , and migration and intra-city mobility variables . The 2012 pre-monsoon serosurvey served as a baseline for the determination of DENV seroprevalence and involved the testing of 1125 individuals . The follow-up serosurvey was carried out during the 2012 post- monsoon ( November ) and households that participated in the baseline serosurvey were revisited . At the time of the follow-up study , 600 of the 1125 ( 53 . 2% ) individuals still resided in the previously recorded locations and were willing to provide another blood sample . Thus , these 600 study participants ( age range: 1–77 years of age ) from 390 households who provided paired serum samples comprised the sample set for determining the extent of seroconversion between the 2012 pre- and post-monsoon time frames . Participation in the serosurvey was on a voluntary basis , and in case of children ( 1–11 years of age ) consent of both individuals and parents ( whenever possible ) was obtained . Blood samples were collected by personnel from the International Centre for Diarrhoeal Disease Research , Bangladesh ( ICDDR , B ) in Dhaka . The ethical guidelines of the Bangladesh Medical Research Council were followed for the collection of 5-mL of intravenous blood from each participant . The serum samples were separated by centrifugation and stored at -20°C in ICDDR , B laboratories; a unique identification number was assigned to each participant to maintain anonymity . All serum samples were sent to the National Microbiology Laboratory ( NML ) in Winnipeg , Canada where they were stored at -80° C until tested . All sera collected during the 2012 pre-monsoon ( n = 1125 ) and post-monsoon ( n = 600 ) were tested for antibodies to DENV using commercial IgM and IgG ELISA kits . Because flavivirus cross reactivity is a concern in ELISAs , and other flaviviruses have been sporadically reported in Bangladesh [77 , 78] , a subset of 100 IgG positive sera from 2012 pre-monsoon and 50 IgG positive sera from post-monsoon were randomly selected and tested for Japanese encephalitis virus ( JEV ) , West Nile Virus ( WNV ) and DENV-neutralizing antibodies by plaque reduction neutralization tests ( PRNT ) as previously described [34] . Serum samples collected during the pre- and post-monsoon were tested for dengue IgG and IgM antibodies by ELISA . Sera were tested at a 1:100 dilution using the DENV IgG Dx Select and DENV IgM Capture Dx Select kits ( Focus Diagnostics Inc . , Cypress , CA , USA ) according to the manufacturer’s instructions . In order to mitigate potential false positive reactions due to heterophilic IgM antibody activity , samples that were initially positive in the IgM ELISA were re-tested using the background subtraction procedure as detailed in the kit instructions . A total of 150 IgG-seropositive samples from across age groups ( 135 adults and 15 children under 12 years of age ) were randomly selected and tested for DENV-neutralizing antibodies using a PRNT assay [34] . A known amount of DEN 2 virus [i . e . , 100 plaque forming units ( PFU ) ] was mixed with various dilutions of test sera and incubated for 90 minutes at 37°C . Following this , 100 μl of the virus-serum mixture was added to fully confluent Vero cell monolayers ( in 6 well tissue culture plates ) and incubated for one hour at 37°C with gentle rocking every fifteen minutes . Following incubation , 3 mls of an agar overlay was applied to each well and the plates were incubated at 37°C , 5% CO2 for 72–120 hours depending on the virus . A second agar overlay containing the vital dye neutral red was added and the presence/absence of plaques was noted after an additional 18 hours of incubation . By determining the final dilution of serum that lead to a 90% reduction in plaque formation ( PRNT90 ) , an end-point titre for virus-specific neutralization activity was calculated . For screening purposes samples with PRNT90 titers ≥ 20 were considered to have neutralizing antibody against DENV or other flaviviruses . End point titres of 4-fold or greater difference between dengue and WNV or JEV were used to identify the specific flavivirus antibodies in the selected serum samples . Among the 1 , 200 selected households for baseline seroprevalence as well as the follow-up survey , data obtained from the participants were analyzed . This study initially targeted one participant from each household; however , multiple serum samples were sometimes collected from each household when more than one household member was willing to participate . Two outcome variables ( i . e . , seroprevalence and seroconversion/ recent DENV infection ) for dengue infection were defined based on serological testing . Univariate and multivariate analyses of potential socioeconomic and demographic risk factors for DENV infection were performed using SPSS ( version 22 . 0; SPSS Inc . , Chicago , IL ) . Age-and gender-specific antibody prevalence , with 95% confidence intervals ( CI ) were calculated . The χ 2 test was used to evaluate the strength of the association between outcome variables and the risk factors , as described in Dhar-Chowdhury et al . [30] , and it was measured by prevalence odds ratios ( OR ) , with 95% CIs . A p-value of <0 . 05 was considered statistically significant . To control for potential confounders and to assess the effect of each risk factor a logistic regression was constructed , and the corresponding ORs with 95% CIs were also estimated . Variables that were related to DENV IgG seroprevalence and seroconversion in the bivariate analysis ( p-value<0 . 25 ) were included in the logistic regression models . A Receiver Operating Characteristic ( ROC ) curve was drawn to assess the goodness-of-fit of our logistic regression model . All analyses were performed using statistical software SAS 9 . 1 . 3 ( SAS Institute . Cary , NC , USA ) . The research was approved by the Bangladesh Medical Research Council ( Bangladesh ) and the Joint Faculty Research Ethics Board of the University of Manitoba ( Canada ) . The purpose and objectives were explained to the head of each household and her/his informed consent was sought to collect demographic , socioeconomic , household infrastructure and ecological information [30] . In Bangladesh , many respondents are illiterate and are unwilling to provide written consent due to fear of forgery or deception . In this regard , both Bangladeshi and Canadian ethics Boards , considering the socio-cultural contexts , approved oral consent procurement , with witnesses and their signatures . Written ( with signature ) or verbal consent ( recorded in tape recorders with consent ) was also obtained from each household member who donated blood samples; in case of children , consent from parent ( s ) was obtained .
A total of 1125 serum samples [1003 ( 89 . 2% ) from adults and 122 ( 10 . 8% ) from children under 12] were collected during the 2012 pre-monsoon from residents of 12 wards of the Dhaka City Corporation from a total of 635 households . The descriptive analysis was performed for each variable of our study population ( Table 1 ) . The age of survey participants ranged from 1 to 77 years with a mean age of 31 . 9 years ( 95% CI: 29 . 0–31 . 0 ) . The male to female ratio of the subjects was 1:1 . 3 and more than half of the participants were housewives ( 34% ) or students ( 19 . 6% ) . During pre-monsoon survey , the overall DENV IgG seroprevalence was 80% ( 900 / 1125 ) . As shown in Fig 1 , among adults , 83 . 1% ( 833 of 1003 ) were positive for IgG while 54 . 9% ( 67 of 122 ) of children had DENV IgG antibodies . Overall , 22 of 1003 ( 2 . 2% ) adults and 1 of 122 ( 0 . 8% ) children had IgM antibodies which suggests that recent infections with DENV were occurring . In general , adults who were 34–44 years of age and older had a high ( 92% ) seroprevalence of DENV antibodies and seroprevalence generally increased with age ( Fig 1; S1 Table ) . During the post-monsoon survey , paired sera were collected from a subset of 600 individuals ( 53 . 3% of the original 1125 participants ) residing in 390 households . The seroprevalence for these 600 individuals increased from 83 . 8% ( 503 / 600 , pre-monsoon ) to 93% ( 559/600 ) during the monsoon period indicating a significantly higher magnitude of DENV circulation and seroconversion during this time frame ( Table 2 ) . Consistent with this pre- to post-monsoon increase in seroprevalence , 57 . 7% ( 56 of 97 ) of individuals ( 600 participants with paired sera ) who were seronegative during the pre-monsoon survey had seroconverted to either IgG ( n = 47 of 56 ) or IgM antibodies ( n = 4 of 56 ) or both immunoglobulins ( n = 5 of 56 ) . Overall , there was a higher number of seroconversions among female ( 60 . 7% CI: 0 . 50–0 . 71 ) participants than males ( 39 . 3% CI: 0 . 29–0 . 50 ) and among adults 75% ( 42 /56 ) than children 25% ( 14/56 ) . For adults , 59% ( 42/71 ) seroconverted between the pre- and post-monsoon periods . Among children ( <12 years of age ) , the extent of seroconversion was somewhat lower at 53 . 8% ( 14/26 ) . For children 6 years of age or more , seroconversion was determined to be 56 . 2% ( 9/16 ) ; while 50% ( 5/10 ) of children less than 6 years old seroconverted . Fig 2 illustrates the geographical distribution of seroprevalence by ward , and the proportion of seroprevalence varied among the wards from moderate/moderate high ( 59 . 8–79 . 8% ) to high/very high ( 79 . 8–99 . 9% ) . The very high degree of seroprevalence was observed in the older settlements , located in the southern and central zones of the city whereas moderate seroprevalence levels prevailed in the fringe areas . There was a strong positive correlation between the results obtained with the IgG ELISAs and the PRNT assay . Of the 150 IgG ELISA-positive samples , 96% and 94% of the pre- and post-monsoon samples , respectively , had neutralizing titers ( ≥ 1:20 ) to DENV-2 . Although a number of samples generated positive titres ( 1:20 dilution ) for all three flaviviruses , endpoint titration performed on these samples were negative for JEV and WNV when the 4-fold difference in virus-specific titres was used to differentiate the identity of the infecting virus . At the lowest screening dilution factor ( i . e . , 1:20 ) cross-reactivity in PRNTs can be observed in flavivirus infections because antibodies elicited against conserved epitopes on the immunogenic envelope protein may cross-react with other flaviviral antigens . Low neutralizing titres to other flaviviruses may therefore occur , leading to the false-positive results on the screening assays . The PRNT results confirmed the absence of JEV and WNV antibodies . Hence , the seroprevalence identified in this study can be regarded as resulting from ‘confirmed dengue infections’ . The univariate analyses [dependent variable was serological status ( i . e . , either positive or negative ) ] confirmed significant association between IgG seroprevalence and all different age groups ( CI: 2 . 70–5 . 90; p-value <0 . 0001 ) ( Table 3 ) . From Fig 3 ( S2 Table ) , we find that among children ( <12 years of age ) , the IgG prevalence increased with age; by the age of 9 years , 77% ( 94 of 122 ) children were IgG seropositive . There was no significant difference in seroprevalence by sex ( p-value = 0 . 372 ) . In addition to age , number of indoor potted plants ( CI: 0 . 34–0 . 77; p-value = 0 . 0011 ) , types of mosquito control measures ( CI: 0 . 93–2 . 93; p-value = 0 . 016 ) , and febrile illness in any family members during the past six months ( CI: 1 . 00–1 . 81; p-value = 0 . 0441 ) showed statistically significant association with seroprevalence ( Table 3 ) . Results of the univariate analyses depicted in Table 4 revealed that seroconversion ( i . e . , seroconversion/ person/ season ) was significantly associated with attendance in mass/public gatherings ( p value = 0 . 032 ) and type of mosquito control measures ( p value = 0 . 029 ) . There was no statistically significant difference in the seroconversion by sex , age , occupation , education , income , number of water vessels indoors , number of indoor potted plants , vegetation/trees nearby and family members suffered from febrile illness . In order to determine whether socioeconomic ( i . e . , income , education and occupation ) , demographic ( age , sex ) , behavioral ( i . e . , attendance in public gatherings , types of mosquito control measures used ) , utensils ( i . e . , water vessels outdoors , number of potted plants ) , and epidemiological ( i . e . , febrile illness among any family members in past six months ) factors were associated with seroprevalence , after adjusting the explanatory variables , a multivariable logistical regression model with stepwise selection was applied . The model outputs shown in Table 5 revealed that age ( χ2 = 18 . 4 , df = 2 , p-value <0 . 0001 ) , number of indoor potted plants ( χ2 = 8 . 3 , df = 1 , p-value = 0 . 004 ) , and type of mosquioto control measures used ( χ2 = 6 . 1 , df = 2 , p-value = 0 . 048 ) were significantly associated with seroprevalence after adjusting all other variables . For the age group 12–44 , the odds ratio is 4 . 13 , implying that the persons in this age group were 4 . 13 times more likely to be exposed ( i . e . , via multiple exposures ) to DENV compared to those who were in the group less than 12 years of age . Similarly , persons in the age group above 45 were more likely ( OR = 5 . 9 ) to be exposed to DENV relative to children ( less than 12 years ) ( Table 5 ) . Persons in household without any indoor potted plants were less likely to be exposed to DENV than those who owned some indoor potted plants ( OR = 0 . 53 ) , suggesting that seroprevalence is positively associated with the presence of indoor potted plants ( Fig 4 ) . Mosquito control measures at the personal level were found to be significantly associated with seroprevalence ( p = 0 . 048 ) . While univariate analysis indicated that persons using bed nets were less likely to be exposed to dengue virus than those using sprays or coils , the degrees of difference for these two comparators versus bed nets were deemed to be nominal in multivariate analysis . A multivariable logistical regression model was used to determine whether the socioeconomic , demographic , behavioral , utensils , ecological ( i . e . , vegetation or trees nearby ) and epidemiological factors ( as defined above ) were associated with seroconversion . Results from the stepwise logistic regression method , shown in Table 6 , revealed that age ( χ2 = 11 . 1 , df = 1 , p-value = 0 . 001 ) , attendance in public gatherings ( χ2 = 17 . 2 , df = 1 , p-value <0 . 0001 ) , and types of mosquito control measures used ( χ2 = 32 . 5 , df = 2 , p-value <0 . 0001 ) were found to be significantly associated with seroconversion after adjusting all other variables . As shown in Fig 5 , the area under the curve ( Receiver Operating Characteristic—ROC ) is significantly different from 0 . 5 since p-value is <0 . 00001 implying that the logistic regression model used to explain the association between seroprevalence and the explanatory variables classifies the ‘status’ variable significantly better than by chance . Persons attending schools or religious institutions were less likely ( OR = 0 . 19 ) to be infected with DENV than individuals who spent time in parks or gardens ( Table 6 ) . The type of mosquito control measures undertaken by household members also significantly affected the patterns in seroconversion; for example , urban residents using mosquito coils were 14 . 6 times more likely to be recently infected with DENV than bed net users [p-value = 0 . 0026 ( Table 6 ) ] . All other explanatory variables were found not to be significantly associated with dengue seroconversion .
This study was subject to several limitations . Firstly , the participation rate was low during the follow-up study . Due to a very high rate of change of residence and relocation spatially among Dhaka city dwellers , we collected only 600 paired sera samples during the post-monsoon survey , causing us to limit the sample size to only 50% of the initial target . A larger sample size in future studies will provide a higher degree of confidence when interpreting serological outcomes . The possibility of information bias in our study overall is nominal . This is because all potential Aedes breeding sites in household premises were inspected with their complete ( total ) counts; for example , field entomologists counted all potted plants , uncovered water tanks , number of trees nearby and thus , there was no possibility of information bias . However , there still could be some degree of information bias in our analysis due to the sampling error in the population estimates . As well , we were unable to determine the infecting DENV serotype responsible for the observed dengue seroconversion and future studies are warranted to identify dengue serotypes and strains currently circulating in the city . Lastly , because this is the first population-based DENV seroprevalence study in Bangladesh , we are unable to compare our findings with other seroprevalence data in the country and we may have underestimated the seroincidence rate , because we relied on collecting specimens from our baseline study population only . | Similar to many other tropical regions of the world , dengue is a major public health problem in Bangladesh where Aedes aegypti mosquitoes are the main vector . Through this serological survey , we present data on the magnitude ( measured in proportions ) of seroprevalence and seroconversion within 12 selected wards in the city of Dhaka , Bangladesh . In 2012 the observed dengue seroprevalence was 93% among individuals tested during post-monsoon with a seroconversion rate between pre- and post-monsoon periods of 57 . 7% . This finding suggests that dengue virus ( DENV ) circulated in the city during the observation period . Past exposure to dengue virus was highly associated with age , possession of indoor potted plants , types of mosquito control measures used , and human movement and attendance in mass gatherings , while the extent of dengue seroprevalence were not associated with the socioeconomic status of the study participants . Our findings suggest that household utilities and water management or storage practices and recognition and elimination of mosquito development sites and participation in mass gatherings are important factors that affect exposure to dengue . Intervention strategies should therefore target these factors for effective prevention and control of dengue infection . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"dengue",
"virus",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"geographical",
"locations",
"tropical",
"diseases",
"microbiology",
"viruses",
"a... | 2017 | Dengue seroprevalence, seroconversion and risk factors in Dhaka, Bangladesh |
Both pulmonary tuberculosis ( PTB ) and intestinal helminth infection ( IHI ) affect millions of individuals every year in China . However , the national-scale estimation of prevalence predictors and prevalence maps for these diseases , as well as co-endemic relative risk ( RR ) maps of both diseases’ prevalence are not well developed . There are co-endemic , high prevalence areas of both diseases , whose delimitation is essential for devising effective control strategies . Bayesian geostatistical logistic regression models including socio-economic , climatic , geographical and environmental predictors were fitted separately for active PTB and IHI based on data from the national surveys for PTB and major human parasitic diseases that were completed in 2010 and 2004 , respectively . Prevalence maps and co-endemic RR maps were constructed for both diseases by means of Bayesian Kriging model and Bayesian shared component model capable of appraising the fraction of variance of spatial RRs shared by both diseases , and those specific for each one , under an assumption that there are unobserved covariates common to both diseases . Our results indicate that gross domestic product ( GDP ) per capita had a negative association , while rural regions , the arid and polar zones and elevation had positive association with active PTB prevalence; for the IHI prevalence , GDP per capita and distance to water bodies had a negative association , the equatorial and warm zones and the normalized difference vegetation index had a positive association . Moderate to high prevalence of active PTB and low prevalence of IHI were predicted in western regions , low to moderate prevalence of active PTB and low prevalence of IHI were predicted in north-central regions and the southeast coastal regions , and moderate to high prevalence of active PTB and high prevalence of IHI were predicted in the south-western regions . Thus , co-endemic areas of active PTB and IHI were located in the south-western regions of China , which might be determined by socio-economic factors , such as GDP per capita .
Pulmonary tuberculosis ( PTB ) and intestinal helminth infection ( IHI ) are still widespread in China . Both diseases are associated with poverty and both seriously impact on people’s health . The latest national epidemiological survey for PTB was conducted in 2010 and showed that the active PTB prevalence was 459 per 100 , 000 among those above 15 years old[1] . The latest national survey for major human parasitic diseases , conducted 2001–2004 reported a total of 26 species of helminth with an overall rate of 21 . 7% . The most common helminth infections found were Ascaris lumbricoides ( 12 . 5% ) , hookworm ( 5 . 7% ) , Trichuris trichiura ( 4 . 2% ) , Clonorchis sinensis ( 0 . 6% ) and Taenia spp . ( 0 . 1% ) [2] . A syndemic , i . e . an aggregation of two or more diseases in a population , in which there is some level of biological interaction could be at work with respect to PTB and IHI[3] . For instance , there are indications that IHI may be one of the risk factors for the development of active PTB in addition to human immunodeficiency virus infection ( HIV ) [4] and aggravation of TB was seen after Opisthorchis infection[5 , 6] . Another study reported a possible link between IHI and dysfunction of the protective immune response to Bacillus Calmette Guérin ( BCG ) vaccine[7] . These findings may have important implications for the strategy to control PTB and IHI in China , which has a high burden of TB and parasitic infections . There have been only two studies about both diseases in China , one of which provided the prevalence maps of PTB without prevalence predictors , and the other provided the prevalence maps of soil-transmitted helminths using data from different surveys[8 , 9] . Prevalence predictors and the geographical distributions of these diseases have not been documented at the same time using uniformly-collected data; nor has joint spatial analysis of both diseases been presented . Therefore , it is felt to be essential to explore local variations in active PTB and IHI with the aim of detecting joint clustering of both diseases using the latest national surveys with uniform approaches .
This study was approved by the Ethics Review Committee ( ERC ) of China CDC and ERC of National Institute of Parasitic Diseases , China CDC . All the data were got from databases or yearbooks , not involved in individuals . Therefore , the informed consent was not necessary in this study . The dataset of active PTB prevalence was obtained from the 2010 , national TB epidemiological survey[1] , which included 176 survey sites with 252 , 940 participants across the country . In addition to the national survey sites , the provinces of Shandong , Henan , Guangdong , Hainan , Sichuan , Gansu , Ningxia and Xinjiang used the same protocol for extra , provincial survey sites including a total of 151 provincial survey sites with 210 , 877 participants . Therefore , active PTB prevalence of total 327 survey sites including national and provincial levels were analyzed in this study ( Fig 1A ) . In the survey , chest X-ray was performed on all subjects , and smear microscopy and culture of sputum specimens were carried out on all subjects who showed symptoms of PTB or an abnormal chest X-ray result[1] . Overall prevalence of IHI is infection with any helminth species . The dataset on overall prevalence of IHI was obtained from the national epidemiological survey for major human parasitic diseases conducted from 2001 to 2004[2] , which included 687 survey sites with 343 , 500 participants across the country that were analyzed in this study ( Fig 1B ) . In the survey , the Kato-Katz technique for stool specimens was used to examine the eggs of intestinal helminths , the test tube filter paper culture method was used to identify Ancylostoma duodenale and Necator americanus and examine other nematode larvae , and the adhesive cellophane anal swab method was used to examine the eggs of Enterobius vermicularis and Taenia spp . During the fecal examinations , eggs or larvae of other parasites other than the above mentioned parasites were also recorded . The survey showed that major intestinal helminths were Ascaris lumbricoides , hookworm , Trichuris trichiura , Clonorchis sinensis and Taenia spp . Proxies of socio-economic , climatic , geographical and environmental factors were extracted as covariates from different readily accessible sources , as shown in Tables 1 and 2 . The gross domestic product ( GDP ) per capita , population density and urban extents with a binary indicator of urban/rural extent were included in the analysis to capture influences of social developments and human activities on both diseases[8 , 10–13] . Climate zones consisting of equatorial , arid , warm , snow and polar zones , precipitation , air temperature and land surface temperature ( LST ) for day and night were used to reflect impacts of climatic factors on both diseases[8 , 10 , 11 , 14–20] , among which air temperature was only included in the analysis of active PTB[19] , and LST only in the analysis of IHI[8 , 11 , 14–16 , 18] . Elevation and water bodies were applied to the evaluation of relationships between geographical factors and both diseases[8–11 , 14–18] , among which Euclidean distances from survey sites to water bodies were only included in the analysis of IHI[8 , 11 , 14–16] . Vertical columnar density ( VCD ) of nitrogen dioxide ( NO2 ) , VCD of sulfur dioxide ( SO2 ) , concentration of particulate matter of 2 . 5 micrometers ( PM2 . 5 ) , soil moisture and normalized difference vegetation index ( NDVI ) were used to assess influences of environmental factors on both diseases[8 , 10 , 11 , 14 , 15 , 17–19 , 21 , 22] , among which VCD of NO2 , VCD of SO2 and PM2 . 5 concentration were only included in the analysis of active PTB[19 , 21 , 22] , and soil moisture and NDVI only in the analysis of IHI[8 , 10 , 11 , 14 , 15 , 17 , 18] . GDP per capita and population density were obtained from the Chinese annual , full-text database , and other data were downloaded from websites providing free geospatial data products . All of the collected covariates for more than one year were averaged . Maps of all covariates can be seen in Fig 2 . All survey sites as well as supporting data for each diseases were converted into feature ESRI datasets ( ESRI Inc . , Redlands , CA , USA ) and then further converted into ESRI raster datasets as needed . All data were processed with ArcGIS 10 ( ESRI ) . The spatial variations in prevalence of active PTB and IHI were modeled using Bayesian geostatistical logistic regression models . The method used is a combination of the logistic regression model and Bayesian Kriging model , which can be used for the analysis of geo-referenced binomial data , e . g . , disease prevalence where the outcome variable is bounded between zero and one[23] . The modeling process describes the variability in the outcome variable as a function of the explanatory variables with the addition of a stochastic spatial effect to model the residual spatial autocorrelation . Exponentiation of the model parameters provides the odds ratio ( OR ) for each covariate which indicates the power and direction of relationships between the explanatory and outcome variables . Detailed descriptions of the structure of the Bayesian geostatistical logistic regression models and the process of model assessment are described in the additional file: S1 Text . Markov chain Monte Carlo ( MCMC ) simulation was used to estimate the univariate and multivariate model parameters by geoRglm package of R statistical software ( R version 3 . 0 . 2 , the R Foundation for Statistical Computing ) . Following a burn-in of 100 , 000 iterations , the chain was run for a further 500 , 000 iterations , with every 100th iteration thereafter stored , resulting in a total of 5 , 000 samples from the posterior distributions , and the convergence was assessed by the Brooks and Roberts diagnostics[24] . The median values from the posterior distribution and their 95% Bayesian credible intervals ( CI ) were calculated and exponentiated to ORs and their respective uncertainty measures . Due to convergence and mixing problems when including all of the covariates in the multivariate model , each of the explanatory variables in Tables 1 and 2 was examined independently using a univariate model . All covariates significantly associated with the prevalence of active PTB or IHI ( i . e . , the 95% Bayesian CI for the OR did not include the value 1 ) in the univariate model were selected for the multivariate parameter estimation to eliminate the collinearity of covariates . Any covariates that were non-significant in the multivariate model were discarded from the final model through inspection of the regression parameters and 95% Bayesian CIs[25] . We tried many cut-points of each continuous variable in the univariate model to find which cut-point is significant . For example , we tried cut-points of 25% , 50% and 75% to observe the P value in the model . If no significance , we continued to try cut-points of 12 . 5% , 37 . 5% , 62 . 5% and 87 . 5% . If also no significance , we continued to narrow the range . If all the cut-points were no significant , the variable was removed . Using the geoRglm package of R statistical software , Bayesian Kriging was employed to produce each smooth prevalence map of active PTB and IHI . A 10 ×10 km spatial resolution prediction grid was created at the national-scale , containing covariate values at each prediction location ( grid cell ) . Samples from the predictive distribution for each prediction location were generated using the above MCMC algorithm given the explanatory variables at each grid cell , and the convergence was assessed by the Brooks and Roberts diagnostics[24] . The posterior medians , lower and upper limits of 95% Bayesian CIs , and posterior standard errors from the predictive distributions were extracted to give predicted prevalence and uncertainty estimates at all locations . Based on the predicted prevalence and population density in each grid cell of the smooth prevalence map , we calculated average prevalence of each county and then created a feature ESRI dataset of prevalence by county for both diseases using ArcGIS software . The Natural Breaks ( Jenks ) method was used to classify the predicted values and their standard errors . Validation of predicted prevalence of active PTB and IHI was undertaken by randomly sampling 15% of total survey sites as validation set and running the model using the remaining 85% survey sites ( training set ) and validating the model with the validation set[26] . The accuracy of the prediction was determined in terms of sensitivity and specificity and by the area under curve ( AUC ) of a receiver-operating characteristic ( ROC ) curve[27] , where the predicted values were compared to the observed values dichotomized at prevalence thresholds of ≥ 20% to assess discriminatory performance of predictions[28] . As a general rule , an AUC between 0 . 5 and 0 . 7 indicates a poor discriminative capacity; 0 . 7–0 . 9 indicate a reasonable capacity; and > 0 . 9 indicate a very good capacity[29] . Moreover , we computed the percentage of test locations with the observed disease risk falling inside 95% Bayesian CI of the predicted posterior distribution , and the predictive mean error ( ME ) between the observed prevalence πiobs and the predicted prevalence πipre at location i , where ME = ∑i = 1 ( πiobs -πipre ) / n ( i = 1 , … , n ) [8 , 17 , 30] . Using the GeoBUGS package , version 1 . 4 . 3 of the WinBUGS software ( Medical Research Council and Imperial College of Science , Technology and Medicine , UK ) , the above feature ESRI dataset of prevalence by county was used to fit a Bayesian shared component model to jointly analyze the spatial variations of both diseases’ prevalence with common latent risk factors . We assumed that the county-specific relative risks ( RRs ) of both diseases’ prevalence depend on a shared latent component common to active PTB and IHI , plus additional latent components specific to each disease[31] . These latent components act as surrogates for unmeasured risk factors of prevalence that affect both or only one of the diseases respectively[31] . Detailed descriptions of the structure of the Bayesian shared component model and the process of model assessment are described in the additional file: S2 Text . Statistical inference of the Bayesian shared component model was made by using the same MCMC algorithm as for the Bayesian geostatistical logistic regression model , and the convergence was assessed using the Brooks and Roberts diagnostics[24] . For this model , the proportion of variability explained by each component for both disease datasets was derived from the empirical variances[32] . The fitting of various models is measured with the deviance information criterion ( DIC ) ; the lower the DIC , the better the model fit[33] . Many studies indicated that Bayesian shared component model was superior in terms of goodness of fit , compared with the individual modeling of diseases[32 , 34–36] . Therefore , we did not compare goodness of fit between the Bayesian shared component model and other relevant models in this study .
It can be seen in Table 2 that the median ( interquartile range [IQR] ) prevalence were 414 / 100 , 000 ( 222–710 / 100 , 000 ) and 10 . 0% ( 2 . 4–27 . 9% ) for active PTB from 327 survey sites and IHI from 687 survey sites , respectively . The geographical distribution of survey sites and the observed prevalence for each disease are shown in Fig 1 . The median ( IQR ) or proportion of socio-economic , climatic , geographical and environmental covariates for survey sites of both diseases are listed in Table 2 and maps of the spatial distribution of all covariates used in Bayesian geostatistical logistic regression model are provided in Fig 2 . In the univariate Bayesian geostatistical logistic regression model , GDP per capita , population density , urban extents , climate zones , elevation , VCD of NO2 and PM2 . 5 concentration were significantly correlated with active PTB prevalence , which can be seen in Table 3 . Similarly , in the univariate spatial regression model , GDP per capita , urban extents , climate zones , LST for day , LST for night , NDVI and distance to water bodies were significantly correlated with prevalence of IHI , which can be seen in Table 4 . In the multivariate Bayesian geostatistical logistic regression model , four variables finally retained significant correlation with active PTB prevalence , which can be seen in Table 3 , where GDP per capita > 18 , 400 RMB Yuan had a protective effect on active PTB prevalence ( OR = 0 . 82 [95% Bayesian CI = 0 . 69–0 . 98] ) , and rural regions , the arid and polar zones and elevation > 100 m had significantly increased risk effects on active PTB prevalence ( OR = 1 . 31 [95% Bayesian CI = 1 . 08–1 . 58] , OR = 1 . 32 [95% Bayesian CI = 1 . 01–1 . 74] and OR = 1 . 29 [95% Bayesian CI = 1 . 02–1 . 66] , respectively ) . Similarly , in the multivariate spatial regression model , four variables finally retained significant correlation with prevalence of IHI , which can be seen in Table 4 , where GDP per capita > 19 , 400 RMB Yuan had a protective effect on prevalence of IHI ( OR = 0 . 77 [95% Bayesian CI = 0 . 62–0 . 95] ) , as did distance to water bodies > 2 , 000 m ( OR = 0 . 78 [95% Bayesian CI = 0 . 63–0 . 95] ) , and the equatorial and warm zones and NDVI > 0 . 61 had significantly increased risk effects on prevalence of IHI ( OR = 1 . 72 [95% Bayesian CI = 1 . 12–2 . 64] and OR = 1 . 24 [95% Bayesian CI = 1 . 03–1 . 52] , respectively ) . For Bayesian geostatistical logistic regression models , an AUC for predicting active PTB prevalence was 0 . 79 ( 95% CI = 0 . 65–0 . 92 ) and an AUC for predicting prevalence of IHI was 0 . 87 ( 95% CI = 0 . 79–0 . 96 ) , which indicated a moderately good predictive performance . Moreover , within 95% Bayesian CI , the spatial regression models were able to correctly estimate 84 . 7% and 94 . 4% for prevalence of active PTB and IHI , respectively . The MEs for predictive prevalence of active PTB and IHI were 90 / 100 , 000 and 1 . 1% respectively , which suggested that the models slightly underestimated prevalence of active PTB and IHI . The predicted prevalence surface of active PTB from the final spatial regression model is illustrated in Fig 3A , 3B and 3C illustrate the lower and upper limits of 95% Bayesian CI for the prediction . High prevalence of active PTB ( ≥ 900 / 100 , 000 ) was predicted in large areas of two provinces including Tibet and Xinjiang and the juncture of four provinces including Guangxi , Sichuan , Guizhou and Yunnan . Low prevalence ( ≤ 391/100 , 000 ) was predicted in most of the south-eastern coastal-line areas , eastern areas of three provinces including Liaoning , Jilin and Heilongjiang , and the juncture of four provinces including Inner Mongolia , Shaanxi , Gansu and Ningxia . Moderate prevalence ( 392-899/100 , 000 ) was predicted between areas of low and high prevalence . Similarly , the predicted prevalence surface of IHI from the final spatial regression model is illustrated in Fig 4A , 4B and 4C illustrate the lower and upper limits of 95% Bayesian CI for the prediction . High prevalence of IHI ( ≥ 27 . 62% ) was predicted in large areas of nine provinces including Fujian , Jiangxi , Hubei , Hunan , Guangxi , Hainan , Chongqing , Sichuan and Guizhou . Low prevalence ( ≤ 7 . 06% ) was predicted in large areas of 11 provinces including Beijing , Tianjin , Hebei , Shanxi , Inner Mongolia , Liaoning , Shanghai , Jiangsu , Shandong , Henan and Xinjiang . Moderate prevalence ( 7 . 07–27 . 61% ) was predicted between areas of low and high prevalence . For prevalence of both active PTB and IHI , the high prediction uncertainties were correlated with high prevalence areas , which can be seen in the additional file: S1A and S1B Fig , respectively . The shared component of RRs for active PTB and IHI derived from Bayesian shared component model is shown in Table 5 and Fig 5 . The shared term captured 28 . 8% ( 95% CI = 26 . 5–30 . 9% ) of the total spatial variation in active PTB , among which 75 . 1% ( 95% CI = 63 . 0–81 . 2% ) was spatially correlated . The shared term captured 69 . 9% ( 95% CI = 63 . 9–74 . 5% ) of the total spatial variation in IHI , among which the same proportion as active PTB ( 75 . 1% [95% CI = 63 . 0–81 . 2%] ) was spatially correlated . Most striking is a large cluster with higher estimation of the shared component ( RR > 1 . 0 ) in 12 provinces including Anhui , Fujian , Jiangxi , Hubei , Hunan , Guangdong , Guangxi , Hainan , Chongqing , Sichuan , Guizhou and Yunnan . The prediction uncertainties of shared component can be seen in the additional file: S2A Fig . The disease-specific components of RRs for active PTB and IHI derived from Bayesian shared component model are shown in Table 5 and Figs 6 and 7 . One disease-specific term captured 71 . 2% ( 95% CI = 69 . 1–73 . 5% ) of the total spatial variation in active PTB , among which 99 . 9% ( 95% CI = 99 . 8–100 . 0% ) was spatially correlated . The other captured 30 . 1% ( 95% CI = 25 . 5–36 . 1% ) of the total spatial variation in IHI , among which 83 . 7% ( 95% CI = 73 . 1–86 . 7% ) was spatially correlated . The disease-specific component for active PTB had a distinct spatial pattern with higher estimation ( RR > 1 . 2 ) in large areas of seven provinces including Guangxi , Sichuan , Guizhou , Yunnan , Tibet , Qinghai and Xinjiang and the juncture of three provinces including Henan , Hunan and Shaanxi , as shown in Fig 6 . The disease-specific component for IHI presented a different spatial pattern with higher estimation ( RR > 1 . 0 ) in large areas of 12 provinces including Anhui , Fujian , Jiangxi , Hubei , Hunan , Guangdong , Guangxi , Hainan , Chongqing , Sichuan , Guizhou and Yunnan and sparse areas of the remaining provinces , as shown in Fig 7 . The prediction uncertainties of disease-specific components can be seen in the additional files: S2B and S2C Fig .
Although the control of both TB and IHI have progressed in China[37 , 38] , there still are millions of new cases of each disease every year . Syndemics of active PTB and IHI may significantly inhibit host immune systems , increase antibacterial therapy intolerance and even alter the protective immune response to vaccination against TB[39] , underlining the importance of exploring co-endemic areas . However , there are few model-based , nation-wide predictive infection risk maps for active PTB and IHI in China[8 , 9] . The estimation of prevalence predictors at the national level and presentation of predictive prevalence maps for active PTB and IHI , as well as co-endemic RR maps of both diseases’ prevalence , are new and accurate as the investigations are based on two recent , national surveys using uniform diagnostic approaches[1 , 2] . Recently , Bayesian geostatistical analysis was extensively applied to the prediction of parasitic diseases , such as schistosomiasis[40–43] , malaria[26 , 30 , 44–47] , leishmaniasis[48] , soil-transmitted helminth infections[8 , 10 , 11 , 14–18] , lymphatic filariasis[47 , 49] , but so far there are only few applications to the prediction of TB[50–53] . In addition , Bayesian geostatistical methods have almost exclusively been focused on spatial modeling of a single disease . Here , Bayesian geostatistical techniques was shown to support separate and joint spatial analysis of two different infections , i . e . active PTB and IHI . The approach used in our analysis identified important predictors related to active PTB and IHI . Model validation suggested moderately good predictive ability of our final models according to the validation results that AUCs of prediction were 0 . 79 and 0 . 87 and proportions of the observed prevalence correctly predicted within 95% Bayesian CI were 84 . 7% and 94 . 4% for active PTB and IHI , respectively . Our final models demonstrated similar , superior predictive performance compared to other studies[8 , 10 , 14 , 16–18] . The MEs ( 90 / 100 , 000 and 1 . 1% for active PTB and IHI , respectively ) in this study indicated a slight underestimation of prevalence , which had also been observed in other studies[8 , 14 , 17 , 43] . Therefore , we believe that our predictions provided stable and reliable information about the prevalence of both diseases . Our results indicated that GDP per capita and population density had negative association with active PTB prevalence in the univariate model , while rural regions , arid and polar zones , elevation , VCD of the NO2 and PM2 . 5 concentrations showed positive associations . No other study presented all these predictors at a time previously[12 , 19–22] . Although all these predictors showed association with active PTB prevalence in the univariate model , results of the multivariate model showed that only GDP per capita , urban extents , climate zones and elevation still retained association with active PTB prevalence , which may suggest that these four predictors had greater impact on active PTB prevalence than other predictors in China . However , other studies showed that there was a negative correlation between altitude and TB prevalence[54–57] . In this study , the positive correlation between altitude and active PTB prevalence possibly indicated that the risk effects of other factors overwhelmed the protective effect of altitude[58] . It is indisputable that socio-economic development can inhibit the transmission of various diseases including IHI[8 , 18 , 42 , 48] and our findings were consistent with an earlier study in China[8] . We also found that GDP per capita and distance to water bodies had a negative association with prevalence of IHI in the univariate model , while rural regions , the equatorial and warm zones , LST for day and night and NDVI had a positive association . Despite that all these factors are correlated with the prevalence of IHI in the univariate model , only GDP per capita , climate zones , NDVI and distance to water bodies still retained correlation with prevalence of IHI in the final multivariate model . This may suggest that these four factors had a greater impact on the prevalence of IHI than other factors in our study . Our predictive prevalence maps for active PTB and IHI presented geographical distribution patterns , which were consistent with previously released maps of both diseases[8 , 9] . Unsurprisingly , there were obviously different geographical distribution patterns of prevalence between active PTB and IHI . For example , moderate to high prevalence of active PTB was predicted in western regions of the country where only low IHI prevalence was predicted; low to moderate prevalence of active PTB was predicted in the more northern parts of China and the south-eastern , coastal regions where low prevalence of IHI was predicted; moderate to high prevalence of active PTB was predicted in the south-western regions where high prevalence of IHI was predicted . The shared component explains the fraction of total variation in spatial RRs for each disease in the shared component model . In this study , for IHI about 70% of the total between-area variation in risk was captured by the shared component , while for active PTB about 29% of the total between-area variation in risk was captured by the shared component . This suggests that the shared component had a slightly weaker association with risk of active PTB than with risk of IHI . Although there was a difference between the fractions of both diseases , the shared component still represented the joint prevalence . The spatial analysis of joint prevalence of active PTB and IHI showed that a large cluster of both diseases was found to be located in the south-western regions of the country , which was consistent with the overlapping areas of high prevalence based on separate predictive maps of both diseases . These findings proved the accuracy and reliability of the shared component model used in this study . The shared component model makes the assumption that there are unobserved covariates that display a spatial structure common to both diseases[31] . The analysis results of the separate multivariate model for active PTB and IHI in our study showed that proxies of socio-economic and climatic factors were simultaneously associated with prevalence of both diseases . The socio-economic factors had the same effects on prevalence of both diseases , while the climatic factors showed the opposite effect including positive correlation between the arid and polar zones and active PTB and positive correlation between the equatorial and warm zones and IHI . Therefore , we inferred that socio-economic factors such as GDP per capita were the main unobserved covariates that determined the co-endemic patterns of active PTB and IHI because they were common to both diseases . Moreover , we also observed that the spatial pattern of disease-specific component for active PTB were similar to the distribution of urban extents , climate ( arid and polar ) zones and elevation in maps , which may indicate that they represented additional risk factors only relevant to active PTB but not to IHI . Similarly , the spatial pattern of disease-specific component for IHI were similar to the distribution of climate ( equatorial and warm ) zones , NDVI and distance to water bodies in maps , which may indicate that these factors were the additional risk factors only relevant to IHI but not to active PTB ( see Fig 8 ) . All the covariates used in this study were extracted from different accessible sources . Hence , the accuracy and spatial resolution were diverse , possibly influencing the capture of disparities of covariates across the country at uniform scale[10] . For example , although precipitation , air temperature , VCD of SO2 and soil moisture were captured in other studies[8 , 10 , 17–21] , we did not find them in our study . One possible reason is that data quality of these factors affected the capture of the model , while another possible reason is that these factors really had no association with both diseases . Additionally , both diseases and covariates had high heterogeneity at the national-scale , but we did not divide these data into three or more groups to present the variations in the Bayesian geostatistical logistic regression equation due to limitations of computing power . Furthermore , the shared component model assumes that the shared and specific component are independent , which ignores the possibility of interaction between the real covariates[31] . In view of these limitations , although we believe that our findings provide a useful approximation for both diseases , we caution against over-interpretation of our findings . In conclusion , our study simultaneously provided prevalence predictors and predictive prevalence maps for active PTB and IHI as well as co-endemic RR maps of both diseases’ prevalence at the national scale . We found that co-endemic areas of active PTB and IHI were located in the south-western regions of China , which may be determined by socio-economic factors such as GDP per capita . Moreover , disease-specific distributions of active PTB may be determined by exclusive factors including urban extents , the arid and polar zones and elevation , while disease-specific distributions of IHI may be determined by exclusive factors including the equatorial and warm zones , NDVI and distance to water bodies . We believe that our estimations provide a valuable assessment of separate and co-endemic situations of active PTB and IHI , therefore we hope that this first effort will contribute useful information to plan syndemic control strategies in co-endemic areas . Additionally , the combination of Bayesian geostatistical techniques in this study provided a new avenue for exploring high prevalence areas of multi-disease syndemics and to understand their interactions at the macro-scale . | Pulmonary tuberculosis ( PTB ) and intestinal helminth infections ( IHI ) are infectious diseases of poverty , and both diseases affect millions of individuals every year in China . However , a neglected topic for both diseases is their co-endemicity , which mostly occurs in poor areas . This is the first time the co-endemicity of PTB and IHI and their risk factors have been explored by means of a Bayesian geostatistical logistic regression model , a Bayesian Kriging model and a Bayesian shared component model , based on data from the national surveys . Our results indicate that active PTB and IHI prevalence are associated with economic and ecological indices , both individually and collectively , with different disease spectra in different ecosystems . Additionally , we find that the south-western regions of China are the largest clustering areas for prevalence of both diseases , where socio-economic factors , such as GDP per capita may be common risk factors . Both socio-economic factors and epidemiological patterns relevant to control strategies for active PTB and IHI are illustrated clearly in this study , so we have reason to believe that they are essential for devising effective control strategies and should be considered in the control programs for both diseases . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"economic",
"analysis",
"china",
"population",
"dynamics",
"tropical",
"diseases",
"geographical",
"locations",
"social",
"sciences",
"parasitic",
"diseases",
"bacterial",
"diseases",
"population",
"biology",
"digestive",
"system",... | 2016 | Co-endemicity of Pulmonary Tuberculosis and Intestinal Helminth Infection in the People’s Republic of China |
Troponin C ( TnC ) is implicated in the initiation of myocyte contraction via binding of cytosolic and subsequent recognition of the Troponin I switch peptide . Mutations of the cardiac TnC N-terminal regulatory domain have been shown to alter both calcium binding and myofilament force generation . We have performed molecular dynamics simulations of engineered TnC variants that increase or decrease sensitivity , in order to understand the structural basis of their impact on TnC function . We will use the distinction for mutants that are associated with increased affinity and for those mutants with reduced affinity . Our studies demonstrate that for GOF mutants V44Q and L48Q , the structure of the physiologically-active site II binding site in the -free ( apo ) state closely resembled the -bound ( holo ) state . In contrast , site II is very labile for LOF mutants E40A and V79Q in the apo form and bears little resemblance with the holo conformation . We hypothesize that these phenomena contribute to the increased association rate , , for the GOF mutants relative to LOF . Furthermore , we observe significant positive and negative positional correlations between helices in the GOF holo mutants that are not found in the LOF mutants . We anticipate these correlations may contribute either directly to affinity or indirectly through TnI association . Our observations based on the structure and dynamics of mutant TnC provide rationale for binding trends observed in GOF and LOF mutants and will guide the development of inotropic drugs that target TnC .
Sarcomeres contract owing to the translocation of the thick filament , comprised of myosin , along actin chains constituting the thin filament ( TF ) . Contraction is initiated and regulated by Troponin proteins tethered to actin , including Troponin C ( TnC ) , Troponin I ( TnI ) and Troponin T ( TnT ) , as well as Tropomyosin ( Tm ) . Specifically , binds to TnC , thereby unveiling a hydrophobic region necessary for binding the TnI switch peptide . Liberation of the TnI regulatory unit from the TF initiates a shift in Tm [1] , thus enabling the weak binding of myosin to actin . Subsequent conversion of Tm to the unblocked state permits a cycle of strong myosin binding and propagation along the TF ( cross-bridge cycling ) . A number of human cardiac diseases including hypertrophic cardiomyopathy ( HCM ) [2] , restrictive cardiomyopathy ( RCM ) [3] and dilated cardiomyopathy ( DCM ) [4] have been attributed to mutations in thin filament , thick filament and associated proteins of the sarcomere . RCM and HCM mutations have been shown to increase sensitivity of force generation as measured by pCa50 , whereas DCM mutations reduce this trend . A large number of mutations leading to HCM , RCM , and DCM phenotypes have been collectively identified [5] but only one LOF , DCM-associated mutation has been found in TnC ( D75Y [6] ) . The prominent role of TnC in force development has thus attracted therapeutic strategies to tune its and TnI affinity including drug-design [7] and protein engineering approaches [8]–[10] . In particular , mutation studies of full-length TnC have revealed engineered variants that shift the equilibrium constant , ( or pCa50 ) , leading to altered force development akin to GOF and LOF [8] . For instance , V44Q and L48Q mutations investigated by [8] , [10] , [11] have been reported to exhibit GOF-like phenotypes in skinned cardiac fibers , with pCa50 values of 6 . 29 and 6 . 13 ( in isolated F27W TnC ) , respectively , relative to the wild-type value of 5 . 48 [8] . Furthermore , Tikunova et al . [10] reported for several GOF mutations including V44Q and L48Q that faster association rates ( 4 . 4 to 5 . 2-fold ) contributed more to the increased rather than slowed dissociation ( approximately 2 . 8-fold ) . In comparison , the E40A and V79Q mutations examined by [8] present LOF-like alterations in force generation with pCa50 values of 5 . 16 and 5 . 30 , respectively . While these studies have implicated binding as the primary factor in reshaping contractile activity , the structural and dynamical basis of the mutations' effect on TnC is largely unknown . Structure determination via X-ray crystallography [12] and NMR [13]–[15] has yielded important insight into the molecular basis of TnC function . These studies indicate that TnC consists of two domains: the C-terminal domain is affixed to the thin filament , while the N-terminal regulatory domain is responsible for binding at physiological concentrations , and TnI . The TnC N-terminal domain consists of five helices , HN ( 4–8 ) , HA ( 14–24 ) , HB ( 38–48 ) , HC ( 54–64 ) and HD ( 73–85 ) , two beta sheets ( 36–37 ) , ( 71–72 ) ( Fig . 1 ) . Loops ( 25–35 ) , ( 65–70 ) of sites I and II form EF hands ( helix-loop-helix ) that selectively bind , although in cardiac TnC , only site II is physiologically active . Within the EF hand , several acidic residues ( D65 , D67 , and E76 ) coordinate , along with S69 and T71; we collectively refer to these amino acids as chelation residues . Prior experimental and theoretical work have leveraged these structural data to probe rapid , nanosecond timescale conformational dynamics that are correlated with binding [14] , [16] . Lim and coworkers [6] characterized a TnC mutant ( D75Y ) isolated from a patient with DCM and demonstrated its decreased binding capacity and disruption of normal structural dynamics . Varughese and Li further investigated via MD changes in the structural dynamics of cardiac troponin , including TnC , upon binding bepridil , a known inotropic agent [17] . Lindert and coworkers [16] combined long time-scale MD simulations and BD simulations to understand the dynamics of wild-type TnC in its -free , -bound , and TnI -bound states , as well as V44Q [18] . Recently , a combined experimental and theoretical approach examined binding and the structural stability of a GOF mutant L48Q [19] . We seek to extend these studies by 1 ) comparing LOF and GOF mutants to better contrast differences in apparent sensitivity and 2 ) explore longer simulation times comparable to dynamics captured by NMR order parameters . By a combination of molecular and Brownian dynamics , our approach identifies structural and dynamic factors impacting binding in light of GOF and LOF mutations . The outcome of this study provides greater insight into the mechanisms of structure/function relationships for N-terminal cardiac TnC that are important to myofilament activation .
A requisite for high affinity binding is competent formation of site II interactions . To gauge the integrity of this coordination amongst the mutants , we report the mean residue/ distance for site II residues D65 , D67 , S69 , T71 and E76 in Table 2 and Fig . S3 . We found close coordination ( distance ) of with D65 , D67 and E76 ( Fig . S3 ) that evidence strong binding and were consistently preserved over the simulation period . We classified the interactions with as mono-dentate versus bi-dentate interactions and found that E65 is primarily mono-dentate , E76 is bidentate , and D67 oscillated between mono- and bi-dentate coordination . S69 and T71 exhibited larger distances of approximately 4 to 5 Å ( not shown ) indicative of comparatively weaker interactions . Altogether , chelation distances were consistent and well-stabilized across holo cases , with no distinction between LOF and GOF mutants .
Our simulations of GOF and LOF variants of the TnC N-terminal regulatory domain recapitulate a number of structural predictions in [8] based on mutants of the full-length F27W TnC For instance , differences in alpha helical character in the apo state were found amongst all mutants , particularly at , with V44Q and E40 presenting the largest increase and decrease , respectively , while V79Q and L48Q had intermediate values . These findings mirror those reported using circular dichroism [8] , although in this study , the measurements assessed the alpha helical content of the entire protein . We also observed a decrease in length upon binding , yet Parvatiyar et al . report an increase in helical secondary structure for the full-length TnC . Since our simulations focus solely on the N-terminal domain , it is possible that binding might increase helical content of the C-terminal domain or linker region , thus offsetting decreases at . Moreover , we discovered that the packing of TnC helices was disturbed for almost all of the mutants we studied . For instance , among the GOF mutants , the V44Q mutation significantly disrupted the hydrophobic contacts between helices and in the holo state ( Fig . S6a ) , as was originally proposed by [8] and [10] . This destabilization may contribute to the larger opening angle and frequency for this mutant . L48Q , on the other hand , presented slight structural changes involving that evidence disruption of the hydrophobic packing , but nevertheless did not cause noticeable opening in our simulations ( Fig . S6b ) . For the LOF mutants , in remarkable agreement with Parvatiyar [8] , we found that the V79Q mutation formed a possible hydrogen bond with D75 , a mutation site involved in ( D75Y [6] ) ( Fig . S6c ) . This alteration in the side chain conformation appeared to disrupt the packing around , which may reduce the ability to form the close D67/E76 contacts observed for the and wild-type cases . In contrast , we found that the E40A mutation did not disrupt the packing , per se , but rather decoupled the WT's apparent hydrogen bond network linking S37 , E40 and R43 ( Fig . S6d ) . Interestingly , its impact appeared to be largely allosteric in nature , as we observed an increase in site II dynamics despite a minimal change at sites 37 and 43 . One explanation for this allosteric change is that the mutation alters the coupling of beta sheets bridging sites I and II , thus perturbing the length and D67/E76 pairing . While these changes amongst mutants are quite apparent from relatively short MD simulations , elucidation of longer timescale conformational changes through direct simulation or extrapolation of principle components [16] may provide additional insight into the role of TnC during the comparatively slow process of sarcomere contraction . In this study , we sought to understand why affinity is enhanced in isolated TnC for GOF mutants and decreased for LOF mutants . If the mutations predominantly impacted the association rate without affecting the equilibrium constant , one might speculate that the mutations alter the transition state barrier to binding . However , the equilibrium constants measured in [10] were in fact found to substantially vary , thus we sought to identify qualitative factors contributing to the free energy of the apo and holo states . We examined this by considering structural and dynamic factors for both apo and bound TnC mutants . Firstly , the binding free energy is determined by the difference between the free energy of the bound substrate-receptor complex versus the isolated apo-receptor and substrate . Our initial hypothesis was that and mutations may either disrupt the native structure of TnC or alter the ability of TnC to coordinate ; however , we found that overall the apo and holo states share remarkable structural similarity , based on RMSD comparison of the entire N-domain . with localized disruptions in the packing about and near , where binds . Second , we evaluated the magnitude of the electrostatic interactions between substrate and receptor as a determinant of ligand binding affinity . Because of the structural similarity amongst the mutants and the remote position of the mutations relative to site II , we did not observe any appreciable changes in the electrostatic potential near the binding site . This is in contrast to the case of D75Y explored by Lim et al . , for which the charged-to-neutral mutation within site II would plausibly cause a significant change in the local electrostatic potential [6] . Instead , we found that strong electrostatic interactions between the ion and its chelating residues afforded tight coordination of as evidenced by the constrained dynamics in , despite minor changes in RMSD amongst the mutants . This demonstrates that TnC is tolerant of changes in backbone conformation , provided that the chelation residues can form optimal interactions with . This is not surprising , as the highly negatively charged chelation residues are strongly attracted to the divalent cation , and thus the driving force for forming the optimal conformation is likely quite large . In so far that the enthalpy of binding is dominated by the charge-charge interactions arising from binding , it is plausible that the enthalpic contribution to binding is roughly comparable amongst the mutants . Estimating this contribution is the subject of a follow-up investigation . Given the qualitative similarity of this electrostatic enthalpic contribution , it is therefore possible that entropic considerations are more significant for discriminating low- from high-affinity mutants , based on our estimated order parameters and correlation data . Common to both mutations types are dynamic regions at sites I , II and as evidenced by our estimates of NMR order parameters . However , a key distinction between the mutation types is that the site II binding domain in the apo state is more rigid , and practically identical to the holo state , for the GOF mutants , in sharp contrast to the LOF mutants and findings for WT TnC ( Fig . 8B [16] ) . In so far that order parameters are a qualitative measure of entropy [23]–[25] , a lesser change in order parameters for the GOF relative to LOF mutants in this study could reflect a smaller entropic cost for binding and thus result in increased affinity . These findings indicate that changes in affinity are driven by alterations in TnC dynamics , not in coordination , in agreement with [8] . They raise the possibility that calcium sensitivity modulation of TnC might be controlled through control of dynamics . Common to both species are also extensive correlated motions in the apo states , as well as a conserved set of correlation patterns in the bound state . GOF and LOF species both exhibited significant anti-correlation between and the C-terminal end of near I26 . However , only for GOF mutants did we observe a complementary anti-correlation between the C-terminal region of and the entire helix . We postulate that this motion alters the angle , for which we speculate that I26 serves as a pivot point though its hydrophobic interactions with . We observed additional correlations between helices and for GOF and wild-type ( see Fig . 6C , [16] ) , but not for LOF mutants , which indicated the V79Q and E40A mutants may have destabilized the hydrophobic residues between these helices in line with previous predictions [8] , [19] . It is further possible that the greater extent of positive correlation between and for the GOF cases relative to the wild-type indicates more extensive stabilization of site II and perhaps contributes to their enhanced affinity . Similarly , we saw a greater number of off-diagonal correlations for GOF relative to LOF , which may evidence greater entropic stabilization of the entire protein for the GOF cases . In comparison to the wild-type correlations reported in Lindert et al . , the comparable off-diagonal correlations for L48Q and considerably more extensive motions for V44Q are in line with the latter having greater affinity . Therefore , in addition to the enthalpic factors stated in the previous section , 1 ) GOF mutants may have a more favorable , entropy-stabilized bound state relative to LOF , 2 ) the distribution of correlated motions may play a role in increasing affinity . Interestingly , two mutations ( L48Q and V44Q ) were reported to have a larger impact on versus , with four- to five-fold larger with respect to wild-type relative to a partially compensating three-fold increase in [10] . A secondary objective of this study was thus to shed light on the molecular basis of altered s for the mutants , and to determine whether complementary mechanisms were at play for the cases . The diffusional component of the association rate , is strongly influenced by electrostatic interactions and geometry , thus we sought to determine to what extent these factors impacted . To explain our results , we resort to the the transient-encounter complex theory formulated by Zhou and others [26] , which suggests that association is divided into two regimes - diffusion limited binding to the protein surface ( transient encounter complex ) , followed by a post-encounter reorganization . The transient encounter complex can be defined as the surface at the cusp of the bound and unbound potential energy surfaces , whereas the post-encounter reorganization corresponds to localized conformational changes that bring the transient complex to the native bound state . Electrostatics are known to often drive association between substrate and receptor to form the encounter complex . Our predictions of were on the order of 109 to 1010 [1/Ms] , which suggests the prominent role of electrostatics in driving the initial states of association , as found for the wild-type [16] . However , since the electrostatic potential and predicted diffusional encounter rates were nearly indistinguishable amongst the GOF and LOF mutants , differences in the experimentally measured values were not likely electrostatics in nature . This leaves the post-encounter regime as a possible discriminator amongst the mutants . Furthermore , as we later argue , post-encounter events may reconcile the discrepancy between the magnitude of the predicted s and the substantially smaller estimates for the overall ( 107 to 108 [1/Ms] [27] , [28] for the wild-type . When reaches site II we postulate that several post-encounter factors impact the rate of binding , including enthalpy and the reorganization time associated with an induced-fit mechanism . Based on the highly solvent-exposed binding site and abundance of negatively charged residues that coordinate cations , we expected a highly favorable enthalpic interaction as approached site II along the reaction coordinate . However , coordination of required the energetically-unfavorable displacement of the solvation shell around ; this partial dehydration appeared to be accompanied by poly-dentate interactions with D67 and E76 . Our PMF calculations indicated that these offsetting energetic factors manifest a modest barrier of 3 [kcal/mol] , which we expected would reduce the association rate by nearly three orders of magnitude ( by Kramer's theory ) . This explanation could align our predicted values of [1/Ms] with experimentally-observed s ( [1/Ms] ) . Since the coordination number of is identical across all mutants , we speculate that the energetic cost of dehydration should be similar for all cases; if dehydration in fact dominates the free energy barrier , we might expect a comparable decrease in association rates amongst the studied mutants . The differences in the time required for the induced fit of from the apo state are expected to impact the overall association rate . Our simulations of the apo state indicated that site II more closely resembled the holo state for GOF relative LOF to mutants , which we attributed to tighter D67/E76 interactions . The close D67/E76 distance likely contributes to , or benefits from , the increased character of V44Q . While the helical length for L48Q is marginally greater than the LOF mutants and wild-type ( data not shown ) , it is apparently sufficiently stable to ensure D67/E76 pairing . We speculate that the close pairing of D67 and E76 , through transient coordination of waters near site II , may contribute to optimal placing of chelation residues and thus constitute a ‘pre-formed’ binding site . Because the GOF mutants appear to spend more time in productive conformations relative to LOF mutants ( as measured by D67-E76 distance ) , the reduced time required to form the binding site could account for their faster experimentally-observed s . This model could explain the findings of Tikunova et al . [10] , who suggested that mutations shift equilibrium between bound-form and apo form , thereby giving rise to enhanced binding affinity for the GOF mutants . In the previous section , we rationalized trends in altered affinity based on apo and -bound TnC , in order to explain experimental results obtained in absence of TnI . In physiological systems , however , TnI binding is known to enhance affinity for TnC , primarily because of the activation of cross-bridges and subsequent alteration of myofilament lattice properties during rigor [29] . In this context , it is possible that the correlated motions observed in holo TnC mutants promote productive TnI binding , although this was not explicitly modeled in our study . Our future work aims to include the contribution of the myofilament lattice to association , at which point more direct comparisons against cooperativity data can be made . Nevertheless , a prerequisite for the formation of the open state supporting TnI binding is the exposure of a hydrophobic patch between residues L29 , A31 , K39 , and E66 [21] . The transition to the open state entails the progression of helices / ( HB ) away from helices , and ( NAD ) , which is associated with enhanced affinity [10] and is suggested to be a common mechanism for the enhanced affinity observed for several mutations along the BC/NAD interface [30] . Our simulation results show substantial exposure of this region for holo V44Q mutant , as evidenced by / angles approaching 100 degrees , versus 130–140 degrees typically observed for the closed apo and holo states of wild-type cardiac TnC . Specifically , for V44Q we observe that significantly deviates from the TnC NAD core , which would be in line with its greater affinity in the intact thin filament . For all other mutants , including the L48Q GOF mutant and the WT ( see Figure 2C in [16] ) , movement is minimal and only transient , sub-nanosecond decreases in the / angle to the 100 degree range are observed , which are likely too fast to be resolved given the temporal resolution of NMR . Thus , either movement is slow relative to our simulation timescale or more likely , TnI plays an active role in stabilizing the open state . In fact , we have recently observed in micro-second timescale simulations of TnC that the V44Q mutation reduces the free energy barrier to decreasing the / angle , which may increase the rate of forming the open state [18] . We now focus on correlated motions that may contribute to an opening event . Prior studies [6] suggest that concerted motions might be implicated in affinity , either directly or indirectly through enhanced interactions with TnI . One speculation based on our correlation data is that the positive correlation between beta sheets and ( of and ) initiates a signal between binding at site II and the exposure of the TnI binding surface . We base this on the correlation/anti-correlation pattern between and , reflect tugging of away from the terminus . By tugging on the , which is directly adjacent to the base of , we would expect a widening between and that exposes the hydrophobic surface for TnI . The strong correlation between and may amplify the effect of the tugging between loops and increase the propensity of widening in mutants . We have identified trends that readily discriminate between GOF and LOF mutants . Our findings suggest that differences in calcium sensitivity cannot easily be explained in terms of large structural changes or differences in the electrostatic potential . Instead , the modulation of calcium binding may be due to dynamic motions ( fluctuations and correlations ) , as well as the time associated with reorganizing the binding site for optimal - TnC interactions . These findings suggest that modulation of the dynamic properties of TnC via mutation may represent an attractive avenue for tuning myofilament contraction . At the same time , theoretical investigations aiming to characterize the thermodynamics of binding from simulation would benefit from considering the contribution of multiple , highly dynamic conformational states in equilibrium . We are currently examining whether these findings are reflected in the -associated TnC mutant D75Y . Moreover , to examine the molecular basis of impairment myosin ATPase inhibition characteristic of and observed for the V44Q and L48Q mutants [8] , we seek to apply our modeling approaches to a model of the intact myofilament [12] .
-bound ( 1AP4 ) and -free ( 1SPY ) NMR structures resolved by [14] were obtained from the Protein Data bank [31] . E40A , V44Q , L48Q , V79Q mutations were performed using the Mutate Residue module in VMD [32] . VMD was used to add a TIP4P water box with a 20 Å boundary on the protein as well as sufficient KCl to obtain charge neutrality and a 0 . 15 M solution . Protein and TIP4P waters were parameterized using the Charmm27 [33] force field . Mutated side chains were subjected to 1000 steps of conjugate gradient minimization with the remainder of the protein held fixed . Using NAMD 2 . 7b [34] , the solvent was equilibrated for 5000 steps of NVT MD protein-fixed simulation using a 2 fs integration step at T = 310 K . 12 . 0 Å cutoffs were used for non-bond terms , and PME tolerance , interpolation order and grid spacing were set to 10e-6 , 4 , and 1 . 0 Å ‘respectively . Harmonic constraints of 10 [kcal/mol] were placed on the protein , which was equilibrated for an additional 5000 NVT steps . 5000 steps of unconstrained NVT , followed by 30000 steps of NPT MD were used for the final equilibration steps . Production runs using the NVT ensemble were at least 120 ns in duration . The Adaptive Biasing Force protocol [35] in NAMD was used to estimate the PMF . Structures and MD parameters follow from the the previous section . The colvars procedure in NAMD was used to guide the atom along the reaction coordinate , here chosen as the distance between and the E76 carboxylic acid through the definition of the AtomDistance colvar group . Lower and upper boundaries of 2 and 12 Å , varied in 0 . 1 Å increments were chosen for the colvar AtomDistance , while upper and lower wall constants were both set to 10 [kcal/mol] . Colvar statistics were collected at 4 ps intervals and the fullSamples parameter was set to collect 500 samples at each window prior to biasing . 11 ns of ABF simulation were performed for the PMF estimation . Simulation files for NAMD are provided at Video S 1 Trajectories were analyzed using the R statistical package Bio3D [36] . Pearson correlation coefficients were computed using the Dynamic Cross Correlation Maps ( dccm ) package implemented in Bio3D . Molecular dynamics snapshots ( taken every 6 ps from the trajectories ) were aligned by the protein's C atoms and subsequently clustered by RMSD using GROMOS++ conformational clustering [37] . A RMSD cutoff of 1 . 5 Å ( apo E40A ) , 1 . 8 Å ( apo L48Q ) , 1 . 7 Å ( apo V44Q , apo V79Q ) , 1 . 6 Å ( holo E40A ) , 1 . 4 Å ( holo L48Q ) , 1 . 9 Å ( holo V44Q ) and 1 . 2 Å ( holo V79Q ) was chosen , respectively . These cutoffs resulted in 3 ( holo E40A , holo L48Q , apo V44Q ) , 4 ( holo V44Q , apo E40A , apo L48Q , apo V79Q ) and 2 ( holo V79Q ) clusters that represented at least 90% of the respective trajectories . Helical content for residues 54 through 68 was computed using dssp [38] and reported in units of amino-acid length . Interhelical angles between HA and HB were calculated using interhlx [39] . APBS [40] was used to compute the electrostatic potential for the proteins in openDX format . An ionic strength of 150 mM was assumed . Backbone N-H order parameters were calculated from the MD simulations of all apo and holo systems applying the isotropic reorientational eigenmode dynamics ( iRED ) approach [41] using a 0 . 5 ns window for averaging . The order parameters are calculated with the mat2s2 . py script using a list of eigenvalues and eigenvectors of all the N-H backbone vectors generated by ptraj . Association rates for the diffusional encounter , were computed using BrownDye [42] . PQR files of the cluster centers were generated using pdb2pqr [43] . The calcium pqr file was generated using a charge of +2 and an ionic radius of 1 . 14 Å . Bd_top was used to generate all necessary input files for the BrownDye runs . A phantom atom of zero charge and negative radius ( −1 . 14 Å ) was introduced after the first execution of bd_top . The phantom atom was placed at the position of the calcium ion from the trajectory frame . Its sole purpose is to define a reaction criterion that is spherically symmetric around the expected binding position of the calcium . The reaction criterion was chosen to be 3 . 5 Å within the calcium binding site . 500 , 000 single trajectory simulations were performed on 8 parallel processors using nam_simulation . The reaction rate constants were calculated using compute_rate_constant from the BrownDye package . A weighted average of the rate constants of each cluster center serves as an estimate of the overall association rate for the system . | Muscle cells contract using a network of thread-like protein assemblies called myofilaments . Contraction is preceded by a signal that causes calcium to rush into the cell cytosol , where it can freely diffuse to and bind the myofilament proteins . Troponin C , a calcium sensor located on the thin filament , initiates and regulates the cascade of changes resulting in the generation of force by the thin and thick filaments comprising the myofilament lattice . In heart tissue , pathological conditions known as dilated and hypertrophic cardiomyopathies ( DCM and HCM , respectively ) are in part associated with abnormalities in the ability of the myofilaments to generate force at normal calcium concentrations . Manipulation of Troponin C calcium-binding through protein engineering and pharmaceutical intervention has thus attracted considerable attention as a therapeutic strategy for ameliorating these cardiac defects . In this study , we uncover a molecular basis of altered calcium handling for several engineered Troponin C variants , which provides further insight into tuning its control of myofilament contraction . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"molecular",
"dynamics",
"chemistry",
"computational",
"chemistry"
] | 2012 | Molecular Basis of Calcium-Sensitizing and Desensitizing Mutations of the Human Cardiac Troponin C Regulatory Domain: A Multi-Scale Simulation Study |
Oscillatory neuronal synchronization between cortical areas has been suggested to constitute a flexible mechanism to coordinate information flow in the human cerebral cortex . However , it remains unclear whether synchronized neuronal activity merely represents an epiphenomenon or whether it is causally involved in the selective gating of information . Here , we combined bilateral high-density transcranial alternating current stimulation ( HD-tACS ) at 40 Hz with simultaneous electroencephalographic ( EEG ) recordings to study immediate electrophysiological effects during the selective entrainment of oscillatory gamma-band signatures . We found that interhemispheric functional connectivity was modulated in a predictable , phase-specific way: In-phase stimulation enhanced synchronization , anti-phase stimulation impaired functional coupling . Perceptual correlates of these connectivity changes were found in an ambiguous motion task , which strongly support the functional relevance of long-range neuronal coupling . Additionally , our results revealed a decrease in oscillatory alpha power in response to the entrainment of gamma band signatures . This finding provides causal evidence for the antagonistic role of alpha and gamma oscillations in the parieto-occipital cortex and confirms that the observed gamma band modulations were physiological in nature . Our results demonstrate that synchronized cortical network activity across several spatiotemporal scales is essential for conscious perception and cognition .
Synchronization of oscillatory brain activity on multiple temporal scales across distant cortical regions is thought to constitute a key mechanism for conscious perception and cognition in humans [1]–[3] . Recently , it has become evident that synchronized cortical networks are dynamically established during cognitive processing to selectively route information to task-relevant cortical sites [4] , [5] . In particular , it has been shown that cortical information flow can selectively be controlled by shifting phase relations between cell groups oscillating at similar frequencies [6] . In the visual system , synchronized gamma-band activity , mediated over cortio-cortical callosal connections , might facilitate feature integration across both visual hemifields [7] . However , most evidence is still correlative in nature [3] . As a consequence , it remains unclear whether the observed rhythmic synchronization patterns merely represent concurrent neuronal activity or whether they are causally relevant for the information flow within cortical networks . The frequency-specific modulation of phase relations with subsequent behavioral alterations would constitute an unequivocal confirmation for the functional role of synchronization processes in large-scale neuronal networks . Up to now , it has been considered difficult to selectively manipulate rhythmic brain activity in the human cerebral cortex . Recently , novel methods for entrainment of perceptually relevant brain oscillations have become available [8] . In particular , transcranial alternating current stimulation ( tACS ) has been shown to entrain cortical oscillations in a frequency-specific manner [9] . Phase-dependent tACS effects have been demonstrated in a variety of human [10] , [11] , animal [12] , [13] , and modeling studies [14] , making it a prime candidate to selectively modulate phase relationships in distant cortical regions [15] . Optimized stimulation electrode montages and novel multi-electrode setups now allow a selective cortical stimulation [16] , [17] . The goal of this study was to test whether perceptually relevant neuronal synchronization in large-scale neuronal networks can selectively be modulated with weak electric currents . Recent studies investigating tACS effects on long-range functional coupling remained equivocal and did not provide conclusive electrophysiological evidence for the causal role of synchronized oscillatory activity [15] , [18] . Here , we focused on the role of interhemispheric gamma-band coherence over parieto-occipital areas during ambiguous motion perception . To this end , we utilized the well-known ambiguous stroboscopic alternative motion ( SAM ) paradigm ( Figure 1A ) [18] . In this paradigm , a physically identical stimulus can be perceived as either horizontally or vertically moving [19] . This perceptual bi-stability makes the stimulus ideally suited to assess percept related network differences . The perceived motion direction might alternate spontaneously depending on the level of interhemispheric gamma-band coherence over motion sensitive cortical areas [20] , [21] , and might be mediated via cortico-cortical callosal fibers [22] . To investigate the functional role of interhemispheric gamma-band coherence in ambiguous motion perception , we combined multi-focal 40 Hz high density ( HD ) -tACS at opposite polarities ( in-/anti-phase stimulation ) with concomitant electroencephalographic ( EEG ) recordings to investigate oscillatory brain activity during stimulation . We expected that the different stimulation protocols might bias network synchrony in opposite directions and therefore facilitate or impair interhemispheric integration of the visual tokens into the horizontal percept .
We assessed the mean duration of horizontal and vertical motion perception by means of a motion ratio ( MR = timehorizontal/timetotal ) ( Table S1 ) and found a nearly balanced distribution during sham ( MR: 48 . 7%±1 . 9% , mean ± standard error of the mean [SEM] ) ( Figure 3A ) . Our results revealed a significant increase of perceived horizontal motion during in-phase stimulation ( +1 . 4%±1 . 9%; change to common baseline ) ( Figure 3A , dashed line ) as compared to anti-phase stimulation ( −1 . 5%±1 . 5%; F1 , 13 = 5 . 17 , p<0 . 05 , Cohen's d = 0 . 59; planned contrasts ) . This difference was not significant during the post condition ( F1 , 13 = 2 . 70 , p = 0 . 12; d = 0 . 05 ) ( Table S1 ) . No significant results were found for the switch rate ( reversals per minute ) ( Figure 3B; all p>0 . 1 ) . Taken together , the behavioral results suggest that the targeted manipulation of the interhemispheric phase difference biased the conscious experience of apparent motion . All analyses for the SAM were conducted without the gamma1-range ( spectral estimates in Figure 4A were obtained after spectral smoothing ) to remove the 40 Hz stimulation artifacts . In order to assess whether stimulation selectively affected the gamma-band , grand average interhemispheric coherence values ( Figure 4A ) were submitted to a two-way repeated measures analysis of variance ( RM-ANOVA ) ( session: sham , in- , and anti-phase stimulation; frequency: six bands ) . We report Greenhouse-Geisser corrected values in case of violation of sphericity . A significant interaction of session × frequency ( F1 . 72 , 22 . 38 = 20 . 5 , p<0 . 005; session: F1 , 13 = 40 . 7 , p<0 . 005; frequency: F1 . 56 , 20 . 22 = 52 . 3 , p<0 . 005 ) indicated a frequency-specific modulation of interhemispheric coherence . Importantly , the increase in gamma-coherence during horizontal motion perception was present in both tACS conditions ( Figure 4B; in-phase: +9 . 3%±3 . 7%; t13 = −2 . 55 , p<0 . 05; anti-phase: +31 . 7%±14 . 5%; t13 = −2 . 18 , p<0 . 05 ) , without any changes in absolute gamma-band power ( Figure 4C; in-phase: t13 = −0 . 55 , p = 0 . 59; anti-phase: t13 = 1 . 20 , p = 0 . 25 ) . Importantly , a two-way RM-ANOVA with factors session ( in- and anti-phase ) and percept ( vertical and horizontal ) calculated on absolute coherence values ( Figure 4A ) indicated that the percept ( F1 , 13 = 8 . 86 , p<0 . 05 ) related increase in gamma coherence was present in both sessions ( F1 , 13 = 52 . 69 , p<0 . 005 ) and was not inflated when absolute coherence values were lower during anti-phase stimulation ( interaction: F1 , 13 = 0 . 81 , p = 0 . 38 ) . In addition , source reconstruction revealed that gamma-band power was again confined to parieto-occipital areas and did not differ between different percepts or stimulation conditions ( all p-values >0 . 05; cluster test ) ( Figure 4I ) . Figure 4A indicated that the coherence modulation was confined to the gamma-band . Thus , we analyzed the gamma coherence change over the three conditions with cluster-based permutation statistics and found a significant difference between in- and anti-phase stimulation over stimulated parieto-occipital cortex ( cluster test: p = 0 . 006 ) ( Figure 4E ) , while no differences where present between both sham or post conditions ( cluster tests: all p>0 . 05 ) . The modulation at the channel pair of interest ( Figure 2A ) highlights the difference between the increase during in-phase stimulation ( +0 . 07±0 . 02; change to common baseline ) ( Figure 4D , dashed line ) and the decrease during anti-phase stimulation ( −0 . 03±0 . 03; effect size: d = 1 . 00 ) . This effect was not present between both sham or post conditions ( sham: d = 0 . 39; post: d = −0 . 03 ) . Interestingly , the coherence modulation was strongest for subjects with an individual gamma coherence peak frequency close to 40 Hz ( Figure 4G; cluster test: p = 0 . 049 ) . The mean gamma peak frequency did not differ between sessions ( t13 = 0 . 17 , p = 0 . 87; PeakInPhase: 51 . 14 Hz±3 . 9 Hz; PeakAntiPhase: 50 . 14 Hz±4 . 02 Hz ) . Importantly , the observed coherence effect was independent from any power changes ( Figure 4H and 4I; all p-values >0 . 2 ) . Taken together , these results indicate that 40 Hz tACS successfully modulated interhemispheric phase synchrony and behavioral outcome , depending on the chosen electrode montage and the subsequently induced phase shifts . The individual percept was influenced by ( i ) the absolute coherence values ( Figure 4A and 4F ) and ( ii ) the relative coherence difference between the vertical and the horizontal percept ( Figures 2B , 2C , and 4B ) . Entrainment would require that effects on the phase of the ongoing activity should outlast stimulation offset [23] . We found significantly modulated gamma coherence values after stimulation ( sham , in- , anti-phase stimulation; RM-ANOVA: F2 , 26 = 3 . 84 , p<0 . 05 ) ( Figure 5A ) . Specifically , we found a significant gamma-band coherence increase following in-phase stimulation ( RS3 ) ( Figure 5B ) as compared to anti-phase stimulation ( t13 = 2 . 62 , p<0 . 05; effect size: d = 0 . 67 ) . We assessed the time course of gamma-band coherence across the six resting state intervals ( time ) with in- and anti-phase stimulation ( session ) with a two-way RM-ANOVA ( Figure 5B ) , and found a significant interaction of session × time ( F3 . 13 , 40 . 65 = 3 . 08 , p<0 . 05; session: F1 , 13 = 3 . 73 , p = 0 . 08; time: F3 . 76 , 48 . 92 = 1 . 21 , p = 0 . 32 ) , indicating that coherence modulation did outlast stimulation offset . This effect was confined to the stimulated parieto-occipital cortical areas ( cluster test: p = 0 . 039 ) ( Figure 5B , inset ) . Importantly , the coherence modulation after stimulation ( RS3 ) was independent from any changes in spectral gamma power ( Figure 5D; t13 = 0 . 07 , p = 0 . 94 ) . Additionally , we found a significant correlation between coherence modulation during stimulation and the outlasting changes after stimulation ( cluster test: p = 0 . 026 ) ( Figure 5C ) . These findings imply a direct relationship between effects during and after stimulation . Outlasting effects ceased after approximately 20 minutes ( Figure 5B ) . Our results indicate that the coherence change during stimulation was positively correlated with the altered MR ( cluster test: p = 0 . 008 ) ( Figure 4F ) , suggesting that increased interhemispheric coherence sustained the horizontal percept . Interestingly , the interhemispheric coherence and behavioral performance returned to baseline values before the post session ( RS5 ) ( Figures 3A and 5B ) . Those findings demonstrate that exogenously induced transient shifts in interhemispheric coherence selectively modulate perception of ambiguous motion . We assessed the phase of the ongoing gamma-band activity during the zero crossing of the external sine wave ( every 30 cycles ) for four distinct frequency bands ( δ , α , β1/2 , γ2/3 ) . For sham and post conditions , a dummy marker was inserted to mimic the tACS trigger events . We tested whether the distribution of instantaneous gamma-band phase angles was non-uniform with Rao's spacing test ( Figure 6A ) and assessed different distributions with Kuiper's tests . Based on a binomial distribution ( p<0 . 0125; Bonferroni-corrected for four frequency bands ) , we assumed statistical significance at group level when >26 out of 28 comparisons were significant ( Kuiper tests: p<0 . 0016 Bonferroni-corrected for 31 electrodes ) . The same analysis applied to the theta/delta , alpha , and beta-bands , indicated that the effect was most prominent in the gamma frequency range and to a lesser degree still present in the beta-band ( Figure 6B ) . The continuous visual stimulus presentation restrained us from an event-related analysis such as inter-trial coherence ( ITC ) and the removal of the γ1-band impeded the analysis of a direct interaction between the ongoing activity and the externally applied sine wave . Thus , we analyzed the Shannon entropy across all spectral estimates as a surrogate marker for network dynamics and neuronal entrainment [24] . As shown previously [25] , [26] , tACS leads to more regular network dynamics and should therefore induce a subsequent entropy decrease . We found that the entropy was significantly reduced during stimulation as compared to sham over parieto-occipital electrodes ( cluster test: p = 0 . 004 ) ( Figure 6C ) . Entropy values returned to baseline during post ( cluster test: p>0 . 05 ) . We observed no differences between in- and anti-phase stimulation ( cluster test: p>0 . 05 ) . Figure 6D depicts the frequency-specific entropy decrease for both comparisons ( stimulation/sham and post/sham ) , with a pronounced decrease in the beta-/gamma-range . These results demonstrated that the entropy decrease was stimulation-specific and indicated that tACS modulated network activity across several temporal scales ( Figure 6D ) . Taken together , the biased phase relationship in gamma-band and more regular network dynamics strongly support entrainment as the putative mechanism of action of tACS [9] . The present findings demonstrate that 40 Hz tACS selectively modulates phase relationships in the gamma band ( Figures 4A , 4D , 6A , and 6B ) , without any concurrent gamma power changes ( Figure 4C , 4H , and 4I ) . In light of the entropy decrease across several temporal scales , we further evaluated their physiological interactions [27] . First , we analyzed the mean alpha power over time ( 8–12 Hz ) ( Figure 7A ) , revealing a prominent alpha decrease during the visual task as compared to fixation ( t13 = 28 . 6 , p<0 . 0005; −73 . 5%±2 . 6% ) . We assessed the average alpha power by means of a three-way RM-ANOVA and found that alpha power was significantly modulated across conditions ( sham/stimulation/post; F1 . 57 , 20 . 44 = 8 . 13 , p<0 . 005 ) ; however , no session ( in-/anti-phase ) or percept ( horizontal/vertical ) related effects were found ( session: F1 , 13 = 0 . 20 , p = 0 . 67; percept: F1 , 13 = 0 . 08 , p = 0 . 79; all interactions: p>0 . 1 ) . Secondly , we assessed the regional specificity of this power decrease and found a significant reduction over lateral parieto-occipital regions in the alpha range during tACS ( cluster test: p = 0 . 014; Figure 7B; effect size: d = 1 . 45 ) . Given the physiological antagonistic role of alpha and gamma oscillations in the parieto-occipital cortex [27] , we investigated cross-frequency interactions between alpha and gamma oscillations . Since no effects on the alpha phase coupling were observed ( Figures 4A and 6B ) , we focused on envelope interactions . We calculated correlations ( Pearson linear correlation and Fisher z-transformation ) between the alpha-amplitude and the gamma2-envelope ( Figure 7C ) . We found a significant cluster when comparing sham and stimulation conditions over lateral parieto-occipital sensors ( cluster test: p = 0 . 005; Figure 7D ) . No differences were present when sham and post conditions were compared ( cluster test: p>0 . 05 ) . In addition , we found no differences between in- and anti-phase stimulation ( cluster test: p>0 . 05 ) . These observations highlight the physiologic interaction between alpha and gamma oscillations: Entrainment of gamma oscillations promoted a secondary alpha power decrease through enhanced cross-frequency interactions as enforced by the external 40 Hz driving source . The phase-specific tACS effect on interhemispheric coherence was robust across several control analyses . First , magnitude squared coherence values can be affected by changes in amplitude correlation . Thus , we repeated all central analyses based on the phase-locking value [28] , which is independent of the amplitude of a given signal . We confirmed the increase in gamma-band phase-locking during horizontal motion perception ( 116 . 3%±5 . 6%; t13 = −2 . 93 , p<0 . 05 ) , as well as the phase-specific significant PLV modulation during stimulation ( t13 = 2 . 53 , p<0 . 05; in-phase: 0 . 08±0 . 02 , anti-phase: −0 . 01±0 . 04 ) . Importantly , the PLV modulation after stimulation offset also confirmed all previous analyses ( t13 = 3 . 14 , p<0 . 05; in-phase: 0 . 6%±0 . 3% , anti-phase: −1 . 1%±0 . 6% ) . Second , EEG studies investigating gamma-band power can be contaminated by microsaccade artifacts [29] . We therefore analyzed eye-tracking data with respect to fixation and occurrence of microsaccades [30] to exclude potential confounds of eye movements on the observed coherence modulation . We found that subjects reliably fixated during both sessions and all conditions , independent of their percept ( three-way RM-ANOVA: all factors and interactions p>0 . 05 ) ( Figure S2 ) . The influence of microsaccades was also assessed in a three-way ANOVA , indicating that the mean number of microsaccades per seconds did not vary between sessions , conditions , or percepts ( three-way RM-ANOVA: all factors and interactions p>0 . 05 ) ( Figure S3 ) .
Cortical information flow between distant regions is dynamically established by selective phase synchronization [2] , [6] and has been demonstrated for a variety of cognitive functions [3] . Especially , synchronization in the gamma-band has been suggested to constitute a fundamental mechanism for feature integration in the brain [1] and might facilitate the emergence of a stable percept of ambiguous stimuli [20] , [31] . In particular , interhemispheric gamma-band coupling might play a crucial role for feature integration across both visual hemifields [7] . However , a number of pitfalls hamper the analysis of long-range synchronization in human EEG studies . Especially , volume spread in the cortical tissue constitutes a severe constraint for the interpretation of human M/EEG data at sensor level [3] . Here , we utilized source reconstruction to highlight two distinct oscillatory gamma sources in the parieto-occipital cortex . The absence of any power differences between conditions or percepts reinforced our conclusion that interhemispheric coherence changes were a direct consequence of altered phase coupling and could not be attributed to differences in source configuration . Novel methods for coupling analysis at source level have successfully been introduced in the past [32] . However , these techniques ideally require the use of many more electrodes than employed in this study . Only relatively few electrodes have been utilized here , to avoid amplitude clipping of adjacent EEG electrodes during tACS , since stimulation currents exceed the usual recording range of EEG amplifiers by several orders of magnitude [9] . Previously , several connectivity measures have been introduced , which suppress coherent activity at 0° phase difference [33] , [34] , thus minimizing the effect of volume conduction at the expense of ignoring physiologic synchronized neuronal activity with 0° phase difference [35] . Here , we reconstructed oscillatory gamma power at source level to rule out that changes in oscillatory power may account for any of the observed coherence or phase-locking differences as observed at scalp level . We utilized sham stimulation as a baseline to control for these effects , since it is unlikely that volume conduction changes as a function of condition . Likewise , the vertical percept served as a baseline for the horizontal percept . Importantly , the directionality of behavioral and electrophysiological effects was directly related to opposite stimulation polarities . Hence , we assume that the key findings of this study were not affected by current methodological limitations . Neocortical spike activity is directly controlled by weak electric fields generated by the cortex itself [36] . Recently , it has been shown that externally applied weak electric fields may mimic endogenous fields and therefore may modulate the temporal structure of large-scale neuronal networks [14] , [25] and synchronize spiking activity [36] to different driving frequencies in a phase-specific manner [13] . Entrainment of perceptually relevant brain oscillations in humans has been demonstrated for repetitive transcranial magnetic stimulation ( rTMS ) [37] and tACS [9] , but the exact mechanisms of action are still largely unknown . The direct interaction of a cortical oscillator and a rhythmic external source by synchronization has been suggested to constitute a key mechanism for entrainment [8] . Computational models indicated that the intrinsic network frequency is ideally suited to entrain the network [25] . Nonetheless , stimulation at adjacent frequencies with higher stimulation intensities can also entrain the network sufficiently [25] , [26] and might explain the observed effects of the 40 Hz stimulation on the γ2/3-range . Importantly , the strongest coherence modulation was observed in subjects with an intrinsic gamma peak frequency close to 40 Hz , emphasizing the need for frequency-matched stimulation protocols in future studies [38] . Previously , the assessment of neuronal activity during stimulation has been hampered by the difficulty to remove stimulation artifacts in concurrent EEG recordings . Recently , a novel approach for artifact rejection has been introduced to remove 10 Hz tACS artifacts [9] . However , its applicability is limited to lower stimulation frequencies , since the jitter in the exact tACS trigger location , caused by an internal-clock-mismatch between tACS and EEG devices , is amplified during 40 Hz tACS . Therefore , a notch-filter was applied to remove the stimulation artifact , a procedure that is commonly used for line noise removal [39] . Hence , we focused the analysis during stimulation on effects in adjacent frequency bands . Importantly , immediate stimulation effects on the phase of the ongoing activity were still present after stimulation offset , thus , making a successful gamma modulation with 40 Hz tACS highly likely . Current limitations concerning the artifact removal might be overcome with synchronized tACS-EEG systems and improved artifact rejection algorithms [9] . In addition , the combination of tACS with different imaging modalities , such as fMRI [40] or magnetoencephalography ( MEG ) [41] might extend our understanding of the physiological efficacy of tACS . The modulation of long-range functional connectivity has been demonstrated for repetitive transcranial magnetic stimulation ( rTMS ) [42] and tACS [15] . However , none of the above studies presented conclusive behavioral and electrophysiological evidence for a successful modulation of perceptually relevant long-range synchronization . The study by Strüber and colleagues [18] also employed the SAM paradigm . However , the authors found a divergent pattern of results . At the behavioral level , their results suggested that only anti-phasic stimulation at 40 Hz effectively modulated the conscious experience of apparent motion . Here we replicated the behavioral key finding , i . e . , that anti-phase stimulation at 40 Hz introduces a bias to vertical motion . We extended the behavioral findings by demonstrating that a HD-tACS electrode montage may lead to a more focal in-phase stimulation ( Figure S1 ) , which biases the SAM perception to horizontal motion . In contrast , the in-phase montage used by Strüber and colleagues mainly targeted the occipital pole and resulted in no perceptual bias ( Figure S1 ) . Taken together , both studies suggest that 40 Hz tACS biases apparent motion perception , irrespective of SAM stimulus parameters or presentation ( foveal or parafoveal ) ( Table S2 ) [43] . At the electrophysiological level , Strüber and coworkers reported that anti-phase stimulation resulted in increased interhemispheric coherence after stimulation , while we found the opposite pattern of results , i . e . , in-phase stimulation enhanced synchronization , anti-phase stimulation impaired functional coupling . Strüber and colleagues had interpreted this apparent contradiction as functional decoupling , i . e . , that two signals with opposite polarities still might be highly coherent as long as the phase shift remains constant . Given the differences in study design , stimulus presentation , and tACS settings ( for a detailed overview please see Table S2 ) , we assume that the divergent patterns might result from the fact that even slight variations in stimulation intensity or electrode montage might lead to opposite network effects [44] . In addition , divergent results might also be explained by the cortical network state dependence of tACS effects [14] , [38] . These findings highlight the need for well-controlled tACS protocols , which should ideally be based on computational models and electric field predictions [25] , [45] . In this study , we based our hypothesis on phase-specific electric field predictions ( Figure S1 ) and subsequently presented evidence for the selective modulation of perceptually relevant interhemispheric gamma synchronization . Crucially , coherence effects , as induced during stimulation , outlasted the offset by approximately 20 minutes . It has been argued that the low spatial specificity might constitute a severe limitation of tACS [46] . However , this characteristic might prove beneficial for the modulation of large-scale networks . Here we utilized bilateral 4×1 ring electrode montages [17] to selectively stimulate regions of extrastriate visual cortex involved in motion perception ( Figure S1 ) [21] , [47] . In the future , multi-channel stimulators and additional modeling work [16] will hopefully allow more focal stimulation settings . Interestingly , we found that network dynamics across multiple temporal scales become more regular during external rhythmic stimulation . In particular , our results reveal that alpha power was selectively reduced after entrainment of gamma signatures . However , effects on phase coupling were mainly observed in the gamma-band . Our results support the idea that tACS effects on the phase of the ongoing activity are frequency-band specific [46] , but ancillary effects might not be constrained to a single frequency-band [11] . At present , it remains unclear whether neuronal entrainment is the only mechanism contributing to the coherence effects or whether mechanisms of neural plasticity are equally important [48] . In absence of effects on alpha phase coupling , it is unlikely that entrainment induced the transient alpha power decrease . Our results rather imply that entrainment of gamma-band activity influenced the physiological antagonistic alpha-gamma interplay in the parieto-occipital cortex [27] , [39] . Conversely , the alpha power decrease might serve as a surrogate marker for the selective entrainment of gamma band signatures , since this modulation resembles a physiologic antagonistic response to pronounced gamma band activity . A number of recent studies have demonstrated that 40 Hz tACS modulates cognitive processing , e . g . , enhances fluid intelligence [49] , induces lucidity in dreams [50] , or modulates the conscious experience of apparent motion [18] . So far , most studies utilized 40 Hz stimulation to entrain the gamma-band , even though some evidence suggested that 60 Hz tACS might be more effective than 40 Hz tACS [51] . However , results by Voss and colleagues [50] indicated that the efficacy of tACS at 25 Hz and 40 Hz might be similar . Currently , it is unclear whether there are distinct sub-bands within the gamma-band [52] and , furthermore , whether they can selectively be entrained with tACS . Previous tACS-EEG studies suggested very narrow-banded effects of tACS [9] , [50] , while our present results indicate that 40 Hz stimulation modulated the gamma-band in a broad frequency-range . Importantly , we found that the effects of 40 Hz tACS were mainly confined to the gamma-band , indicating that tACS operates within canonical frequency boundaries constituting the rhythmic brain architecture [53] . These findings imply that tACS might be a powerful tool to assess causal contributions of certain frequency-bands to distinct cognitive processes . Given that we observed clear cross-frequency interactions operating within physiologic boundaries , we urge caution when interpreting tACS effects in absence of electrophysiological recordings [54] . Interestingly , our data also indicated that subjects with a gamma coherence peak close to 40 Hz exhibited the strongest coherence modulation during 40 Hz tACS ( Figure 4G ) . Thus , previously observed effects might be contorted due to a large intersubject variability . These findings further highlight the need for a rational design of tACS protocols [55] . Recently , it has become evident that cortical dynamics across multiple spatiotemporal scales influence conscious perception [3] , [56] . Multi-stable phenomena have been related to changes in oscillatory activity in large-scale neuronal networks and might therefore reflect the periodical and constant reevaluation of sensory input [57] . Crucially , it has been shown that both , local ongoing activity [58] and interregional network activity [4] , influence subsequent perception . Bistable perception is ideally suited to study underlying network activity , since identical sensory input can lead to distinct percepts , depending on the current network state [4] . Previously , perception of the SAM has been linked to different spectral features . In particular , it has been demonstrated that frontal gamma power increments [59] together with parieto-occipital alpha power decrements [60] precede a perceptual switch . Importantly , the subjects' perceptual bias was influenced by the level of interhemispheric gamma-band coherence ( Figure 2C ) [20] and the individual percept was enforced by selectively modulating coherence levels by tACS . Frontal gamma-band increments might trigger changes in parieto-occipital networks [61] and thereby influence perception . Our data suggest that the selective entrainment of interhemispheric gamma-band synchrony might mimic the physiologic mechanism of top-down controlled percept reversals . A similar mechanism has previously been demonstrated for bottom-up processes with lower level neuron populations entraining rhythmic patterns in higher cortical areas in a feed-forward fashion [62] . Complementary coupling analyses in source space employing human M/EEG [3] will be necessary to determine whether top-down control of the SAM is associated with the selective entrainment of interhemispheric phase synchrony . Interestingly , decreased parieto-occipital alpha activity has been linked to a destabilization of the SAM [60] . In our study , we found a decrease in alpha activity in response to the entrainment of physiologic gamma-band signatures , however , without any accompanying switch rate changes ( Figure 3B ) . This observation indicates that the alpha decrease might actually reflect a secondary process in response to the gamma mediated perceptual reversals , thus , highlighting the antagonistic role of alpha and gamma oscillations [27] . Taken together , our findings strongly support the idea that neuronal interactions across different cortical regions are encoded at multiple temporal scales [56] by selective synchronization between task-relevant areas [3] . tACS studies have a seemingly endless search space: Electrode placement , stimulation frequency , intensity , and duration are obvious concerns . Here we based our hypothesis on two recent SAM studies [18] , [20] and subsequently reproduced their main findings . We documented the electrophysiological signatures of SAM perception in absence ( Figure 2A–2C ) and in presence of tACS ( Figure 4A and 4B ) . However , a number of limitations do apply . ( i ) We only applied tACS for 20 minutes , since it had previously been shown that 20 minutes of 40 Hz tACS effectively modulated apparent motion perception [18] . While 20 minutes of stimulation are well within current safety limits [63] , it is unclear how stimulation duration impacts the behavioral and electrophysiological outcome . ( ii ) Furthermore , our results suggest that stimulation at individual peak frequencies might actually be more efficient ( Figure 4G ) than using a fixed stimulation frequency . ( iii ) Another limitation is the preceding sham session . In accordance with previous results [9] , [38] , we observed outlasting stimulation effects for approximately 20–30 minutes , which impede an inverse procedure . Here we utilized a post condition to validate the sham condition and to control for outlasting behavioral and electrophysiological effects . Throughout the study we did not find any differences between sham and post conditions . Furthermore , the counter-balanced within-subject design in two sessions allowed us to study directionality effects , thus , minimizing concerns that the subjects could distinguish between real and sham stimulation . In fact , it was impossible for subjects to distinguish in- and anti-phase stimulation . ( iv ) Another clear restriction is the unbalanced electric field distribution ( Figure S1 ) . While our in-phase stimulation was very focal , the anti-phase montage led to a more distributed electric field . However , in contrast to the study by Strüber and colleagues [18] , both montages targeted the extrastriate visual cortex and modulated the behavioral outcome . We believe that multi-channel stimulators and optimized electric field models will improve the focality of tACS in the future . We expect that a rational design of tACS experiments using frequency-matched and neuro-navigated protocols will improve the efficacy in human tACS studies [55] . So far , the basic physiological principles behind the efficacy of tACS are still largely unknown [46] . Further work in human , animal , and modeling studies will hopefully advance our understanding of tACS and its interactions with neuronal circuits . In the future , complementary modeling approaches may guide individually tailored stimulation protocols and electrode features to overcome current limitations , such as the unclear electric field distribution in the head , the cortical state dependence , and the high intersubject variability .
In summary , our results demonstrate that interhemispheric gamma-band coherence can be selectively modulated by tACS . In this study , we established a causal role of synchronized gamma band oscillations for feature integration across both hemispheres and confirmed the antagonistic role of alpha and gamma oscillations in the parieto-occipital cortex [27] . Our results demonstrate the ability of tACS to selectively entrain cortical oscillations and add to a growing body of evidence indicating that synchronized oscillatory activity in large-scale neuronal networks is a key mechanism for conscious perception and cognition [3] , [56] . Disturbances of synchronized network activity have previously been related to schizophrenia , autism spectrum disorders ( ASDs ) , and Parkinson's disease [64] . In particular , ASD have been associated with impaired feature integration across both hemispheres [65] . Future research might therefore offer the possibility to individually tailor therapeutic interventions by means of non-invasive brain stimulation [66] . In particular , the frequency specificity of tACS makes it an ideal candidate for treatment of rhythmic cortical disturbances , as recently demonstrated for tremor suppression in patients with Parkinson's disease [10] .
In accordance with previous studies [18] , [20] , 14 healthy volunteers ( eight females , six males , mean age: 27 . 5±6 . 7 years ) were recruited from the University Medical Center in Hamburg , Germany , including two of the authors ( RFH , HK ) . All participants ( including participating authors ) were blinded towards the stimulation sequence . All subjects were right-handed , reported no history of neurological or psychiatric disease , and were medication-free during the experiments . They all had normal or corrected-to-normal vision . All participants gave written informed consent according to the local ethics committee and the Declaration of Helsinki . This study has been approved by the local Ethics Committee of the Medical Association in Hamburg , Germany ( IRB number: PV4335 ) . The SAM ( Figure 1A ) was generated with the Psychophysics Toolbox [67] implemented in MatLab ( The MathWorks Inc . ) and presented on a BenQ XL2420T screen ( 1 , 920×1 , 080 , 120 Hz ) . The display was 60 cm away from the participants . The horizontal dot distance was 5 . 1° , the vertical 6 . 9° at a constant dot size of 0 . 35° . We introduced a shorter horizontal than vertical distance to compensate the vertical bias of equidistant SAMs [19] . Every trial of continuous visual stimulation lasted 1 minute . Participants reported their percept by pressing two different buttons with their right hand . All volunteers participated in two sessions of the experiment carried out within 1 week . After preparation of the tACS and EEG electrodes ( see below ) , participants completed a training session with ambiguous and non-ambiguous trials to familiarize volunteers with the stimulus and to ensure that all participants could reliably track their percepts . On both days all participants completed ten trials during sham , 20 trials during stimulation ( in- or anti-phasic tACS ) , and ten trials during the post condition . Whether participants were stimulated with in- or anti-phase stimulation on the first day was counterbalanced across subjects ( Figure 1B , the participating experimenters were blinded whether they received in- or anti-phasic stimulation ) . A sham condition always preceded electrical stimulation to avoid carry-over effects [38] . Altogether , six resting state epochs , 3 minutes each , were recorded during stable fixation of a central dot to assess outlasting changes . | Brain activity is profoundly rhythmic and exhibits seemingly random fluctuations across a very broad frequency range ( <0 . 1 Hz to >600 Hz ) . Recently , it has become evident that these brain rhythms are not just a generic sign of the brain-at-work , but actually reflect a highly flexible mechanism for information encoding and transfer . In particular , it has been suggested that oscillatory synchronization between different areas of the cortex underlies the establishment of task-relevant networks . Here , we investigated whether gamma-band synchronization ( ∼40 Hz ) is causally involved in the integration between the two brain hemispheres of alternating visual tokens into a coherent motion percept . We utilized transcranial alternating current stimulation ( tACS ) , a novel non-invasive brain stimulation technique , which allows frequency-specific entrainment of cortical areas . In a combined tACS-electroencephalography study , we selectively up- and down-regulated interhemispheric coherence , resulting in a directed bias in apparent motion perception: Increased interhemispheric connectivity sustained the horizontal motion percept , while decreased connectivity reinforced the vertical percept . Thus , our data suggest that the level of interhemispheric gamma-band coherence directly influenced the instantaneous motion percept . From these results , we conclude that synchronized neuronal activity is essential for conscious perception and cognition . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Conclusions",
"Material",
"and",
"Methods"
] | [
"systems",
"biology",
"computer",
"and",
"information",
"sciences",
"psychology",
"medicine",
"and",
"health",
"sciences",
"radiology",
"and",
"imaging",
"neural",
"networks",
"network",
"analysis",
"physiology",
"neurology",
"biology",
"and",
"life",
"sciences",
"neu... | 2014 | Selective Modulation of Interhemispheric Functional Connectivity by HD-tACS Shapes Perception |
While conventional LDL-C , HDL-C , and triglyceride measurements reflect aggregate properties of plasma lipoprotein fractions , NMR-based measurements more accurately reflect lipoprotein particle concentrations according to class ( LDL , HDL , and VLDL ) and particle size ( small , medium , and large ) . The concentrations of these lipoprotein sub-fractions may be related to risk of cardiovascular disease and related metabolic disorders . We performed a genome-wide association study of 17 lipoprotein measures determined by NMR together with LDL-C , HDL-C , triglycerides , ApoA1 , and ApoB in 17 , 296 women from the Women's Genome Health Study ( WGHS ) . Among 36 loci with genome-wide significance ( P<5×10−8 ) in primary and secondary analysis , ten ( PCCB/STAG1 ( 3q22 . 3 ) , GMPR/MYLIP ( 6p22 . 3 ) , BTNL2 ( 6p21 . 32 ) , KLF14 ( 7q32 . 2 ) , 8p23 . 1 , JMJD1C ( 10q21 . 3 ) , SBF2 ( 11p15 . 4 ) , 12q23 . 2 , CCDC92/DNAH10/ZNF664 ( 12q24 . 31 . B ) , and WIPI1 ( 17q24 . 2 ) ) have not been reported in prior genome-wide association studies for plasma lipid concentration . Associations with mean lipoprotein particle size but not cholesterol content were found for LDL at four loci ( 7q11 . 23 , LPL ( 8p21 . 3 ) , 12q24 . 31 . B , and LIPG ( 18q21 . 1 ) ) and for HDL at one locus ( GCKR ( 2p23 . 3 ) ) . In addition , genetic determinants of total IDL and total VLDL concentration were found at many loci , most strongly at LIPC ( 15q22 . 1 ) and APOC-APOE complex ( 19q13 . 32 ) , respectively . Associations at seven more loci previously known for effects on conventional plasma lipid measures reveal additional genetic influences on lipoprotein profiles and bring the total number of loci to 43 . Thus , genome-wide associations identified novel loci involved with lipoprotein metabolism—including loci that affect the NMR-based measures of concentration or size of LDL , HDL , and VLDL particles—all characteristics of lipoprotein profiles that may impact disease risk but are not available by conventional assay .
Standard measures of plasma lipoprotein concentration do not reveal heterogeneity in the size of lipoprotein particles or their content of cholesterol and triglycerides . Yet recognizing this heterogeneity may be essential for understanding qualitative differences in lipid metabolism among individuals . Some reports identify a pattern in the size distribution of lipoprotein sub-fractions as intimately connected with coronary heart disease [1] , [2] . Related findings identify a link between lipoprotein profile and metabolic syndrome , and by inference to diabetes [3] . While these observations remain controversial for prognostic use [4] , they point to alterations in lipoprotein metabolism in disease . The variation in particle size and lipid content can be quantified accurately by NMR-based methods that determine lipoprotein particle concentration according to lipid class and particle size . Thus , NMR methods can measure concentration of large and small low density lipoprotein ( LDL ) particles as well as concentration of the related intermediate density lipoprotein ( IDL ) particles , and similarly concentration of small , medium , and large high density lipoprotein ( HDL ) or very low density lipoprotein ( VLDL ) particles . HDL and LDL particle concentration can also be estimated by chemical measures of apolipoprotein A1 ( ApoA1 ) and apolipoprotein B ( ApoB ) protein concentration , respectively , but neither these assays nor other standard clinical assays provide information about particle size distribution , and consequently the apportionment of cholesterol and triglycerides to different sized particles . The greater precision in characterizing lipoprotein profiles using NMR-based techniques provides an opportunity for correspondingly greater detail in understanding lipid metabolism , for example by genome-wide genetic analysis , as has been done recently for plasma concentration LDL-C , HDL-C , triglycerides , ApoA1 , and ApoB [5]–[13] .
Among 17 , 296 WGHS participants with confirmed European ancestry ( Table 1 ) , we performed genome-wide association analysis assuming an additive genetic model for 22 plasma lipoprotein measures determined either by NMR methods or by standard clinical assay . On the basis of genome-wide significance ( P<5×10−8 ) , genetic variation at total of 31 loci was associated with at least one of the lipoprotein fractions ( Table 2 ) . Thirty of these 31 loci derive from analysis in the whole sample , while the remaining locus was identified with genome-wide significance in a subset of 12 , 489 ( 72% ) strictly fasting participants , for whom there were small but significant differences in lipoprotein profiles compared with non-fasting participants ( Table 1 ) . Nearly all of the associations with genome-wide significance level in the fasting subsample also had genome-wide significance in the larger , better powered whole sample . One exception was the genome-wide significant association with ApoA1 at ABCA1 ( 9q31 . 1 ) , a locus that was identified in the whole sample on the basis of genome-wide significant associations with HDL-C and medium HDL particles but not for ApoA1 . The other was an association with mean VLDL size at 8p23 . 1 , a locus that appears only in analysis in the fasting sub-sample ( Table 2 ) . These additional associations remain strongly suggestive in the whole sample ( P<1 . 6×10−5 ) even though they do not reach genome-wide significance . Statistics for the most significant genome-wide associations with P<5×10−8 at each of the candidate loci are shown in the Table S1 . Seven of the 31 unique loci reveal novel genome-wide significant associations with the plasma lipoprotein fractions ( see bold font type , Table 2 ) . The associations at 3q22 . 3 ( PCCB/STAG1 ) , 6p21 . 32 ( BTNL2 ) , 7q32 . 2 ( KLF14 ) , 12q24 . 31 . B ( CCDC9/DNAH10/ZNF664 ) and 17q24 . 2 ( WIPI1 ) are all near genes ( Figure S1 ) , while genome-wide significant associations at the remaining two novel loci , 8p23 . 1 and 12q23 . 2 are remote ( i . e . >150kb ) from known genic regions . Among the standard clinical measures LDL-C , HDL-C , and triglycerides only , novel genome-wide loci were found at KLF14 ( 7q32 . 2 ) and CCDC9/DNAH10/ZNF664 ( 12q24 . 31 . B ) , both for triglycerides . The association at the novel locus 8p23 . 1 ( which differentiated the fasting sample from the whole sample on the basis of mean VLDL particle size ) is over 1 . 8 Mb from a recently described association at 8p23 . 1 between SNP rs7819412 and triglycerides [6] . The remaining 24 unique loci suggested genes recognized for a diversity of roles in lipid metabolism , broadly defined ( Figure S1 ) . Thus , SNPs with genome-wide significance , were confirmed in or near PCSK9 ( at 1p32 . 3 ) , APOA2 ( 1q23 . 3 ) , APOB ( 2p24 . 1 ) , ABCG5/8 ( 2p21 ) , HMGCR ( 5q13 . 3 ) , LPL ( 8p21 . 3 ) , APOA1-A5 ( 11q23 . 3 ) , ABCA1 ( 9q31 . 1 ) , FADS1-3 ( 11q12 . 2 ) , LIPC ( 15q22 . 1 ) , CETP ( 16q13 ) , LIPG ( 18q21 . 1 ) , LDLR ( 19p13 . 2 ) , the APOC-APOE complex ( 19q13 . 32 ) , and PLTP ( 20q13 . 12 ) . Similarly , association at 9q34 . 2 implicating the ABO gene recapitulates and extends the known association between blood group antigen and total cholesterol [14] , [15] . Less well characterized genic regions , which nonetheless have been validated recently for roles in lipid metabolism , were confirmed for ANGPTL3 ( 1p31 . 3 ) , CELSR2/MYBPHL/PSRC1/SORT1 ( 1p13 . 3 ) , GCKR ( 2p23 . 3 ) , MLXIPL ( 7q11 . 23 ) , and TRIB1 ( 8q24 . 13 ) , HNF1A ( 12q24 . 31 . A ) , and HNF4A ( 20q13 . 12 ) . The association at COBLL1/GRB14 ( 2q24 . 3 ) with HDL-C was recently described elsewhere in this same cohort and validated by replication [16] . The previous study found much stronger association in women than men , suggesting a potential interaction with gender . At this locus , the gene GRB14 is thought to inhibit receptors in the insulin receptor class [17] , [18] . The current analysis extends associations at this locus to concentrations of LDL , HDL , and VLDL particles according to size ( Table S1 ) . Consistent with a high degree of correlation among the lipoprotein measures ( Table S2 ) , the rank order by p-value among the highly significant SNPs was similar for each measure with at least one genome-wide significant association ( Figure S1 ) . A notable exception was the APOB gene ( 2p24 . 1 ) , where the ordering of the p-values , conditional analysis , and patterns of linkage disequlibrium ( LD ) among the top SNPs ( Table S1 ) revealed three classes of associations . One class included VLDL-related fractions , triglycerides , and mean LDL size for which either rs673548 or rs676210 ( LD r2 = 1 . 0 ) had the strongest association; a second class included ApoB , large LDL particles , and total LDL particles for which either rs1713222 or rs506585 ( LD r2 = 0 . 5 ) had the strongest association; and a final class including only LDL-C for which rs137117 was most strongly associated ( Figure 1A ) . Between SNPs in different classes , maximum LD ranged from r2 = 0 . 04–0 . 11 . Similarly , at APOA5-APOA1 ( 11q23 . 3 ) , p-values revealed two classes of associations seemingly segregating between effects nearer the APOA5 gene involving triglycerides and effects nearer the APOA1 gene involving HDL related lipoprotein fractions ( Figure 1B ) . Large , well-characterized cohorts with NMR-based measurement of lipoprotein fractions are scant , but sub-samples of about 2700 participants in the Framingham Heart Study Offspring cohort ( FHS ) [19] and about 2000 total CHD cases and controls from PROCARDIS [20] had both the NMR-based lipoprotein measures and genome-wide genetic data already determined . Among all candidate loci , concordance of direction of effects was observed respectively at 124 out of 146 ( 84% ) [84% in fasting sub-sample] and 125 out of 133 ( 94% ) [99% in fasting sub-sample] of the candidate associations for which there was genotype information in FHS and PROCARDIS ( Table S3 [whole WGHS sample candidates] , Table S4 [fasting WGHS subsample candidates] ) . For each of the previously known loci except ABCA1 ( 9q31 ) , at least one of the candidate associations was nominally significant ( P<0 . 05 , two-sided ) in at least one of the replication cohorts or in analysis combining p-values from the two replication cohorts when effect estimates ( beta coefficients ) indicated trends in lipoprotein measure consistent with the effects observed in the WGHS . Among the 7 novel loci from the primary analysis only , where the effect estimates for the WGHS were generally smaller and power for replication was less , concordance of the direction of effects remained high for the PROCARDIS sample [86% ( 25/29 ) ] , although only modest for the FHS sample [58% ( 22/38 ) ] , but these associations were not significant ( two-sided P>0 . 05; Table S3 ) . However , a recent genome-wide meta-analysis of LDL-C , HDL-C , and triglycerides found significant , but not genome-wide significant , associations among these fractions with candidate SNPs from the WGHS at PCCB/STAG1 ( 3q22 . 3 ) , BTNL2 ( 6p21 . 32 ) , KLF14 ( 7q32 . 2 ) , and 8p23 . 1 [10] , although the significant SNP associations at PCCB/STAG1 ( 3q22 . 3 ) and BTNL2 ( 6p21 . 32 ) were not fully concordant between the two studies ( Table 3 ) . Independent evidence for functional consequence of the candidate SNP ( rs10778213 ) at 12q23 . 2 is its genome-wide significant association in a smaller sample from the WGHS with plasma C-reactive protein ( CRP ) , a biomarker of inflammation that is slightly correlated if at all with the two HDL measures associated at this locus ( total HDL particle concentration [HDL:T] , Spearman r = 0 . 22; HDL cholesterol estimated by the NMR [HDL:N] , Spearman r = −0 . 04 ) [21] . With the larger sample of WGHS genotype information in the current study , the association with plasma CRP is more significant ( P<5×10−15 ) . Finally , the associations at CCDC92/DNAH10/ZNF664 [12q24 . 31 . B] and WIPI1 ( 17q24 . 2 ) were not confirmed either in the meta-analysis shown in Table 3 or in a second genome-wide meta-analysis of LDL-C , HDL-C , and triglycerides that also evaluated gender stratified association [11] ( data not shown ) . Nevertheless , ongoing genotyping in the WGHS of an additional 4639 samples ( 3305 with fasting status ) completed subsequent to the main analysis provided significant support for these last two loci on the basis of internal replication , as well as significant or borderline significant support for four others , confirming directions of effects for all novel candidate associations , and leading to smaller p-values in analysis combining the main WGHS sample with the additional samples for all but three entries in Table 3 and at least one lipoprotein measure for each locus ( compare to Table S1 ) . To assess the contribution of common genetic variation at each of the candidate loci to each of the adjusted lipoprotein fractions , we constructed regression models by stepwise selection of SNPs in the vicinity of the primary genome-wide significant associations . Most of these models explain less than 1% of the variation in the adjusted lipoprotein fractions ( Figure 2 , Table S5 , and Table S6 ) . The top three effects , all at APOC-APOE complex ( 19q13 . 32 ) , explain 8 . 9% , 8 . 4% , and 7 . 1% of the variance in ApoB particle concentration , the related total LDL particle concentration , and LDL-C , respectively . Fasting status had an influence on retention of SNPs in the model selection procedure , but only for loci with modest effects ( Compare Table S5 and Table S6 ) . There were no genetic contributions remaining from the model selection procedure for any of LDL-C , HDL-C , triglycerides , ApoA1 , or ApoB concentration at APOA2 ( 1q23 . 3 ) in the whole sample and at WIPI1 ( 17q24 . 2 ) in the fasting subsample , suggesting that these loci would not have been identified for genome-wide association with the five conventional lipoprotein fractions even in a much larger sample with the genome-wide SNP genotyping panel used in this study . Clustering loci on the basis of the profile of associated lipoprotein fractions suggests sub-groups of loci with related patterns of effects ( Figure S2 , Figure S3 ) , perhaps suggesting distinct but possibly overlapping biological pathways for lipoprotein metabolism . For example , HNF1A , LDLR , ABCG5/8 , PCSK9 , and CELSR2/PSRC1/SARS/SORT1 largely share associations with IDL , small VLDL , total VLDL large LDL , LDL-C , total LDL , and ApoB . The total genetic effects for each lipoprotein determined by summing over the effects at all loci ranged from 2 . 1% for mean VLDL size to 17 . 2% for ApoB ( Table 4 ) . The effects were not substantially different when the entire model selection procedure was performed in the fasting subsample ( Table 4 ) , and only slightly smaller in general among the unadjusted lipoprotein fractions ( Table S7 ) . Notably , the common genetic variation in this study at the genome-wide loci had a greater total effect on mean particle size than on standard clinical cholesterol measures for HDL but not for LDL or VLDL ( Table 4 ) . To examine the possibility that other loci might include SNPs with genome-wide significant association conditional on effects at the primary loci , we adjusted the primary lipoprotein fraction measurements ( which were already adjusted for clinical covariates ) for SNPs retained by the model selection procedure at the candidate loci , and repeated the genome-wide association testing . Quantile-quantile analysis confirmed that all of the excess of extremely small p-values in the original analysis could be explained by the variation at the candidate loci ( not shown ) . Similarly , genotype-based statistical models ( as opposed to the allele-based additive models used in the primary analysis ) did not reveal other loci with genetic influences at the genome-wide significance level in the whole sample . While we adjusted the lipoprotein measures with a full set of clinical characteristics to reduce variance and enhance power in the primary analysis , it remained possible that relevant SNPs would be overlooked if they acted through effects on the adjustment covariates . Similarly , subtle effects on the association estimates due to non-normality of the ( possibly log-transformed ) adjusted lipoprotein measures or due sub-European population stratification might confound hypothesis testing . To evaluate whether our discovery procedure was robust , we performed secondary analyses repeating the entire genome-wide discovery procedure for alternative nested subsets of clinical covariates with and without further adjustment for population structure and quantile normalization ( Table S8 ) . Comparing the full adjustment procedure to alternatives using either a reduced set of clinical covariates or age only , with or without additional adjustment for potential sub-European population stratification and quantile normalization yielded further genome-wide significant associations at three loci with known lipid metabolic genes , LPA ( 6q25 . 3 ) , LCAT ( 16q22 . 1 ) , and APOH ( 17q24 . 2 ) , and two additional loci , 6p22 . 3 and 10q21 . 3 . All of the additional loci were present in the age-adjusted analysis . Associations at 6p22 . 3 and 10q21 . 3 appear to be novel and implicate , respectively the GMPR or MYLIP genes and the JMJD1C gene . The lead SNPs at each of these loci were significantly associated with at least one of LDL-C , HDL-C or triglycerides in the recently published meta-analysis ( Table 5 ) [10] . Similarly , in internal replication among the additional 4639 WGHS samples with genotype available after the main analysis was complete , associations at the candidate SNPs were all significant and the trends of effects were all consistent with effects in the discovery sample ( Table 5 ) . We note that at JMJD1C ( 10q21 . 3 ) , the candidate SNPs have minor allele frequency near 0 . 5 , and that available data does not allow us to determine whether the differences in the direction of the minor allele effect on VLDL fractions in the WGHS and triglycerides in the previously published replication study are truly physiological or rather that the frequency of the coded ( i . e . minor ) allele from the WGHS is greater than 0 . 5 in the replication cohort resulting in an opposite sign of the effect estimates . Since lipoprotein particle size is closely related to triglyceride content , we also performed secondary analysis examining genome-wide significant associations after adjustment of the lipoprotein fractions by the full set of clinical covariates and ( log-transformed ) triglyceride levels ( Table 5 and Table S8 ) . This analysis identified only one new genome-wide significant association . At 11p15 . 4 , rs7938647 in the intron of the SBF2 gene was associated with full-plus-triglyceride adjusted total HDL particle concentration . Again , internal replication provided support for this association although there was no association ( P>0 . 05 ) with LDL-C , HDL-C , or triglycerides in the recent meta-analysis for replication . Among its unique characteristics , the NMR-based methodology provides information about IDL and VLDL particle concentration , both aspects of lipoprotein profiles that are difficult to measure by conventional methods . For IDL , genetic associations were observed at many of the candidate loci ( Figure 2 , Table 2 , Table S1 ) and most strongly at LIPC ( 15q22 . 1 ) , where rs1532085 had an estimated 0 . 11 nmol/l shift in particle concentration for each copy of the minor allele ( p = 1 . 5×10−20 ) . For total VLDL concentration , association with genetic variation was observed at many loci but none more strongly than at the APOC-APOE complex where rs439401 , which is in perfect LD with rs7412 ( the SNP that distinguishes APOE alleles E2 and E3 ) , had an estimated −2 . 4nmol/l shift in concentration per copy of the minor allele ( p = 2 . 1×10−12; Table S1 ) . Loci strongly affecting the relative concentration of NMR-based estimates of small , medium , and large particle size could be identified on the basis of genome-wide effects on mean particle size , and these associations were of special interest when there was no accompanying association with the corresponding cholesterol measure retained in the model selection procedures ( Table 6 , Figure S4 ) . For LDL , mean particle size was associated with genome-wide significance at 12 loci ( Table 2 ) , among which the model selection procedures failed to identify any association with LDL-C at MLXIPL ( 7q11 . 23 ) , LPL ( 8p21 . 3 ) , CCDC92/DNAH10/ZNF664 ( 12q24 . 31 . B ) , and LIPG ( 18q21 . 1 ) . These loci implicate genes related to glucose or triglyceride metabolism as well as unrecognized biological function at one novel locus ( CCDC92/DNAH10/ZNF664 [12q24 . 31 . B] ) . The associations with mean LDL particle size were a consequence of strong inverse effects on large and small LDL particles ( MLXIPL [7q11 . 23] , LPL [8p21 . 3] , LIPG [18q21 . 1] ) or of exclusive effects on small LDL ( CCDC92/DNAH10/ZNF664 [12q24 . 31 . B] ) [see Figure S4] . In the fasting subsample , the associations with the NMR based measures at LPL ( 8p21 . 3 ) and LIPG ( 18q21 . 1 ) also met genome-wide significance , but the associations at MLXIPL ( 7q11 . 23 ) and CCDC92/DNAH10/ZNF664 ( 12q24 . 31 . B ) did not . For HDL , 9 loci had genome-wide significance for mean particle size ( Table 2 ) , among which the clinical measure of HDL-C was not associated with genetic variation only at GCKR ( 2p23 . 3 ) , as was also found in the fasting subsample ( Figure 2 , Table 6 ) . The discordant effects on LDL size and cholesterol content at LPL ( 8p21 . 3 ) , CCDC92/DNAH10/ZNF664 ( 12q24 . 31 . B ) , and LIPG ( 18q21 . 1 ) but not those of HDL size and cholesterol content were independent of triglyceride level in as much as associations persisted in analysis that further adjusted the lipoprotein fractions for ( log-transformed ) triglycerides , although only at nominal significance rather than genome-wide significance ( Table 6 ) . By the same standards , loci could be identified with effects on mean particle size but not total particle concentration ( Table 6 ) . Thus , SNPs at LIPC ( 15q22 . 1 ) and LIPG ( 18q21 . 1 ) had genome-wide significant associations for mean LDL particle size , but were null for particle concentration in model selection procedures in both the whole sample and the fasting subsample . These loci are characterized by genes known to influence triglyceride metabolism . Similarly , for HDL , comparison of associations with mean particle size and total particle concentration identified variation at KLF14 ( 7q32 . 2 ) , FADS1-3 ( 11q12 . 2 ) , CCDC92/DNAH10/ZNF664 ( 12q24 . 31 . B ) and LIPC ( 15q22 . 1 ) , implicating roles for known lipid candidate genes as well as loci with unknown functions . Variation at the novel locus WIPI1 ( 17q24 . 2 ) , while not affecting mean HDL particle size , was associated with the concentration of medium-sized HDL , but not large or small HDL , total HDL particle concentration , or HDL-C ( Table 2 , Figure 2 , Figure S4 ) . In addition , associations at LPL ( 8p21 . 3 ) in the fasting subsample distinguished total HDL particle concentration from HDL-C ( Table 6 ) . VLDL particle size but not concentration was influenced by variation at 8p23 . 1 in the fasting subsample but there were no genome-wide significant associations at this locus in the whole sample . Again , in triglyceride-adjusted analysis , discordant effects on mean particle size and total concentration persisted but at some of the candidate loci in the analysis of LDL and HDL ( Table 6 ) . Recent genome-wide meta-analysis of lipoprotein LDL-C , HDL-C , and triglycerides identified and validated 17 loci that were not found at the genome-wide significance level in the current population [10]–[12] in spite of comparable statistical power . We examined SNPs within 100kb of each of these additional candidate loci to extend associations to each of the NMR-based lipoprotein fractions . The choice of a threshold p-value for significance is a controversial issue in these analyses: although all of the candidate loci had been validated previously , the current analysis was performed in the context of a genome-wide association study . We present all locus associations when statistical significance of Bonferroni corrected p-value for the most significant locus association was less than 0 . 05 , accounting for the product of the number of lipoprotein fractions tested ( 22 ) and the number of locus SNPs considered ( range 8–125 ) ( Table S9 and Table S10 ) . Seven loci ( TMEM57 , GALNT2 , TIMD4/HAVCR1 , MADD/FOLH1/NR1H3 , MVK/MMAB , LCAT , CLIP2/PBX4/NCAN/SF4 ) met this criterion in the whole WGHS sample and , at the same standard , one more locus ( MAFB ) could be added in the fasting subsample . Among these loci , associations with lipoprotein size measures were found for LDL at GALNT2 , and for HDL at GALNT2 , MADD/FOLH1/NR1H3 , MVK/MMAB , CLIP2/PBX4/NCAN/SF4 . No associations at the stringent significance level were found with mean VLDL size or total IDL concentration . Associations with HDL and LDL total particle concentration were largely consistent with parallel associations with ApoA1 and ApoB respectively .
By performing genome-wide association analysis among 17 , 296 Women with European ancestry for 22 NMR-based and conventional lipoprotein fractions , we identified 36 loci in the primary and secondary analyses for roles in lipoprotein metabolism , broadly defined . Ten of these loci have not been reported in other recent genome-wide association studies , including one identified only after adjustment for triglyceride levels . The functional bases for the associations are uncertain for five , including associations at 8p32 . 1 and 12q23 . 2 that map to intergenic regions . In spite of the high degree of correlation among some of the NMR-based and conventional measures , two of the novel loci ( PCCB [3q22 . 3] and PPP1R3B [8p23 . 1] ) could not have been found at the genome-wide significance standard solely with conventional measures ( or their NMR-based equivalents ) of lipoprotein profile in the WGHS . Replication in independent cohorts of men and women as well as other observations provided confirmatory evidence for candidate variation at all novel , although only through internal replication at SBF2 ( 11p15 . 4 ) , CCDC92/DNAH10/ZNF664 ( 12q24 . 31 . B ) , and WIPI1 ( 17q24 . 2 ) . The failure of external replication to validate these two novel loci may simply reflect intrinsic differences from the WGHS in NMR-based assay protocols ( FHS ) or clinical features of the cohort ( e . g . lipid lowering treatment in PROCARDIS ) as well as limiting power; alternatively , the associations observed in the WGHS may not reflect true genetic effects . Among the primary loci , total genetic effects were largest and appreciable for ApoB , total LDL , and others . They were the smallest for mean VLDL size . While the heritability for the NMR-based fractions has not been thoroughly explored , the present analysis suggests some aspects of lipoprotein profiles may be much less affected by common genetic variation than others . Combining the 31 loci in the primary analysis , the five loci in the secondary analysis ( three novel loci plus APOH and LCAT ) , and the seven previously recognized loci for which the WGHS extends associations to the NMR-based lipoprotein measures brings the total to 43 loci characterized by the present study . As important as the total number of candidate loci , some loci harbored variation exclusively correlated with the size of lipoprotein particles rather their cholesterol or total concentration ( Table 6 ) . A priori , one might have argued that triglyceride metabolic processes would be critical in this respect . This notion was confirmed by several candidate genes with known function in triglyceride metabolism , for example the enzymes encoded by LPL , LIPC , LIPG , and GCKR as well as the transcriptional regulatory protein encoded MLXIPL all have activities that may alter equilibrium pools of triglycerides and hence particle size or concentration . Other loci with only partly understood function were also identified , and these loci may now be further characterized through the current analysis . While it remains possible that the loci in Table 6 contain genetic variants not evaluated in this study and yet associated with cholesterol content or total particle concentration , the discordant effects on particle size compared with cholesterol or total particle concentration suggest biochemical pathways impinging on aspects of lipoprotein metabolism that are overlooked by standard clinical testing . To the extent that the pathophysiology of cardiovascular disease and related metabolic disorders , e . g . diabetes , is influenced by the distribution of lipoprotein particle size there may be therapeutic opportunities targeting the biochemical pathways identified by the discordant associations . The procedures in the primary analysis enforced a genome-wide significance standard of P<5×10−8 for each lipoprotein measure . This standard was likely adequate for performing separate tests in the whole sample and the fasting subsample ( see Materials and Methods ) but does not explicitly address the multiplicity of testing the 22 lipoprotein measures at once . In part , the burden of significance is attenuated by correlations between the lipoprotein measures ( Table S2 ) , but the correlations are not exact and independent aspects of each measure are revealed by the diversity of effects shown in Figure 2 as well as by the discordant associations of Table 6 . However , the choice of P<5×10−8 for genome-wide significance can be further justified by false discovery rate ( FDR ) analysis . For p-values from all of the lipoprotein measures considered at once , the conventional standard requiring FDR<0 . 05 implied a P<2×10−5 , more than two order of magnitude less significant than the genome-wide p-value threshold . Similarly , among the individual lipoprotein measures , FDR<0 . 05 implied at worst P<7×10−7 for the case of IDL , still less significant than our genome-wide standard by over an order of magnitude . Thus , on a post-hoc basis , applying the conventional genome-wide standard P<5×10−8 for all fractions appears to have been justified . Four of the 10 novel loci ( 7 from primary analysis , 3 from secondary analysis ) have functional links to lipoprotein metabolism or disease status , even if strict biochemical roles of the candidate genes and protein are not yet known . Variation at BTNL2 ( 6p21 . 32 ) has been associated with Grave's disease , multiple sclerosis , and sarcoidosis , apparently independent of the neighboring HLA class DR genes [22]–[24] . In addition , the lipoprotein association at this locus is within 780kb of a recently reported association of rs2254387 with LDL-C attributed to the B3GALT4 gene encoding a galactosyltransferase [6] . At STAG1/PCCB ( 3q22 . 3 ) , the genome-wide significant association with small HDL particle concentration is in the STAG1 gene , but a more likely candidate for lipid metabolism may be the adjacent PCCB gene encoding the propionyl coenzyme A carboxylase beta subunit , in which substitutions cause Mendelian forms of proprionic acidemia ( see , for example [25] ) . At 8p23 . 1 , over 150kb from the candidate SNP rs983309 , PPP1R3B encodes a phosphatase regulating glycogen phosphorylase , a plausible regulator of glucose and triglycerides . At 17q24 . 3 , the connection to lipid metabolism can be made through an encoded domain of WIPI1 protein , the WD40 domain , which is a structural motif thought to interact with phospholipids [26] . The strongest association at this locus is over 2Mb away but statistically independent from the associations of rs1801689 with full-plus-triglyceride-adjusted total LDL particle concentration or rs2909207 with age-adjusted medium HDL particles ( Table 5 ) , both adjacent to the lipid candidate gene APOH [27] . The remaining six loci have intergenic status , or are proximal to genes with unresolved connections to lipoprotein metabolism . Nevertheless , association at one of these six loci , 12q23 . 2 , between rs7307277 and HDL-C measured by NMR involves the same SNP we previously reported for genome-wide significant association with plasma C-Reactive Protein ( CRP ) in a subset of the current population [21] , an association that remains highly significant in the current sample ( P = 4 . 5×10−15 ) . Previous reports , including our own , had also identified associations at GCKR , APOC-APOE complex , and HNF1A with both lipid fractions and CRP [21] . We could now also add HNF4A to this list since rs4810479 at 20q13 . 12 . A is associated in the WGHS with both CRP and the lipoprotein fractions ( Table 2 , Table S1 ) . These links between lipoprotein metabolism and CRP are particularly intriguing given the efficacy of lipid lowering therapy with statins among individuals identified as at risk on the basis of elevated CRP [28] . The etiology of cardiovascular disease is complex , and is believed to include an interplay between cell-based processes , including inflammation , and blood components , including lipoprotein fractions . The latter aspect may be summarized by clinical measures of cholesterol or triglycerides , or by ApoA1 and ApoB concentration . However , none of these aggregate measures reflects the full diversity of lipoprotein species in blood . The current investigation not only identifies novel loci for lipid metabolism in general , but may also help delineate the impact of lipoprotein metabolic genes on lipoprotein profile viewed with the highest resolution currently available .
All analyses were performed with approval of local institutional review boards ( IRBs ) . All samples in the discovery analysis derive from the Women's Genome Health Study ( WGHS ) , a prospective cohort of North American women with phenotypes related to cardiovascular disease , extensive clinical and demographic data , blood samples at baseline , and ongoing genome-wide genotyping [29] . The current data derive from 17 , 296 WGHS participants with confirmed , self-reported European ancestry who were non-diabetic , not using lipid lowering therapy at baseline , and for whom genotype information was available . Within this group , 12 , 489 ( 72% ) provided the baseline blood sample at least 8 hours after a meal and these participants constitute the fasting subsample . Samples in the replication analysis derive from PROCARDIS , an ongoing European study of premature coronary artery disease [30] , [31] , and from the Framingham Heart Study ( FHS ) [32] , an ongoing , family-based longitudinal cohort designed to identify correlates with cardiovascular health , including subgroup analysis of the impact of plasma lipoprotein fractions . The FHS samples with NMR-based lipoprotein measurements for replication derive from the Offspring cohort within the FHS [19] . In the WGHS , lipoprotein determinations were performed on baseline plasma samples that had been stored in liquid nitrogen ( −170°C ) since collection . LDL-C , HDL-C , triglycerides , ApoA1 , and ApoB100 levels were all measured by direct assay and had low coefficients of variation [29] . NMR-based lipoprotein fractions were determined as described by proton NMR spectroscopy ( LipoProtein-II assay , Liposcience Inc . , Raleigh , NC ) [33] . The coefficients of variation for these measures were also low ( range 0 . 4–7 . 1% ) , except for the concentration of medium HDL particles ( CV<30% ) and IDL particle concentration ( CV = 13 . 1% ) [4] . PROCARDIS measurements were also performed with LipoProtein-II assays . Lipoprotein fractions for the FHS [19] samples were measured with the LipoProtein-I assay ( Liposcience Inc . Raleigh , NC ) , which provides less accuracy for some measurements but is otherwise similar to LipoProtein-II . Genotyping in the WGHS sample was performed using the HumanHap300 Duo “+” chips or the combination of the HumanHuman300 Duo and iSelect chips ( Illumina , San Diego , CA ) with the Infinium II protocol . In either case , the custom SNP content was the same; these custom SNPs were chosen without regard to minor allele frequency ( MAF ) to saturate candidate genes for cardiovascular disease as well as to increase coverage of SNPs with known or suspected biological function , e . g . disease association , non-synonymous changes , substitutions at splice sites , etc . For quality control , all samples were required to have successful genotyping using the BeadStudio v . 3 . 3 software ( Illumina , San Diego , CA ) for at least 98% of the SNPs . In the final dataset , SNPs were retained with MAF >1% , successful genotyping in 90% of the subjects , and deviations from Hardy-Weinberg equilibrium not exceeding P = 10−6 in significance . A total of 335 , 603 unique SNPs , of which 32 , 521 derive from the custom content , remained in the final data . Although assays for two non-synonymous SNPs at the APOE locus ( 19q13 . 32 ) , rs429358 and rs7412 , which determine ApoE isotype , failed in the design of the Illumina custom content , genotypes for these two SNPs were determined separately by an allele-specific , PCR based method ( Celera , Alameda , CA ) [34] . These additional SNPs are in linkage disequilibrium with SNPs in the Illumina panel . The targeted genotypes for APOE were included during the model selection procedures but not during the primary analysis to discover loci with genome-wide significant associations . Primary analysis to discover loci with highly significant associations in the WGHS discovery cohort was performed by linear regression in PLINK [35] assuming an additive relationship between the number of copies of the minor allele of each SNP and the mean values of the adjusted lipoprotein measures . A conservative threshold of P<5×10−8 was assumed for genome-wide significance [36] . For each lipoprotein measure , a full adjustment was performed by linear regression using the clinical covariates: age at baseline ( continuous ) , BMI ( continuous ) , menopausal status ( yes/no ) , current smoking status ( yes/no ) , and use of hormone replacement therapy ( yes/no ) . Concentrations of IDL particles , total LDL particles , medium HDL particles , triglycerides determined by NMR , and triglycerides determined by chemical assay were log-transformed before adjustment to approximate normality . Self-reported European ancestry was confirmed among the WGHS participants included in the primary analysis by clustering in a principal component analysis in PLINK with 1443 ancestry informative SNPs chosen for large Fst values ( >0 . 4 ) among the HapMap CEU , YRI , and JPN+CHB populations [37] . Discrepancy between self reported European ancestry and the clustering pattern was observed only for 68 samples ( <0 . 5% ) , and these samples were excluded from the analysis . In addition , genomic control parameters for the primary analysis were close to unity , ranging from 1 . 013–1 . 061 . There was an estimated 80% power at the genome-wide significance level to detect effects explaining 0 . 23% and 0 . 32% of the variance in the adjusted lipoprotein measures respectively in the whole sample and the fasting subsample . The primary analysis also included association testing in a nested subset of 72% of the study participants who reported fasting for at least eight hours before providing the baseline blood sample . Analysis in this subset was expected to differ from the analysis in the whole sample by opposing trends: a loss of power due to reduced sample size was contrasted with possibly smaller variance among lipoprotein fractions that are influenced by prandial status , e . g . triglycerides . Because the majority of the sample was fasting , the association statistics in the two samples were expected to be highly correlated , and the statistical penalty for this additional testing in the Bonferroni framework was expected to be less than a factor of two . Our genome-wide significance threshold ( P<5×10−8 ) was already smaller than required by correction for the number of SNPs tested by a factor of three , and justified including testing the fasting subset in the primary analysis . Once loci having at least one genome-wide significant association with at least one lipoprotein fraction had been identified , a non-redundant set of SNPs contributing to each lipoprotein fraction at each locus was constructed by forward-backward stepwise selection using the Bayesian Information Criterion ( BIC ) from among all genotyped locus SNPs within 100 kb of the locus genome-wide SNP associations . Separately , these model selection procedures were performed also at each of the candidate loci with the unadjusted , but possibly log-transformed , lipoprotein fractions to estimate the proportion of variance explained without adjustment . To assess the degree to which the adjustment procedure or sub-European population stratification might influence the identification of genome-wide loci , we performed a secondary analysis to evaluate the sensitivity of the locus discovery procedure to the adjustments applied to lipoprotein fractions before association testing . First , we adjusted for all of the clinical covariates as well as ten eigenvectors corresponding to a principal component analysis of genotype frequency in EIGENSTRAT [38] among 64 , 208 SNPs chosen with inter-SNP LD r2<0 . 2 and followed by quantile normalization of the residuals . Second , we adjusted with all of the clinical covariates except BMI , either with or without inclusion of the eigenvectors and subsequent quantile normalization . Finally , we adjusted only for baseline age , again either with or without inclusion of the eigenvectors and subsequent quantile normalization . In an additional secondary analysis , the genome-wide association procedures were performed with lipoprotein fractions transformed and fully adjusted as for the primary analysis , including also log transformed triglyceride levels among the adjustment variables ( see text ) . Additional analytic procedures , including the hierarchical clustering of loci according to effects on lipoprotein fractions , as well as the graphical representations were programmed in R [39] , and included the False Discovery Rate analysis with the R-package QVALUE [40] . All annotations derive from human genome reference sequence hg18 ( NCBI build 36 . 1 ) , the UCSC Refseq as of October 27 , 2008 , and the dbSNP database ( build 129 ) as represented by the UCSC database . In the Framingham Heart Study ( FHS ) sample , residual lipoprotein fractions were created by adjusting for gender , age at exam lipoprotein fraction collection ( continuous ) , age-squared ( continuous ) , and the top ten principal components from EIGENSTRAT [38] before analysis . When appropriate , log transformations were applied to approximate normality before computing residuals . Association testing was performed in R [39] using a linear mixed effect regression model with a kinship matrix to account for the family structure in the sample . Genotype data were derived by imputation using MACH 1 . 0 ( http://www . sph . umich . edu/csg/abecasis/mach/ ) from raw genotypes collected with the Affymetrix ( Santa Clara , CA ) 500K array , and the regression models assumed a linear relationship between the dosage of the minor allele ( ranging from 0 to 2 ) and the lipoprotein measures [10] . Only SNPs with high quality imputation measures ( squared correlation of imputed and true genotype >0 . 3 ) were used in the analysis . In the PROCARDIS study [20] , where genotype data derive from the Illumina ( San Diego , CA ) Human 1M platform representing a superset of the SNPs in the WGHS data , lipoprotein fractions were adjusted for case/control specific effects of age at baseline ( continuous ) , gender , country of recruitment ( Germany , Italy , Sweden , United Kingdom ) , self-reported hypertension ( yes/no ) , diabetes ( yes/no ) , current smoking status by questionnaire ( yes/no ) , and statin therapy ( yes/no ) . Regression models assumed a linear relationship between the number of copies of the minor allele and adjusted mean lipoprotein measure . | Genome-wide association studies ( GWAS ) of plasma lipoprotein fractions hold great promise for understanding lipid metabolism and its central role in cardiovascular disease and related disorders . Conventional assays for lipoprotein status determine total cholesterol content of low- or high-density lipoprotein particles ( LDL-C or HDL-C , respectively ) or total plasma triglyceride content ( as an estimate of very-low density lipoprotein particle concentration [VLDL] ) . All three measures have been targets for recent GWAS . However , a more precise target for GWAS of lipoprotein metabolism would be the concentration of the individual lipoprotein particles according to class ( LDL , HDL , VLDL ) and size ( small , medium , and large ) , all of which can be measured by NMR-based methods . In a population of 17 , 296 women of European ancestry from the Women's Genome Health Study , we have performed a GWAS for 22 lipoprotein measures derived from NMR-based and conventional assays . We find 43 genetic loci involved in lipoprotein metabolism , including 10 novel loci . The results offer a clearer picture of common genetic influences on lipoprotein metabolism than available previously , including genetic effects on the distribution of LDL , HDL , and VLDL particle size , as well as on IDL and VLDL particle concentration , neither of which can be assessed by conventional measures . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/gene",
"discovery",
"genetics",
"and",
"genomics/complex",
"traits",
"cardiovascular",
"disorders/cardiovascular",
"diseases",
"in",
"women",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"genetics",
"and",
"genomics/medical",
"geneti... | 2009 | Forty-Three Loci Associated with Plasma Lipoprotein Size, Concentration, and Cholesterol Content in Genome-Wide Analysis |
Hereditary factors are presumed to play a role in one third of colorectal cancer ( CRC ) cases . However , in the majority of familial CRC cases the genetic basis of predisposition remains unexplained . This is particularly true for families with few affected individuals . To identify susceptibility genes for this common phenotype , we examined familial cases derived from a consecutive series of 1514 Finnish CRC patients . Ninety-six familial CRC patients with no previous diagnosis of a hereditary CRC syndrome were included in the analysis . Eighty-six patients had one affected first-degree relative , and ten patients had two or more . Exome sequencing was utilized to search for genes harboring putative loss-of-function variants , because such alterations are likely candidates for disease-causing mutations . Eleven genes with rare truncating variants in two or three familial CRC cases were identified: UACA , SFXN4 , TWSG1 , PSPH , NUDT7 , ZNF490 , PRSS37 , CCDC18 , PRADC1 , MRPL3 , and AKR1C4 . Loss of heterozygosity was examined in all respective cancer samples , and was detected in seven occasions involving four of the candidate genes . In all seven occasions the wild-type allele was lost ( P = 0 . 0078 ) providing additional evidence that these eleven genes are likely to include true culprits . The study provides a set of candidate predisposition genes which may explain a subset of common familial CRC . Additional genetic validation in other populations is required to provide firm evidence for causality , as well as to characterize the natural history of the respective phenotypes .
Colorectal cancer ( CRC ) ( MIM 114500 ) is a major cancer type , with over one million new cases diagnosed worldwide each year . It is the third most common malignancy [1] , and the second most common cause of cancer mortality [2] . Inherited factors are estimated to play a crucial role in at least one third of all CRC cases [3] . However , high-penetrance mutations in known CRC predisposing genes , such as the mismatch repair ( MMR ) genes , APC , MUTYH ( MYH ) , SMAD4 , BMPR1A , STK11/LKB1 , PTEN , AXIN2 , POLE , and POLD1 explain only around 5% of these cases [4]–[6] . There are a few examples of rare variants in CRC predisposing genes conferring moderate or low carrier risk , such as APC ( I1307K ) [7] , BLM [8] and GALNT12 [9] . Of these , the APC I1307K variant has been most extensively studied and occurs almost exclusively in the Ashkenazi Jewish population [7] . In addition to these , genome-wide association ( GWA ) studies have identified common low-penetrance variants at approximately 20 genomic loci associated with CRC susceptibility . However , the identified common variants at these loci exert only a modest effect on CRC risk [10]–[12] . Unknown variants of moderate or low penetrance are likely to explain at least part of the missing heritability in CRC . CRC families with few affected individuals are an attractive patient group to search for such genetic factors , but tools for such work have been poor . These families are relatively common but too small for linkage analyses , and the culprit variants are likely to be too diverse and rare to be detected in GWA studies . One approach has been to study the additive contribution of low-penetrance variants on familial risk . A previous study has estimated that ten known low-penetrance CRC variants collectively explain around 9% of the variance in familial risk [13] . Advances in sequencing technologies have made exome sequencing a feasible approach to search for rare coding variants of varying penetrance . In this study , we aimed at identifying variants predisposing to common familial CRC by performing exome sequencing on 96 independent familial CRC cases derived from a consecutive collection of unselected patients . Here , familial CRC is characterized as having at least one first-degree relative diagnosed with CRC; indeed the great majority of the 96 familial cases displayed only one first-degree relative with CRC . All patients were from Finland , known for its relatively homogenous population [14] , [15] . This empowers the analysis since affected individuals are more likely to share ancestral predisposition mutations and haplotypes , stemming from a handful of founders . To our knowledge , this is the largest effort to date where exome sequencing has been applied to familial forms of cancer to identify novel predisposing genes .
Exome sequencing analysis was performed on germline DNA from 96 independent familial CRC cases . The clinical and histopathological features of the cases are summarized in Table 1 and in more detail in Table S1 . The average read depth attained for target regions was 43 and at least 86% of the captured target regions were covered by four or more sequence reads for all the samples . We identified a total of 76 , 487 nonsynonymous variants in the exome data ( Figure 1 ) . Sequence data were first evaluated for known predisposing genes ( MLH1 , MSH2 , MSH6 , PMS2 , APC , MUTYH , SMAD4 , BMPR1A , LKB1/STK11 , PTEN , AXIN2 , POLE , and POLD1 ) . No clear pathogenic mutations were found in these genes . The following missense variants were identified ( not confirmed by Sanger sequencing ) ; MSH6 c . 2800G>C p . D934H , and PTEN c . 1016C>A P339Q . However , the patients did not present typical clinical phenotypes; in the case of the MSH6 variant the tumor did not display microsatellite instability and in the case of the PTEN variant patient records revealed no features suggestive of Cowden syndrome ( MIM 158350 ) . Thus , these variants remain of unknown clinical significance . We hypothesized that predisposing germline variants would likely be rare in the general population , and predicted to truncate the protein product . We therefore filtered the data to prioritize such variants ( Figure 1 ) . First , variants had to be protein truncating with putative loss-of-function alteration; including nonsense , frameshift ( insertion and deletion ) or splice-site variants ( IVS +1 , +2 , −1 , and −2 ) . A total number of 3 , 654 truncating variants were found in the exome data . Second , variants were excluded if present in the 1000 Genomes Project [16] or population matched exome control data ( n = 212 ) at minor allele frequency ( MAF ) >0 . 001 . After control filtering , 2 , 090 truncating variants remained . Third , genes with truncating variants in more than one familial CRC case were selected for further analysis . There were a total of 588 such variants of which 422 were frameshift , 115 nonsense , and 51 splice site variants ( Figure 1 ) . Frameshift variants were grossly overrepresented in the list of truncating variants due to sequencing artifacts . Finally , manual filtering was performed on all variants to further remove artifacts due to duplicated regions , mapping errors , and systematic sequence specific errors . The filtering procedure resulted in a shortlist of 29 genes with 46 truncating variants . These were subsequently validated by Sanger sequencing ( Figure 1 ) . Sanger sequencing was successful for all amplicons , and 23 truncating variants in 18 genes were confirmed . Of these seven were frameshift , 12 nonsense , and four splice-site variants . To further exclude neutral polymorphisms , the confirmed variants were screened in 310 Finnish population matched controls , of which approximately two-thirds were also regionally matched . Variants with MAF>0 . 001 in the overall discovery phase control set ( including Finnish control exome data and Sanger sequenced controls ) were excluded ( Figure 1 ) . In total , we identified 11 candidate predisposing genes with 14 truncating germline variants in at least two familial CRC cases ( Table 2 ) ; UACA , SFXN4 , TWSG1 , PSPH , NUDT7 , ZNF490 , PRSS37 , CCDC18 , PRADC1 , MRPL3 , and AKR1C4 . A summary of all these variants and respective frequencies are presented in Table 2 . Gene descriptions and proposed functions of the identified genes are listed in Table S2 . Typically , the same truncating variant was detected in several patients . However , three genes harbored two different types of truncating germline variants ( Table 2 ) . Nine genes showed truncating variants in 2/96 familial cases . Two genes had truncating variants in 3/96 cases; UACA ( uveal autoantigen with coiled-coil domains and ankyrin repeats ) ( 3/96 , 3 . 1% ) and SFXN4 ( sideroflexin 4 ) ( 3/96 , 3 . 1% ) . In UACA , p . Q1116X was identified in two out of 96 familial cases and present in 522 Finnish population matched controls with a MAF of 0 . 001 . UACA p . R1292X was found in one out of 96 cases and the variant was not found in controls ( Figure 2 ) . In SFXN4 , three out of 96 cases had c . 32delC . This variant had a MAF of 0 . 001 in population matched controls . None of the other identified truncating variants were identified in population matched controls , except for c . 389_390insA in PSPH which was found in 1/502 controls ( MAF 0 . 001 ) . To further explore the frequency of these variants in controls , we referred to the Exome Variant Server ( NHLBI GO Exome Sequencing Project ( ESP ) , Seattle , WA , http://evs . gs . washington . edu/EVS/ [July 2013] ) . Three of the identified germline variants , SFXN4 c . 32delC , NUDT7 c . 111T>A , and PRSS37 c . 176+1G>A , were reported , however , at a MAF of less than 0 . 0003 . The exome data was also searched for missense variants in the 11 candidate predisposition genes; five missense variants were observed in five genes ( Table S3 ) . All of the missense variants were present in one case only , except for p . Q83H in PSPH which was identified in two out of the 96 familial cases . None of the missense variants were predicted to have a damaging effect on the protein by either of the prediction programs used ( Table S3 ) . The identified missense variants were very rare in population matched controls ( MAF<0 . 001 ) . Loss of heterozygosity ( LOH ) was examined in cancers of CRC cases with candidate predisposing germline variants ( Figure 1 ) . The following genes displayed LOH in at least one cancer: UACA , TWSG1 , PSPH , and ZNF490 ( Table 2 ) . Seven LOH events were observed and all targeted the wild-type allele ( P = 0 . 0078 ) . In UACA three out of six examined tumors showed loss of the wild-type allele and in TWSG1 ( twisted gastrulation protein homolog 1 ) both of the tumors showed loss of the wild-type allele ( Figure 2 ) . Variants in genes showing loss of the wild-type allele in tumor tissue were genotyped in an independent set of validation phase samples ( Figure 1 ) . This set included 954 Finnish population matched CRC cases and 586 Finnish population matched controls . UACA p . Q1116X was identified in two additional unrelated CRC cases and one control ( Table 2 ) . The ages at diagnosis were 67 and 58 years for the two cases . In the overall set of Finnish population matched controls used in this study , two out of 1 , 108 controls had UACA p . Q1116X ( MAF = 0 . 0009 ) . UACA p . R1292X was found in one additional case ( diagnosis at the age of 61 ) and no controls were heterozygous for this variant . The variant p . R350X in ZNF490 was found in one additional case ( diagnosed at the age of 58 ) and remained absent in controls ( Table 2 ) . TWSG1 p . Q41X was not present in any additional cases or controls . Genotyping was not successful for PSPH c . 389_390insA . Next , LOH was analyzed in the tumors of the four additional cases with truncating variants ( Table 2 ) . One of the additional cases with UACA p . Q1116X showed LOH involving the wild-type allele ( Figure 2 ) . Segregation analysis of the identified truncating variants was performed for all the affected first degree relatives for whom samples were available . In total , segregation was analyzed in seven families for five of the identified truncating variants; c . 32delC in SFXN4 , p . Q41X in TWSG1 , p . R350X in ZNF490 , c . 168+1G>A in PRADC1 , and c . 620delA in AKR1C4 ( Figure 3 and Figure S1 ) . The following variants showed segregation; c . 32delC in SFXN4 , c . 168+1G>A in PRADC1 , and c . 620delA in AKR1C4 . The variant p . Q41X in TWSG1 segregated in one family but not the other ( Figure 3 ) and p . R350X in ZNF490 did not segregate ( Figure S1 ) .
Exome sequencing is a powerful tool for discovering novel genetic variants that predispose to disease [17] . To examine the genetic basis of common familial CRC we exome sequenced 96 independent cases ( Table 1 ) derived from a previously described population-based collection of patients [4] , [18] and from an additional unselected collection ( unpublished ) . To our knowledge , this is the largest effort to date where familial CRC has been studied by exome sequencing to identify novel CRC predisposing genes . Several strategies were applied to improve the power of gene discovery . First , a large set of familial CRC cases ( at least one first-degree relative diagnosed with CRC ) was utilized , negative for any known high penetrance CRC mutation . Second , the cases were from Finland , known for its isolated population with reduced genetic heterogeneity . Such isolated populations are enriched for rare founder variants , facilitating identification of disease genes [15] . Third , tumor tissue availability for all the CRC cases allowed for the assessment of somatic allelic imbalance , which gave important additional information related to pathogenicity of the variants . Fourth , genotyping of selected variants was performed in a set of validation phase population matched samples , consisting of 954 cases and 586 controls . In total , we identified 11 novel candidate CRC susceptibility genes with rare truncating variants in two or three familial CRC cases; UACA , SFXN4 , TWSG1 , PSPH , NUDT7 , ZNF490 , PRSS37 , CCDC18 , PRADC1 , MRPL3 , and AKR1C4 ( Table 2 and Table S2 ) . They were absent or rare ( MAF≤0 . 001 ) in the general population . The results fit with the “rare variant hypothesis” that proposes that a significant proportion of the missing heritability of complex diseases is due to a series of rare variants , each conferring a moderate increase in risk . Typically , such risk alleles function dominantly and independently [19] , [20] . The “rare variant hypothesis” is strongly supported by evolutionary theory , which argues that variants that promote disease are selected against and are therefore rare . Another argument for the hypothesis comes from recent empirical population genetic data which shows that rare variants are enriched for deleterious mutations [21] . The question remains whether the identified candidate genes act as classical tumor suppressors with second hits or show alternative characteristics , such as haploinsufficiency or dominant-negative effects . Of the genes identified , four out of 11 showed loss of the wild-type allele in at least one tumor . In total , seven LOH events were observed and none showed loss of the mutant allele ( P = 0 . 0078 ) . This suggests that complete inactivation of these genes seems to be preferentially selected for in tumor evolution and that these germline variants are prime candidates for CRC susceptibility . Perhaps the strongest candidate predisposition gene , in view of the LOH data and case frequency , was the apoptosis-associated gene UACA . Three of the 96 familial CRC cases were found to carry heterozygous truncating variants ( p . Q1116X and p . R1292X ) in UACA ( Table 2 ) . We performed genotyping to screen the variants in a set of validation phase samples . We identified three additional unrelated cases who were heterozygous for the variants encoding either p . Q1116X or p . R1292X . Second hits by LOH involving the germline wild-type allele were found in three of the six tumors ( Figure 2 ) . The average age of onset of CRC in the familial cases was 54 years ( 58 , 54 and 50 ) ( Figure 3 ) , younger than the mean age of onset of 71 in familial cases without the UACA truncating variants ( Table S1 ) . UACA has recently been identified as a novel regulator of apoptosis . It is known to reside within the Apaf-1/procaspase-9 complex and regulate apoptosis activating factor ( APAF-1 ) . It also regulates the apoptotic pathway by controlling the activation of nuclear factor ( NF ) -κB [22] . In addition , UACA gene expression has been shown to be down-regulated in non-small cell lung carcinoma ( MIM 211980 ) [23] . Taken together , the loss of UACA in cancer cells might result in altered activation of apoptotic pathways , ultimately promoting genesis of CRC . Another gene of particular interest was TWSG1 . The detected truncating germline variant ( p . Q41X ) was present in two familial CRC cases ( 2/96 cases ) and completely absent in 1 , 039 Finnish population matched controls ( Table 2 ) . Loss of the remaining normal TWSG1 allele was observed in both tumors indicating that the gene might act as a classical tumor suppressor gene ( Figure 2 ) . The index case with Q41X in family 1 developed CRC at the age of 53 and segregation analysis showed that the variant was inherited from the affected mother ( Figure 3 ) . The mother had developed CRC at the age of 68 and lung cancer at the age of 77 . The variant did not segregate with CRC in family 2 . Rare risk alleles of moderate penetrance are usually over-represented in familial cases; however co-segregation of disease is not always observed [19] . Previous studies have shown TWSG1 to be a regulator of BMP-signaling [24] . It is known to act downstream of TGF-β , inducing SMAD2 phosphorylation and mediating DNA binding on Smad3/4 consensus sites [25] . TWSG1 functions in cellular pathways that are essential in genesis of CRC , however , its exact role in these pathways remains to be clarified . In summary , exome sequencing is a well-justified strategy for discovering cancer predisposing variants . The identification of predisposing variants has substantial implications for disease risk assessment and surveillance in family members . Here , we identified eleven candidate predisposing genes with truncating variations in familial CRC . A key challenge is how to identify predisposing variants in the background of non-pathogenic polymorphisms . Screening the eleven genes in familial CRC cases representing different populations will be important to gain robust evidence for pathogenicity , as well as to characterize the natural history of the respective phenotypes . This information , then , can be translated into tools for cancer prevention and early diagnosis in individuals carrying true predisposition alleles .
This study was reviewed and approved by the Ethics Committee of the Hospital district of Helsinki and Uusimaa ( HUS ) . Signed informed consent or authorization from the National Supervisory Authority for Welfare and Health was obtained for all the study participants . The Agilent SureSelect Human All Exon Kit v1 ( Agilent , Santa Clara , CA , USA ) was used to capture exomic regions . Paired end short reads were sequenced on either Illumina GAII or HiSeq platform ( Illumina Inc . , San Diego , CA , USA ) . Raw sequence data was received in FASTQ format and quality checked with FASTQC ( http://www . bioinformatics . bbsrc . ac . uk/projects/fastqc ) . All exomes passed the quality control . 3′ ends with high adapter similarity were removed by an in-house script whereafter reads were mapped to the human reference genome GRCh37 by BWA ( Burrow-Wheelers Aligner ) . Duplicates were removed with Picard Tools ( http://picard . sourceforge . net ) MarkDuplicates . Local realignment was done by Genome Analysis Toolkit ( GATK ) IndelRealigner to improve the detection of small insertions and deletions . The initial single nucleotide variant ( SNV ) and indel calls needed for creating the GATK realignment intervals were made using samtools mpileup and downloaded from the 1000 genomes project Phase I indel calls ( [16]the August 2010 release ) . Final SNV and indel calls were made using the GATK UnifiedGenotyper with a low variant quality score threshold ( 1 . 0 ) . Exome data analysis was performed in “Rikurator” , an in-house visualization and comparative analysis tool ( unpublished ) . The tool allowed for simultaneous analysis of all the 96 exomes and interactive quality/control filtering . The following quality filters were used: ( i ) variants had to have a quality score ≥50 , ( ii ) coverage had to be ≥6 , and ( iii ) the percentage of mutated reads had to be ≥30 . Truncating variants , including nonsense , frameshifting insertion and deletion , or splice-site alteration IVS +1 , +2 , −1 , and −2 , were extracted . Data was control filtered against population matched exome control data ( n = 212 ) and data from the 1000 Genomes Project ( Phase 1 release ) [16] . Variants were excluded if present in the 1000 Genomes Project or exome control data at MAF>0 . 001 ( Figure 1 ) . Genes with truncating variants present in at least 2/96 cases were studied further . Manual filtering was performed on all variants to further remove artifacts due to duplicated regions , mapping errors , and systematic errors . Systematic errors , both position specific and sequence specific , in high-throughput sequence data have been described previously by Meacham et al [26] . Finally , outputs were generated for Ensembl canonical transcripts ( Ensembl build 37 ) . Potential loss-of-function variants were verified by Sanger sequencing from DNA extracted from normal tissue samples . Sequencing primers were designed with the Primer3-program ( http://frodo . wi . mit . edu/primer3/ ) using NCBI37/Hg19 as the reference sequence . The primer sequences can be found in Text S1 . The fragments were amplified with the AmpliTaqGold enzyme ( Applied Biosystems , Foster City , CA ) . The PCR products were purified using the ExoSAP-IT PCR purification kit ( USB Corporation , Cleveland , OH , USA ) . Electrophoresis was run on a 3730xl DNA Analyzer ( Applied Biosystems at Institute for Molecular Medicine Finland , FIMM ) . The sequencing reactions were performed utilizing the Big Dye Terminator v . 3 . 1 kit ( Applied Biosystems , Foster City , USA ) , Sanger sequencing was performed implementing the ABI3100×l technology ( Applied Biosystems ) , and the sequence graphs were visualized with the Chromas – software ( version 2 . 33 , Technelysium Pty Ltd , Helensvale , Australia ) . The results were analyzed both manually and with the Mutation Surveyor –software ( version v3 , 30 , Softgenetics , State College , PA , USA ) . Confirmed truncating variants were Sanger sequenced in 310 Finnish population matched healthy controls , of whom about two-thirds were regionally matched . Sanger sequencing was performed as described above . All variants that had a MAF>0 . 001 in the discovery phase control set were excluded . Sanger sequencing was also performed on DNA extracted from tumor tissue in cases carrying validated truncating variants . All tumors had been microscopically evaluated by a pathologist and all except one contained ≥50% of carcinoma tissue . Loss of heterozygosity was analyzed by comparing allelic ratios of tumor and respective normal tissue DNA , as previously described [27] . Peak heights were manually measured from sequence graphs based on which allelic ratios were calculated . Variants in genes showing loss of the wild-type allele in tumor tissue were genotyped in a set of validation phase samples , comprising 954 population matched CRC cases and 586 population matched controls . Genotyping was carried out by using the 7900HT Fast Real-Time PCR System ( Applied Biosystems ) and was performed at the Estonian Genome Center , University of Tartu . The variant p . Q41X in TWSG1 was genotyped using massARRAY iPLEX Gold ( Sequenom , San Diego , CA ) and performed at the Institute for Molecular Medicine Finland ( FIMM ) , University of Helsinki . The genotyping conditions and primers utilized can be found in Text S1 . Genotyping success rates were over 90% for all the variants , except for PSPH where the genotyping assay failed . All the variants identified by genotyping were further confirmed by Sanger sequencing . The exome data was searched for missense variants at the 11 candidate predisposition loci . The same filtering criteria were utilized as for truncating variants . The variants were excluded if present in Exome Variant Server ( NHLBI GO Exome Sequencing Project ( ESP ) , Seattle , WA , http://evs . gs . washington . edu/EVS/ [July 2013] ) with MAF>0 . 001 . The functional effects of the identified missense variants were predicted by SIFT ( http://sift . jcvi . org/ ) and PolyPhen 2 ( http://genetics . bwh . harvard . edu/pph2/ ) . Archived Formalin-fixed , Paraffin-embedded ( FFPE ) tissue samples were ordered for first degree relatives with CRC whenever possible . Genomic DNA was extracted from all available FFPE samples . Sanger sequencing was performed on identified truncating variants to test for segregation . In total , segregation was analyzed in seven families for five of the identified truncating variants . One-tailed exact binominal test was used for P-value calculations . | Many individuals with a family history of colorectal cancer have no detectable germline mutation in the known cancer predisposing genes . We aimed to identify novel susceptibility genes for this common phenotype by performing exome sequencing on 96 independent cases with familial colorectal cancer . Eighty-six patients had one affected first-degree relative , and ten patients had two or more . None of the patients had a previous diagnosis of a hereditary syndrome . We focused our search on genes with rare variants , predicted to truncate the protein product , since these are likely candidates for disease predisposition . Using this approach we identified truncating germline variants in eleven genes , present in two or three independent familial colorectal cancer cases . We analyzed the respective tumor DNAs and found loss of the wild-type allele in seven out of seven occasions , involving four genes . No tumor showed loss of the mutant allele which provides us with additional evidence for disease causality . Further studies are required to provide firm evidence for pathogenicity . Genetic knowledge on confirmed predisposing genes can ultimately be translated into tools for cancer prevention and early diagnosis in individuals carrying predisposition alleles . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | Eleven Candidate Susceptibility Genes for Common Familial Colorectal Cancer |
The rapidly growing amount of publicly available information from biomedical research is readily accessible on the Internet , providing a powerful resource for predictive biomolecular modeling . The accumulated data on experimentally determined structures transformed structure prediction of proteins and protein complexes . Instead of exploring the enormous search space , predictive tools can simply proceed to the solution based on similarity to the existing , previously determined structures . A similar major paradigm shift is emerging due to the rapidly expanding amount of information , other than experimentally determined structures , which still can be used as constraints in biomolecular structure prediction . Automated text mining has been widely used in recreating protein interaction networks , as well as in detecting small ligand binding sites on protein structures . Combining and expanding these two well-developed areas of research , we applied the text mining to structural modeling of protein-protein complexes ( protein docking ) . Protein docking can be significantly improved when constraints on the docking mode are available . We developed a procedure that retrieves published abstracts on a specific protein-protein interaction and extracts information relevant to docking . The procedure was assessed on protein complexes from Dockground ( http://dockground . compbio . ku . edu ) . The results show that correct information on binding residues can be extracted for about half of the complexes . The amount of irrelevant information was reduced by conceptual analysis of a subset of the retrieved abstracts , based on the bag-of-words ( features ) approach . Support Vector Machine models were trained and validated on the subset . The remaining abstracts were filtered by the best-performing models , which decreased the irrelevant information for ~ 25% complexes in the dataset . The extracted constraints were incorporated in the docking protocol and tested on the Dockground unbound benchmark set , significantly increasing the docking success rate .
The rapidly growing amount of publicly available information from biomedical research is a modern day phenomena that is likely to continue and accelerate in the future . Most of this information is readily accessible on the Internet , providing a powerful resource for predictive biomolecular modeling . The accumulated data revolutionized structure prediction of proteins in the 80’s [1] and , recently , of protein complexes [2–4] due to the growth of Protein Data Bank ( PDB ) [5] , providing enough structural “templates” for the prediction targets . Instead of painstaking and generally unreliable exploration of the enormous search space , based on the physical “first principles , ” nowadays tools can simply proceed to the solution based on similarity to the existing , previously determined structures . In our opinion , the next stage of this revolution is brewing due to the rapidly expanding amount of information , other than experimentally determined structures , which still can be used as constraints in biomolecular structure prediction [6] . In this paper we present the first , to our knowledge , approach to structural modeling of protein-protein ( PP ) complexes ( protein docking ) , based on the input from automated text mining ( TM ) of publications on the Internet . Protein-protein interactions ( PPI ) are central for many cellular processes . Structural characterization of PPI is essential for fundamental understanding of life processes and applications in biology and medicine . Because of the inherent limitations of experimental techniques and rapid development of computational power and methodology , protein docking is a tool of choice in many studies . One of the main problems in protein docking [7] is identification of a near-native match among the large , often overwhelming , number of putative matches produced by a global docking scan . To detect the near-native matches at the docking post-processing stage , a scoring procedure is performed by re-ranking of the scan output matches , typically using energy/scoring functions , which are too computationally expensive or impossible/impractical to include in the global search . Such scoring schemes may be based on structural , physicochemical , or evolutionary considerations [8] . For some PPI , information on the docking mode ( e . g . one or more residues at the PP interfaces ) is available prior to the docking . If this information is certain , there is no need for the docking global scan , and the search can be performed in the sub-space that satisfies the constraints . However , if the probability of such information is <100% , it may rather be included in the post-processing of the global scan , as part of the scoring . Given the inherent uncertainties of the global-search docking predictions , such independent information on the binding modes is extremely valuable [9] . Such information may be available on the case-by-case basis . However , for docking server predictions that can be used by the broad biological community an automated search for such information can be of great value . The PPI research is an extremely active field , yielding a vast amount of publications on interacting proteins [10] . These publications quickly become available online ( e . g . through PubMed , http://www . ncbi . nlm . nih . gov/pubmed ) , and are a growing resource for automated mining of the PP binding mode . Many applications ( PubMed ENTREZ , NLProt , MedMiner , etc . ) utilizing TM techniques have been developed to improve access to the published knowledge [11] . TM converts textual information into database content and complex networks , facilitating development of novel working hypothesis [12] . In biology , TM tools have been used to mine generic or specific information on genes , proteins and their functional relationships . Natural Language Processing ( NLP ) and Support Vector Machines ( SVM ) have been used to extract information on connection between proteins in PPI networks [13–19] . Along with the networks of interacting proteins , TM tools have been used to generate a dataset of non-interacting proteins [20] . Full-text articles on metabolic reactions and mutation impacts have been mined by rule-based parser , pattern matching , and entity taggers ( protein and gene names along with specific keywords ) [21 , 22] . Automated TM procedure was developed to predict subcellular localization and function of proteins [23] . TM in combination with the dynamic perturbation analysis has been employed to increase confidence in predicted protein functional site [24] , suggesting that if a residue is mentioned in an abstract on the protein structure , it is likely to be in the functional site . TM approaches are also implemented in many Web-based applications . There are different TM tools for identification of interacting proteins from biological literature and databases [25] . GENIA corpus ( a collection of semantically annotated documents ) has been specifically designed for testing NLP approaches [26] . PESCADOR extracts a network of interactions from a user-provided set of PubMed abstracts [27] . LAITOR can further filter this mined interactome according to the specific user needs [28] . CRAB extracts data from MEDLINE abstracts , which are relevant to tumor-related chemicals posing risk to human health [29] . PIE utilizes word and syntactic features to effectively capture PPI patterns from biomedical literature [30 , 31] . eFIP mines information on phosphorylation and related interactions of a given protein using rule-based NLP [32] . PPInterFinder extracts Medline abstracts on human proteins using co-occurrences of protein names , specific keyword dictionary , and pattern matching [33] . BioQRator can annotate PPI-relevant entity relationships from the biomedical publications [34] . In this paper , we propose the first , to our knowledge , approach to TM constraints for PP docking . Our methodology , by design , is a combination and expansion of two well-developed TM fields: ( 1 ) identification of interactors in PPI networks , and ( 2 ) detection of protein functional ( small ligand ) sites . We use the first one as the source of expertise on TM of PPI ( existing approaches are concerned with the fact of interaction , not the mode of interaction ) , and the second one as the source of expertise on TM for structural prediction of the binding sites on proteins ( existing approaches are for small non-protein ligands ) . The method was tested on PubMed abstracts of publications on protein complexes from Dockground ( http://dockground . compbio . ku . edu ) and showed a significant improvement of the docking success rates .
The principal stages of the TM protocol are shown in Fig 1 . We divide our procedure into two parts , information retrieval ( selecting abstracts containing names of both or either proteins in a complex ) and information extraction ( detecting occurrence of residues in the retrieved abstracts ) . The abstracts were further filtered by SVM model with optimal sets of features . The TM tool was benchmarked on 579 PP complexes with known bound X-ray structures from Dockground and applied for re-scoring of the initial docking models for 99 protein pairs from the Dockground unbound benchmark set 3 . Protein name and UniProtKB ID , corresponding to the particular PDB code and protein chain , were obtained from PDB . To simulate the “real case scenario” when the structure of the target protein is unknown , PubMed ID ( PMID ) of the direct citation ( publication describing the X-ray structure of the complex ) was extracted from the PDB and the publication excluded from the further consideration . To further test our methodology , we also restricted the analysis to abstracts published prior to the direct citation paper . Using the UniProtKB ID , protein information in XML format was acquired from the UniProtKB [35] . Information from both PDB and UniProtKB was accessed through REST ( REpresentational State Transfer ) Web Services ( http://www . rcsb . org/pdb/software/rest . do ) , ( http://www . uniprot . org/help/programmatic_access ) . For the query construction , at the current stage , we used only recommended , short and alternative protein names , ignoring organism name , classification of the monoclonal antibodies ( “CD_antigen” tag ) , all parts of gene information ( name , synonyms , ordered locus names , and open reading frame ) , E . C ( Enzyme Commission ) numbers , as well as UniProtKB terms “Uncharacterized protein . ” Inclusion of all this additional information into the search queries requires implementation of deep parsers , which is in our plans for the future research . Protein names were normalized by replacing reserved characters with their URL encodings ( spaces replaced by %20 , etc . ) , by removing extra ( trailing ) spaces , and by hyphen replacement . The query for a protein with a hyphen in its name contains OR-connected versions of the name with hyphen , hyphen removed and replaced by space . For example , query for “IL-15R-alpha” ( PDB: 2z3q , chain B ) also includes following variations: “IL-15Ralpha” , “IL15R-alpha” , “IL15Ralpha” , “IL%2015Ralpha” , “IL15R%20alpha” , and “IL%2015R%20alpha” . Short names with < 3 symbols were ignored and additional AND-connected keyword “protein” was added to the 3-symbols names . For 13 PP complexes in the set , protein names coincided with generic , frequently used words ( “act” for 1yrt , chain A , UniProtKB: P0DKX7 or “hot” for 2ido , chain B , UniProtKB: Q71T70 ) . To reduce noise , search queries for such complexes contained also MESH terms [36] ( combinations of the MESH terms under heading “Biochemical Phenomena” ) . For 87 proteins , UniProtKB had section “Cleaved into the following X chains” referring to different domains . In such cases , we considered several scenarios . In the case of exact match between PDB and UniProtKB recommended protein names ( 29 proteins ) , we assumed that the PDB structure comprised all the domains mentioned in UniProtKB and included into the query OR-connected recommended name for the entire protein and the names of all the domains . For example , X-ray structure for cationic trypsin ( PDB 2xtt , chain B , UniProtKB: P00760 ) contains both domains ( Alpha-trypsin chain 1 and Alpha-trypsin chain 2 ) mentioned in UniProtKB and the PDB name matches exactly the UniProtKB recommended name . If PDB name matched exactly only one of the domain names ( 17 proteins ) , than the query included the name of only this domain along with the recommended protein name . For example , “Protease inhibitor SGPI-1” ( PDB 2xtt , chain A , UniProtKB: O46162 ) is the PDB protein name , which matches exactly one of the cleaved components and does not match recommended UniProtKB name “Serine protease inhibitor I/II” . If the PDB protein name did not match exactly neither UniProtKB recommended name , nor any domain names ( 41 proteins ) , we considered only the recommended and the PDB names . For example , “epithelial-cadherin” is the PDB protein name ( PDB 2omz , chain B , UniProtKB: P12830 ) , which is not the same as the recommended UniProtKB name “Cadherin-1” and none of the cleaved components ( “E-Cad/CTF1” , “E-Cad/CTF2” , “E-Cad/CTF3” ) . String comparison was performed using Perl module Text::Levenshtein , which implements Levenshtein similarity string matching algorithm ( http://search . cpan . org/dist/Text-Levenshtein/lib/Text/Levenshtein . pm ) . After constructing two queries , “query1” and “query2” , one for each protein in a particular complex , two final queries were assembled: “query1 AND query2” ( termed here as AND-query ) and “query1 OR query2” ( OR-query ) . The AND- and OR-queries were submitted to ESearch and EFetch modules of NCBI EUtilies tool ( http://www . ncbi . nlm . nih . gov/books/NBK25501 ) . To keep track on which protein is studied in the retrieved abstracts , two parts of the OR-query were submitted separately . Maximum of 100 , 000 PubMed abstracts with publication dates between January 1 , 1971 and November 30 , 2014 were retrieved for each submitted query . Abstracts of publications corresponding to 579 complexes , retrieved by the E-utilities from the PubMed ( the number of abstracts varies for different types of queries , see Results and Discussion ) , were searched for the residues using regular expressions ( Table 1 ) obtained by the manual inspection of 100 abstracts that mention residues . We considered patterns with only three-letter or full residue names , since mining of one-letter residue abbreviations requires deep parsing of the surrounding text , which is beyond the scope of our current study . However , if keywords related to mutagenesis studies ( “mutation” , “mutagenesis” , “mutagen” , “mutant” , “substitution” ) were spotted , one-letter abbreviations for mutation ( e . g . , “S4A” ) were included in the search patterns . For the mutations , both original and substitution residues were taken ( e . g . , for the pattern “S4A” , both Serine 4 and Alanine 4 were considered as the mined residues ) . Since residues participating in docking are on the protein surface , the names and numbers of the extracted residues were checked against the names and numbers of the surface residues from the original PDB file . For the AND-query , the check was performed against both chains of the original complex , whereas for the OR-query , the examination was done only for the protein mentioned in the retrieved abstract ( to reduce noise due to the accidental match of the residue name and numbers ) . Surface residues were defined as those with ≥ 25% of their surface exposed to solvent [37] . The solvent accessible area was calculated by the program surfv [38] . Only the residues with both name and number matching the residues from the original PDB file were considered further ( we termed them "identified residues" ) . In the case of mismatch between PDB and UniProt sequence numbering , we mapped the UniProt sequence on the PDB one as in Ref . [39] . An identified residue was considered correct if any of its heavy atoms was ≤ 6 Å from any heavy atom of the interacting protein in the co-crystallized complex , which means that the residue is at the PP interface . Performance of the TM protocol for a particular PPI , for which a query extracted N abstracts containing residues , was quantified as a fraction of correct ( interface ) residues among all identified residues PTM=∑i=1NNiint∑i=1N ( Niint+Ninon ) , ( 1 ) where Niint and Ninon are numbers of interface ( correct ) and non-interface ( incorrect ) residues in abstract i . We generated a set of features by handpicking 60 words from carefully read randomly selected 21 PPI abstracts and 43 non-PPI abstracts ( Table 2 ) . Subsets of 50 , 40 , 30 , 20 and 10 features were also selected based on our understanding of importance of a feature for PPI description . We refer to these sets of features as manually selected , abbreviated as MFxx , where xx is the number of features in the set . We also generated a set of features by automated counting of words in the abstracts . We refer to this set and all of its subsets as automatically selected , abbreviate as AFxx , where xx is the number of features in the set . For this set , Lpos = 450 positive and Lneg = 855 negative abstracts , respectively satisfying conditions [Niint−Ninon>4]OR[Ninon=0ANDNiint>0]and[Ninon−Niint>4]OR[Ninon>0ANDNiint=0] , ( 2 ) were selected from 1 , 523 abstracts retrieved by the AND-queries . Positive and negative abstracts were further randomly split into training ( 80% of abstracts , or Lpostrain = 360 positive and Lnegtrain = 684 negative abstracts ) and validation ( remaining 20% ) sets , and features were selected from the training set . Specific protein and amino acid names were excluded from the counting , as they were part of the queries in the TM protocol . Stop words ( “and” , “as” , “because” , “the” ) were also purged from the abstracts . The abstracts were subjected to the tokenizer [40] for the suffix stripping by the Porters stemming algorithm [41] in order to get the stem ( root ) forms of the remaining abstract words . We slightly modified the original algorithm so that a root of a word would accommodate wider variability in the spelling of words with the same meaning . For example , words “include” and “inclusion” are counted by the root “inclu-” , words “mutant” , “mutagenesis” , “mutation” , “mutagen” , “mutated” , “mutations” , “mutation” are accounted for by the root “muta-” , etc . The normalized counts for each stem ( feature ) were calculated separately for the positive ( k = pos ) and the negative ( k = neg ) abstracts in the training set fk ( m ) =1Lktrain∑i=1LktrainJi ( m ) , ( 3 ) where Ji ( m ) is the number of times feature m appears in abstract i . All features satisfying conditions |fpos ( m ) −fneg ( m ) |>0 . 02andfpos ( m ) +fneg ( m ) >0 . 2 ( 4 ) were selected for the full set AF143 ( automatically selected 143 features ) . The criteria are meant to balance a maximal number of features and a strong signal . The full set was sorted based on the ratio δ ( m ) =fpos ( m ) −fneg ( m ) fpos ( m ) +fneg ( m ) ( 5 ) and consists of 76 PPI-relevant ( δ ( m ) > 0 ) and 67 PPI-non-relevant ( δ ( m ) < 0 ) features ( Table 2 ) . We define an SVM model as an SVM classifier with a kernel function , trained and validated with a particular set of features . For the training and validation of the SVM models we used readily available SVMLight [42 , 43] with polynomial , K ( Xi , Xj ) = ( αXiXj + C ) d , and radial-base ( RBF ) , K ( Xi , Xj ) = exp ( −γ|Xi − Xj|2 ) kernel functions ( Xi and Xj are support and test feature vectors ) . We tested different values of parameters d and γ while parameters α and C had default values ( α = 1 and C = 0 ) , and distinguished a particular case of the polynomial kernel with d = 1 , as the linear kernel . We have also investigated how results are affected by varying degree d of polynomial and parameter γ of RBF kernels . Validation of the SVM models was carried out in the classification mode where an abstract was identified as positive or negative depending on the sign of the SVM-score . In some cases , abstracts with SVM scores close to zero ( within a margin ) were considered as “unclassified” and excluded from the performance evaluation . All SVM models were trained on 1044 abstracts and validated on different 261 abstracts ( see above ) . Performance of an SVM model was evaluated in usual terms of precision P , recall R , accuracy A [44] , and Matthews’ correlation coefficient MCC [45] P=TPTP+FP , R=TPTP+FN , A=TP+TNTP+FN+TN+FPandMCC=TP×TN+FP×FN ( TP+FP ) ( TP+FN ) ( TN+FP ) ( TN+FN ) , ( 6 ) where TP , FP , TN , and FN are , correspondingly , the numbers of correctly identified positive , incorrectly identified positive , correctly identified negative and incorrectly identified negative abstracts in the validation set . Basic TM protocol with the OR-queries was used to mine residues for 99 complexes from the Dockground benchmark set 3 [46] , containing the unbound X-ray structures for the co-crystallized complexes ( bound structures ) . Queries for individual proteins and 63 binary complexes were generated as described above . For 36 multimeric complexes , queries were generated using OR-combinations of queries for all monomers in a multimeric chain ( e . g . , for complex AB: CDE the OR-query was “ ( queryA OR queryB ) OR ( queryC OR queryD OR queryE ) ” ) . Abstract of publications on the X-ray structure of the co-crystallized complex were excluded from consideration using corresponding PMID from the PDB entry . For validation , the extracted residues were matched to the residues in the bound structures of the dataset ( numbering and chain IDs in the bound and the unbound structures is often different ) . Extracted residues were ranked , in descending order , separately for each interactor ( single or multimeric ) by the confidence function f ( R ) =min ( 10 , ∑i=1NRai ) , ( 7 ) where NR is the total number of distinct abstracts , in which residue R is mentioned , and ai = 2 , if abstract i was retrieved by the AND-query and ai = 1 , if the abstract was retrieved by the OR-query only . Top five residues for each interactor were used as constraints in our GRAMM docking program [47] giving an extra weight ( proportional to f ( R ) ) to the scoring function if the identified residue was at the interface of a docking model . The upper limit of 10 in Eq 7 was chosen to balance the diversity of low confidence ( f = 1 ) vs . high confidence ( f = 10 ) constraints and potential overrepresentation of a residue in publications ( very high f values ) . If > 5 residues had the highest f values , then preference was given to the residues with scores containing more contributions from the abstracts retrieved by the AND-queries . Otherwise , the excess residues were removed from the list randomly . For validation , the residues at the crystallographically determined interface ( reference residues ) were extracted from the co-crystallized complexes using 6 Å distance cutoff between the heavy atoms of the proteins in the complex . All pairs of these interface residues were ranked in ascending order by the distance between their Cα atoms . The top three pairs were submitted to GRAMM with the highest possible confidence score 10 ( reference constraints ) . The unbound structures were docked by GRAMM once using the TM constraints and then , for comparison , the reference constraints . The output of the global low-resolution docking scan consisted of 20 , 000 matches , with no post-processing ( except for the removal of redundant matches ) . These matches were subjected to scoring by the sum of the f values ( Eq 7 ) if constraints were generated for the complex . If no constraints were generated , the score was zero . All matches were then re-sorted according to these scores . The quality of a match was assessed by Cα ligand interface root-mean-square deviation , i-RMSD ( ligand and receptor are the smaller and the larger proteins in the complex , respectively ) , calculated between the interface of the docked unbound ligand and corresponding atoms of the unbound ligand superimposed on the co-crystallized bound structure .
We ran the free docking by GRAMM to model complexes of unbound proteins from the Dockground X-ray benchmark 3 [46] using constraints generated by the basic TM protocol with OR-queries ( see Methods ) . In the unbound set of 99 complexes , by design the component proteins have both the co-crystallized and separately resolved X-structures , and as such were presumably on average more extensively studied than the complexes from the main bound set of 579 complexes used in this study for TM evaluation . This resulted in a significantly larger pool of publications extracted by the OR-queries ( 68 abstracts per complex for the unbound set , compared to 32 abstracts per complex for the bound set ) . Thus , a significantly larger number of residues per complex were identified ( S13 Fig ) and the TM performed better on the unbound set than on the larger bound one ( last row in Table 3 ) . However , the number of irrelevant ( non-interface ) residues was also significantly larger , reducing TM effectiveness ( S14 Fig ) . The AND-queries retrieved abstracts with residues for 37 complexes only and TM protocol with the AND-queries was not used here separately . However , for residue ranking ( Eq 7 ) , we kept track of which residues were retrieved by the AND-queries . The TM results based on OR-queries for the top 10 residues per complex ( 5 for each protein , ranked by the frequency of the residue occurrence , Eq 7 ) were significantly better ( S14 Fig ) . Thus , these residues were submitted to GRAMM docking program for scoring of the docking scan output . To single out the role of the TM constraints , they were applied to re-rank unrefined and otherwise unscored docking models output directly from the GRAMM scan ( the baseline for evaluating the impact of the TM constraints ) . For comparison , the re-ranking was also done separately with the correct interface residues as constraints ( see Methods ) . We used strict ( at least , one model with i-RMSD ≤ 5 Å in top 10 predictions ) and relaxed ( at least , one model with i-RMSD ≤ 8 Å in top 100 predictions ) success criteria . The TM scoring significantly increased docking success rates , by 71% ( compared to the baseline shown as blue columns in Fig 6 ) according to the stricter criterion , and by 32% according to the relaxed one ( Fig 6 ) . The results on the reference set of constraints , corresponding to the correct interface residues , showed that 27 complexes have near-native matches in the top 20 , 000 scan predictions according to the strict criterion , and 62 according to the relaxed one . The RMSD was calculated between the unbound ligand predicted match and the unbound ligand structurally aligned with the bound in the complex . Such alignment has significant mismatches with the receptor in a number of complexes , due to the conformational change upon binding . So the near-native matches for such complexes cannot be predicted by the surface complementarity-based rigid body free docking ( this correlates with the docking decoys results [53] where a near-native match was found only for 61 complexes in 500 , 000 top scan matches ) . The observed increase of the docking success rate is the result of constraints from the basic TM only . One can assume that the deep parsing/NLP will lead to further improvement of the docking quality , closer to the level of the reference constraints ( Fig 6 ) . TM has been widely used in recreating PPI networks , as well as in detecting functional sites ( small ligand binding sites ) on protein structures . Combining and expanding these two well-developed research areas , we applied TM to structural modeling of protein-protein complexes ( protein docking ) . Abstracts of publications on 579 protein complexes from Dockground were retrieved from PubMed , using AND- and OR-queries ( both proteins and at least one protein mentioned in the text , correspondingly ) . The AND-queries identified more correct residues than the OR-queries , but retrieved abstracts with residues for significantly less complexes . SVM was used to improve the performance of OR-queries . The SVM models generated using simple bag-of-words representation of the text , removed irrelevant information extracted by the OR-queries , albeit not enough for an accurate discrimination of non-interface from the interface residues , as shown by the inconsistent performance of different SVM models . Whereas human expertise can consistently distinguish relevant from non-relevant to the interface information ( as shown by our evaluation of a small subset of abstracts ) , a reliable and accurate automated procedure requires greater sophistication than the basic one used in our study . The basic TM was used to generate constraints for docking , and tested on the protein-protein unbound docking benchmark set . TM significantly increased the docking success rates . Contextual analysis by deep parsing on sentence/residue level ( an on-going study in our group ) should improve the detection of the interface residues , and further increase the docking success rates . The preliminary results in this proof-of-concept study showed that TM is a promising approach to protein docking , with its utility increasing along with the rapidly growing amount of publicly available information on protein complexes . | Protein interactions are central for many cellular processes . Physical characterization of these interactions is essential for understanding of life processes and applications in biology and medicine . Because of the inherent limitations of experimental techniques and rapid development of computational power and methodology , computer modeling is a tool of choice in many studies . Publicly available information from biomedical research is readily accessible on the Internet , providing a powerful resource for modeling of proteins and protein complexes . A major paradigm shift in modeling of protein complexes is emerging due to the rapidly expanding amount of such information , which can be used as modeling constraints . Text mining has been widely used in recreating networks of protein interactions , as well as in detecting small molecule binding sites on proteins . Combining and expanding these two well-developed areas of research , we applied the text mining to physical modeling of protein complexes ( protein docking ) . Our procedure retrieves published abstracts on a protein-protein interaction and extracts the relevant information . The results show that correct information on binding can be obtained for about half of protein complexes . The extracted constraints were incorporated in a modeling procedure , significantly improving its performance . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"Discussions"
] | [] | 2015 | Text Mining for Protein Docking |
Human T-lymphotropic virus type 1 ( HTLV-1 ) -associated myelopathy/tropical spastic paraparesis ( HAM/TSP ) is a rare chronic neuroinflammatory disease . Since the disease course of HAM/TSP varies among patients , there is a dire need for biomarkers capable of predicting the rate of disease progression . However , there have been no studies to date that have compared the prognostic values of multiple potential biomarkers for HAM/TSP . Peripheral blood and cerebrospinal fluid ( CSF ) samples from HAM/TSP patients and HTLV-1-infected control subjects were obtained and tested retrospectively for several potential biomarkers , including chemokines and other cytokines , and nine optimal candidates were selected based on receiver operating characteristic ( ROC ) analysis . Next , we evaluated the relationship between these candidates and the rate of disease progression in HAM/TSP patients , beginning with a first cohort of 30 patients ( Training Set ) and proceeding to a second cohort of 23 patients ( Test Set ) . We defined “deteriorating HAM/TSP” as distinctly worsening function ( ≥3 grades on Osame's Motor Disability Score ( OMDS ) ) over four years and “stable HAM/TSP” as unchanged or only slightly worsened function ( 1 grade on OMDS ) over four years , and we compared the levels of the candidate biomarkers in patients divided into these two groups . The CSF levels of chemokine ( C-X-C motif ) ligand 10 ( CXCL10 ) , CXCL9 , and neopterin were well-correlated with disease progression , better even than HTLV-1 proviral load in PBMCs . Importantly , these results were validated using the Test Set . As the CSF levels of CXCL10 , CXCL9 , and neopterin were the most strongly correlated with rate of disease progression , they represent the most viable candidates for HAM/TSP prognostic biomarkers . The identification of effective prognostic biomarkers could lead to earlier detection of high-risk patients , more patient-specific treatment options , and more productive clinical trials .
Human T-lymphotropic virus type 1 ( HTLV-1 ) is a human retrovirus associated with persistent infection of T-cells [1] . While the majority of HTLV-1-infected individuals remain asymptomatic , approximately 2 . 5–5% develop an aggressive T-cell malignancy , termed adult T-cell leukemia ( ATL ) [2] , [3] and 0 . 3–3 . 8% develop a serious chronic neuroinflammatory disease , termed HTLV-1-associated myelopathy/tropical spastic paraparesis ( HAM/TSP ) [4]–[6] . Aside from Japan , endemic areas for this virus and the associated disorders are mostly located in developing countries in the Caribbean , South America , Africa , the Middle East , and Melanesia [7] , [8] , which may explain why these conditions have remained ill-defined and virtually untreatable for so long [9] . HAM/TSP is characterized by unremitting myelopathic symptoms such as spastic paraparesis , lower limb sensory disturbance , and bladder/bowel dysfunction [10] , [11] . Although the symptoms of HAM/TSP have been well documented for quite some time , the rate at which these symptoms progress has only recently become a point of interest . The clinical course of HAM/TSP has classically been described very simply as insidious onset and continuous progression [12] , but recent reports have hinted at a more complex , heterogeneous pool of patients with differing clinical needs . Recent studies have shown that although HAM/TSP usually progresses slowly and without remission as per the classical description , there is a subgroup of patients whose conditions decline unusually quickly and who may be unable to walk within two years of onset and another subgroup whose conditions decline unusually slowly and who may only display very mild symptoms [13]–[15] . It is only logical that these patients should receive treatments tailored to suit their individual needs rather than identically aggressive treatments . Unfortunately , clinicians are currently only able to distinguish between these different groups by observing the way a patient's disease progresses over time , usually years; clinicians often decide to treat the patients immediately and identically rather than wait and allow the disease to progress further . Therein lies the dire need for biomarkers with the power to forecast the rate and extent of disease progression and enable clinicians to make more accurate prognoses and prescribe the most appropriate and effective treatments in a timely manner . Several candidate prognostic biomarkers with elevated levels in HAM/TSP patients have already been identified in the peripheral blood and cerebrospinal fluid ( CSF ) . In the peripheral blood , such candidates include the HTLV-1 proviral load in peripheral blood mononuclear cells ( PBMCs ) and serum levels of the soluble IL-2 receptor ( sIL-2R ) [16] , [17] . The level of neopterin in the CSF has been reported to be a useful parameter for detecting cell-mediated immune responses in the spinal cord of HAM/TSP patients and the CSF anti-HTLV-1 antibody titer has been shown to be associated both with CSF neopterin levels and the severity of clinical symptoms [18]–[20] . In addition , several cytokines have been detected in the CSF and/or spinal cord of HAM/TSP patients , including interleukin ( IL ) -1β , granulocyte-macrophage colony-stimulating factor ( GM-CSF ) , interferon ( IFN ) -γ , and tumor necrosis factor ( TNF ) -α [21]–[24] . Some chemokines , such as chemokine ( C-X-C motif ) ligand ( CXCL ) 9 , CXCL10 , and chemokine ( C-C motif ) ligand ( CCL ) 5 , have been shown to be substantially elevated in both the blood and the CSF with respect to asymptomatic carriers ( ACs ) or patients with other neurological diseases such as multiple sclerosis [25]–[28] . This is the first study to compare the adequacies of several of these candidate biomarkers for forecasting the rate of disease progression . We hypothesized the existence of biomarkers capable of differentiating stable and deteriorating HAM/TSP patients . In this retrospective study , a preliminary experiment was first conducted to select the most promising candidate biomarkers by comparing blood and CSF levels in HAM/TSP patients and control subjects ( Figure S1 ) . Four candidate blood markers ( sIL-2R , CXCL9 , CXCL10 , and proviral load ) and five candidate CSF markers ( CXCL9 , CXCL10 , neopterin , cell count , and anti-HTLV-1 antibody titer ) were selected . To evaluate the relative effectiveness of these candidate biomarkers for predicting rate of disease progression , a classification system was created and HAM/TSP patients were designated as either deteriorating or relatively stable . The levels of candidate biomarkers were then compared between the two patient groups . In the current study , we identified three viable candidates for HAM/TSP prognostic biomarkers that could lead to more accurate prognoses and more prudent , patient-specific treatment plans .
The study was designed and conducted in accordance with the tenets of the Declaration of Helsinki . The protocol in this study was approved by the Ethics Review Committee of St . Marianna University School of Medicine ( No . 1646 ) . Prior to the collection of blood or CSF samples , all subjects gave written informed consent permitting the analysis of their samples for research purposes as part of their clinical care . Between April 2007 and February 2013 , we enrolled 53 HAM/TSP patients according to the inclusion and exclusion criteria shown in Table 1 , and divided them into two cohorts based on the chronological order of their doctor's visits: a 30-patient Training set and a 23-patient Test set . Demographics and clinical characteristics of the Training set and Test set are shown in Table 2 and Table 3 , respectively . Between April 2007 and December 2009 , we enrolled 22 HTLV-1-infected ACs as control subjects for blood analysis and eight HTLV-1-infected subjects ( seven ACs , one patient with smoldering ATL ) as control subjects for CSF analysis according to the inclusion and exclusion criteria shown in Table 1 . These two groups were not mutually exclusive; some ACs donated both blood and CSF to this study . Demographics of control subjects as compared to the HAM/TSP patients are shown in Table S1 . Blood and/or CSF samples were obtained within a one-hour window for each subject . Peripheral blood samples were collected in heparin-containing blood collection tubes and serum-separating tubes . Plasma and PBMCs were obtained from the former tubes and serum was obtained from the latter . PBMCs were isolated with standard procedures using Pancoll® density gradient centrifugation ( density 1 . 077 g/mL; PAN-Biotech GmbH , Aidenbach , Germany ) . Plasma and serum samples were stored at −80°C until use . CSF was collected in polypropylene tubes . A small amount of CSF was used for routine laboratory tests , which included total protein , cell count , and IgG level . The remaining CSF was aliquoted into cryotubes and stored at −80°C until undergoing further analysis . All tests in this study were performed on samples from these frozen stocks . The serum concentration of sIL-2R was determined using an ELISA ( Cell Free N IL-2R; Kyowa Medex Ltd . , Tokyo , Japan ) . HTLV-1 proviral load was measured using real-time PCR , following DNA extraction from PBMCs , as previously described [29]–[31] . Plasma levels of IL-1β , TNF-α , and IFN-γ were measured using a cytometric bead array ( CBA ) ( BD Biosciences , Franklin Lakes , NJ USA ) , which was used according to the manufacturer's instructions . Plasma concentrations of CXCL9 , CXCL10 , CXCL11 , and CCL5 were also measured using a CBA ( BD Biosciences ) . CSF cell count was determined using the Fuchs–Rosenthal chamber ( Hausser Scientific Company , Horsham PA USA ) . Total protein and IgG levels in the CSF were measured using a pyrogallol red assay and a turbidimetric immunoassay , respectively . The anti-HTLV-1 antibody titer was determined using the gelatin particle agglutination test ( Serodia-HTLV-1; Fujirebio , Tokyo , Japan ) . CSF concentration of sIL-2R was determined using an ELISA ( Cell Free N IL-2R; Kyowa Medex ) . CSF neopterin level was measured using high-performance liquid chromatography . IFN-γ and six chemokines ( CXCL9 , CXCL10 , CXCL11 , CCL3 , CCL4 , and CCL5 ) were measured using a CBA ( BD Biosciences ) . The CSF concentrations of three chemokines ( CCL17 , CCL20 , and CCL22 ) and IL-17A were measured using commercially available ELISA kits ( CCL17 , CCL20 , and CCL22: TECHNE/R&D Systems , Minneapolis , MN USA; IL-17A: Gen-Probe , San Diego , CA USA ) . All assays were conducted according to the respective manufacturers' instructions . The 53 total HAM/TSP patients without any history of HAM/TSP-targeting treatments were interviewed using a questionnaire ( Figure S2 ) to determine the changes in Osame's Motor Disability Score ( OMDS ) over time ( Figure S3 ) . OMDS is a standardized neurological rating scale as a measure of disability [10] ( Figure S1 ) . Based on the changes in OMDS , “deteriorating cases” and “stable cases” were identified in both the Training set and Test set patient cohorts . Patients with deteriorating HAM/TSP were defined as those whose OMDS worsened ≥3 grades over four years and patients with stable HAM/TSP were defined as those whose OMDS remained unchanged or worsened 1 grade over four years . Patients whose OMDS worsened 2 grades over four years were excluded from the patient cohort in order to create a larger gap between the deteriorating and stable patient groups . GraphPad Prism 5 ( GraphPad Software , Inc . , La Jolla , CA USA ) was used to plot graphs and perform statistical analyses . Differences between the two subject groups were tested using the Mann-Whitney U-test . Receiver operating characteristic ( ROC ) analysis was performed to examine the sensitivity and specificity of individual biomarkers . For the ROC analyses , an area under the ROC curve ( AUC ) of 1 . 0 was used to represent a perfect test with 100% sensitivity and 100% specificity , whereas an area of 0 . 5 was used to represent random discrimination . Spearman's rank correlation test was employed to investigate the correlation between the four CSF markers ( CXCL10 , CXCL9 , neopterin , and cell count ) and the proviral load in PBMCs . To compare the four CSF markers between three groups ( HTLV-1-infected control , n = 8; stable HAM/TSP , n = 25; and deteriorating HAM/TSP , n = 20 ) , we used the Kruskal–Wallis test followed by Dunn's post-hoc tests . P-values<0 . 05 were considered statistically significant .
In order to identify candidate blood markers for HAM/TSP , the concentrations of IL-1β , TNF-α , and IFN-γ were measured in plasma samples from four ACs and four HAM/TSP patients . Plasma levels of IL-1β and TNFα were below the detection limits ( <2 . 3 pg/mL and <1 . 2 pg/mL , respectively ) except in one patient with HAM/TSP . Plasma IFN-γ levels showed no significant differences between ACs and HAM/TSP patients ( median 10 . 4 pg/mL and 13 . 9 pg/mL , respectively ) . Therefore , these quantities were not measured in additional samples ( Figure S1 ) . The proviral DNA load in PBMCs , serum sIL-2R , and plasma levels of the chemokines CXCL9 , CXCL10 , CXCL11 , and CCL5 were also measured in 22 ACs and 30 HAM/TSP patients without any history of immunomodulating treatments , including corticosteroids , IFN-α , and immunosuppressive drugs . The results revealed that serum levels of sIL-2R , plasma levels of CXCL10 and CXCL9 , and proviral DNA load in PBMCs were markedly higher in HAM/TSP patients compared to ACs ( p≤0 . 0001 , Figure 1A ) . These quantities were then compared using ROC analysis to determine which parameters were superior markers for HAM/TSP . From the results of the ROC analysis , we determined that serum sIL-2R and plasma CXCL10 had the highest potential for distinguishing HAM/TSP patients from ACs with high sensitivity and specificity ( area under the ROC curve [AUC]>0 . 9 ) , followed by plasma CXCL9 and HTLV-1 proviral load in PBMCs ( 0 . 8<AUC<0 . 9 ) ( Figure 1B ) . Thus , four candidate blood biomarkers were selected for further investigation: serum sIL-2R , plasma CXCL10 , plasma CXCL9 , and HTLV-1 proviral load in PBMCs . In order to identify candidate CSF markers for HAM/TSP , elevated levels of various potential markers were screened for in CSF samples from HAM/TSP patients . CSF IL-17A was detectable ( >3 . 0 pg/mL ) in only one of eight HAM/TSP patients screened ( including six deteriorating-type patients ) , and the level in this one patient ( deteriorating-type ) was negligible ( 4 . 0 pg/mL ) . CSF IFN-γ was detectable ( >1 . 8 pg/mL ) in only 3 of 10 HAM/TSP patients screened ( six deteriorating patients ) , and the levels in all three were negligible ( range 3 . 3–4 . 2 pg/mL ) . Therefore , these cytokines were not measured in additional patients . Total protein , cell count , IgG , neopterin , sIL-2R , and nine chemokines ( CXCR3 ligands: CXCL9 , CXCL10 , and CXCL11; CCR5 ligands: CCL3 , CCL4 , and CCL5; CCR4 ligands: CCL17 and CCL22; CCR6 ligand: CCL20 ) were also measured in the CSF of 30 untreated HAM/TSP patients and in eight HTLV-1-infected control subjects ( seven ACs and one patient with smoldering ATL ) . The results indicated that CSF levels of CXCL10 , neopterin , and CXCL9 were remarkably higher in HAM/TSP patients compared to control subjects ( p<0 . 0001 overall , Figures 2A and S4 ) and that CSF levels of cell count and CCL5 were less so but still significantly higher ( p = 0 . 0019 and p = 0 . 0119 , respectively; Figure 2A ) . By contrast , there were no differences in the CSF levels of IgG and total protein between HAM/TSP patients and control subjects , and CSF sIL-2R levels were only detectable in a single HAM/TSP patient ( data not shown ) . ROC analysis showed that the CSF levels of CXCL10 , neopterin , CXCL9 , and CSF cell count could be used to relatively accurately distinguish HAM/TSP patients from control subjects ( AUC>0 . 8 ) ( Figure 2B ) . Therefore , these four CSF markers were selected as candidates for further investigation . It should be noted that the sensitivity of CSF cell count was very low ( 36 . 7% ) when compared to the other three: CXCL10 ( 83 . 3% ) , CXCL9 ( 86 . 7% ) , and neopterin ( 76 . 7% ) ( Figure S5 ) . In short , we selected nine markers: eight markers chosen based on the analyses described above and CSF anti-HTLV-1 antibody titer , which is a known diagnostic marker for HAM/TSP . To determine which biomarkers were associated with HAM/TSP disease progression , the levels of these nine markers were compared between the deteriorating and stable HAM/TSP patient groups ( see Methods for definitions of deteriorating and stable ) . The results revealed that all five CSF markers were significantly higher in the deteriorating group compared to the stable group ( Figure 3A ) , but that none of the four blood markers , including proviral load , were significantly different between the two groups . The deteriorating group included three patients with particularly rapidly progressive HAM/TSP , defined as those who had been confined to wheelchairs ( OMDS: ≥ grade 6 ) within two years after the onset of symptoms [13] , [14] ( black circles in Figures 3A and S3B ) . These rapid progressors exhibited high levels of the CSF markers and high proviral loads . ROC analysis revealed that the levels of the CSF markers ( CXCL10 , CXCL9 , neopterin , and cell count ) , but not anti-HTLV-1 antibody titer , distinguished clearly between patients with deteriorating HAM/TSP and stable HAM/TSP ( AUC>0 . 8 , Figure 3B ) . To validate the results obtained using the Training Set , the same nine markers were compared between deteriorating and stable patients using the Test Set ( a second cohort of 23 HAM/TSP patients that had not undergone HAM/TSP-targeting treatment ) . As shown in Figure 4A , the results indicated that the levels of five CSF markers , proviral load in PBMCs , and serum sIL-2R were significantly higher in deteriorating cases than in stable cases . Among them , CSF levels of CXCL10 , CXCL9 , neopterin , and CSF cell count exhibited particularly high sensitivities and specificities for detecting the deteriorating HAM/TSP cases in the Test set as well as Training set ( AUC>0 . 8 , Figures 4B and S1 ) . The demographics of the HAM/TSP patients versus the control subjects for both the blood tests and CSF analyses were compared and evaluated for statistical significance ( Table S1 ) . There were no significant differences in age or gender distribution between the HAM/TSP patients and either control subject group . Similarly , the demographic and clinical characteristics of stable versus deteriorating HAM/TSP subjects in both the Training and Test sets are shown in Tables 2 and 3 , respectively . There were no significant differences in age or gender distribution among either set , but deteriorating patients in both sets were significantly older at disease onset and had been living with the disease for shorter periods of time . Deteriorating patients in the Training set scored higher OMDS values than their stable counterparts ( p<0 . 01 ) , but there was no such significant difference in the Test set . To investigate the potential influence of disease duration as a secondary variable , a new test group was created containing only those patients for whom the disease onset date was 7–13 years prior to the sample collection day . Patients fitting this criterion were selected from the 53 total available from both the Training and Test sets: eight stable patients and ten deteriorating patients; we confirmed that there was no significant difference in disease duration between these two groups . The results remained consistent with our previous findings: CSF CXCL10 , CXCL9 , and neopterin were all elevated in deteriorating patients with respect to stable patients ( p<0 . 01 , Figure 5 ) . Four stable HAM/TSP patients were left completely untreated and followed for a period of three to five years . Within this time , one patient rose one grade on the OMDS scale , and the other three experienced no change in OMDS grade at all . The levels of CSF CXCL10 and neopterin remained consistently low over time ( Figure S6 ) .
To date , there have been few well-designed studies that have evaluated the relationship between biomarkers and HAM/TSP disease progression . In a previous retrospective study with 100 untreated HAM/TSP patients , a significant association was demonstrated to exist between higher HTLV-1 proviral load in PBMCs and poor long-term prognosis; however , the predictive value of high proviral load appeared to be too low to qualify it as a marker for disease progression in clinical practice [32] . Here we conducted a retrospective study to compare for the first time the relationships of PBMC proviral load and several inflammatory biomarker candidates to disease progression in untreated HAM/TSP patients . In this study , elevated CSF cell count , neopterin concentration , and CSF levels of CXCL9 and CXCL10 were well-correlated with disease progression over the four year period under study , better even than HTLV-1 proviral load in PBMCs ( Figures 3 and 4 ) . As CSF pleocytosis , CSF CXCL10 , CSF CXCL9 , and CSF neopterin are known indicators of inflammation in the central nervous system [33] , [34] , our findings indicate that the rate of HAM/TSP progression is more closely reflected by the amount of inflammatory activity in the spinal cord than by the PBMC proviral load . However , we also found a significant correlation between PBMC proviral load and the levels of the CSF markers identified in this study ( Figure S7 ) , indicating that a higher PBMC proviral load does indeed suggest more inflammation in the spinal cord and therefore a poorer long-term prognosis . These findings are consistent with the theory that HAM/TSP is the result of an excess of inflammatory mediators caused by the presence of HTLV-1-infected T-cells [35]–[37] . The HTLV-1 proviral load in the CSF as well as the ratio of the proviral load in the CSF to that in PBMCs have been reported to be effective for discriminating HAM/TSP patients from ACs or multiple sclerosis patients infected with HTLV-1 [38] , [39] . Some researchers have suggested that these values might be associated with the rate of disease progression , but there has been only one small cohort study and one case report investigating this point , and so the significance of this experimental evidence is still questionable [40] , [41] . In addition to statistical validation with multiple , larger cohorts , it would also be beneficial to use precise definitions for progressive versus stable patients , as we have done in this study . Although the volume of CSF available per sample was too limited to measure CSF proviral load in the present study , we plan to incorporate CSF proviral load in a future prospective study and compare its usefulness to that of other biomarker candidates . From our results , we concluded that of the potential biomarkers under study , CXCL10 , CXCL9 , and neopterin are the most fit for determining the level of spinal cord inflammation , and thus the most fit for predicting disease progression in HAM/TSP patients . Although the CSF cell count is an easily measurable inflammatory marker , it is not sensitive enough to reliably detect the level of spinal cord inflammation . Numerous patients with CSF cell counts within the normal range exhibited high levels of other inflammatory markers , such as neopterin and CXCL10 ( Figure S5 ) . In fact , it has been reported that CSF pleocytosis is present in only approximately 30% of HAM/TSP patients [42] . Furthermore , in our study , there was no significant difference in CSF cell count between the control subjects and the stable HAM/TSP patients ( Figure S8 ) . We also explored the possibility of combining multiple biomarkers via multiple logistic regression to form a combination more sensitive and specific than individual markers , but the results indicated that there is not much to be gained from combinations ( data not shown ) . While there were no significant demographic differences between subject groups , the clinical characteristics of stable versus deteriorating HAM/TSP patients of course differed widely ( Tables 2 , 3 , and S2 ) . We confirmed the already well-reported statistic that deteriorating patients experience HAM/TSP onset relatively late in life [12] , [14] , [20] ; our data also reflected the short disease duration expected of deteriorating patients , who by definition progress through the disease more rapidly than their stable counterparts . As patients in all groups were of similar age at sample collection , the significant difference in age of onset should not have any impact on our findings . However , it was necessary to consider the possibility that those patients in a later stage of the disease ( i . e . those listed with longer disease durations ) might possess elevated or diminished biomarker levels regardless of rate of disease progression . We confirmed that this difference in disease duration was not a confounding factor in our selection of candidate biomarkers by comparing stable and deteriorating HAM/TSP patients with similar disease durations ( 7–13 years ) , and we were able to obtain results consistent with our earlier findings ( Figure 5 ) . Finally , the OMDS values for the stable and deteriorating patient groups in the Test set were perfectly identical , eliminating the need to consider the possibility that the biomarkers could have been elevated according to disease severity regardless of rate of progression . The main limitation of our retrospective study is that our samples were collected from patients at the end of the four year period during which the extent of progression was analyzed as opposed to the beginning of the four year period , which would have been optimal for directly measuring their prognostic powers . Of course , the patients with severe HAM/TSP symptoms began undergoing treatment soon after sample collection , rendering any observations on disease course after sample collection un-useable for analysis in this study . While this situation is non-ideal , we hypothesize that biomarker levels in a given patient do not substantially change over a few years' time . We were actually able to monitor the biomarker levels of four untreated HAM/TSP patients over 3–5 years , and the levels remained relatively stable in all four subjects over time ( Figure S6 ) , supporting our hypothesis . However , these were all stable HAM/TSP patients ( hence the lack of treatment ) , and so we cannot rule out the possibility that biomarker levels in untreated deteriorating patients may dramatically rise , fall , or fluctuate . The results of the analysis of patients with similar disease durations ( Figure 5 ) also support our hypothesis that disease duration is not an important determinant of biomarker levels , but it is of course not conclusive . We expect that a prospective study in the future will reveal the answer to this question . The results of this study indicate that CXCL9 and/or CXCL10 may play a key role in the pathogenesis of HAM/TSP by recruiting more inflammatory cells to the spinal cord lesions . In this study , we measured the levels of the chemokines in the CSF that might play a part in inducing the migration of T-helper ( Th ) cells . CD4+ Th cells differentiate from naïve T-cells to members of the Th subset ( e . g . , Th1 , Th2 , Th17 , or Treg cells ) , and each one expresses its own characteristic chemokine receptors [43] . Usually , Th1 cell express CCR5/CXCR3 receptors , Th2 and Treg cells express CCR4 , and Th17 express CCR6 . Interestingly , CCR4 ligands ( CCL17 and , CCL22 ) and the CCR6 ligand ( CCL20 ) were not detected in the CSF of HAM/TSP patients . Moreover , of the CCR5 ligands , only CCL5 was elevated , but only slightly , and there was no association with rate of disease progression . Of the CXCR3 ligands , only CXCL9 and CXCL10 were correlated with the rate of disease progression . These results show that the pathology of HAM/TSP is unique among immune disorders in that , unlike other inflammatory disorders such as multiple sclerosis or rheumatoid arthritis that exhibit Th17 as well as Th1 involvement , the chemokine involvement in HAM/TSP is Th1-dominant . In a previous study , cytokines produced by HTLV-1-infected T-cells in HAM/TSP patients were analyzed , and the results showed that IFN-γ was elevated and IL-17 reduced [43] , [44] . Taken together , the results of these studies indicate that the characteristics of HTLV-1-infected T-cells themselves may be responsible for the Th1-dominant chemokine production observed in HAM/TSP . Also , these results suggest that the CXCR3-ligand ( CXCL9 and CXCL10 ) interactions play an important role in the pathophysiology of HAM/TSP . Recently it was established that these CXCR3-ligand interactions are extremely important for the pathogenesis of several neurological disorders [33] . Therefore , future research on the significance of these interactions in the pathogenic process of HAM/TSP will be important for clarifying the suitability of CXCL9 and CXCL10 as biomarkers or therapeutic targets . In conclusion , in this retrospective study , we have demonstrated that CSF levels of CXCL10 , CXCL9 , and neopterin are promising candidate prognostic biomarkers for HAM/TSP . These biomarkers may provide a means for the early identification of patients at increased risk of debilitating disease progression , those that may need anti-inflammatory therapies to limit or prevent this , and for evaluating the efficacy of such therapies . This initial identification of prognostic biomarkers for HAM/TSP should be followed by a future multicenter prospective clinical study . | HTLV-1-associated myelopathy/tropical spastic paraparesis ( HAM/TSP ) is a rare neurodegenerative disease caused by infection with human T-lymphotropic virus type 1 ( HTLV-1 ) . HTLV-1 infects 10–20 million people worldwide , and , depending on the region , 0 . 25–3 . 8% of infected individuals develop HAM/TSP . As the disease progresses , chronic inflammation damages the spinal cord and lower limb and bladder function gradually decline . In the worst cases , even middle-aged patients can become perpetually bedridden . Today , there are treatments that may alleviate the symptoms to a certain degree , but there is no cure that can halt disease progression , and there are no known biomarkers to indicate the level and speed of disease progression . In this study , we successfully identified three promising candidate biomarkers . We believe that the use of these biomarkers could lead to more accurate prognoses and more prudent , patient-specific treatment plans . We not only hope that these biomarkers are sensitive enough to use as selection criteria for clinical trials , but also that measurements of these biomarkers can be used to accurately evaluate drug effectiveness . In short , the biomarkers we identified have the potential to help more effectively treat current HAM/TSP patients and to pave the way for new drugs to potentially cure future HAM/TSP patients . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2013 | CSF CXCL10, CXCL9, and Neopterin as Candidate Prognostic Biomarkers for HTLV-1-Associated Myelopathy/Tropical Spastic Paraparesis |
Over 1 . 1 billion people worldwide are at risk for lymphatic filariasis ( LF ) , and the global burden of LF-associated lymphedema is estimated at 16 million affected people , yet country-specific estimates are poor . A house-to-house morbidity census was conducted to assess the burden and severity of lymphedema in a population of 1 , 298 , 576 persons living in the LF-endemic district of Khurda in Odisha State , India . The burden of lymphedema in Khurda is widespread geographically , and 1 . 3% ( 17 , 036 ) of the total population report lymphedema . 51 . 3% of the patients reporting lymphedema were female , mean age 49 . 4 years ( 1–99 ) . Early lymphedema ( Dreyer stages 1 & 2 ) was reported in two-thirds of the patients . Poisson regression analysis was conducted in order to determine risk factors for advanced lymphedema ( Dreyer stages 4–7 ) . Increasing age was significantly associated with advanced lymphedema , and persons 70 years and older had a prevalence three times greater than individuals ages 15–29 ( aPR: 3 . 21 , 95% CI 2 . 45 , 4 . 21 ) . The number of adenolymphangitis ( ADL ) episodes reported in the previous year was also significantly associated with advanced lymphedema ( aPR 4 . 65 , 95% CI 2 . 97–7 . 30 ) . This analysis is one of the first to look at potential risk factors for advanced lymphedema using morbidity census data from an entire district in Odisha State , India . These data highlight the magnitude of lymphedema in LF-endemic areas and emphasize the need to develop robust estimates of numbers of individuals with lymphedema in order to identify the extent of lymphedema management services needed in these regions .
Lymphatic filariasis ( LF ) is a mosquito-borne parasitic infection that damages the lymphatic system and can cause chronic and debilitating swelling of the limbs known as lymphedema or , in its more advanced form , elephantiasis . Approximately 1 . 1 billion people are at risk for LF in more than 73 countries worldwide [1] , 600 million of whom reside in 250 districts in India accounting for over 40% of the global LF burden [2] . An estimated 50% of India’s LF endemic population resides in northeastern India in areas of poverty [3–5] . Among the LF endemic regions , Odisha state ( Fig 1 ) , one of the poorest states in India , was recognized historically as one of the most highly endemic and remains as such [4 , 6 , 7] . Similar to the World Health Organization’s ( WHO ) global goal to eliminate LF as a public health problem by 2020 , India has set a national goal of LF elimination which focuses on two pillars: 1 ) interruption of disease transmission through mass drug administration ( MDA ) and 2 ) reduction of LF-associated morbidity for patients already impacted by clinical disease . MDA campaigns have been occurring in India since 1997 and have progressed well with some areas having demonstrated interruption of transmission [3 , 8 , 9] . However , it is recognized that MDA does not treat the chronic clinical manifestations , specifically lymphedema , elephantiasis and hydrocele , seen in some LF-infected individuals [2 , 10] . Therefore , for individuals with lymphedema and elephantiasis , lymphedema management programs are necessary to mitigate the symptoms associated with lymphedema , and to prevent the development of episodes of adenolymphangitis ( ADL ) known to lead to the worsening of lymphedema [11 , 12] . Accurate estimates of lymphedema and hydrocele patients are needed by national LF elimination programs to inform service provision , yet robust assessments have been rarely performed [10 , 13] . An emphasis needs to be placed on obtaining accurate patient estimates in order to properly scale up morbidity management and disability prevention services for LF patients with clinical disease . Enumeration of individuals with early lymphedema ( Dreyer stages 1 and 2 ) is particularly valuable as these individuals have the potential to gain the most benefit , compared to those with later stages of lymphedema , from lymphedema management programs [14] . Although outside the scope of this survey , estimates of number of patients with hydrocele are also necessary for country programs to ensure that sufficient surgical resources are available for these individuals . The objective of this analysis was to describe and analyze data obtained from a thorough method of enumerating individuals with lower limb swelling in a highly-LF endemic district in Odisha State , India .
Permission for the morbidity census was obtained from the Odisha State Department of Health and Family Welfare . The morbidity census was determined to be program evaluation under CDC policy . The morbidity census took place in rural and peri-urban areas of Khurda district in Odisha state ( Fig 2 ) over a four month period in 2005 ( approximate population 1 . 3 million ) . All individuals living in Khurda district yet outside the capital city of Bhubaneswar were included in the morbidity census . Prior to this census it was well established that Khurda was highly endemic for Bancroftian filariasis; however , more robust estimates of individuals with lymphedema were needed in preparation for the initiation of a district-wide community-based lymphedema management program [6 , 8 , 11 , 12 , 15] . In 2005 , a local non-governmental organization ( NGO ) , Church’s Auxiliary for Social Action ( CASA ) in partnership with approximately forty other local NGO’s working in Khurda district , conducted a house-to-house morbidity census to enumerate the number of individuals with lymphedema/elephantiasis in the district . Approximately 200 community health workers and NGO volunteers were trained by government workers and other filarial experts in patient enumeration and clinical diagnosis in accordance with the Dreyer manual on basic lymphedema management and the Indian government’s manual for Operational Guidelines on Elimination of Lymphatic Filariasis [16 , 17] . Staging was not validated for all the patients enumerated , but a small sample of patients did undergo validation of staging as part of an evaluation of the program . The details on these patients are discussed in the following publications [16 , 17] . Trained staff documented persons reporting lymphedema in each household , and staged the degree of lymphedema according to the Dreyer staging system [16] . Communication with individuals occurred in the local language , Oriya . Standardized questionnaires were administered by trained staff and data on basic demographics and risk factors for lymphedema were collected for all lymphedema patients . Additionally , trained staff asked individuals with lymphedema to report the occurrence and number of ADL episodes they may have experienced at any time during the last twelve months . We recognize that recalling the number of ADL episodes over one year may be difficult for patients , particularly those who have had numerous ADL episodes and subsequent work we have done would recommend asking about ADL episodes over a shorter period of time , such as one month [11 , 12] . Data on individuals reporting lymphedema were recorded on paper registers by NGO volunteers . Over a period of two months following completion of the morbidity census , data were double data entered into Microsoft Excel by CASA data management staff . The morbidity census enumerated the number of individuals with lymphedema of the leg in Khurda district by village . All households in the district were asked about the presence of individuals with lymphedema in the household . Data collected on individuals with lymphedema included gender , age , caste , religion , level of education , village of residence , rural/peri-urban residence , presence of leg swelling , lymphedema stage ( Dreyer scale ) , number of ADL episodes in the past year , and presence of another family member with lymphedema . For the purposes of this paper , lymphedema refers to swelling observed in the lower limbs—right or left leg , as this is the most commonly observed location for filarial-associated lymphedema . Lymphedema severity was established based on the seven stage classification system developed by Dreyer , et al . [16] . ADL episodes were defined as any period of pain , redness , and/or swelling of the affected leg , which may have been accompanied by fever and/or chills [18 , 19] .
Approximately 68% of the villages listed in the CASA census were correlated to the Indian census , the remaining 32% either used smaller naming references than the Indian census ( such as hamlet or ward ) or were referred to by a different name at the time of the CASA census . Fig 3 illustrates that persons with lymphedema are present in over 50% of villages in Khurda district . Furthermore , Fig 4 shows the distribution of individuals with early lymphedema ( stage 1 and stage 2 ) by village throughout the district . Mapping of lymphedema prevalence demonstrates the widespread burden of this chronic condition throughout Khurda district . The geographic distribution of individuals with lymphedema including those with early stage ( Dreyer stages 1 & 2 ) is widespread throughout the district . Data on 17 , 036 individuals reporting lymphedema were recorded . 51 . 3% of the individuals were female , mean age 49 . 4 years ( Table 1 ) . Two-thirds ( 67% ) of individuals reported early lymphedema ( Dreyer stages 1 & 2 ) , 22% reported moderate lymphedema ( Dreyer stage 3 ) , and 10% reported advanced lymphedema ( Dreyer stages 4–7 ) . Mean lymphedema stage was 2 . 1 ( range 1–7 ) . The majority of the study population ( 83 . 9% ) reported a history of ADL episodes; mean number of ADL episodes per year was 1 . 05 ( range 0–3 ) . Additionally , 21 . 5% of the individuals reported family members with lymphedema as well . In univariate analysis , the prevalence of lymphedema among females was less than the prevalence of lymphedema among males ( PR: 0 . 84 , 95% CI 0 . 77 , 0 . 93 ) ( Table 2 ) . Additionally , there was a significantly higher prevalence of advanced lymphedema with increasing age ( p-value <0 . 0001 ) . When compared with persons aged 15–29 years , persons 70 years and older had more than a three-fold higher prevalence of lymphedema ( PR: 3 . 47 , 95% CI 2 . 56 , 4 . 71 ) . Individuals who reported two ADL episodes in the previous year were more than five times as likely as individuals who reported no ADL episodes in the previous year to have advanced lymphedema . Interestingly , individuals who reported having a family member with lymphedema had on average a 30% greater prevalence of advanced lymphedema compared with persons who did not report a family member with lymphedema ( PR: 1 . 30 , 95% CI 1 . 12 , 1 . 49 ) . The results from the multivariate analysis support the associations observed in univariate analysis ( Table 2 ) .
This house-to-house morbidity census provides an example of a comprehensive method for estimation of individuals with lymphedema in a LF-endemic region . Conducted at the district level in India , with a population of more than 1 . 2 million people , this census demonstrates that enumeration of individuals with lymphedema is possible on a large scale . Our findings suggest that the prevalence of lymphedema in Odisha , India in 2005 was considerable , affecting 1 . 3% of the population . Previously , the prevalence of lymphedema was estimated to affect 3 . 4% of the population , while 2014 burden assessments estimate the number of lymphedema patients in Odisha State to be approximately 80 , 000 [15 , 22] . While an MDA campaign had occurred in Odisha in 1997 , consistent annual MDA in Odisha did not begin until 2004 . Therefore the prevalence of lymphedema calculated from the results of the morbidity census would not have been significantly impacted by MDA . Conducting the morbidity census as a house-to-house survey allowed for the creation of maps detailing the distribution and prevalence of lymphedema at the village level , which had not previously been achieved . The greatest proportion ( 67 . 5% ) of participants reported early lymphedema , the burden of which is widespread throughout the district ( Fig 4 ) . The significant number of individuals with early stage lymphedema highlights the large percent of the population that may benefit from the establishment of lymphedema management programs that target modifiable disease risk factors such as ADL episodes [1 , 13 , 23 , 24] . Identifying that the prevalence of lymphedema is geographically widespread emphasizes the need for scaling up lymphedema management services and training health care workers throughout the district . Furthermore , the results of this morbidity census informed communities of the burden of lymphedema in their villages and identified the need for lymphedema management resources in individual communities . The census was necessary in order to determine the number of human and financial resources needed to implement the community-based lymphedema management program in Khurda , specifically the number of NGO volunteers , commodities for lymphedema management , training materials , etc . Based on prior experiences with lymphedema management programs , CASA determined that one NGO volunteer was needed to train and follow-up with twenty lymphedema patients . In addition to providing mass drug administration ( MDA ) to prevent LF infection , the Indian national lymphatic filariasis elimination program has implemented an LF morbidity management and disability prevention program that involves training of health care professionals and of community health workers ( P . K . Srivastava , personal communication ) . The establishment of community-based lymphedema management programs in Odisha State has demonstrated success in the following areas: prevention of ADL episodes , halting lymphedema progression , improving quality of life , improving patient perceptions of their disability , and improving work productivity for lymphedema patients living in this region [11 , 12] . The efforts of partnering NGOs and the response of the Indian national LF elimination program can serve as an example to other LF elimination programs in scaling up their morbidity management and disability prevention ( MMDP ) efforts . We recognize that a house-to-house morbidity census for LF-related disease may not be feasible for all LF-endemic districts given human and fiscal resource constraints . Other approaches to morbidity assessments have involved the use a community-led standard messaging system ( SMS ) reporting tools piloted in Malawi and Ghana , as well as a 30 cluster survey that was conducted to estimate the lymphedema and hydrocele burden in Bangladesh [25 , 26] . We recommend that further operational research be performed to develop more cost-effective tools for assessment of LF-related disease burden to inform countries preparing to scale up MMDP services . Analysis of the risk factors for advanced lymphedema provided further evidence to support the need for lymphedema management programs in this region . Among individuals aged 15 and older who reported lymphedema , older age ( 70 years and older ) was highly associated with an increased prevalence of advanced stage lymphedema; which is consistent with findings in other LF endemic areas , such as Haiti [24] . The increase in prevalence of lymphedema with increasing age corresponds to the fact that lymphedema symptoms often do not appear until years or even decades after an individual is first infected and they tend to worsen over time and with the number of ADL episodes a person experiences . In addition , this in-depth morbidity census confirmed the observation reported previously in smaller studies that advanced lymphedema is associated with the number of ADL episodes reported during the previous year [11–13] . These findings emphasize the opportunity for prevention of lymphedema progression and possibly disability through implementation of programs that reduce the severity and frequency of ADL episodes in LF-endemic regions . Quantifying the number of individuals with lymphedema and hydrocele through morbidity censuses such as this one can be used to calculate the economic impact of the disability associated with the chronic manifestations of LF infection . Numerous studies have evaluated the economic impact of the chronic manifestations of LF infection for both estimated financial loss and number of work days lost [2 , 27–30] . Ramaiah et al . [29 , 30] provided evidence for the significant costs associated with chronic manifestations as well as ADL episodes , including an estimated $842 lost annually to families needing to fund treatments and losing work days [28] . Additionally , calculations using average hours worked per day , average hours of work-time lost , and patient estimates for India determined that 1 , 098 million days of work were lost due to chronic LF infection in India annually [28] . The decrease in morbidity associated with lymphedema management programs has been shown to save on average 1 . 7–3 . 5 days of labor per patient resulting in significant financial savings for families and communities , especially in highly LF-endemic areas [11 , 12] . By calculating accurate estimates of those with chronic manifestations of LF infection , countries can better understand the local economic impact of LF- related disease . These data can inform interventions that effectively reduce this burden both at local and national levels as a complement to ongoing efforts focused on interruption of LF transmission . This analysis was subject to a number of limitations . It is uncertain whether all patients identified in the census had LF-associated lymphedema or lymphedema due to another cause given that no serologic testing was performed . However , LF is the leading cause of lymphedema in Odisha , and lymphedema management programs are needed for individuals affected by lymphedema regardless of the cause . Additionally , strengthening country health systems through the integration of LF lymphedema management programs with leprosy , Buruli ulcer and podoconiosis programs has been recommended by WHO in the aide-memoire [31] . Secondly , because of the stigma associated with this chronic manifestation of LF infection , it is possible that patients were unwilling or reluctant to divulge their status to surveyors . The results of this census , then , may underestimate the true burden of lymphedema . As the staff performing the census were not trained health care professionals , they did not perform extensive physical examinations of individuals with lymphedema which may have impacted the lymphedema staging leading to an overestimate of stage 1 patients given that stage 1 corresponds to any minor swelling that may not have been LF-related . Additionally , in this census there is the potential for recall bias when patients were asked to report the number of ADL episodes they experienced in the last year which may have been an underestimate given other publications that have reported higher numbers of ADL episodes within 30 day to 6 month time frames [24 , 26] . Lastly , the morbidity census and Indian census data were reported at different levels of region names . This explains why only 67% of the morbidity census data was able to be mapped , the remaining data included levels of residence that did not directly match the Indian census and due to the limitations of the time since the morbidity census occurred it was not feasible to reconcile these discordant villages . MDA campaigns have demonstrated great success in interrupting the transmission of LF , and as we continue to move forward towards the goal of elimination of LF as a public health problem countries will need to ensure morbidity management services are available for those patients already experiencing chronic sequelae associated with LF infection . In most LF-endemic countries the burden of filarial disease is poorly defined and is often underestimated . The data obtained from a morbidity census can provide program managers with a sense of the widespread geographic burden of disease . The methodology of this census can serve as an example for other LF-endemic countries seeking to enumerate lymphedema patients at the district and national level to meet the dossier requirements for morbidity management and disability prevention set by WHO for LF elimination . Scaling-up morbidity management programs is an essential component of global LF elimination as it was the recognition of the suffering of patients with chronic infection that inspired setting of the goal of global LF elimination in the first place . | The results presented in this article demonstrate the need to assess the clinical disease burden due to lymphatic filariasis ( LF ) in areas of the world where this disease occurs . There are effective strategies that can be implemented to reduce the suffering and morbidity associated with chronic filarial infections . These include lymphedema management programs that teach basic hygiene to decrease limb infections that contribute to lymphedema progression . These interventions should be implemented on a broader scale and can be integrated with other chronic disease prevention programs at subnational and national levels . However , knowledge of the scale and distribution of individuals suffering with lymphedema in countries with LF is required to decide where and how these health services should be implemented . Countries with LF are encouraged to conduct burden assessments to decide how best to implement lymphedema management programs . Providing health services for those with lymphedema is an important component of reducing LF as a public health problem . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"cancer",
"detection",
"and",
"diagnosis",
"medicine",
"and",
"health",
"sciences",
"lymphatic",
"mapping",
"pathology",
"and",
"laboratory",
"medicine",
"geographical",
"locations",
"tropical",
"diseases",
"india",
"census",
"parasitic",
"diseases",
"health",
"care",
... | 2016 | Evaluating the Burden of Lymphedema Due to Lymphatic Filariasis in 2005 in Khurda District, Odisha State, India |
Leptospirosis essentially affects human following contact with rodent urine-contaminated water . As such , it was mainly found associated with rice culture , recreational activities and flooding . This is also the reason why it has mainly been investigated in temperate as well as warm and humid regions , while arid zones have been only very occasionally monitored for this disease . In particular , data for West African countries are extremely scarce . Here , we took advantage of an extensive survey of urban rodents in Niamey , Niger , in order to look for rodent-borne pathogenic Leptospira species presence and distribution across the city . To do so , we used high throughput bacterial 16S-based metabarcoding , lipL32 gene-targeting RT-PCR , rrs gene sequencing and VNTR typing as well as GIS-based multivariate spatial analysis . Our results show that leptospires seem absent from the core city where usual Leptospira reservoir rodent species ( namely R . rattus and M . natalensis ) are yet abundant . On the contrary , L . kirschneri was detected in Arvicanthis niloticus and Cricetomys gambianus , two rodent species that are restricted to irrigated cultures within the city . Moreover , the VNTR profiles showed that rodent-borne leptospires in Niamey belong to previously undescribed serovars . Altogether , our study points towards the importance of market gardening in maintain and circulation of leptospirosis within Sahelian cities . In Africa , irrigated urban agriculture constitutes a pivotal source of food supply , especially in the context of the ongoing extensive urbanization of the continent . With this in mind , we speculate that leptospirosis may represent a zoonotic disease of concern also in arid regions that would deserve to be more rigorously surveyed , especially in urban agricultural settings .
Leptospira is a genus of spirochetes which comprises three lineages , one of which grouping pathogenic species for both animal and human [1] . Leptospirosis is a major zoonotic disease that may affect at least 500 , 000 and potentially up to 1 million persons , and kill ~60 , 000 ones per year worldwide [2–5] . Its incidence remains poorly documented because leptospirosis leads to clinical signs that are difficult to distinguish from other widespread endemic pathologies such as dengue , malaria , influenza , etc . [6] . In addition , many countries where it has an obvious burden lack appropriate diagnostic facilities , thus strongly suggesting that cases may be massively underreported [2 , 4] . Among other mammals , rodents , especially rats , constitute major reservoirs of Leptospira spp . : the bacterium resides in the host renal tubules and is then excreted into the environment through its urine . Leptospirosis is thought to be essentially associated with water where humans get contaminated following contact with the pathogen through skin abrasions or mucous membranes ( reviews in [4 , 7] ) . In particular , rice culture , recreational water activities and flooding have been massively linked to leptospirosis . This is the reason why the disease was essentially looked for , and found in temperate as well as warm and humid tropical regions ( reviewed in [8] ) . Surveys in arid zones are rare , although some mentions exist from desert to sub-desert areas ( e . g . , Somalia: [9]; Arizona: [10]; Mexico: [11]; Brazil: [12] ) , thus suggesting that Leptospira may be much more widespread than currently thought and could also extend to dry regions . As an example , prevalence in wild Malagasy mammals was found higher in Northern areas of the island where rainfalls are weaker [13] . Mentions of Leptospira in Africa ( review in [14–16] ) are quite scattered , and even very rare for some particular regions ( see Fig . 2 in [14] ) . For instance , in the West African Sahel zone , some data are available for Senegal ( two investigations in both humans and cattle in the 1970s ) , Chad ( one report in a dog in 2008 ) and Mali ( one human case report in the 1990s , and one investigation in cattle in the early 1970s ) while no monitoring has ever been conducted in Burkina-Faso or Niger [14] . Yet , reported sporadic epidemics in various parts of the continent reflect a lack of knowledge of the disease rather than a truly narrow distribution of Leptospira [8] . This suggests that further investigations in Africa in general , and in arid zones in particular are required . Moreover , leptospirosis is often associated with disadvantaged urban areas where poor sanitation together with elevated rodent-human interactions increase the risk of rodent-to-human transmission ( e . g . , [17–19]; reviewed in [4] ) . Taking into account the impressive growth of African cities [20] , there is little doubt that leptospirosis will be a major ( re ) emerging disease on the continent [14] . Niger , focus of the present study , ranks last of the World for the Human Development Index ( 187 out of 187; [21] ) . The capital city , Niamey , lies on the Niger River in the western part of the country , and is located in the typical Sahelian bioclimatic zone . As such , it is characterized by high temperatures ( monthly average temperatures between 22–36°C ) and low rainfalls ( ~540 mm per year ) with a single rainy season between May and September . It was created ex nihilo at the very end of the nineteenth century by French colonizers ( reviewed in [22] ) . During the last decades , the city has been experiencing an explosive spatial and demographic growth with its population increasing from >30 , 000 in the late 1950s , to 707 , 000 in 2001 , and currently reaching more than 1 , 000 , 000 inhabitants [22–24] . As often in such cases , this rapid urbanization is characterized by many informal settlements and insufficient sanitary services . Accordingly , data hence knowledge about zoonotic pathogens that may circulate in Niger are extremely scarce , potentially explaining why so many fevers are misdiagnosed as malaria ( i . e . , more than 55% in the rainy season , and up to 95% during the dry season; [25] ) . Expectedly , Leptospira appears among the top candidate pathogens that may explain these so many fevers of unknown origin [26] . These are the reasons why we took advantage of a monitoring of urban rodents conducted in Niamey , the main town of Niger [27] , to perform the first survey of rodent-borne Leptospira in this very poor country where zoonoses are dramatically under-documented .
The whole rodent trapping campaign was validated by national and local authorities ( scientific partnership agreement number 301027/00 between IRD and the Republic of Niger ) . At the French level , all sampling procedures were approved by the “Comité d’Ethique pour l’Expérimentation Animale—Languedoc Roussillon” ( agreement number C34-169-1 , valid until 25th July 2017 ) and were conducted by biologists from the CBGP holding certificates to carry out experiments on live animals ( agreement number C34-488 ) . None of the rodent species investigated in the present study has protected status ( see UICN and CITES lists ) . All animals were treated in a humane manner in accordance with guidelines of the American Society of Mammalogists . All rodents were euthanized through cervical dislocation . Permit to enter and work within private properties were systematically obtained through oral but explicit agreement from adequate institutional ( research agreement quoted above; mayor ) and traditional authorities ( both neighborhood and family chiefs ) . From October 2009 to February 2011 , an extensive survey of urban rodent assemblages was conducted in 52 localities of Niamey , Niger , thus allowing the exploration of more than 215 trapping sites with an effort of >14 , 500 night-traps ( see details in [27] ) . Among the 987 rodents captured , 578 were included in the present screening of Leptospira . They consisted in 66 Arvicanthis niloticus , 12 Cricetomys gambianus , 350 Mastomys natalensis , 50 Mus musculus and 100 Rattus rattus originating from 49 localities sites within the city ( Table 1 and Fig 1 ) . African rodent species identification may sometimes be difficult due to the frequent co-occurrence of sibling taxa , notably in the genera Rattus [28] , Arvicanthis and Mastomys [29] . This is the reason why a special attention was paid to taxonomic diagnosis which relied on karyotyping ( for Arvicanthis , Mus and Mastomys ) , cytochrome b gene sequencing ( for Arvicanthis and Rattus ) , PCR and species-specific RFLP ( for Mastomys ) and genotyping ( for Mastomys and Rattus ) . All these procedures have been described in details elsewhere ( see [27] , and references therein ) . Individual genomic DNA was extracted from ethanol-preserved kidney tissue using the Qiagen DNeasy Blood and Tissue Kit , and was quantified using Nanodrop technology ( Thermoscientific ) . Kidney DNA samples were then prepared in equimolar concentration . Pools grouping 50 rodent individual DNA samples each were then arranged by species as follows: ( i ) one pool made of 50 A . niloticus from 7 localities , ( ii ) one pool of 50 M . musculus from 3 localities , ( iii ) two pools with 50 black rats each from 11 localities , respectively , and ( iv ) seven pools of 50 M . natalensis each and representing 32 localities . Samples were chosen in order to cover most ( when not all ) localities where each species had been found during a recent broader survey of urban rodents of Niamey ( Table 1; see [27] ) . The eleven pools of DNA were then screened for the presence of bacteria using universal PCR primers targeting the hypervariable region V4 of the 16S rRNA gene ( 251bp ) via Illumina MiSeq ( Illumina ) high throughput sequencing . The V4 region has been proven to offer excellent taxonomic resolution for bacteria at the genus level [30] . A multiplexing strategy enabled the identification of bacterial genera in each pool sample . We followed the method detailed in Kozich et al . [31] for PCR amplification , indexing , pooling of PCR products and de-multiplexing . Bacteria taxonomic identifications at the generic level were performed using the Silva SSU Ref NR 119 database ( http://www . arb-silva . de/projects/ssu-ref-nr/ ) as a reference [32] . Each DNA pool was analyzed in triplicate using three independent PCRs and three amplicon libraries in the same next generation sequencing ( NGS ) run using a MiSeq sequencer ( Illumina ) . Rodents that belonged to metabarcoding Leptospira-positive pools as well as 16 A . niloticus and 12 C . gambianus which had not been included in the latter NGS-based survey were all individually screened for pathogenic Leptospira species using a dedicated Real Time PCR-based test . To do so , sequences of lipL32 gene from Leptospira kirschneri ( AF121192 ) , L . interrogans ( AF181553 , AF245281 , AF366366 , LIU89708 ) , L . borgpetersenii ( AF181554 ) , L . santarosai ( AF181555 ) and L . noguchii ( AF181556 ) were aligned , and a consensus sequence was determined using BioEdit v . 7 . 1 . 9 . New forward ( LIP32BF: 5’-AGC TCT TTT GTT CTG AGC GA-3’ ) and reverse ( LIP32BR: 5’-TAC GAA CTC CCA TTT CAG CGA TTA-3’ ) primers were designed from this consensus sequence using the Light Cycler Probe design software v . 2 . 0 ( Roche ) . This new set of primers was proved to detect most known pathogenic Leptospira species ( namely L . interrogans , L . borgpetersenii and L . kirschneri in ‘wet lab’ , as well as L . santarosai and L . noguchii in silico ) with lower Ct values than the primers used in recent lipL32 RT-PCR-based survey ( e . g . , [33 , 34] ) . We used the TaqMan probe ( FAM-5′-AAA GCC AGG ACA AGC GCC G-3′-BHQ1 ) previously described in Stoddard et al . [33] , thus allowing us to amplify a 199 pb-long fragment of the leptospiral lipL32 gene . RT-PCR reactions were performed using a LightCycler 480 ( Roche ) in 96-well microtitre plates with 10μL as final volume for each reaction . Optimal amplification conditions were obtained with 0 . 5μM of each primer , 0 . 2μM of probe , 2X of Probe Master buffer ( Roche ) , 0 . 5U of Fast start Taq DNA polymerase ( Roche ) and 2μL of sample DNA . RT-PCR program consisted in an initial denaturing step at 95°C for 10 min , followed by 50 cycles of 95°C for 15s , 60°C for 30s and 72°C for 1s , and a final cooling step to 40°C . All samples were investigated in independent duplicates . Genomic DNA isolated from L . interrogans serovar Canicola and L . borgpetersenii serovar Tarassovi were used as positive controls . The Beta Actin gene was amplified from all samples as an internal RT-PCR control in order to detect false negative results [35] . A 330 pb-long fragment of the rrs gene was amplified from genomic DNA of the RT-PCR-positive rodents . Primers A and B were used for a first amplification; when the PCR was negative , a nested PCR was performed with primers C and D [36] . PCR products were sequenced in both directions at Eurofins Genomics . Species-specific identification was performed through Blastn ( option Megablast for highly similar sequences ) procedure under NCBI database . Identification at the subspecies level was performed by multiple-locus variable-number tandem repeat analysis ( MLVA ) using the loci VNTR4 , VNTR7 and VNTR10 as previously described [37] with the following modifications . MLVA was performed on DNA extracts using 70 cycles of amplification with a higher concentration of Taq polymerase ( GE Healthcare ) . The sizes of the amplified products were then analysed using a 1% agarose gel electrophoresis , and the profiles were compared with the database of the National Reference Center for Leptospirosis ( Institut Pasteur , Paris , France ) . Our purpose here was to map the most suitable within-city areas for Leptospira-carrying rodent species , as identified by molecular methods . Since reservoir rodents in Niamey were all found to belong to rural-like species ( i . e . A . niloticus and C . gambianus; see below ) and since these latter species strictly segregate spatially from true commensal ones throughout the town [27] , we chose to focus on rural-like species only . For such a purpose , a Geographic Information System ( GIS ) of Niamey was implemented from a SPOT satellite image ( CNES 2008 ) using the following seven land cover categories ( LCC ) : Niger river , ponds , bare soils , tarred areas , trees , other greenings and sheet steel-made roofs . The local urban landscape was described in the vicinity of each of the 11 sampling points where C . gambianus and/or A . niloticus specimens were caught ( Table 1 ) . To do so , circular buffers of 30 m radius were centered upon each sampling location , and the corresponding landscape was extracted using the R software [38] and the package “raster” [39] . Each circular landscape was described using the percentage of landscape ( PLAND ) composition metric computed for each LCC [40] using the R package SDMTools [41] . This led to a set of 7 PLAND values ( one for each LCC ) for each sampling location . The second step of the analysis consisted in processing these compositional data through a Principal Component Analysis [42] using the R package “ade4” [43] . The first Principal Component ( PC1 ) partly , but highly significantly separated the locations with and without trapped rodents ( Monte-Carlo test of between-group inertia , 999 replicates , p = 0 . 009; [44] ) . The locations with rodents were associated to high values of PLAND for trees and greenings and low proportion of bare soil and sheet steel-made roofs . As a third step , we rasterized the GIS of Niamey into 60x60 meters cells ( N = 67 , 077 ) within which the percentage of each PLAND was computed . These pixels were then projected onto PC1 as supplementary rows [42] . Their coordinates onto PC1 thus represented their relative position with regards to the gradient of habitat suitability for Leptospira-carrying rodent species . The pixel coordinate comprised within the range of the coordinates of locations where Leptospira-carrying rodent species were standardized to range between 0 and 1 and subsequently mapped ( Fig 2 ) . As such , this map depicts the city-wide spatial variation of the similarity between local habitat and average landscape composition of locations where Leptospira-carrying rodent species were caught . In other words , it shows the distribution of suitable habitats for rodent-borne Leptospira within Niamey . Expectedly , most of the build-up areas of the city were retrieved as unsuitable for rural-like hence Leptospira-carrying rodent species ( Fig 2 ) .
In total , 578 rodents from 49 localities ( Fig 1 ) and five main categories of habitats ( i . e . , households , markets , coach station , gardens and factories , the latter including a slaughter house , a husking rice industry and an industrial storeroom; Table 1 ) were investigated for the presence of Leptospira using one to four complementary molecular approaches ( i . e . , metabarcoding , RT-PCR , sequencing , VNTR profiles; Table 1 ) . First , 550 individuals were screened in triplicated species-specific pools through bacterial 16S metabarcoding ( Table 1 ) . A total amount of 287 , 057 16S sequences was obtained . Among them ( which include some bacterial genera of potential medical interest such as Helicobacter , Orientia , Mycoplasma , Streptococcus , Ignatzschineria ) , the three replicates of the A . niloticus-specific pool were found positive for Leptospira ( 3385 , 3144 and 3050 Leptospira sequences , respectively ) while no Leptospira sequence were retrieved for the other pools , with the only exception of one Mastomys-specific pool in which 37 , 58 and 64 Leptospira sequences were found . Such a low amount of sequences was intriguing and , after close verifications , we found that that one Leptospira-positive Arvicanthis individual had been added by error to this slightly positive Mastomys-pool ( NB: this had been noted on the bench book but this pipetting mistake was then omitted ) . In order to unambiguously confirm that this lab error was responsible for the few Leptospira sequences retrieved within this particular pool , each Mastomys individual was screened using the lipL32 RT-PCR-based procedure: no positive Mastomys could be found . Second , the 50 Arvicanthis niloticus specimens from the NGS positive pool as well as 16 additional A . niloticus and 12 C . gambianus individuals ( that had not been included in the metabarcoding survey ) were investigated individually using duplicated lipL32-targeting qPCR ( Table 1 ) . These 78 animals originated from sites J-LMO1 , J-LMO2 , J-NOG , J-CYA , J-DAR , J-GAM , J-KIR1 and CRA-3 ( Table 1 ) . Among them , seven animals trapped in J-GAM , J-KIR1 and J-LMO2 appeared Leptospira-positive twice ( with Ct ≤ 31 ) , while one from J-GAM was found positive in only one of the two duplicates ( Ct = 38 . 2 ) . In addition , 12 Cricetomys gambianus from sites J-LMO1 , J-KIR2 , CRA-1 and CRA-2 were also investigated through RT-PCR: one of them ( from CRA-2 ) was found twice Leptospira-positive ( Ct = 20 . 3 and 20 . 5 ) . Third , the DNA of the seven A . niloticus and the C . gambianus qPCR-positive individuals were successfully amplified and sequenced for the Leptospira rrs gene ( only the A . niloticus that was qPCR-positive in one of the two duplicates could not be amplified ) . All eight sequences ( Genbank accession numbers KT583752 to KT593759 ) were found strictly identical ( whatever the rodent host species ) and , following a Blastn procedure , strictly identical to L . kirschneri sequences ( 100% identity; 100% sequence cover; E value = 4 . e-136; the subsequent most similar sequences belonged to L . interrogans with 99% identity , 100% sequence cover and E value = 1 . e-134 ) . MLVA is a simple and rapid PCR-based method for the identification of most of the serovars of L . interrogans and L . kirschneri [37] . MLVA of the VNTR-4 , VNTR-7 , and VNTR-10 loci were performed in all nine RT-PCR-positive individuals . No PCR product was obtained for the sample that had been found positive in only one out of the two RT-PCR duplicated screenings while two different patterns were retrieved for the remaining ones . First , all Arvicanthis samples belonged to genotype I ( i . e . , VNTR4: 450bp , VNTR7: 320bp , VNTR10: 350bp ) . Second , genotype II was only represented in the single Leptospira-positive C . gambianus specimen ( i . e . , VNTR4: 370bp , VNTR7: 320bp , VNTR10: no amplified product ) . None of these genotypes I and II have been described previously . All the Leptospira-carrying rodents identified in Niamey were trapped in February , October and November . These months all correspond to the dry and cool season . Nevertheless , our sampling did not allow us to investigate seasonality in a satisfying manner , especially within the urban gardens where most rodents were caught in February , October and November , except for one individual trapped in July and two specimens trapped in March .
Our study allows us to highlight for the first time the presence of pathogenic leptospires in Niger . At a wider scale , our data also add to the very rare mentions of Leptospira spp . in the Sahel [14] , thus confirming that these bacteria do circulate in Sub-Saharan Africa more extensively than currently thought . Moreover , our molecular investigations showed that rodent-borne Leptospira in Niamey belonged to L . kirschneri and to a genotype that had never been identified previously . Its biological features and medical impact , including its virulence in human , remain to be studied in details . Leptospirosis is one of the most widespread zoonotic diseases around the World . In tropical areas , contact with contaminated water following heavy rainfall and flooding episodes is thought to be a major risk of exposure to pathogenic Leptospira spp . [45] . In temperate regions , infection mode is less clear , with recreational water activities and animal caretaking potentially also being of epidemiological importance [4] . In developing countries , high infection rates were also found in cities , essentially within disadvantaged urban areas that usually show poor sanitation and where rodents are numerous ( e . g . , [17–19 , 46 , 47] ) . Here , we point towards a potential other major context of Leptospira infection risk in the tropics , namely the market garden areas that surround most cities in developing countries , including those that lie within semi-arid regions . Indeed , rats are usually considered as the major rodent reservoirs for leptospires worldwide [48] . In Eastern Africa , Mastomys natalensis is thought to be the principal source of human infection [49] . Rattus rattus and M . natalensis are from far the most abundant species that were found within Niamey [27] . Yet , out of the 450 specimens of these two species that were tested here , none could be found Leptospira-positive . On the contrary , only Arvicanthis niloticus and Cricetomys gambianus specimens , all trapped within urban market gardens , were detected as carrying Leptospira . This strongly suggests that Leptospira spp . circulate mostly , if not only in these particular habitats . This is tempting to speculate that irrigated gardens and rice fields along the Niger River provide the warm and moist environmental conditions that favor the bacterium circulation with both the presence of mammalian hosts such as rodents , human-maintained humidity of soils and free water . The absence of rodent-borne leptospires elsewhere in town despite the abundance of potential competent hosts ( especially Rattus rattus and Mastomys natalensis; [27] ) as well as poor sanitation conditions would be explained by long-term aridity , thus strongly contrasting with the situation observed in other wetter tropical cities . The importance of environmental factors in the epidemiology of pathogenic Leptospira species has already been suggested in Thailand where the commensal species Rattus exulans was found infected much less frequently than other rural / wild species [34] . Ganoza and colleagues [46] further suggested that anthropogenic modification of the urban habitat was a major driver of leptospiral transmission to human . With this in mind , our study emphases the potentially highly critical role of urban market gardening in leptospirosis epidemiology since horticulture rapidly extends within and around towns of most developing countries . In sub-Saharan Africa , these so-called green cities are considered as a trump card to reach the “zero hunger” challenge [50] . For instance , urban and peri-urban horticulture produces most of all leafy vegetables that are consumed in Accra ( Ghana ) , Dakar ( Senegal ) , Bangui ( Central African Republic ) , Brazzaville ( Congo ) , Ibadan ( Nigeria ) , Kinshasa ( Democratic Republic of Congo ) and Yaoundé ( Cameroon ) , which represent a total population of 22 . 5 million inhabitants [50] . Yet , the setup of agricultural spaces in close proximity to , when not inside cities or villages raise public health issues since they may favor the maintaining of some pathogenic agents and eventually their vectors or reservoirs , hence potentially increasing the risk of human exposure to the associated diseases , such as malaria ( e . g . , in Benin: [51 , 52]; in Ghana: [53] ) , various gastro-intestinal infections ( e . g . , in Benin: [54] ) schistosomiasis ( e . g . , in Ivory Coast: [55 , 56] in Niger: [57] ) , leptospirosis ( this study ) or potentially toxoplasmosis ( e . g . , in Niamey , Niger: [58] ) . Fine-scale studies show that the impact of these infectious agents may vary at very local scale , depending on the habitat structure and use ( e . g . , [55 , 56] ) . In the same manner , in Brazilian slums , human cases of leptospirosis seems to aggregate at the very local scale of some households [59] , thus suggesting that city-scale studies are inadequate to fully understand the disease epidemiology [48] . These findings , together with our first description of rodent-borne pathogenic Leptospira within urban market gardens of Niamey , suggest that investigations are now required in order to ( i ) provide a more precise picture of Leptospira circulation within the urban farming zones of this Sahelian city , and ( ii ) to look whether human transmission evidence indeed exists in Niger . If this was to be the case , leptospirosis may well represent an important amount of the numerous cases of “fever of unknown origin” that mimic malaria in this semi-arid area . Our GIS-based inferences of suitable areas for Leptospira-carrying rodent species in Niamey clearly correspond to intra-city agricultural zones , especially those along the Niger River and the Gountou Yéna wadi ( Fig 2 ) . This suggests that human populations at higher risk may well be urban farmers as well as all people that are in close contact with the river waters for their everyday activities ( e . g . , fishing , clothes and dish washing , bathing , etc ) . This is the reason why we recommend that investigations about human prevalence are conducted in these areas where leptospires may represent a very impacting though under-diagnosed health issue . Finally , climatic change together with human-mediated modifications of land use accentuates Niger River-associated flooding events ( see , for instance , the dramatic episodes that occurred in Niamey in 2010 , 2012 and 2013; [60 , 61] ) . From there , we anticipate an increase of leptospirosis’ impact on human health in Niamey in a near future . | We surveyed rodent-borne Leptospira in rodents from Niamey , the capital town of Niger , using bacterial metabarcoding , RT-PCR , sequencing , VNTR typing and GIS-based geostatistics . Two new serovars of Leptospira kirschneri were identified in Arvicanthis niloticus and Cricetomys gambianus , two species that inhabit exclusively urban irrigated gardens . Since no rodent-borne leptospires could be found in the core city , our results point towards the importance of urban agriculture in the maintaining and the circulation of these bacteria in cities from semi-arid regions where they are usually poorly documented and even hardly looked for . Accordingly , this is one of the very rare mentions of these zoonotic agents in Sahel , and the first one in Niger . Keeping in mind the critical role of urban gardening for food security in extensively growing West African cities , we believe that leptospirosis should be more closely scrutinized in Sahelian countries where numerous cases of human fevers are of unknown origin . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Urban Market Gardening and Rodent-Borne Pathogenic Leptospira in Arid Zones: A Case Study in Niamey, Niger |
The transformation of synaptic input into patterns of spike output is a fundamental operation that is determined by the particular complement of ion channels that a neuron expresses . Although it is well established that individual ion channel proteins make stochastic transitions between conducting and non-conducting states , most models of synaptic integration are deterministic , and relatively little is known about the functional consequences of interactions between stochastically gating ion channels . Here , we show that a model of stellate neurons from layer II of the medial entorhinal cortex implemented with either stochastic or deterministically gating ion channels can reproduce the resting membrane properties of stellate neurons , but only the stochastic version of the model can fully account for perithreshold membrane potential fluctuations and clustered patterns of spike output that are recorded from stellate neurons during depolarized states . We demonstrate that the stochastic model implements an example of a general mechanism for patterning of neuronal output through activity-dependent changes in the probability of spike firing . Unlike deterministic mechanisms that generate spike patterns through slow changes in the state of model parameters , this general stochastic mechanism does not require retention of information beyond the duration of a single spike and its associated afterhyperpolarization . Instead , clustered patterns of spikes emerge in the stochastic model of stellate neurons as a result of a transient increase in firing probability driven by activation of HCN channels during recovery from the spike afterhyperpolarization . Using this model , we infer conditions in which stochastic ion channel gating may influence firing patterns in vivo and predict consequences of modifications of HCN channel function for in vivo firing patterns .
Thermal fluctuations in the conformation of an ion channel protein can cause it to make spontaneous transitions between discrete conducting and non-conducting states [1] , [2] . Nevertheless , computational models of ionic conductances in a neuron generally assume the behavior of a population of ion channels to be deterministic and stochastic gating of ion channels is usually neglected in models of synaptic integration and spike initiation [3] , [4] . For a typical cortical principal neuron , this assumption can be justified by the very small amplitude of the conductance change and resulting membrane current caused by opening of a single ion channel compared to either the resting membrane conductance or the threshold current for firing of an action potential . However , when neurons are depolarized to membrane potentials around the threshold for initiation of action potentials , the biophysical mechanisms that underlie spike generation dictate that the effective membrane conductance becomes very low [5] . As a result , even small fluctuations in ionic current through relatively few ion channels could significantly alter the membrane potential and the initiation of action potentials [6] , [7] . Consistent with this possibility stochastic gating of membrane ion channels that determine the threshold for action potential initiation can influence the dynamic electrical properties of neurons [8]–[11] . However , little attention has been given to the consequences of stochastic ion channel gating for the patterns of spike output produced during active states in which the membrane potential is depolarized to near threshold . We have focused on understanding the influence of stochastic ion channel gating on the integrative properties of stellate neurons from Layer II of the medial entorhinal cortex ( MEC ) . These glutamatergic neurons provide cortical input to the hippocampal dentate gyrus [12] , [13] . Electrophysiological recordings reveal two unusual integrative properties of stellate neurons from the MEC [14]–[17] . First , during prolonged periods of excitation stellate neurons fire action potentials in stereotypical clustered patterns . The frequency of spikes within a cluster is approximately 8–14 Hz and is relatively independent of the average spike frequency , while the intervals between spike clusters are typically hundreds of milliseconds or longer [18] . The organization of clustered spike patterns appears to depend on a large and slow spike afterhyperpolarization ( AHP ) that is also independent of the overall average spike frequency [18] . A second distinctive feature of stellate neurons is the emergence of prominent ( ∼3–5 mV in amplitude ) intrinsic membrane potential fluctuations upon membrane depolarization [14] . These fluctuations have been proposed to contribute to network rhythmicity due to their power in the theta frequency range ( 4–12 Hz ) , the prominent oscillatory frequency of entorhinal and hippocampal network activity during exploratory behavior and REM sleep [19] . Previous models and experimental results indicate that stochastic gating of persistent Na+ channels may be essential for the sub-threshold oscillations observed in stellate neurons [20] , [21] . However , the consequences of stochastic gating of other classes of ion channel expressed by stellate neurons have not been explored . Moreover , while sub-threshold oscillations have been suggested to drive clustered spike patterns [14] , [22] , [23] , the mechanisms underlying oscillations and clustered spike firing can be dissociated experimentally [18] , [24] , and therefore it is not clear if stochastic fluctuations in ion channel opening play any role in the generation of clustered spike firing . Hyperpolarization-activated , cation non-selective ( HCN ) channels play a central role in determining subthreshold integration and the pattern of action potential initiation in stellate neurons from the MEC [18] , [25] . The substantial hyperpolarization-activated current ( Ih ) in stellate cells is mediated in large part by HCN1 channels and is a major determinant of the effective membrane conductance of the neuron at rest and at more depolarized potentials close to the threshold for initiation of action potentials [18] . Experiments using pharmacological and genetic manipulations suggest that HCN channels increase the probability that clustered patterns of action potentials will be generated and increase the frequency of action potentials within each cluster [18] . However , the mechanisms through which HCN channels influence these patterns of spike firing are not clear . Computational models of stellate neurons have suggested either that Ih plays an essential role in perithreshold oscillations and clustered patterns of spike firing [26] or that Ih is not required [21] . Moreover , numerous studies suggest that the effects of Ih on the integrative properties of a neuron are highly context dependent [18] , [27]–[36] . Thus , the role of Ih is determined by interactions with other ion channels . Depending upon the cell type and even the subcellular compartment studied , Ih can lead to varied properties , from prevention of bistability [37] to regulation of dendritic spiking [38] . Therefore , understanding the properties of stellate neurons and their sensitivity to manipulations of Ih will likely require an account of the interactions between multiples classes of ion channel . To better understand the impact of stochastic ion channel gating on the patterns of spike output from stellate neurons and to reconcile the contrasting views of the role of Ih in perithreshold oscillations and clustered patterns of spike firing , we addressed two questions . How do interactions of HCN channels with other membrane ion channels lead to the emergence of membrane potential oscillations and spike firing patterns recorded from entorhinal stellate cells ? Could stochastic ion channel gating at potentials close to spike threshold influence the patterns of spike output generated by stellate neurons ? We demonstrate that whereas a deterministic model of channel gating is sufficient to account for many of the properties of entorhinal stellate neurons at hyperpolarized membrane potentials , including the consequences of HCN1 deletion , a model with stochastically gating ion channels is necessary to reproduce the distinctive properties of stellate neurons near threshold . Examination of the model reveals that spike initiation is probabilistic and that the tendency to emit clustered spikes can be explained by a transient increase in the probability of spike initiation following recovery from the action potential AHP . We find that this transient increase in spike probability is primarily due to Ih and explains the role of HCN channels in the emergence of clustered patterns of spikes . Finally , we ask whether stochastic ion channel gating could contribute to patterns of spike output observed in vivo . We propose that stochastic gating of ion channels expressed by stellate neurons is crucial to their transformation of synaptic input into a patterned spiking output and places constraints on the development of models of entorhinal cortex function [39] .
At membrane potentials just below the threshold for initiation of action potentials , stellate cells generate membrane potential fluctuations with a dominant frequency typically in the 5–10 Hz range [14] , [16] . Our previous experimental studies using HCN1 knockout mice indicate that , at any given membrane potential , HCN1 channels are not required for fluctuations in this frequency range , but rather HCN1 channels suppress low-frequency components of membrane potential activity [18] . However , the amplitude of the theta frequency fluctuations becomes larger with depolarization towards the spike threshold and if the absolute value of the membrane potential is not accounted for , then deletion of HCN1 channels can appear to reduce the amplitude of membrane potential fluctuations by lowering the most depolarized potential at which fluctuations can be maintained without triggering action potentials [18] . These results contradict proposed deterministic models for the generation of theta frequency fluctuations by stellate cells [17] , [26] and also suggest how failure to account for differences in membrane potential could lead to the conclusion that block of HCN channels abolishes theta frequency fluctuations [25] . Nevertheless , it has yet to be shown whether these experimental observations can be accounted for in a theoretical model . We first examined the membrane potential of the stochastic models during injection of constant current of amplitude adjusted to the maximum possible without triggering action potentials ( Figure 2 ) . For the wild-type and knockout versions of the model this corresponded to respective mean membrane potentials of −51 . 6 and −53 . 4 mV . At these membrane potentials , the stochastic stellate neuron models show large fluctuations in membrane potential ( ∼3–4 mV peak to peak; Figure 2 ) , whereas the otherwise identical deterministic models show no fluctuations ( Figure 2C ) . We found that the membrane potential fluctuations recorded over long epochs ( 20 s ) are spectrally complex , but show peak activation between 3–10 Hz consistent with previous observations in stellate neurons in vitro [18] , [42] . Some previous studies have analyzed brief epochs in which the membrane potential fluctuations appear to be coherent oscillations [16] , [25] , [43] . Consistent with these studies , we also find that short epochs of membrane potential , recorded from simulations with the stochastic models , reveal clear autocorrelation peaks ( Figure 2B ) and dominant frequency components in the theta frequency range ( Figure 2C and 2D ) . Removal of the fast and large component of Ih in the knockout model resulted in an apparent shift in the peak of the spectral density to lower frequencies ( <5 Hz ) similar to previous experimental results in HCN1 knockout mice [18] ( Figure 2C and 2D ) . By contrast , measurements made when controlling for membrane potential between the models , reveal that the knockout model has larger amplitude fluctuations ( Vavg = −53 . 7 mV , simulation time = 3 s; σWT = 0 . 37 mV σKO = 0 . 47 mV; see also Figure S2 ) , also consistent with experimental data [18] . In further agreement with previous experimental data [18] , these effects can be explained by the ability of HCN channels to reduce the membrane impedance at low frequencies ( Figure S2 ) . As predicted by changes in impedance , responses to a white noise current stimulus , with standard deviation matched to the current noise recorded in the stochastic model , were enhanced in the deterministic knockout model compared with the equivalent wild-type model ( Figure 2C ) . Phase plots of the relationship between membrane current and voltage during perithreshold fluctuations , revealed that Ih is a minor contributor ( Figures S3 ) to the net membrane current changes that drive fluctuations . Thus , the stochastic model accounts well for the properties of subthreshold fluctuations and their dependence upon HCN1 channels reported previously [16] , [18] , [21] , [25] , [42] . This model is consistent with perithreshold fluctuations arising from interaction of stochastically gating ion channels other than HCN channels ( Figure S4 ) [21] , but with the amplitude and spectral properties of the fluctuations shaped by the presence of HCN channels and dependent on the average membrane potential at which the fluctuations are examined . The most depolarized average membrane potential that can be maintained without initiation of an action potential appears to determine the maximal observable amplitude of membrane potential fluctuations and is altered both in the HCN1 knockout model ( Figure 2 ) and in experimental recordings of stellate cells from HCN1 knockout mice [18] . To further assess the stability of the membrane potential prior to action potential initiation we injected slow , ramp-like currents that crossed spike threshold for both the wild-type ( Figure 3A ) and knockout ( Figure 3B ) versions of the model . We averaged the membrane potential from several sweeps in a time window 0 . 1–0 . 5 s before the initial action potential for each trial ( Figure 3E ) . The spike-triggered averages ( Figure 3D ) revealed that removal of the HCN1-like current from the model causes spikes to initiate from a more hyperpolarized membrane potential ( wild-type: −51 . 15+/−0 . 12 mV; knockout: −52 . 72+/−0 . 12 mV; P = 4×10−11; N = 20 total trials; Figure 3E ) . This difference between the wild-type and knockout models is independent of stochastic channel gating ( Figure 3D ) , but is to be expected from the increased rate of depolarization resulting from the reduced membrane conductance following removal of HCN1 channels . However , for both of the deterministic models the membrane potential follows a more depolarized trajectory than in the corresponding stochastic models ( Figure 3D ) . This is consistent with spontaneous membrane potential fluctuations in the stochastic models triggering action potentials relatively early during the ramp current . Consistent with the difference in responses to DC current injection ( Figure 2 ) , during the time-window preceding the spike , the more depolarized potentials in the wild-type model are associated with an increased standard deviation of the membrane potential due to stochastic channel gating ( wild-type: 0 . 90+/−0 . 06 mV; knockout: 0 . 69+/−0 . 04 mV; P = 0 . 005; Figure 3E ) . The shift in membrane potential stability was accompanied by a small increase in the standard deviation of the time of the first action potential in the stochastic HCN1 knockout model ( wild-type: 0 . 119±0 . 008 s; knockout: 0 . 152±0 . 015 s; Figure 3C; P<0 . 05 , N = 60 simulations ) , suggesting that HCN1 channels may increase the reliability of spike timing as well as the stability of the sub-threshold membrane potential . When stellate cells experience maintained depolarizing currents that drive action potential firing at mean frequencies less than 5 Hz , the pattern of firing is characterized by clusters of action potentials at a relatively high frequency ( 8–14 Hz ) interspersed with silent periods [14] , [18] , [24] . We determined the conditions for initiation of spikes with mean frequencies less than 5 Hz , at which clustered spike patterns might be expected . In the deterministic model the transition from silence to continuous action potential firing occurs when the amplitude of the injected current is increased above 258 . 4 pA and 320 . 5 pA for wild-type and knockout configurations , respectively . For the deterministic models this transition corresponds to a sharp transition from silence to repetitive spiking at ∼6 Hz ( wild type ) and ∼3 Hz ( HCN1 knockout ) and clustered spike patterns were not observed ( Figure S5 ) . By contrast , the current threshold for the transition between silent and spiking states was ∼246 pA and ∼308 pA for the stochastic versions of the wild-type and HCN1 knockout models , respectively . In both stochastic models , arbitrarily low firing frequencies could be obtained when the injected current was just above this threshold . When the mean frequency of action potentials was less than approximately 5 Hz , then both stochastic models generated clustered patterns of spikes ( Figure 4 ) . Thus , stochastic ion channel gating enables clustered patterns of spikes to emerge during firing at low frequencies in response to input currents that are of insufficient amplitude to initiate action potentials in the corresponding deterministic model . We next examined in detail the patterns of spiking that emerge when constant current injected into the stochastic model drives low-frequency action potential firing ( Figure 4 ) . Consistent with electrophysiological results [18] , [24] , [44] , we find that the interspike interval ( ISI ) distribution of the stochastic model in response to constant current injection is multimodal , being characterized by both a dominant , short ISI mode as well as a wide distribution of long ISIs ( Figure 4A ) . However , in the knock-out model this short latency peak is much broader than in the wild-type model ( Figure 4A ) . Closer examination of the model behavior across a range of average firing frequencies revealed the characteristic tendency of stellate neurons to fire clustered action potentials ( Figure 4B and 4D ) . The knockout version of the model reveals a lesser tendency to fire spikes in clusters ( Figure 4C and 4D ) , consistent with the broadening of the short latency peak in the ISI histogram ( Figure 4A ) . We quantified the probability of clustering ( Pc ) with definitions used previously for experimental data ( see Materials and Methods; [18] ) . For the wild-type stellate neuron model Pc depends upon average firing frequency and peaks at intermediate ( 1–3 Hz ) frequencies ( Figure 4D ) . Importantly , Pc is significantly reduced in the knockout model at intermediate average firing rates ( Figure 4D ) . Finally , as in experimental recordings , the average number of spikes per cluster in the stochastic models is quite variable and depends on the average firing frequency ( Figure 4E ) . We previously demonstrated that Ih accelerates the repolarization from the AHP in stellate neurons , while overall shorter AHPs predict an increased tendency of neurons to fire clustered patterns of action potentials [18] . Similarly , the half-width of the AHP in the wild-type stochastic model was independent of the average frequency of spike firing ( Figure 5C ) . In contrast , after the simulated removal of HCN1 channels , the AHP half-duration was broader and varied as a function of average spike frequency ( Figure 5C ) , just as in experimental recordings from stellate neurons in HCN1 knockout mice [18] . The increase in duration of the AHP following removal of HCN1 channels was found in both stochastic and deterministic ( Figure S5 ) versions of the model indicating that this role of Ih does not require stochastic gating of the membrane ion channels . To quantify spike initiation following the AHP we calculated the conditional probability that a spike occurred at a time t following a previous spike at time t0 ( P ( st|st0 ) ) [45] . For spike trains generated by the knockout model ( Figure 5B , right panels ) , the latency to the increase in P ( st|st0 ) following a spike was increased and the magnitude of the change in P ( st|st0 ) was reduced from more than 6 fold to less than 3 fold compared with spike trains generated by the wild-type model ( Figure 5A , right panel ) . These changes are correlated with the reduction in Pc observed in simulations of the knockout model across a range of firing frequencies ( Figure 4D ) . Together , these simulations indicate that deterministic or stochastic versions of our model stellate neuron are sufficient to account for the resting membrane properties , subthreshold stability of the membrane potential and the sensitivity of these properties to alteration of Ih . However , only the version of our model containing stochastically gating ion channels is able to further account for the spontaneous emergence of membrane potential fluctuations at potentials near threshold . Moreover , the stochastic models produce clustered patterns of action potentials similar to spike patterns recorded from stellate neurons from wild-type and HCN1 knockout mice . Since our characterization of the stochastic models suggest that they provide a remarkably good account of experimental observations of both the resting and active properties of entorhinal stellate neurons , we went on to use these models to investigate how stochastic ion channel gating influences spike initiation and the generation of distinctive clustered patterns of action potentials . How do the clustered patterns of action potentials emerge and why do they require stochastic ion channel gating ? In a deterministic neuron , clusters or bursts of action potentials arise through modulation of spiking by slow changes in the state of one or more ion channels [46] , [47] . Indeed , such a deterministic mechanism has previously been proposed to account for clustered patterns of action potentials fired by entorhinal stellate neurons [26] . By this account , stochastic ion channel gating may lower the threshold for spike generation , but is not essential for the generation of clustered patterns of activity . However , stochastic ion channel gating may permit mechanisms for control of spike patterns that are not possible in deterministic models . In particular , whereas initiation of an action potential in a deterministic neuron is binary , with a clearly defined threshold , for stochastic neurons fluctuations in ion channel activity can lead to cancellation of a spike even when the deterministic threshold is crossed . At the other extreme spikes can be initiated in conditions that are well below the deterministic spike threshold [8] , [9] . Therefore , in a stochastic neuron there is no clearly defined boundary between a spiking and a non-spiking state and thus spike initiation should be considered probablistic rather than binary . The probabilistic nature of spiking in the stochastic model leads to a simple alternative mechanism for generation of clustered patterns of spikes , whereby the transient elevation in the probability of spiking following a previous action potential is sufficient to produce patterned output ( Figure 6 ) . According to this mechanism , changes in the recovery from a spike would alter the pattern of spikes by modifying the spike probability immediately following the refractory period ( Figure 6B ) . As a result , the activation of ion channels during each action potential and its associated AHP can be independent of the position of the action potential within or outside a cluster . Several lines of evidence support this probabilistic mechanism . First , conditional probability distributions , P ( st|st0 ) ( Figure 5A and 5B , right panels; also see Materials and Methods ) , reveal that the wild-type version of the model produces clustered action potentials by elevating the conditional probability of firing a spike , P ( st|st0 ) , over the steady-state probability , P ( st ) , for a brief period of ∼50 ms following a spike ( Figure 5A , right panels ) . Moreover , the reduction in Pc in the HCN1 knockout model is correlated with a decrease in P ( st|st0 ) ( Figure 5B , right panels ) as required by a probabilistic mechanism for clustered firing ( Figure 6B ) . Second , the number of spikes within a cluster is variable for a particular firing rate ( e . g . 3 . 11±1 . 7 spikes per cluster for 1 . 6 Hz ) and depends upon the average firing rate in both our model ( Figure 4E ) and experimental data [18] . This suggests that the number of spikes in a cluster is probabilistic and is consistent with a stochastic model of spike generation , but distinct from previous deterministic models [26] . Third , in a deterministic mechanism the half-width of the AHP should systematically vary with position in the cluster and should determine the succeeding ISI when terminating a cluster . Thus , on a spike-by-spike basis we would expect the AHP to correlate with the subsequent ISI . However , we find no such correlation in spike trains from either the wild-type or knockout models ( Figure 5D ) . Nonetheless , in both population data from experiments and in different versions of the stochastic model the AHP half-width correlates with Pc . These observations therefore support our conceptual model of spike patterning and suggest that there may be a common ionic basis that regulates the time course of both the AHP and P ( st|st0 ) . Fourth , to generate activity patterns that take place over relatively long time scales , such as spike clusters , a deterministic model requires relatively slow changes in the state of the model and at least one of the model parameters must vary as a function of a spike's location within a cluster . By contrast , the probabilistic mechanism of spike clustering does not require slow changes in model parameters beyond the recovery period from the AHP ( Figure 6 ) . Consistent with this prediction we find that the distribution of currents during AHP recovery is not different between the first spike in a cluster and all other spikes regardless of their position ( see below , Figures 7 and S8 ) . Fifth , the conditional spike probabilities ( P ( st|st0 ) ) are sufficient to generate spike trains with interspike interval histograms and clustered patterns of spikes that are indistinguishable from spike trains generated by the biophysical neuronal models ( Figure 6B and 6C ) . Thus , P ( st|st0 ) can fully characterize the spike train . By contrast , if there were higher-order correlations in the spike probabilities , as would be the case in any deterministic model of clustered spiking , then the conditional probabilities would differ for each spike and no single set of conditional spike probabilities would fully characterize the spike train [45] . In principle , the transient increase in the probability of action potential firing that occurs following recovery from the AHP could arise though a number of mechanisms: ( 1 ) A transient shift in the balance of membrane currents that together determine the overall direction and rate of change of the membrane potential; ( 2 ) A change in the stochastic current fluctuations that act as the noise source that enables probabilistic spike firing; or ( 3 ) A reduction in the threshold for spike initiation . To address the first possibility , we evaluated the membrane current at a narrow range of membrane potentials ( −50 . 5 to −49 . 5 mV ) , just below the voltage threshold for spike initiation ( Figure S8 ) . In the stochastic model the membrane potential enters this range during silent epochs when spikes are not initiated , immediately before initiation of the first spike in a cluster and in the epoch following recovery from the AHP when a subsequent spike may or may not be triggered . We therefore assigned each membrane potential measurement to one of three different classes ( Figure 7A ) : a 50 ms windows prior to spike initiation from steady state ( red ) ; during AHP recovery of all spikes without regard to their position within a cluster ( blue ) ; and silent epochs during which no spiking occurred during or in the subsequent 100 ms ( black ) . For each point within these time windows we sampled the membrane current if the membrane voltage was within the range −50 . 5 mV to −49 . 5 mV and then generated histograms of the membrane current for each epoch . Comparison of these three cases revealed that spike initiation from steady state is associated with a small but significant inward shift in the net ionic current relative to periods of silence ( Figure 7A ) . The small shift in the mean current is consistent with the low average firing frequency at steady state ( i . e . low P ( st ) ) . By contrast , the recovery from the AHP is associated with a larger shift ( ∼8 pA ) of the net membrane current in the inward direction ( Figure 7A ) , consistent with the increase in P ( st|st0 ) relative to P ( st ) following AHP recovery and with the shift observed for spikes that initiate clusters ( Figure S8 ) . Thus , during the period following recovery of the AHP , the membrane experiences on balance a greater net inward current at potentials approaching threshold , driving further depolarization of the membrane potential and spiking . We also evaluated whether other mechanisms might contribute to the change in firing probability following recovery from the AHP . Importantly , we found no difference in the standard deviation ( σ ) of the membrane current prior to initiation of spikes from steady state ( σ = 16 . 6 pA ) , compared with AHP recovery ( σ = 16 . 6 pA ) or silence ( σ = 16 . 5 pA ) , indicating that stochastic current fluctuations have a similar magnitude in each condition ( Figure 7A ) . Moreover , there was no correlation between the membrane potential at which we detected spike initiation ( see Materials and Methods ) and the preceding ISI for either the wild-type or knockout models ( R<1×10−4; Figure S7 ) , indicating that the brief elevation in P ( st|st0 ) is not due to an alteration in the voltage threshold following a previous spike . Thus , the shift in average membrane current , as opposed to a change in the stochastic current fluctuations or spike threshold , appears to be the major determinant of increased firing probability following the AHP . Since our experimental and modeling data indicates that HCN channels influence both the AHP and clustered spiking , we asked whether changes in Ih during the AHP could account for the shift in membrane current that underlies the increase in P ( st|st0 ) relative to P ( st ) . Importantly , the shift can be fully explained by an increase in the amplitude of Ih during AHP recovery ( Figure 7B ) . By comparison another current important for spike initiation , the persistent sodium current ( INaP ) , shows no change ( Figure 7C ) . Consistent with this explanation , phase plots for Ih ( Figure 7E ) and INaP ( Figure 7F ) during an action potential , reveal an increased Ih density associated with recovery from the AHP . Are the kinetics of Ih important for the relatively brief increase in P ( st|st0 ) that appears to underlie generation of clustered patterns of activity ( Figures 5 and 6 ) ? Simulated voltage-clamp of isolated Ih using a command potential based upon the action potential waveform ( Figure 8 ) , revealed an increased density of Ih following recovery from the AHP ( Figure 8B ) . Comparison of the observed Ih ( Iobs ) with the current density predicted from the steady-state I–V relationship for Ih ( Iss ) , revealed that while Iobs was less than Iss at time points corresponding with the peak of the AHP , during the return phase of the AHP Iobs is larger than Iss ( Figure 8C amd 8D ) . This transient elevation in Ih relative to steady-state precedes the time course of P ( st|st0 ) with an expected lag for action potential initiation and detection ( Figure 8D ) . To determine if this shift in net membrane current could cause the shift in firing probability , we simulated an increase in the injected current by the peak value of Iobs−Iss . The increase in P ( st ) ( dashed red line; Figure 8D ) during this simulation relative to P ( st ) under the control simulation ( dashed blue line; Figure 8D ) accurately predicts the peak of P ( st|st0 ) . Thus , a brief change in the net inward current due to Ih during the AHP appears to be sufficient to explain the magnitude and time course of P ( st|st0 ) . To directly test the influence of the slow gating kinetics of Ih on action potential clustering we scaled the forward and reverse rates of the closed-open transition of Ih ( Figure S9 ) . While scaling the kinetics did not alter the magnitude of the steady-state current , it did allow Ih to equilibrate to the membrane potential during recovery from the AHP ( Figures 8E and S9 ) and significantly reduced the short-latency ( ∼100 ms ) peak in P ( st|st0 ) ( Figures 8F and S9 ) . This reduction in spike probability following a prior spike resulted in a 33% reduction in Pc for a 1–2 Hz average firing rate . However , changing the kinetics of Ih complicates this analysis and likely leads to an underestimate of the effect . For example , the change in kinetics leads to a 10% reduction in the AHP half-width and increases the stochastic fluctuations in Ih about its mean , both of which effects could increase Pc . Stochastic gating of HCN currents is not necessary for clustered spiking ( Figure S6 ) . Thus , we also ran simulations with fast , deterministic HCN channels to prevent the increase in fluctuations and found that Pc was reduced 40% to 0 . 33 , close to the theoretical minimum of 0 . 29 for a refractory Poisson process where P ( st|st0 ) is equal to P ( st ) ( Figure 6 ) . Together , these data suggest that activation of Ih during the AHP is an important determinant of both the AHP half-width and the clustering of action potentials . Given the relatively slow kinetics of Ih the closing of HCN channels lags the depolarization of the membrane on the tail of the AHP and Ih fails to equilibrate to the membrane potential . As a result , the AHP recovery is associated with a transient increase in Ih relative to steady-state that contributes to an increase in the probability of action potential initiation . Moreover , this effect is robust across a range of channel kinetics tested ( Figure S9 ) . However , due to their relatively small single channel conductance [48] , changes in mean HCN current act primarily as a DC bias current , rather than as a noise source . Could the stochastic model that we outline here also explain aspects of the firing patterns of neurons in behaving animals ? Consistent with this possibility , spike times obtained from in vivo single unit recordings [49] show elevations ( made clear by exponential bin spacing [50] ) in their ISI distribution at around 100 ms ( Figure 9E ) . This ISI resembles the peak of P ( st|st0 ) in simulations of our stochastic model , but unlike the responses of our model to constant current input , the in vivo spike trains contain a much broader overall distribution of ISIs . To provide a more realistic comparison between the model and in vivo data , we therefore carried out stimulations of the response of the model neuron to simulated synaptic drive . To reduce the uncertainty of comparing the model output with in vivo recordings during which the physiologically relevant inputs are unknown , we first examined a wide region of stimulus space by varying the standard deviation and offset of a band-limited , white noise stimulus ( Fmax = 50 Hz ) . In this way , we obtained a description of the relationship between properties of the simulated input to the model and the mean frequencies ( Figure 9A ) and coefficient of variation ( CV; Figure 9B ) of the ISI distributions generated by the spike outputs from the model . Based on comparison of these data with the frequency and CV of spike trains recorded in vivo ( Figure 9C ) , we selected for use in further simulations parameters that generated spike trains with CV and ISI spanning the space covered by the in vivo spike data ( Figure 9D ) . For inputs with a large standard deviation and a small offset there were only small differences between output responses of the stochastic and deterministic versions of the wild-type model ( Figure 9F; χ2-test , P = 0 . 01 ) . In contrast , for inputs with a large offset and small standard deviation , striking differences were apparent between the responses of the stochastic and deterministic models ( Figure 9G; χ2-test , P<0 . 0001 ) . In both cases the stochastic model tends to redistribute the average ISI distribution such that it is enriched for 100–200 ms ISIs , but this effect is greater for the responses to weakly varying inputs ( Figure 9H ) . Unlike the in vivo experimental data , the simulations above did not generate high frequency ( >25 Hz ) bursts of spikes . However , examination of the stimulus space indicated that high variance stimuli with substantial DC offsets could produce spikes at high frequency ( Figure 9A ) . Since recordings of the local field potential in the medial temporal lobe in vivo indicate that the network is characterized by long periods of relatively uncorrelated activity interspersed with brief epochs of highly correlated activity [51] , we attempted to mimic these stimulus statistics by assuming that the stimulus can be characterized by a relatively low average variance ( characteristic of uncorrelated presynaptic activity ) interspersed at random ( Poisson ) delays ( λ = 1 s ) with random duration ( λ = 200 ms ) epochs of high average variance ( characteristic of correlated presynaptic activity ) . This pattern of stimulation is illustrated graphically as a transition between two points in stimulus space ( Figure 9D ) and resulted in a much broader ISI distribution that more closely matched the in vivo data ( Figure 9I ) . Under these stimulus conditions , simulations of the stochastic model also resulted in an ISI histogram enriched for intervals around 100 ms consistent with clustered spiking ( Figure 9J; χ2-test , P<0 . 0001 ) . Finally , we sought to determine predictions the model could make for the in vivo distribution of ISIs for MEC stellate neurons in HCN1 knockout mice . Assuming the stimulus conditions reflect properties of the inputs to the MEC in vivo , the stochastic model predicts that stellate neurons from HCN1 knockout mice should show reduced average firing rates ( FWT = 3 . 77 Hz; FKO = 1 . 01 Hz ) and less clustered firing ( Figure 9K and 9L ) , but an increase in the fraction of spikes emitted in high frequency bursts ( Figure 9K ) . While we have not found evidence for compensatory changes following HCN1 deletion [18] , [27] , [28] , [38] , we nevertheless also considered the possibility that differences in excitability between wild-type and HCN1 knockout mice may be compensated for by homeostatic changes in the average strength of synaptic inputs in vivo [52] . Thus , compensating for the shift in the current threshold for spike firing following HCN1 deletion by altering the average offset amplitude of the simulated in vivo synaptic input , the stochastic model predicts that MEC stellate neurons from HCN1 knockout mice should show a slight shift in the peak of their ISI distribution of approximately +100 ms ( Figure 9K and 9L ) . In addition , over a range of compensation values all of our simulations ( data not shown ) suggest that the peak of the ISI distribution in the range of clustered spiking should actually increase in the knockout mice presumably due to the increased impedance of the membrane near threshold ( Figures 2 , 3 , and S2 ) .
The rules that determine transformation of synaptic input into patterns of spike output are fundamental to computations carried out within the central nervous system . While models of many cortical neurons take advantage of simplifying assumptions that characterize spike output as an invariant function of synaptic input ( e . g . [53] , [54] ) , experimental recordings suggest that stellate neurons from layer II of the MEC generate clustered patterns of spike output through intrinsic mechanisms that may not be reducible in this way . In the biophysical model of a stellate neuron that we develop here , a brief increase in spike probability immediately following recovery from a preceding action potential can substantially modify the pattern of spike output . In the low firing frequency regime , spikes can be initiated by random fluctuations of the net membrane current . As a result of the balance of currents near threshold , the low effective membrane conductance , and the relatively large currents that can be produced by individual ion channels , small bias currents can substantially alter the probability of firing by shifting the mean of the net membrane current . This model is sufficient to explain the clustered patterns of spikes that are recorded from stellate cells during injection of constant current ( Figure 6 ) . This mechanistic account also provides some suggestion that the tendency of neurons in layer II of the MEC in behaving animals to fire spikes at 5–10 Hz may result from the transient , spike-dependent increase in spike probability that can influence spiking even in the presence of a continuously varying barrage of synaptic inputs ( Figure 9 ) . The model that we develop here differs from a number of other models proposed to explain the integrative properties of stellate neurons . Two previous , biophysically-detailed deterministic models have proposed that cyclic interactions between INaP and Ih are necessary and sufficient to produce perithreshold oscillations [25] , [26] . However , this conclusion is not supported by experimental observations from stellate neurons following genetic deletion of HCN1 [18] , or pharmacological block of Ih [22] . One of these previous biophysical models also produces patterned spiking , although quantitative comparisons of the patterns produced with experimental data have not been reported [26] . This previous model requires slow deterministic changes in model parameters to produce clustered patterns of spiking , whereas the model we propose here demonstrates that such slow changes are not necessary for the emergence of clustered spike firing . Nevertheless , it is possible that in entorhinal stellate neurons slow changes in ion channel states could further influence spike firing patterns in addition to the activity-dependent changes in spike probability that we describe here . A conductance-based stochastic model [21] and a more abstract stochastic resonate-and-fire ( SIF ) model [44] have also been developed to account for the properties of stellate neurons . These models successfully account for the complex spectral properties of perithreshold fluctuations of membrane potential that are recorded experimentally and that are also generated by the stochastic model we describe here . It was previously suggested that a simplified , stochastic INaP is sufficient to produce patterened spiking [21] through random threshold crossings and spike omissions [8] , [9] . However , the spiking patterns produced due to a stochastic INaP alone are nearly identical to a stochastic point process with a refractory period and thus do not provide a good match to the patterning observed experimentally ( see Figure 4 in [21] ) . By contrast , the stochastic model we describe here produces more complex spike patterning that is a better match to the characteristics of clustered firing observed experimentally and not well described by a refractory Poisson process ( Figure 6 ) . The mechanism that we suggest for generation of clustered firing patterns also differs markedly from a recently proposed resonate-and-fire model that also reproduces clustered firing patterns of stellate cells [44] . The resonate and fire model explicitly states that sub-threshold resonance mechanisms are required to generate clustered spike patterns , whereas recent experimental studies clearly dissociate sub-threshold resonance from clustered spike firing patterns of stellate cells [18] , [24] . Consistent with this data , the probabilistic model that we propose does not require sub-threshold resonance for generation of clustered spike firing and can provide a mechanistic explanation for dissociation of these two properties . There remain features of the firing patterns recorded experimentally from stellate neurons that are not well captured by any model proposed so far . A striking feature of some stellate neurons is a fairly regular intercluster interval even in the absence of coherent subthreshold oscillations ( e . g . Figure S10 ) . Our model stellate neuron , however , appears to exhibit more widely distributed intercluster intervals . One likely cause of this discrepancy is the simplification of the AHP current used in the model . Indeed , early results demonstrated that blockade of calcium entry can reduce the tendency of spiking to be clustered [15] . Since neuronal morphology can influence patterns of spike output [55] , a further important limitation to the model that we propose here is that it is composed of only a single compartment . On the one hand , this could lead to an underestimate of the influence of stochastic gating , as in an extended dendritic structure fewer ion channels would contribute to ionic currents in any single compartment and thus the influence of stochastic channel gating on the membrane potential would be greater , as has been argued to be the case for thin axons [56] . On the other hand , comparison between our simulations and experimental data suggest that the magnitude of perithreshold oscillations and extent of spike clustering are comparable or perhaps larger in the model . Assessing the contribution of stochastic ion channel gating to the spatially distributed properties of stellate neurons will require future studies with more detailed computational models developed in parallel with more detailed electrical measurements from spatially distinct regions of the neuron . Nonetheless , the general principles that we establish here are likely to be robust to differences in morphology and although further morphological data may improve the similarity between our model data and the experimental data , simple models of neurons , neural circuits , and behavior can provide important functional insights in the absence of exhaustive detail [57] . Analysis of the stochastic model supports a key role for HCN channels in controlling the pattern of spike output from stellate neurons and suggests how the unique biophysical properties of HCN channels enable this role to be achieved . Thus , HCN channels active during the AHP fail to completely deactivate as the membrane potential returns to the steady-state ( Figures 7 and 8 ) . As a result , HCN channels briefly introduce a small bias current that substantially increases the probability of initiating a subsequent action potential ( Figure 8 ) . To the best of our knowledge this is a unique function of HCN channels that depends critically upon both their activation by membrane hyperpolarization and their deactivation kinetics ( Figures 7 , 8 , and S9 ) . Such an interaction , between a bias current introduced by a slowly gating ionic current with a small single channel conductance such as Ih and rapidly varying currents composed of ion channels with larger single channel conductances , may be a general mechanism by which neurons produce changes in firing properties that pattern action potential output . Importantly , under naturalistic stimulus conditions , patterned spiking in the stochastic model can still provide significant modifications to the response properties of stellate neurons ( Figure 9 ) . Several neuronal subtypes have been reported to display perithreshold oscillations of membrane potential [58] , [59] . If intrinsic oscillations in other neurons also arise from stochastic channel gating , then patterned action potential firing driven by the interaction between multiple stochastic currents may also be a more general feature of neuronal spiking . The entorhinal cortex is the last stage at which cortical information is processed prior to entering the hippocampal formation . Stellate cells in layer II constitute a major excitatory projection to the dentate gyrus and may correspond to the recently discovered ‘grid cells’ , which encode an animal's location in its environment through grid-like spatial firing fields [49] , [60] , [61] . While unreliable synaptic transmission is often considered as a noise source in neural circuits [6] , less attention is usually given to the possible impact of stochastic ion channel fluctuations . Using stimulus parameters selected to obtain output firing properties similar to those recorded in vivo , we found that the presence of stochastically gating ion channels reliably increased the number of action potentials emitted with an ISI characteristic of intra-cluster intervals ( Figure 9 ) . This tendency depended on the stimulus statistics used , but is consistent with the peak in the in vivo ISI histograms around 100 ms and with our explanation for clustering as a transient increase in spike probability during the ∼70–150 ms following an action potential . Since the trains of synaptic stimuli used for these simulations have random statistics , these data support the idea that the effects of stochastic ion channel gating may in some conditions be superimposed on , rather than overwhelmed by , synaptic noise sources . Thus , stochastic ion channel gating may have to be accounted for in order to explain the firing of grid cells in behaving animals . However , further evaluation of this hypothesis will require much more information about the actual synaptic inputs received by grid cells . In addition , to better compare in vitro and in vivo data future studies will be required to establish whether in vivo data sets obtained from superficial layers of the MEC are indeed enriched for stellate neurons [49] , [60] . We also attempted to predict the responses of wild-type and HCN1 knockout neurons to naturalistic stimuli . These simulations suggested that in the absence of changes to the input stimulus , stellate neurons lacking HCN1 will have an approximately 65% reduction in average firing rate ( Figure 9 ) . This reduced firing rate is characterized by an increase in the fraction of spikes emitted in high frequency bursts , or a sparsening of the response properties . There have been several suggestions that high frequency bursts convey unique information [62]–[64] about input stimuli and thus , this change could contribute to the enhancement of hippocampus-dependent learning in mice with deletion of HCN1 channels [28] . Whereas initiation of action potentials in deterministic model neurons is a binary process with a clearly definable threshold , in more realistic neuronal models containing stochastically gating ion channels spike initiation is probablistic . Here we show that one general consequence of stochastic ion channel gating is that firing of an action potential can transiently modify the spike probability leading to the emergence of intrinsically generated patterns of spike output . In the case of the model we develop here , activation of HCN channels , during recovery from the action potential afterhyperpolarization , drives a brief increase in spike probability that leads to the emergence of clustered patterns of spike firing . As well as providing an account of both the resting and active integrative properties of stellate neurons in the medial entorhinal cortex , analysis of responses of this model to simulated in vivo synaptic inputs , suggests conditions in which stochastic ion channel gating might impact firing patterns of behaving animals . Thus , our results suggest a mechanism by which random changes in the conformation of small numbers of individual ion channel proteins could impact neural computations that underlie cognitive processes such as spatial navigation and memory .
Modeling experiments were implemented in Matlab 7 ( Natick , MA ) using kinetic formalisms described in Text S1 . The model has also been completely replicated in NEURON 5 . 9 , but Matlab simulations were used for the data reported . The model cell was a sphere with a diameter of 50 µm and a specific capacitance of 1 . 67 µF/cm2 ( to account for the lack of a dendritic arbor ) . The model included implementations of a fast , transient sodium current ( NaT ) , a persistent sodium current ( NaP ) , a delayed rectifier-type postssium current ( Kdr ) , a fast inactivating A-type potassium current ( KaF ) and a slowly inactivating potassium current ( KaS ) , a “calcium-activated” potassium current ( KCa ) , a linear potassium leak ( Kl ) and a fast or slow hyperpolarization-activated current ( Hf or Hs ) . Hf , Hs and KCa are implemented as two-state channels , which is sufficient to capture their dominant kinetics , although additional states would be required to more fully capture details of their gating . NaP , KaF , and KaS , were modeled with a cyclical four state inactivation model . NaT and Kdr currents were modeled according to the original Hodgkin-Huxley formalism with 5 and 8 states , respectively . The total current density of each channel was closely matched to existing data . In order to model stochastic channels , it was assumed that the states obeyed a first order Markov-type probabilistic description [2] . To track channel populations in each state a random number was generated for each channel in a given state ( a “particle” ) at each time step ( Δt ) . Assuming that the time step is sufficiently small the probability of a transition is equal to rate×Δt , with a transition occurring in the event that a random number , evenly distributed between 0 and 1 is less than rate×Δt . For particles with multiple possible transitions ( i . e . multistate channels that have multiple transitions into and out of a given state ) , a unique transition was chosen using non-overlapping distributions of transition probabilities . Briefly , a uniformly distributed collection of random numbers between zero and one , thresholded by the value P ( transition ) will give N , the number of transitions that occur . In the case where multiple transitions are possible , we observe that a given “particle” can only undergo a single transition . We know from probability theory that:However , if there can only be a single transition then:and thus , The probability that a given transition occurs is then the sum of the elementary probabilities . Dividing the probability space between 0 and 1 into bins of size P ( A ) , P ( B ) and 1- P ( A∪B ) , and placing random variables uniformly distributed between 0 and 1 , gives the desired values for the number of transitions . This brute force method is similar to the simple Monte Carlo method described elsewhere [65] and to the method used elsewhere to model stochastic channels [9] . The time step used was 10 µs ( corresponding to the approximate minimum dwell time of NaT ) and numerical integration was accomplished using a 4th order Runge-Kutta method ( most results were confirmed using the Backward Euler integration method ) . Simulations were run in Matlab and all analysis was completed using Igor Pro ( Wavemetrics; Eugene , OR ) . A complete description of parameters used for the model currents and justification of parameters can be found in Text S1 . Further , each channel was implemented as either stochastic or deterministic and it was ensured that in all cases the two solutions converged . For some simulations a partially stochastic model was used to speed simulation times and provide a good estimate of the fully stochastic model ( data in Figures 4 and 9 ) . This was justified by directly examining the contribution of each conductance ( Figures S4 and S6 ) . Throughout the text we have made reference to a number of descriptions of the biophysical properties of the neuron that are elaborated upon here for clarity . The passive membrane properties we characterize are the resting membrane conductance and resting membrane potential . Typically these values are obtained by analyzing the response of the membrane potential to small current steps . By convention we assume that the state of the voltage-dependent currents is unaltered . The values are then obtained by application of Ohm's Law . However , during active states , when the neuron or model is depolarized away from its resting potential , the assumption that the underlying conductances are unaltered by small changes in injected current are generally less safe . At depolarized potentials we use a modified definition of the membrane conductance and consider the “effective” membrane conductance . Here we define the effective membrane conductance as the slope of the relationship between the membrane current and the membrane potential . This definition thus explicitly takes in to account the change in membrane conductance in response to a change in membrane voltage [2] , [5] . All simulation data were analyzed in IGOR Pro ( Wavemetrics ) using both built-in analysis functions and custom written routines . Unless indicated otherwise mean values are ±standard error of the mean ( SEM ) . Statistical tests were accomplished using Excel ( Microsoft ) and IGOR Pro ( Wavemetrics ) . Analysis of in vivo recordings of cortical neurons from the superficial layers of the medial entorhinal cortex was based upon data obtained from: http://commonweb . ntnu . no/cbm/moser/gridcell . For Figure 9 we attempted to provide a general , readily parameterized model of synaptic drive that might occur in vivo . Because we used a single compartment model , appropriately scaled current stimulation can be equivalent to conductance-based stimuli [66] . Further , in order to provide a readily parameterized stimulus to explore the space of possible responses we chose to use colored white noise . Again , over the range of frequencies where the impedance of the cell membrane is maximal , random barrages of synaptic input show approximately white stimulus statistics [66] . For our stimulus we thus create a broadband , white noise stimulus that was bandlimited to 50 Hz . The standard deviation and DC offset of the current stimulus were scaled according to the parameters in Figure 9A and 9B and applied directly to the model . The ISI histogram of responses to the broadband stimulus was not as broad as for the in vivo experimental data . Examining the experimental data revealed that this was primarily due to a lack of high-frequency ( >50 Hz ) bursting in the model . We made the assumption that occasional changes in stimulus statistics could give rise to this high frequency bursting . By examining the approximate length of such periods we determined that ∼200 ms long changes in stimulus statistics were consistent with the experimental data . We assumed a Poisson distribution for the duration of these epochs of high frequency activity . We chose an interval between the high frequency epochs that gave an approximately correct balance in the ISI distribution ( mean = 1 s; Poisson distributed ) . Finally , the amplitude of the changes in the DC component and standard deviation were taken from the survey of parameter space to match the central peak of the bursting ISIs ( see Figure 9 ) . | Neurons use electrical impulses called action potentials to transmit signals from their cell body to their axon terminals , where the impulses trigger release of neurotransmitter . Initiation of an action potential is determined by the balance of currents through ion channels in a neuron's membrane . Although it is well established that membrane ion channels randomly fluctuate between open and closed states , most models of action potentials account for the average current through these channels but not for the current fluctuations caused by this stochastic opening and closing . Here , we examine the consequences of stochastic ion channel gating for stellate neurons found in the entorhinal cortex . The intrinsic properties of these neurons cause characteristic clustered patterns of spiking . We find that in a model of a single stellate neuron that is constrained by previous experimental data clustered action potential patterns are produced only when the model accounts for the random opening and closing of individual ion channels . This stochastic model provides an example of a general mechanism for patterning of neuronal activity and may help to explain the patterns of spikes fired by entorhinal neurons that encode spatial location in behaving animals . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biophysics/theory",
"and",
"simulation",
"computational",
"biology/computational",
"neuroscience",
"neuroscience/neuronal",
"signaling",
"mechanisms",
"neuroscience/neuronal",
"and",
"glial",
"cell",
"biology",
"neuroscience/theoretical",
"neuroscience"
] | 2009 | Stochastically Gating Ion Channels Enable Patterned Spike Firing
through Activity-Dependent Modulation of Spike Probability |
Cystic echinococcosis ( CE ) is a complex disease caused by Echinococcus granulosus ( E . granulosus ) , and its immunophatogenesis is still not clearly defined . A peculiar feature of chronic CE is the coexistence of Th1 and Th2 responses . It has been suggested that Th1 cytokines are related to disease resistance , whereas Th2 cytokines are related to disease susceptibility and chronicity . The aim of this study was to evaluate , by multi-parametric flow cytometry ( FACS ) , the presence of CE specific immune signatures . We enrolled 54 subjects with suspected CE; 42 of them had a confirmed diagnosis , whereas 12 were classified as NO-CE . Based on the ultrasonography images , CE patients were further categorized as being in "active stages" ( 25 ) and "inactive stages" ( 17 ) . The ability of CD4+ T-cells to produce IFN-γ , IL-2 , TNF-α , Th2 cytokines or IL-10 was assessed by FACS on antigen-specific T-cells after overnight stimulation with Antigen B ( AgB ) of E . granulosus . Cytokine profiles were evaluated in all the enrolled subjects . The results show that none of the NO-CE subjects had a detectable AgB-specific response . Among the CE patients , the frequency and proportions of AgB-specific CD4+ T-cells producing IL-2+TNF-α+Th2+ or TNF-α+Th2+ were significantly increased in the “active stages” group compared to the “inactive stages” group . Moreover , an increased proportion of the total polyfunctional subsets , as triple-and double-functional CD4 T-cells , was found in CE patients with active disease . The response to the mitogen , used as a control stimulus to evaluate the immune competence status , was characterized by the same cytokine subsets in all the subjects enrolled , independent of CE . We demonstrate , for the first time to our knowledge , that polyfunctional T-cell subsets as IL-2+TNF-α+Th2+ triple-positive and TNF-α+Th2+ double-positive specific T-cells associate with cyst biological activity . These results contribute to increase knowledge of CE immunophatogenesis and the disease outcome in terms of control and persistence .
Cystic echinococcosis ( CE ) is a widespread zoonosis caused by the larval stage of the tapeworm Echinococcus granulosus ( E . granulosus ) [1] . CE is also a complex disease , and several aspects , such as its natural history , parasite-host interplay , poor response to treatment , and predisposition to persistence are still not clearly defined . An important question is how the parasite may influence the quality of the host’s immune response . A peculiar feature of chronic CE is the coexistence of Th1 and Th2 responses . It has been suggested that Th1 cytokines are related to disease resistance and in contrast , Th2 cytokines are associated with disease susceptibility and chronicity [2]; high levels of Th1 cytokines are found in patients who were successfully responding to treatment , whereas high levels of IL-4 and IL-10 occur in patients who did not [3–5] . This result indicates that the IL-10/IL-4 endogenous production induced by CE may impair Th1 response , allowing for E . granulosus persistence [6] . The nature and amount of antigens released by the parasite may play key roles in these immunoregulation mechanisms . For instance , E . granulosus Antigen B ( AgB ) , one of the most abundant antigens in the hydatid cyst fluid , modulates the host’s response , inhibiting neutrophil recruitment [7 , 8] and altering dendritic cell maturation to prime T lymphocytes into a non-protective Th2 response [9] . Notably , AgB skewed Th1/Th2 cytokine ratios towards a preferentially Th2 polarization , mainly in patients with active stages [8 , 10 , 11] . However , despite the high number of studies on the immune response induced by E . granulosus antigens , a comprehensive analysis of the ability of AgB-specific T-cells to co-express multiple functions has not yet been performed . Better understanding of the induction of multifunctional T-cells in the human disease may help to clarify the disease outcome , as also shown in other diseases such as HIV and TB [12–16] . This could facilitate the development of new diagnostic tools and/or the clinical management of CE patients . Therefore , the aim of this study was to simultaneously characterize the E . granulosus-specific immune response in terms of cytokine production by flow cytometry in peripheral blood mononuclear cells ( PBMC ) derived from prospectively enrolled CE patients with active and inactive disease after in vitro stimulation with AgB .
Patients admitted to the “L . Spallanzani” National Institute for Infectious Diseases ( INMI ) and Sant’Andrea Hospital with suspected CE [risk factors for CE at the interview ( Table 1 ) and the presence of abdominal or lung cysts at the time of the visit or in the past] were evaluated for enrollment . CE was diagnosed based on the characteristics of images [ultraonography ( US ) , nuclear magnetic resonance or both] , and serology as a confirmatory test . Information regarding demographic data , risk factors for CE , laboratory data , symptoms , treatment and cyst description were collected . Hydatid cysts were staged according to the WHO classification [17] . Patients having multiple cysts were classified according to the more active stage [18] . The CE patients with active ( CE1 and CE2 ) and transitional ( CE3a and CE3b ) cysts were considered as a whole group because no significant differences were found in terms of the IL-4 specific immune response detected in whole blood ( p = 0 . 1 ) [11] . Patients were then further classified into the categories of “active stages” ( active and transitional cysts ) and “inactive stages” ( inactive cysts ) ( Fig 1 ) . The study was approved by the Ethical Committees of INMI ( parere 34/2010; parere 28/2014 ) and the Sant’Andrea Hospital ( Prot . C . E . n . 436/11 ) , and all enrolled individuals provided written informed consent . The following stimuli were used for PBMC stimulation: native AgB at 10 ug/mL , ( produced at the Istituto Superiore di Sanità , as previously reported in [9] ) , costimulatory molecules anti-CD28 and anti-CD49d monoclonal antibodies ( mAb ) at 1ug/mL each ( BD Bioscence , San Jose , USA ) , staphylococcal enterotoxin B ( SEB ) at 200 ng/mL ( Sigma , St . Louis , MO , USA ) . The fluorescently conjugated mAb used in this study were: AQUA DYE- AmCyan ( Invitrogen Life Technology , Monza , IT ) , anti-CD4 peridinin chlorophyllprotein ( PerCP ) -Cy5 . 5-conjugated ( Miltenyi Biotec S . r . l . , BO , Italy ) , anti-CD3 allophycocyanin ( APC ) H7-conjugated ( Miltenyi ) , anti-TNF-α phycoerythrin ( PE ) -Cy7-conjugated ( eBiosceience , San Diego , CA , USA ) , anti- IFN-γ Horizon V450-conjugated ( BD Biosciences ) , anti-IL-2 fluorescein isothiocyanate ( FITC ) -conjugated ( BD Biosciences ) , anti-IL-4 PE ( BD Biosciences ) , anti-IL-5 PE ( Biolegend , San Diego , CA , USA ) , anti-IL-13 PE ( Biolegend ) , anti-IL-10 allophycocyanin ( APC ) -conjugated ( BD Biosciences ) . Heparinized WB was collected and processed within 2 hours . PBMC were isolated by standard methods on Ficoll-Paque Plus ( GE Healthcare Bio-Sciences AB , Uppsala , Sweden ) and incubated with stimuli at 37°C . Brefeldin at 10ug/mL ( Sigma ) was added after 1 h or 20 h of stimulation . ICS was performed after 24 h of incubation . Unstimulated cells were used as a negative control . PBMC were stained for vitality and then fixed in 2% paraformaldehyde . Therefore , the cells were resuspended in the PBS-2% FCS-0 . 5% saponin-2mM EDTA-1% FcR- binding inhibitor ( eBioscience ) buffer and stained with mAbs for surface markers and intracellular cytokines . At least 300 , 000 events were acquired using a FACSCanto II flow cytometer ( BD Biosciences ) . Multiple-parameter flow cytometry data were analyzed using FlowJo ( Tree Star Inc . , San Carlos , CA ) and SPICE software ( provided by Dr . Roederer , Vaccine Research Center , NIAID , NIH , USA30 ) . Cells were gated according to forward and side scatter plots and the frequency of single , double , triple , quadruple and quintuple cytokines producing CD4+ T-cells was evaluated using boolean combination gates . As the anti-IL-4 , anti-IL-5 , anti-IL-13 mAbs were conjugated with the same fluorochrome , we evaluated these cytokines as a whole , identifying them as “Th2 cytokines” . After subtracting the background values , the total cytokine production and the different cytokine subsets were expressed as frequency or percentages ( proportions ) of the total cytokine response . The positive CD4+ T-cell response was defined as the production of any cytokines ( IFN-γ and/or IL-2 and/or TNF-α and/or Th2 cytokines and/or IL-10 ) , with 0 . 03% as the detection limit corresponding to at least 30 analyzed events . Functional characterization of the cytokine-producing subsets was performed only in subjects with a positive AgB cytokine response . The FACS results were generated by LP and blindly re-evaluated by a co-author , EP . The agreement of the results was high ( k = 0 . 9 ) and the discrepancies were solved by discussion . Data were analyzed using SPSS v . 20 for Windows ( SPSS Italia SRL , Bologna , Italy ) and Prism 6 software ( Graphpad Software 6 . 0 , San Diego , CA , USA ) . Medians and interquartile ranges ( IQR ) were calculated for continuous measures; chi square for dichotomous measures . The Kruskal-Wallis test and Mann-Whitney U test were used for comparisons among several groups or pairwise comparisons , respectively . Bonferroni correction was used if needed . P values as ≤0 . 05 or as ≤0 . 016 after the Bonferroni correction were considered significant .
Between April 2013 and May 2015 we prospectively enrolled 54 subjects ( Fig 1 ) . Among them , 42 ( 77 . 8% ) had a confirmed CE diagnosis whereas 12 ( 22 . 2% ) were classified as “NO-CE subjects” , having cysts that were not related to CE . Based on the cyst stage activity , the CE patients were further classified into “active stages” [25 ( 59 . 5% ) ] or “inactive stages” [17 ( 40 . 5 ) ] groups . Demographic and clinical features are shown in Table 2 . CE patients were mainly Italian , coming from the central regions [25 ( 73 . 5% ) ] . Serology was scored positive in 31 patients ( 73 . 8% ) . The 11 subjects who scored negative were characterized by a US , showing mainly inactive cysts ( CE4 and CE5 ) [6 ( 54 . 5% ) ] . Seventeen ( 40 . 5% ) CE patients were treated with albendazole ( ABZ ) prior to inclusion in the study , whereas 16 patients ( 38 . 1% ) were going to start ABZ after blood collection . More than 40% of the evaluated cysts were small ( diameter <5 cm ) , with a hepatic localization . Among the enrolled patients , only 6 ( 14 . 3% ) had a farming-related job , although the majority of them [35 ( 83 . 3% ) ] reported risk factors , such as contact with shepherd dogs . Twenty-five ( 59 . 5% ) CE patients reported symptoms , with abdominal discomfort being the most common symptom recorded . The 12 NO-CE subjects were sex-and age-matched donors . As the CE patients , they came from Italy ( Table 2 ) , mainly from the central regions . The serology was scored negative in almost all of the subjects with the exception of 1 person . To evaluate the AgB-specific T-cell responses , the cytokine profiles were assessed by ICS . All patients scored positive to the mitogen ( SEB ) . To evaluate the AgB-specific responses , we focused our analysis on the CD4+ T-cells , as in the set-up experiments , performed on a limited number of CE patients , we did not detect any CD8 T-cell specific response ( none of the subjects tested ) . The ability of CD4+ T-cells to produce IFN-γ , IL-2 , TNF-α , Th2 cytokines or IL-10 was assessed in all the enrolled subjects , however , none of the NO-CE subjects had a detectable AgB-specific response . Among the CE patients , the magnitude of the cytokine response to AgB ( considering the production of any cytokine ) was higher in the “active stages” group ( median: 0 . 06 , IQR: 0–0 . 3 ) compared to the CE patients included in the “inactive stages” group ( median: 0 . 02; IQR: 0–0 . 2 ) , although it was not significant ( p = 0 . 8 ) ( Fig 2A ) . In addition , the proportion of responders to AgB was also higher in the “active stages” group than in the “inactive stages” patients ( 56% vs 47 . 1% ) ; however , the difference was not significant ( Fig 1 ) . We investigated the AgB-specific CD4+ T-cells in terms of IFN-γ , IL-2 , TNF-α , IL-10 and Th2 cytokines frequency independently of the simultaneous production of the cytokines ( Fig 2B ) . The total response in the active stages group is characterized by the production of IL-2 , TNF-α and Th2 cytokines , whereas in the “inactive stages” group it is characterized by the production of IL-10 in addition to IL-2 , TNF-α and Th2 cytokines . However , no significant differences were found for any of the comparisons performed . To better define the specificity of the results obtained , we compared the AgB cytokine response with that elicited by the positive control SEB ( Fig 2C ) . The “total cytokine response” to SEB was mostly characterized by IFN-γ , IL-2 and TNF-α cytokines in all evaluated subjects . To note , both the CE patients and the NO-CE subjects had the same cytokine profile in response to the SEB antigen . A boolean gating analysis was then performed to categorize cytokine-positive cells into 31 different subsets consisting of quintuple , quadruple , triple , double or single cytokine-expressing populations . The frequency of AgB-specific CD4+ T-cells characterized to be IL-2+TNF-α+Th2+ ( triple- positive ) or TNF-α+Th2+ ( double-positive ) was increased in the “active stages” group compared to the “inactive stages” group ( p = 0 . 02 and p = 0 . 006 , respectively ) ( Fig 3A ) . Similar results were found when the proportions of the AgB-specific response were analyzed ( p = 0 . 03 and p = 0 . 008 , respectively ) ( Fig 4A ) . Moreover , the monofunctional CD4+ subset producing Th2 cytokines was increased in the “active stages” group compared to the “inactive stages” group , although the difference was not significant ( Figs 3A and 4A ) . In contrast , the “inactive stages" group showed a higher frequency of the IL-10 monofunctional CD4+ T-cells subset compared to the “active stages” group . However , the difference was not significant ( Fig 3A ) . To better define the specificity of the result obtained , we compared the functional profile of the AgB cytokine response with that elicited by SEB . The SEB response , evaluated as frequency or proportion , was mostly characterized by the same cytokine subsets in all the subjects enrolled , independent of the CE status ( Figs 3B and 4B ) . No significant differences were found for any of the comparisons performed in response to SEB among the two groups analyzed . All these data suggest that only the response to E . granulosus antigens as AgB-specific triple functional IL-2+TNF-α+Th2+ cytokines and double functional TNF-α+Th2+ cytokines CD4+ T-cells associated with cyst biological activity . Finally , we evaluated if the proportion of the monofunctional and polyfunctional subsets are differently represented in the two groups of CE patients evaluated . A trend for higher proportions of cells exerting 2 or 3 functions was found in the “active cysts” group compared to the “inactive cysts” group ( Fig 5 ) , supporting the previous data . In contrast , the proportion of the monofunctional subsets was similar between the two groups ( Fig 5 ) .
In this prospective study , we characterize , for the first time to our knowledge , the specific immune response to AgB of E . granulosus in patients with the active and inactive clinical forms of CE . We demonstrate that IL-2+TNF-α+Th2+ triple-positive AgB-specific CD4+ T-cells and TNF-α+Th2+ double-positive AgB-specific CD4+ T-cells associate with cyst biological activity , being significantly increased in the active CE stages . We also found an increased ( but not significant ) proportion of the total polyfunctional subsets in CE patients with active disease . Altogether , these results help to contribute to the knowledge of the CE immunophatogenesis , as well as the mechanisms associated with disease control and persistence . Moreover , comprehension of the pathways involved in host protection or parasite survival could help to find biomarkers for developing new tools for CE diagnosis and/or therapy monitoring . In the last few years , polyfunctional T-cells have been intensively studied in viral , bacterial and parasitic diseases . In chronic HIV [19–21] and hepatitis C [22] viral infections , polyfunctional T-cells have been correlated with the immune protection . In contrast , their role during bacterial chronic infections such as tuberculosis is still controversial [13–16 , 23] and there is currently no consensus whether polyfunctional T-cells represent a marker of protective immunity or disease activity . Similar to viral infections , during protozoan Th1-mediated parasitic diseases , such as Leishmaniasis or Malaria , polyfunctional T-cells have been suggested to have a role in the induction of a protective immunity [24–27] . In Chagas disease , Trypanosoma cruzi-infected children , at early stages of infection , displayed mainly double- or triple-functional CD4+ T-cells whereas chronically infected adults showed monofunctional T-cell specific-responses [28] . In agreement with these findings , in the present study we found that the polyfunctional T-cell subsets producing Th2 cytokines associate with the active stages of CE . These results suggest that the cells characterized by a superior functional capacity are linked to an increased biological cyst activity rather than to a protective role . Regarding the Th2 monofunctional T-cell subset , both the frequency and proportion were found increased in CE patients with active stages . These results are in agreement with the finding that active and transitional cysts are characterized by elevated IL-4 levels either in serum [29] or in AgB-stimulated blood [11] . Moreover , in patients with inactive stages , the IL-10 monofunctional T-cells subset was increased , suggesting that this cytokine has a role in parasite persistence , as previously speculated [6] . Additional studies on a larger cohort of CE patients may help to clarify if these T-cell subsets could be considered as a signature of active stages and inactive stages , respectively . The potential limits of this study should be considered . First , we performed a cross-sectional study , analyzing a relatively small number of subjects within each group . Moreover , the small sample size hampered us from performing any intra-group analysis , evaluating each WHO CE stage . However , although a larger population size is needed to confirm these observations , the results generated here seem to be robust , as confirmed by the functional profile obtained in all the subjects enrolled , independent of CE status in response to the control stimulus SEB . In addition , the AgB cross-reactions have not been evaluated; the antigen specificity was tested in NO-CE subjects , as the prevalence of Alveolar Echinococcosis or Taeniasis is low in Italy . Finally , the analysis was restricted to those scored positive to AgB , who are not all CE patients . However , this is a limit of all the immune-based assays that measure antigen-specific responses [15 , 16 , 30] . The use of more antigenic molecules or peptides or different readouts or biological samples different from blood [31–34] may overcome this issue and we are currently working on this . In conclusion , we demonstrated , for the first time , that polyfunctional T-cells subsets as IL-2+TNF-α+Th2+ triple-positive and TNF-α+Th2+ double-positive specific T-cells associate with biological cyst activity . Although additional studies on patients successfully responding to chemotherapy , or on patients followed over time are needed for a complete understanding of the role polyfunctional T-cells play in CE , these results may contribute to better characterizing CE immune responses and may open the door to new opportunities for generating tools for CE diagnosis and treatment monitoring . | Cystic echinococcosis ( CE ) is a widespread zoonosis caused by the tapeworm Echinococcus granulosus ( E . granulosus ) . CE is a complex disease , and several aspects of its immunophatogenesis are still not clearly defined . An important question is how the parasite influences the quality of the host’s immune response . A peculiar feature of chronic CE is the coexistence of Th1 and Th2 responses , and Th1 cytokines are related to disease resistance , whereas Th2 cytokines are related to disease susceptibility and chronicity . In the last few years , polyfunctional T-cells have been intensively studied in viral , bacterial and parasitic diseases to better understand if they represent a marker of protective immunity or disease activity . In the present study it is shown that the polyfunctional T-cell subsets producing Th2 cytokines associate with the active stages of CE . These results suggest that the cells characterized by a superior functional capacity are linked to an increased biological cyst activity rather than to a protective role . These results may contribute to increase the knowledge of CE immunophatogenesis and the disease outcome in terms of control or persistence . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Polyfunctional Specific Response to Echinococcus Granulosus Associates to the Biological Activity of the Cysts |
Trp-cage is a designed 20-residue polypeptide that , in spite of its size , shares several features with larger globular proteins . Although the system has been intensively investigated experimentally and theoretically , its folding mechanism is not yet fully understood . Indeed , some experiments suggest a two-state behavior , while others point to the presence of intermediates . In this work we show that the results of a bias-exchange metadynamics simulation can be used for constructing a detailed thermodynamic and kinetic model of the system . The model , although constructed from a biased simulation , has a quality similar to those extracted from the analysis of long unbiased molecular dynamics trajectories . This is demonstrated by a careful benchmark of the approach on a smaller system , the solvated Ace-Ala3-Nme peptide . For the Trp-cage folding , the model predicts that the relaxation time of 3100 ns observed experimentally is due to the presence of a compact molten globule-like conformation . This state has an occupancy of only 3% at 300 K , but acts as a kinetic trap . Instead , non-compact structures relax to the folded state on the sub-microsecond timescale . The model also predicts the presence of a state at of 4 . 4 Å from the NMR structure in which the Trp strongly interacts with Pro12 . This state can explain the abnormal temperature dependence of the and chemical shifts . The structures of the two most stable misfolded intermediates are in agreement with NMR experiments on the unfolded protein . Our work shows that , using biased molecular dynamics trajectories , it is possible to construct a model describing in detail the Trp-cage folding kinetics and thermodynamics in agreement with experimental data .
Understanding protein folding thermodynamics and kinetics is a central issue in molecular biology [1]–[3] and computer-aided modeling is becoming increasingly useful also in this field . Direct comparison between simulations and experiments requires both an accurate description of the system and the possibility to sample extensively the configuration space . In order to observe folding with molecular dynamics , it is necessary to use very large computers [4] , [5] , worldwide distributed computing [6] , or an enhanced sampling technique [7]–[16] . A system that is almost ideal for theoretical investigation is the Trp-cage ( TC5b ) [17] , a designed 20-residue miniprotein that folds rapidly [18] and spontaneously to a globular structure . The NMR structure ( 1L2Y ) [17] reveals a compact hydrophobic core , in which the Trp side chain is buried . The secondary structure elements include a short ( residues 2–8 ) , a 310-helix ( residues 11–14 ) and a polyproline II helix at the C-terminus . The folding mechanism of this system has been studied with several experimental techniques . Calorimetry , circular dichroism spectroscopy ( CD ) [19] and fluorescence [18] show a cooperative two-state folding behavior with transition midpoint at approximately 314 K and a relaxation time of 3 . 1 µs at 296 K [18] . UV-Resonance Raman [20] reveals a more complex unfolding behavior , with the presence of a compact intermediate that retains an character and in which the hydrophobic core is even more compact . NMR experiments [17] , [21] show a substantially cooperative thermal unfolding , but the large negative chemical shift deviations of and suggest that those residues might pack more tightly as the temperature is raised . Also fluorescence correlation spectroscopy experiments cannot be interpreted in terms of a simple two-state folding and the formation of a molten-globule-like intermediate has been proposed [22] . By atomistic modeling the Trp-cage folding has been studied using several different approaches [23]–[33] . In particular , with an all-atom explicit-solvent description , the folding of Trp-cage has been studied by replica exchange molecular dynamics ( REMD ) [31] , [34] . Starting from an extended configuration , a structure with a root mean square deviation ( RMSD ) <2 Å from the NMR reference structure is obtained after 100 ns of simulation on 40 replicas [34] . A relatively high melting temperature of 440 K is predicted . Other studies suggested that , even if Trp-cage is a rather small system , achieving statistical convergence in a REMD simulation may require much longer simulation times [35] , [36] . The kinetics of Trp-cage folding was studied , in explicit solvent , by transition path sampling ( TPS ) [36] and transition interface sampling ( TIS ) [37] . The folding of Trp-cage was also investigated by two of us using the bias exchange metadynamics approach ( BE ) [38] , in which metadynamics potentials acting on different collective variables ( CVs ) are exchanged among molecular dynamics ( MD ) simulations performed at the same temperature . Using this method it is possible to explore simultaneously a virtually unlimited number of CVs . Since all the MD simulations are performed at the same temperature the number of replicas does not grow with the system size like in REMD and in the approach of Ref . [39] . Using BE it was possible to reversibly fold Trp-cage [38] , villin headpiece , advillin headpiece together with two of their mutants [40] and Insulin chain B [41] using an explicit solvent force field , in less than 100 nanoseconds of simulation with only eight replicas . Recently this method was also used for exploring the mechanism of enzyme reactions [42] . In atomistic simulations of biological systems , after an exhaustive exploration is achieved , it is necessary to extract from the trajectory the relevant metastable conformations , to assign their occupation probability , and to compute the rates for transitions among them . Several methods have been developed for this scope [43]–[48] . These methods have the big advantage of reducing a complex dynamics in a high-dimensional configuration space to a Markov process describing transitions among a finite number of metastable states . They are suitable for analyzing an ergodic molecular dynamics trajectory , but they cannot be straightforwardly applied if the system is evolved under the action of an external bias . In this paper we present a method that allows exploiting the statistics accumulated in a bias exchange metadynamics run [38] for constructing a detailed kinetic and thermodynamic model of a complex process such as the Trp-cage folding . The approach presented here aims at extracting the same information from a BE simulation as one can obtain from the analysis of a long ergodic MD run or of several shorter runs [43]–[48] . The method relies on the projection of the BE trajectory on the space defined by a set of variables , which are assumed to describe the relevant physics of the system . These variables are not necessarily the ones that are used for the BE simulation and can be chosen . Once the CVs are selected , the rate model is constructed following three steps: The model constructed in this manner is designed to optimally reproduce the long time scale dynamics of the system . It can be used , for example , for characterizing the metastable misfolded intermediates of the folding process . The advantage of using biased trajectories , besides the acceleration of slow transitions , is a greatly enhanced accuracy of the estimated free energy at transition state regions . This approach is first illustrated on the Ace-Ala3-Nme peptide ( hereafter Ala3 ) . This system is simple enough to allow benchmarking the results against a long standard MD simulation . For this system the model is capable of reproducing with excellent accuracy the kinetics and thermodynamics observed in the unbiased run . The same approach is then applied to the Trp-cage miniprotein . A model is built that allows describing the folding process , computing the folding rates and the NMR spectra , simulating a T-jump experiment , etc . The scenario that emerges is in good agreement with the available experimental data . By kinetic Monte Carlo ( KMC ) [53] , [54] and Markov cluster analysis ( MCL ) [51] , [52] several metastable sets ( clusters ) are identified . These states , except for the folded cluster , can be considered misfolded intermediates of the folding process . At 298 K two main clusters are present , with a population of 58% and 25% , respectively . The most populated is the folded state and its structural properties are very close to the NMR ensemble . The second most populated cluster retains a significant amount of secondary structure , but has a from the native state of approximately 4 . 4 Å . In this cluster , the Trp is trapped in a hydrophobic pocket and its distance from Pro12 and Gly11 is reduced . The presence of this cluster in the thermal ensemble of the system can explain some anomalies in the temperature behavior observed in NMR [17] and UV-Raman [20] experiments . The structures of the most populated misfolded intermediates are in good agreement with the unfolded states distances reported in Ref . [21] . Using the kinetic model a fluorescence T-jump experiment is also simulated . In agreement with the experimental results [18] , a relaxation time of 2 . 3±0 . 7 µs is found . This time is primarily determined by the relaxation towards the folded state of a compact molten globule-like structure , which acts as a kinetic trap . Relaxation times among all the other clusters , including transitions between fully unstructured states and the folded state , are all in the sub-microsecond time domain . Thus , surprisingly , the relaxation time measured by fluorescence may not be directly related to the ‘folding’ transition , if one calls ‘folding’ the transition from a random coil to the native state .
In the BE approach [38] a large set of CVs that are expected to be relevant for the process under investigation is chosen . A number NR ( number of replica ) of MD simulations ( walkers ) are run in parallel , biasing each walker with a metadynamics bias acting on just one or two collective variables . In BE the sampling is enhanced by attempting , at fixed time intervals of a few ps , swaps of the bias potentials between pairs of walkers . The swap is accepted with a probability ( 1 ) where and are the coordinates of walker a and b and is the metadynamics potential acting on the walker a ( b ) . In this manner , each trajectory evolves through the high dimensional free energy landscape in the space of the CVs sequentially biased by different low dimensional potentials acting on one or two CVs at each time . The results of the simulation are NR low dimensional projections of the free energy [38] . In BE the convergence of the bias potential to the corresponding free energy projection is monitored like in standard metadynamics: if the CVs are properly chosen and describe all the “slow” degrees of freedom , after a transient time , reaches a stationary state in which it grows evenly fluctuating around an average that estimates the free energy [55] . Convergence of metadynamics has been demonstrated analytically for a Langevin model [56] , and numerically for several realistic systems [55] , also in the presence of exchanges between different replicas [39] . Low dimensional free energy projections are often not very insightful , as in complicated processes like protein conformational transitions each minimum in a low dimensional profile may correspond to several different structures . In order to estimate the relative probability of the different structures one should find a manner to estimate the free energy in a higher dimensional space ( e . g NR ) . In this section a novel method to address this issue is described . The idea is to exploit the low-dimensional free energies obtained from BE to estimate , by a weighted-histogram procedure , the free energy of a finite number of structures that are representative of all the configurations explored by the system . These structures are determined by performing a cluster analysis , namely grouping all the frames of the BE trajectories in sets ( bins ) in which all the elements are close to each other in CV space . Since the scope of the overall procedure is constructing a model that describes also the kinetic properties of the system , it is important that the bins are defined in such a way that they satisfy three properties: A set of bins that satisfy these properties is here defined dividing the CV space in small hypercubes forming a regular grid . The size of the hypercube is defined by its side in each direction: where is the number of collective variables . This determines directly how far the bin centers are . Each frame of the BE trajectory is assigned to the hypercube to which it belongs and the set of frames contained in a hypercube defines a bin . This very simple approach is used here only in order to keep directly under control the distance between the bins , but the results presented in this Section apply also if the cluster analysis is performed with one of the other approaches that have been developed for this scope [43] , [44] , [57] . The canonical weight of each bin is estimated by a weighted histogram procedure based on the metadynamics bias potentials . The derivation that we report follows ref . [49] . Denote by the history-dependent potential generated by the walker up to time expressed in Boltzmann constant units . After a certain time ( 5 ns for Ala3 and 22 ns for Trp-cage ) , metadynamics has explored all the available CV space . At the end of the simulation , an estimate of the free energy is the average of after [55] , [58]: ( 2 ) where is the total simulation time . During the last part of the BE run fluctuates around ( except for an irrelevant additive constant that grows linearly with time ) , but these fluctuations are small if the deposition rate of the Gaussians is not excessive . In order to keep the error induced by these fluctuations under control it is convenient to consider two different bias potentials of the form of Eq . 2 , one obtained extending the integral from up to , the other from up to . Only the configurations collected after in which the two bias potentials are consistent within few ( for Ala3 and for the Trp-cage ) are retained for further analysis . The unbiased probability to observe bin is estimated on walker using the standard umbrella sampling reweighting formula: ( 3 ) where is a parameter that fixes the normalization and is the set of frames in the walker that are assigned to bin The are used to construct the best possible estimate of the probability of observing bin . This requires estimating the error on . Here it is assumed that the error on a bin free energy estimate is: ( 4 ) where is a constant that takes into account the correlation time and ( 5 ) In order to simplify the notation we have neglected the position-dependence of . For both Ala3 and Trp-cage we used an upper bound for ( = 1 and 10 , respectively , considering that the trajectory is saved every ps ) estimated from several unbiased MD simulations started from different configurations . In the last passage in Eq . ( 4 ) the fact that is an unbiased estimator of is assumed . The combined probability is now written as a linear combination of the , namely , where the weights are parameters that have to be determined and is normalization constant . The expected error on is . The optimal weights for each bin are determined separately minimizing this error with the constraint . This gives and , finally , ( 6 ) with . The constants are obtained iteratively from the condition ( 7 ) The free energy estimate given by Eq . 6 is affected by an error ( 8 ) consistently with what is found in the normal weighted histogram analysis method . Within this framework , the average value of an observable can be calculated , using the estimated free energies , as ( 9 ) where the sums run over all the bins , is the temperature and is the average value of in the bin . If the bin size is small enough , the bias potentials are approximately constant for the configurations belonging to the same bin [40] . Thus can be reliably estimated as the arithmetic average of in all the configurations explored by the BE trajectory belonging to the bin . Corrections deriving from the variation of the bias potentials inside a bin have also been considered but they lead to negligible effects for small . The enthalpy of bin is obtained averaging the enthalpy over the structures belonging to the bin . The entropy is estimated as . Neglecting the dependence of the entropy on the temperature , the free energy at a temperature different from is estimated as ( 10 ) with an error of . Using Eq . 9 together with Eq . 10 allows extrapolating the average value of the observables for a few tens of K around the temperature at which the simulation is performed . The uncertainty on can be derived at each temperature from the error on , , and using error propagation on Eqs . 9 and 10: ( 11 ) where is the standard deviation of inside bin . In this section we describe a manner for constructing an approximate kinetic model describing transitions between the bins introduced in the previous Section . Constructing the model requires estimating the rates for a transition between every pair of neighboring bins and . As BE trajectories are biased , the transition probabilities observed in the BE run cannot be taken as a direct measure of the true transition rates . The kinetic model is constructed assuming that the transitions between bins are described by rates of the form introduced in Ref . [47] , [50] , namely by diffusion with a bias determined by their free energy difference: ( 12 ) where are the rates associated to simple diffusion on a flat free energy surface . This form of the transition rates ensures that the limiting probability distribution of the dynamics is correct , namely that the probability to observe bin at long times scales is proportional to . If the bins form a hypercubic grid in CV space the rates can be exactly expressed as a function of the ( possibly position-dependent ) diffusion matrix and of the hypercube side [47] . In the following to simplify the notation we denote by the diffusion matrix appearing in the transition rate between two bins and assuming that is the average of and [47] . In one dimension the bins are labelled by a single integer ( ) and , following Refs [47] , [50] , and zero otherwise . In dimensions the bins are labelled by integers . If is diagonal , the one-dimensional expression for the rates can be generalized straightforwardly . If is non-diagonal the only rates different from zero are those in which one or two of the components of vary by one: ( 13 ) This form of the rates can be derived discretizing the Fokker-Planck equation for diffusion on the regular grid defined by the hypercube centers . The derivatives are discretized as centred differences , in such a way that if is a positive-definite matrix all the resulting rates are positive , as is required in a kinetic model . The error of this procedure scales as the square of the distance between neighbouring bins [47] . At finite grid spacing the accuracy can be improved allowing transitions between non-neighbouring bins . It can be verified that if the system is evolved with the rate equation 12 using , then the Einstein relation is satisfied , namely ( 14 ) The rates given by Eq . 12 are used in a KMC algorithm [51] , [52] to generate a dynamics between bins . If the bins size is small enough the KMC kinetics resembles the kinetics of an overdamped Langevin dynamics [47] . If the free energy is flat , by construction the model gives the correct diffusive behaviour but if deviations from this behavior are observed when the bin size is too large . On the other hand , a small bin size can hinder the accuracy of the free energies . Thus , both large and small bin size may alter the quality of the kinetic model due to bad description of the underlying free energy surface or inaccurate sampling . Moreover even if there are no problems related to the bin size , describing the dynamics with Eq . 12 amounts to neglecting memory effects . This approximation can be particularly severe if an important variable is not included explicitly in the model . The model is expected to be reasonably accurate if the memory time is much smaller than the typical transition time ( usually between metastable sets ) that one wants to measure . The diffusion matrix entering in Eq . 13 is estimated using the approach of Ref . [47] , in which one maximizes the likelihood that a given MD trajectory is generated by a rate equation of the form Eq . 12 . Computing requires first generating at least one MD trajectory without the metadynamics bias . The accuracy of the procedure can be improved , if the relevant metastable states of the system are known , by running several independent MDs starting from these states . Otherwise one can select at random a few conformations along the BE trajectory and use these as the initial conditions for MD . The trajectory ( or the set of trajectories ) is then mapped at a time lag onto the bins . Then several KMC trajectories are run with an initial guess for , starting from the bins visited by the MD trajectory . Using the KMC trajectories one computes the conditional transition probabilities at a time lag among all the pairs of bins , visited by the trajectory . This is evaluated by counting transitions between the bins: where is the number of times the KMC trajectory is found in bin at time being in bin at time zero , and is the number of times the trajectory visits bin . This procedure is slightly different from the one used in Ref . [47] , where is calculated by diagonalizing the rate matrix , which in the cases considered here has a very large size ( of the order of 105×105 ) . The notation indicates that these probabilities depend parametrically on . Using these probabilities one evaluates the logarithm of the likelihood to observe the sequence of bins obtained by MD . This is given by ( 15 ) is then maximized as a function of . This can be done by simulated annealing , starting from an initial guess of and iterating until the likelihood reaches a plateau . As outlined in Ref [48] , the diffusion matrix found in this way depends in general by the chosen time lag . A common behavior is that by increasing the time lag the elements of the diffusion matrix converge to a well defined value . This means that after this the dynamics between bins is close to Markovian and is well approximated by the model proposed . As a consequence only transition that occur on a time scale bigger than are correctly described by this model . Applying this procedure the prefactor of the rate Eq . 12 , which has the form of a jump process among a discrete set of states , is directly optimized . This is a clear advantage with respect to other methods for computing , in which a continuous evolution of the collective variables is assumed . Moreover , as the free energies are known , the only variational parameter is and comparably short trajectories are sufficient to determine it with a good statistical accuracy . The approach described in the previous two sections has been carefully benchmarked on solvated Ala3 . For this system , it was possible to compare the predictions of the kinetic model , with the results of a very long ( ∼2 µs ) molecular dynamics trajectory . All the BE and MD simulations were performed using the GROMACS suite of programs [59] , [60] and the AMBER03 [61] force field . Ala3 was placed in a periodic cubic box containing 1052 TIP3P water [62] molecules . The time step was set to 2 fs and the LINCS [63] algorithm was used to fix the bond lengths of Ala3 . The SETTLE algorithm [64] was used to fix angle and bond length of water molecules . Electrostatic and Lennard-Jones interactions were calculated with a cutoff of 1 . 0 nm . Lennard-Jones interactions are switched off smoothly from 0 . 9 nm to 1 . 0 nm . The neighboring list was updated every 5 steps and the cut-off distance for the short-range neighbor list was set to 1 . 1 nm . The Particle Mesh Ewald method [65] , [66] was used to treat long-range electrostatic interactions with a maximum grid spacing for the fast fourier transform of 0 . 12 nm and an interpolation order of 4 . A constant temperature of 300 K was achieved by coupling the system to a Berendsen thermostat [67] with a characteristic time of 0 . 1 ps . A constant pressure of 1 bar was achieved by coupling the system to a Berendsen barostat [67] with a characteristic time of 2 . 5 ps . Several independent MD simulations were performed , with a length varying between ∼30 ns and ∼30 ns , for a cumulative time of 1 . 8 µs . The conformations of Ala3 are specified by its six backbone dihedral angles ( , where ) ( see Fig . S1 , inset ) . Following Refs . [68]–[70] , and ( central Ramachandran angles of Ala3 ) were considered in order to assign the main conformations of the system , denoted by ( , ) , ( , ) , ( , ) , and ( , ) . Besides the latter conformational states , eight different states were also considered in order to analyze the results of the kinetic model . These are the free energy minima with the three dihedrals in the or region of the Ramachandran plane , namely , , etc . ( see Fig . S1 ) . The system was also simulated using bias exchange metadynamics ( BE ) [38] exploiting the six dihedral angles ( see Fig . S1 , inset ) as CVs . Each CV was biased in a different walker . Hence , NR = 6 , and each walker evolved under the action of a one-dimensional metadynamics potential acting on one of the six CVs . The width and the height of the Gaussians used in metadynamics were 0 . 1 rad and 0 . 1 kJ/mol respectively . A new Gaussian was added to the metadynamics potential every 1 ps . Exchanges of the bias potentials between pairs of walkers are attempted every 10 ps . Three independent BE simulations of 30 ns each ( one simulation consist of 30 ns for each replica ) were carried out in order to check the reproducibility of the results . The computational setup used in Ref . [38] is briefly summarized here . The simulations were performed with the GROMACS suite of programs [59] , [60] and the AMBER03 force field [61] , at a temperature of 298 K . The initial structure ( pdb entry 1L2Y ) [17] was solvated with 2075 TIP3P [62] water molecules in a 40×40×40 Å water box . The system was simulated using BE [38] . Five collective variables ( CVs ) were biased according to the bias exchange scheme [38] . CV1: number of contacts; CV2: number of contacts; CV3: number of backbone h-bonds . CV1 , CV2 , and CV3 are defined as where the sum runs over the appropriate set of atoms ( all the for CV1 , all the for CV2 and all the backbone H and O for CV3 ) and , 6 . 5 and 2 Å for CV1 , CV2 , and CV3 respectively . CV4: fraction of dihedrals belonging to the region in the Ramachandran plot , defined as . CV5: correlation between successive dihedrals , defined as . The sums in CV4 and CV5 run over all the residues . All the variables are dimensionless and none of them requires the a priori knowledge of the folded state . The Gaussian widths chosen for CV1 , CV2 , CV3 , CV4 , CV5 were , , , , and , respectively . Simulations were performed with 8 walkers: one for each variable plus two walkers reconstructing a free energy surface in two dimensions: CV3-CV4 and CV4-CV5 . The last walker , the “neutral walker” , is not biased by any metadynamics potential , but is allowed to exchange conformations with the others . A Gaussian of height 0 . 1 kJ/mol was added every 1 ps to the bias potential for all the walkers except the neutral walker . The total length of the simulations was 50 ns . In Ref . [38] it was shown that the neutral walker statistics is approximately canonical , and all the averages were there computed using only its configurations , while the trajectories of the biased walkers were not used at all . The converged free energy profiles for each walker can be found in Ref . [38] . The MD simulations used for calculating the diffusion matrix and the NMR properties were run with the same computational setup of BE simulation ( except for specified changes in temperature ) .
Ala3 is a simple polypeptide that has been extensively used as a benchmark system . Although small , this system shows several protein-like features , such as intramolecular hydrogen bonds and a fragment of structure . Since the system is small , it is possible to characterize carefully its equilibrium and kinetic properties by extended MD simulations . In this section the results obtained by applying the approach presented in the Methods section to the Ala3 system will be exposed . The results presented here were obtained analyzing , with the method introduced in the Methods section the BE trajectory of Trp-cage from Ref . [38] . The rate model described in the Methods section has the form of a generalized rate equation with the rates given by Eq . 12 . The presence of metastable sets ( “clusters” ) was detected applying the MCL [53] , [54] method to the Trp-cage kinetic model . The algorithm requires choosing a parameter that tunes the granularity of the description: for only one cluster is detected , while for large all the bins are assigned to different clusters . Several choices of the parameter are attempted ( in Ref . [53] , [54] the value is considered ) . At 298 K , for only two relevant clusters are found , one with an occupancy of ≈90% and one of ≈5% . The RMSD among the structures belonging to the big cluster is very large , indicating that , for this system , is not appropriate . For the large cluster splits in two clusters with populations of ≈12% and ≈77% . Still the larger cluster includes qualitatively different structures . At the larger cluster splits further in three , while the other clusters remain approximately unchanged . Increasing further up to 1 . 17 does not modify significantly the three most populated clusters , whereas for the system is fragmented in more than 10 clusters . At , only 5 significantly populated ( >1% ) clusters are found , the two larger ones having a population of ≈58% and ≈25% respectively ( Table 2 ) . The average RMSD between the clusters structures and the NMR ensemble is ≈1 . 8 Å for cluster 1 and >4 . 4 Å for cluster 2 and the other clusters . Moreover , all the bins with RMSD <2 Å belong to cluster 1 . This allows concluding that MCL analysis using is able to identify a folded cluster with structural properties similar to the NMR ensemble . Its occupancy is of 58% at 298 K . Remarkably , at this temperature it exists another cluster with non-negligible population ( 25% ) that contains structures that are different from the structural ensemble generated from the NMR data ( RMSD = 4 . 4 Å ) . In the next section the consequences of the existence of this second cluster in the thermal ensemble at 300 K are discussed . It is worth to note that in the MD simulations used for the calculation of , if the trajectory starts from a structure belonging to a cluster , it remains there for most of the simulation ( few tens of ns ) . This means that MD simulations are consistent with the description of metastable states given by the MCL algorithm . In Fig . 4A , the most populated clusters obtained for are shown using a projection on three variables , the contacts , the fraction , and the correlations between consecutive dihedrals . Each color corresponds to a different cluster , and the lowest free energy bin ( attractor ) of each cluster is depicted as a sphere of the same color . The properties of the clusters depicted in Fig . 4A are summarized in Table 2 . In Fig . 5 , the hydrophobic contacts and the hydrogen bonds with the Trp6 are shown schematically for each attractor . Selected proton distances are also displayed for the three most populated clusters . A good agreement with the NMR unfolded state distances reported in Ref . [21] is found . Cluster 1 , as already anticipated , resembles very closely the NMR structure . More details will be provided in the following section ( the atomic cartesian coordinates for the reference structure of cluster 1 are reported in Dataset S1 ) . Cluster 2 has a RMSD of ∼4 . 4 Å with respect to the NMR structure , but it retains at least part of the native . The Trp SASA in this cluster is 70 . 5±1 Å2 , which compares with the value of 47 . 1±0 . 6 Å2 observed in the folded cluster . This indicates that Trp is shielded from the solvent also in cluster 2 . Arg16 forms a with Tyr3 ( see Fig . 4A ) while Trp6 is in contact with Pro12 , Pro18 , Gly11 and the aliphatic chain of Arg16 ( see Fig . 5 ) . As outlined in Fig . 5 , except for the Arg16 distance , the cluster 2 attractor ( reference structure ) shows Pro12 and Arg16 distances shorter than those in the folded cluster . The nearest hyperpolarized [21] Trp6 proton can be different in each cluster ( e . g . in cluster 1 the Arg16 distance is shorter than Arg16 ) . These distances are in very good agreement with those found in the NMR experiments [21] for the unfolded state . This cluster resembles the intermediate observed in a 100 ns implicit solvent simulation ( Ref . [24] , the atomic cartesian coordinates for the reference structure of cluster 2 are reported in Dataset S2 ) . Cluster 3 ( orange ) still contains a short . The contacts are reduced with respect to the folded cluster and the Trp is partially solvent exposed . The reference structure of cluster 3 is similar to the state I of Ref . [36] and to the intermediate structure found in Ref . [31] , with the difference that the Asp9-Arg16 salt bridge in cluster 3 is formed only in a fraction of the bins belonging to the cluster . This may indicate that the salt bridge is rather unstable . The Leu7 distance in the cluster 3 attractor is shorter than that in the folded state . Also in this case the distance compare well with the NMR experiments value [21] . This imply that the presence of cluster 2 and cluster 3 ( the two most populated misfolded clusters ) is consistent with the unfolded state ensemble information reported in Ref . [21] ( the atomic cartesian coordinates for the reference structure of cluster 3 are reported in Dataset S3 ) . The other clusters show only a small residual secondary content and can be generically referred to as “unfolded states” . The attractor of cluster 4 is stabilized by the formation of the Asp9-Arg16 salt bridge ( the atomic cartesian coordinates for the reference structure of cluster 4 are reported in Dataset S4 ) . The bins belonging to cluster 5 are mostly compact molten globule structures characterized by the presence of several hydrophobic and contacts ( even more than in the native state ) but small secondary content ( see Fig . 4A and Fig . S8 ) . In the most stable bin of this cluster Trp6 is in contact with Pro17 and Pro18 residues ( see Fig . 5 , the atomic cartesian coordinates for the reference structure of cluster 5 are reported in Dataset S5 ) . In Fig . 4B the occupancies of cluster 1 , 2 , and 5 are plotted as a function of temperature . As expected the folded cluster ( cluster 1 ) increases its occupancy as the temperature decreases . Its population is 50% at 310 K , a temperature that is consistent with the experimental melting point of 317 K [19] , [20] . The error on the occupancies becomes large at , indicating that the temperature extrapolation based on Eq . 10 is unreliable after this temperature . The occupancy of cluster 5 is almost negligible at 300 K ( 2 . 8% ) , but it grows significantly with temperature ( see Fig . 4B ) . The importance of this will become clear when the kinetic properties of the system will be discussed . The helical content decreases only slowly with temperature , consistently with REMD results in explicit solvent [34] . On the average , only ∼1 residue melts between 290 and 320 K . In order to characterize in more detail the nature of the clusters described in the previous section , it is useful to consider their NMR properties . As only cluster 1 and 2 are compact and show a significant content of secondary structure , the investigation is here restricted to these two clusters . In Fig . 6A the protons CSDs of cluster 1 are compared with the experimental results ( full circles ) . The shifts are estimated as described in the Method section . The correlation between theoretical and experimental NMR CSDs is rather good ( ) , while cluster 2 shows a much smaller correlation with experiments , especially for protons that have negative CSDs . The correlation with NMR data is even smaller for all the other clusters . This confirms that the cluster classification deriving from Markov cluster analysis accurately discriminates between the folded state ( cluster 1 ) , an unfolded state with several native-like features ( cluster 2 ) , and all the rest . The correlation with experiments is retained using in the average the full ensemble of bin ( ) . Even if correlation is good , it has to be noted that the proportionality factor between theoretical and experimental CSDs is 0 . 46 in the full ensemble of bins and 0 . 6 in cluster 1 . To investigate the origin of the variations in the proportionality factor two 20 ns equilibrium MD simulations have been performed , at 282 K ( experimental temperature ) and at 300 K , starting from the NMR structure and with the same computational setup used in the BE simulation . At both temperatures the proportionality factor with experimental CSDs is 0 . 8 instead of 1 , therefore 0 . 8 has to be considered the reference value for our computational setup . The optimal proportionality factor of 0 . 8 is obtained if the CSDs are computed on the lowest free energy bin of cluster 1 . The slope difference between 0 . 6 ( cluster 1 ) and 0 . 8 may be ascribed to small inconsistencies between the ensemble of structures generated with BE and by an unbiased MD starting from the NMR structure . The further slope variation when the calculation is extended to the full ensemble of bins is most likely a consequence of calculating NMR properties at 298 K instead of at the experimental temperature of 282 K where the population of cluster 1 is larger . Using a similar procedure ( see Methods ) RCS and its temperature derivative were also computed . It is worth to note that most of the large CSD are due to the Trp RCS [17] . The protons whose RCS is large are also those whose RCS depends more strongly on , in excellent agreement with the experimental data [17] . The protons RCS temperature derivatives as a function of the RCS are plotted in Fig . 6B . The results are plotted as a function of the RCS estimated at 298 K . The comparison is performed at 298 K and not at the experimental temperature of 282 K in order to avoid error propagation that is unavoidable if Eq . 10 is used for extrapolating the results for a large temperature difference . Despite of this , the two observables correlate linearly ( for the ) , consistently with experiments [17] . Side chain protons in the C-terminal part of the protein fall on the same correlation line , also in agreement with the experiments [17] . A few protons deviate significantly from this linear behavior . The most significant deviation are observed for , , and , the last two being also reported experimentally [17] . The RCS of and is large , while their RCS derivative is almost zero . The cluster decomposition proposed here can be used to elucidate the presence of these outliers . In fact , the RCS of is −0 . 53±0 . 01 p . p . m . and −0 . 97±0 . 02 p . p . m in cluster 1 and 2 respectively , while other protons ( except and ) have RCS which are less negative in cluster 2 than in cluster 1 or similar in the two clusters . The RCS of has a similar value in both clusters . This significant difference derives from the fact that and in cluster 2 are much closer to Trp than in cluster 1 . Since , increasing the temperature , the relative population of cluster 2 and 1 changes ( see Fig . 4B ) , the RCS of , and changes with temperature less than the RCS of other protons . In view of these results , the anomalous behavior of and observed experimentally can be considered a signature of the presence of cluster 2 in the thermal ensemble of Trp-cage . The fluorescence relaxation after a temperature jump ( T-jump ) was estimated according to the procedure outlined in the Methods section . This observable is used in Ref . [18] to infer information on the Trp cage folding kinetics . The fluorescence properties of the system are here estimated by computing the Trp SASA , which is known to correlate with fluorescence [76] . The result shows a smooth decay to an asymptotic value on the time scale of the microseconds . A double exponential decay model describes very accurately the data ( , see Fig . S4 ) . The two time constants are , and . The large gap between the first and the second time constant is a strong indication of two-state behavior . The value of is in agreement with the experimental relaxation time of 3 . 1 µs for the florescence T-jump [18] . This shows that the rate model is capable of reproducing accurately the dynamics of the real system , at least for what concerns the relaxation of fluorescence . The microscopic rearrangements that determine will be discussed in detail in the next section . The influence that the error on the free energies and on the enthalpies has on the results is ∼500 ns ( see Methods ) . The error deriving from neglecting the position dependence of is ∼500 ns ( see section Application to the Trp cage folding and Text S1 ) . Thus the overall error on the relaxation time is . Including the correction suggested in Ref . [77] to take into account the unphysical viscosity of TIP3P water [78] the relaxation time is , still in fair agreement with experiments . Here the dynamics of the system is investigated in more details , still using the rate model introduced in the Methods section . The characteristic times of the system are related to the eigenvalues of the rate constant matrix . Consistently with what is found for the Trp SASA relaxation , the second largest eigenvalue corresponds to a characteristic time of 2447 ns . The third eigenvalue corresponds to 434 ns , with a gap of 2013 ns from the first , consistently with a two state behavior [18] . The second eigenvector has large positive components in cluster 1 and 2 and large negative components in cluster 5 . This suggests that the longest relaxation time of the system is associated to a transition between these states . In order to analyze more quantitatively this issue , the rates for the transitions between the clusters found by Markov cluster analysis were extracted from a very long KMC simulation ( ) . For two clusters A and B with occupancy and , the rate constant to go from A to B was calculated counting the number of times that a trajectory goes from A to B without passing from any other cluster during the KMC simulation . The rate to go from A to B was estimated as . To minimize the number of recrossing , the KMC trajectory is assumed to visit a cluster any time it visits any bin belonging to the group of lowest free energy bins containing 70% of the cluster population . Bins that do not fall in this definition were considered as transition states . The transition rates obtained in this manner are represented in Fig . 7 . For clarity , all the clusters whose occupancy is below 1% are omitted from the figure . The equilibration between cluster 1 and 2 is rather fast and transition times to cluster 3 are also in the sub-microsecond domain , but when the system reaches cluster 5 on average ∼2 µs are necessary to return to the folded cluster . The folding pathways schematized in figure are consistent with the two routes proposed by Ref . [36] , except for the transitions involving cluster 5 . The folding pathway initiating from cluster 4 and passing from cluster 3 is characterized by the early formation of an and resembles the pathway passing from state I in Ref . [36] . The pathway passing from cluster 2 is instead characterized by the formation of several hydrophobic contacts , while the content remains on average lower . This resembles the pathway passing from state L in Ref . [36] . If the molten-globule state ( cluster 5 ) is neglected the folding and unfolding rates are compatible with those reported in Ref . [37] , considering the difference in the force field .
The approach presented here exploits the trajectories of multiple metadynamics simulations for building a thermodynamic and kinetic model of complex processes ( e . g . protein folding ) whose description requires a large number of collective variables . The aim of the model is to reproduce the long time scale dynamics of the system and to extract the metastable sets ( clusters ) of the kinetic process . These states may correspond , for example , to misfolded conformations . The model is constructed as follows: in a first step the equilibrium probabilities of a finite set of conformational states , or bins , are determined by a weighted-histogram procedure exploiting the low-dimensional free energies estimated by metadynamics . In a second step an approximated description of the kinetics is obtained estimating the transition rates among the bins . The diffusion matrix entering in the model is estimated by a maximum-likelihood procedure [47] employing relatively short unbiased MD trajectories . The approach was tested on the Ace-Ala3-Nme peptide in explicit solvent using the six backbone dihedral angles as CVs . For this system equilibrium MD trajectories on the microsecond timescale are sufficient to sample the relevant conformational space and were used as a reference to evaluate the accuracy of the kinetic model obtained from the BE results . The bins free energies obtained with the method presented here are in excellent agreement with free energies computed from equilibrium MD . The transition rates among neighboring bins are used to run a long KMC . The mean first passage times among selected states obtained in this way are in agreement with those extracted from the reference MD simulations . Trp-cage is a designed miniprotein that , due to its small size and fast folding rate , has been the object of several theoretical investigations . Here this system is analyzed with a new method , introduced in this paper , that allows deriving a kinetic model of the system by analyzing a set of biased MD trajectories . The model shows the presence of several metastable states ( clusters ) . The most populated one can be classified as the folded state . The second most populated cluster has a RMSD of ∼4 . 4 Å from the NMR structure and retains part of its secondary structure ( see Fig . 4A ) . In this cluster the Trp is more strongly packed between Gly11 and Pro12 than in the NMR structure and its population relative to cluster 1 increases with temperature ( see Fig . 4B ) . This can explain the anomalous behavior of the temperature dependence of the CSD of hydrogen atom observed both experimentally [17] and in the simulated NMR experiment ( see Fig . 6B ) . The cluster 2 and cluster 3 reference structures are consistent with experimental unfolded state distances [21] ( see Fig . 5 ) . The presence of these two clusters is also in agreement with the strengthening of proline ( s ) -Trp excitonic interactions with temperature and the broad melting observed in Ref . [20] . In spite of the presence of several intermediates both the simulated T-jump experiment ( see Fig . S4 ) and the spectrum of the kinetic matrix associated with the rate model are consistent with a two state kinetics [18] . The calculated time constant of the folding process is ∼2 . 3±0 . 7 µs ( or ∼3 . 8±1 . 2 µs including the correction of Ref . [77] ) in fair agreement with the experimental relaxation time [18] . To investigate the folding dynamics using the kinetic model we derived a folding mechanism which involves the detected intermediates ( see Fig . 7 ) . Starting from open structures , the folding process can follow two main routes . One of them consists in an earlier formation of the N-terminal ( cluster 3 ) followed by the hydrophobic collapse , while the other involves first the formation of hydrophobic contacts with less helical content ( cluster 2 ) and then the completion of both secondary and tertiary structure . This is in agreement with the pathways found in Ref . [36] . The time required to undergo these transitions is in the sub-microsecond time domain , which is less than the slowest relaxation time found in the simulated T-jump experiment and more consistent with the third eigenvalue of the kinetic matrix . Indeed , the folding mechanism ( see Fig . 7 ) shows that , if Trp-cage reaches the molten globule state , more than 2 µs are necessary to reach the folded state . This implies that the experimental folding time is ultimately determined by the slow equilibration between the first two clusters and the compact molten globule state that acts as a kinetic trap . In this state no secondary structure element is present , but a hydrophobic core with several tertiary contacts is formed . In Ref . [79] the Pro12Trp mutation brings to an increased stability of the folded state and a faster folding time of ∼1 µs . This seems to be in agreement with the folding mechanism presented here , since the mutation would strongly stabilize cluster 1 and cluster 2 but not the molten globule cluster . A possible way to assess experimentally the presence of the molten globule could be a mutation of Pro17 to a more polar residue ( e . g . Asn ) or a chemical modification of this residue as the lower rigidity associated to the absence of the Pro17 ring could destabilize the folded state [80] . In fact in the attractor of cluster 5 Pro17 shows a strong interaction with Trp6 , and this interaction does not play a key role in other relevant clusters ( see Fig . 5 ) . In conclusion , we have presented an approach aimed at constructing a rate model for complex biomolecular processes starting from a set of biased MD trajectories . One could argue that other approaches aimed at the same purpose are based on less severe assumptions . Distributed simulation techniques allow computing the folding rates directly , and have been applied successfully for studying folding in explicit solvent of even larger systems [32] , [43] , [45] . Normal replica exchange [29] , [34] , when converging , provides a direct measure of the equilibrium distribution , and does not require a complicated reweighting procedure . Finally , if one would use an implicit solvent description of the system , one could observe several folding/unfolding events by simple finite-temperature molecular dynamics , and it would not be necessary to use an enhanced sampling technique . In this framework , a rate model for the system could be constructed in a more rigorous manner [43] , [44] , [46] . Still , despite of the approximations that are done , the approach presented here provides a picture of the dynamics and thermodynamics of the system that is detailed and in agreement with all the experimental evidences presented so far . We believe that this result ultimately derives from the combined use of an accurate ( but expensive ) force field , and of a method that , at the price of generating non-equilibrium trajectories , allows an efficient exploration of configuration space and the accurate calculation of free energies . | Understanding the mechanism by which proteins find their folded state is a holy grail of computational biology . Accurate all-atom simulations have the potential to describe such a process in great detail , but , unfortunately , folding of most proteins takes place on a time scale that is still not accessible to routine computer simulations . We introduce here an approach that allows for constructing an accurate kinetic and thermodynamic model of folding ( or other complex biological processes ) using trajectories in which the process under investigation is forced to happen in a short simulation time by an appropriate external bias . An important strength of this approach is the possibility of identifying and characterizing misfolded conformations that , in some proteins , are related to important diseases . We use this method to study the folding of Trp-cage , predicting the structure of the folded state and the presence of several intermediates . We find that , surprisingly , fully unstructured “unfolded” states relax towards the folded conformation rather quickly . The slowest relaxation time of the system is instead related to the equilibration between the folded state and another compact structure that acts as a kinetic trap . Thus , the experimental folding time would be determined primarily by this process . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"computational",
"biology/molecular",
"dynamics",
"biophysics/theory",
"and",
"simulation",
"biophysics/protein",
"folding"
] | 2009 | A Kinetic Model of Trp-Cage Folding from Multiple Biased Molecular Dynamics Simulations |
There is increasing interest in the microbiome of the hepatobiliary system . This study investigated the influence of infection with the fish-borne liver fluke , Opisthorchis felineus on the biliary microbiome of residents of the Tomsk region of western Siberia . Samples of bile were provided by 56 study participants , half of who were infected with O . felineus , and all of who were diagnosed with gallstone disease . The microbiota of the bile was investigated using high throughput , Illumina-based sequencing targeting the prokaryotic 16S rRNA gene . About 2 , 797 , discrete phylotypes of prokaryotes were detected . At the level of phylum , bile from participants with opisthorchiasis showed greater numbers of Synergistetes , Spirochaetes , Planctomycetes , TM7 and Verrucomicrobia . Numbers of > 20 phylotypes differed in bile of the O . felineus-infected compared to non-infected participants , including presence of species of the genera Mycoplana , Cellulosimicrobium , Microlunatus and Phycicoccus , and the Archaeans genus , Halogeometricum , and increased numbers of Selenomonas , Bacteroides , Rothia , Leptotrichia , Lactobacillus , Treponema and Klebsiella . Overall , infection with the liver fluke O . felineus modified the biliary microbiome , increasing abundance of bacterial and archaeal phylotypes .
There is increasing interest in the microbiota with respect to diseases of the gastroenterological system [1 , 2] including the liver and biliary tree [3 , 4] . Many reports have detailed the colorectal/ fecal microbiota , given that samples of feces are readily accessible using non-invasive approaches . Modifications of the gut microbiota have been documented for a number of liver diseases including primary biliary cirrhosis , primary sclerosing cholangitis , cholelithiasis , [4–6] . Moreover , information is becoming available on the microbial composition of the bile during liver disease [6–8] . Conversely , it had long been considered that during good health the bile was sterile or at least that bile was inimical to bacteria [9] , with a few reports indicating colonization of the gallbladder and bile as the consequence of reflux of the duodenal contents , blood-borne infection and infection spread through the portal-venous channels [10] . High throughput sequencing of bacterial 16S rDNA genes has provided information on the present of complex microbiota in the bile environment even in absence of biliary tract morbidity . Studies of pigs show that bacteria from the phyla Proteobacteria , Firmicutes and Bacteroidetes populate the gall bladder ecosystem [9] . The investigation of the feces , bile and gallstones from patients diagnosed with cholelithiasis ( gallstones ) revealed higher bacterial diversity in the biliary system in the comparison with feces; the biliary tract microbiome of gallstone patients includes >100 bacterial OTUs belonging to six bacterial phyla [6] . On the other hand , dysbiosis of biliary microbiome may play key role in the biliary inflammation , supporting the concept that factors that affect bile duct composition can be associated with liver diseases [4] . In this context , findings in hamsters infected with Opisthorchis viverrini demonstrated that infection with this liver fluke not only modifies the intestinal microbiota but revealed the presence of >60 phylotypes of nine phyla in the biliary system associated with the parasites [11] . Moreover , infection of hamsters with O . viverrini positively correlated with increased co-infection with Helicobacter pylori and H . bilis , both in the fecal microbiota and in the biliary tract within the gut of the liver flukes [12] . Liver flukes excrete and secrete mediators [13] , altering liver functions that may modify the biliary environment [14] and which , in turn , may modify the composition of the microbiota [12 , 15] . Indeed , interactions between liver flukes and the microbiome can be expected to be dynamic and to modify the metabolic responses specific to opisthorchiasis , as known during infection with other helminths [16 , 17] . Molecular markers of inflection with the blood fluke Schistosoma mansoni infection were found in urine to be primarily linked to changes in gut microflora , energy metabolism and liver function [18] , and infection with Schistosoma haematobium leads to changes in bacterial pathobionts in the urinary bladder [19] . Other metabolites known to arise from the activities of helminths including catechol estrogens , oxysterols and their adducts involving host cell DNA and other macromolecules likely also influence the ecology of the microbiome [20 , 21] . Also , helminth parasites can harbor endosymbiotic microbes , in particular the rickettsia-like bacteria of trematodes [22] and symbiotic Wolbachia of filariae [23] . This study investigated the influence of infection with the fish-borne liver fluke Opisthorchis felineus on the biliary microbiome , within a background clinical setting of cholelithiasis .
The Ethics Committee of the Siberian State Medical University approved this study . All participants provided written informed consent . Participants ranged in age from 40 to 61 years . Prospective participants who had used antibiotics or probiotics within the previous six months were excluded from the study . Fifty-six participants who had been diagnosed with cholelithiasis ( gallstone disease ) but who were in disease remission provided samples of bile . Gallstone disease had been diagnosed by B-mode ultrasonography . Thirty of these 56 participants were concomitantly diagnosed with infection with the fish-borne liver fluke , Opisthorchis felineus , whereas the remainder ( 26 persons ) was not infected with O . felineus ( below ) . The bile samples were obtained from the study participants during therapeutic intervention for cholelithiasis involving open or laparoscopic cholecystectomy at the 3-d City Tomsk Hospital , Tomsk , western Siberia . During cholecystectomy , 5–10 ml of bile was aspirated from the gallbladder under sterile conditions , three to five ml dispensed into in a sterile tube , and thereafter dispatched immediately to the laboratory . Two ml bile was clarified by centrifugation ( 10 , 000 g , 10 min ) , the supernatant removed , and the pellet was stored at -80°C until processing . Other aliquots of these biles , ~3 ml were subjected to centrifugation at 5 , 000 g , 10 min , after which the pellet was examined for eggs of O . felineus . The pellet was diluted into Buffer ASL QIAamp Stool Mini Kit ( QIAGEN , Hilden , Germany ) , 25 mg glass beads ( 0 . 1 mm diameter ) added to the suspension , the mixture vortexed for 10 seconds , and then subjected to bead-beating ( Mini-Beadbeater-24 , Bio Spec Products Inc ) for three minutes . A second bead beating was performed after incubating the suspension at 70°C , after which phenol-chloroform extraction was undertaken to recover genomic DNAs . Subsequently , the DNA was dissolved in 20 μl TE , and DNA yield was measured using a NanoDrop ND-1000 UV spectrophotometer ( Nano-Drop Technologies , Wilmington , DE ) . DNA was aliquoted to perform the PCR to confirm or not infection with O . felineus ( exclusion ) and for the 16S rRNA sequence-based survey of biliary prokaryotes . Control DNA extractions in which 100 μl sterile water replaced biliary DNA were undertaken , in order to address laboratory and sequence-based artifacts that can occur with reagents and kits [24] . Status of infection liver fluke infection was established by the microscopic examination for eggs of O . felineus in the material pelleted from several ml of bile and by PCR to identify the presence of DNA of O . felineus in the pellet . To confirm the infection , we employed a PCR-real time commercial kit for identification of O . felineus ( Medico-biological Union , Novosibirsk , Russia ) [25] following the manufacturer’s guidelines . PCR using bile pellet DNA , above , was performed in a thermal cycler ( LightCycler 480 , Roche ) . The DNA samples were used for a 16S rRNA sequence-based survey of bacterial diversity . Amplicons that cover V3 and V4 hypervariable region of 16S rRNA genes ( Escherichia coli positions 341–805 ) were generated by PCR with using Primers Next-16S-1st-F 5’- TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG CCT ACG GGN GGC WGC AG -3’ and Next-16S-1st-R 5’- GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA GGA CTA CHV GGG TAT CTA ATC C-3’ . These primers contain gene-specific sequence ( bold-face font ) and Illumina adapter sequences . The initial PCR cycles were carried out in MJ Mini thermal cycler ( MJ Research ) . The PCR reactions were performed in the following program: initiation enzyme activation at 95°C for 3 min , followed by 25 cycles consisting of denaturation at 95°C for 30 sec , annealing at 55°C for 30 sec and extension at 72°C for 30 sec . After 25 cycles , the reaction was completed with a final extension of 5 min at 72°C . PCR products were recovered by chromatography on Ampure XP beads ( Thermo Fisher Scientific ) and deployed in a second PCR . The Illumina Nextera XT Index kit ( Illumina Inc . , San Diego . CA , USA ) were used for multiplexing . Two unique indices located on either end of the amplicon were chosen based on the Nextera dual-indexing strategy . To incorporate the indices to the 16S amplicons , PCR reactions were performed on MJ Mini thermal cycler ( MJ Research ) . Cycling conditions consisted of one cycle of 95°C for 3 min , followed by eight cycles of 95°C for 30 sec , 55°C for 30 sec and 72°C for 30 sec , followed by a final extension cycle of 72°C for 5 min . After purification of PCR-products on Ampure beads ( Thermo Fisher Scientific ) , the concentrations were measured using Qubit technology ( Thermo Fisher Scientific ) . The libraries were sequenced by 2 × 300 bp paired-end sequencing on the MiSeq platform using MiSeq v3 Reagent Kit ( Illumina ) at the Faculty of Bioengineering and Bioinformatics , Lomonosov Moscow State University . Analysis of the 16S rRNA gene reads was performed using the QIIME ( quantitative insight into molecular ecology ) pipeline , version 1 . 9 . 0 [26] . The Operational Taxonomic Units ( OTUs ) picking strategy consisted in usage of the open QIIME reference OTU picking algorithm with the OTU-picking method UCLUST [27] . Chimera-checked GreenGenes taxonomy v13 . 5 was used as the reference base for taxonomic assignment [28] . After taxonomic assignment and demultiplexing , OTUs present only in reagent control samples were subtracted from O . felineus infected and O . felineus non-infected groups to eliminate reads due to contamination . Samples with ≥200 counts were included in the analysis . Alpha diversity within and between groups ( infected or not-infected with O . felineus ) samples of was calculated in QIIME using Chao1 , Shannon and Simpson alpha metrics at depth of 200 sequences per sample . Alpha diversity comparisons were calculated using a two-sample non-parametric t-test and 999 Monte Carlo permutations . Beta diversity was investigated by principal components analysis ( PCoA ) both on non-normalized and normalized ( CSS-algorithm [29] ) data with the usage of unweighted Unifrac distance and validated with ANOSIM in QIIME . To examine significance of variation among groups at levels of phylum , we used fitZig model , a metagenomeSeq-package in the R statistical environment [29] . Metagenomic prediction was undertaken using the Galaxy-based PICRUSt algorithm [30] against KEGG database , with statistical analysis of variation among groups analyzed using the Mann-Whitney-Wilcoxon test and logistic regression . S1 Table outlines the pipeline . To visualize genera , we associated each genus with the OTU from GreenGenes DB and prepared a table of the OTU findings that represents all genera identified . The taxonomic tree , which was created during taxonomic classification stage , was pruned with the usage of the genus list by filter_tree . py script in QIIME workflow . Radial phylograms were constructed using FigTree 1 . 4 . 2 and MEGA 7 http://www . megasoftware . net/ [31] .
Patients hospitalized with cholelithiasis and who were diagnosed also with ( n = 30 ) and without ( n = 26 ) infection with O . felineus participated in the study . These two groups were similar in age and gender , and relatively similar numbers in each group presented with the comorbidities of pancreatitis and infection with hepatitis C virus ( Table 1 ) . The Illumina sequencing produced 1 , 547 , 628 reads . Demultiplexing showed 628 , 111 reads were suitable for further analysis . After extraction of reagent contamination controls , there are 81 , 627 reads and 2 , 797 discrete OTUs were identified . Taxonomic composition consisted of archaeal and bacterial super-kingdoms , 25 different phyla , 55 classes , 84 orders , 147 families , 246 genera ( Fig 1 ) , along with 77 species-level phylotypes that were well supported . Supplementary S2 Table lists several of these latter phylotypes . The median number of reads per sample was 585 ( range , 5–10037 ) . However , this wide range in numbers of reads per sample , which spanned two orders of magnitude , hindered comparison among the samples . Accordingly , samples with < 200 reads were not included in the subsequent analysis . After this filtration , reads from the remaining 37 samples were analyzed in depth ( Table 1 ) . Alpha diversity was estimated after rarefaction at a depth of 200 sequences per sample by using richness metrics ( Chao1 , the Shannon and Simpson diversity index ) [32] . Analyses of microbial communities did not reveal differences in richness ( Chao1 ( Fig 2A ) ) , Shannon and Simpson indices ( S1 Fig ) between participants infected with O . felineus and non-infected individuals . Principal components analysis ( PCoA ) of the beta diversity , i . e . community diversity ( compositional heterogeneity ) / divergence among samples was undertaken using QIIME , wherein unweighted UniFrac distances ascertained beta diversity . In the case where we used non-normalized phylogenetic data , the first principal coordinate , PC1 accounted for 14 . 97% of total variance , and after CSS normalization was , PC1 accounted for 19 . 65% of total variance ( Fig 2B and 2C ) . This difference in bacterial communities between the O . felineus-infected and uninfected participants was significant , although not robust , and was confirmed using the non-parametric statistical test analysis of similarity ( ANOSIM ) , unweighted Unifrac—R = 0 . 12 , P = 0 . 02 ( normalized data ) . As presented in S2 Fig , hierarchical clustering analysis confirmed these modest differences among the bacterial communities . To consider the influence of host sex on cholelithiasis [33] , we examined the richness metrics ( Chao1 , the Shannon and Simpson diversity index ) after rarefaction at a depth of 200 sequences per sample from the female versus male participants . Chao1 analysis revealed that the diversity was higher in the female in comparison with the male participants ( p = 0 . 0461 ) ( S3 Fig ) . Four phyla , the Proteobacteria , Firmicutes , Bacteroidetes and Actinobacteria dominated the biliary microbiota in the participants of this study , all of whom were diagnosed also with cholelithiasis ( Fig 3; S4 Fig ) . However , the contribution by members of phylum Spirochaetes was significantly increased during infection with O . felineus . At the level of genus , this was exemplified by increases in Treponema ( Fig 4; Table 2 ) . Also at the phylum level , higher proportions of Planctomycetes ( P ≤ 0 . 01 ) , Synergistetes ( P ≤ 0 . 01 ) , Verrucomicrobia ( P ≤ 0 . 01 ) , and TM7 ( P ≤ 0 . 01 ) were evident in the group of participants infected with O . felineus in comparison with the uninfected participants . We identified all significant taxa aggregated to OTUs in the bile microbiota associated with the liver fluke infection . Differences were apparent at taxonomic levels from phylum to genus . At the level of genus , 22 phylotypes differed between these two groups . Most phylotypes that differed were detected in higher abundance ( i . e . absolute read counts ) in bile from the O . felineus-infected participants ( Fig 4; Table 2 ) . Among specific examples , there was elevated abundance of Klebsiella spp . , Aggregatibacter spp . , Lactobacillus spp . , Treponema spp . , Haemophilus parainfluenzae and Staphylococcus equorum in bile of participants infected with liver flukes . In addition , Veillonella dispar , Paracoccus aminovorans , Parabacteroides distasonis , Sphingomonas changbaiensis , Cellulosimicrobium sp . , Phycicoccus sp . and others were detected solely in bile from persons infected with O . felineus ( Fig 4; Table 2 ) , whereas Flectobacillus sp . , Xanthobacter sp . , Burkholderia sp . , Streptomyces sp . , Jeotgalicoccus psychrophilus and Treponema socranskii increased in the uninfected group vs the group with infection with O . felineus ( Table 2 ) . Reads assigned to the super-kingdom Archaea were identified in the microbial community of bile from one of the O . felineus-infected persons; these reads aggregated with a phylotype from the Phylum Euryarcheota , genus Halogeometricum . Given the potential for pathogenic microbes for involvement in cholelithiasis [6 , 34] , a list of phylotypes identified in bile samples is presented . Also , we searched the list of phylotypes for the presence of bacteria that had been described as associated with the human biliary tract by Shen and coworkers [7] . We compared the list of phylotypes detected in the present study in the bile of participants presenting with gallstone disease ( 37 individuals ) with the list of microbes recently described in human gallstones and bile [7] . About 9% of the same species were identified here , including Rothia aeria , Haemophilus influenza , Veillonella dispar , Acinetobacter johnsonii , Acinetobacter lwoffii and Streptococcus anginosus ( S2 Table ) . The prediction of functional KEGG pathway abundances from the 16S rDNA-based metagenomes was accomplished using PICRUSt . The same predicted functional pathways characterized the O . felineus infected and O . felineus non-infected bile , so that functional differences were not evident . Predicted metagenomes at the three hierarchical KEGG pathway levels revealed the functional categories represented in the bile microbiota of patients with cholelithiasis . Membrane transport , carbohydrate metabolism and amino acid accounted for more than one third of the hypothetical functions from the KEGG pathways at level 2 ( S3 Table ) . Sequence data obtained have been deposited to the European Nucleotide Archive , accession number PRJEB12755 , http://www . ebi . ac . uk/ena/data/view/PRJEB12755 .
Although it had been assumed that the biliary system in a healthy person is a sterile organ , it is now apparent that bile supports a complex microbiome in otherwise healthy individuals [4 , 9] . Nonetheless it has long been known that cholelithiasis , cholecystitis and cholangitis lead to bacteriobilia [35 , 36] . The presence of bacteria in the bile and gallbladder/gallstones has been diagnosed by microbial culture , where positive culture of bile during cholelithiasis and chronic cholecystitis ranges from 0–81% [37 , 38] . Frequently identified are Escherichia coli , and species of Enterococcus , Klebsiella , and Pseudomonas [38–40] . Analysis by pyro-sequencing targeting the bacterial 16S rRNA gene revealed that phylotypes of the phylum Firmicutes were dominate the bile of healthy pigs , with Proteobacteria and Actinobacteria also prominent , and with lesser contributions from other phyla . Firmicutes , Proteobacteria and Bacterioidetes , dominate the human biliary microbiome of gallstones and bile during cholelithiasis [6] . Our present findings accord with these reports [6] . Biliary tract microbiota of participants with cholelithiasis showed substantial person-to-person variation; the relative abundance of phylum Firmicutes varies 0–92% through the different samples . Similar phenomena have been reported for microbiota of gallstones from residents of Kunming , China [7] . Nonetheless , species contributing to biliary microbiota of the participants from Siberia differed markedly from microbes reported form China . Phylotypes previously identified in bile also were present , including and Haemophilus parainfluenzae , Enterobacter cloacae [41 , 42] , and Streptococcus anginosus , which is associated with pyogenic liver abscess [43] . In addition , microbes associated with periodontal disease , including Treponema socranskii [44] , T . amylovorum [45] , Veillonella dispar , [46] , Aggregatibacter segnis [47] , and Bacteroides eggerthii [48] were identified . Others more usually known from the external environment , including soil , plants , and rivers , also were identified including Sphingomonas changbaiensis , Rathayibacter caricis , Bacillus flexus , Methylobacterium adhaesivum , Psychrobacter pacificensis , and Pseudomonas umsongensis . Although alpha diversity of the biliary microbiome did not appear to be impacted during infection with O . felineus . A diverse often contradictory literature has accumulated over past decade on the influence of helminth infection on the microbial diversity of the intestines . Among other examples , polyparasitsm by soil-transmitted nematodes ( Ascaris , Tichuris , hookworms ) results in increased diversity of gut microbiota in indigenous Malaysians and microbial diversity decreases following deworming [49] . By contrast , in other situations , increasing alpha diversity is not apparent during trichuriasis [50] . In comparison , infection with O . felineus lead to the modification of composition of the bile microbiome . Specifically , most of the phylotypes that differed were detected in higher abundance in bile during opisthorchiasis although some phylotypes decreased; Jeotgalicoccus psychrophilus , a Gram-positive halophile [51 , 52] was included among the latter . Lactobacillus spp . increased in richness in O . felineus-infected bile . Colonization of the gut by nematodes has been shown to be associated with increasing prominence of Lactobacillaceae . Mice parasitized by the intestinal nematode Heligmosomoides polygyrus exhibit increased numbers of Lactobacillaceae in the ileum [53] and in the duodenum [54] . Chronic infection of mice with the whipworm Trichuris muris also increases the abundance of Lactobacillus spp . [55] , and similarly hamsters infected with O . viverrini-infected exhibit more Lactobacillus in the colon [15] . Intriguingly , Lactobacillus species may contribute probiotic defense against allergies [56 , 57] . In regions endemic for opisthorchiasis felinea , specifically in western Siberia , liver fluke infection modifies genetic risk of atopic bronchial asthma [58] . Furthermore , in urban regions , the presence of antibodies to O . felineus negatively correlates with the atopic sensitization [59] . There is evidence that the modification of the microbiota by helminths contributes to modulation of allergic inflammation [60 , 61] . Our data provided additional support that helminth infection promotes the increase in numbers of Lactobacillus species that , in turn , influences the paradoxical relationship between allergic diseases and helminthiasis . Haemophilus parainfluenzae also increased in O . felineus infected samples; this pathogen is associated with liver abscess [62] , and liver abscess represents a serious complication of opisthorchiasis felinea [63] . Veillonella dispar , Paracoccus aminovorans , Parabacteroides distasonis , Sphingomonas changbaiensis , among others , were constituents of the biliary microbiome of the liver fluke-positive participants . Although V . dispar is known from bile [53] , P . distasonis has been described from feces as a risk factor for obesity [64] . These two phylotypes represent microbes typically seen in the human alimentary tract . By contrast , S . changbaiensis is known from forest soils [12] and Paracoccus aminovorans associates with the skin of fish [65] [66] . In addition , we identified reads that aggregated with the archaeal genus Halogeometricum ( phylum Euryarcheota ) . Flesh of salted , dried river fishes represents a dietary stable in regions of Siberia [67] . We speculate that Halogeometricum and Paracoccus aminovorans may have been transported to the biliary tract with ingested dried fish and/or other fish products contaminated with metacercarie of O . felineus . Other phylotypes of the Euryarcheota occur in bile of hamsters infected with metacercariae of O . viverrini [15] . Conveyance of these environmental microbes from the outside world to the human alimentary tract may have been accomplished during establishment of infection by the liver flukes . Notwithstanding the novelty and complexity of the findings , our study has limitations . The findings associated with O . felineus took place in the setting of concomitant gallstone disease . The microbial profile of the bile may differ in the absence of cholelithiasis , and furthermore , the pH of the bile ( which was not measured here ) may have influenced the microbiome [68 , 69] . Metabolic changes associated with gallstone formation can lead to microflora discrete from that of healthy individuals [6] . Moreover , we cannot exclude that participants in the non-liver fluke infected cohort had not previously been infected given elevated prevalence of opisthorchiasis felinea in the Tomsk region [70] . Nonetheless , these findings appear to be novel in the context of the biliary microbiome during opisthorchiasis . It will be informative to investigate this phenomenon further , including in people without gallstone disease living in regions where liver flukes are endemic and infection with which represents increased risk for bile duct cancer . | The microbiota of the alimentary tract and other sites of the body influences human health . Contrary to popular belief , the bile within the liver is not sterile , and may host a microbiome consisting of diverse species of microbes . The spectrum of microbial species and their numbers within the biliary system may be influenced by disease including infection with pathogens such as parasitic worms and with gallstone disease , liver cancer and other ailments . Here we examined the microbes in the bile of patients from western Siberia , Russia who were concurrently infected with a food-borne parasitic worm , the liver fluke Opisthorchis felineus . Infection with this liver fluke is common in western Siberia , as a consequence of dietary preference for undercooked or smoked fresh-water fishes that often carry the larva of the liver fluke . Using high throughput sequencing targeting a conserved bacterial gene and statistical analyses , numerous bacterial species were identified in the bile of the patients . Infection with the liver fluke modified the biliary microbiome , resulting in abundant and diverse species of bacteria and Archaea . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"microbiome",
"gut",
"bacteria",
"bile",
"microbiology",
"tropical",
"diseases",
"parasitic",
"diseases",
"animals",
"trematodes",
"foodborne",
"trematodiases",
"gastroenterology",
"and",
"hepatol... | 2016 | Biliary Microbiota, Gallstone Disease and Infection with Opisthorchis felineus |
Programs for schistosomiasis control are advancing worldwide , with many benefits noted in terms of disease reduction . Yet risk of reinfection and recurrent disease remain , even in areas with high treatment coverage . In the search for means to better prevent new Schistosoma infections , attention has returned to an older strategy for transmission control , i . e . , chemical mollusciciding , to suppress intermediate host snail species responsible for S . mansoni and S . haematobium transmission . The objective of this systematic review and meta-analysis was to summarize prior experience in molluscicide-based control of Bulinus and Biomphalaria spp . snails , and estimate its impact on local human Schistosoma infection . The review was registered at inception with PROSPERO ( CRD42013006869 ) . Studies were identified by online database searches and hand searches of private archives . Eligible studies included published or unpublished mollusciciding field trials performed before January 2014 involving host snails for S . mansoni or S . haematobium , with a primary focus on the use of niclosamide . Among 63 included papers , there was large variability in terms of molluscicide dosing , and treatment intervals varied from 3–52 weeks depending on location , water source , and type of application . Among 35 studies reporting on prevalence , random effects meta-analysis indicated that , on average , odds of infection were reduced 77% ( OR 0 . 23 , CI95% 0 . 17 , 0 . 31 ) during the course of mollusciciding , with increased impact if combined with drug therapy , and progressively greater impact over time . In 17 studies reporting local incidence , risk of new infection was reduced 64% ( RR 0 . 36 CI95% 0 . 25 , 0 . 5 ) , but additional drug treatment did not appear to influence incidence effects . While there are hurdles to implementing molluscicide control , its impact on local transmission is typically strong , albeit incomplete . Based on past experience , regular focal mollusciciding is likely to contribute significantly to the move toward elimination of schistosomiasis in high risk areas .
Schistosomiasis , the chronic human disease caused by Schistosoma spp . parasite infections , is a preventable illness that , if left untreated , is associated with long-term undernutrition , anemia , organ scarring and fibrosis , resulting in disabling patient symptoms [1 , 2] . Current anti-schistosomiasis chemotherapy programs focus on controlling or preventing morbidity by treating school-age children who typically have the highest levels of Schistosoma infection [3] . However , because pre-school infection [4–6] and recurrent infection during childhood [7 , 8] are associated with significant risk for disease , optimal disease prevention can occur only when parasite infection or reinfection can be effectively blocked [9] . By themselves , Preventive Chemotherapy ( PCT ) campaigns [3] using mass drug administration have not been very successful in limiting transmission in high-risk areas [4 , 10–13] . The WHO roadmap’s new focus on 'transmission control , wherever possible' [14] means it is appropriate to re-examine the efficacy of intermediate-host snail control for prevention of human-to-snail-to-human parasite transmission . Reduction in infected snail numbers at the places where humans come into contact with freshwater could substantially reduce each patient's frequency of exposure to infecting parasite larvae ( cercariae ) , and , hence , reduce the frequency of reinfection . In the 1960s early theoretical modelling [15] suggested that a greater than 90% reduction in snail numbers , in conjunction with population drug treatment , had the potential to extinguish Schistosoma populations from local ecosystems . Chemical molluscicides , including copper sulfate , sodium pentachlorophenate ( NaPCP ) , N-tritylmorpholine ( Frescon ) , and niclosamide ( Bayluscide , Bayer 73 ) were used extensively in the 1950s , 1960s , and 1970s for schistosomiasis control in Africa , South America and Asia [16] , but following the introduction of oral drug therapies , molluscicides have not seen as much use in the last 30 years [17] . As present-day programs contemplate integrated strategies for schistosomiasis control , it is important to systematically review the efficacy of molluscicide use in snail suppression and its effectiveness for infection prevention , so that planners can project likely costs and impacts when targeting parasite elimination in at-risk locations . In the present systematic review and meta-analysis , we compiled the results of field trials of chemical mollusciciding focusing primarily on control of S . mansoni and S . haematobium species and the use of niclosamide molluscicide , now the most commonly used agent for host snail control . The decision to include just two parasite species was derived from the African focus of our sponsor , the Schistosomiasis Consortium for Operational Research and Evaluation ( SCORE ) , and was taken in light of previous publication of a meta-analysis of mollusciciding impacts in China [18 , 19] . While we encountered a number of limitations in the available study literature , we found sufficient quantitative evidence that routine mollusciciding can effectively reduce snail numbers in a manner that significantly reduces reinfection or new Schistosoma infection in typical at-risk human populations .
The data used in this project were aggregated , anonymized data from previously published studies; as such , this study does not constitute human subjects research according to U . S . Department of Health and Human Services guidelines ( http://www . hhs . gov/ohrp/policy/checklists ) . The protocol for this project was developed prospectively by the authors , then registered and published in the International Prospective Register of Systemic Reviews ( PROSPERO ) online database , http://www . crd . york . ac . uk/prospero/index . asp , number CRD42013006869 , on 16 December 2013 . Our a priori review question was , “Does chemical mollusciciding effectively reduce snail numbers in a manner to prevent reinfection or new Schistosoma infection in at risk human populations ? ” focusing primarily on control of S . mansoni and S . haematobium species and the use of niclosamide molluscicide ( 2-amino ethanol salt of 2' , 5'-dichloro-4'-nitro salicylanilide , sold as Bayluscide , Mollutox , and other names ) . The PRISMA checklist and the PROSPERO protocol for this study are provided as Supporting Information files S1 and S2 Files . To quantify the effects of repetitive use of chemical mollusciciding , we aimed to include any available published or unpublished reports on its use for control of Bulinus or Biomphalaria species for prevention of S . haematobium or S . mansoni infection . No limits were placed in terms of location or language of the report . However , we did not include studies of S . japonicum or S . mekongi control , which was the topic of a recently published meta-analysis from China [18 , 19] . Studies had to include periodic application of chemical compounds to transmission water contact sites or experimental locations , as well as the names of the snail species treated , and treatment doses , frequency , habitat ( static vs . flowing water ) , region , and season of application . Information about local human prevalence and incidence of Schistosoma infection , before and after intervention , was also sought as secondary outcomes for meta-analysis . We aimed to include any studies performed after the development of niclosamide molluscicide compounds , ( i . e . , after January 1961 ) to the close of the search phase of the project , 1 January 2014 . Historical perspectives , observational studies and prospective trials were eligible for inclusion if they provided the necessary quantitative data . We identified published studies using PubMed , Google Scholar , Web of Science , SCIELO , African Journals Online , as well as resources such as WHO technical reports and archived files at Case Western Reserve University and the Schistosomiasis Consortium for Operational Research and Evaluation ( SCORE ) . Where published bibliographies of the recovered studies were found to contain promising citations ( including grey literature ) not included in online searches , these papers were obtained , whenever possible , and screened for inclusion in the meta-analysis . We examined the available electronic database literature using combination searches of the following terms: 'molluscicide'; 'snail control/prevention'; 'Biomphalaria ( Australorbis ) ’; 'Bulinus'; 'field [trial]' ‘schistosomiasis/prevention and control’ ‘transmission’ , and/or 'niclosamide’ . Secondary report finding was done by scanning PubMed 'similar articles' feature , and by using the Google- and PubMed-generated listings of papers that cited papers that we found to contain well-conducted snail control intervention trails . As relevant articles were identified , we broadened our search by accessing additional titles through the online databases’ automated ‘related articles’ links . Full titles and abstracts were recovered for the initial screening phase of study selection . Review of titles and abstracts was performed by two trained reviewers , searching for data content meeting study requirements . The studies found suitable for inclusion—including historical , observational , and prospective studies—were then obtained for full-text review from online or library sources . Where a single report contained data on multiple individual community surveys , each survey was also separately abstracted for inclusion in some of the sub-group comparison analysis . We excluded studies where sufficient details of snail control measures were not reported , or when the data on the community or individual Schistosoma infection levels were not sufficiently detailed to confirm the reported incidence and prevalence of infection following the implementation of niclosamide or other molluscicide treatments . Cases of duplicate publication or extended analysis of previously published data were also excluded . Full listings of included and excluded studies are provided as Supporting Information files S3 and S4 Files . Included papers were abstracted and their relevant features entered into a purpose-built database created in Microsoft Excel 2013 software ( Redmond , WA ) . These papers were archived by the authors in both paper and electronic ( pdf ) formats at the Center for Global Health and Diseases , Case Western Reserve University . In addition to full citation information and year of publication , information was collected on the country and region where the study was performed , along with snail genus and species , chemical mollusciciding treatment , whether the study was performed in a research laboratory or in the field , the concentration range of molluscicide delivered , and percentage kill of the observed molluscs . The effective days of molluscicide-mediated snail control were also captured , as well as the beginning and end values for population-wide incidence and prevalence . Data entries were fully verified by the second reviewer before final data analysis was carried out . Reported data on snail outcomes was quite diverse in terms of delivery , metrics , and timing of interventions and follow ups . Those results were thus summarized only qualitatively . The impact of snail control interventions on local prevalence and incidence of infection among at-risk human populations could be compared across studies , however . Results from each study and sub-study were entered into Comprehensive Meta-Analysis software , v . 3 ( CMA , Biostat , Englewood , NJ ) for calculation of summary estimates of treatment impacts , along with their confidence intervals . Potential modifying factors assessed in preliminary analysis included parasite species , starting infection prevalence , age-group monitored , study era , region , duration of control , type of habitat , and impact of added drug treatments . Heterogeneity among studies was expected due to differences in habitat , snail and parasite species , and human populations involved [16] . Heterogeneity levels were scored using Higgins’s and Thompson’s I2 statistic [20] . Summary estimates of intervention effects were computed using Der Simonian and Laird random-effects modeling [21] implemented in CMA software . The data sets used for this analysis are provided as Supporting Information files S1 and S2 Datasets . Linear meta-regression by CMA was also used to estimate the impact of duration of control on the odds/risk of Schistosoma infection . Potential publication bias of studies reporting impact on human incidence or prevalence of Schistosoma infections was assessed by visual inspection of funnel plots , and calculation of the Egger test for plot asymmetry . These funnel plots and statistics are provided as Supporting Information file S5 File .
Fig 1 contains a flow chart that details the results of the search and selection strategy for the studies included in this systemic review . Of the 357 listings recovered , 315 were obtained from online database searches , and 42 from bibliographies and archived materials . After titles and abstracts were assessed , 140 reports were selected for full review , ultimately yielding 63 studies ( from 40 papers ) to be included in the qualitative or quantitative analysis reported here ( see Supporting Information files: S3 File ‘Listing of Included Studies’ and S4 File ‘Listing of Excluded Studies’ ) . As detailed in the Methods section , the included studies were not limited in terms of publication date or language . The study focus was limited , however , to the control of Biomphalaria or Bulinus snail species for prevention of S . mansoni and S . haematobium transmission . Studies of S . japonicum and other human and animal schistosome species were not included . Sixty-three reports provided information on mollusciciding’s impact on snail numbers and/or the duration of its suppressive effects . Of these , one was a graduate student thesis and the remainder were reports published in peer-reviewed journals . Twenty-seven ( published ) papers reported on the impact of mollusciciding campaigns on Schistosoma spp . prevalence among the local human population , and 12 papers provided estimates of the effects on local incidence of infection among children and/or adults . Whereas most studies reported on the use of niclosamide compounds for snail control , 6 studies used either sodium pentachlorophenate ( NaPCP ) [22–24] or N-tritylmorpholine ( Frescon ) [25] as molluscicides . For studies of niclosamide effects , publication dates ranged from 1960 to 2013 ( median 1981 ) . Of the 63 studies , 33% were from East Africa ( Tanzania , Kenya , Sudan , Ethiopia ) , 17% were from southern Africa ( Zimbabwe and South Africa ) , 17% were from Brazil or the Caribbean ( St . Lucia ) , 11% were from North Africa ( Egypt and Morocco ) , 11% were from West Africa ( Gambia , Ghana , Mali ) , 6% were from Central Africa ( Burundi , Cameroon ) , and 4% were from Iran . Among 35 studies reporting impact on human Schistosoma spp . infection prevalence , 14 ( 40% ) focused on prevalence among school age children , 20 ( 57% ) reported results for the general population , and 1 ( 3% ) reported separate results for adults and children . Eleven ( 31% ) of these studies reported the impact of snail control as used alone , 20 programs ( 57% ) employed snail control plus some form of community-based screening and treatment intervention ( i . e . , only egg-positive persons were treated ) , 2 studies ( 6% ) used snail control plus a school-based treatment strategy , and 2 ( 6% ) used snail control combined with targeted mass drug administration . Research performed in the pre-1990s era often did not report quality-related trial details in peer-reviewed publications , and therefore , we did not pursue quality weighting in the present analysis . Many of the researchers who were involved are now deceased or retired , so access to primary data was not available . Most of the reported studies involved a single intervention site , the studies were non-randomized , and for the most part the comparison of intervention effects involved historical and not concurrent comparison data ( i . e . , they used a one-group , pre-test/post-test study design [26] ) . It is possible that the selection of study sites favored extremely high or low transmission locations . In the studies where concurrent untreated comparison sites were monitored , non-homogeneity of risk between the treated and untreated areas was often likely ( e . g . , Tamiem et al . [27] ) . Threats to validity in assessing molluscicide-related impact on Schistosoma transmission included secular trends among other interventions , and maturation of environments [23 , 28] , heterogeneities within landscapes and populations [29] , loss to follow up , and the lesser reliability of egg-count diagnostics as prevalence and intensity of Schistosoma infections decline [30–33] . In addition , there was potential unreliability of treatment implementation , and unknown risk of reinfection from population movement within or outside the control areas [34 , 35] . The role of diffusion of information about schistosomiasis was also unknown . Observation ( Hawthorne ) effects were possible during implementation , particularly on plantation estate programs run by employers [36 , 37] . However , in our analysis of reporting bias , we did not find evidence of underreporting of negative results or adverse outcomes ( see Supporting Information file S5 File for funnel plot and regression analysis ) . A wide variety of molluscicide delivery approaches were used in the included studies , depending primarily on the speed of water flow in the snail habitat , the extent of the water area to be treated , and the multiplicity of local human water contact sites in the area targeted for control . More rapidly flowing water bodies were often treated by drip-feed delivery systems that provided a constant dose of molluscicide to a stream or canal over a period of 1–2 days [38–41] . In irrigation schemes , where water flow could be diverted and controlled , some projects utilized impoundments of molluscicide-containing water that could be slowly transferred through the canal system to fully treat the entire irrigation system [42 , 43] . Slow moving streams and static and seasonal ponds were often treated with focal spraying of shallows and vegetation at the water’s edge—the primary habitat of the intermediate host Biomphalaria and Bulinus spp . snails . Large lakes presented a special problem for treatment dosing because of rapid molluscicide dispersal by currents and wave action [44] . In one study , plastic sheeting was employed to isolate lake shore areas to retain molluscicide for a sufficient time to effect snail control [45] . Reporting on the lethal impact of molluscicide treatment and duration of its effects varied widely among studies . Early laboratory and field trials established that niclosamide concentrations of 0 . 1 to 3 mg/L ( units equivalent to ‘parts per million’ ( ppm ) , a term often used in the older literature ) could kill over 90% of intermediate host snails [46] , and that the dose effect was dependent on the duration of exposure—lower concentrations ( 0 . 04–0 . 53 mg/L ) were effective if applied for 24h , higher concentrations ( 0 . 5–1 . 2 mg/L ) could be lethal if applied for only 6h [47–49] . In the schistosomiasis control campaigns included in our meta-analysis , the estimated concentrations delivered to local water bodies ranged from 0 . 025 [37 , 50] to 10 mg/L [51] . The median and most common dose target in our included studies was 1 mg/L , consonant with the experience summarized by Andrews , et al . in their detailed 1983 review [46] . Immediate impact of mollusciciding was usually assessed at breeding sites 24 h after delivery [52] , and then if living snails were still present , reapplication of chemical , sometimes at higher concentrations [53] , was frequently used to maximize snail suppression . Where drip feed administration was applied to running streams , significant molluscicide effects on snail numbers were detectable 900 meters [54] , 1375 meters [55] , 1700 meters [40] , even up to 10 km [38] downstream . Snail mortality was assessed in many different ways . Some studies reported live/dead snails at various intervals after treatment , others reported on treatment impact on caged sentinel snails placed in the treated water habitats [38 , 54] . Because early reapplication was used in several programs but not clearly detailed , we could not determine a summary estimate of niclosamide efficacy in terms of ( dose X exposure time ) [46 , 47 , 49 , 54 , 56–60] across the reported field studies . Vegetation , wave action , and debris were noted in several studies to be confounding factors affecting the molluscicidal efficacy of chemical applications [24 , 44 , 45 , 61] . Many projects achieved 100% elimination of targeted snails for periods lasting from several weeks to several months . Some projects failed to reach 100% snail elimination following mollusciciding [35 , 39 , 55 , 58 , 61–64] . Nevertheless , these sites were able to obtain 88–99% immediate reduction in snail numbers . When snail re-emergence occurred , repopulation times ranged from 2 weeks [62] up to 18 months [65] , depending on location and habitat . Serial monitoring for significant snail repopulation at human water contact sites was an important part of implementation , and was most often used to decide the intervals needed for repeated mollusciciding . Fig 2 indicates the between-treatment mollusciciding intervals reported by 47 studies of S . mansoni and S . haematobium control in different areas of Asia ( Iran ) , Africa , South America ( Brazil ) and the Caribbean ( St . Lucia and Puerto Rico ) . There was a large range of working between-treatment intervals reported ( 21 days to 365 days ) , depending in part on the seasonality of transmission , the type of water treated ( flowing , static , or canal ) , and the desired lethality of the treatment applied ( suppression vs . elimination of snails ) . Decisions regarding treatment intervals were most often based on snail repopulation detected on regular waterside surveys at treated water contact sites . The median interval used in the reported studies was 90 days , with an inter-quartile range ( IQR ) of 42–90 days . Thirty-five studies reported in 28 publications [8 , 22–24 , 27 , 28 , 34 , 35 , 39 , 41 , 42 , 66–81] reported on the interval impact of programs that included snail control campaigns on the prevalence of detectable Schistosoma infection among local human populations . Heterogeneity in prevalence outcomes was quite high ( I2 = 99 . 9 ) among studies . Fig 3 indicates the pre- and post-intervention levels of Schistosoma prevalence for individual studies; the median pre-control prevalence for all studies was 45% , ( Range: 5% to 92% , IQR 26% to 58% ) while the median post-control prevalence was significantly lower , 17 . 5% ( Range: 0% to 53% , IQR 6% to 30% , ( P < 0 . 001 by Wilcoxon signed rank test ) ) . Supporting Information file S1 Fig shows the forest plot for the prevalence studies . Fig 4 graphs summary estimates of the odds ratio for infection after molluscicide treatment intervention as compared to pre-treatment levels ( numeric details for this graph are provided in Supporting Information file S1 Table ) . Overall , surveyed populations had a significantly reduced odds of infection ( OR 0 . 23 , CI95% 0 . 169 , 0 . 309 ) following snail control intervention . The impact was less strong where snail control was used alone ( OR 0 . 47 , CI95% 0 . 276 , 0 . 800 ) , and greatest among studies where snail control was combined with community-based screening and treatment programs ( OR 0 . 162 , CI95% 0 . 116 , 0 . 225 ) . There was not a significant difference in terms of impact between S . mansoni- and S . haematobium-endemic locations . Treatment of natural water sites had greater overall impact than treatment of irrigation systems , which may account for the observation that North Africa ( primarily irrigation locations ) had less improvement than sites elsewhere in Africa , Asia , South America , and the Caribbean . Studies focused only on school age children reported lesser gains in terms of post-intervention prevalence when compared to general population studies , likely reflective of school age groups’ much greater risk for infection/reinfection . Not shown , starting prevalence of infection did not have a clear effect on the size of prevalence reductions obtained during a mollusciciding program ( for details , see Supporting Information file S3 Fig , which shows a forest plot of the range of outcomes data ( ORs ) arranged according to starting prevalence of Schistosoma infection ) . Of note , three ( 9% ) of the 35 studies [66 , 77 , 81] , two in Egypt and one in Zimbabwe , did not demonstrate reductions in local Schistosoma prevalence . In addition , another three studies reported less than a five percentage point drop in local human Schistosoma prevalence during their mollusciciding trial period [24 , 39 , 73] . These three less successful studies were performed in Egypt and in Liberia and involved both S . mansoni and S . haematobium areas . A summary of the implementation , population , and environmental features of these six projects having relatively limited mollusciciding impact is included in Supporting Information file S2 Table . Their individual reports provided several possible explanations for their limited program impact . These included i ) having only a short duration of follow-up ( i . e . , 1 year after mollusciciding implementation ) [24]; ii ) a lesser impact of supplemental drug treatments on S . mansoni as compared to S . haematobium [39 , 77]; iii ) inability of the implemented molluscicide program to reduce snail numbers at transmission sites [66 , 81]; and iv ) incorrect timing of mollusciciding application relative to maximal seasonal transmission [73] . In consideration of the long-term effects of multi-year programs , a meta-regression of mollusciciding effects on prevalence odds ratios vs . time is presented in Fig 5 . It suggests progressively greater reductions in infection prevalence as mollusciciding programs extend beyond the first few years of snail control . Seventeen studies reported in 12 publications [28 , 35 , 38 , 41 , 50 , 61 , 67 , 69 , 71 , 73 , 76 , 82] reported on the impact of mollusciciding programs on the incidence of new Schistosoma infections before and during control . As for prevalence , above , heterogeneity in incidence outcomes was high ( I2 statistic = 93 . 2 ) among the reported studies . From our random effects meta-analysis of all 17 studies , the risk of new infection was estimated to be reduced by 64% ( relative risk = 0 . 36 , CI95% 0 . 25 , 0 . 50 ) in schistosomiasis control programs that included mollusciciding . The forest plot for studies reporting incidence outcomes is provided in Supporting Information file S2 Fig . Yearly incidence dropped from a median 22% ( Range: 4% to 78% , IQR 14% to 54% ) before intervention to a median 8% ( Range: 2% to 80% , IQR 4% to 12% , P = 0 . 001 for the difference ) during the course of these mollusciciding intervention programs . Fig 6 graphs summary estimates of the risk ratio for infection after molluscicide treatment intervention as compared to pre-treatment levels ( numeric details are provided in Supporting Information file S3 Table ) . Of note , reduction of incidence was greater ( i . e . , RR was lower ) for areas with natural water sources as compared to irrigation schemes ( RR 0 . 36 vs . 0 . 55 ) . There was also no apparent difference in incidence reduction effect when drug treatments were included in the control programs ( RR for snail control alone was 0 . 33 vs . 0 . 32 for snail control plus community screening and drug treatment ) . Fig 7 shows the shift in pre- and post- incidence values for individual studies . Fig 8 graphs a meta-regression of the observed impact of mollusciciding on incidence ( in terms of log-risk ratio ) according to the duration of program implementation , indicating what appears to be an effect on incidence in most locations within 1–3 years . As noted earlier , most included studies focused on one area using historical control data to assess impact . Nine studies reported in eight publications [22 , 23 , 27 , 28 , 38 , 61 , 71 , 76] reported on concurrent infection outcomes in untreated areas near the molluscicide trial site . Table 1 summarizes the observed effects on prevalence and incidence of Schistosoma infection in the treated and untreated zones in each study . Of note , treatment and comparison zones had only one area unit each , and assignment was not randomized . In general , all molluscicide-treated areas saw greater declines in prevalence or incidence than untreated areas . Remarkably , though , prevalence fell significantly without molluscicide intervention ( or population-based drug treatments ) over a period 8 years in the Brazil study region [28] and over 13 years in Puerto Rican districts [22 , 23] , indicating interval changes in local risk for transmission that were unrelated to the snail control intervention . In the Ghana-2101 project , Lyons [76] observed a spontaneous drop in S . haematobium incidence over a 2 year period in an untreated comparison area , which was associated with an unexplained interval disappearance of local bulinid snails . Five study areas saw no change in prevalence in their untreated comparison areas but concurrent reductions in prevalence of 24% to 89% within snail control areas .
This systematic review and meta-analysis summarizes what has been a broad and lengthy experience with the use of mollusciciding for control of S . mansoni and S . haematobium transmission . Results of our analysis suggest that chemical-based snail control , particularly with the compound niclosamide , can effectively reduce local transmission of Schistosoma parasites when delivered at regular interval and under skilled supervision [17] . Direct treatment effects on snails were difficult to summarize , because of the many differences in sampling and reporting used in the included studies . However , where snail reductions were quantified , most programs saw very significant reductions or complete disappearance of local Schistosoma host snails during program implementation . Earlier studies tended to favor broad , intensive mollusciciding in an attempt to eliminate intermediate host Biomphalaria or Bulinus spp . snails . With such approaches , and particularly where water flow could be controlled in irrigation systems and transmission was more seasonal , treatment intervals could be extended to 6–12 months [8 , 79 , 83] . Later studies , often dealing with natural water bodies and more focal human water contact , tended to favor more frequent focal administration of mollusciciding [13 , 41 , 43 , 62 , 66 , 84] , allowing snails to persist elsewhere outside the main human water contact zones . While not fully explored , several preliminary studies suggested that slow-release strips or pellet formulations [85 , 86] or delayed-release molluscicide capsules [87] might better focus the impact of chemical molluscicide and extend its duration of impact , with concomitant cost-savings due to a reduced need for frequent delivery . Given the cumulative experience of the programs summarized here , it becomes clear that focality of snail habitats , combined with overlap into human water contact zones , represent factors in successful Schistosoma transmission . Not all waterbodies within a control area are suitable for host snails [88] , but human movement among water contact sites can strongly facilitate regional persistence of transmission [89 , 90] . For these reasons , implementation of focal snail control requires an adequate surveillance component to have an accurate working knowledge of local snail habitat , of human water contact zones , and of the seasonal factors affecting the abundance of snails and the likelihood of transmission [66 , 73 , 91] . As Shiff [92] points out , the ultimate value of a mollusciciding campaign is measured by its impact on human infection . Where snail control was used alone , early reductions in human incidence and later reductions in human Schistosoma infection prevalence could usually be obtained . When snail control was combined with population screening and selective or mass drug therapy , prevalence was reduced more quickly and incidence diminished . However , transmission was frequently not eliminated . When most successful , programs involving snail control achieved 85–100% reductions in local prevalence of Schistosoma infections [8 , 22 , 23 , 68–71 , 75 , 78] . However , some control programs appeared to have only minimal impact on local prevalence [39 , 66 , 73 , 77 , 81] . While there were some apparent differences in effects by region and by parasite species , the snail species named in the included studies were too diverse to draw meaningful comparisons for prevalence or incidence outcomes stratified at the intermediate snail host species level . Authors cited the fact that established control is vulnerable to resurgence of snail populations from local refugia [93–95] . The homing characteristics of miracidia for host snails , combined with the homing of cercariae towards human skin and the high degree of asexual multiplication within infected snails , strongly favor persistence of transmission [70] . Given the presence of untreated human individuals within the control area , whether from refusal of drug therapy or in-migration from ( or temporary travel to ) areas not under Schistosoma transmission control [64 , 70 , 76 , 77 , 96] , new infections are likely to continue to occur . Habitat changes , growth in human populations , and breakdowns in program performance are other factors that can contribute to limit the impact of snail control programs—whereas Egypt did very well with mollusciciding campaigns in the 1960s [71] , by the 1980s their programs were having limited effectiveness [66 , 73] . An independent on site WHO review performed in 1985 identified gaps in communication between snail control teams and health personnel , errors in selection of snail sampling sites , inefficiencies in snail testing , and an over-reliance on infrequent area-wide mollusciciding ( as opposed to more frequent focal mollusciciding ) as contributing causes to poor performance in that era [66] . The evidence summarized in this meta-analysis appears , in general , to favor mollusciciding as an effective method to reduce Schistosoma infections over time , with an additive effect on prevalence where population-based drug control is also given . However , the quality of the reported evidence is limited . The studies included in the analysis were non-randomized interventional trials , often with only historical data used for comparison in assessing the magnitude of snail control outcomes [26] . Where concurrent comparison areas were used , transmission was inconsistent in some areas , suggesting that secular trends or temporal fluctuations were occurring , which means that there is risk of over- or under-estimating the impact of mollusciciding in studies that use only historical controls . As such , and in view of the variability of ecology in Schistosoma transmission settings , the formal scientific evidence for a ‘generalizable’ consistent effect of snail control can be considered only minimally strong at this time . Other limitations in performing our analysis stem from study-to-study differences in snail control implementation , measures of snail impact , monitoring of human population outcomes , and duration of control . For the reader considering implementation of a snail control program based on niclosamide mollusciciding , the 1983 monograph by Andrews , et al . [46] provides extensive information on the chemistry , biology , and toxicology of niclosamide , as well as its effects on non-targeted plant and animal species . Niclosamide’s environmental impacts have been more recently reviewed by Dawson [97] , who concluded that there is minimal risk to humans and the environment , provided its application is appropriately dose-limited , informed , and supervised . It is important to be aware , however , that niclosamide is harmful to fish , amphibia , certain insect larvae , and in higher doses , to aquatic vegetation [46 , 97 , 98] . Because niclosamide quickly decays over 24 hours , animals that can rapidly move away from an area of application may return in a matter of days [62] . In aggregate , it appears that metered and very focal niclosamide administration at human water contact sites has the potential to provide the greatest impact on Schistosoma transmission with the least impact on local ecosystems . Overall , the impacts reported in the included studies predominantly lean toward a positive effect of mollusciciding in reducing Schistosoma transmission , with longer duration of control leading to a greater impact . These findings hold promise that the benefits of mollusciciding could be further defined in modern , well-designed comparison trials . A randomized comparison trial of mass drug administration ± snail control is in now progress in Zanzibar as part of a program that is attempting local elimination of S . haematobium [99] . Additional mollusciciding trials for S . mansoni control and elimination are under development . We look forward to the results of these efforts , which are expected to provide valuable evidence to further inform Schistosoma control policy . Based on past experience , regular focal mollusciciding is likely to contribute significantly to the move toward elimination of schistosomiasis in high risk areas . | Infection with Schistosoma blood flukes is a leading cause of chronic parasitic disease in at-risk areas of Africa , South America , Asia , and the Philippines . Over past decades , many national programs have implemented regular drug treatment to control or prevent the advanced complications of Schistosoma infection . However , these periodic treatments do not stop transmission of the parasite , which occurs when human sewage contaminates local water bodies and parasite eggs infect intermediate host snails . In this systematic review , we collated past experience of using chemically-mediated snail control for prevention of schistosomiasis . This approach , used in many Schistosoma-affected countries before the advent of the current oral drug regimens , has the potential to significantly reduce transmission if properly applied . Our meta-analysis of 63 studies ( performed 1953–1981 ) catalogued a wide variety of water treatments and schedules employed . Among studies reporting on human infection , we found that snail control reduced local human prevalence and incidence of infection in most , but not all locations . Estimates from the aggregated studies indicate that snail control ( alone ) typically reduced new infections by 64% and local prevalence declined over a period of years . This decline was accelerated and more profound ( 84% reduction ) if drug treatment was also made available . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Systematic Review and Meta-analysis of the Impact of Chemical-Based Mollusciciding for Control of Schistosoma mansoni and S. haematobium Transmission |
Reovirus infection leads to apoptosis in both cultured cells and the murine central nervous system ( CNS ) . NF-κB-driven transcription of proapoptotic cellular genes is required for the effector phase of the apoptotic response . Although both extrinsic death-receptor signaling pathways and intrinsic pathways involving mitochondrial injury are implicated in reovirus-induced apoptosis , mechanisms by which either of these pathways are activated and their relationship to NF-κB signaling following reovirus infection are unknown . The proapoptotic Bcl-2 family member , Bid , is activated by proteolytic cleavage following reovirus infection . To understand how reovirus integrates host signaling circuits to induce apoptosis , we examined proapoptotic signaling following infection of Bid-deficient cells . Although reovirus growth was not affected by the absence of Bid , cells lacking Bid failed to undergo apoptosis . Furthermore , we found that NF-κB activation is required for Bid cleavage and subsequent proapoptotic signaling . To examine the functional significance of Bid-dependent apoptosis in reovirus disease , we monitored fatal encephalitis caused by reovirus in the presence and absence of Bid . Survival of Bid-deficient mice was significantly enhanced in comparison to wild-type mice following either peroral or intracranial inoculation of reovirus . Decreased reovirus virulence in Bid-null mice was accompanied by a reduction in viral yield . These findings define a role for NF-κB-dependent cleavage of Bid in the cell death program initiated by viral infection and link Bid to viral virulence .
Tissue injury in response to infections by many viruses occurs as a consequence of apoptosis . Multiple studies using animal models of viral disease demonstrate a correlation between apoptotic potential and disease severity [1] , [2] , [3] , [4] . These observations highlight proapoptotic signaling following virus infection as an attractive target for antiviral therapy . However , despite its central importance in viral pathogenesis , gaps in knowledge about the identity of death signaling pathways that modulate virus-induced apoptosis in vivo , along with an incomplete understanding of how these signaling cascades are activated during virus infection , have hampered the deployment of this strategy for treatment of viral disease . Mammalian reoviruses injure infected cells via apoptosis both in culture and in tissues of infected animals . As such , studies of these viruses have contributed to an understanding of how virus infection culminates in apoptotic cell death . Unlike other viruses in which virulence correlates with cell-death capacity , the identity of viral and cellular factors that regulate reovirus-induced apoptosis in cell culture are for the most part known [4] , [5] , [6] , [7] , [8] , [9] , [10] . Moreover , many of these intermediaries also modulate reovirus-induced apoptosis in vivo [4] , [7] , [11] , [12] . Studies using reassortant reoviruses [5] , [6] , ectopically expressed proteins [13] , and genetically engineered reovirus mutants [4] , [7] highlight a critical role for reovirus outer-capsid protein µ1 in apoptosis induction . Collectively , these studies indicate that prodeath signaling evoked by µ1 occurs subsequent to membrane penetration but prior to synthesis of viral RNA or protein [4] , [7] , [14] , [15] . Classical death-receptor-mediated extrinsic apoptotic pathways stimulated by reovirus infection execute the death response [16] . Treatment of cells with soluble TRAIL receptors or expression of a dominant-negative form of Fas-associated death domain ( FADD ) protein blocks apoptosis , demonstrating that signaling via death receptors is required for execution of the apoptotic program [17] . In keeping with the function of extrinsic apoptotic signaling in reovirus infection , caspase-8 activation [18] and Bid cleavage [19] are observed in cells infected with reovirus [16] . Reovirus infection also stimulates intrinsic apoptotic pathways , as evidenced by release of cytochrome c and Smac/DIABLO from the mitochondria and activation of caspase-9 [16] , [20] , [21] , [22] , [23] . Concordantly , reovirus-induced apoptosis is dampened by over-expression of Bcl-2 [24] , which inhibits mitochondrial apoptotic pathway activation [25] . Bid is a proapoptotic BH3-only member of the Bcl-2 family that functions to link the extrinsic apoptotic pathway and the mitochondrial amplification loop of the intrinsic pathway . Following death-receptor signaling , cytoplasmically resident Bid is cleaved by activated caspase-8 to generate a truncated form of Bid known as tBid [26] . tBid translocates to the mitochondria and triggers the release of cytochrome c and activation of the core mitochondrial apoptotic machinery [19] , [27] . It is not known whether Bid plays a functional role in apoptosis induction by reovirus . Moreover , the relationship between apoptosis effector pathways and early events in viral replication are not understood . In addition to these classical apoptotic pathways , the innate immune response transcription factor , NF-κB , is activated following reovirus infection [8] . NF-κB activation by reovirus depends on the viral µ1 protein and can be accomplished by genome-deficient reovirus particles [4] , [7] , [8] . Blockade of NF-κB signaling using chemical inhibitors or cell lines genetically deficient in NF-κB p50 , NF-κB p65/RelA , IκB kinase ( IKK ) -α , or IKK adaptor IKKγ/NEMO significantly diminishes reovirus-induced apoptosis [8] , [28] . Consistent with these findings , activation of NF-κB occurs within the first few hours of reovirus infection and precedes the biochemical and morphological hallmarks of apoptotic cell death [8] , [28] . These observations suggest that NF-κB couples µ1-mediated events to the cellular apoptotic machinery . Although regulation and function of NF-κB has been extensively studied , the precise relationship between NF-κB and the cell-death machinery remains undefined . In this study , we examined the function of cellular apoptosis regulator Bid using genetically deficient murine embryo fibroblasts ( MEFs ) and mice . We found that while Bid is dispensable for reovirus replication in cell culture , its function is required for reovirus-induced apoptosis . Blockade of NF-κB signaling , which diminishes apoptosis induction by reovirus [8] , [28] , prevents cleavage of Bid . In comparison to wild-type mice , Bid-deficient mice display diminished susceptibility to reovirus-induced CNS disease following either peroral ( PO ) or intracranial ( IC ) inoculation . Attenuated reovirus virulence in the absence of Bid is associated with decreased reovirus replication in the murine CNS . These results define an important role for Bid in virus-induced apoptosis and disease and illuminate Bid-dependent prodeath signaling as a viable target for antiviral therapy .
Reovirus infection of HEK293 epithelial cells leads to a biphasic loss of full-length ( FL ) Bid [16] . Since a mitochondrial amplification loop through Bid is required for apoptosis only in some cell types , such as hepatocytes [29] , [30] , it is not known if Bid is cleaved in all cell types infected by reovirus . In addition , although calpains [31] , [32] , caspases [26] , [33] , [34] , [35] , and cathepsins [36] , [37] , [38] , [39] can mediate Bid cleavage and have been implicated in apoptosis induction by reovirus [15] , [16] , [40] , the precise identity of the protease that generates tBid following reovirus infection is not known . To determine whether tBid is generated following reovirus infection of fibroblasts , and to define the mechanism of Bid cleavage following reovirus infection , we infected murine L929 fibroblasts with reovirus strain type 3 Dearing ( T3D ) and monitored levels of FL Bid and tBid over 48 h ( Figure 1A ) . While levels of FL Bid remained unchanged in mock-infected cells , we observed loss of FL Bid between 24 and 48 h post-infection . Decreased levels of FL Bid correlated with a corresponding increase in levels of tBid . To determine whether the generation of tBid results in activation of the mitochondrial loop of the intrinsic apoptotic pathway , we assessed levels of procaspase-9 as a surrogate for the formation of the caspase-9-containing apoptosome ( Figure 1A ) . In a time frame consistent with cleavage-induced generation of tBid , we observed a decrease in procaspase-9 levels in reovirus-infected cells . These findings suggest that following reovirus infection of murine fibroblasts , Bid serves to activate the mitochondrial apoptotic pathway . The adaptor molecule FADD is required for cleavage of Bid following reovirus infection of HEK293 cells [16] . Based on these data , we hypothesized that caspase-8 activity as a consequence of extrinsic prodeath signaling , is required for cleavage and activation of Bid . To test this hypothesis , we assessed the capacity of reovirus to mediate Bid cleavage in L929 cells treated with caspase-8 inhibitor Z-IETD-FMK ( Figure 1B ) . As anticipated , Bid cleavage was not observed in mock-infected cells or mock-infected cells treated with Z-IETD-FMK ( data not shown ) . Although tBid was generated at 36–48 h following reovirus infection of vehicle-treated cells , reovirus failed to efficiently induce activation of Bid in Z-IETD-FMK-treated cells until 48 h post-infection , providing evidence that reovirus evokes cleavage of Bid via caspase-8 . In response to a variety of death agonists , Bid amplifies death signaling by linking the extrinsic ( caspase-8 ) and intrinsic ( caspase-9 ) apoptotic pathways [30] . Since our findings with reovirus parallel this pattern , our results suggest that Bid functions similarly following reovirus infection by linking the death-receptor and mitochondrial apoptotic pathways . Signaling via the intrinsic pathway is essential for reovirus-induced apoptosis [16] . This observation , along with the dependence of mitochondrial apoptotic signaling on cleavage of Bid , suggests that Bid serves an essential function in reovirus-induced apoptosis . To directly test whether Bid is required for apoptosis induction following reovirus infection , we compared reovirus-induced apoptosis in wild-type and Bid-deficient MEFs . For these experiments , MEFs were infected with T3D , and apoptosis was assessed by chemiluminescent measurement of the activity of caspase-3 and caspase-7 , which serve as effector caspases for both the extrinsic and intrinsic apoptotic pathways ( Figure 2A ) . In comparison to mock-infected cells , infection of wild-type cells resulted in a significant increase in caspase-3/7 activity at 24 h post-infection . Since MEFs are poorly permissive for reovirus infection [41] , staining of infected cells by indirect immunofluorescence indicated that adsorption with 100 PFU/cell of T3D resulted in infection of only ∼8% of cells at 20 h post infection ( data not shown ) . Despite a low frequency of infection , this MOI resulted in an ∼3-fold increase in caspase-3/7 activity . When infection was initiated at 1000 PFU/cell , ∼20% cells were infected ( data not shown ) , and caspase-3/7 activity increased ∼5-fold . In contrast , infection of Bid-deficient cells resulted in minimal caspase-3/7 activity following infection at either MOI . Increase in caspase-3/7 activity following treatment of each cell type with a broad-spectrum protein kinase inhibitor , staurosporine , was equivalent ( ∼5-fold ) , demonstrating that although Bid-deficient cells possess functional death-signaling pathways , they resist apoptosis induction by reovirus . As an alternative means to quantify apoptosis , we compared wild-type and Bid-deficient MEFs for the onset of morphological characteristics of apoptosis following reovirus infection using an acridine orange ( AO ) staining assay ( Figure 2B ) . Infection of wild-type cells resulted in a significant increase in the fraction of apoptotic cells at 48 h post-infection with 40% and 100% of the cells exhibiting apoptotic features at MOIs of 100 and 1000 PFU per cell , respectively . In contrast , Bid-deficient cells infected with T3D at either MOI displayed levels of apoptosis equivalent to mock-infected cells , ∼10% . Similar results were obtained following infection with another apoptosis-proficient reovirus strain , T3SA+ ( data not shown ) . These data indicate that Bid is required for apoptosis induction following reovirus infection . To determine whether decreased apoptosis in Bid-deficient cells is attributable to alterations in reovirus infection in the absence of Bid , we compared reovirus infectivity in wild-type and Bid-deficient cells using an indirect immunofluorescence staining assay ( Figure 2C ) . An equivalent proportion of reovirus antigen-positive cells was detected at 20 h post-adsorption of wild-type and Bid-deficient cells . These data indicate that reovirus is capable of initiating infection in Bid-deficient cells . To determine whether reovirus completes a full infectious cycle in Bid-deficient cells , wild-type and Bid-deficient cells were adsorbed with T3D , and viral titers were determined by plaque assay at 0 , 12 , 24 , and 48 h after infection ( Figure 2D ) . Reovirus replicated with similar kinetics and produced equivalent yields in wild-type and Bid-deficient cells . Thus , the failure of Bid-deficient cells to undergo apoptosis in response to reovirus is not a consequence of diminished reovirus infection of these cells . We conclude that Bid is a key regulator of reovirus-induced apoptotic cell death . The identification of an essential role for Bid in apoptosis induction following reovirus infection allowed us to examine the relationship between NF-κB activation and Bid cleavage . To determine whether Bid is required for activation of NF-κB following reovirus infection , we compared reovirus-induced NF-κB activation in wild-type and Bid-deficient cells using a reporter assay . Wild-type and Bid-deficient MEFs were transfected with an NF-κB-luciferase reporter plasmid and infected with reovirus . Analogous to treatment with TNFα , a control NF-κB agonist , reovirus infection resulted in equivalent ( ∼2- to 3-fold ) activation of NF-κB-driven gene expression in wild-type and Bid-deficient cells ( Figure 3A ) . These results indicate that Bid is dispensable for NF-κB activation following reovirus infection and suggest that either reovirus-induced NF-κB activation occurs prior to Bid cleavage or that NF-κB activation and Bid cleavage occur in parallel but independent pathways that both function in apoptosis induction by reovirus . To determine whether cleavage-induced Bid activation is dependent on NF-κB , we examined Bid cleavage in cells lacking p65/RelA , an NF-κB subunit required for apoptosis induction following reovirus infection [8] . Infection of wild-type MEFs with reovirus results in generation of tBid at 36–48 h after infection ( Figure 3B ) . In contrast , infection of p65/RelA-deficient MEFs with reovirus did not lead to tBid generation even though efficient viral replication is observed in these cells [8] . Treatment of both wild-type and p65/RelA-deficient MEFs with apoptotic agonists TNFα and cycloheximide resulted in efficient cleavage of Bid , indicating that cell-death pathways leading to Bid cleavage are intact in both cell types . These findings suggest that cleavage and activation of Bid following reovirus infection requires NF-κB and place Bid cleavage subsequent to NF-κB signaling in response to reovirus infection . Moreover , since Bid amplifies death responses from the extrinsic apoptosis pathway by activating the mitochondrial loop , these findings suggest that death-receptor signaling during reovirus infection occurs in an NF-κB-dependent manner . Apoptosis-signaling pathways involving death receptors DR4 and DR5 and death ligand TRAIL , as well as Fas and FasL , have been implicated in apoptosis induction by reovirus [17] , [42] , [43] . However , it is not known which of these pathways mediates cleavage-induced activation of Bid . It is also not understood whether NF-κB regulates the activation of these pathways . Since upregulation of Fas following reovirus infection is dependent on prodeath signaling via c-Jun N terminal kinase ( JNK ) [43] , and because JNK is activated via a mechanism distinct from NF-κB following reovirus infection [44] , we focused our efforts on assessing the regulation and function of death-receptor signaling via TRAIL following reovirus infection . For these studies , we assessed the capacity of reovirus to induce apoptosis in MEFs lacking TRAIL-R , the only known receptor for TRAIL on murine cells [45] , [46] , [47] ( Figure 4 ) . In comparison to mock infection , T3D infection of wild-type cells resulted in an MOI-dependent ∼5- to 20-fold increase in caspase-3/7 activity at 24 h post-infection ( Figure 4A ) . Although T3D infection of TRAIL-R-deficient cells also resulted in an increase in caspase-3/7 activity in comparison to mock-infection , the magnitude of this increase was only ∼2- to 6-fold . Assessment of apoptosis in wild-type and TRAIL-R-deficient MEFs using AO staining also showed an increase in apoptosis both in wild-type and TRAIL-R-deficient cells in comparison to mock-infected cells ( Figure 4B ) . However , a substantially greater fraction of wild-type cells showed morphologic features of apoptosis in comparison to TRAIL-R-deficient cells infected at equivalent MOI , suggesting that efficient induction of apoptosis by reovirus requires TRAIL-R . T3D displayed comparable replication kinetics and produced equivalent yields in wild-type and TRAIL-R-deficient cells ( Figure 4C ) . Thus , differences in the apoptotic potential of reovirus in wild-type and TRAIL-R-deficient cells are not associated with differences in reovirus growth in these cells . To determine whether reovirus-induced cleavage of Bid is dependent on signaling via TRAIL-R , we monitored Bid cleavage following infection of TRAIL-R-deficient cells ( Figure 4D ) . At 48 h post-infection of wild-type cells with T3D , FL Bid was cleaved to generate tBid . In contrast , FL Bid was not cleaved in T3D-infected TRAIL-R-deficient cells . While the apparent difference in the levels of FL Bid in wild-type and TRAIL-R-deficient cells was not reproducible , we consistently observed that levels of FL Bid remained unchanged in TRAIL-R-deficient cells following reovirus infection . These data indicate that TRAIL-R contributes to the induction of apoptosis by reovirus and suggest that cleavage of Bid following reovirus infection is dependent on TRAIL-R signaling . Reovirus virulence correlates with its capacity to cause apoptosis [4] , [7] , [11] , [12] , [48] , [49] . Given the central role of Bid in apoptosis induction by reovirus in cell culture , we hypothesized that reovirus apoptosis and virulence would be diminished in the absence of Bid . To test this hypothesis , we inoculated two-day-old wild-type and Bid-deficient mice perorally with a highly virulent , enteric , neurotropic reovirus strain , T3SA+ [50] , and monitored infected animals for signs of neurological disease and infection-induced morbidity over a period of 21 days ( Figure 5A ) . Following inoculation with 104 PFU of T3SA+ , most wild-type mice developed paralysis and respiratory distress . In contrast , the majority of Bid-deficient mice were asymptomatic . Consistent with this observation , ∼91% of wild-type mice succumbed to reovirus infection with a median survival time of 11 days , whereas only ∼30% of Bid-deficient mice died . Due to the relative resistance of Bid-deficient mice to reovirus-induced encephalitis , a median survival time could not be determined . Thus , the cellular apoptotic regulator Bid modulates reovirus-induced encephalitis . To determine whether the enhanced survival of Bid-deficient mice in comparison to wild-type mice following T3SA+ infection results from reduced reovirus replication , we compared titers of reovirus at sites of primary and secondary replication at 4 , 8 , and 12 d post-inoculation ( Figure 5B–E ) . Peak titers of reovirus were comparable or slightly higher ( ∼5- to 10-fold ) in the intestine , liver , and heart of wild-type mice in comparison to Bid-deficient animals . In contrast , substantially greater differences in peak reovirus titers were observed in the brain , with wild-type animals showing ∼25- to 100-fold higher titers in comparison to those in Bid-deficient mice at 8 d post-inoculation . However , by 12 d post-inoculation , titers of reovirus in wild-type and Bid-deficient mouse brains were equivalent . These findings suggest that reovirus infection is inefficient in the absence of Bid , especially in the CNS . Although titers of reovirus in the CNS were decreased in Bid-null mice following PO inoculation , it was not clear whether reduced reovirus titer in the CNS was a consequence of diminished reovirus dissemination to the CNS or diminished reovirus replication at that site . To distinguish between these possibilities , we inoculated wild-type and Bid-deficient mice intracranially with 100 PFU of T3SA+ and monitored infected animals for signs of CNS disease and mortality for 21 days ( Figure 6A ) . At this dose of T3SA+ , most wild-type and Bid-deficient mice displayed symptoms of neurological disease . Concordantly , both strains of mice succumbed to reovirus-induced disease with equivalent frequency and a median survival time of 13 days . Reovirus titers in the brains of wild-type and Bid-deficient mice also were comparable at 4 , 8 , and 10 d post-inoculation ( Figure 6B ) . These results indicate that following a high-dose inoculation , Bid is dispensable for reovirus growth in the murine CNS and attendant encephalitis . Peak titers of reovirus in the brains of intracranially-inoculated wild-type mice were ∼1000-fold higher than those in perorally-inoculated wild-type animals ( compare Figures 5E and 6B ) . We thought it possible that this difference in viral load might contribute to the dramatic difference in the requirement for Bid in the pathogenesis of reovirus-induced CNS disease following PO and IC inoculation . To test this hypothesis , we inoculated wild-type and Bid-deficient animals intracranially with a considerably lower but still lethal dose of T3SA+ , 5 PFU , and monitored infected animals for signs of reovirus encephalitis ( Figure 6C ) . In comparison to wild-type mice in which ∼95% succumbed to disease , ∼70% of Bid-deficient mice developed lethal encephalitis . Moreover , the median survival time of wild-type mice infected with T3SA+ was significantly less ( 13 days ) than that of Bid-deficient mice ( 15 days ) . To determine whether this difference in survival correlates with the efficiency of reovirus replication in the CNS , we compared titers of reovirus in brains resected from infected mice at 4 , 8 , and 12 d post-inoculation ( Figure 6D ) . Titers of reovirus in brains of wild-type mice were substantially higher ( ∼10- to 100-fold ) at each interval in comparison to those in Bid-deficient mice , with those at 4 and 12 d post-inoculation reaching statistical significance . These findings indicate that at a lower viral inoculum , Bid promotes efficient replication of reovirus in the CNS . Collectively , these data suggest that Bid influences reovirus virulence by regulating the growth of reovirus in the brain . To assess the capacity of T3SA+ to produce neurological injury in the presence and absence of Bid , we examined hematoxylin and eosin ( H&E ) -stained coronal brain sections prepared from wild-type and Bid-deficient mice euthanized 10 d following IC inoculation with 5 PFU of T3SA+ ( Figure 7 ) . This time point was chosen to coincide with the presence of maximal viral titers following inoculation by this route . Since the inoculum used for these experiments was at least ∼10-fold lower than that used for most other studies of reovirus CNS pathogenesis [4] , [7] , [11] , [12] , [43] , [48] , [51] , the extent of injury following infection of wild-type mice was not as extensive . Nonetheless , inoculation of wild-type mice with T3SA+ resulted in neuronal death in the cerebral cortex , hippocampus , thalamus , and hypothalamus , consistent with previous reports [4] , [7] , [11] , [12] , [43] , [48] , [51] . While the majority of infected wild-type mouse brains showed signs of injury , tissue damage was minimal in all of the brains examined from similarly infected Bid-deficient animals . Examination of the hippocampal region of a representative wild-type mouse brain at higher magnification showed damage to the CA3 region , with the pyramidal cells showing condensed nuclei characteristic of apoptosis ( Figure 7B ) . In contrast , little damage was detected in an equivalent region of a Bid-deficient mouse brain ( Figure 7B ) . These findings indicate that Bid is required for neurological injury produced by reovirus in mice . To determine whether these differences in neurological injury are attributable to alterations in tropism of reovirus in the absence of Bid , sections of mouse brain were stained for reovirus antigen . Reovirus displayed similar tissue distribution in wild-type and Bid-deficient mouse brains , indicating that Bid expression does not influence reovirus tropism ( data not shown ) . The CA3 region of a wild-type mouse brain showed reovirus antigen in areas coincident with extensive neuronal damage ( Figure 7B ) . Regions positive for reovirus antigen also stained with an antibody for activated caspase-3 . Although similar regions of Bid-deficient mouse brains contained reovirus antigen , staining was of diminished intensity and frequency ( Figure 7B ) , consistent with the decreased efficiency of reovirus replication in the CNS ( Figure 6B ) . Accordingly , few cells showing intense caspase-3 staining were observed in regions that contained reovirus . These data suggest that neuronal apoptosis following reovirus infection is diminished in the absence of Bid . Thus , Bid links reovirus replication and apoptosis induction in the production of fatal encephalitis .
Early steps in reovirus replication elicit apoptosis via a signaling pathway dependent on NF-κB [4] , [7] , [8] , [14] , [15] . It is not understood how virus-induced NF-κB activation leads to cell death . In this study , we evaluated the function and regulation of Bid in apoptosis caused by reovirus . We found that although Bid is dispensable for reovirus replication in cell culture , it is required for the induction of apoptotic cell death following reovirus infection . In this context , Bid is converted to its active , proapoptotic form , tBid , in an NF-κB-dependent manner . Generation of tBid in reovirus-infected cells requires signaling via TRAIL-R and caspase-8 . These findings indicate that NF-κB signaling following reovirus infection results in activation of the extrinsic apoptotic pathway . In turn , the extrinsic apoptotic pathway evokes the mitochondrial apoptotic cascade via cleavage-induced activation of Bid . Together , these events culminate in the induction of apoptotic cell death . Many viruses induce apoptosis via activation of host-encoded apoptosis-regulating factors . For example , the VSV M protein induces apoptosis by inhibiting the transcription of antiapoptotic factors such as Bcl-xl [52] . In other cases , virus-encoded polypeptides insert into mitochondrial membranes and trigger cytochrome c release , leading to activation of the mitochondrial apoptotic pathway . For example , influenza A virus PB1-F2 is thought to directly activate proapoptotic signaling by interaction with the mitochondrial membrane-associated factors ANT3 and VDAC1 [53] . Although one model for apoptosis induction by reovirus suggests that the φ fragment of reovirus µ1 protein induces apoptosis by directly targeting mitochondria analogous to PB1-F2 [13] , our studies using apoptosis-defective reovirus mutants [4] , [7] , coupled with data presented here , support the idea that φ-mediated NF-κB signaling activates the mitochondrial apoptotic pathway indirectly via death-receptor signaling and Bid cleavage . This indirect mechanism of mitochondrial pathway activation by a proximal signal transducer also explains the timing of prodeath signaling in the reovirus replication cycle . We think that events associated with viral entry into cells , which are mediated by the µ1 φ fragment subsequent to membrane penetration , activate NF-κB within 1 h of infection [7] , [28] . Unlike other NF-κB agonists such as TNFα which rapidly and transiently activate NF-κB , activation of NF-κB following reovirus infection is gradual and sustained and occurs maximally at 6–8 h post infection [28] . Activated NF-κB complexes lead to expression of genes that promote cleavage-induced activation of Bid at 24–36 h post infection and elicit characteristic features of apoptosis , including effector caspase activation and DNA fragmentation . These changes occur subsequent to completion of viral replication , and , therefore , apoptosis appears to have little detectable effect on viral growth in cell culture [4] , [7] , [8] , [28] . Although unusual , NF-κB-dependent apoptotic pathways are also utilized by other viruses such as Dengue virus [54] , HIV [55] , infectious bursal disease virus [56] , and Sindbis virus [57] . Thus , our studies may have uncovered a potentially conserved signaling pathway utilized by viruses to induce apoptosis via NF-κB . It is not known how activation of NF-κB by reovirus culminates in cell death . Three previous studies have attempted to identify proapoptotic host genes that serve as effectors of the death response following reovirus infection . In the first , gene-expression profiles following infection with reovirus strains type 1 Lang ( T1L ) and type 3 Abney ( T3A ) , which differ in the capacity to induce apoptosis [58] , were compared by microarray analysis [59] . These experiments did not demonstrate differences in expression of death ligands or their respective receptors following infection by either strain . Thus , it was concluded that expression of these death mediators by reovirus is unlikely to contribute to apoptosis induction by reovirus . However , some differences were observed in expression of regulators of death-receptor signaling [59] . But since T1L and T3A display significant genetic diversity and vary in the modulation of multiple signaling pathways [9] , [60] , the contribution of NF-κB to the expression of prodeath genes could not be established . In the second study , gene-expression profiles following T3D infection in the presence and absence of functional NF-κB were compared [61] . Although this study identified several NF-κB-dependent genes that coordinate the cellular antiviral immune response , including numerous interferon-stimulated genes ( ISGs ) , no classical components of death receptor-mediated signaling pathways or proapoptotic Bcl-2 family members were significantly upregulated in response to reovirus infection . In the third study , gene-expression profiles of reovirus strains that differ in the capacity to elicit translational shutoff were compared [62] . This study also demonstrated an increase in ISG expression but did not identify obvious NF-κB-dependent candidates that could serve to activate death receptor signaling . We think there are three possibilities to explain why apoptosis-regulating , transcriptional targets of NF-κB , such as death ligands ( e . g . , FasL and TRAIL ) [63] , [64] , death receptors ( e . g . , Fas and DR ) [65] , [66] , and death-signaling regulators ( e . g . , Bax and Bcl-Xs , ) [67] , were not identified in these studies . First , changes in the expression of prodeath genes activated by reovirus infection may be too transient to have been detected in the intervals selected for analysis . Second , the transformed nature of the cell lines used in these studies may not have been amenable to detection of alterations in gene expression induced by reovirus infection . Third , NF-κB activation following reovirus infection may regulate death signaling at a post-transcriptional level by an as yet unknown mechanism . In support of this idea , levels of DR5 protein increase following reovirus infection [17] but not its mRNA [59] . Additional studies using primary , non-transformed cell lines and genetically engineered viruses that differ only in the capacity to activate NF-κB are required to define how reovirus activates extrinsic apoptotic pathways to evoke cell death . In addition to enhancing an understanding of mechanisms by which virus-induced signaling leads to activation of Bid , our studies highlight a critical role for Bid in controlling the pathogenesis of a viral disease . We found that Bid-deficient mice are less susceptible to lethal encephalitis produced by a neurotropic reovirus strain following either PO or IC inoculation . Reovirus replicates with slower kinetics in the absence of Bid , and virus-induced apoptosis and CNS injury are diminished in Bid-deficient animals . Although Bid contributes significantly to reovirus pathogenesis , our data do not allow us to determine whether diminished reovirus virulence in Bid-deficient animals is attributable to reduced capacity of reovirus to replicate in the CNS , diminished capacity of reovirus to injure neurons by apoptosis , or both effects . It is also not clear whether the decreased capacity of reovirus to evoke apoptosis in the CNS is a cause or effect of the lower viral titers at that site . Because Bid serves to amplify the death response , it is not universally required for apoptosis induction . In some cell types , known as type I cells , caspase-8 activation results in direct , Bid-independent activation of the apoptosis effectors , caspase-3 and caspase-7 [29] . In others , known as type II cells , apoptosis requires amplification of death signals through stimulation of the mitochondrial pathway . In these cases , Bid serves to link the extrinsic and intrinsic apoptotic pathways [30] . Since the requirement for Bid in reovirus virulence is dependent on viral dose , we think that the role of Bid as an apoptosis regulator contributes to viral replication and consequent neurovirulence . Thus , we hypothesize that neurons function like type I cells when infected at a higher dose of virus and do not require the amplification of the mitochondrial apoptotic pathway via Bid to undergo apoptotic cell death . However , at lower infectious doses , neurons function like type II cells and require Bid-driven activation of the intrinsic mitochondrial apoptotic cascade to elicit cell death . This model also may explain why primary neuronal cultures infected with reovirus at a high MOI do not appear to require cytochrome c release and caspase-9 activation for apoptosis induction [42] . It is not known how Bid controls the efficiency of reovirus replication in the CNS . One possibility is that Bid-regulated apoptosis is required for efficient release of virus from neurons . Therefore , cell-to-cell spread of reovirus within the CNS may be inefficient in the absence of Bid . As an extension of this idea , blockade of apoptosis by other means also should cause a delay in reovirus replication . However , although symptoms of encephalitis are alleviated in NF-κB p50-deficient mice or in wild-type mice treated with a JNK inhibitor due to a reduction in virus-induced neuronal apoptosis , reovirus replication kinetics are not substantially diminished [11] , [12] . Consistent with these findings , diminished virulence of apoptosis-defective reovirus mutants is not accompanied by significant decreases in reovirus replication efficiency in the CNS [4] , [7] . Apoptosis following reovirus infection can occur in absence of p50 , albeit at low efficiency [8] . Similarly , apoptosis-defective reovirus mutants retain some capacity to induce apoptotic cell death [4] , [7] . Therefore , it is possible that in comparison to reovirus infection of Bid-deficient mice , CNS apoptosis was incompletely blocked in these other studies . Such a difference in the efficiency of apoptosis inhibition could explain the observed discrepancy in the requirement for Bid and other host or viral modulators of apoptosis for efficient replication of reovirus . A second possibility is that a Bid function not related to its capacity to regulate apoptosis contributes to reovirus replication in the CNS . Analogous to its role in reovirus-induced cell death , Bid is implicated in apoptosis caused by many viruses [68] , [69] , [70] , [71] , [72] , [73] , [74] , [75] , [76] , [77] , [78] , [79] . However , prior to our study , it was not known whether Bid modulates the pathogenesis of viral disease . The function of Bid in viral pathogenesis has been examined in a previous study , which found that the BH3-only protein , Puma , but not Bid , contributes to apoptosis-mediated elimination of antigen-specific T cells following acute infection with herpes simplex virus-1 [80] . Here , we demonstrate a pathogenic function for Bid in viral infection . Should Bid similarly modulate disease outcomes following infection by other virulent viruses , antiapoptotic compounds targeting Bid [81] , [82] , [83] may serve as useful antiviral therapeutics .
Murine L929 cells were maintained in Joklik's minimal essential medium supplemented to contain 10% fetal bovine serum ( FBS ) , 2 mM L-glutamine , 100 U/ml penicillin , 100 µg/ml streptomycin , and 25 ng/ml amphotericin B ( Invitrogen ) . Wild-type and Bid-deficient MEFs were maintained in Dulbecco's minimal essential medium ( DMEM ) supplemented to contain 10% FBS , 2 mM L-glutamine , 100 U/ml penicillin , 100 µg/ml streptomycin , and 25 ng/ml amphotericin B . TRAIL-R-deficient MEFs , prepared from D13 embryos , were maintained in DMEM supplemented to contain 10% FBS , 2 mM L-glutamine , 1× MEM nonessential amino acids , 0 . 1 mM 2-mercaptoethanol , 20 mM HEPES , 100 U/ml penicillin , 100 µg/ml streptomycin , and 25 ng/ml amphotericin B . Reovirus strain T3D is a laboratory stock . T3SA+ was generated by reassortment of reovirus strains T1L and type 3 clone 44-MA as described [84] . Purified reovirus virions were generated from second- or third-passage L-cell lysate stocks of twice-plaque-purified reovirus [85] . Viral particles were Freon-extracted from infected cell lysates , layered onto 1 . 2- to 1 . 4-g/cm3 CsCl gradients , and centrifuged at 62 , 000×g for 18 h . Bands corresponding to virions ( 1 . 36 g/cm3 ) were collected and dialyzed in virion-storage buffer ( 150 mM NaCl , 15 mM MgCl2 , 10 mM Tris-HCl [pH 7 . 4] ) [86] . Rabbit antisera raised against T1L and T3D have been described [87] . Rabbit antiserum specific for procaspase-9 was purchased from Cell Signaling . Goat antiserum specific for Bid was purchased from R & D systems , and goat antiserum specific for actin was purchased from Santa Cruz Biotechnology . HRP-conjugated anti-rabbit and anti-goat secondary antibodies were purchased from Amersham GE Biosciences . Alexa Fluor-conjugated anti-mouse immunoglobulin ( Ig ) G , anti-rabbit IgG , and anti-goat IgG secondary antibodies were purchased from Invitrogen . Plasmids pRenilla-Luc and pNF-κB-Luc [88] were obtained from Dr . Dean Ballard ( Vanderbilt University ) . L929 cells or wild-type , Bid-deficient , or TRAIL-R-deficient MEFs were either adsorbed with reovirus at an MOI of 100 PFU/cell or mock-infected in serum-free medium at 4°C for 1 h , followed by incubation in serum-containing medium at 37°C for various intervals . Whole cell lysates were prepared by washing cells in phosphate-buffered saline ( PBS ) followed by lysis using 1× RIPA buffer ( 50 mM Tris [pH 7 . 5] , 50 mM NaCl , 1% TX-100 , 1% DOC , 0 . 1% SDS , and 1 mM EDTA ) containing a protease inhibitor cocktail ( Roche ) . Following centrifugation at 15 , 000×g to remove debris , the lysates were resolved by electrophoresis in polyacrylamide gels and transferred to nitrocellulose membranes . Membranes were blocked for at least 1 h in blocking buffer ( PBS containing 5% milk or 2 . 5% BSA ) and incubated with antisera against Bid ( 1∶1000 ) , actin ( 1∶2000 ) , or procaspase-9 ( 1∶1000 ) either at room temperature for 1 h or 4°C overnight . Membranes were washed three times for 10 min each with washing buffer ( PBS containing 0 . 1% Tween-20 ) and incubated with1∶2000 dilution of horseradish peroxidase ( HRP ) -conjugated or Alexa Fluor-conjugated goat anti-rabbit Ig ( for procaspase-9 ) or donkey anti-goat Ig ( for Bid and actin ) in blocking buffer . Following three washes , membranes were incubated for 1 min with chemiluminescent peroxidase substrate ( Amersham Biosciences ) and either exposed to film ( for HRP-conjugated secondary antibodies ) or scanned using an Odyssey Infrared Imager ( LiCor ) . Wild-type , Bid-deficient , or TRAIL-R-deficient MEFs ( 104 ) were seeded into black clear-bottom 96-well plates ( Costar ) and adsorbed with reovirus in serum-free medium at room temperature for 1 h . Following incubation of cells at 37°C for 24 h , caspase-3/7 activity was quantified using the Caspase-Glo-3/7 assay ( Promega ) . Wild-type , Bid-deficient , or TRAIL-R-deficient MEFs ( 5×104 ) were grown in 24-well plates ( Costar ) and adsorbed with reovirus at room temperature for 1 h . The percentage of apoptotic cells after 48 h incubation was determined using AO staining as described [6] . For each experiment , >200 cells were counted , and the percentage of cells exhibiting condensed chromatin was determined by epi-illumination fluorescence microscopy using a fluorescein filter set ( Zeiss Photomicroscope III; Thornwood , NY ) . Wild-type or Bid-deficient cells ( 2×105 ) were grown in 24-well plates and adsorbed with reovirus at room temperature for 1 h . Following removal of the inoculum , cells were washed with PBS and incubated in complete medium at 37°C for 18 h . Monolayers were fixed with methanol , washed twice with PBS , blocked with 2 . 5% Ig-free bovine serum albumin ( Sigma-Aldrich ) in PBS , and incubated successively for 1 h with polyclonal rabbit anti-reovirus serum at a 1∶1000 dilution and for 1 h with Alexa Fluor 546-labeled anti-rabbit IgG at a 1∶1000 dilution . Monolayers were washed with PBS , and infected cells were visualized by indirect immunofluorescence using a Zeiss Axiovert 200 fluorescence microscope . Reovirus antigen-positive cells were quantified by counting fluorescent cells in at least two random fields of view in triplicate wells at a magnification of 20× . Wild-type , Bid-deficient , or TRAIL-R-deficient MEFs ( 2×105 ) in 24-well plates were adsorbed with reovirus at room temperature for 1 h in serum-free medium , washed once with PBS , and incubated in serum-containing medium for various intervals . Cells were frozen and thawed twice prior to determination of viral titer by plaque assay using L929 cells [89] . Wild-type and Bid-deficient cells in 24-well plates were transfected with 0 . 72 µg/well of an NF-κB reporter plasmid , which expresses firefly luciferase under NF-κB control ( pNF-κB-Luc ) , and 0 . 08 µg/well of control plasmid pRenilla-Luc , which expresses Renilla luciferase constitutively , using Fugene6 ( Roche ) . After incubation for 24 h , transfected cells were adsorbed with reovirus in serum-free medium at room temperature for 1 h and incubated at 37°C in serum-containing medium for 24 h . Luciferase activity in the cultures was quantified using the Dual-Luciferase Assay Kit ( Promega ) according to the manufacturer's instructions . Wild-type C57BL/6J mice were obtained from Jackson Laboratory . Bid-deficient mice backcrossed on to a C57BL/6J background for at least 8 generations have been previously described [30] . Two-day-old mice were inoculated either perorally or intracranially with purified virus diluted in PBS . PO inoculations were delivered in a volume of 50 µl by passage of a polyethylene catheter 0 . 61 mm in diameter ( BD ) through the esophagus and into the stomach [90] . The inoculum contained 0 . 3% ( vol/vol ) green food coloring to allow the accuracy of delivery to be judged . IC inoculations were delivered to the left cerebral hemisphere in a volume of 5 µl using a Hamilton syringe and a 30-gauge needle ( BD Biosciences ) [91] . For analysis of viral virulence , mice were monitored for weight loss and symptoms of disease for 21 days . For survival experiments , mice were euthanized when found to be moribund ( defined by rapid or shallow breathing , lethargy , or paralysis ) . For determination of viral titer and immunohistochemical staining , mice were euthanized at various intervals following inoculation and organs were resected . For analysis of virus growth , organs were collected into 1 ml of PBS and homogenized by freezing , thawing , and sonication . Viral titers in organ homogenates were determined by plaque assay using L929 cells [89] . For immunohistochemical staining , organs were fixed overnight in 10% formalin , followed by incubation in 70% ethanol . Fixed organs were embedded in paraffin , and 5-µm histological sections were prepared . Consecutive sections were stained with H&E for evaluation of histopathologic changes or processed for immunohistochemical detection of reovirus antigens or activated caspase-3 [11] . Animal husbandry and experimental procedures were performed in accordance with Public Health Service policy and the recommendations of the Association for Assessment and Accreditation of Laboratory Animal Care and approved by the Vanderbilt University School of Medicine Institutional Animal Care and Use Committee . | Viruses injure host tissues by activating signaling pathways that trigger cell death by a process called apoptosis . Hence , blockade of apoptosis may serve as a useful strategy to dampen the severity of viral disease . However , deployment of such a strategy requires identification of host signaling networks that control cell death and a detailed molecular blueprint of how these pathways are activated by a virus . In this study , we used mammalian reovirus , an important experimental model for studies of viral encephalitis , to elucidate how cell death pathways are activated following viral infection and whether these signaling cascades influence the capacity of a virus to produce lethal CNS disease . We found that Bid , a host regulator of cell death , influences apoptosis induction by reovirus . Moreover , Bid is required for efficient reovirus replication in the CNS and modulates reovirus neurological disease . These findings highlight Bid as a critical regulator of viral pathogenesis and illuminate a potential new target for development of antiviral therapeutics . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"virology/virulence",
"factors",
"and",
"mechanisms"
] | 2010 | Bid Regulates the Pathogenesis of Neurotropic Reovirus |
Ethylene is one of the most important hormones for plant developmental processes and stress responses . However , the phosphorylation regulation in the ethylene signaling pathway is largely unknown . Here we report the phosphorylation of cap binding protein 20 ( CBP20 ) at Ser245 is regulated by ethylene , and the phosphorylation is involved in root growth . The constitutive phosphorylation mimic form of CBP20 ( CBP20S245E or CBP20S245D ) , while not the constitutive de-phosphorylation form of CBP20 ( CBP20S245A ) is able to rescue the root ethylene responsive phenotype of cbp20 . By genome wide study with ethylene regulated gene expression and microRNA ( miRNA ) expression in the roots and shoots of both Col-0 and cbp20 , we found miR319b is up regulated in roots while not in shoots , and its target MYB33 is specifically down regulated in roots with ethylene treatment . We described both the phenotypic and molecular consequences of transgenic over-expression of miR319b . Increased levels of miR319b ( miR319bOE ) leads to enhanced ethylene responsive root phenotype and reduction of MYB33 transcription level in roots; over expression of MYB33 , which carrying mutated miR319b target site ( mMYB33 ) in miR319bOE is able to recover both the root phenotype and the expression level of MYB33 . Taken together , we proposed that ethylene regulated phosphorylation of CBP20 is involved in the root growth and one pathway is through the regulation of miR319b and its target MYB33 in roots .
The plant hormone ethylene ( C2H4 ) is essential for a myriad of physiological and developmental processes [1–3] . A linear ethylene signaling pathway has been established [4] that plants perceive ethylene by ER-located receptors , which are similar to the bacterial two component histidine kinases [5 , 6] . In the absence of ethylene , the receptors activate a Raf-like protein kinase CONSTITUTIVE TRIPLE RESPONSE 1 ( CTR1 ) [7] . Activated CTR1 inhibit an ER-tethered protein ETHYLENE INSENSITIVE 2 ( EIN2 ) through phosphorylation [8 , 9] . EIN2 is degraded and the degradation is mediated by two F-box proteins: ETP1 and ETP2 [10] . In the presence of ethylene , the EIN2 C-terminal ( EIN2-C ) is dephosphorylated , cleaved and translocated into both the nucleus and P-body [9 , 11 , 12] . In the nucleus , the EIN2 CEND transduces signals to the transcription factors ETHYLENE INSENSTIVE3 ( EIN3 ) and ETHYLENE INSENSITIVE3-LIKE1 ( EIL1 ) , which are sufficient and necessary for activation of all ethylene-response genes [13 , 14] . In P-body , EIN2C mediates translation repression of EBF1 and EBF2 [15 , 16] , which are the two F-box proteins , which target EIN3 for degradation [17 , 18] . Recently new study discovered that noncanonical histone acetylation H3K23Ac is involved in ethylene regulated gene activation in an EIN2 and partial EIN3 dependent manner [19] . Protein phosphorylation plays critical roles in ethylene response . Such as ethylene receptors are similar in sequence and structure to bacterial two-component histidine kinases , and ethylene controls autophosphorylation of the histidine kinase domain in ethylene receptor ETR1 [20] , the histidine kinase activity of ETR1 is not required for but plays a modulating role in the regulation of ethylene responses [21] . Furthermore , biochemical and functional analysis of CTR1 , a protein similar to the Raf family protein kinases that negatively regulates ethylene signaling in Arabidopsis [22] . Recently study has demonstrated that in the absence of ethylene , CTR1 targets to EIN2 C-terminal end for phosphorylation with its kinase domain [9] . However , the phosphorylation regulation in ethylene signaling is still under developed . In this study , we found that ethylene induces phosphorylation of CBP20 at Ser245 . Constitutive phosphorylation mimic form of CBP20S245D or CBP20S245E rescues the root less sensitive phenotype of cbp20 mutant in ethylene , but the constitutive dephosphorylation mimic form of CBP20S245A is unable to rescue cbp20 mutant phenotype . Through small RNA sequencing and mRNA sequencing , we found a set of miRNAs and their targets are specifically regulated in roots by ethylene in a CBP20 dependent manner . Among them , the expressions of miR319b and its potential target MYB33 display anti-correlation pattern in a CBP20 dependent manner in Col-0 roots . Small RNA northern blot in roots shows that miR319 is specific up regulated by ethylene treatment , which requires CBP20 phosphorylation . Genetic study shows that over expression of miR319b leads to the down regulation of MYB33 , resulting in enhanced ethylene sensitive phenotype in roots , which is similar to its target myb33 mutant phenotype . The phenotype of miR319bOE is rescued by adding mutated MYB33 ( mMYB33 ) , which containing mutation at miR319 targeting site . Furthermore , we provided evidence showing that miR319b , while not miR159 influences the expression of MYB33 in the presence of ethylene in roots . Overall , our results demonstrate that ethylene regulated phosphorylation of CBP20 is involved in the root growth . One model is through the regulation of miR319b and its target MYB33 in roots in response to ethylene , providing a new link of cap binding protein phosphorylation associated microRNA to root growth in the ethylene response .
Previous studies have shown that phosphorylation plays critical roles in ethylene signaling and many ethylene regulated phosphorylation proteins have been identified [11 , 23] . By searching the phosphorylation MS/LS data under ethylene treatment , we found that the CBP20 , a component of cap-binding complex , is highly phosphorylated at Ser245 site with ethylene treatment ( Fig 1A and 1B and S1A Fig ) . Protein alignment with CBP20s from different species showed that CBP20 is highly conserved , and the Ser245s are all identical through different species examined ( S1B Fig ) , suggesting the function of CBP20 is conserved and the phosphorylation at S245 of CBP20 is potentially important for CBP20 . To study the role of phosphorylation of CBP20 in the regulation of ethylene response , the ethylene response phenotype of cbp20 mutant was examined . We found cbp20 mutant displayed partial reduced ethylene responsive phenotype in the roots , but not in the hypocotyls and apical hooks ( Fig 1C–1E ) . To examine the connection between phosphorylation state of CBP20 and the ethylene responsive phenotype of cbp20 , 35S promoter driven phosphorylation mimic form of CBP20 ( 35S:CBP20S245D and 35S:CBP20S245E ) and dephosphorylation mimic form of CBP20 ( 35S:CBP20S245A ) ( S2A Fig ) , were generated and introduced into cbp20 mutant to obtain CBP20S245A/cbp20 , CBP20S245D/cbp20 and CBP20S245E/cbp20 . Additionally , 35S promoter driven wild type CBP20 ( 35S:CBP20 ) was introduced to cbp20 as control ( S2A Fig ) . The full length of CBP20 , CBP20S245E and CBP20S245D were able to rescue the cbp20 mutant phenotype in the presence of ethylene . However , CBP20S245A was unable to rescue the phenotype ( Fig 1C–1E ) , which suggests that the phosphorylation of CBP20 is involved in the root growth in the presence of ethylene . To explore how the phosphorylation of CBP20 is involved in the regulation in ethylene response , we first examined the root phenotype of cbp80 mutant in response to ethylene . We found that cbp80 displays similar phenotype as cbp20 in the presence of ethylene ( S2B Fig ) . We next tested the interaction between wild type CBP80 with CBP20 , CBP20S245A , CBP20S245D or CBP20S245E by yeast two-hybrid . In consistent with previous study [24] , we were able to detect the interaction between CBP20 and CBP80 , however , the interaction was not influenced by the phosphorylation states of CBP20 ( S2C Fig ) . Generally , CBP80 interacts with CBP20 and in assisting CBP20 transfer into nucleus [25] , we then examined cellular localization of CBP20S245A-YFP and CBP20S245D-YFP or CBP20S245E-YFP into cbp20 with or without the presence of ethylene . Both wild type CBP20 and mutated CBP20 were mainly localized in the nucleus and their localizations were not altered by ethylene treatment ( S2D Fig ) . CBC complex has a key role in several gene expression mechanisms [26–28] and CBP20 is essential for miRNA biogenesis [29–31] . We speculated that CBP20 is required for the biogenesis of miRNAs in the root growth in ethylene response . To address this question , we conducted small RNA sequencing using the roots and shoots isolated from 3-day old etiolated seedlings of Col-0 or cbp20 mutant treated with air or ethylene ( S3A and S3B Fig ) . In consistent with previous study [30] , most of species of miRNA detected were down regulated in cbp20 ( S1–S4 Tables ) . By comparing the miRNA expressions in the roots and shoots of Col-0 and cbp20 treated with air or 4 hours ethylene gas . We found that ethylene altered miRNA expressions in a tissue specific manner ( Fig 2A ) . As shown in Fig 2A and 2B , almost no shared ethylene induced differential expressed miRNAs were detected in the shoots and roots in Col-0 or cbp20 mutant . Given ethylene regulated miRNA expression is tissue specific , and cbp20 root specific phenotype with ethylene treatment , we speculated that the ethylene responsive root phenotype of cbp20 potentially due to the alteration of miRNAs in roots . Through further analysis , 13 ethylene regulated miRNAs ( P<0 . 05 ) were identified specifically in Col-0 roots ( Fig 2A ) . Among them , 7 miRNAs were up regulated and 5 were down regulated , and most of them are uncharacterized miRNAs ( Fig 2B and S1 Table ) . Overall these results demonstrate that ethylene alters miRNA expression in a tissue specific manner , and we are able to identify ethylene regulated miRNAs specifically in roots in a CBP20 depend manner . In plants , the main function of miRNAs appears to be in gene regulation . Therefore , we expected the expression of the potential targets of 12 miRNAs identified above is anti-correlated with their miRNAs in response to ethylene in Col-0 roots . We conducted RNA sequencing use the same tissues as mentioned in small RNA seq with two biology duplications ( S4A and S4B Fig ) . Comparable numbers of ethylene-regulated genes were detected in the roots and shoots of Col-0; however , only about 20–30% of genes were overlapped between these two type tissues ( Fig 2C and S1–S4 Datasets ) , showing the tissue specific in ethylene response . Further GO analysis showed that ethylene related GO terms were enriched in those ethylene regulated genes shared between shoots and roots , and root development related GO terms were enriched in the genes specifically regulated in roots ( S5A–S5C Fig ) . Similarly , in cbp20 mutant the gene regulation also showed tissue specificity in response to ethylene ( Fig 2C and S1–S4 Datasets ) . We then compared ethylene regulated gene expression in the roots of Col-0 and cbp20 . As shown in Fig 2D , about 60% of up regulated genes and 75% of down regulated genes in Col-0 roots were not altered in cbp20 roots in the presence of ethylene ( S5 and S6 Datasets ) , showing CBP20 dependency . We then studied the association between 12 microRNAs and their targets genes in ethylene response . In total 841 potential target genes were identified and 203 with high confidence ( T score < = 5 ) ( S7 Fig and S8 Dataset ) . Among them , only 8 target genes were differentially regulated by ethylene in the roots of Col-0 , while their differential expressions were impaired in cbp20 ( Fig 2E ) . By comparing the expressions of ethylene altered miRNAs and their target genes in Col-0 roots , two miRNAs ( miR319b , miR863-3p ) were identified that up regulated by ethylene and the expression of their potential target genes was down regulated in the presence of ethylene in the roots of Col-0 , while not in the roots of cbp20 ( Fig 2F ) . To validate the function of miRNAs identified above in response to ethylene , and the connection between their expressions with the phosphorylation state of CBP20 , we examined mature miR319 by northern blot in cbp20 mutant , CBP20/cbp20 , CBP20S245A/cbp20 and CBP20S245E/cbp20 with or without ethylene treatment . In consistent with small RNA sequencing result , miR319 indeed was up regulated by ethylene in the roots of Col-0 ( Fig 3A ) , and the elevation was impaired in the roots of cbp20 mutant . Furthermore , constitutive phosphorylated CBP20S245E , while not dephosphorylated CBP20S245A were able to recover the ethylene induced elevation of miR319 expression ( Fig 3A ) , indicating that ethylene induced phosphorylation of CBP20 potentially required for the elevation of miR319b expression in roots . We next examined the expression of pri-miR319b in the roots of Col-0 , cbp20 , CBP20/cbp20 , CBP20S245A/cbp20 , CBP20S245D/cbp20 and CBP20S245E/cbp20 treated with air or 4 hours ethylene gas by quantitative RT-PCR . As shown in Fig 3B , the expression of pri-miR319b was decreased with the ethylene treatment and the down regulation was impaired in cbp20 mutant . The down regulation of pri-miR319b was detected in the roots of CBP20S245D/cbp20 or CBP20S245E/cbp20 , while not in that of CBP20S245A/cbp20 ( Fig 3B ) , indicating that the phosphorylation is required for the down regulation of pri-miRNA , further suggesting that the elevation of miR319b in response to ethylene due to the biogenesis of miRNA , while not due to the elevation of pri-miR319b . To further examine how the phosphorylation of CBP20 influences the gene expression of MYB33 in response to ethylene , we conducted qRT-PCR in the roots of Col-0 , CBP20/cbp20 , CBP20S245A/cbp20 , CBP20S245D/cbp20 and CBP20S245E/cbp20 with or without ethylene treatment . The expression level of MYB33 was indeed decreased by ethylene treatment , which is consistent with RNA-seq result ( Fig 3C ) . In cbp20 and CBP20S245A/cbp20 plants , the down regulation of MYB33 was impaired , however , in CBP20S245D/cbp20 or CBP20S245E/cbp20 ( Fig 3C ) , the expression of MYB33 is recovered as that of in Col-0 . Overall , the result shows that the expression of MYB33 is anti-correlated with the expression of miR319b specifically in roots in ethylene response , indicating that ethylene induced phosphorylation of CBP20 inhibits the expression of MYB33 , which potentially through CBP20 regulated biogenesis of miR319b in roots . To further examine whether miR319b plays a role in root growth in ethylene response , we generated the miR319b overexpression ( miR319bOE ) plants , and examined their phenotype in response to ethylene . As expected , the roots of miR319bOE plants were more sensitive to ethylene than that of wild type ( Fig 4A–4C ) . MYB33 is one of potential targets of miR319b , we therefore obtained myb33 mutant to examine its phenotype in response to ethylene . As expected , the roots of myb33 mutant displayed similar phenotype as that of miR319bOE in the presence of ethylene ( Fig 4A–4C and S6A and S6B Fig ) . Comparing to wild type , the pri-miR319b was increased ( Fig 4D ) , while the expression of MYB33 was decreased in miR319bOE plants ( Fig 4E ) , showing that the elevation of miR319b is not due to the elevation of its precursor in the presence of ethylene , but potentially due to the miRNA biogenesis process in response to ethylene . We then conducted a 5′ RNA Ligase-Mediated ( RLM ) -Rapid Amplification of cDNA ends ( RACE ) in both Col-0 and miR319OE plants to evaluate that MYB33 is one of targets of mir319b in vivo , In Col-0 , no cleavage event was detected between the 10th nucleotide U and the 11nd nucleotide C from the 5′ end of the miRNA in Col-0 . However , in miR31bOE , 5 out of 15 cleavage events were detected between nucleotides 10 and 11 from the 5′ end of the miRNA ( S6C Fig ) , which are in consistent with the published data [32] . Taken all together , these results indicate that miR319b is involved in root growth by targeting MYB33 for degradation in a CBP20 dependent manner . Because MYB33 is a shared target between miR319 and miR159 , we examined the expression of miR159 in the roots of Col-0 and cbp20 treated with air and ethylene by northern blot . Inconsistent with small RNA-seq result , no differential expression of miR159 was detected in Col-0 roots between air and ethylene treatments ( Fig 5A ) . As previous published data [30] , we detected the reduction of miR159 in cbp20 comparing to that of in Col-0 , which is consistent with published data [30] . However , no ethylene induced alteration for miR159 was detected in the roots of both Col-0 or cbp20 mutant ( Fig 5A ) . In addition , no significant difference of pri-miR159 was detected in both Col-0 and cbp20 roots by ethylene treatment as well ( Fig 5B ) . Furthermore , the phosphorylation mimic forms of CBP20S245D and CBP20S245E , while not dephosphorylation mimic form of CBP20S245A behaved as wild type CBP20 ( Fig 5B ) . We further examined the pri-miR159a in two independent miR319bOE plants and found pri-miR159a was not affected by the overexpression of miR319b ( Fig 5C ) , indicating that in the presence of ethylene , the down regulation of MYB33 is associated with the up regulation of miR319b , while not miR159 . To evaluate whether the phenotype of miR319bOE is caused by down regulation of MYB33 , we constructed mutated MYB33 ( mMYB33 ) carrying mutation in miR319 targeting site ( S7A Fig ) and introduced it into miR319bOE plants to obtain miR319bOE/mMYB33OE . The ethylene responsive phenotypes of both the roots and shoots of miR31bOE were recovered in miR319bOE/mMYB33OE plants ( Fig 6A–6C ) . We then examined the expression of MYB33 in the roots of miR319bOE and miR319bOE/mMYB33OE . In consistent with Fig 5E , the expression of MYB33 was down regulated in miR319bOE plants , and was highly up regulated in miR319bOE/mMYB33OE in comparing to that of in Col-0 ( Fig 6D ) , while the expression of pri-miR319b in miR319bOE and in miR319bOE/mMYB33OE plants were comparable ( S7B Fig ) . TCPs are known targets of miR319 , however , no ethylene induced alteration of TCPs was detected in our RNA-seq result , to further confirm the result , we examined the gene expression of TCPs by qRT-PCR in Col-0 treated with or without ethylene . No significant change was detected for those gene expressions in response to ethylene ( S7C Fig ) . In addition , gene expression of TCP24 was not altered in miR319bOE/mMYB33 in comparing to that of in miR319bOE plants ( S7D Fig ) , and the expression of other TCPs displayed similar patterns as TCP24 both in Col-0 and miR319OE plants ( S7E Fig ) . Finally we conducted Agrobacterium-mediated transient co-expression assay with MYB33 or mutated mMYB33 CDS fused to 35S::LUC 3’UTR with or without the miR319b precursor . Comparing to the assay without miR319b precursor , MYB33 expression was significantly lower in the presence of miR319b precursor ( Fig 7A and 7B ) . However , no significant change was detected for mMYB33 expression between with and without the presence of miR319b precursor ( Fig 7A and 7B ) . The similar assay was also conducted using YFP-HA-tagged MYB33 or mutated MYB33 ( mMYB33 ) with or without miR319b precursor to examine how miR319b influence gene expression of MYB33 and its protein level . The gene expression of MYB33 , while not mMYB33 , is down regulated in the presence of miR319b ( S8A Fig ) . In consistent with gene expression , MYB33 protein is also lower in the assay with the presence of miR319b . However , mMYB33 protein was not altered by miR319b ( S8B Fig ) . Taken all together our data support that miR319b targets to MYB33 for degradation in roots .
It has been well known that protein phosphorylation are involved in many different plant hormones such as phosphorylation regulates the polarity of PIN in auxin [33–37] , gibberellins [33 , 38 , 39] , cytokinin [40] , ABA [41] and in BR signaling [42–44] . Many studies have demonstrated that MPKKK cascade promotes ACS6 and EIN3 phosphorylation [45 , 46] . Recently study has demonstrated that in the absence of ethylene , the receptors activate CTR1 , which phosphorylates EIN2 C-terminus [9] . With the presence of ethylene , the EIN2C is dephosphorylated and then cleaved and translocated into nucleus to activate the downstream signaling pathway [11 , 12] . However , the phosphorylation regulation in ethylene signaling is still largely unknown . Genome wide phospho-peptide survey in 3-day old etiolated seedlings treated with air or ethylene was done previously [11] , we found the phosphorylation of CBP20 is highly regulated by ethylene gas ( Fig 1 ) . In the absence of ethylene , no phosphorylated peptides of CBP20 were detected , while in the presence of ethylene gas , 17 spectrum counts of phosphorylation peptide ( 239aa-253aa ) was detected . Further genetics study demonstrated that the phosphorylation of CBP20 is involved in the growth of root in the presence of ethylene ( Fig 1C–1E ) . CBP20 is a subunit of CBC complex , which is vital for plant development . Previous study has demonstrated that CBP80 , the other subunit of CBC is involved in the regulation of hypocotyl in ethylene signaling through regulation the biogenesis of small RNAs [47] . In mammalians , it has demonstrated that growth factors mTORC1 kinase regulated S6 kinases able to phosphorylates CBP80 , activating the CBC affinity for 7mG [48 , 49] . However , no evidence has shown CBC complex is regulated by phosphorylation in response to hormones . Here , for the first time we provide evidence showing that CBP20 Ser245 site is highly phosphorylated with ethylene treatment . The constitutive phosphorylated CBP20 , while not the constitutive non-phosphorylated CBP20 is able to rescue the ethylene root phenotype of cbp20 , strongly suggesting that the phosphorylation of CBP20 is involved in ethylene response . Yet , how CBP20 phosphorylation occurs in the presence of ethylene is still undetermined . In our precious study of phosphopeptides in etr1-1 mutant with or without ethylene treatment , no phosphorylation was detected for CBP20 , showing that the phosphorylation of CBP20 is ethylene dependent . Therefore , the identification of kinases that regulate CBP20 phosphorylation specifically in the presence of ethylene will be an immediately interest . CBC complex regulate many aspects of biological processes including transcription regulation , pre-mRNA splicing , pre-mRNA 3’end processing , miRNA biogenesis , mRNA stability , mRNA and snRNA nuclear export , the pioneer round of translation and nonsense-mediated RNA decay [28] . However , no evidence has shown that CBP20 is involved in ethylene response . In our study , through high throughput sequencing for small RNAs and mRNAs in different plants treated with or without ethylene , we identified ethylene regulated miRNAs . In addition , we found that CBP20 regulates many species of miRNA expressions in response to ethylene with a tissue specific manner ( Fig 2 ) . miRNAs are involved in many different aspects of plants . Specifically in plant hormones , such as miR160 targets to several ARF family members to activate auxin signaling pathway for root cap formation [50]; miR159 targets to MYB33 to activate ABA signaling pathway for seed germination [30] . In ethylene signaling pathway , it have been reported that EIN3 represses miR164 transcription and up regulates the transcript level of NAC2 to regulate leaf senescence [51] . Here we provide evidence showing for the first time that miRNAs are differentially regulated by ethylene in a tissue specific manner ( Fig 2A and 2B ) , and many of the differential regulations are abolished in cbp20 mutant ( Fig 2A and 2B ) . Previous studies have shown that CBP20 is required for the biogenesis of many miRNAs [30] . Interestingly , our data showed that some miRNA species are down regulated in cbp20 mutant , which indicating non-CBP20 dependent miRNA biogenesis is potentially involved in ethylene response . By comparing the miRNAs in Col-0 and in that of cbp20 , we found many miRNA species are up regulated in cbp20 in the presence of ethylene . One possibility is that the precursors of those miRNAs are elevated by ethylene , resulting in the elevation of their miRNAs . Alternatively , CBP20 independent miRNA biogenesis machinery is elevated in the presence of ethylene , resulting in the increase of the miRNAs . However , recently study has shown that small RNA biogenesis machinery component Dicers are not involved in ethylene response [15] . Therefore , further comprehensive studies will be critical to characterize the newly identified ethylene regulated , while not CBP20 dependent miRNAs and uncover the mechanistic details that how biogenesis occurs specifically in the presence of ethylene . MicroRNAs in plant are small RNAs , which are approximately 21 nucleotides in length . Normally , they are negative regulators of gene expression through base pairing to the complementary sequence within the target mRNAs , leading to the target mRNA degradation through RISC-mediated cleavage . In comparing to the ethylene altered small RNAs with ethylene regulated genes , we found that miR319b was up-regulated while MYB33 was down-regulated in Col-0 roots with ethylene treatment , and the regulation is CBP20 dependent . Small RNA northern blot shows that miR319 is indeed up-regulated under ethylene treatment and the regulation is dependent on CBP20 phosphorylation . However , it is well known that MYB33 is a shared target between miR159a and miR319b . The miR159 and miR319 families are similar in sequence , but they have distinct target genes: miR159 is specific for MYB transcription factors , mainly MYB33 and MYB65 . In contrast , miR319 mainly targets TCP transcription factors , predominantly TCP2 and TCP4 . MiR319 also targets MYB33 and MYB65 , but due to its low abundance , this regulation is negligible . However , in our study we provided multiple lines of evidence showing that in the ethylene response , miR319b targets MYB33 for degradation specifically in roots , leading to the ethylene regulated cbp20 root phenotype: ( 1 ) , miR319 was specifically up regulated by ethylene in Col-0 roots , while not in cbp20 mutant ( Fig 4A ) . However , miR159 was not regulated by ethylene ( Fig 5A ) ; ( 2 ) MYB33 was down regulated in Col-0 roots in response to ethylene , while not in cbp20 roots ( Fig 4C ) . ( 3 ) The pri-miR319b was down regulated in Col-0 roots , while not in cbp20 roots ( Fig 4B ) ; ( 4 ) The pri-miR159a was not regulated by ethylene in Col-0 ( Fig 5B ) ; ( 5 ) In the over expression miR319b plants , MYB33 was largely down regulated , while miR159 and pri-miR159a were not altered ( Figs 4C , 5B and 5C ) ; ( 6 ) Overexpression of the mMYB33 containing mutated miR319b target site is able to recover phenotype caused by miR319bOE ( Fig 6 ) ; ( 7 ) Our data ( Fig 7 ) and published data has shown that miR319b targets MYB33 for cleavage [32] . In summary , our study discovered that ethylene regulates the phosphorylation of CBP20 , and the phosphorylation is required for the elevation of miR319 , which leading to the down regulation of MYB33 expression in roots , resulting in root growth inhibition in the presence of ethylene .
All mutants were in the Columbia-0 ( Col-0 ) background , cbp80 ( CS878659 ) , myb33-1 ( SALK_065473 ) , myb33-2 ( SALKseq_056201 ) are ordered from ABRC . cbp20 has been described in [52] . Seeds were sterilized with 4% bleach and then washed as least three times with sterilized water , then the seeds were sown on MS medium . Plants were grown in long days ( 16h light/8h dark ) at 22°C on soil . To construct CBP20 overexpression vectors for complementing cbp20 mutant phenotype , the CBP20 full length and phosphorylation site mutated CDS sequences of CBP20 were amplified using the Phusion High-Fidelity DNA Polymerase ( NEB ) . The PCR products were cut with KpnI and SalI , and then the corresponding fragments were ligated into the KpnI-SalI site of the pCHF3 vector to give rise to 35S:CBP20-YFP , 35S:CBP20S245A-YFP , 35S:CBP20S245E-YFP and 35S:CBP20S245D-YFP . To construct vectors for yeast two-hybrid , the CDS of CBP80 was amplified using the Phusion High-Fidelity DNA Polymerase . The PCR product was cut with SalI and XbaI , and then the corresponding fragment was ligated into the SalI-SpeI site of the pDBLeu vector ( Invitrogen ) to give rise to pBD-CBP80 . The CDSs of CBP20 , CBP20S245A , CBP20S245E and CBP20S245D were amplified using the High-Fidelity DNA Polymerase . The PCR products were digested by SalI and SpeI , and the corresponding fragments were ligated into the SalI-SpeI sites of the pEXP-AD502 vector ( Invitrogen ) to give rise to pAD-CBP20 , pAD-CBP20S245A , pAD-CBP20S245E and pAD-CBP20S245D . To construct miR319b overexpression vector , a 1kb genomic DNA contain the full length of pri-miR319b was amplified using the Phusion High-Fidelity DNA Polymerase . The PCR product was digested with KpnI and SalI , and then the corresponding fragment was ligated into the KpnI-SalI site of the pCHF3 vector to give rise to pCHF3-miR319b . All the sequences above were verified by sequencing . The binary constructs were introduced into Agrobacterium tumefaciens strain GV3101 by electroporation and then introduced into Col-0 or cbp20 mutant plants by the floral dip method [53] . Transgenic plants were screened on MS plates in the presence of 50 mg/L kanamycin , and homozygous lines were verified by antibiotic selection . For each construct , multiple independent lines were examined with similar results , and as least one representative line was shown . The data has been collected from previous study and the calculation was also followed the method as published [11] . Arabidopsis seeds were sown on MS medium plates with or without addition of 1 μM or 10 μM 1-aminocyclopropane-1-carboxylic acid ( ACC , Sigma ) , the biosynthetic precursor of ethylene . After 3 days of cold treatment , the plates were wrapped in foil and kept in 22°C dark chamber for 3 days . The hypocotyls and roots were measured using NIH Image ( http://rsb . info . nih . gov/nih-image/ ) . The yeast two-hybrid assay was performed according to the ProQuest™ Two-Hybrid System ( Invitrogen ) . Briefly , pBD-CBP80 and pAD-CBP20 , -CBP20S245A , -CBP20S245E or -CBP20S245D were co-transformed into the yeast strain Mav203 ( Invitrogen ) . The transformants were grown on SD/-Trp-Leu medium or SD/-Trp-Leu-His with 10mM 3AT dropout medium . The transformants growing on SD/-Trp-Leu-His with 10mM 3AT dropout medium indicates interaction between corresponding proteins . Primers used in this assay were listed in S5 Table . The seedlings of 35S:CBP20-YFP , 35S:CBP20S245A-YFP , 35S:CBP20S245E-YFP and 35S:CBP20S245D-YFP transgenic plants were grown on MS medium with or without addition of 10 μM ACC in dark for 3 days in 22°C . Then the YFP fluorescence of root tips was observed under Zeiss LSM 710 Confocal microscopy . Arabidopsis seeds were grown on MS medium in the air-tight containers in the dark at 22°C supplied with a flow of hydrocarbon-free air ( Zero grade air , AirGas ) for 3 days . The plants tissues were harvest after with continually flow of hydrocarbon-free air or hydrocarbon-free air with 10 parts per million ( ppm ) ethylene gas for 4 hours as previously described [7] . Total RNA was extracted using a RNeasy Plant Kit ( Qiagen ) from 3 days etiolated seedlings treated with air or 4 hours ethylene gas . First-strand cDNA was synthesized using Superscript III First-Strand cDNA Synthesis Kit ( Invitrogen ) . Real time PCR was performed with the LightCycler 480 SYBR Green I Master ( Roche ) following the manufacturer’s instructions . PCR reactions were performed in triplicate on a Roche 96 Thermal cycler . The expression level was normalized to UBQ10 control . Total RNA was isolated from roots or shoots of 3-day old etiolated seedlings treated with air or 4 hours ethylene gas using TRIzol reagent ( Invitrogen ) . For mRNA library construction , in briefly , the mRNA was isolated by NEBNext Poly ( A ) mRNA Magnetic Isolation Module and fragmented at 94°C for 15mins . Then the cDNA was synthesis by NEBNext Ultra Directional RNA Library Prep Kit for Illumina . The PCR reactions were conducted by using different index primers ( NEBNext Multiplex Oligos for Illumina ) . The PCR products were purified by Agencourt AMPure XP Beads ( Beckman Coulter ) . The quality of the libraries was assessed by Bioanalyzer ( Agilent High Sensitivity Chip ) . The libraries then were sequenced on Hiseq 4000 Systems ( Illumina ) . For small RNA library construction , in briefly , the cDNAs were synthesized using NEBNext Small RNA Library Prep Set for Illumina . The PCR reaction was amplified by different Index primers and the PCR products were first purified by the Agencourt AMPure XP Beads , and then selected the size using 6% PolyAcrylamide Gel . The ~140 bp bands corresponding to miRNAs were isolated . The library quality was assessed on Bioanalyzer ( Agilent High Sensitivity Chip ) . The libraries were sequenced on Hiseq 4000 Systems ( Illumina ) after assessed on Bioanalyzer . miRNA prediction pipeline was written by Python scripting language . High-quality small RNA reads were obtained from raw reads through filtering out poor quality reads and removing adaptor sequences using FASTX toolkit [54] . Adaptor-trimmed unique sequences were aligned to TAIR10 Arabidopsis genome using bowtie [55] and structural RNAs such as tRNA , rRNA , snRNA , and snoRNA were excluded . The perfect matched reads between 18–28 nucleotides ( nts ) in length were selected . To obtain the precursor sequences , potential miRNA sequences ( reads ≥ 50 ) were extended upstream and downstream of 100 to 500 nts with a step size of 100 nts . Each putative precursor sequence was folded using RNA fold from Vienna RNA software package [56] , and the potential miRNA* sequences were selected with mismatch ratio of 0 . 3 or less . The region of these putative precursor sequences with addition of 15 nts marginal sequences were re-folded using RNA fold to check whether miRNA/miRNA* duplex was suitable for primary criteria for annotation of plant miRNAs [57] . The miRNA candidates were essentially grouped into families by mature sequence similarity and/or loci . Using the miRNA annotation information of Arabidopsis thaliana in miRBase 21 ( http://www . mirbase . org ) , all members of miRNA candidate families of the known miRNAs were selected . The putative target sites of miRNAs were identified by aligning mature miRNA sequences with the Arabidopsis cDNA sequences using TargetFinder ( http://carringtonlab . org/resources/targetfinder ) . miRNA targets were computationally predicted essentially as described [58–60] . Briefly , potential targets from FASTA searches were scored using a position-dependent , mispairing penalty system . Penalties were assessed for mismatches , bulges , and gaps ( +1 per position ) and G:U pairs ( +0 . 5 per position ) . Penalties were doubled if the mismatch , bulge , gap , or G:U pair occurred at positions 2 to 13 relative to the 5’-end of the miRNAs . Only one single-nucleotide bulge or single-nucleotide gap was allowed , and the targets with penalty scores of six or less were considered to be putative miRNA targets . RNA-seq raw reads were aligned to TAIR10 genome release using Top Hat version 2 . 0 . 9 [61] with default parameters . Differential expressed genes were calculated by Cufflinks version 2 . 2 . 1 following the workflow with default parameters [62] . Differentially expressed genes were those for which relative fold change values of larger than 1 . 5 and RPKM value larger than 1 were observed . To evaluate reproducibility of the RNA-seq data , the expression levels between two replicates for each sample and conditions were compared for all genes with FPKM > 0 . 5 in both replicates . The log2 transformed FPKM values ( log2 ( FPKM + 1 ) was calculated , then R scripts were used to analyze the correlation between biological replicates . Total RNA was isolated from root of 3 days etiolated seedlings treated with air or 4 hours ethylene gas using TRIzol reagent ( Invitrogen ) . 10ug RNA of each sample was separated on 15% denaturing 8M urea-PAGE gel and then transferred and UV crosslinked onto BrightStar®-Plus Positively Charged Nylon Membrane ( Ambion ) . The membrane was pre-hybridized by ULTRAhyb®-Oligo Hybridization Buffer . miRNA probes were end-labelled by T4 Polynucleotide Kinase ( NEB ) with r-P32 ATP . The membrane was hybridized with probe overnight and then wash by 2xSSC for two times . Then the membrane was exposed to a phosphor imager screen and the relative abundance levels were measured by ImageQuant TL software . 5’RLM-RACE was performed following the manufacturer’s instructions of FirstChoice RLM-RACE Kit ( Ambion ) . Briefly , total RNA ( 10 μg ) from root of Col-0 and miR319b OE line was directly ligated to the 5’RACE Adapter by T4 RNA ligase ( Ambion ) . cDNA was synthesized using Superscript III First-Strand cDNA Synthesis Kit ( Invitrogen ) use Oligo ( dT ) primer . Gene-specific reaction was first done with the 5’RACE Outer Primer and gene-specific primer MYB33 SpeI-R . Then the PCR product was purified and performed by the second round of PCR using 5’RACE inner Primer and gene-specific primer MYB33 SpeI-R ( S5 Table ) . The 5RLM-RACE product was gel purified , digested with Sal I and Spe I and then cloned into pDBLeu vector for sequencing . Transient expression assay in N . benthamiana were performed by infiltrating 4-week-old N . benthamiana plants with Agrobacterium containing MYB33 or mutated MYB33 CDS with or without Agrobacterium harbouring constructs containing the miR319b precursor . Leaf tissue was collected 3 days later for RNA and protein analysis . For luciferase assay , Agrobacterium containing MYB33 or mutated mMYB33 CDS fused to 35S::LUC 3’UTR with or without Agrobacterium harbouring constructs containing the miR319b precursor were injected in to N . benthamiana plants . After 3 days , The leaves were sprayed with 500 μM luciferin ( Promega , Madison , Wisconsin ) and placed in the dark for 5 min . Luciferase activity was observed using NightOWL LB 983 in vivo Imaging System ( Berthold , Oak Ridge , Tennessee ) . Primers used in this study were listed in S5 Table . | Ethylene is one of the most essential hormones for plant developmental processes and stress responses . However , the phosphorylation regulation in the ethylene signaling pathway is largely unknown . Here we found that ethylene induces the phosphorylation of CBP20 at S245 , and the phosphorylation is involved in root growth . Genome wide study on ethylene regulated gene expression and microRNA expression together with genetic validation suggest that ethylene- induced phosphorylation of CBP20 is involved in root growth and one pathway is through the regulation of miR319b and its target gene MYB33 . This study provides evidence showing a new link of cap binding protein phosphorylation associated microRNA to root growth in the ethylene response . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"phosphorylation",
"sequencing",
"techniques",
"chemical",
"compounds",
"plant",
"growth",
"and",
"development",
"ethylene",
"gene",
"regulation",
"organic",
"compounds",
"hormones",
"developmental",
"biology",
"plant",
"science",
"micrornas",
"plant",
"hormones",
"molecu... | 2016 | Phosphorylation of CBP20 Links MicroRNA to Root Growth in the Ethylene Response |
Shuni virus ( SHUV ) is an orthobunyavirus that belongs to the Simbu serogroup . SHUV was isolated from diverse species of domesticated animals and wildlife , and is associated with neurological disease , abortions , and congenital malformations . Recently , SHUV caused outbreaks among ruminants in Israel , representing the first incursions outside the African continent . The isolation of SHUV from a febrile child in Nigeria and seroprevalence among veterinarians in South Africa suggests that the virus may have zoonotic potential as well . The high pathogenicity , extremely broad tropism , potential transmission via both biting midges and mosquitoes , and zoonotic features warrants prioritization of SHUV for further research . Additional knowledge is essential to accurately determine the risk for animal and human health , and to assess the risk of future epizootics and epidemics . To gain first insights into the potential involvement of arthropod vectors in SHUV transmission , we have investigated the ability of SHUV to infect and disseminate in laboratory-reared biting midges and mosquitoes . Culicoides nubeculosus , C . sonorensis , Culex pipiens pipiens , and Aedes aegypti were orally exposed to SHUV by providing an infectious blood meal . Biting midges showed high infection rates of approximately 40–60% , whereas infection rates of mosquitoes were lower than 2% . SHUV successfully disseminated in both species of biting midges , but no evidence of transmission in orally exposed mosquitoes was found . The results of this study show that different species of Culicoides biting midges are susceptible to infection and dissemination of SHUV , whereas the two mosquito species tested were found not to be susceptible .
Arthropod-borne ( arbo ) viruses continue to pose a threat to human and animal health [1 , 2] . In particular the order Bunyavirales comprises emerging pathogens such as Crimean-Congo haemorrhagic fever virus ( CCHFV ) and Rift Valley fever virus ( RVFV ) [3 , 4] . The World Health Organization ( WHO ) has included both CCHFV and RVFV to the “Blueprint” list of ten prioritized viruses likely to cause future epidemics and for which insufficient countermeasures are available [5] . In the veterinary field , prioritized viral diseases of animals , including RVFV , are notifiable to the World Organization for Animal Health ( Office International des Epizooties , OIE ) . Apart from pathogens that are recognised as major threats by WHO and OIE , many have remained largely neglected . Before the turn of the century , West Nile virus , chikungunya virus , and Zika virus were among these neglected viruses until they reminded us how fast arboviruses can spread in immunologically naïve populations [2] . Although these outbreaks came as a surprise , in hindsight , smaller outbreaks in previously unaffected areas could have been recognised as warning signs . Shuni virus ( SHUV; family Peribunyaviridae , genus Orthobunyavirus , Simbu serogroup ) recently emerged in two very distant areas of the world [6] . SHUV was isolated for the first time from a slaughtered cow in the 1960s in Nigeria [7] . During subsequent years , the virus was isolated on several occasions from domestic animals including cattle , sheep , goats , and horses [7–10] , from wild animals including crocodiles and rhinoceros [10] , and from field-collected Culicoides biting midges and mosquitoes [8 , 11 , 12] . More recently , SHUV was associated with malformed ruminants in Israel [13 , 14] . Emergence of SHUV in areas outside Sub-Saharan Africa shows the potential of this virus to spread to new areas , and increases the risk for SHUV outbreaks in bordering territories such as Europe . Isolation of SHUV from a febrile child and detection of antibodies in 3 . 9% of serum samples from veterinarians in South Africa shows that SHUV can infect humans as well , although its ability to cause human disease is still uncertain [7 , 15 , 16] . Proper risk assessments rely on accurate knowledge of disease transmission cycles . Arbovirus transmission cycles can only become established when competent vectors and susceptible hosts encounter under suitable climatic conditions . Although SHUV has been isolated from pools of field-collected Culicoides biting midges and mosquitoes [7 , 11 , 12] , the role of both insect groups as actual vectors remains to be confirmed . Detection of virus in field-collected insects is not sufficient to prove their ability to transmit the virus . Arboviruses need to overcome several barriers ( i . e . midgut and salivary gland barriers ) inside their vector , before they can be transmitted [17 , 18] . In addition to virus isolation from field-collected vectors , laboratory studies are therefore needed to experimentally test the ability of blood-feeding insects to become infected with , maintain , and successfully transmit arboviruses ( i . e . , vector competence ) [19] . To gain insights into the potential of Culicoides biting midges and mosquitoes to function as vectors of SHUV , we studied the susceptibility of four main arbovirus vector species ( Culicoides nubeculosus and C . sonorensis biting midges , and Culex pipiens biotype pipiens and Aedes aegypti mosquitoes ) for SHUV .
African green monkey kidney cells ( Vero E6; ATCC CRL-1586 ) were cultured in Eagle’s minimum essential medium ( Gibco , Carlsbad , CA , United States ) supplemented with 5% fetal bovine serum ( FBS; Gibco ) , 1% non-essential amino acids ( Gibco ) , 1% L-glutamine ( Gibco ) , and 1% antibiotic/antimycotic ( Gibco ) . Cells were cultured as monolayers and maintained at 37°C with 5% CO2 . Vero E6 cells that were used in biting midge and mosquito infection experiments in the biosafety level 3 ( BSL3 ) facility were cultured in Dulbecco's modified Eagle medium ( Gibco ) supplemented with 10% FBS , penicillin ( 100 U/ml; Sigma-Aldrich , Saint Louis , MO , United States ) , and streptomycin ( 100 μg/ml; Sigma-Aldrich ) . Prior to infections in the BSL3 facility , Vero E6 cells were seeded in 4- ( 2-hydroxyethyl ) -1-piperazineethanesulfonic acid-buffered DMEM medium ( HEPES-DMEM; Gibco ) supplemented with 10% FBS , penicillin ( 100 U/ml ) , and streptomycin ( 100 μg/ml ) , fungizone ( 50 μg/ml; Invitrogen , Carlsbad , United States ) , and gentamycin ( 50 μg/ml; Gibco ) . C6/36 cells ( ATCC CRL-1660 ) , derived from Ae . albopictus mosquitoes , were cultured in Leibovitz-15 ( L-15 ) growth medium ( Sigma-Aldrich ) supplemented with 10% FBS , 2% Tryptose Phosphate Broth ( Gibco ) , 1% non-essential amino acids solution , and 1% antibiotic/antimycotic . Cells were cultured as monolayers and incubated at 28°C in absence of CO2 . KC cells , derived from embryos of colonized C . sonorensis biting midges [20] , were cultured as monolayers in modified Schneider’s Drosophila medium ( Lonza , Basel , Switzerland ) with 15% FBS , and 1% antibiotic/antimycotic at 28°C in absence of CO2 . SHUV ( strain An10107 , P2 Vero , 1980 ) was kindly provided by the World Reference Center for Emerging Viruses and Arboviruses ( WRCEVA ) . The virus was originally isolated from the blood of a slaughtered cow in 1966 in Nigeria by inoculation of neonatal mice , and passaged twice in Vero cells [21] . The passage 3 ( P3 ) stock was generated by inoculation of Vero E6 cells with the P2 stock at a multiplicity of infection ( MOI ) of 0 . 001 . The supernatant was harvested at 6 days post inoculation , centrifuged , and stored in aliquots at -80°C . The P4 stock was generated by inoculating Vero E6 cells at MOI 0 . 01 using the P3 stock . At this MOI , full cytopathic effect ( CPE ) was present at 3 days post infection . Virus titers were determined using endpoint dilution assays ( EPDA ) on Vero E6 cells [22] . Titers were calculated using the Spearman-Kärber algorithm and expressed as 50% tissue culture infective dose ( TCID50 ) [23 , 24] . The virus detection and titration procedure was validated using a SHUV-specific reverse transcriptase quantitative PCR ( RT-qPCR; S1 Supporting Information ) . Cells were seeded in T25 cell culture flasks at densities of 7 . 5 × 105 ( Vero E6 ) , 1 . 5 × 106 ( C6/36 ) , or 2 . 5 × 106 ( KC cells ) per flask in 10 ml complete medium . After overnight incubation , the flasks were inoculated with SHUV at an MOI of 0 . 01 ( P4 stock ) . The MOI calculation for each cell line was based on the virus titer that was determined on Vero E6 cells . One hour after inoculation , the medium was removed and replaced with fresh medium . At time points 0 ( sample taken directly after medium replacement ) , 24 , 48 , and 72 h post infection , 200 μl samples were taken and stored at -80°C for later analysis . For each cell line , virus titers were determined in triplicate per time point by EPDA using Vero E6 cells , which showed distinct CPE [22] . Culicoides nubeculosus were kindly provided by The Pirbright Institute , Pirbright laboratories , United Kingdom , in 2012 [25] , and were maintained at 23°C with 16:8 light:dark cycle and 60% relative humidity . Culicoides sonorensis were kindly provided by the Arthropod-Borne Animal Diseases Research Laboratory , USDA-ARS ( courtesy of Dr . Barbara Drolet ) in 2017 [26] , and were maintained at 25°C with 16:8 light:dark cycle and 70% relative humidity . Similar rearing protocols were used for both biting midge species . Eggs were transferred to square larval holding trays ( C . nubeculosus: 25 x 25 x 8 cm , Kartell , Noviglio , Italy; C . sonorensis: 19 x 19 x 20 cm , Jokey , Wipperfürth , Germany ) with filter wool ( Europet Bernina International , Gemert-Bakel , The Netherlands ) attached with double-sided tape to the bottom . Trays were filled with tap water , a few millilitres of rearing water in which larvae had completed their life cycle , and two drops of Liquifry No . 1 ( Interpet , Dorking , United Kingdom ) . Larvae were fed with a 1:1:1 mixture of bovine liver powder ( MP biomedicals , Irvine , CA , US ) , ground rabbit food ( Pets Place , Ede , The Netherlands ) , and ground koi food ( Tetra , Melle , Germany ) . Culicoides nubeculosus larvae were additionally fed with nutrient broth No . 2 ( Oxoid , Hampshire , UK ) . Pupae were transferred to plastic buckets ( diameter: 12 . 2 cm , height: 12 . 2 cm; Jokey ) and closed with netting on the top through which the biting midges could feed . Emerged adults were provided with 6% glucose solution ad libitum . Cow blood ( Carus , Wageningen , The Netherlands ) was provided through a Parafilm M membrane using the Hemotek PS5 feeding system ( Discovery Workshops , Lancashire , United Kingdom ) for egg production . The Cx . pipiens pipiens colony was established in the laboratory from egg rafts collected in the field in The Netherlands during August 2016 . Egg rafts were individually hatched in tubes . Pools of approximately 10 first instar larvae were identified to the biotype level using real-time PCR [27] . The colony was started by grouping larvae from 93 egg rafts identified as the pipiens biotype . Mosquitoes were maintained at 23°C with 16:8 light:dark cycle and 60% relative humidity [28 , 29] . Adult mosquitoes were kept in Bugdorm-1 rearing cages and maintained on 6% glucose solution ad libitum . Cow blood or chicken blood ( Kemperkip , Uden , The Netherlands ) was collected in BC Vacutainer lithium heparin-coated blood collection tubes ( Becton Dickinson , Breda , The Netherlands ) , and stored at 4°C . Blood was provided through a Parafilm M membrane using the Hemotek PS5 feeding system for egg production . Egg rafts were transferred to square larval holding trays ( 25 x 25 x 8 cm , Kartell ) filled with tap water and two drops of Liquifry No . 1 . Hatched larvae were fed with a 1:1:1 mixture of bovine liver powder , ground rabbit food , and ground koi food . Pupae were collected every 2 days and placed in Bugdorm-1 insect rearing cages . Aedes aegypti mosquitoes from the Rockefeller strain ( Bayer AG , Monheim , Germany ) were used in all experiments . The mosquito colony was maintained as described before [30] . In short , mosquitoes were maintained at 27°C with 12:12 light:dark cycle and 70% relative humidity . Adult mosquitoes were kept in Bugdorm-1 rearing cages and maintained on 6% glucose solution ad libitum . Human blood ( Sanquin Blood Supply Foundation , Nijmegen , The Netherlands ) was provided through a Parafilm M membrane using the Hemotek PS5 feeding system for egg production . Eggs were transferred to transparent square larval holding trays ( 19 x 19 x 20 cm , Jokey ) , filled for approximately one-third with tap water and three drops of Liquifry No . 1 . Hatched larvae were fed with Tetramin Baby fish food ( Tetra ) . Larval trays were closed with fine-meshed netting , to allow adult mosquitoes to emerge inside larval trays . Twice a week , adults were aspirated from the larval trays and collected in Bugdorm-1 insect rearing cages . Groups of adult C . nubeculosus ( 1–7 days old ) , C . sonorensis ( 1–11 days old ) , Cx . p . pipiens ( 4–20 days old ) , and Ae . aegypti ( 4–7 days old ) were transferred to plastic buckets ( diameter: 12 . 2 cm , height: 12 . 2 cm; Jokey ) and closed with netting before being taken to the BSL3 facility . Culex p . pipiens mosquitoes were kept on water for 3 days , whereas the other species were maintained on 6% glucose solution until being offered an infectious blood meal . SHUV P3 stock with a mean titer of 3 . 0 x 106 TCID50/ml was mixed 1:1 with cow blood . The used cow blood was tested negative for Schmallenberg virus ( SBV ) antibodies , to prevent cross-neutralisation with SHUV . The infectious blood meal was provided through a Parafilm M membrane using the Hemotek PS5 feeding system , under dark conditions at 24°C and 70% relative humidity . After 1 h , insects were anesthetized with 100% CO2 and kept on a CO2-pad to select fully engorged females . For each species , five fully engorged females were directly stored at -80°C for each replicate . These samples were used to determine the ingested amounts of SHUV for each species . All remaining and fully engorged females were placed back into buckets with a maximum group size of 110 individuals per species per bucket . All insects were provided with 6% glucose solution via a soaked ball of cotton wool on top of the netting ad libitum . Culicoides sonorensis and Ae . aegypti were kept at 28°C for 10 days , whereas C . nubeculosus and Cx . p . pipiens were kept at 25°C for 10 days . These temperatures were selected for optimal replication of the virus , and to reflect differences in the rearing temperature for each species . Three replicate experiments of C . nubeculosus ( N1 = 84 , N2 = 82 , N3 = 77 , Ntotal = 243 ) , C . sonorensis ( N1 = 9 , N2 = 9 , N3 = 30 , Ntotal = 48 ) , and Cx . p . pipiens ( N1 = 89 , N2 = 57 , N3 = 65 , Ntotal = 211 ) were carried out , and two replicate experiments of Ae . aegypti ( N1 = 72 , N2 = 77 , Ntotal = 149 ) . During each replicate , biting midges and mosquitoes were fed in parallel with the same infectious blood meal . Adult female Cx . p . pipiens ( 3–9 days old ) and Ae . aegypti ( 4–6 days old ) mosquitoes were injected with SHUV into the thorax to investigate the role of mosquito barriers on dissemination of SHUV . Mosquitoes were anesthetized with 100% CO2 and positioned on the CO2-pad . Female mosquitoes were intrathoracically injected with 69 nl of SHUV ( P3 stock with a titer of 3 . 0 x 106 TCID50/ml ) using a Drummond Nanoject II Auto-Nanoliter injector ( Drummond Scientific , Broomall , Unites States ) . Injected Cx . p . pipiens were maintained at 25°C and injected Ae . aegypti were maintained at 28°C . Mosquitoes were incubated for 10 days at the respective temperatures , and had access to 6% glucose solution ad libitum . Injections were done during a single replicate experiment for Cx . p . pipiens ( N = 50 ) and Ae . aegypti ( N = 50 ) . After 10 days of incubation at the respective incubation temperatures , samples from surviving biting midges and mosquitoes were collected . Biting midges were anesthetized with 100% CO2 and transferred individually to 1 . 5 ml Safe-Seal micro tubes ( Sarstedt , Nümbrecht , Germany ) containing 0 . 5 mm zirconium beads ( Next Advance , Averill Park , NY , United States ) . For a selection of C . nubeculosus ( N = 77 ) and C . sonorensis ( N = 30 ) from one replicate experiment , heads were removed from bodies and separately stored in tubes . All samples were stored at -80°C until further processing . Mosquitoes were anesthetized with 100% CO2 to remove legs and wings . Mosquito saliva was then collected by inserting the proboscis into a 200 μl yellow pipette tip ( Greiner Bio-One ) containing 5 μl of a 1:1 solution of 50% glucose solution and FBS . The saliva sample was transferred to a 1 . 5 ml micro tube containing 55 μl of fully supplemented HEPES-DMEM medium . Mosquito bodies were individually stored in 1 . 5 ml Safe-Seal micro tubes containing 0 . 5 mm zirconium beads . Frozen biting midge and mosquito tissues were homogenized for 2 min at maximum speed ( setting 10 ) in the Bullet Blender Storm ( Next advance ) , centrifuged for 30 seconds at 14 , 500 rpm in the Eppendorf minispin plus ( Eppendorf , Hamburg , Germany ) , and suspended in 100 μl of fully supplemented HEPES-DMEM medium . After addition of the medium , samples were blended again for 2 min at maximum speed , and centrifuged for 2 min at 14 , 500 rpm . Mosquito saliva samples were thawed at RT and vortexed before further use . In total 30 μl of each body or saliva sample was inoculated on a monolayer of Vero E6 cells in a 96 wells plate . SHUV stock or infectious blood mixture was included as positive control and wells to which no sample was added were included as negative controls . After 2–3 h the inoculum was removed and replaced by 100 μl of fully supplemented HEPES-DMEM medium . Wells were scored for virus induced CPE at 3 and 7 days post inoculation , with full CPE being observed at the latter time point . Afterwards , virus titers for positive samples of biting midge bodies and heads , as well as mosquito bodies and saliva were determined with single EPDA on Vero E6 cells [30] . Virus titers were determined using the Reed & Muench algorithm [31] . A subset of samples was validated by RT-qPCR , to confirm that observed CPE was induced by SHUV ( S1 Supporting Information ) . Infection rate ( virus-infected whole body ) and dissemination efficiency ( virus-infected head ) were determined for biting midges , whereas infection rate ( virus-infected whole body ) and transmission efficiency ( virus-infected saliva ) were determined for mosquitoes . Infection rate , dissemination efficiency , and transmission efficiency were calculated , respectively , by dividing the number of females with virus-infected bodies ( infection ) , virus-infected heads ( dissemination ) , or virus-infected saliva ( transmission ) by the total number of females tested in the respective treatment and that survived the incubation period . The values were subsequently expressed as percentages by multiplying with 100 . Two biting midge samples of which only the head was virus-positive , but not the body , were considered to be uninfected .
Mammalian , mosquito , and midge cells were inoculated with SHUV to gain insight into the replicative fitness of this virus and strain in different host cell types . The results show that SHUV is capable to produce progeny in all three cell types ( Fig 1 and S1 Data ) . Of note , a strong CPE was observed in the Vero E6 cells upon infection whereas no CPE was observed in the insect cell lines . Therefore , Vero E6 cells were used to determine titers by EPDA . To evaluate the susceptibility of two species of biting midges ( C . nubeculosus and C . sonorensis ) for SHUV , groups of individuals of both species were orally exposed to an infectious blood meal with a mean SHUV titer of 3 . 0 x 106 TCID50/ml . SHUV titers of ingested blood were determined for a selection of 10 fully engorged females for each species , that were directly stored at -80°C after feeding . Both species ingested low amounts of SHUV that were below the detection limit of the endpoint dilution assay of 103 TCID50/ml . Infection rates were also determined after 10 days of incubation at temperatures of 25°C ( C . nubeculosus and Cx . p . pipiens ) or 28°C ( C . sonorensis and Ae . aegypti; Fig 2 and S2 Data ) . Both biting midge species showed high infection rates of 44% for C . nubeculosus ( N = 243 ) , and 60% for C . sonorensis ( N = 48; Fig 2A ) . SHUV replicated to median titers of 2 . 4 x 103 TCID50/ml in body samples of C . nubeculosus and 1 . 1 x 104 TCID50/ml in body samples of C . sonorensis ( Fig 2E ) . For one replicate experiment , heads were separated from the bodies and tested for presence of SHUV to assess whether the virus successfully passed from the midgut to the haemocoel , indicative of dissemination throughout the body . Dissemination efficiencies were 18% ( N = 77 ) for C . nubeculosus and 10% ( N = 30 ) for C . sonorensis ( Fig 2C ) . In all virus-positive heads that induced CPE , SHUV titers were lower than 103 TCID50/ml . Because only very low amounts of SHUV were detected in biting midge heads , the actual percentage of disseminated infections might be higher . A subset of the samples was additionally tested by RT-qPCR to confirm that CPE was induced by SHUV ( S1 Supporting Information ) . The relatively high infection rates and dissemination efficiencies observed in this study and the absence of a salivary glands barrier in biting midges as shown in previous studies [17 , 32] , suggests that both C . nubeculosus and C . sonorensis have the potential to transmit SHUV . SHUV was previously isolated from field-collected mosquitoes [8] . Therefore , we determined vector competence for two mosquito species ( Cx . p . pipiens and Ae . aegypti ) which are important vectors for several arboviruses [22 , 28 , 30] . SHUV titers of ingested blood were determined for a selection of 10 fully engorged female mosquitoes that were directly stored at -80°C after feeding on an infectious blood meal with a SHUV titer of 3 . 0 x 106 TCID50/ml . Similar to results obtained with the biting midges , the amounts of SHUV ingested by both mosquito species was less than 103 TCID50/ml . No SHUV infection was observed in the Cx . p . pipiens mosquitoes ( N = 211 ) following oral exposure , whereas infection rates of 2% were found for orally exposed Ae . aegypti mosquitoes ( N = 149; Fig 2B ) . SHUV replicated to median titers of 6 . 3 x 103 TCID50/ml in body samples of Ae . aegypti ( Fig 2F ) , which was comparable to titers found in biting midges . No SHUV was detected in any of the saliva samples taken from either Cx . p . pipiens or Ae . aegypti ( Fig 2D ) . Thus , SHUV was able to successfully infect a small proportion of Ae . aegypti mosquitoes but not Cx . p . pipiens , and no evidence was found for transmission of SHUV by mosquitoes . The very low infection rates of mosquitoes triggered further investigation into potential mosquito barriers against SHUV infection . To this end , Cx . p . pipiens and Ae . aegypti mosquitoes were intrathoracically injected with SHUV , to bypass the potential midgut barrier . Direct injection of SHUV into the thorax resulted in high infection rates of 70% for Cx . p . pipiens ( N = 50 ) , and 100% for Ae . aegypti ( N = 50; Fig 3A ) . Transmission efficiency of 32% ( N = 50 ) was found for Cx . p . pipiens and 8% ( N = 50 ) for Ae . aegypti ( Fig 3B ) . Interestingly , although infection rates of Cx . p . pipiens were below 100% , we found a relatively high transmission efficiency . This may indicate a relatively weaker salivary gland barrier in Cx . p . pipiens compared to Ae . aegypti mosquitoes that had 100% infection rate , but relatively low transmission efficiency . To gain more insight in replication of SHUV , virus titers were determined for virus-infected mosquito body and saliva samples . Titers of virus-infected Cx . p . pipiens body samples were almost all below the detection limit of 103 TCID50/ml of the endpoint dilution assay ( Fig 3C ) . This indicates that even when SHUV is injected into the thorax , there is no productive virus replication . In contrast , we found median titers of 7 . 1 x 104 TCID50/ml for virus-infected Ae . aegypti body samples . This shows that SHUV is able to successfully replicate in Ae . aegypti when the midgut barrier is bypassed . In the majority of mosquito saliva samples , SHUV titers were less than 103 TCID50/ml ( Fig 3D ) . Taken together , SHUV is able to disseminate in mosquitoes , but both the midgut and salivary glands form a barrier for SHUV .
SHUV was previously isolated from field-collected pools of Culicoides biting midges and from mosquitoes , but their involvement in SHUV transmission remained to be confirmed [8 , 11 , 12] . Here , we show for the first time that SHUV is able to infect and replicate in biting midges as well as in mosquitoes , but only the biting midge species evaluated in the present study can be considered highly susceptible to infection . Both C . nubeculosus and C . sonorensis showed high infection rates of 44% and 60% when incubated for 10 days at 25°C and 28°C , respectively . It has been demonstrated that a salivary gland barrier is absent for Orbiviruses and Schmallenberg virus in biting midges [17 , 32] . This knowledge , in combination with evidence of successful dissemination of SHUV to the heads indicates that the biting midge species evaluated in the present study are likely competent vectors of SHUV . Importantly , the finding that SHUV replicates efficiently in two biting midge species from a different geographic background suggests that various species of Culicoides may function as vectors of SHUV . SHUV infection and replication in biting midges seems more efficient compared to other biting midge-borne viruses such as SBV and bluetongue virus ( BTV ) , which generally show infection rates up to 30% [32–36] . Both SBV and BTV have caused sudden and large-scale epizootics in Europe , with devastating consequences for the livestock sector [37 , 38] . The relatively high susceptibility and efficiency of replication in biting midges , and recent spread of SHUV to areas outside Sub-Saharan Africa [13] , should therefore be interpreted as a warning for its epizootic potential . In contrast to the high infection rates in biting midges , only few orally exposed Ae . aegypti mosquitoes became infected with SHUV during 10 days of incubation at 28°C . In addition , no evidence of successful dissemination to the salivary glands of the two mosquito species was found . SHUV replication and transmission ( 8% ) was observed when the virus was directly injected into the thorax of Ae . aegypti mosquitoes . This indicates that both the midgut infection barrier and the salivary gland barrier prevent infection and subsequent transmission of SHUV by Ae . aegypti mosquitoes . Of the Cx . p . pipiens mosquitoes that were orally exposed to SHUV , none became infected during 10 days of incubation at 25°C . Moreover , replication of SHUV was low in Cx . p . pipiens , as evidenced by low titers when it was directly injected into the thorax . However , a relatively high percentage of mosquito saliva samples contained SHUV . We therefore conclude that the midgut barrier is the main barrier that prevents infection of Cx . p . pipiens with SHUV . Our findings are in line with an earlier study on the closely-related SBV , which showed no evidence for involvement of Cx . pipiens in virus transmission , although SBV was able to infect Cx . pipiens mosquitoes [39] . However , as Cx . theileri has been identified as a vector of several other bunyaviruses , this mosquito may also be a possible vector of SHUV [40 , 41] . Thus , vector competence studies with additional mosquito species collected from the field are needed to fully understand the possible role of mosquitoes in natural transmission cycles of SHUV . In this study , we determined infection , dissemination , and transmission of SHUV by infectivity assays and virus titers by EPDA ( i . e . assays based on inoculation of samples on Vero cells which are then screened for CPE ) . Such infectivity assays and EPDAs have the advantage of detecting infectious virus particles , whereas other methods like qPCR that quantify genome equivalents , may include defective virus particles and thereby not accurately represent infectious virus . Of note , observed CPE in the infectivity assays and EPDAs was found to invariably correspond with SHUV RNA as determined by RT-qPCR ( S1 Supporting Information ) . Recent outbreaks of SBV and BTV exemplified the tremendous impact of midge-borne viruses on animal health [37 , 38] . Our study demonstrates highly efficient infection , replication , and dissemination of SHUV in two biting midge species ( C . nubeculosus and C . sonorensis ) . However , conclusive evidence for SHUV transmission by biting midges should be provided by experiments with infected biting midges and susceptible mammals , although these kind of experiments are costly and complex . We cannot exclude that results obtained with laboratory-reared vectors are different from those obtained with field-collected vectors . Therefore , future studies should test vector competence of field-collected Culicoides biting midge and mosquito species exposed to different quantities of SHUV , to more accurately predict the risk of SHUV transmission in specific areas . These experiments in combination with behavioural and ecological research will contribute to our understanding of the transmission cycle of SHUV . | Arthropod-borne ( arbo ) viruses are notorious for causing unpredictable and large-scale epidemics and epizootics . Apart from viruses such as West Nile virus and Rift Valley fever virus that are well known to have a significant impact on human and animal health , many arboviruses remain neglected . Shuni virus ( SHUV ) is a neglected virus with zoonotic potential that was recently associated with severe disease in livestock and wildlife . Isolations of SHUV from field-collected biting midges and mosquitoes suggests that SHUV may be transmitted by these insects . Laboratory-reared biting midge species ( Culicoides nubeculosus and C . sonorensis ) and mosquito species ( Culex pipiens pipiens and Aedes aegypti ) , that are known to transmit other arboviruses , were exposed to SHUV via an infectious blood meal . SHUV was able to successfully disseminate in both biting midge species , whereas no evidence of infection or transmission in both mosquito species was found . Our results show that SHUV infects and disseminates in two different Culicoides species , suggesting that these insects could play an important role in the disease transmission cycle . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"culicoides",
"microbiology",
"saliva",
"animals",
"viruses",
"insect",
"vectors",
"veterinary",
"science",
"infectious",
"diseases",
"veterinary",
"diseases",
"aedes",
"aegypti",
"viral",
"repl... | 2018 | Vector competence of biting midges and mosquitoes for Shuni virus |
Bank voles are uniquely susceptible to a wide range of prion strains isolated from many different species . To determine if this enhanced susceptibility to interspecies prion transmission is encoded within the sequence of the bank vole prion protein ( BVPrP ) , we inoculated Tg ( M109 ) and Tg ( I109 ) mice , which express BVPrP containing either methionine or isoleucine at polymorphic codon 109 , with 16 prion isolates from 8 different species: humans , cattle , elk , sheep , guinea pigs , hamsters , mice , and meadow voles . Efficient disease transmission was observed in both Tg ( M109 ) and Tg ( I109 ) mice . For instance , inoculation of the most common human prion strain , sporadic Creutzfeldt-Jakob disease ( sCJD ) subtype MM1 , into Tg ( M109 ) mice gave incubation periods of ∼200 days that were shortened slightly on second passage . Chronic wasting disease prions exhibited an incubation time of ∼250 days , which shortened to ∼150 days upon second passage in Tg ( M109 ) mice . Unexpectedly , bovine spongiform encephalopathy and variant CJD prions caused rapid neurological dysfunction in Tg ( M109 ) mice upon second passage , with incubation periods of 64 and 40 days , respectively . Despite the rapid incubation periods , other strain-specified properties of many prion isolates—including the size of proteinase K–resistant PrPSc , the pattern of cerebral PrPSc deposition , and the conformational stability—were remarkably conserved upon serial passage in Tg ( M109 ) mice . Our results demonstrate that expression of BVPrP is sufficient to engender enhanced susceptibility to a diverse range of prion isolates , suggesting that BVPrP may be a universal acceptor for prions .
Prions , or proteinaceous infectious particles , are self-propagating protein conformations that cause a variety of fatal neurodegenerative illnesses . Prions composed of the prion protein ( PrP ) cause Creutzfeldt-Jakob disease ( CJD ) in humans , scrapie in sheep , chronic wasting disease ( CWD ) in cervids , and bovine spongiform encephalopathy ( BSE ) [1] , [2] , [3] . In these diseases , cellular PrP ( PrPC ) , which is a glycosylphosphatidylinositol ( GPI ) -anchored membrane protein , undergoes a conformational conversion into a β-sheet-rich , aggregation-prone isoform , termed PrPSc [4] , [5] . Accumulation of PrPSc within the central nervous system ( CNS ) results in profound neurological dysfunction as well as neuropathological changes , which include spongiform ( vacuolar ) degeneration , astrocytic gliosis , and neuronal loss . In contrast to PrPC , which is sensitive to protease digestion , the most commonly encountered forms of PrPSc are partially resistant to digestion with proteases , producing a truncated fragment referred to as PrP 27–30 [6] . Distinct strains of prions can be distinguished and classified by the incubation periods upon inoculation of laboratory animals , differences in neuroanatomic target areas and patterns of PrPSc deposition within the brain , and biochemical properties , including the molecular weight of PrP 27–30 [7] , [8] . It is believed that the properties of distinct prion strains are enciphered within the conformation of PrPSc [9] , [10] . In some instances , it is more appropriate to refer to prion strains as “isolates” if they have not been serially passaged . The intraspecies transmission of various prion strains or isolates is generally an efficient process , in which 100% of the inoculated animals develop CNS disease , the incubation period is relatively uniform , neuropathologic patterns are similar , and biochemical properties of PrPSc are maintained . In contrast , the interspecies transmission of prions is usually an inefficient process , in which only a fraction of inoculated animals develop signs of neurologic dysfunction , resulting in more variable and prolonged incubation periods [11] , [12] . Furthermore , the properties of prion strains or isolates are frequently altered upon initial passage in a different species [13] . Upon second passage , the incubation periods are shorter , and the biochemical and neuropathological properties of the prion isolate are stabilized . This phenomenon is what is referred to as the “species barrier” for prion replication [14] . At the molecular level , the species barrier was initially believed to be governed entirely by the sequence of PrP , with interspecies differences in the amino acid sequence of PrP hindering disease transmission [15] . For instance , transgenic ( Tg ) mice expressing hamster or human PrP are susceptible to hamster or human prions , respectively , whereas wild-type ( wt ) mice are largely resistant to prions from either species [16] , [17] . With further study , it became clear that the initially monastic view of the species barrier was incomplete: in particular , differences in the sequences of PrPSc in the inoculum and PrPC in the host were insufficient to explain all aspects of prion transmission from one host to another . For example , the MM1 subtype of sporadic ( s ) CJD prions transmitted to Tg mice expressing the M129 variant of human PrP in ∼200 days while variant ( v ) CJD prions required more than 600 days [18] . Conversely , Tg mice expressing bovine PrP exhibited signs of neurological dysfunction at ∼270 days after inoculation with vCJD prions , but remained well for greater than 500 days after inoculation with sCJD ( MM1 ) prions [19] , [20] , [21] . Importantly , the proteins comprising vCJD and sCJD ( MM1 ) prions had the same amino acid sequence , arguing that an additional “barrier” must be invoked to explain the differences in transmission efficiency described above , which might best be called a “strain barrier” to reflect distinct conformations of PrPSc molecules . Together , the species and strain barriers have been called “transmission barriers , ” where a given PrPC sequence is capable of propagating only a distinct subset of PrPSc conformations [20] , [22] . When the PrPC and PrPSc conformations are compatible , efficient disease transmission occurs . Unlike other rodents , bank voles ( Myodes glareolus ) are susceptible to prions from a diverse range of species , including humans [23] , [24] , [25] , [26] , [27] . This suggests that species and possibly strain barriers are greatly attenuated in bank voles , an observation that has been recapitulated in vitro using protein misfolding cyclic amplification ( PMCA ) [28] . Two explanations seem plausible for the promiscuity of bank voles for replicating prions originating in diverse species: first , the presence of an especially permissive prion replication cofactor [17] , [29] , [30] , [31] , [32] and second , a broadly compatible bank vole PrP ( BVPrP ) sequence . The latter would seem to be the more parsimonious explanation where the amino acid sequence of BVPrP facilitates adoption of self-propagating conformations both spontaneously and upon exposure to exogenous prions [24] , [33] . The sequence of the mature processed form of BVPrP , in which the N- and C-terminal signal peptides have been removed , differs from that of mouse PrP at only eight positions [25] . Notably , the high-resolution structure of bank vole PrPC revealed the presence of a “rigid loop” but no remarkable characteristics that might confer its unique replication behavior [34] . To determine the range of prion susceptibility conferred by BVPrP expression , we challenged Tg mice expressing BVPrP with 16 prion isolates from 8 different species . BVPrP is polymorphic at codon 109 , where either a methionine ( M ) or isoleucine ( I ) residue can be present [35] . Tg ( BVPrP ) mice expressing either the M109 or I109 polymorphic variant of BVPrP were susceptible to a wide range of prion isolates derived from many species , confirming that the enhanced susceptibility of bank voles to prions with different PrP sequences is mediated by the sequence of BVPrP .
Because Tg mice that express the I109 allotype of BVPrP develop age-dependent signs of spontaneous neurologic illness [33] , we initially focused our studies on Tg mice expressing the M109 allotype . Tg ( BVPrP , M109 ) 22019 mice , denoted Tg ( M109 ) mice , express BVPrP at ∼5 times the level of PrP expression found in wt mice and did not develop any signs of spontaneous neurologic illness up to 500 days of age [33] . We inoculated the Tg ( M109 ) mice with 16 different prion isolates derived from humans , cattle , elk , sheep , guinea pigs , hamsters , mice or meadow voles ( MV ) ( Table 1 ) . The following prion isolates were tested: sCJD ( three subtypes: MM1 , MM2 , and VV2 ) ; vCJD; sCJD ( MM1 ) prions passaged in Tg mice expressing the M129 variant of human PrP [Tg ( HuPrP ) mice]; cattle BSE; elk CWD; sheep scrapie isolate SSBP/1; sCJD ( MM1 ) prions passaged in guinea pigs; hamster-adapted scrapie strain Sc237; mouse-adapted scrapie strain RML; mouse-adapted BSE strain 301V [maintained in mice expressing either the PrP-A or PrP-B allotype of mouse PrP and denoted 301V ( A ) and 301V ( B ) , respectively]; MV-adapted RML; and MV-adapted Sc237 prions . Remarkably , 119 of 120 inoculated Tg ( M109 ) mice developed signs of neurologic dysfunction consistent with prion disease , with mean incubation periods ranging from 50 days to just under 400 days ( Table 1 ) . The relatively short incubation periods and high transmission efficiencies for this diverse set of prion isolates in Tg ( M109 ) mice suggest that these mice , like bank voles , do not impose a barrier for interspecies prion transmission . To determine the reproducibility of these findings , we utilized another Tg line denoted Tg ( BVPrP , M109 ) 3118 mice , which express BVPrP at ∼2 . 5 times the level of PrP expression in wt mice . Like the Tg ( M109 ) 22019 mice , the Tg ( M109 ) 3118 mice were also susceptible to MV- , mouse- , hamster- , and human-derived prion isolates ( Table S1 ) . Tg ( MoPrP ) mice , which overexpress mouse PrP at ∼4–5 times the level in wt mice , did not exhibit a general susceptibility to prions ( Table S2 ) , arguing that the increased susceptibility of Tg ( M109 ) mice to diverse prion isolates cannot be attributed to PrP overexpression . In the Tg ( M109 ) mice , two cases of sCJD ( MM1 ) prions produced incubation times of ∼200 days; on second passage , the incubation period decreased modestly to 175 days . Relatively larger reductions in the incubation times on second passage were observed with other prion isolates . For example , sCJD ( MM2 ) prions gave an incubation time of ∼240 days on first passage , which decreased to 90 days on second passage , and sCJD ( VV2 ) prions decreased from ∼400 days on first passage to ∼100 days on second passage . Most interesting among the human isolates was vCJD , with an initial incubation time of 330 days , which decreased to 40 days on second passage in Tg ( M109 ) mice , an 8-fold reduction upon serial transmission . Notably , BSE prions , from which vCJD prions are derived , exhibited an incubation time of ∼370 days on first passage and decreased to ∼65 days on second passage . CWD prions from elk produced neurological dysfunction in ∼250 days on first passage and ∼150 days on second passage in Tg ( M109 ) mice . Sc237 prions from Syrian hamsters inoculated into Tg ( M109 ) mice produced an incubation time of ∼95 days on first passage , which decreased to ∼75 days on second passage . When RML prions from wt mice were inoculated into Tg ( M109 ) mice , the incubation time was ∼75 days but decreased to ∼50 days on second passage . These results suggest that transmission barriers still exist for some strains despite a general susceptibility of Tg ( M109 ) mice to many different prion isolates . Proteinase K ( PK ) -resistant PrPSc was found in the brains of all clinically ill Tg ( M109 ) mice inoculated with each of the prion isolates tested ( Figure 1A ) . Furthermore , spongiform degeneration and prominent astrocytic gliosis were found in the brains of Tg ( M109 ) mice inoculated with each of the different isolates ( Figure S1 ) , confirming that these mice developed prion disease . However , levels of PK-resistant PrPSc were much lower in RML-inoculated Tg ( M109 ) mice than in RML-inoculated wt mice or in RML-infected Tg ( MoPrP ) mice ( Figure S2A ) . Similarly , levels of PK-resistant PrPSc following challenge of Tg ( M109 ) mice with Sc237 prions were much lower than in Sc237-infected hamsters or in Sc237-inoculated Tg mice overexpressing hamster PrP [Tg ( SHaPrP ) mice] ( Figure S2B ) . Levels of PK-resistant PrPSc remained low upon second passage of RML or Sc237 prions in Tg ( M109 ) mice . However , this was not true for all the prion isolates analyzed: substantially higher levels of PK-resistant PrPSc were observed in the brains of BSE- and vCJD-inoculated Tg ( M109 ) mice ( Figure S2C ) . Prion strains can be classified according to the electrophoretic mobility of the unglycosylated band of PK-resistant PrPSc , migrating to either ∼21 kDa or ∼19 kDa , respectively termed type 1 and type 2 strains , similar to the nomenclature for sCJD prions [8] . In Tg ( M109 ) mice , type 1 strains migrated to ∼20 kDa and type 2 strains migrated to ∼19 kDa ( Figure 1A ) . In general , the electrophoretic mobilities observed for the original prion isolates were conserved upon transmission to Tg ( M109 ) mice ( Figure 1B–K ) . For instance , RML , Sc237 , CWD , scrapie SSBP/1 , and sCJD ( MM1 ) are type 1 strains and exhibited a type 1 pattern upon transmission to Tg ( M109 ) mice . Similarly , BSE , sCJD ( MM2 ) , sCJD ( VV2 ) , and vCJD are type 2 strains and generated type 2 strains following transmission to Tg ( M109 ) mice . Slight alterations in the size of PK-resistant PrPSc were observed for the sCJD ( MM1 ) , CWD , and SSBP/1 isolates upon propagation in Tg ( M109 ) mice ( Figure 1B , G–H ) , and type 2 PrPSc in Tg ( M109 ) mice had a slightly larger molecular mass compared to the type 2 PrPSc in the original human inocula ( Figure 1C–F ) . However , of the 11 isolates analyzed , only 2 clearly changed strain type upon passage in Tg ( M109 ) mice: both the 301V ( A ) and 301V ( B ) isolates exhibited a type 2 pattern in Tg ( M109 ) mice whereas the original isolates were type 1 strains ( Figure 1K ) . Thus , Tg ( M109 ) -passaged 301V prions more closely resembled the BSE isolate from which the 301V strain was originally derived . Another method for discriminating prion strains is the comparison of the relative abundances of di- , mono- , and unglycosylated PK-resistant PrPSc . The most abundant glycoform for all prion isolates was diglycosylated PrPSc following passage in Tg ( M109 ) mice ( Figure 1A ) . For prion isolates with high levels of diglycosylated PrPSc ( such as Sc237 , 301V , CWD , BSE , and vCJD ) , the glycoform ratios appeared to be conserved upon serial passage in Tg ( M109 ) mice ( Figure 1E–G , I , K ) . In contrast , for prion isolates that did not exhibit high levels of diglycosylated PrPSc , such as sCJD ( MM1 ) , sCJD ( MM2 ) , sCJD ( VV2 ) , SSBP/1 , and RML , the relative abundance of diglycosylated PrPSc increased upon propagation in Tg ( M109 ) mice ( Figure 1B–D , H , J ) . To further investigate whether the properties of the prion isolates were conserved upon transmission in Tg ( M109 ) mice , we examined the patterns of PrPSc deposition in the brains of prion-inoculated Tg ( M109 ) mice . PrPSc deposition was found in the brains of all clinically ill Tg ( M109 ) mice inoculated with each of the prion isolates tested ( Figure 2A–R ) , although the level of PrPSc deposition in Tg ( M109 ) mice was typically less than what is observed in other experimentally inoculated laboratory animals . Generally , the characteristic pattern of PrPSc deposition for a given prion isolate was conserved following one or two passages in Tg ( M109 ) mice . For the human inocula , the “synaptic” pattern of PrPSc deposition observed with the sCJD ( MM1 ) subtype and the plaque-like deposition of PrPSc commonly observed with sCJD ( VV2 ) were both recapitulated in Tg ( M109 ) mice ( Figure 2A , C ) . PrPSc plaques were observed in the vicinity of vacuolation in vCJD-inoculated Tg ( M109 ) mice ( Figure 2E ) , which is somewhat reminiscent of the “florid” PrPSc plaques present in the brains of vCJD patients [36] . Notably , the presence of florid plaques in vCJD-inoculated animals is species-dependent and their absence does not necessarily imply lack of strain fidelity [37] , [38] . The neuropathological signature of RML prions in mice is the diffuse deposition of PrPSc in the hippocampus; this pattern of PrPSc deposition was also observed in RML-inoculated Tg ( M109 ) mice ( Figure 2O ) . Similarly , plaque-like PrPSc aggregates in the corpus callosum , which is the hallmark of the Sc237 strain , were observed in Sc237-inoculated Tg ( M109 ) mice ( Figure 2M ) , and the thalamic plaque-like PrPSc deposits in CWD-inoculated Tg ( M109 ) mice ( Figure 2I ) resembled those present in CWD-inoculated Tg mice expressing elk PrP [39] . We conclude that for the majority of prion isolates , the neuropathological signatures of PrPSc deposition were maintained upon transmission to Tg ( M109 ) mice . Additionally , the pattern of cerebral PrPSc deposition on first passage was indistinguishable for each isolate when compared to the second passage in Tg ( M109 ) mice ( Figure 2 , compare left and right columns ) . The conformational stability of PrPSc molecules , which is a measure of their ability to resist denaturation by guanidine hydrochloride ( GdnHCl ) [40] , was used to characterize the prion strains transmitted to Tg ( M109 ) mice . We performed conformational stability assays on the original inocula and after serial transmission through Tg ( M109 ) mice by titrating the stability of protease-resistant PrPSc using GdnHCl denaturation ( Figure 3 ) . Before and after two passages in Tg ( M109 ) mice , sCJD ( MM1 ) , vCJD , BSE , and Sc237 prions exhibited GdnHCl1/2 values of ∼2 M ( Figure 3A , C , D , F ) while sCJD ( VV2 ) prions had GdnHCl1/2 values of ∼2 . 8 M ( Figure 3B ) . CWD prions , either before or after passaging in Tg ( M109 ) mice , were intermediate with GdnHCl1/2 values of 2 . 4 M ( Figure 3E ) , whereas RML prions exhibited the lowest conformational stability of ∼1 . 5 M , which was unchanged upon propagation in Tg ( M109 ) mice ( Figure 3G ) . These findings argue that the conformations of these seven prion isolates were unaltered upon serial passage in Tg ( M109 ) mice .
As a fourth test to assess the fidelity of prion strain replication upon passage in Tg ( M109 ) mice , we performed retrotransmission experiments for the sCJD ( MM1 ) , CWD , Sc237 , RML , and 301V ( A ) isolates . In these experiments , Tg ( M109 ) -passaged prions were reintroduced into Tg mice expressing the PrP sequence of the species from which the prion isolate was originally derived . Inoculation of Tg ( HuPrP ) mice with Tg ( M109 ) -passaged sCJD ( MM1 ) prions , Tg ( SHaPrP ) mice with Tg ( M109 ) -passaged Sc237 prions , and Tg ( MoPrP ) mice with Tg ( M109 ) -passaged RML or 301V ( A ) prions resulted in clinical signs of prion disease in all of the inoculated animals ( Table 2 ) . In contrast , none of the Tg mice expressing elk PrP developed signs of neurologic illness following challenge with Tg ( M109 ) -passaged CWD prions , suggesting that a substantial species barrier exists when attempting to convert elk PrPC using bank vole PrPSc . For the experiments in which successful retrotransmission was achieved , the PK-resistant PrPSc in ill recipient mice was identical to that of the original isolate passaged into the same respective Tg line , as judged by the electrophoretic mobilities and relative glycoform ratios ( Figure 4A–D ) . Furthermore , the patterns of cerebral PrPSc deposition from the original isolate were recapitulated following retrotransmission ( Figure 4E–J ) . Based on the conservation of biochemical , neuropathological , and conformational properties of the prion isolates upon transmission to Tg ( M109 ) mice and upon retrotransmission after passage into Tg ( M109 ) mice , we conclude that prion strain fidelity was often maintained upon transmission to Tg ( M109 ) mice .
We inoculated Tg ( BVPrP , I109 ) 3574 mice , denoted Tg ( I109 ) , with 7 prion isolates from 5 different species: sCJD ( MM1 ) [2 human cases and 1 case passaged in Tg ( HuPrP ) mice] , CWD ( elk ) , Sc237 ( hamster ) , and RML ( mouse and MV-passaged ) . Hemizygous Tg ( I109 ) mice express PrP at ∼4 times the level of PrP expression found in wt mice and developed spontaneous signs of neurological dysfunction at a mean age of ∼340 days [33] . Similar to the results obtained in Tg ( M109 ) mice , all inoculated Tg ( I109 ) mice developed signs of progressive neurologic dysfunction ( Figure 5A ) , with mean incubation periods ranging from ∼50 days for MV-passaged RML prions to ∼260 days for each of the 3 sCJD ( MM1 ) isolates ( Table 3 ) . The mean incubation periods were slightly longer in Tg ( I109 ) mice than in Tg ( M109 ) mice on first passage of these isolates , which was likely due to the lower level of PrP expression in the Tg ( I109 ) line . PK-resistant PrPSc ( Figure 5B ) , vacuolation ( Figure S3A–E ) , astrocytic gliosis ( Figure S3F–J ) , and cerebral PrPSc deposition ( Figure S3K–O ) were observed in the brains of the ill Tg ( I109 ) mice , which confirmed the diagnosis of prion disease . The ages at which the sCJD ( MM1 ) - and CWD-inoculated Tg ( I109 ) mice developed neurologic disease partially overlapped with the onset of spontaneous illness in this line ( Figure 5A ) . However , we could distinguish the spontaneous disease phenotype from the inoculated disease because the spontaneously ill animals did not exhibit PrP 27–30 in their brains [33] . Thus , any inoculated animal that developed signs of neurologic illness but lacked detectable levels of PrP 27–30 in its brain was excluded from the study . Importantly , only four such mice were found , and the vast majority of inoculated animals ( 49 of 53 ) exhibited PrP 27–30 in their brains ( Figure 5C–H ) . We next sought to determine whether the biochemical and neuropathological properties of the various prion isolates were conserved upon transmission to Tg ( I109 ) mice . Tg ( I109 ) mice inoculated with two cases of sCJD ( MM1 ) prions or with sCJD ( MM1 ) prions previously passaged in Tg ( HuPrP ) mice exhibited considerable prion strain diversity among individual animals ( Figure 5C–E ) . Whereas some of the animals exhibited type 1 PrPSc similar to that observed in Tg ( M109 ) mice ( Figure 5C , lanes “a” and “e” ) , others displayed a type 2 pattern ( Figure 5C , lane “b” ) or even a mixed type 1/type 2 phenotype ( Figure 5C , lanes “c” and “d” ) . Similarly , passage of CWD into Tg ( I109 ) mice resulted in 6 of 7 animals harboring PK-resistant PrPSc similar to that observed in CWD-inoculated Tg ( M109 ) mice ( Figure 5F , lane “a” ) ; one Tg ( I109 ) mouse showed PrPSc of slower electrophoretic mobility ( Figure 5F , lane “b” ) . In contrast to replication in Tg ( I109 ) mice , neither sCJD ( MM1 ) nor CWD prions underwent any detectable biochemical changes in PrPSc during multiplication in Tg ( M109 ) mice ( Figure S4 ) . The PK-resistant PrPSc present in the brains of RML- and Sc237-inoculated Tg ( I109 ) mice were similar to the those observed in Tg ( M109 ) mice inoculated with the same isolates ( Figure 5G , H ) , as judged by glycoform ratios and type 1 electrophoretic mobility . The patterns of cerebral PrPSc deposition in the brains of sCJD ( MM1 ) - , Sc237- , and RML-inoculated Tg ( I109 ) mice were similar to those observed in Tg ( M109 ) mice inoculated with the same prion isolates ( compare Figure S3K , M , N with Figure 2A , M , O ) . In contrast , a Tg ( I109 ) mouse infected with CWD ( also shown in Figure 5F , lane “a” ) harbored small amounts of diffuse PrPSc in the thalamus ( Figure S3L ) whereas CWD-inoculated Tg ( M109 ) mice had large plaque-like deposits of PrPSc ( Figure 2I ) . Collectively , these results argue that passage of sCJD ( MM1 ) and CWD prions through Tg ( I109 ) mice resulted in alterations to these prion strains .
The study of human prions in mice has been hindered traditionally by long incubation periods . For example , sCJD ( MM1 ) prions transmit poorly to wt mice; only a few inoculated mice ever develop prion disease and those that do exhibit incubation times of 600 days or more [12] . In Tg ( HuPrP ) mice , the incubation periods were ∼160 days for sCJD ( MM1 ) prions [45] and ∼700 days for vCJD prions [18] . Reductions in the incubation times were achieved when human-specific residues in PrP were reverted to those of the mouse . For instance , in Tg mice expressing a chimeric human/mouse PrP containing 7 human residues , the incubation times for sCJD ( MM1 ) and vCJD prions were ∼110 and ∼360 days , respectively [45] . Reversion of an additional human-specific residue to its mouse equivalent in Tg1014 mice further reduced the incubation periods to ∼80 days for sCJD ( MM1 ) prions and ∼200 days for vCJD prions , but a change in strain type was apparent in some vCJD-inoculated animals [21] . Incubation times of ∼200 days on first passage and ∼175 days on serial passage for sCJD ( MM1 ) prions were substantially longer in the Tg ( M109 ) mice compared to Tg1014 mice . Notably , vCJD prions transmitted disease in ∼40 days on second passage in Tg ( M109 ) mice , and the fidelity of the vCJD strain was maintained . To the best of our knowledge , this is the most rapid human prion strain isolated to date . The incubation periods for the MM2 and VV2 subtypes of sCJD prions upon serial passage in Tg ( M109 ) mice were also considerably more rapid than those observed in mice expressing human PrP or chimeric human/mouse PrP [45] , [46] , [47] . We speculate that Tg ( M109 ) mice inoculated with BVPrP-adapted sCJD or vCJD prions may constitute an excellent system for performing initial assessments of the in vivo efficacy of candidate CJD therapeutics , although weak therapeutic effects may be harder to discern in mice with such rapid incubation periods and positive results would need to be confirmed in Tg mice expressing human PrP . Based on the studies with chimeric human/mouse PrP described above , constructing chimeric human/bank vole PrP transgenes may lead to even shorter incubation times for CJD prions . The rapid incubation periods and apparent strain fidelity observed for most prion isolates upon serial passage in Tg ( M109 ) mice should greatly facilitate the study of the biochemical and structural basis of prion strains . For instance , the incubation periods for BSE prions in Tg mice expressing bovine PrP is ∼250 days [19] , but merely ∼60 days upon second passage in Tg ( M109 ) mice . Thus , Tg ( M109 ) mice may be useful for rapidly producing BSE prions for structural studies . Tg ( M109 ) mice should also facilitate accurate comparisons between various prion strains or isolates . For instance , there has been considerable debate as to whether the conformational stability of a given prion strain is related to its incubation period . Although some of us ( S . J . D . and S . B . P . ) as well as others found that there was a direct correlation between conformational stability and incubation period , with less stable strains propagating more rapidly [48] , [49] , another study found the opposite , namely that strains with short incubation periods exhibited higher conformational stabilities [50] . We did not observe a definitive relationship between conformational stability and incubation period for seven different prion isolates serially propagated in Tg ( M109 ) mice . One caveat of this conclusion is that we did not include any synthetic or anchorless prion strains in our study , which exhibit the highest conformational stabilities [48] , [49] , [51] , [52] . Although Tg ( M109 ) mice developed signs of neurologic illness following challenge with a diverse range of prion isolates , for many of the strains tested , levels of PK-resistant PrPSc , cerebral PrPSc deposition , and vacuolation were lower than those generally found in prion-infected rodents . Similarly low levels of PrPSc were reported in I109 bank voles inoculated with CWD prions [27] . Several explanations seem plausible: one possibility might be that BVPrPSc replicates in a few critical regions in the CNS that produce progressive neurological deficits before widespread accumulation of BVPrPSc occurs [53] , [54] . A second possibility is that the amino acid sequence of BVPrP favors protease-sensitive conformations more readily than most other PrPs , similar to the predominance of protease-sensitive prions in the brains of CJD patients [55] . A third possible explanation is that during prion replication , BVPrPSc may exhibit a greater propensity for generating highly neurotoxic PrP conformers , such as the hypothetical PrPL entity [22] , [56] , compared to PrPs from other species . The rapid production of highly neurotoxic but PK-sensitive BVPrPSc conformers may be sufficient to elicit signs of neurological deficits prior to the extensive accumulation of PK-resistant PrPSc in the brain . Indeed , Tg ( I109 ) mice developed spontaneous signs of neurologic disease and prion-specified neuropathological changes in the absence of detectable levels of PrP 27–30 [33] , suggesting that BVPrP may be inherently prone to adopting neurotoxic conformations . Although BVPrP is overexpressed in the brains of Tg ( BVPrP ) mice , protein overexpression is insufficient to explain the general susceptibility of these mice to prions because bank voles , which express physiological levels of BVPrP , are also highly susceptible to a diverse range of prion isolates [24] , [25] , [27] . Therefore , an important unanswered question is what structural feature of BVPrPC makes it so susceptible to forming PrPSc when exposed to PrPSc molecules from many other species ? Because the mature forms of BVPrP and MoPrP differ at only eight positions [25] , our results argue that at most , eight residues in PrP mediate this phenomenon . At these eight positions , six of the BVPrP residues are also found in the sequence of hamster PrP ( Figure S5 ) . Because hamsters do not exhibit a bank vole–like general susceptibility to prions [57] , [58] , it seems reasonable to speculate that the other two residues ( Glu227 and Ser230 ) in BVPrP may play an important role in its unique behavior , especially because Glu227 is not found in other mammalian PrPs ( Figure S5 ) . Indeed , these two residues are located near the C-terminal end of the protein , in proximity to the GPI anchor attachment site . Although it is unclear how these residues influence the behavior of BVPrP , two possibilities include perturbation of PrP shedding from the membrane by ADAM proteases [59] and modulation of the interaction of BVPrP with other proteins or membrane lipids [32] , [60] , [61] . Although C-terminal residues in BVPrP may contribute to its unique properties , other BVPrP residues , either alone or in combination , may also be important . For example , unlike mouse PrPC , the structure of BVPrPC includes a so-called “rigid loop” in the region connecting β-strand 2 to α-helix 2 [34] . Tg mice expressing either a chimeric elk/mouse PrP , a chimeric horse/mouse PrP , or the I109 variant of BVPrP , all of which contain a rigid loop , develop a spontaneous neurologic illness reminiscent of prion disease [33] , [62] , [63] , suggesting that the presence of a rigid loop may render PrP more prone to misfolding . However , although the existence of a rigid loop in the structure of PrPC can modulate the interspecies transmission of prions in some instances [64] , it does not in other cases [65] . Thus , while the rigid loop in BVPrP may contribute to its unique promiscuity for diverse prion strains , it is unlikely to be the sole factor . Although the mechanism by which BVPrPC seems to act as a “universal acceptor” of prions is unknown , the structure of BVPrPC might permit it to bind promiscuously to PrPSc molecules from many different species , enabling prion replication . An alternate explanation is that a misfolding intermediate on the pathway to PrPSc formation is more readily populated or is stabilized by the BVPrP sequence [66] , [67] . Such a replication intermediate may be partially unfolded and thus exhibit a lower energy barrier to conversion by PrPSc from different species . This hypothesis would also explain the increased propensity for BVPrP to spontaneously adopt an infectious , neurotoxic conformation [33] . The mechanism by which some prion isolates , such as sCJD ( MM1 ) and CWD , underwent changes upon propagation in Tg ( I109 ) mice remains enigmatic . One possibility is that the simultaneous presence of injected prions and spontaneously formed prions in Tg ( I109 ) mice could alter the properties of the inoculated prion isolates , because the incubation periods for CWD and sCJD ( MM1 ) prions overlapped substantially with the occurrence of spontaneous disease in this line ( Figure 5A ) . However , this explanation seems unlikely for CWD prions because their properties were also clearly altered upon serial propagation in I109 bank voles [27] , which do not develop spontaneous neurologic disease . A second possibility is that the natural CWD and sCJD ( MM1 ) isolates used in transmission experiments are not homogeneous and that less abundant conformers present in the inocula may preferentially propagate in Tg ( I109 ) mice . Mixtures of strains have been described in both sCJD patients and in CWD-infected cervids [68] , [69] , [70] . In this scenario , these substrains fail to emerge as the dominant species in Tg ( M109 ) mice or in Tg mice expressing homotypic PrP due to prion strain interference effects . Indeed , there are documented examples in which the replication of a faster but less abundant prion strain is suppressed by the presence of a slower , but more abundant strain [71] , [72] . Thus , the presence of isoleucine at codon 109 of BVPrP may hinder prion interference effects , allowing less abundant but more rapid strains to gradually emerge upon serial passage . Still another hypothesis is that prion strains are actually “quasi-species” that are composed of a collection of substrains that can interconvert [73] . The energy landscape of prion replication may be very different for BVPrP ( I109 ) , allowing substrains that are not densely populated in the original host to emerge . The extraordinary promiscuity of BVPrP demonstrates that a small number of amino acid differences in PrP can profoundly alter the properties of prions . It is interesting to consider whether BVPrP-like versions of other aggregation-prone proteins may exist in certain species . With the recent convergence of scientific evidence that many , if not most , neurodegenerative diseases are caused by proteins that become prions [74] , [75] , the identification of organisms expressing Aβ , tau , or α-synuclein proteins that exhibit an increased propensity to misfold may facilitate studies on the transmissibility of Alzheimer's disease and Parkinson's disease .
All mouse studies were carried out in accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals ( Institute of Laboratory Animal Resources , National Academies Press , Washington , DC ) ; protocols were reviewed and approved by the UCSF Institutional Animal Care and Use Committee: “Breeding colony and production of transgenic rats and mice” ( AN084871 ) and “Incubation periods of prion and other neurodegenerative diseases” ( AN084950 ) . Hemizygous Tg ( BVPrP , M109 ) 22019 [“Tg ( M109 ) ”] , Tg ( BVPrP , M109 ) 3118 mice , and Tg ( BVPrP , I109 ) 3574 [“Tg ( I109 ) ”] mice express BVPrP under the control of the hamster PrP promoter [33] and were maintained by backcrossing to FVB mice lacking murine PrP expression ( Prnp0/0 mice ) [76] . Tg ( SHaPrP ) 7 mice that express hamster PrP [16] , Tg ( ElkPrP ) 12584 mice expressing elk PrP [39] , and Tg ( HuPrP ) 2669 mice expressing human PrP containing the M129 polymorphism [77] were also maintained on a Prnp0/0 background . Tg ( MoPrP ) 4053 mice expressing the PrP-A allotype of mouse PrP [78] were maintained on a wild-type ( Prnp+/+ ) background . The following prion isolates were used in this study: mouse-adapted scrapie strain RML ( maintained in wild-type CD-1 mice expressing the PrP-A allotype ) ; hamster-adapted scrapie strain Sc237; MV-passaged RML or Sc237 prions [79]; mouse-adapted BSE strain 301V ( passaged in mice expressing either PrP-A or PrP-B ) ; SSBP/1 sheep scrapie prions derived from a pool of scrapie-infected sheep brains , which were a generous gift from Dr . Nora Hunter; CWD prions derived from the brain of a naturally infected elk [Elk1 isolate; [39]]; BSE prions derived from the brain of a naturally-infected cow and then passaged 4 times in Tg mice expressing bovine PrP; human sCJD prions obtained from the brains of patients exhibiting either the MM1 , MM2 , or VV2 disease subtypes; sCJD ( MM1 ) prions that were passaged in either Tg ( HuPrP ) mice or in guinea pigs [80]; and human prions obtained from the brain of a variant CJD patient , provided by the UK National CJD Surveillance Unit . Brain homogenates [10% ( wt/vol ) in calcium- and magnesium-free PBS] were diluted to 1% ( wt/vol ) using 5% bovine serum albumin ( BSA ) . Weanling mice ( ∼2-month-old ) were anesthetized with isoflurane and then inoculated with 30 µL of the 1% brain homogenate into the right parietal lobe using a 27-gauge syringe . Inoculated animals were assessed daily for routine health and checked three times weekly for the presence of signs of neurologic illness . Mice were euthanized once two or more neurologic signs were apparent , using the standard diagnostic criteria for assessing prion disease in mice [81] . Brains were then removed , and either snap-frozen on dry ice and then stored at −80°C for biochemical analyses or fixed in 10% buffered formalin for neuropathological studies . Ten percent ( wt/vol ) brain homogenates in calcium- and magnesium-free PBS were generated using either an OmniTip ( Omni International ) with a PowerGen homogenizer ( Fisher Scientific ) or with a bead beater ( Precellys ) . Nine volumes of 10% brain homogenate were added to one volume of 10× detergent buffer [5% ( vol/vol ) NP-40 , 5% ( wt/vol ) sodium deoxycholate in PBS] and then incubated on ice for 20 min followed by centrifugation at 1 , 000 × g for 5 min to remove cellular debris . Protein concentrations in the supernatant were then determined using the BCA assay ( Thermo Scientific ) . One mg of detergent-extracted protein was diluted to a final volume of 398 µL using 1× detergent buffer [0 . 5% ( v/v ) NP-40 , 0 . 5% ( w/v ) sodium deoxycholate in PBS . ] Two µL of a 10 mg/mL PK stock solution ( Fermentas ) was then added to samples to be digested , resulting in a final PK concentration of 50 µg/mL ( a PK:protein ratio of 1∶50 ) . Samples were then incubated at 37°C with vigorous shaking for 1 h . PK digestions were terminated by the addition of phenylmethylsulfonyl fluoride ( PMSF ) to a final concentration of 2 mM . One hundred µL of a 10% ( vol/vol ) solution of sarkosyl was then added to bring the final sarkosyl concentration to 2% . Samples were then ultracentrifuged at 100 , 000× g for 1 h at 4°C , and the supernatants removed by aspiration . Pellets were resuspended in 1× NuPAGE loading buffer ( Life Technologies ) containing 2 . 5% ( vol/vol ) β-mercaptoethanol by vortexing , boiled for 10 min , and then analyzed by immunoblotting . PK-digested brain homogenate samples ( containing 200–500 µg of digested total protein ) were prepared as described above and then loaded onto 10% NuPAGE gels ( Life Technologies ) . Undigested samples ( typically 10 µg total protein ) were prepared by diluting detergent-extracted brain homogenate directly into 1× NuPAGE loading buffer containing β-mercaptoethanol and then boiling for 5 min . SDS-PAGE was performed using the MES buffer system , and gels were subsequently transferred to PVDF membranes using a wet blotting system . Membranes were blocked for 2 h at room temperature using blocking buffer [5% ( w/v ) nonfat milk in Tris-buffered saline containing 0 . 05% ( v/v ) Tween-20 ( TBST ) ] and then incubated with horseradish peroxidase ( HRP ) -conjugated primary antibody overnight at 4°C . Blots were washed three times with TBST , developed using the enhanced chemiluminescent detection system ( GE Healthcare ) and then exposed to x-ray film . PrP was detected using the antibody HuM-P [82] . Twenty µL of detergent-extracted brain homogenate was mixed with 2× stocks of GdnHCl to give final concentrations of 1 , 1 . 5 , 2 , 2 . 5 , 3 , 3 . 5 , or 4 M GdnHCl . For the 4 . 5- and 5-M samples , only 10 µL of brain homogenate was used . Samples were incubated at 22°C with shaking ( 800 rpm ) for 2 h and then diluted to 0 . 4 M GdnHCl in 1× detergent buffer . PK was added to a final concentration of 20 µg/mL , and the samples were digested at 37°C with shaking for 1 h . Digestions were then terminated by adding PMSF to a final concentration of 2 mM . One hundred µL of a 12% ( vol/vol ) sarkosyl solution was then added to give a final concentration of 2% . Samples were then ultracentrifuged at 100 , 000× g for 1 h at 4°C , and the supernatants removed by gentle aspiration . Pellets were resuspended in 1× NuPAGE loading buffer containing β-mercaptoethanol , boiled for 10 min , and then analyzed by immunoblotting as described above . Films were scanned using a CCD camera ( FluorChem 880; Alpha Innotech ) and then densitometry performed using Image J . GdnHCl1/2 values were calculated using the variable slope ( four parameter ) function in Prism 5 . Brains were removed , immersion-fixed in 10% buffered formalin , and then embedded in paraffin . Sections were cut at 8 µm , mounted on glass slides , deparaffinized , and then processed for immunohistochemistry or stained with hematoxylin and eosin ( H&E ) . Endogenous tissue peroxidases were inhibited by incubating the slides in a 3% hydrogen peroxide solution ( prepared in methanol ) for 30 min . Sections to be stained with anti-PrP antibodies were subjected to hydrolytic autoclaving ( 121°C for 10 min in citrate buffer ) . Slides were then blocked with 10% ( vol/vol ) normal goat serum for 1 h and then incubated with primary antibody overnight at 4°C . The following primary antibodies were used: anti-GFAP rabbit polyclonal antibody Z0334 ( Dako , 1∶500 dilution ) to detect astrocytic gliosis , and anti-PrP antibodies 3F4 ( 1∶1 , 000 dilution ) [83] or HuM-D18 ( 1∶500 dilution ) [84] to detect PrPSc deposition . Bound antibody was detected using a Vectastain ABC peroxidase kit ( Vector Laboratories ) and visualized using 3-3′-diaminobenzidine ( DAB ) . Slides were counterstained with hematoxylin and then photographed using an AxioImager . A1 microscope ( Carl Zeiss ) . | Prions are infectious proteins that cause devastating neurodegenerative diseases in both humans and animals . Unlike other rodents , bank voles are highly susceptible to prions from many different species , suggesting that bank voles do not impose a “species barrier , ” which normally restricts the transmission of prions from one species to another . We were curious as to whether the unprecedented promiscuity of bank voles for prions is due to the specific prion protein sequence expressed , or to some other factor inherent to bank vole physiology . To answer this question , we inoculated transgenic mice that express bank vole prion protein [Tg ( BVPrP ) mice] with a diverse set of prions deriving from eight different species . Like bank voles , Tg ( BVPrP ) mice were highly susceptible to prions from all species tested , demonstrating that the BVPrP sequence mediates the enhanced susceptibility of bank voles to prions . Because the amino acid sequences of mouse and BVPrP differ at only eight positions , our results demonstrate that alterations to a small subset of residues within PrP can have a profound effect on the susceptibility of an organism to prions from another species . | [
"Abstract",
"Introduction",
"Results",
"Retrotransmission",
"of",
"prion",
"strains",
"Transmission",
"of",
"prions",
"to",
"Tg(I109)",
"mice",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"veterinary",
"prion",
"diseases",
"infectious",
"diseases",
"veterinary",
"diseases",
"genetically",
"modified",
"organisms",
"zoonoses",
"medicine",
"and",
"health",
"sciences",
"transgenic",
"animals",
"biology",
"and",
"life",
"sciences",
"agricultural",
"biotechnolog... | 2014 | Evidence That Bank Vole PrP Is a Universal Acceptor for Prions |
Three-dimensional microscopy is increasingly prevalent in biology due to the development of techniques such as multiphoton , spinning disk confocal , and light sheet fluorescence microscopies . These methods enable unprecedented studies of life at the microscale , but bring with them larger and more complex datasets . New image processing techniques are therefore called for to analyze the resulting images in an accurate and efficient manner . Convolutional neural networks are becoming the standard for classification of objects within images due to their accuracy and generalizability compared to traditional techniques . Their application to data derived from 3D imaging , however , is relatively new and has mostly been in areas of magnetic resonance imaging and computer tomography . It remains unclear , for images of discrete cells in variable backgrounds as are commonly encountered in fluorescence microscopy , whether convolutional neural networks provide sufficient performance to warrant their adoption , especially given the challenges of human comprehension of their classification criteria and their requirements of large training datasets . We therefore applied a 3D convolutional neural network to distinguish bacteria and non-bacterial objects in 3D light sheet fluorescence microscopy images of larval zebrafish intestines . We find that the neural network is as accurate as human experts , outperforms random forest and support vector machine classifiers , and generalizes well to a different bacterial species through the use of transfer learning . We also discuss network design considerations , and describe the dependence of accuracy on dataset size and data augmentation . We provide source code , labeled data , and descriptions of our analysis pipeline to facilitate adoption of convolutional neural network analysis for three-dimensional microscopy data .
The continued development and widespread adoption of three-dimensional microscopy methods enables insightful observations into the structure and time-evolution of living systems . Techniques such as confocal microscopy [1 , 2] , two-photon excitation microscopy [3–6] , and light sheet fluorescence microscopy [6–12] have provided insights into neural activity , embryonic morphogenesis , plant root growth , gut bacterial competition , and more . Extracting quantitative information from biological image data often calls for identification of objects such as cells , organs , or organelles in an array of pixels , a task that can especially challenging for three-dimensional datasets from live imaging due to their large size and potentially complex backgrounds . Aberrations and scattering in deep tissue can , for example , introduce noise and distortions , and live animals often contain autofluorescent biomaterials that complicate the discrimination of labeled features of interest . Moreover , traditional image processing techniques tend to require considerable manual curation , as well as user input regarding which features , such as cell size , homogeneity , or aspect ratio , should guide and parameterize analysis algorithms . These features may be difficult to know a priori , and need not be the features that lead to the greatest classification accuracy . As data grow in both size and complexity , and as imaging methods are applied to an ever-greater variety of systems , standard approaches become increasingly unwieldy , motivating work on better computational methods . Machine learning methods , in particular convolutional neural networks ( ConvNets ) , are increasingly used in many fields and have achieved unprecedented accuracies in image classification tasks [13–16] . The objective of supervised machine learning is to use a labeled dataset to train a computer to make classifications or predictions given new , unlabeled data . Traditional feature-based machine learning algorithms , such as support vector machines and random forests , make use of manually determined characteristics , which in the context of image data could be the eccentricity of objects , their size , their median pixel intensity , etc . The first stages in the implementation of these algorithms , therefore , are the identification of objects by image segmentation methods and the calculation of the desired feature values . In contrast , convolutional neural networks use the raw pixel values as inputs , eliminating the need for determination of object features by the user . Convolutional neural networks use layers consisting of multiple kernels , numerical arrays acting as filters , which are convolved across the input taking advantage of locally correlated information . These kernels are updated as the algorithm is fed labeled data , converging by numerical optimization methods on the weights that best match the training data . ConvNets can contain hundreds of kernels over tens or hundreds of layers which leads to hundreds of thousands of parameters to be learned , requiring considerable computation and , importantly , large labeled datasets to constrain the parameters . Over the past decade , the use of ConvNets has been enabled by advances in GPU technology , the availability of large labeled datasets in many fields , and user-friendly deep learning software such as TensorFlow [17] , Theano [18] , Keras [19] , and Torch [20] . In addition to high accuracy , ConvNets tend to have fast classification speeds compared to traditional image processing methods . There are drawbacks , however , to neural network approaches . As noted , they require large amounts of manually labeled data for training the network . Furthermore , their selection criteria , in other words the meanings of the kernels’ parameters , are not easily understandable by humans [21] . There have been several notable examples of machine learning methods applied to biological optical microscopy data [22 , 23] , including bacterial identification from 2D images using deep learning [24] , pixel-level image segmentation using deep learning [25–27] , subcellular protein classification [28] , detection of structures within C . elegans from 2D projections of 3D image stacks using support vector machines [29] , and more [30–34] . Nonetheless , it is unclear whether ConvNet approaches are successful for thick , three-dimensional microscopy datasets , whether their potentially greater accuracy outweighs the drawbacks noted above , and what design principles should guide the implementation of ConvNets for 3D microscopy data . To address these issues , we applied a deep convolutional neural network to analyze three-dimensional light sheet fluorescence microscopy datasets of gut bacteria in larval zebrafish ( Fig 1a and 1b ) and compared its performance to that of other methods . These image sets , in addition to representing a major research focus of our lab related to the aim of understanding the structure and dynamics of gut microbial communities [10 , 35–37] , serve as exemplars of the large , complex data types increasingly enabled by new imaging methods . Each 3D image occupies roughly 5 GB of storage space and consists of approximately 300 slices separated by 1 micron , each slice consisting of 6000 x 2000 pixel 2D images ( 975x325 microns ) . These images include discrete bacterial cells , strong and variable autofluorescence from the mucus-rich intestinal interior [38] , autofluorescent zebrafish cells , inhomogeneous illumination due to shadowing of the light sheet by pigment cells , and noise of various sorts . The bacteria examined here exist predominantly as discrete , planktonic individuals . Other species in the zebrafish gut exhibit pronounced aggregation; identification of aggregates is outside the scope of this work , though we note that the segmentation of aggregates is much less challenging than identification of discrete bacterial cells , due to their overall brightness and size . The goal of the analysis described here is to correctly classify regions of high intensity as bacteria or as non-bacterial objects . Using multiple testing image sets , we compared the performance of the convolutional neural network to that of humans as well as random forest and support vector machine classifiers . In brief , the ConvNet’s accuracy is similar to that of humans , and it outperforms the other machine classifiers in both accuracy and speed across all tested datasets . In addition , the ConvNet performs well when applied to planktonic bacteria of a different genus through the use of transfer learning . Transfer learning has been shown to be effective in biological image data in which partial transference of network weights from 2D images dramatically lowers the amount of new labeled data that is required [15 , 28 , 31 , 39] . We explored the impacts on the ConvNet’s performance of network structure , the degree of data augmentation using rotations and reflections of the input data , and the size of the training data set , providing insights that will facilitate the use of ConvNets in other biological imaging contexts . Analysis code as well as all ∼ 21 , 000 manually labeled 3D image regions-of-interest are provided; see Methods for details and urls to data locations .
The image data we sought to classify consist of three-dimensional arrays of pixels obtained from light sheet fluorescence microscopy of bacteria in the intestines of larval zebrafish [10 , 35–37] . Fig 1B shows a typical optical section from an initially germ-free larval zebrafish , colonized by a single labeled bacterial species made up of discrete , planktonic individuals expressing green fluorescent protein; a three-dimensional scan is provided as Supplementary Movie 1 . All the data assessed here were derived from fish that were reared germ free ( devoid of any microbes ) [40] and then either mono-associated with a commensal bacterial species or left germ free . Nine scans are of fish mono-associated with the commensal species ZWU0020 of the genus Vibrio [10 , 41 , 42] , two scans are of fish in which the zebrafish remained germ-free , and a single scan is from a fish mono-associated with Pseudomonas ZWU0006 [36] . For each 3D scan , we first determined the intestinal space of the zebrafish using simple thresholding and detected bright objects ( “blobs” ) using a difference of Gaussians method described further in Methods . From each blob , we extracted 28x28x8 pixel arrays ( 4 . 5x4 . 5x8 microns ) , which served as the input data to the neural network , to be classified as bacterial or non-bacterial . Since there is no way to obtain ground truth values for bacterial identity in images , we manually classified blobs to serve as the training data for the neural network , using our expertise derived from considerable prior work on three dimensional bacterial imaging . Notably , in prior work we showed that the total bacterial abundance determined by manually corroborated feature-based bacterial identification from light sheet data corresponds well with the total bacterial abundance as measured through gut dissection and serial plating assays [35] . In Fig 1C–1F we show representative images of blobs corresponding to bacteria and noise . In order to estimate an upper bound on the classification accuracy we can expect from the learning algorithms , we chose a single image scan which we judged to be typical of a noisy , complex 3D image of the intestine of a larval zebrafish colonized by bacteria . We then had six lab members with least two years’ experience with light sheet microscopy of bacteria individually label each of the detected potential objects as either a bacterium or not . We show in Fig 2A the agreement between lab members . Excluding human 3 the agreement between any pair of humans is always above 0 . 87 . The outlier , human 3 , is the person with the least experience with the imaging data , namely the principal investigator . We next created a set of labeled data by manual classification of blobs from the 9 Vibrio scans and 2 scans of germ-free fish , consisting in total of over 20 , 000 objects . Including scans from germ-free fish is particularly important to enable accurate counting of low numbers of bacteria , which arise naturally due to extinction events [10] and population bottlenecks [41] . As detailed in Methods , we used Google’s open-source Tensorflow framework [17] to create , test , and implement 3D convolutional neural networks . Such networks have many design parameters and options , including the number , size , and type of layers , the kernel size , the downsizing of convolution output by pooling , and parameter regularization . In general , overly small networks can lack the complexity to characterize image data , though their limited parameter space is less likely to lead to overfitting . Conversely , larger networks can tackle more complex classification schemes , but demand more training data to constrain the large number of parameters , and also carry a greater computational load . In between these extremes , many design variations will typically give similar classification accuracy . We chose a simple architecture consisting of two convolutional layers followed by a fully connected layer . The first and second convolutional layers contain 16 and 32 5x5x2 kernels , respectively . Each layer is followed by 2x2x2 max pooling as further described in Methods . The final layer is a fully connected layer consisting of 1024 neurons with a dropout rate of 0 . 5 during training . After this , softmax regression is used for binary classification . We explored various alterations of our network architecture , and illustrate here the effect of simply varying the number of kernels per convolutional layer . We assessed the classification accuracy as a function of the number of kernels in layer 1 , with the number of kernels in layer 2 being double this . Accuracy was calculated using cross validation , training on all but one image dataset ( where an image dataset is a complete three-dimensional scan of the gut of one zebrafish ) , testing on the remaining image dataset , and repeating with different train/test combinations . The network accuracy initially increases with kernel number and plateaus at roughly 16 kernels , beyond which the variance in accuracy increases ( Fig 2B ) . Therefore , increasing the number of kernels beyond approximately 16 gives little or no improvement in accuracy at the expense of model complexity and increased variability . We note that there are many ways to alter network complexity , for example adding or removing layers , all of which may be interesting to investigate . Here , a rather small model consisting of two layers is sufficient to achieve human-level accuracy , suggesting that adding layers is unlikely to be useful . We trained the ConvNet using manually labeled data from eight of the Vibrio image datasets and the two datasets from germ-free fish ( devoid of gut bacteria ) and then tested it on the remaining manually labeled Vibrio image dataset that was used to assess inter-human variability , described above . The agreement between the neural network and humans ( mean ± std . dev . 0 . 89 ± 0 . 01 ) was indistinguishable from the inter-human agreement ( mean ± std . dev . 0 . 90 ± 0 . 02 ) , again excluding human 3 , indicating that the ConvNet achieves the practical maximum of bacterial classification accuracy ( Fig 2A ) . Examples of images for which all humans agreed on the classification , and in which there was disagreement , are provided in S1 Fig . To further test the network’s consistency across different imaging conditions we applied it separately to each of the 3D image datasets of larval zebrafish intestines . We also tested , with the same procedure and data , random forest and support vector machine classifiers to address the question of whether or not the ConvNet outperforms typical feature based learning algorithms . We first consider two experiment types: zebrafish intestines mono-associated with Vibrio ZWU0020 ( 9 image datasets , i . e . 9 complete three-dimensional scans from of different zebrafish ) and germ-free zebrafish ( 2 image datasets ) . Classifier accuracy for each Vibrio-colonized or empty-gut image scan was determined by cross-validation , training the network using all of the other image datasets , and testing on the dataset of interest . To test the variance in accuracy due to the training process , we performed three repetitions of each train/test combination using the same data . We found that the neural network outperforms the feature based algorithms on every image dataset ( Fig 3 ) , and also shows less variation in accuracy between the different datasets . The enhanced accuracy from the neural network is especially dramatic for germ-free datasets , for which it achieves over 90% accuracy , in contrast to less than 75% for feature based methods . For a given test dataset , the training variance for the convolutional neural network is small but nonzero , indicating that the network training algorithm finds similar , but not identical , minima with different ( random ) initializations on the same training data . It is also small for the random forest classifier . Interestingly , it is zero for the SVM classifier , indicating that given the same dataset , the algorithm is finding the same minimum . To further verify the robustness of our accuracy measures , we performed tests using a manually labeled image dataset that was completely distinct from those previously considered , and that therefore played no role in cross-validation or other prior work . This new test set consisted of 1302 images of bacteria ( 482 images ) or noise ( 840 images ) . We determined the classification accuracy of our convolutional neural network to be 89 . 3% , the support vector classifier to be 83 . 1% , and the random forest classifier to be 78 . 5% , in agreement with the prior assessments . The random forest , support vector machine , and neural network classifiers process roughly 300 , 400 , and 950 images per second , respectively; i . e . the neural network runs 2-3 times faster than the feature based learning algorithms on the same data . Convolutional Neural Networks famously require large amounts of training data which must often , as is the case here , be evaluated and curated by hand . To assess the scale of manual classification required for good algorithm performance , which is a key issue for future adoption of neural networks in biological image analysis , we explored the effect on the network’s accuracy of varying the amount of training data . We set aside 25% of the images from each of the Vibrio and germ-free fish image scans and trained the network using an increasing number of images from the remaining data . We increased the amount of training data in two different ways . First , we consecutively added to the training set all images from each image dataset excluding a subset of the images previously reserved for testing ( labeled “New datasets” in Fig 4A ) . Second , we randomly shuffled the training images from all the image scans , adding 1500 images to the training set over each iteration ( labeled “Train/test split” in Fig 4A ) . For the first method , enlargement of the training set corresponds to a greater amount of data as well as data from more diverse biological sources . For the second , data size increases but the biological variation sampled is held constant . In both cases , accuracy plateaus at a number of images on the order of 10 , 000 ( Fig 4A ) . The rise in accuracy with increasing training data size is only slightly more shallow with the first method , surprisingly , demonstrating that within-sample variation is sufficient to train the network . Data augmentation , the alteration of input images through mirror reflections , rotations , cropping , and the addition of noise , etc . , is commonly used in machine learning to enhance training dataset size and enable robust training of neural networks . To characterize the utility of data augmentation for 3D bacterial images , we focused in particular on image rotations and reflections , because the bacteria have no preferred orientation and hence augmentation by these methods creates realistic training images . We note that data augmentation is not necessary for feature based learning methods in which parity and rotational invariance can be built into the features used for classification . Obviously , augmented data is not independent of the actual training data , and so does not supply wholly new information . We were curious as to how including rotated and reflected versions of previously seen data compares , in terms of network performance , to adding entirely new data , a comparison that is useful if evaluating the necessity of performing additional imaging experiments . To test this , we compared the accuracies of the network when adding new data to that when adding rotated and reflected versions of existing data . We started with a fixed number of 1500 total objects randomly sampled from the entire set and , in the case of including new data , added another random 1500 objects at each iteration . For the augmented data , we applied random rotations and reflections to the original 1500 objects to iteratively increase the training size by 1500 objects . Each trained network was tested on the same test set of objects as that of Fig 4A . As shown in Fig 4B , the addition of new data leads to a plateau in accuracy of roughly 90% while for augmented data the plateau value is around 88% . This result demonstrates that , in the context of our network , simply augmenting existing data can raise classification accuracy to nearly the optimal level achieved by new , independent data . We assessed the accuracy of the convolutional neural network on images of discrete gut bacteria of another species , of the genus Pseudomonas . Training solely on the Vibrio images and testing on Pseudomonas gives ∼ 75% accuracy ( Fig 5 ) . However , this is much lower than the ∼ 85 − 95% accuracy obtained on Vibrio images ( Fig 4 ) ; the Pseudomonas species is not an exact morphological mimic of the Vibrio species . The Pseudomonas dataset is small ( 1190 images ) ; using 80% of its images for de novo neural network training gives ∼ 72% accuracy in identifying Pseudomonas in test datasets ( Fig 5 ) . We suspected that the general similarity of each species as rod-like , few-micron-long cells would allow transfer learning , in which a model trained for one task is used as the starting point for training for another task [43 , 44] . Using the network weights from training on Vibrio image datasets , as before , as the starting values for training on the small Pseudomonas dataset gives over 85% accuracy in classifying Pseudomonas ( Fig 5 ) .
We find that a 3D convolutional neural network for binary classification of bacteria and non-bacterial objects in 3D microscopy data of the larval zebrafish gut yields high accuracy without unreasonably large demands on the amount of manually curated training data . Specifically , the convolutional neural network obtains human-expert-level accuracy , runs 2-3 times faster than other standard machine learning methods , and is consistent across different datasets and across planktonic bacteria from two different genera through the use of transfer learning . It reaches these performance metrics after training on fewer than 10 , 000 human-classified images , which require approximately 20 person-hours of manual curation to generate . Moreover , augmented data in the form of rotations and reflections of real data contributes effectively to network training , further reducing the required manual labor . Experiments of the sort presented here typically involve many weeks of laboratory work . Neural network training , therefore , is a relatively small fraction of the total required time . In many biological imaging experiments , including our own , variety and similarity are both present . Multiple distinct species or cell types may exist , each different , but with some morphological similarities . It is therefore useful to ask whether such similarities can be exploited to constrain the demands of neural network training . The concept of transfer learning addresses this issue , and we find that applying it to our bacterial images achieves high accuracy despite small labeled datasets , an observation that we suspect will apply to many image-based studies . Transfer learning is a rapidly growing area of interest , with an increasing number of tools and methods available . There are likely many possibilities for further performance enhancements to network performance via transfer learning , beyond the scope of this study . One commonly used approach is to train initially on a large , publicly available , annotated dataset such as ImageNet . It is not likely that ImageNet’s set of two-dimensional images of commonplace objects will be better than actual 3D bacterial data for classifying 3D bacterial images . Nonetheless , it would be interesting to examine whether training using ImageNet or other standard datasets could establish primitive filters on which 3D convolutional neural networks could build . In addition , given the rapid growth of machine learning approaches in biology , it is likely that large , annotated datasets of particular relevance to tasks such as those described here will be developed , further enabling transfer learning . Though the data presented here came from a particular experimental system , consisting of fluorescently labeled bacterial species within a larval zebrafish intestine imaged with light sheet fluorescence microscopy , they exemplify general features of many contemporary three-dimensional live imaging applications , including large data size , high and variable backgrounds , optical aberrations , and morphological heterogeneity . As such , we suggest that the lessons and analysis tools provided here should be widely applicable to microbial communities [45] as well as eukaryotic multicellular organisms . We expect the use of convolutional neural networks in biological image analysis to become increasingly widespread due to the combination of efficacy , as illustrated here , and the existence of user-friendly tools , such as TensorFlow , that make their implementation straightforward . We can imagine several extensions of the work we have described . Considering gut bacteria in particular , extending neural network methods to handle bacterial aggregates is called for by observations of a continuum of planktonic and aggregated morphologies [36] . Considering 3D images more generally , we note that the approach illustrated has as its first step detection of candidate objects ( “blobs” ) , which requires choices of thresholding and filtering parameters . Alternatively , pixel-by-pixel segmentation is in principle possible using recently developed network architectures [13 , 46] , which could enable completely automated processing of 3D fluorescence images . In addition , pixel-based identification of overall morphology ( for example , the location of the zebrafish gut ) could further enhance classification accuracy , by incorporating anatomical information that constrains the possible locations of particular cell types .
Three-dimensional scans of the intestines of larval zebrafish , derived germ-free and colonized by fluorescently labeled bacteria prior to imaging , were obtained using light sheet fluorescence microscopy as described in Refs . [10 , 35 , 36] . All experiments involving zebrafish were carried out in accordance with protocols approved by the University of Oregon Institutional Animal Care and Use Committee . The microscope was based on the design from Keller et al [6] , and has been described elsewhere [35 , 45] . In brief: a laser is rapidly oscillated creating a thin sheet of light used to illuminate a section of the specimen , in this case , a larval zebrafish . An objective lens is seated perpendicular to the laser sheet , focusing two-dimensional images onto a sCMOS camera . The specimen is scanned through the sheet along the detection axis , thereby constructing a 3D image . The camera exposure time was 30 ms , and the laser power of the laser was 5 mW as measured between the theta-lens and excitation objective . Of the twelve image datasets used for this work , nine were of the zebrafish commensal bacterium Vibrio sp . ZWU0020 , one was of a Pseudomonas commensal sp . ZWU0006 , and two were from germ-free fish , devoid of any bacteria . An example 3D image dataset of the anterior “bulb” of one larval zebrafish gut is available at the link noted in the README . md file at github: https://github . com/rplab/Bacterial-Identification , together with the 6 lab members’ labels for each detected object in the volume , the convolutional neural network’s classification , and each of the extracted region-of-interest voxels . Other image sets are available upon request; for each zebrafish gut , the full image dataset is roughly 1 GB in size . Rough segmentation of the intestine was performed using histogram equalization of each individual z-stack followed by a moving average over 30 consecutive images in the z-stack followed by hard thresholding to create a binary mask that overestimated the size of the intestine . While extremely rough , this technique requires no manual editing or outlining . After this , blob detection was performed using the difference of Gaussians technique from the scikit-image library on each two-dimensional image , and the blobs were linked together across consecutive images in each stack . Regions 28x28x8 pixels in size centered at each detected blob were then saved to be labeled by hand as either a bacterium or noise . The code for extracting the regions of interest is publicly available on Github at https://github . com/rplab/Bacterial-Identification . From the 12 datasets , 20 , 929 images were hand labeled of which 38% were bacteria and 62% were noise . Hand labeling took roughly 1-2 hours per scan . All of the 28x28x8 pixel images and the corresponding labels are available from links in the README . md file at the Github repository https://github . com/rplab/Bacterial-Identification . All code for the project was written in Python . Over sixty features were created initially . These were assessed using scikit-learn’s feature_importances_ , from which the thirty one most helpful features were retained . The features used included geometric properties obtained by ellipse-fitting and texture-based characteristics; a detailed list is provided in the python code features . py provided on Github: https://github . com/rplab/Bacterial-Identification . The data were tested using both a random forest and support vector classifier from the scikit-learn library . The random forest used 500 estimators . The support vector classifier from sci-kit learn , sklearn . svm . SVC ( ) , was tested over a range of parameters and kernels using scikit-learn’s GridSearchCV which yielded highest accuracy when using a radial basis function kernel with penalty C = 1 . The 3D convolutional neural network was created using Google’s TensorFlow . Each input image was 28x28x8 pixels . The network consisted of two convolutional layers followed by a fully connected layer . The first layer was composed of 16 , 5x5x2 kernels of stride 2 and same padding followed by 2x2x2 max pooling , the second layer contained 32 5x5x2 kernels of the same stride and padding and was also followed by 2x2x2 max pooling . We chose to double the number of kernels after max pooling as in [47] . After the final convolutional layer we employed a fully connected layer consisting of 1024 neurons . The classes were then determined using a softmax layer . The network had a dropout of 0 . 5 , a learning rate of 0 . 0001 and the data was trained over 120 epochs randomly rotating and reflecting each image over each epoch unless otherwise specified . The weights were updated using the Adam optimization method and we use leaky-ReLu activation functions . During each epoch of training , each input image has a fifty percent probability of receiving a reflection in x , y and z followed by a fifty percent probability of subsequently being transposed . This particular scheme was chosen due to its low computational load . We have made the code for this convolutional neural network available on Github at https://github . com/rplab/Bacterial-Identification . The code was implemented on using python 3 . 5 on Ubuntu 16 . 04 , with a Intel Core i7-4790 CPU with an Nvidia GeForce GTX 1060 graphics card on a computer with 32 GB of RAM . With this hardware it took roughly one minute to train and create the features for the RF and SVC using about 17 , 000 images , and roughly one hour to train the 3D ConvNet on the same number of images . | The abundance of complex , three dimensional image datasets in biology calls for new image processing techniques that are both accurate and fast . Deep learning techniques , in particular convolutional neural networks , have achieved unprecedented accuracies and speeds across a large variety of image classification tasks . However , it is unclear whether or not their use is warranted in noisy , heterogeneous 3D microscopy datasets , especially considering their requirements of large , labeled datasets and their lack of comprehensible features . To asses this , we provide a case study , applying convolutional neural networks as well as feature-based methods to light sheet fluorescence microscopy datasets of bacteria in the intestines of larval zebrafish . We find that the neural network is as accurate as human experts , outperforms the feature-based methods , and generalizes well to a different bacterial species through the use of transfer learning . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"fish",
"medicine",
"and",
"health",
"sciences",
"neural",
"networks",
"vibrio",
"vertebrates",
"light",
"microscopy",
"neuroscience",
"animals",
"animal",
"models",
"osteichthyes",
"model",
"organisms",
"microscopy",
"artificial",
"intelligence",
"experimental",
"organis... | 2018 | Performance of convolutional neural networks for identification of bacteria in 3D microscopy datasets |
Soil grains harbor an astonishing diversity of Streptomyces strains producing diverse secondary metabolites . However , it is not understood how this genotypic and chemical diversity is ecologically maintained . While secondary metabolites are known to mediate signaling and warfare among strains , no systematic measurement of the resulting interaction networks has been available . We developed a high-throughput platform to measure all pairwise interactions among 64 Streptomyces strains isolated from several individual grains of soil . We acquired more than 10 , 000 time-lapse movies of colony development of each isolate on media containing compounds produced by each of the other isolates . We observed a rich set of such sender-receiver interactions , including inhibition and promotion of growth and aerial mycelium formation . The probability that two random isolates interact is balanced; it is neither close to zero nor one . The interactions are not random: the distribution of the number of interactions per sender is bimodal and there is enrichment for reciprocity—if strain A inhibits or promotes B , it is likely that B also inhibits or promotes A . Such reciprocity is further enriched in strains derived from the same soil grain , suggesting that it may be a property of coexisting communities . Interactions appear to evolve rapidly: isolates with identical 16S rRNA sequences can have very different interaction patterns . A simple eco-evolutionary model of bacteria interacting through antibiotic production shows how fast evolution of production and resistance can lead to the observed statistical properties of the network . In the model , communities are evolutionarily unstable—they are constantly being invaded by strains with new sets of interactions . This combination of experimental and theoretical observations suggests that diverse Streptomyces communities do not represent a stable ecological state but an intrinsically dynamic eco-evolutionary phenomenon .
Sampling DNA from diverse ecosystems has revealed a breathtaking diversity of microbial life [1] , [2] , especially in soil [3]–[5] . But we have barely begun to explore , both experimentally and theoretically , how these complex communities coexist and function . We know that microbes can interact via secretion of a wide array of small molecules , most notably antibiotic compounds . But how prevalent , diverse , and specific are such interactions ? How is the incredible diversity of microbes and their natural products maintained and promoted by complex and spatially structured networks of interactions ? To tackle these questions , we isolated bacterial strains from individual grains of soil and systematically measured all pair-wise interactions among them . We measured compound-mediated interactions , where a “sender” strain affects a “receiver” strain by secreting metabolites , antibiotics , or other compounds ( Figure 1AB ) . We focused on bacteria from the genus Streptomyces , which are the most prolific producers of small molecules , are abundant in soil [6] , and exhibit diverse production and resistance capabilities [6]–[8] that are modular and prone to Horizontal Gene Transfer ( HGT ) [9] . Sixty-four Streptomyces from four individual grains of soil were isolated ( Figure 1C ) , phenotyped for all possible pair-wise interactions , and genotyped for 16S rRNA . We explored the statistical properties of the resulting network and juxtaposed them with those emerging from a simple ecological model of bacteria evolving production of and resistance to antibiotics [10] , [11] . We developed a high-throughput platform for measuring directional pairwise interactions by observing how the products of one bacterial strain affect the colony growth of another . A fine-pored filter is placed on a nutrient agar surface , a lawn of the sender strain is grown on top , and the filter is removed—leaving behind sterile agar that has been altered or conditioned by the sender strain . The conditioned agar is then resupplied with concentrated liquid nutrients to compensate for the nutrients consumed by the donor . A receiver strain is point-inoculated onto the sterile conditioned agar and time-lapse movie of the growing colony is taken . A high-throughput implementation of this assay allowed us to acquire 11 , 500 movies along 15 d at 4 h time resolution , covering all pairwise interactions within our collection of 64 strains in duplicate ( Figure 1D , Materials and Methods ) . By comparing colony growth of the receiver strains on conditioned and non-conditioned agar , we identify interactions between strains . We quantified the first time point in which each colony becomes visible on the images ( appearance time ) to identify growth inhibitory and growth promotion interactions . In addition , we visually scored instances of inhibition of aerial mycelium formation ( and subsequent sporulation ) .
We found a rich and complex interaction matrix among our collection of strains with multiple cases of growth inhibition , growth enhancement , and inhibition of aerial mycelium ( Figure 2A; and see also all the 11 , 500 time lapse movies underlying this matrix in Figure S8 ) . This matrix showed two immediately apparent special properties: it is “balanced” and it is “sender determined , ” as we now explain . The first noteworthy property of the matrix is that the frequency of interactions ( a . k . a . connectance ) is balanced: the probability that two random isolates interact is neither close to zero nor close to one . While there are many strong interactions , the matrix is far from the limit in which interactions are non-specific because everyone interacts with everyone else: of the 64 isolates , there are at least 42 different interaction profiles . We found 45% of growth or aerial mycelium inhibitory interactions ( 25% complete inhibitions of growth ) and 19% growth promotion ( see Materials and Methods for additional details ) . The frequency of inhibitory interactions is significantly higher ( and more balanced ) than previous estimates based on zones of inhibition [12] . The balanced frequency of interactions makes this network more highly connected than most known ecological networks ( only for a few food webs the density approaches 30% ) [13]–[16] . The second striking property of the measured matrix is that it is very different along the sender and receiver axes , with characteristic stripes of inhibition and non-inhibition running along the receiver ( vertical ) axis . This asymmetry is surprising because the existence of an inhibitory interaction is , in general , influenced by both the sender , which needs to produce a toxic compound , and the receiver , which needs to be sensitive to the compound produced . In the extremes , a matrix that is determined purely by the properties of the sender would exhibit perfect stripes in the vertical direction , while a receiver-determined matrix would exhibit stripes along the horizontal ( sender ) axis . Thus , the network we observed is more sender-determined . This sender-receiver asymmetry can be quantified by comparing the distribution of the fraction of isolates that each isolate inhibits ( sender degree ) with the distribution for the fraction of isolates that inhibit each isolate ( receiver degree ) ( see Figure 2B ) . The sender degree is broad and peaks near its extreme values , while the receiver degree is narrower and unimodal . The difference of the variances of the receiver and sender distributions is a measure of the sender-receiver asymmetry ( Q = −0 . 37; Figure 2C ) . The negative value of this quantity means that information gathered about a sender from a few interactions let us predict far better the rest of its interactions than the corresponding information about a receiver . The sender-receiver asymmetry is pronounced but not extreme , indicating the importance of resistance to antibiotics that strains do not themselves produce . The bimodality of the sender degree distribution makes this network very different from networks with nodes randomly connected with a fixed probability ( Erdös–Renyi random graphs ) , scale-free ( social ) networks [13] , and food webs with exponential-tailed distributions [17] . It is unclear how coexistence between strains that inhibit almost everyone and strains that inhibit almost no one is maintained . One possibility is the presence of an ecological tradeoff between ability to inhibit and ability to resist , which would imply a positive correlation between the sender and receiver degrees . But no such correlation exists; on the contrary , isolates that inhibit most are also among the most resistant ( Figure S1A , p<10−4 ) . There is also no correlation between growth rate on non-conditioned media and the sender or receiver degree ( Figure S1B ) . We decided to look for hints about the maintenance of a diverse sender-determined network in the network evolution . We sequenced the 16S rRNA of all isolates and found that closely related isolates are less likely to inhibit each other ( Figure 3B ) , but there is a poor overall correlation between phenotypic and phylogenetic distances ( Figure 3A , C ) . Even isolates with identical 16S rRNA sequences can have very different interaction profiles . This lack of strong correlation between phylogeny and inhibition profiles is consistent with previous work [18] . To further exploit this phylogenetic signal , we compared the phenotypic divergence of sender and receiver profiles for isolates with the same 16S , and contrasted it with the null expectation of isolates with different 16S ( Figure 3D ) . Interestingly , the sender profiles diverge disproportionately more than the receiver profiles for closely related strains even after controlling for the overall sender-dominated nature of the matrix ( P = 2⋅10−4 ) . So it seems that the Streptomyces community is in a state in which frequent evolutionary changes in production ( mediated for example by transfer of plasmids carrying antibiotic production genes ) cause dramatic changes to ecological interactions . The coupling between ecology and evolution is therefore important for understanding the network properties . Is the balanced frequency of interactions and sender-determined nature accidental or a natural outcome of the ecological and evolutionary dynamics of interacting Streptomyces communities ? Can we account for the large changes in interaction patterns over short evolutionary distances ? We consider a simple in silico model of communities of strains producing and resisting a set of antibiotics . A strain inhibits another if it produces at least one antibiotic to which the other is sensitive . Communities of strains with randomly assigned production of and resistance to antibiotics exhibit a diverse set of qualitatively different interaction matrices , depending on the frequency of production and resistance ( Figure 4A ) . With many antibiotics , matrices similar to the observed—balanced frequency of interactions and moderately sender-determined—occupy a small region of the parameter space , and require low frequency of production and high frequency of resistance . This raises the question of whether introducing evolution into the model can inherently direct it into the regime of balanced and sender-determined interactions . We imposed simple non-spatial ecological dynamics that implicitly incorporates the importance of spatial relations between bacteria over short time scales ( the antibiotics stay near their producers ) . The fitness of each strain depends on the weighted sum of its interactions with all other strains , incorporating the following contributions: ( i ) a negative effect of being inhibited by others , ( ii ) an advantage of inhibiting others , and ( iii ) a reduced ability to inhibit if being inhibited ( protection by inhibition ) . A cost for production or resistance of any antibiotic is also added . The resulting mathematical structure is that of a discrete time Lotka-Volterra model with coefficients derived from the pairwise interaction matrix ( Figure 4B and Materials and Methods ) . For simplicity , the model ignores that antibiotics might also function as a “common good” reducing competition from non-resistant Streptomyces and non-Streptomyces strains . In addition , it ignores the possibility of resistant neighbors extending their protection to non-resistant strains [19] . We also ignore positive interactions and primary metabolism differences ( utilization of different resources and cross-feeding ) , which are potentially important . Some of these effects can be incorporated by adding terms with higher order interactions ( for several model extensions , see Materials and Methods and Figure S7 ) . To capture the long-term effects of the interplay between ecology and evolution on the statistical properties of interactions , we added mutations to the above model . Mutations allow acquisition or loss of production and resistance to any of the antibiotics . Turnover of production and resistance capabilities is indeed expected to be important for Streptomyces , as evidenced by the modular nature of antibiotic production and the vectors through which it spreads . The simulation starts from a single strain that is sensitive to all antibiotics and follows the dynamics until a statistical steady state is reached ( Figure 4C ) . We systematically explored the behavior of the model for a range of costs of production and resistance ( Figure 4D ) . The results show a maximum cost of production above which no antibiotics are produced ( Figure 4D , white area ) . Strikingly , below this threshold we see balanced sender-determined matrices ( Figure 4D , green shades ) , as long as the production costs are higher than the resistance costs ( above the dashed blue line ) . There is an inherent feedback that keeps the frequency of interactions from becoming too low or too high: an increase of interaction frequency selects for an increase in resistance levels , which then leads to a decrease of the interaction frequency . This qualitative picture holds provided that the level of protection by inhibition is below a certain threshold ( Figures S2 , S3A ) . On the other hand , if inhibition is an effective defense , then when resistance cost is high the system collapses into a state in which most strains inhibit each other ( Figures S2 , S3A ) . While different outcomes are possible in the model , we observe balanced and sender-determined matrices over a large region of the parameter space ( Figure S3 ) . We also explored the relation between interaction and phylogeny in the simulations . In agreement with our experimental observation , in the balanced and sender-determined region , we find that in the resulting interaction matrices ( Figure S4A ) strains are more likely to interact when they are phylogenetically distant ( Figure S4B ) , and there is a weak overall correlation between phylogenetic and phenotypic distance ( Figure S4C ) . Community diversity requires both ecology and evolution . The functional diversity of the system increases sharply with both the evolutionary rate and the population size , and turning off the ecological interactions or reducing the mutation rate leads to a loss of diversity ( Figure 4E and Figure S5 ) . The community steady state is characterized by a continuous turnover of different interaction phenotypes ( Figure S4D ) , indicating its evolutionary instability . To investigate statistical properties beyond those captured by the degree distributions , we followed an established procedure for identifying interaction motifs—local patterns of interactions that are more frequent than expected by chance [20] . We discovered that the simulated interaction networks are strongly enriched for mutual inhibition when compared with random networks with the same sender and receiver degrees for each isolate ( Figure 5A ) . This is not surprising since mutual inhibition is an important mechanism for ecological balance . We , therefore , looked for reciprocity in the experimental data . In the experimental data , unlike the model , there is an extra complexity due to positive interactions ( growth promotions are not included in the model ) . With both positive and negative interactions there are six two-isolate motifs ( Figure 5B , C ) . We compared the six motif frequencies with those for random matrices that have the same sender and receiver degrees for each isolate , and which preserve the corresponding degrees for the growth promotion interactions . Since an obvious source of reciprocity structure is the presence of identical isolates , we excluded strains with identical 16S and interaction profiles from this analysis . As we observe in the model , the analysis of the experimental data revealed statistically significant enrichment for reciprocal interactions—there are more mutual inhibitory interactions and mutual growth promotions than expected and fewer asymmetric relationships ( Figure 5B ) . If reciprocity of interactions among pairs of strains is a property of coexisting communities , we may expect that it will be more enriched in strains coming from the same soil grain than for strains isolated from different grains . We found that while the frequency and strength of positive and negative interactions does not differ within and between grains , interactions of pairs of strains within grains do indeed tend to be more reciprocal than interaction of strains from different grains ( Figure 5C ) . This result is significant only if we include the inhibitions of aerial mycelium . The motif distributions are also sensitive to the choice of thresholds for defining interactions . A threshold independent analysis of the continuous data shows again enrichment for reciprocity ( Figure S6 , p = 0 . 001 ) . The apparent enrichment for reciprocity remains if we control for a tendency to have isolates with more similar 16S within a grain . A larger dataset will be required to distinguish between different underlying causes for the patterns of interactions within and between soil grains .
We find that Streptomyces isolates from soil grains exhibit diverse and rich interaction patterns . The interaction matrix they form has a balanced frequency of interactions—the probability that two random strains interact is neither close to zero nor to one . The sender-degree distribution is broad and bimodal—isolates tend to inhibit almost everyone or almost no one , which makes the interactions statistically controlled more by the properties of the sender than the receiver . This sender-receiver asymmetry , while pronounced , is not extreme , indicating the importance of resistance to compounds produced by others . These properties make this network very different from other ecological networks , which have monotonic degree distributions , and typically exhibit much lower interaction frequency . Finally , the community is enriched in reciprocal interactions—interaction pairs are enriched in mutual inhibitory interactions and mutual growth promotions , while it is rare to find cases in which one strain promotes a second , but this second strain inhibits the first . This reciprocity is further enriched among strains derived from the same grain of soil , thus revealing spatial structuring of interactions . These properties of the interaction network have emerged from a long evolutionary process , which we probed by juxtaposing interactions and phylogeny . We found that the interactions of an isolate can change dramatically even over short evolutionary time ( indicated by very close 16S sequences ) , with evolution changing the production profiles more than the resistance profiles . Incorporating such fast evolution in a dynamic ecological model of antibiotic interactions , we find that most of the observed properties of the network are reproduced under a broad range of parameters . The community compositions are not static—increase in production of an antibiotic promotes resistance , which promotes sensitivity , and invites production again . As the community undergoes cycles with respect to different antibiotics , different combinations of production and resistance become favorable , which makes it evolutionary unstable . In our model both ecological interactions and continuous turnover of interaction phenotypes are required to maintain functional diversity . Our work has several important limitations . Perhaps the main one is that interactions are measured in the lab and actual interactions in the soil may be more complex or different . We were also limited to studying only the interactions among Streptomyces strains; interactions between Streptomyces and other microbes could be of major importance . Higher order interactions , such as synergy or antagonism between natural products , or induction of small molecule production by other small molecules , are not captured by the pairwise measurements . Many of these shortcomings are inherent to most current studies of microbial species interactions . However , the systematic and high-throughput nature of the current study allows us to ask questions at the statistical level , and might therefore be less prone to some of these difficulties . Furthermore , the high-throughput interaction platform developed here and the simulations offer a natural foundation for many subsequent studies of microbial communities , which will address some of the above concerns , potentially yielding important biological insights . For example , it is now possible to probe how the statistical properties of networks , such as the relative significance of positive and negative interactions , are affected by media composition and the presence of other small molecules . This enables investigations of the regulatory roles of and epistatic effects between small molecules . It would also be interesting to see whether the effects of a sender on a receiver will be modified if the sender is co-incubated with the receiver . Finally , the interaction platform can be used to follow the evolutionary and ecological dynamics of synthetic laboratory communities of interacting microbial strains . The observed network properties do not seem to correspond to an ecologically stable state maintained by antibiotic interactions alone . Instead , the model and observations suggest that they are supported by a constant evolutionary change . The distribution of production and resistance in the community is poised so that simple changes in production capabilities of a strain can alter its interactions with many other strains potentially to a great ecological advantage . This evolutionarily unstable ecological state seems complemented by the modular nature of the secondary metabolite gene clusters , which enable such changes and , thus , lead to turnover of interaction phenotypes of different strains and species . This continuous turnover might in turn be important for the emergence and maintenance of the modularity and clustering of small molecule production and resistance genes and their recruitment to mobile genetic elements [21] . This reasoning suggests a unified view of network structure , network evolution , and modularity of secondary metabolism to be further explored .
We sampled four soil grains of soil by touching the soil with a dry needle tip , and lifting particles of less than 1 mg of wet weight . Three of the grains were 1 cm away from each other in one soil core , and the fourth grain was 10 cm away from a second soil core . The depth was approximately 2 cm below the surface . The sampling was performed in December from foliage-covered soil away from visible roots . Each grain was dried for 2 d , then suspended in dH2O , vortexed , sonicated , diluted , and plated on Streptomyces Isolation Media [7] . Plates containing five colonies or fewer were sampled in order to minimize potentially biasing interactions between emerging colonies . Isolates that exhibited the characteristic aerial mycelium pattern of Streptomyces were selected at random after 2 wk , and their genus identity later verified by sequencing . Five of the isolates were classified as genus Kitasatospora within the family Streptomycetaceae by the Ribosome Database Project [22] . Each isolate was restreaked once , then grown in TSB for 3 d , and 300 µl/plate was spread on four petri dishes containing Bennett's agar [7] . Plates were incubated for 14 d at 28°C . Spore lawns were harvested in 12 ml of 0 . 01% Tween 80 , vortexed for 2 min , and filtered through 5 µm syringe filter to separate the spores from mycelium . The filtrate was centrifuged at 1 , 000 g for 10 min , and the spore pellet was resuspended in 1 . 1 ml of 20% glycerol , aliquoted , and frozen at −80°C . Each spore stock that we used was thawed only once . During stock preparation , tubes were kept on ice . Bulk soil was sent to the Soil and Plant Tissue Testing Lab at the University of Massachusetts at Amherst . The soil pH is 5 . 5 . The texture is loam with 46 . 7% sand , 42 . 1% silt , and 11 . 2% clay . Organic matter , 12 . 6% . NO3-N , 0 ppm . Mineral content: P , 7 ppm; K , 230 ppm; Ca , 1 , 511 ppm; Mg , 157 ppm . Micronutrients: B , 0 . 3 ppm; Mn , 7 . 1 ppm; Zn , 9 . 3 ppm; Cu , 0 . 3 ppm; Fe , 32 . 4 ppm; S , 28 . 8 ppm . Cation Exch Cap , 21 . 7 Meq/100 g . Media for interactions: 15 g purified agar in 1 L d H2O , 2 g potato starch , 0 . 8 g casein , 1 g KNO3 , 0 . 4 g K2HPO4 , 0 . 2 g MgSO4 , 30 mg CaCl2·2H2O , pH 7 . 2 . All components were autoclaved separately in concentrated form , and all agar plates were made from the same autoclaved stocks . Resupply media was 18× concentrated interaction media with the exception of KHPO4 , which was 36× concentrated , pH 7 . 0 . Black 96-well agar plates were robotically over-filled with agar , and before solidification a glass plate was lowered to 1 . 5 mm above the plate to flatten the agar meniscus . The glass plate was slid sideways upon solidification of the agar . The resulting agar columns were flat on top ( to ensure good filter contact and high image quality ) , protruded above the edge of the plate ( to ensure good contact with filter during conditioning ) , and well separated from neighboring wells ( to prevent cross-talk ) . Since high pipetting accuracy was required , the aspirated amount was automatically adjusted based on the instantaneous agar temperature ( ∼50°C ) , care was taken to dip the pipette tips to the same depth in the agar reservoir , and room ventilation was turned off to prevent asymmetric cooling of the agar in the tips . Rectangular filters—polycarbonate 0 . 03 µm pore size—were placed over the agar plates . Each well was inoculated with 8 µl of spore stock . Due to the hydrophobicity of the filters , droplets above neighboring wells were well separated . After 8 d of incubation , growth on each filter was imaged , and the filter removed . Filter images were used to discard data from defectively conditioned or contaminated wells . Plates were resupplied with a 20 µl droplet of resupply media , and dried in a fume hood for 90 min . Each plate was pinned from a source 96-well plate containing 100 µl/well of spore stock ( ∼107 spores/ml ) . Source plates were kept between 4 and 8°C during pinning . Pins were sterilized between plates to prevent contamination of the source plate due to accidental contamination of the agar plates . The time of pinning of each plate was recorded , and it was placed upside down on a flatbed scanner so that the agar surface is 2 mm away from the scanner glass surface . The focusing plane of the scanners was correspondingly adjusted . To minimize agar drying , plates were sealed to the scanners with packing tape . Colonies were scanned approximately every 4 h . Temperature was maintained at 28°C , but jumped temporarily by 1°C after each scan . Plates were scanned for at least 15 d . The appearance time for each colony ( the first time point at which a colony becomes visible on the images ) was manually determined using custom interactive software . Colonies associated with agar defects or contaminations were discarded . Aerial mycelium is apparent on the images as a fuzzy texture on top of the colonies ( Figure 1D ) . Aerial mycelium inhibition was scored if there was no or very little ( in comparison to non-conditioned ) aerial mycelium coverage of the colony after 15 d . Twenty-five percent of the interactions are complete inhibitions , i . e . no visible growth of receiver colonies . One isolate inhibits itself . An additional 10% of interactions are partial inhibitions with colonies appearing at least 1 d later on conditioned media ( for a total of 35% ) . The fraction of inhibitory interactions is 45% , if inhibitions of aerial mycelium formation are included . Colonies were grown in TSB for 3 d , centrifuged at 1 , 000 g for 10 min , and resuspended in dH20 three times . Cells were then resuspended in lyses buffed ( PrepMan ) and heated to 100°C for 10 min , centrifuged , and the supernatant was frozen at −20°C . 3 µl of this supernatant was added to 60 µl PCR mix containing 12 µl Qiagen Q-solution , 2 . 4 µl of 10 µM forward primer GAG AGT TTG ATC CTG GCT CAG , and reverse primer CGG CTA CCT TGT TAC GAC TTC . Samples were PCR amplified ( 95°C for 3 min , 35 cycles of 95°C for 1 min , 55°C for 1 min , 72°C for 1:30 min , and final extension at 72 deg for 7 min ) , and PCR products were sent for sequencing upon confirmation of existence of a product of the expected size ( ∼1 . 5 kb ) . Sequences from the forward and reverse primers had a significant overlap . Sequences are available through Genbank , accession numbers: JN020489–JN020551 . The grain and isolate number within a grain is specified in the description for each sequence; e . g . G4_6 is the sixth isolate from grain four . We considered the profiles of two isolates distinct if they differed by more than 2 d in appearance time ( the first time point in which a colony becomes visible on the images ) for both replicates in at least three sender or three receiver positions . According to this measure , there are 42 distinct phenotypic profiles . For a N×N binary interaction matrix ( one indicates an interaction , and zero no interaction ) , the frequency of interactions is , and the sender-receiver asymmetry is defined as . Matrices with negative Q are sender-determined , and with positive Q are receiver-determined . We obtain negative Q independently of how we threshold the inhibitory interactions and of whether or not we include aerial mycelium inhibitions . The fraction of differences between profiles was calculated ( after discarding defective and inconsistent replicas ) . The profiles were taken from a binary interaction matrix in which inhibitions were defined as delays in colony appearance time of more than 1 d . Increasing the threshold to 3 d ( i . e . strong inhibitions ) did not change the qualitative findings of Figure 3 . However , inclusion of aerial mycelium inhibitions renders the statistics of Figure 3D insignificant . Sequences were aligned to a universal 16S rRNA template using the Ribsomal Database Project website [22] . For each pair of sequences , only positions for which both sequences have high-quality values from the sequencing trace were considered; the rest were treated as missing values . Phylogenetic distance was computed as the fraction of differences ( in high-quality positions ) . Alignment gaps were counted as normal differences . For the measured network of positive and negative interactions ( without weights ) , we generated an ensemble of random networks that have the same number of ingoing and outgoing arrows of positive and negative interactions for each isolate . Networks were randomized by taking random pairs of single arrows ( between different isolates ) and swapping the isolates on which they end , provided the two arrows created by the swap do not exist already or correspond to missing or defective experimental values . ( In this way , the missing or defective values of the matrix were kept in place . ) Each cycle consisted of swapping one pair of positive and one pair of negative arrows . This operation was performed thousands of times before selecting each random ensemble representative . For each random network the frequency of each of the six pairwise motifs was calculated , without counting any of the diagonal matrix elements . The motif significance ( p value ) was calculated as the fraction of random networks that have more extreme motif frequency than that for the observed network . The protocol was analogous for the matrices resulting from the eco-evolutionary model , which had only negative interactions and no missing or defective values . Ecology: Each strain i is characterized by an array Ziα , specifying whether it is producer ( P ) , resistant ( R ) , or sensitive ( S ) to antibiotic α . Let Ai←j be the binary matrix of inhibitory interactions . Strain j inhibits strain i , i . e . Ai←j = 1 , if Zjα = P and Ziα = S for any α . Let ni be the fractional abundances of different strains ( summing to one ) . The “fitness” of i is , where . At each time step N individuals are drawn from different species with relative probafobilities . λ and η are ( positive ) ecological parameters controlling the direct benefit of inhibiting neighbors and the consequence of mutual inhibition ( the level of protection by inhibition is 1−η ) , and is the intensity of selection within an ecological cycle ( which we specify through by ) . λ = 0 means that inhibition is a zero-sum game; in the other extreme λ = 1 means complete spite ( no direct benefit for the inhibitor ) . Production or resistance of an antibiotic incurs a multiplicative fitness cost , so that is the bare fitness reflecting the costs and of production and resistance ( antibiotic dependent ) , and δ is the Kronecker delta . Evolution: each antibiotic position in each of the N individuals mutates within the SRP space of possibilities with probability specified by a set of transition rates: , , , , , . N = 106 , = 0 . 05 , λ = 0 . 15 , η = 0 . 7 , 40 antibiotics . , , , and . The relative mutation rates assume that loss of function is more likely than gain of function , and gain of resistance is easier than gain of production . The no interaction case in Figure 4E corresponds to = 0 . Figures S2 and S3 explore the behavior of the model for other parameters . We examined the behavior of the model when different antibiotics have different production costs rather than identical costs . The production costs were uniformly distributed in the interval ranging from the resistance cost up to the maximal cost for which a producer can invade a sensitive strain . We discovered that this extends the region over which we observed balanced interaction matrices , and leads to receiver-determined matrices at large resistance costs ( Figure S7A ) . We also added an evolutionary operator that mimics more closely within-population HGT—rates of change towards production and resistance of an antibiotic are proportional to the abundance of production and resistance to that antibiotic in the population ( rather than being constants ) . With probability of 10−4 an organism pairs with another random organism and gains a production or resistance for an arbitrary antibiotic of the donor . Adding within-population HGT ( while keeping the mutations ) did not qualitatively change the results ( Figure S7B ) . | Soil harbors a diverse spectrum of bacteria that secrete small molecules such as antibiotics . Streptomyces bacteria , considered the most prolific producers , have been mined for decades for novel products with therapeutic applications , yet little is known about the properties of the interaction networks these compounds mediate . These networks can hold clues about how the diversity of small molecules and of Streptomyces strains with different production and resistance capabilities is maintained and promoted . To explore the network properties , we developed a high-throughput platform for measuring pairwise phenotypic interactions mediated by secreted metabolites , and used it to measure the interaction network among 64 random Streptomyces isolates from several grains of soil . We found many strong but specific interactions that are on average determined more by metabolite production than by metabolite sensitivity . We found reciprocity between strains , whereby if one strain inhibits or promotes the growth of a second strain , it's likely that the second strain affects the first strain in a similar manner . These interactions are not correlated with phylogeny , as very closely related strains exhibit different interaction patterns . We could explain these findings with a mathematical model requiring interplay between ecological dynamics and evolution of antibiotic production and resistance , suggesting that the bacterial and small molecule diversity of these communities is maintained by constant evolutionary turnover of interaction phenotypes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"systems",
"biology",
"microbial",
"evolution",
"biology",
"microbiology",
"microbial",
"growth",
"and",
"development",
"genetics",
"and",
"genomics",
"microbial",
"ecology"
] | 2011 | Structure and Evolution of Streptomyces Interaction Networks in Soil and In Silico |
Publication bias in animal research , its extent , its predictors , and its potential countermeasures are increasingly discussed . Recent reports and conferences highlight the potential strengths of animal study registries ( ASRs ) in this regard . Others have warned that prospective registration of animal studies could diminish creativity , add administrative burdens , and complicate intellectual property issues in translational research . A literature review and 21 international key-informant interviews were conducted and thematically analyzed to develop a comprehensive matrix of main- and subcategories for potential ASR-related strengths , weaknesses , facilitators , and barriers ( SWFBs ) . We identified 130 potential SWFBs . All stakeholder groups agreed that ASRs could in various ways improve the quality and refinement of animal studies while allowing their number to be reduced , as well as supporting meta-research on animal studies . However , all stakeholder groups also highlighted the potential for theft of ideas , higher administrative burdens , and reduced creativity and serendipity in animal studies . Much more detailed reasoning was captured in the interviews than is currently found in the literature , providing a comprehensive account of the issues and arguments around ASRs . All stakeholder groups highlighted compelling potential strengths of ASRs . Although substantial weaknesses and implementation barriers were highlighted as well , different governance measures might help to minimize or even eliminate their impact . Such measures might include confidentiality time frames for accessing prospectively registered protocols , harmonized reporting requirements across ASRs , ethics reviews , lab notebooks , and journal submissions . The comprehensive information gathered in this study could help to guide a more evidence-based debate and to design pilot tests for ASRs .
In recent years , several reports have questioned the way animal research is conducted and reported , citing a lack of reproducibility of preclinical animal research data and poor translation of published preclinical data into the human setting [1 , 2] . Because results from preclinical animal research inform other preclinical research , early clinical research , and , in cases in which evidence from clinical trials is missing , even off-label clinical practice , accurate and complete reporting of animal research is essential to reduce harm to trial participants and patients , to optimize funding allocation , and to effectively reduce and refine animal research . Incomplete reporting of studies and study results has been described as “publication bias , ” “selective reporting , ” and “dissemination bias” [3] . Whereas this bias has been studied in depth for clinical trials , less data is available for preclinical animal research . Various data sources have been used for the analysis of the publication rate of clinical trials: study protocols approved by an ethics committee [4–6] and study protocols registered in a trial registry [7] . The latter has become a much more useful data source since 2004 , when the International Committee of Medical Journal Editors ( ICMJE ) made registration of clinical studies a prerequisite for publication in their journals [8] . In 2007 the United States Food and Drug Administration ( FDA ) Amendment Act further supported trial registration . The comparison of trial registries and journal publications has allowed new kinds of analysis: following the rate and time of journal publication , analyzing possible factors influencing publication status ( such as study type and funding source ) , as well as finding discrepancies between registry entries and data published in journals [9 , 10] . This approach has helped to quantify dissemination bias in clinical research and to define possible reasons and solutions [11 , 12] . Similar studies are not available for animal research , because there are no registries for this type of study . Instead , analyses of dissemination bias in this field are more indirectly based on ( A ) data from survey research with animal researchers [13] , ( B ) inferences from the debate about the relatively low reproducibility of preclinical research [14] , ( C ) the high failure rate of early human trials [1 , 15] , ( D ) statistical methods calculating the probability of bias in the available data [16 , 17] , and ( E ) study abstracts published in conference proceedings or on their websites [18] . By analogy with clinical research , the implementation of prospective animal study registries ( ASRs ) has recently been suggested as one measure that might help to directly assess and substantially reduce dissemination bias in animal research [19] . A workshop on “publication bias in animal research” organized by the National Centre for the Replacement , Refinement & Reduction of Animals in Research ( NC3R ) in 2015 also focused on the issue of ASRs . A panel debate in this workshop demonstrated the broad lines of , and strong contrasts in , argumentation for and against ASRs . All panel participants agreed , however , that future decision-making on the issue of ASRs depends strongly on context , such as registry characteristics and knowledge about conflicting stakeholder interests . This study aimed to address that need . The primary objective of this stakeholder analysis was to systematically and transparently assess the full spectrum of potential ASR-related strengths , weaknesses , facilitators , and barriers ( SWFBs ) .
We present here an overview of the results of our stakeholder analysis; further interpretation of key results is then given in the Discussion section . Altogether , we identified 518 relevant text passages in the 21 interview transcripts ( based on 13 hours of interview with stakeholders from four different groups , see the Material and Methods section ) and 11 references [13 , 19–28] ( see also S2 Table ) , from which we derived 130 subcodes grouped under the four broad categories for ASR-related SWFBs . Table 1 presents our definitions for ASR-related SWFBs . The Material and Methods section further explains how the 130 subcodes were derived . Fig 1 visualizes core results and S1 Table presents the full spectrum of all 130 codes for the four SWFB categories together with sample quotations for each code ( Fig 1 , S1 Table ) . For example , for the broad category “strengths , ” our thematic analysis provided the subcode “Quality improvement , research support” by referring to the following two quotations: ( A ) “At the moment , I can publish a study with ten animals per cohort and you don’t know that I just ignored the other ten animals in the same cohort because I didn’t like what happened . There’s no way you could know that , ” and ( B ) “In animal research , it’s really unusual for people to articulate a primary hypothesis . If you require study registration and the registries require that you stipulate a primary hypothesis , you are now creating a very strong motivation for people to actually start designing studies having thought through what their primary hypothesis is going to be . ” S1 Table gives sample quotations for each of the 130 codes . The thematic analysis reached saturation for the first level of subcategories , meaning that analyzing further interviews would not reveal other SWFB subcategories . For each SWFB category , we list subcategories separately by interest group: ( a ) animals , ( b ) preclinical/clinical research , ( c ) industry , ( d ) regulators , ( e ) public/patients , and ( f ) overarching issues . Most codes were related to “preclinical/clinical research” ( n = 63 ) , followed by “overarching issues” ( n = 25 ) , “industry” ( n = 15 ) , “public/patients” ( n = 10 ) , “animals” ( n = 7 ) , and “regulators” ( n = 4 ) .
Interviewees across all stakeholder groups said that a registry would most probably function as an incentive for ( i ) more rigorous study protocols for animal research ( codes S9a and S9b in S1 Table ) , ( ii ) less selective reporting in journal publications ( code S10 ) , and ( iii ) less biased reporting of preclinical data in protocols and investigator brochures for early human studies that are submitted to Research Ethics Committees ( RECs ) or regulatory bodies ( code S34 ) . It was also undisputed that a registry could help to better disseminate evidence ( code S4a ) , thereby promoting transparency ( codes S33 and S5 ) and also facilitating network building among researchers who work on similar research questions ( code S6a ) . How strongly these potential strengths would materialize was difficult for the interviewees to anticipate , but it was highlighted that if a qualitatively appropriate and informative registry such as clinicaltrials . gov existed for preclinical research , this could serve as a core information source for searching and refining ideas ( code S4b ) , fostering networks ( code S6b ) , disseminating findings ( codes S34 ) , and promoting trust within the scientific community ( codes S19a and S19b ) . Interviewees also unanimously described as a strength the role of registries in facilitating meta-research ( research on preclinical research ) that could , for example , help to quantify publication bias in a more robust way ( codes S14 and S15a ) . Based on experience of clinical trial registries , all stakeholder groups were ambivalent about whether ASRs would directly reduce publication bias ( codes S10 and S15b ) or whether it would rather be a tool for other more indirect methods , as described above . Sounder research funding allocation ( code S32 ) was also highlighted as a potential strength . This is because reviewing research grants can also be affected by biased publication; decreased bias could thus improve judgments on whether a research proposal really suggests new and relevant investigations . Similarly , some interviewees pointed to a positive impact on replacement , reduction , and refinement ( 3R principles ) of animal experiments ( codes S2a , S2b , and S3 ) . However , others feared that stricter regulations following an ASR implementation , e . g . , requirements to meet sample size calculations , might in the end lead to larger study group sizes and therefore to more animals being used in preclinical animal research ( code B4 ) . On the other hand , this could increase the statistical power of animal studies performed ( code S9a ) . The overall numbers of animals as well as the cost might be balanced by the fact that fewer redundant experiments would be performed ( code W2 ) . With regard to this concern , other interviewees highlighted the need to better differentiate between exploratory and confirmatory animal studies [29] . Whereas the registration of confirmatory animal studies would help reduce waste in clinical research ( codes S26a and S26b ) , the registration of exploratory studies would rather help avoid the redundant use of animals ( codes S2b and S2c ) and provide information that might help in the refinement of animal studies ( code S3 ) . Another controversial topic was centered on whether and how ASRs might affect financial , human , and time resources ( code S18a ) . We describe potential strengths in this paragraph and related weaknesses in the next section . Some interviewees suggested as a potential strength the savings in time and money if ASRs helped avoid experiments already performed elsewhere ( codes S2b and S18b ) . Some industry representatives argued that this might also reduce the cost of drug development ( code S25 ) , whereas others were skeptical due to the complex reasons for high costs in competitive drug development ( code W10 ) . It was also controversial whether public trust in biomedical research more generally and in animal research specifically could be increased by transparency resulting from ASRs ( codes B1a and B2a ) . It was mentioned that the public will always welcome more transparency ( code S28b ) , but interviewees remained skeptical about whether ASR-related transparency will finally result in more or less public trust in research ( codes S28a and B1b ) . See the “Barriers” section for more information on this issue . A clear weakness for researchers in both industry and academia was the negative impact on intellectual property ( code W3 ) and the associated potential theft of ideas ( code W4 ) . See the section “Uncontroversial Potential Facilitators” below for potential solutions offered by interviewees to this issue . Addressing this concern requires balancing intellectual property interests that demand a confidentiality time frame as long as possible with the usefulness of a registry that requires detailed and up-to-date information . Another fear expressed by animal researchers was of the additional administration and the time needed to accomplish it ( see codes W1a and W1b ) . Some mentioned the considerable amount of time they already have to invest in complying with regulation and documentation requirements . However , participants assumed any registry would be time consuming , and they recognized that there may be ways to reduce expenditure of time , as explained in the section “Uncontroversial Potential Facilitators” below . One interviewee highlighted the difference between the additional time needed for administration and the time saved by reducing unnecessary repetitions of experiments ( code S18c ) , which might explain the initially surprising emergence of time issues in both the strengths and the weaknesses category . Some interviewees argued that ASRs will negatively affect creativity and serendipitous findings in animal research ( code W6 ) , fearing that a registry could preemptively define a structure for studies that might be too rigid to capture some project ideas and thereby prevent certain types of research ( code W7 ) . However , those who agreed that a registry could negatively affect creativity mainly referred to exploratory research , which they did not want to be limited by a time-consuming registry . Others also highlighted how more comprehensive and less biased information on previous research might even facilitate creativity and inspire innovative research questions ( code S7 ) . The possible theft of ideas and the associated competitive disadvantage was an issue that worried both academic and industrial researchers . However , when it was suggested that registry entries could be set for disclosure at some future moment to protect intellectual property , many interviewees agreed that , depending on the time frame ( for which one to two years was often suggested ) , this would facilitate an ASR implementation ( codes F9a and F9b ) . Although the administrative burden was a frequently mentioned weakness , one possible solution to this was mentioned as well: if the “core data” for an ASR was congruent with the data that need to be submitted for research funding , approval processes , journal submissions , and other purposes ( e . g . , using the ARRIVE guidelines as a basis ) , the time constraints would be lessened greatly ( code F40 ) . In addition , some interviewees highlighted that digital lab notebooks ( DLNs ) could play a central role in quick and efficient registration of studies: “If you have all your study characteristics on your DLN , then you just need to press the submit button to upload your protocol to a registry” ( code F11b ) . However , interviewees were rather pessimistic as to whether such harmonization of core data and the implementation of DLNs are realizable in the next few years . The importance of the resource question is illustrated by the fact that several facilitators mentioned by animal researchers were related to financial or staff support to cope with the additional workload ( codes F12 and F13 ) . Again , regarding the topic of intellectual property , a few interviewees suggested retrospective registration as a possible solution ( code F8 ) . However , the use of such a retrospective registry was questioned when it came to possible strengths in quality improvement , such as reduction of biased data or incentives for better study design ( codes S8–13 ) . A strong debate arose around the question of voluntariness . Interviewees did not agree whether voluntary registration would yield a database contributed to and used by many people ( code F27 ) or whether enforcement , e . g . , by publishers or legislation , was needed to create a well-populated registry useful to the research community as a whole ( code F19 ) . The example of clinical study registration , in which after the publication of the ICMJE statement on registration as a prerequisite for journal publication an enormous increase in clinical trial registration was observed [30] , suggests that some kind of incentive or enforcement is needed to push forward the implementation of such a tool . A point mentioned by some of the animal researchers was the fear of disadvantages in funding or career development , especially for scientists appearing with many “negative” results or failed studies in such a registry ( codes 15a and 15b ) . This often led to comments on the lack of a proper “culture of error” in preclinical research ( codes 14a and 14b ) . Many interviewees affirmed this comment and highlighted in this context that registries could help to shed more light on the obvious issue that research only improves via failures [31] . The feared transparency of failure could also influence the creativity issue , because researchers might prioritize “safe/low-risk” research questions rather than “innovative/high-risk” questions in order to avoid “negative” registry entries . The question of time and personnel resources also emerged as a possible barrier due to the lack of time needed to effectively use a registry both by researchers ( codes B6 and B30 ) or by RECs ( code B28 ) . However , at least among the researchers , many interviewees emphasized the possible time savings from avoidance of studies that had already been performed or experimental set-ups that have proven unsuccessful ( see “Uncontroversial Potential Strengths” ) . Some researchers feared the irrational use of the registry by animal rights activists ( B1c ) and therefore would prefer a registry that either doesn’t show names of the scientists or that is generally not open to the public ( codes F25 and F26 ) . Other interviewees countered these proposals with the arguments that names of animal researchers and the experiments they have performed are increasingly publicly available , e . g . , through open access publications , and that transparency and proactive information of the public may increase public trust ( codes B2b and B2c ) . As a more general barrier , some of the potential strengths of a registry , such as networking possibilities , better visibility , and improved resource allocation , seem to be more attractive to young researchers rather than established group leaders for whom the benefits might be outweighed by disadvantages , such as the possible loss of competitive edge and the fear of more standardization , regulation , and administration ( codes S19c , F22 , and B18 ) . A similar effect was mentioned in the industrial realm , in which the big , established companies are more likely to see competitive disadvantages in a registry , whereas for smaller companies , the benefits might prevail ( code F33 ) . In summary , our interview study showed that there is broad interest among all stakeholders in increased transparency in preclinical animal research . ASRs might play a crucial role in this regard . As usual , the devil is in the detail , and it depends on the registry structure and on implementation and framing conditions how well this tool would balance potential strengths and weaknesses and how it would be accepted in the scientific community . Furthermore , the question arises whether there are subgroups of animal studies that are more or less suited for registration , or that benefit the different stakeholder groups distinctly . Although some kind of regulation may be needed to put this into practice , it is also important to protect the interests of the affected stakeholder groups , maybe by setting a confidentiality time frame in which prospectively registered information is not accessible to others and competitive advantages are not compromised . Whether more transparency via ASRs could speed up the process of drug development is hard to predict . This , of course , would be in line with the interests of all stakeholders and would add to current developments in the pharma industry to stop the expensive development of ineffective pharmaceuticals as early as possible . Finally , there are already efforts from the pharma industry similar to the idea of ASRs , albeit only for certain fields , such as toxicology ( e . g . , Registry of Industrial Toxicology Animal-data [RITA]; see http://reni . item . fraunhofer . de/reni/public/rita/ ) . As a next step , pilot registries could be tested to assess the kind of information and the level of detail needed in an effective and efficient ASR . As one interviewee said about ASRs , “One cannot kill good ideas , and the idea of transparency is a great one . ” Improved transparency is currently being discussed in several research domains [29 , 30] , but of course the research community needs time to become familiar with the associated concepts . Therefore , the stakeholders involved in animal research and affected by an ASR implementation should take the chance to participate in the discussion and to shape the future of their field .
The Hannover Medical School REC approved the study , and all interview participants provided written informed consent . An exploratory search of relevant literature was made using PubMed in June 2015 with two search strings , “preclinical stud* regist*” and “animal stud* regist* , ” resulting in 175 and 388 hits , respectively . All titles were screened to identify papers addressing the issue of ASRs or closely related issues . Further papers were obtained via consultation with experts in the field . Snowball strategies ( reference check and citation check ) for the included papers were applied using Scopus and Google Scholar but did not reveal other relevant literature . Because of the small number of finally included references ( n = 11 , see S2 Table ) , of which the majority did not provide detailed information but merely stated that an ASR was a potentially important means of addressing the problem of publication bias in preclinical studies , we conducted key-informant interviews and used them as our main data source for the stakeholder analysis . We performed semi-structured , open-ended interviews with “key informants , ” that is , persons being an “expert source of information” [32] . The following five criteria helped to define key informants: role in the community , knowledge , willingness , communicability , and impartiality [32] . We invited experts from different stakeholder groups that proved via authorship in relevant publications that they are “knowledgeable . ” From workshops on the topics , such as the NC3R workshop in 2015 , we were also able to identify potential interview participants that are “communicable” and “willing” to express their viewpoints on the topics . Our key informants belonged to four different stakeholder groups , intended to represent the groups most affected by or most influential on the implementation of an ASR , and were selected by purposive sampling . Our sampling was purposive , as we aimed to address the diversity of existing viewpoints and relevant expertise , and we aimed to interview “information-rich” stakeholders . In our case , we wanted to recruit key actors from industry , targeting product developers/researchers , as well as people from overarching organizations . For researchers , it was important to include different research areas as well as career stages . The interview guide ( S1 Text ) started with open questions on potential strengths and weaknesses of ASRs from the viewpoint of the respective stakeholder group . Over the course of the interview , further questions regarding potential facilitators of and barriers to ASR implementation were added . The interview guide was discussed with external experts and pilot-tested for feasibility and acceptability in two cognitive interviews . All interviewees were sent the same invitation and signed a written consent form ( S2 Text and S3 Text ) . Of the 21 interviewees , 14 were from Germany , 3 from Great Britain , 2 from the United States , 1 from Singapore , and 1 from Canada . Interviews were performed in person or via telephone and lasted 40–45 minutes . 13 interviews were conducted by DS , 5 by SW , and 3 by DSS . All interviews were audiotaped and transcribed . To extract , analyze , and synthesize the relevant information on SWFBs for ASRs , thematic text analysis was applied to all 21 transcripts and 11 references using MaxQDA [33] . See Table 1 for our definitions of SWFB . Having conducted 15 interviews , first , 3 transcripts were systematically analyzed by all authors independently . Interview passages mentioning SWFBs were identified , and a descriptive code was applied . Second , the findings were compared to identify potential differences in coding . However , only minor differences occurred and were solved by discussion . Third , aspects mentioned in one interview were matched with those from another in order to collate the various codes and to cluster the findings into an initial matrix of categories and subcategories for SWFBs . This matrix served as a starting point for the further thematic analysis of the other 12 transcripts and the 11 references . One researcher ( SW ) employed the above-described approach to add and modify codes until preliminary thematic saturation was achieved for the main categories and first-order subcategories . Thematic saturation implies that no new categories ( themes ) can be generated for the SWFB matrix , which is itself the primary result of the thematic analysis [34] . This resulted in a matrix of broad and narrow categories for SWFBs . Another researcher ( DS ) then checked all transcripts and references and the resulting matrix and proposed changes . After agreement on the preliminary SWFB matrix , we conducted six more interviews to verify that we had reached thematic saturation . The analysis of the six additional interviews added information that could be grouped under the already existing categories , and we also slightly modified the wording of existing categories . However , the second round of interviews did not result in additional categories or major modifications to the SWFB matrix . We thus confirmed thematic saturation . All researchers discussed and slightly modified the matrix for internal consistency and agreed the final matrix . | The manifold contributions over the last years on “publication bias” and “reproducibility crisis” in animal research initiated a debate on whether and how prospective animal study registries ( ASRs ) should be established in analogy to clinical trial registries . All recent debate , however , followed rather broad lines of argumentation and concluded that future decision-making on the issue of ASRs depends strongly on better knowledge about relevant characteristics of ASRs and about conflicting stakeholder interests . More qualitative but systematically developed evidence in this regard is needed . The primary objective of this study , therefore , was to present a systematically derived spectrum of all relevant strengths , weaknesses , facilitators and barriers ( SWFBs ) for ASRs . A systematic literature review and 21 key-informant interviews with experts from preclinical and clinical research , industry , and regulatory bodies were conducted to fulfill this objective . Our investigations resulted in a comprehensive and structured account of 130 issues and arguments around ASRs . Future debate and decision-making on ASRs might be heavily influenced by arguments and reasoning from individual experts and thus result in “eminence-based” policy making that relies on expert opinion . This study’s comprehensive spectrum of arguments and issues around ASR , developed through systematic and transparent methods , helps to balance the ongoing debate and thus facilitate a more evidence-based policy making . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
] | [
"research",
"funding",
"medicine",
"and",
"health",
"sciences",
"decision",
"making",
"engineering",
"and",
"technology",
"social",
"sciences",
"neuroscience",
"commercial",
"law",
"intellectual",
"property",
"clinical",
"medicine",
"cognitive",
"psychology",
"cognition",... | 2016 | Animal Study Registries: Results from a Stakeholder Analysis on Potential Strengths, Weaknesses, Facilitators, and Barriers |
Non-structural protein 2 ( NS2 ) plays an important role in hepatitis C virus ( HCV ) assembly , but neither the exact contribution of this protein to the assembly process nor its complete structure are known . In this study we used a combination of genetic , biochemical and structural methods to decipher the role of NS2 in infectious virus particle formation . A large panel of NS2 mutations targeting the N-terminal membrane binding region was generated . They were selected based on a membrane topology model that we established by determining the NMR structures of N-terminal NS2 transmembrane segments . Mutants affected in virion assembly , but not RNA replication , were selected for pseudoreversion in cell culture . Rescue mutations restoring virus assembly to various degrees emerged in E2 , p7 , NS3 and NS2 itself arguing for an interaction between these proteins . To confirm this assumption we developed a fully functional JFH1 genome expressing an N-terminally tagged NS2 demonstrating efficient pull-down of NS2 with p7 , E2 and NS3 and , to a lower extent , NS5A . Several of the mutations blocking virus assembly disrupted some of these interactions that were restored to various degrees by those pseudoreversions that also restored assembly . Immunofluorescence analyses revealed a time-dependent NS2 colocalization with E2 at sites close to lipid droplets ( LDs ) together with NS3 and NS5A . Importantly , NS2 of a mutant defective in assembly abrogates NS2 colocalization around LDs with E2 and NS3 , which is restored by a pseudoreversion in p7 , whereas NS5A is recruited to LDs in an NS2-independent manner . In conclusion , our results suggest that NS2 orchestrates HCV particle formation by participation in multiple protein-protein interactions required for their recruitment to assembly sites in close proximity of LDs .
Chronic infection with the hepatitis C virus ( HCV ) is amongst the most frequent causes of liver cirrhosis and hepatocellular carcinoma [1] . About 3% of the world population is persistently infected with this virus and inspite of significant decline of new infections , owing to the long incubation period , a profound rise in the frequency of long-term complications such as steatosis , cirrhosis and liver cancer is expected [2] . HCV is the predominant member of the genus Hepacivirus in the family Flaviviridae . These viruses are enveloped and possess a single strand RNA of positive polarity . In case of HCV the genome has a length of ∼9 . 6 kb and it encodes a single polyprotein that is cleaved co- and post-translationally by cellular and viral proteases into 10 different products [3] , [4]: core , envelope protein 1 ( E1 ) , E2 , p7 , nonstructural protein 2 ( NS2 ) , NS3 , NS4A , NS4B , NS5A and NS5B . Core , E1 and E2 are the main viral constituents of the HCV particle . P7 and NS2 are essential ‘co-factors’ for virus assembly [5] , [6] , but dispensable for RNA replication [7] . This process is catalyzed by the concerted action of NS3 to NS5B proteins forming –together with cellular proteins- a membrane-associated replicase complex [8] . Studies of HCV assembly and release have become possible with the identification of the genotype 2a isolate JFH1 that efficiently replicates in the human hepatoma cell line Huh-7 and supports production of infectious virus particles [9] . This culture system has been improved by the identification of virus titer-enhancing mutations increasing infectivity yields by up to 1 , 000-fold [10]–[12] and the construction of JFH1 chimeras in which the region encoding core to NS2 has been replaced by analogous genome fragments from other HCV isolates [13] , [14] . With the advent of these cell culture systems , first insights into HCV assembly and the roles of p7 and NS2 in this process could be gained . P7 is a small hydrophobic protein composed of two transmembrane segments ( TMS ) [15] , [16] . It is capable to form hexa- or heptameric complexes that can act as a viroporin [17]–[19] . P7 is dispensable for RNA replication [7] , but crucial for infectivity in vivo [20] likely because of its critical role in virus particle assembly [5] , [21] . Whether p7 is a component of the virion is discussed controversially [21] , [22] . NS2 is a 217 amino acids ( aa ) long cysteine-protease composed of a highly hydrophobic N-terminal membrane binding domain ( MBD ) and a C-terminal globular and cytosolic protease subdomain . The latter is capable to form dimers creating a composite active site [23] . This protease is not directly required for RNA replication , but has to be cleaved off the N-terminus of NS3 to allow formation of an active replicase [24] . Recently it was shown that NS2 is essential for HCV assembly [5] , [6] . Interestingly , protease activity is not required for particle formation , but rather global integrity of both NS2 subdomains [25] , [26] . Although the precise mode-of-action of NS2 during HCV assembly is not known , a recent study suggests that this protein acts at a late stage of infectious particle formation [27] . The exact membrane topology and architecture of the N-terminal MBD of NS2 is not known . It might be composed of three trans-membrane segments ( TMS ) [28] , but alternative models are possible . We have recently shown that TMS1 ( aa 1–23 ) adopts an overall helical fold interrupted by flexible glycine residues at position 10 and 11 [6] . While TMS-1 is clearly predicted as a single membrane-spanning trans-membrane helix , this is less clear for TMS2 and TMS3 that may span the membrane bilayer or reside on its cytosolic surface in a helix-loop-helix configuration . By using combinations of reverse genetic and biochemical approaches , convincing evidence has been obtained that also factors of the viral replicase are essential for particle formation , most notably NS3 and NS5A [25] , [29]–[32] . The latter is a highly phosphorylated RNA binding protein composed of an N-terminal amphipathic alpha-helix serving as a membrane anchor and contributing to targeting of the protein to lipid droplets ( LDs ) [33] , [34] , and three domains [35] . Domain I forms a dimer and is essential for RNA replication [36] . Most of domain II is dispensable for replication [30] whereas the C-terminal domain III is essential for virus production , most likely via interaction with the core protein [32] . This interaction appears to be regulated by casein kinase II-mediated phosphorylation of NS5A [29] . Assembly of HCV particles is tightly linked to lipid metabolism , LDs and the machinery required for production and secretion of very-low-density lipoproteins ( VLDL ) [31] , [37]–[39] . Several models of HCV assembly have been put forward , but the precise details are unknown ( reviewed in [40] ) . While these models can explain the early steps of nucleocapsid formation , it is unclear how these nucleocapsids acquire the membranous viral envelope and the envelope glycoproteins and how this process is linked to VLDL formation and secretion . NS2 may play a central role in these reactions , but the precise mechanisms are not known [25] , [27] . In this study we undertook a detailed structural and functional characterization of the N-terminal MBD of NS2 . We solved the NMR-structures of TMS2 and TMS3 and propose a model of NS2 membrane topology . In addition , we performed a structure-activity study of the MBD and established an interaction map of NS2 . The data reveal that NS2 serves as a key organizer participating in multiple protein-protein interactions that are required for the assembly of infectious HCV particles .
We reported recently that a transmembrane segment denoted TMS1 was almost invariably predicted in the very N-terminal region ( aa 1–23 ) of NS2 , irrespective of the analyzed genotypes and subtypes [6] . TMS in the 23–102 region ( [6] and references therein ) yielded inconsistent results that depended both on the genotype examined and the method used ( data not shown ) . By using secondary structure predictions and the algorithm developed by Wimley and White to calculate the propensity of an aa sequence to interact with membranes ( Figure S1 , A and B ) we could deduce that the consensus segments 17–45 and 72–96 exhibit a clear propensity to partition into the membrane bilayer and likely include transmembrane helical passages ( Figure 1A and supplementary Figure S1 ) . In contrast , the aa segment 49–71 is predicted not to show such properties . Based on these results , the NS2 MBD sequence was divided into the three segments: 1–27 , 27–59 , and 60–99 , each containing a putative transmembrane helix ( Figure 1A ) . To determine the capacity of these segments to associate with membranes , we analyzed proteins comprising full length NS2 or putative NS2 TMS that were C-terminally fused to green fluorescent protein ( GFP ) by fluorescence microscopy . The NS2-GFP fusion protein showed a fluorescence pattern that included the nuclear membrane , was strongest in the perinuclear region , and extended in a reticular pattern throughout the cytoplasm ( Figure 1B ) . This pattern corresponds to the endoplasmic reticulum ( ER ) , as corroborated by the colocalization with protein disulfide isomerase ( PDI ) . Each of the predicted TMS of NS2 showed a very similar subcellular localization whereas GFP expressed individually was diffusely distributed throughout the cell including the nucleus . These observations indicate a clear propensity of each of the three segments to associate with membranes . To gain insight into the structure and membranotropic properties of NS2 segments 27–59 and 60–99 , the corresponding peptides of the Con1 strain ( 1b ) designated NS2[27]-[59] and NS2[60-99] were chemically synthesized , purified to homogeneity , and their structures were analyzed by circular dichroism and nuclear magnetic resonance in membrane mimetic environments ( for details see supplementary Figure S1 and materials and methods S1 ) . The 3D model structures obtained for both peptides identified one α-helical segment in case of NS2[27]-[59] and three well defined helical segments in case of NS2[60-99] ( Figure 1C ) . Based on physicochemical considerations , a transmembrane orientation of the amphipathic α-helix in TMS2 could only be achieved upon interaction with another complementary TMS neutralizing the polar and charged residues located in the hydrophobic core of the membrane . In case of NS2[60-99] , the 35 aa segment including all three helices would be too long for a single transmembrane passage given the average length of transmembrane helices ( 16 to 25 aa ) . As the most hydrophobic stretch extends between aa 82–93 , and considering that both edges of this stretch include large hydrophobic residues , we assume that segment ∼77–97 forms a TMS . The first small helix ( 64–69 ) , which includes the short hydrophobic stretch VILL might be located in the membrane interface , possibly in-plane of the membrane . These considerations together with the available structural data for NS2 TMS1 [6] allow us to propose a model for the membrane association and topology of NS2 MBD ( Figure 1D ) . It would contain three transmembrane , mainly helical segments ( TMS1: 4–23; TMS2: 27–49; and TMS3: 72–94 ) , connected by a small cytosolic loop ( aa 24–26 ) and by a luminal segment ( aa 50–71 ) containing a short helix supposed to interact with the membrane interface . Although in this model the three TMS and protease ectodomain are represented as separated entities , given the dimeric structure of the protease domain [23] we expect a packed overall NS2 structure and eventually higher-order complexes of NS2 dimers mediated by intermolecular interactions between MBDs . To correlate the structure of NS2 MBD with function , we conducted an extensive mutagenesis of this domain based on the NMR secondary structures in order to identify residues and structural determinants that are most critical for HCV assembly without affecting RNA replication . Mutants with a very low , but still detectable assembly competence were then used to select for pseudoreversions capable to rescue the assembly defect , which was achieved by serial passage of virus in Huh7 . 5 cells . We anticipated that these pseudoreversions would reside either within NS2 , which might be used to refine the structure model , or in other viral proteins that thus would be candidates for interaction with NS2 . For the reverse genetic studies we used the JFH1 derivative JFH1mut4-6 containing three mutations ( V2153A and V2440L in NS5A and V2941M in NS5B ) elevating virus titers close to the level of the highly efficient chimera Jc1 ( ∼106 TCID50/ml ) without affecting RNA replication [10] ( Figure 2A ) . We chose JFH1mut4-6 for several reasons: first , assembly efficiency of the parental JFH1 genome is very low ( ∼103 TCID50/ml ) , which precludes its use to select for pseudoreversions; second , the titer enhancing mutations reside in the replicase , thus avoiding possible effects on the structural proteins and the assembly factors p7 and NS2; third , the Jc1 chimera has a cross-over site of two different HCV genomes within NS2 [14] , which may confound phenotypes caused by mutations within NS2 . In the first set of experiments we generated a panel of NS2 mutants in which individual α-helices of the different TMS of JFH1mut4-6 were replaced by those of the genotype 1b isolate Con1 . These ‘helix-swap’ mutations should not affect the secondary structure and thus preserve overall structure , folding and topology of NS2 , but disrupt genotype-specific protein-protein interactions . Importantly , since each of the exchanged α-helices of JFH1 and Con1 differ by several aa residues the risk to select for revertants rather than pseudorevertants was very low . Based on the NMR structures reported earlier [6] and in this study , we constructed 7 helix-swap mutants ( aa sequences of affected helices are boxed in Figure 1A ) : JFH1-CT1 . h and CT2 . h in which we exchanged the α-helix of TMS1 ( aa 11–22 of NS2 ) or TMS2 ( aa 34–46 ) , respectively; JFH1-CT3 . h1 , JFH1-CT3 . h2 and JFH1- CT3 . h3 in which the individual short α-helices of TMS3 were exchanged ( aa 62–69; 75–85 and 89–97 , respectively ) . In addition , we generated constructs JFH1-CT3 . h12 and CT3 . h23 in which two short α-helices of TMS3 were exchanged at the same time ( aa 62–85 and 75–97 , respectively ) . As shown in Figure 2B , save for the mutant in which α-helix1 in TMS3 was exchanged ( mutant JFH1-CT3 . h1 ) , all mutants were profoundly impaired in virus production and infectivity titers were reduced up to 1 , 000 fold at 72 h after transfection of Huh7 . 5 cells . This impairment correlated with the degree of sequence conservation between the exchanged helices . Aa sequence alignments revealed that Con1 – JFH1 sequence similarities were in the range of 58–81% , but in case of helix1 of TMS3 similarity was only 37% . Since this helix swap mutant was unaffected in assembly , this region most likely is required for interactions with the membrane and thus genotype independent . Analysis of intra- and extracellular infectivity revealed that in all cases reduced titers were due to impaired assembly rather than virus release ( Figure 2C ) . With the exception of mutant CT3 . h23 producing lower amounts of NS2 assembly defects could not be ascribed to gross alterations of NS2 abundance ( Figure 2D ) . The protein of higher molecular weight detected with mutant JFH1-CT2 . h corresponded to uncleaved p7-NS2 arguing for a processing defect of this mutant , which was not the case for all the other mutants . In addition , a distinct product of smaller size ( about 16 kDa ) was detected on longer exposure ( not shown ) and this protein most likely corresponds to an N-terminal cleavage product of NS2 designated tNS2 [6] ( see below ) . In summary , these results show that the integrity of the N-terminal MBD of NS2 is important for HCV assembly and that all 3 TMS are required . To further narrow down the regions within individual TMS of NS2 that are most crucial for assembly , we generated a panel of single aa substitutions that were designed on the basis of the degree of conservation across the different genotypes , aa size , charge , polarity and hydrophobicity ( Figure 1A ) . Large and hydrophobic aa were replaced by smaller and less hydrophobic ones ( Y26P , Y39A , W51A , F77A or LL83-84AA ) ; small aa probably serving as flexible linkers between individual α-helices were replaced by larger aa assumed to reduce flexibility of the N-terminal MBD ( G10A/S/T/P , P53I , P73I and G88L ) ; charged aa potentially involved in electrostatic interactions were replaced by aa with opposite charge thus possibly introducing repulsive forces ( K27E , E45R , R58E , R61E and K81E ) ; finally , the polar threonine residue at position 80 was replaced by the larger and hydrophobic aa leucine . All mutants were tested for protein expression , kinetic of infectivity release as well as amounts of intra- and extracellular infectivity . The results shown in Figure 3A and B demonstrate that both infectivity release and intracellular infectivity levels were reduced with all mutants , albeit to very different degrees . In agreement with our earlier results demonstrating the important role of the glycine residue at position 10 for virion assembly [6] , we found that mutants G10P and G10T did not support virus production and even less drastic alanine or serine substitutions reduced infectivity titers up to 300-fold . Substitutions residing in loop1 that connects α-helix 1 and 2 strongly reduced or completely abolished production of infectious HCV particles ( Y26P and K27E , respectively ) . In contrast , aa substitutions in loop2 ( W51A , P53I , R58E and R61E ) slowed down the kinetic of infectivity release , which was best detected at 24 h post transfection , whereas intra- and extracellular infectivity titers were reduced only moderately at later time points as compared to the parental construct JFH1mut4-6 ( wt ) . Alanine substitutions of aromatic aa residues in TMS2 ( Y39 ) or TMS3 ( F77 ) reduced infectivity titers up to 1 , 000-fold whereas mutations introducing electrostatic repulsion in TMS2 or TMS3 ( E45R and K81E , respectively ) blocked virus production almost completely . Substitutions targeting the flexible region between helix one and two or helix two and three in TMS3 ( P73I and G88L , respectively ) strongly reduced infectivity titers ( 1 , 000-fold at 72 h p . e . ) . Surprisingly substitutions affecting the highly conserved polar aa residue at position 80 ( T80L ) or the two leucine residues at positions 83 and 84 ( LL83-84AA ) did not give rise to a detectable phenotype . Interestingly , as shown in Figure 1D , residues 80 , 83 and 84 are located on one helix side , suggesting that it is not important for protein – protein interactions . In contrast , residues 77 and 81 are located on the opposite side , arguing that this helix side might be involved in interactions . Western blot analysis of NS2 proteins expressed in cells after transfection with each of the mutants or the parental construct revealed no gross difference in the abundance of this protein and the other HCV proteins arguing for similar replication levels and protein stabilities ( Figure 3C ) . Nevertheless , some variations in abundance of individual HCV proteins were detected such as lower amounts of NS2 in case of K81E and reduced amounts of core protein in case of LL83AA . However , these rather subtle differences are very unlikely to account for the often drastic impairment of HCV assembly . In agreement with an earlier report , ‘truncated NS2’ [6] was detected to variable levels , but its abundance did not correlate with assembly phenotypes . Given the important role of tryptophan in protein-protein interaction within a membrane [41] and their preferred location at the membrane interface [42] , [43] , we also analyzed a panel of mutations affecting the two fully conserved W35 and W36 residues . A striking correlation was found between reduction of aromaticity as well as size of residues at these sites and reduction of virus production ( data not shown ) arguing that the aromatic side chains of W35 and W36 are involved in essential interactions such as membrane tethering of TMS2 via aromatic ring stacking . Having generated a panel of NS2 mutants with a selective assembly defect , and –in some cases- an additional virus release defect ( e . g . Y26A; Figure 3B ) we wanted to establish a genetic interaction map of NS2 . For this purpose we used a cell culture adaptation strategy , which was possible , because the parental construct JFH1mut4-6 that was used for mutagenesis already supports high level virus production and therefore , selection for pseudoreversion would not give rise to undesired mutations enhancing assembly in general [10] . Culture supernatants collected from cells 72 h after transfection with a given NS2 mutant were concentrated and used to infect naive Huh7 . 5 cells that were passaged 6 times . After 4 additional passages of culture supernatants , they were used to inoculate naïve Huh7 . 5 cells and virus titers produced therefrom were determined by TCID50 assay . In case of mutants with elevated virus titers , cell lysates were used to prepare total RNA , HCV genomes were amplified by RT-PCR and amplicons spanning most of the 5′ NTR up to the middle of NS3 were either directly sequenced or cloned prior to sequence analysis . In the latter case at least two independent clones were analyzed and only mutations conserved between the two cDNA clones were considered in order to discriminate against mutations that might have been introduced by PCR . Pseudoreversions outside the analyzed region , including NS5A , were not considered because we used the JFH1mut4-6 genome that already contained titer-enhancing mutations in NS5A to allow adaptation . Mutations identified by this approach were inserted into the corresponding parental NS2 mutant and replication as well as assembly properties were analyzed by Western blot and TCID50 assays . A summary of all pseudoreversions identified in this way along with their degree of titer enhancement is given in Table 1 . Several assembly deficient mutants ( G10T , G10P , K81D , G88L and JFH1-CT3 . h23 ) could not be adapted , because the virus was rapidly lost during cell passages suggesting that the genetic barrier was too high and assembly impairment too strong . Nevertheless , for most mutants we could select for pseudoreversions with the exception of G10A and K81E where reversion to wild type occured , which was achieved by just one nucleotide substitution . Since we had inserted in addition a silent nucleotide exchange in the subsequent codon that was retained in the selected virus , we could rule out a contamination with wild type virus . All of the other adapted mutants contained pseudoreversions . They resided primarily within NS2 and two ‘classes’ of pseudoreversions could be discriminated: first , those residing at the same position as the primary mutation , but with a different substituting aa residue; second , pseudoreversions at a different site than the primary mutation . Pseudoreversions belonging to the first class ( R45G and I73S ) enhanced virus production 88 , 000- and 3 , 100-fold and thus back to the level of the parental genome JFH1mut4-6 ( Figure 4A and Table 1 ) . Pseudoreversions belonging to the second class ( double mutants G10S-T23N , W36F-Q32R and Y39A-G25R ) increased infectivity titers ∼220- , 450- and 10-fold , respectively , with G10S-T23N also reaching wild type levels ( Figure 4A and Table 1 ) . For this reason the T23N substitution was also combined with mutants G10T and G10P , but titer increase was very moderate ( Figure 4A ) . In addition to pseudoreversions in NS2 , one particular mutation was identified in NS3 ( Q221L ) that rescued infectivity titers of the W35F and the W36L mutants by 1 , 400- and 38 , 000-fold ( Figure 4B ) . This NS3 mutation has previously been described as a general titer enhancing substitution [25] , [27] . For this reason we inserted the Q221L mutation into mutants W35A , W35L and W36A that are completely defective in assembly and that could not be adapted with our approach . As shown in Figure 4B , infectivity titers were enhanced in all cases corroborating the more general assembly enhancing phenotype exerted by this particular NS3 mutation . Moreover , the same mutation was also capable to rescue virus production of several NS2 mutants even in trans ( supplementary Figure S2 ) . Although in this case rescue efficiency was lower as compared to direct insertion of the mutation into the NS2 mutant genome , this observation suggests that a genetic separation of the replication and assembly function of NS3 is possible . In case of the F77A mutation residing in TMS3 . 2 of NS2 two mutations located in E2 could be selected: Y215S located in the ectodomain and V341A in the N-terminus of the TMS of E2 ( Table 1 ) . When we tested these substitions in the context of the parental NS2 mutant we found that Y215S completely abrogated infectious virus production concomitant with a strong reduction in intracellular NS2 amounts ( Figure 4A ) . However , the V341A substitution enhanced virus titer ∼100-fold arguing for a ( direct or indirect ) interaction between these two proteins for efficient particle production . A more complex pattern of mutations was found upon selection for pseudoreversions in case of helix-swap mutants JFH1-CT1 . h , JFH1-CT2 . h , JFH1-CT3 . h2 and JFH1-CT3 . h3 ( Table 1 and Figure 4C ) . For CT1 . h , we found E3D residing in the N-terminal helix of p7 [16] , F14L in TMS1 of NS2 and I17M in the membrane-binding amphipathic helix α0 of the NS3 protease domain . When inserted into the parental virus , strongest rescue of assembly was found with the p7 mutation and infectivity titers were further increased 10-fold at 24 h or 2-fold at 72 h p . t . when combined with the NS2 pseudoreversion ( CT1 . h - p7NS2 ) . In contrast , the NS3 mutation had a slightly negative effect ( CT1 . h-NS3-I17M ) . For the JFH1-CT2 . h mutant we detected two pseudoreversions residing in the turn connecting the N-terminal helix and the TM1 helix of p7 ( N15D ) [16] and the loop connecting TMS1 and TMS2 of NS2 ( G25R ) . In this case , the mutation in NS2 rescued HCV assembly strongest ( ∼265-fold ) whereas the p7 mutation had a very moderate effect and in combination with the NS2 mutation did not enhance virus titers further . In case of pseudoreversions detected with helix-swap mutants JFH1-CT3 . h2 and JFH1-CT3 . h3 , the only titer enhancing mutations were found in NS2 ( K172R , T21A ) with T21A arguing for an interaction between TMS1 and TMS3 of NS2 . The double mutation identified in E1 of JFH1-CT3 . h2 construct ( E151D/I181S ) had no effect . For all tested single aa mutants and helix-swap mutants and their corresponding pseudorevertants , the enhancement of infectivity titers in cell culture supernatants correlated with increased amounts of intracellular infectivity showing that the pseudoreversions rescued primarily assembly rather than virus release ( supplementary Figure S3 ) . Moreover , with the exception of the F77A-E2-Y215S double mutant , amounts of NS2 as well as NS5A , NS3 and core protein were not grossly altered ( Figure 4 and supplementary Figure S4 , respectively ) suggesting that overall protein stabilities were not profoundly affected by the mutations . We note however that for mutant F77A already producing somewhat lower amounts of NS2 as compared to the wildtype , NS2 abundance was reduced much more by the Y215S substitution in E2 . Moreover , in case of helix-swap mutant CT2 . h and the corresponding pseudorevertants cleavage between p7 and NS2 was impaired ( Figure 4C ) . Interestingly , even in case of rescue mutant CT2 . h-NS2 , the substitution in NS2 enhancing virus production about 250-fold does not affect the amounts of this uncleaved precursor arguing that assembly competence of this mutant is restored in a manner that still allows delayed p7-NS2 cleavage , thus compensating e . g . an impaired p7-NS2 interaction ( see below ) . Although in most cases , selection for pseudoreversion resulted in compensatory mutations within NS2 itself , we also identified pseudoreversions in E2 , p7 and NS3 . In case of helix-swap mutant CT1 . h , the E3D substitution in p7 restored almost wild type infectivity titers ( Table 1 ) . Likewise , the assembly defect of mutant W35F in NS2 was compensated by Q221L in NS3 and NS2 mutant F77A was compensated by the V341A substitution in E2 . These results suggested that NS2 might interact ( directly or indirectly ) with each of these proteins . To support this assumption by pull-down assays , we first generated a fully functional JFH1-derivative with a tagged NS2 protein suitable for efficient immunoprecipitation and allowing capture of NS2 independent from any mutation that might affect recognition with the NS2-specific antibody . To this end we constructed a series of mutants in which NS2 was fused N- or C-terminally with several tags such as the FLAG- , hexa-histidine ( His ) - or hemagglutinin ( HA ) -tag . All genomes with C-terminal fusions no longer supported HCV particle production ( data not shown ) . Moreover , when we tried to select for titer enhancing pseudorevertants of assembly deficient NS2-tagged mutants , in all cases the tag was partially or completely deleted ( not shown ) . In contrast , viable mutants were obtained with N-terminally tagged NS2 versions in which the first 5 codons of NS2 were duplicated upstream of the heterologous sequence that was composed of a single copy of the tag and a linker sequence encoding for Gly-Ser-Gly preceeding NS2 . In addition , variants were generated with a second insertion of the tag sequence to increase efficiency of immunoprecipitation ( Figure 5A ) . Analysis of the kinetics of virus production revealed that both the single Flag-tagged ( F-NS2 ) and the HA-Flag-double tagged variant ( HAF-NS2 ) produced amounts of intra- and extracellular virions that were comparable to the parental genome JFH1mut4-6 ( data not shown and Figure 5B ) . In contrast , the variants with the tandem Flag-tag ( FF-NS2 ) or the His-Flag-tag combination ( HisF-NS2 ) produced lower amounts of extra- and intracellular infectious particles arguing for an assembly defect . Western blot analysis revealed comparable replication of all constructs and no defect of polyprotein processing was detected ( Figure 5C ) . The sizes of the various NS2 proteins and their immunoreactivities confirmed that the tag ( s ) remained fused to mature ( fully processed ) NS2 . Taking advantage of these assembly-competent tagged NS2 constructs , we selected JFH1mut4-6HAF-NS2 to determine NS2 interactants . Huh7 . 5 cells were transfected with this construct and NS2-containing immunocomplexes captured from lysates that were prepared 72 h p . t . were analyzed by Western-blot for coprecipitation of core , E2 , p7 , NS3 and NS5A . Specificity of immunoprecipitation and Western blot analysis was determined by using the parental JFHmut4-6 construct that lacked the N-terminal NS2-tag . The results in Figure 5D show that the tagged NS2 protein co-precipitated with E2 , p7 , NS3 and NS5A , but not with core . In contrast , no signal was found in case of the non-tagged genome inspite of comparable amounts of viral proteins in cell lysates , demonstrating specificity of these co-precipitations . To analyze whether the pseudoreversions in E2 , p7 and NS3 affect the NS2 interaction pattern , we chose those NS2 mutants for which infectivity titers were enhanced by a pseudoreversion outside of NS2: CT1 . h and CT1 . h-p7-E3D; W35F and W35F-NS3-Q221L; F77A and F77A-E2-V341A . Lysates of all samples harvested 72 h after electroporation together with positive and negative controls were subjected to HA-specific pull-down and immunocomplexes were analyzed by Western blot ( Figure 6A ) . Pull-down efficiencies were quantified by densitometry scanning and normalized to protein amounts detected in the corresponding cell lysate ( Figure 6B ) ; based on this quantification fold enhancement of coimmunoprecipitation achieved by the pseudoreversion was determined ( Figure 6C ) . For all mutants , NS2 interaction with the other viral proteins was reduced , but to very different extents . Most pronounced was the impairment of NS2 interaction with E2 , p7 and NS3 , whereas interaction with NS5A was less affected . Importantly , the E3D pseudoreversion in p7 introduced into CT1 . h enhanced interaction with E2 and NS3 back to wild type levels correlating well with the rescue of assembly competence of this helix-swap mutant . Unfortunately this pseudoreversion disrupted the epitope recognized by the p7-specific antibody and therefore , the degree of coprecipitation of this p7 with NS2 could not be determined . Interaction of NS2 with NS5A was elevated even above the wild type level ( Figure 6B , C ) . As expected , the Q221L pseudoreversion in NS3 introduced into the W35F NS2 mutant increased NS2 – NS3 interaction , but surprisingly had little or no effect on NS2 interaction with E2 or p7 , respectively , and even a negative effect on interaction with NS5A ( Figure 6B , C ) . The pseudoreversion in the TMS of E2 ( V341A ) introduced into the F77A mutant moderately enhanced interaction of NS2 with E2 and NS3 , but no significant enhancement of interaction with p7 and NS5A was detected . Interestingly , NS2 containing this F77A substitution coprecipitated with both phospho-variants of NS5A to the same extent , whereas all other NS2 proteins tested preferentially interacted with the basal phosphorylated form p56 ( Figure 6A ) . This phenotype of the F77A mutant was not altered by the adaptive mutation residing in E2 . To support and extend the interaction patterns described above with an alternative assay we performed colocalization studies of NS2 with structural and other nonstructural proteins . In the initial set of experiments , we determined the subcellular localization of NS2 ( Figure 7A ) and observed a profound change from a reticular ER staining pattern 36 h post transfection ( NS2 colocalization with the ER marker PDI is not shown ) to a strong punctate NS2 stain accumulating in close proximity of LDs 72 h post transfection . By counting ∼200 cells we defined two phenotypes , based on the number of LDs with NS2 accumulation: phenotype 1 with less than 10 NS2-positive LD structures per cell and phenotype 2 with more than 10 . A time-dependent increase of phenotype 2 was also observed although the overall percentage was lower , which was probably due to lower replication as compared to RNA transfection ( supplementary Figure S5 ) . The functional relevance of these two phenotypes is supported by the analysis of the NS2 mutants and their respective pseudorevertants ( Figure 7B ) . We found that NS2 decorated LDs were much less frequent in an assembly deficient mutant and even 72 h after transfection the majority of NS2 was localized at the ER . Importantly , upon insertion of the corresponding pseudoreversion a shift back to phenotype 2 representing higher abundance of NS2 ‘positive’ LDs 72 h p . t . was detected ( Figure 7B ) . To determine whether other viral proteins might be recruited to LDs in an NS2-dependent manner we performed colocalization studies . As shown in Figure 8A , at each analyzed time point we found a striking colocalization of NS2 and E2 in case of the wild type , consistent with the coimmunoprecipitation results . In addition , we detected a strong accumulation of both proteins around LDs 72 h p . t . ( Figure 8A ) . A similar pattern , but less colocalization as determined by Pearson's correlation coefficient , was found for NS2 with NS3 ( Figure 8B ) . Interestingly , a lower degree of colocalization of NS2 and NS5A predominated 36 h p . t . and NS5A localized in close proximity of LDs independent of NS2 . Accumulation of NS2 around LDs at the later time point coincided with increased NS2-NS5A colocalization at these sites . No significant colocalization was detected between NS2 and core protein at LDs . However , a small amount of core colocalized with NS2 in a reticular , presumably ER-derived compartment ( supplementary Figure S6 ) . Attempts to detect p7 by immunofluorescence were not successful with the available antibodies and insertion of tags into p7 very much impaired assembly ( not shown ) . Therefore , p7 – NS2 colocalization could not be studied . Given the most pronounced loss of NS2 accumulation around LDs ( i . e . low frequency of phenotype 1 ) with mutant CT1 . h we determined for this construct and the corresponding pseudorevertant NS2 colocalization with E2 , NS3 and NS5A as well as HCV protein accumulation at LDs . For the parental NS2 mutant we found that E2 no longer localized to LDs and localization of NS3 to these sites was strongly impaired ( Figure 9A ) . However , recruitment of these viral proteins to LDs and strong colocalization at LDs was restored by the pseudoreversion in p7 ( E3D; Figure 9A , lower panel ) . In contrast , NS5A was recruited to LDs independent from NS2 and NS2 – NS5A colocalization was also restored by this pseudoreversion . When analyzing a larger panel of NS2 mutants and their corresponding pseudorevertants for colocalization of these HCV proteins in a quantitative manner ( Figure 9B ) we found for CT1 . h a slight reduction of NS2 colocalization with E2 , NS3 and NS5A that was partially restored by the pseudoreversion residing in p7 ( CT1 . h-p7 ) . In case of NS2 mutants W35F and F77A only NS2 colocalization with NS5A was impaired , but restored by the corresponding pseudoreversion in NS3 or E2 ( W35F-NS3-Q221L or F77A-E2-V341A , respectively ) . In contrast , colocalization of NS2 with E2 or NS3 was unaffected by these NS2 mutations ( Figure 9B ) . This result could be explained by the fact that the mutations in NS2 might impair accumulation around LDs and thus would lead to accumulation of NS2 at ER membranes where also the majority of E2 and NS3 reside . Therefore , colocalization of these NS2 proteins with E2 and NS3 ( at the ER membrane ) might be strong . In contrast , NS5A is recruited to LDs independent of NS2 and therefore , NS2 mutations that no longer are recruited to LDs might have lower colocalization rate that would be restored by the pseudoreversion that rescues ‘LD targeting’ of NS2 .
By using NMR of synthetic peptides we solved the secondary structures of TMS2 and 3 and propose a membrane topology model of the overall N-terminal MBD ( Figure 1D ) . This model supports and very much extends an earlier report [28] and proposes 3 transmembrane α-helices , connected by flexible loop regions . While TMS1 and 2 consist of one α-helix , TMS3 is composed of three . Each of TMS1 - 3 is capable to mediate membrane association on its own . To determine whether individual helices within TMS3 are sufficient for membrane targeting we analyzed subcellular localization patterns of NS2-GFP fusion proteins comprising NS2 aa residues 60–88 or 74–99 . However , these proteins displayed a predominantly diffuse fluorescence signal , arguing that all 3 helices of TMS3 are required for membrane targeting ( J . G . and D . M . , unpublished ) . The model of three TMS is consistent with homo-intramolecular TMS interactions revealed by the pseudoreversions indicating that TMS1 interacts with TMS2 and TMS3 . The model is also consistent with hetero-intermolecular interactions by which TMS1 and TMS2 could interact with p7 whereas TMS3 could interact with the TMS of E1 and E2 ( Figure 10 ) . Moreover , the fact that point mutations in the long and variable connecting loop between TMS2 and TMS3 had no effect on virus production is in keeping with its ER luminal location . Conversely , the sensitivity to mutation of the small loop between TMS1 and TMS2 is consistent with its cytosolic localization However , it should be stressed that helices observed in TMS2 and TMS3 are not classical membrane anchoring TM helices , since they contain polar and charged residues . In addition , the TMS2 helix exhibits an amphipathic character suggesting that it could associate with the membrane interface , at least transiently . Based on physicochemical considerations , a transmembrane orientation of this helix is expected to be achieved only upon interaction with another complementary transmembrane segment neutralizing the polar and charged residues located in the hydrophobic core of the membrane . In this context , it is possible that the transmembrane association of TMS2 and TMS3 occurs in the translocon during NS2 biosynthesis . Alternatively , these TMS might be first released into the cytosol where they could interact at the membrane interface and then associate with the membrane to adopt their final transmembrane topology . Interestingly , the length of the connecting loop between TMS2 and TMS3 and the absence of an interaction between these TMS suggest that TMS3 might be an independent entity possibly interacting with distant partners . The fact that chimeric genomes with high assembly competence can be obtained when using a cross-over site right after TMS1 of NS2 indicates that TMS1 is functionally separated from the remainder of the NS2 MBD [14] . Overall , the MBD of NS2 appears to be composed of a series of structural elements with own functional properties , but with the capacity to acquire new functions upon intra- and intermolecular interactions . This structural plasticity is likely essential to ensure the multiple interactions mediated by the NS2 MBD . Almost all helix-swap mutations reduced assembly competence arguing for genotype specific incompatibilities between individual TMS of either NS2 or other viral proteins , such as p7 and E2 . The only exception was mutant CT3 . h1 affecting helix1 of TMS3 that acts most likely as an adaptable linker between TMS2 and TMS3 . This helix is the least conserved sequence of the NS2 MBD suggesting that it mediates interactions with the membrane in a genotype-independent manner . Selection of assembly-impaired NS2 mutants for titer enhancing mutations compensating the assembly defect to the most part lead to pseudoreversions within NS2 . This was the case for all helix-swap mutants and several single aa exchanges . Six out of 9 pseudoreversions within NS2 were found in the loop region connecting TMS1 and 2 . This loop resides on the cytosolic side of the ER membrane and by interaction with membrane phospholipids it may stabilize membrane association of NS2 or is involved in intra- or intermolecular protein-protein interactions . Overall , these pseudoreversions most likely restore structural alterations induced by the primary mutation . Based on our model of the NS2 MBD ( Figure 1D ) , at least some of these mutations could be explained . One example is the Y39A pseudoreversion compensating the assembly defect caused by the G25R mutation suggesting that aa residues 25 and 39 , which are located on TMS1 and TMS2 , respectively , might be in contact ( Figure 10 ) . We assume that the “hole” created in TMS2 by the Y39A substitution is compensated by a bulky aa in the interacting TMS1 counterpart , thus ‘filling up’ the hole in the mutated TMS2 . This interaction likely occurs at or close to the membrane interface , where the charge of Arg is well tolerated . Importantly , this assumption is corroborated by the G25R pseudoreversion that was selected with helix-swap mutant JFH1-CT2 . h , which has a histidine residue at position 25 . We therefore conclude that the beginning of the loop between TMS1 and TMS2 likely interacts with the helix in TMS2 . Another example are the pseudoreversions K172R and T21A in NS2 that were selected with helix-swap mutants JFH1-CT3 . h2 and JFH1-CT3 . h3 , respectively , suggesting interactions between TMS1 and TMS3 ( Figure 10 ) . While this can be easily explained for position 21 , the aa at position 172 is more remote from the membrane surface . Nevertheless , this residue is at the junction between the two subdomains of the NS2 protease domain and thus still suitable to contact TMS3 . We tried to integrate all these informations into our NMR-based structure model of NS2 MBD , but these attempts were confounded by the fact that NS2 is a dimer , which most likely forms higher-order oligomeric complexes . Therefore , we do not know whether a given mutation restores intra- or intermolecular interactions . Nevertheless , the tight correlation between structural integrity of NS2 and its role in assembly is underlined by the fact that titer-enhaning mutations within NS2 have also been found by us and others when using JFH1 wild type or various virus chimeras with low assembly competence [10] , [44]–[49] . Apart from pseudoreversions within NS2 , we also identified two in p7 . Importantly , in case of the helix-swap mutant affecting TMS1 , the pseudoreversion in p7 ( E3D ) enhanced virus production almost back to wild type level . This result argues for an interaction between TMS1 of NS2 and p7 ( Figure 10 ) . Unfortunately , this assumption could not be tested directly , because this mutation destroyed the epitope recognized by the p7-specific antibody . However , we have earlier described that for most virus chimeras the best junction for fusion of the genome segments resides after TMS1 of NS2 , whereas an intergenotypic fusion right after p7 was severely impaired in assembly [14] . Thus , genotypic compatibility between TMS1 of NS2 and the structural proteins as well as p7 appears to be required for efficient assembly . A direct interaction between NS2 and envelope glycoproteins might be suggested by the pseudoreversions V341A residing in the N-terminus of the TMS of E2 and mutation I181S in E1 ( Figure 10 ) . We note that V341 in E2 and the primary NS2 mutation F77 responsible for the assembly defect are both most likely located in the membrane hydrophobic core , close to the ER membrane interface ( Figure 1D for NS2 and Figure 1 in [50] for E2 ) . Moreover , I181S in E1 selected as pseudoreversion with the helix-swap mutant CT3 . h2 resides in the center of the TMS of E1 and thus could also directly form a stable in-membrane interaction with NS2 ( Figure 10 ) . W35F and W36L independently adapted via the Q221L pseudoreversion in NS3 that has also been found in two earlier reports [25] , [51] . This reversion is highly potent and restores viral infectivity up to ∼38 , 000-fold . Interestingly , this NS3 mutation also rescues assembly in trans showing that the replication and assembly function of NS3 can be separated genetically . While the mechanism by which Q221L enhances assembly is not known , we note that this residue resides on the helicase NTPase subdomain surface in a basic patch and is well accessible . This positively charged surface area might interact with the membrane surface by electrostatic interactions . In this way the aa residue at position 221 could contact the NS2 MBD at the membrane interface , at least transiently ( Figure 10 ) . According to this hypothesis , the replacement of the polar residue ( Q ) by a large hydrophobic aa ( Leu ) might reinforce membrane binding . Co-immunoprecipitation studies revealed stable interactions of NS2 with NS3 , p7 and E2 whereas interaction with NS5A was rather weak . Importantly , none of the tested conditions revealed NS2 interaction with core . These results were well supported by immunofluorescence studies demonstrating a profound and rapid colocalization of NS2 with E2 and NS3 at the ER or an ER-derived membrane compartment prior to accumulation around LDs . Several lines of evidence suggest that NS2 recruits E2 –and thus most likely also E1 that forms a very stable E1/E2 heterodimer [52]– and eventually also p7 to assembly sites in close proximity of LDs . First , we detected a profound colocalization of NS2 and E2 for each time point after infection or transfection; second , in NS2 assembly-defective viruses E2 localized primarily to the ER; third , upon insertion of the corresponding pseudoreversion E2 and NS2 colocalized again to LDs . The NS2-independent LD localization of NS5A and its weak interaction with NS2 is in agreement with previous data showing that NS5A expressed on its own is targeted to LDs , for which the N-terminal amphipathic helix appears to be most critical [33] . No significant colocalization of NS2 and core at LDs was detected . However , a small fraction of core protein presumably residing at the ER colocalized with NS2 both at the early and the late time points after transfection . Although LDs have been described as sites of HCV assembly [31] the weak NS2 – core colocalization is not in contradiction to this observation . In fact , it is speculated that the early steps of HCV assembly ( nucleocapsid formation ) might take place at LDs whereas the envelopement is thought to occur at the ER or an ER-derived compartment . Since NS2 probably acts at a late step of assembly [27] and might be involved in envelopment of the nucleocapsid , the colocalization of core and NS2 at the ER in close proximity of LDs rather than directly on LDs would support such a model . Moreover , given the complex membrane topology of NS2 this protein most likely can not move onto the surface of LDs that is formed by a membrane monolayer . The results described in this study together with earlier reports [25] , [27] invite speculation how NS2 might contribute to assembly . It is assumed that the early steps ( nucleocapsid formation ) occur in close proximity of LDs that may serve as assembly platforms [31] . By interaction between core and ( RNA-containing ) NS5A , capsid formation might be triggered [32] . How the envelope is acquired is not known , but we assume that NS2 plays a central role in this step . Since the TMS of E1 and E2 lack a cytosolic domain that could interact with the core protein , an adaptor protein such as NS2 that in turn efficiently binds p7 and E2 ( and the latter forming heterodimers with E1 ) , might be required to ‘deliver’ the envelope proteins to assembly sites in close proximity of LDs . This process could be facilitated by a particular membrane lipid environment supporting recruitment of the NS2 complex as well as the ( lipid-binding ) nucleocapsid . Alternatively , one or several host cell factors such as CIDE-B , described as an NS2 interactant [53] and required for lipid homeostasis [54] , might be recruited by NS2 and aid in assembly . Moreover , NS2 efficiently binds to NS3 arguing that NS2 can form an additional complex with the replicase . How such a complex would contribute to assembly is unclear , but it may ‘tether’ the replicase to the assembly sites thus facilitating core – NS5A interaction . Alternatively , NS2 may form just one higher-order protein complex including in addition to E1/E2 and p7 NS3 . This is probably facilitated by the N-terminal MBD that might form ‘clusters’ within the membrane . Finally , it is possible that the strong NS2 - NS3 interaction affects cleavage at the NS2-3 site , in this case contributing to assembly in a rather indirect manner . In conclusion , our results point to a central role of NS2 in HCV assembly by formation of ( a ) multiprotein complex ( es ) with structural and eventually also nonstructural proteins and recruiting them to assembly sites in close proximity of LDs . In this respect , NS2 acts as a central organizer of HCV virion formation .
Sequence analyses were performed by using Network Protein Sequence Analysis ( NPSA ) ( http://npsa-pbil . ibcp . fr [55] ) and European HCV Database ( http://euhcvdb . ibcp . fr [56] ) . Multiple-sequence alignments were performed with CLUSTAL W [57] , by using the default options . Protein secondary structures were deduced from a large set of prediction methods available at the NPSA website , including HNNC , SIMPA96 , MLRC , SOPM , PHD , and Predator ( http://npsa-pbil . ibcp . fr/NPSA and references therein ) . Octanol hydrophobicity plots were generated with MPEx ( http://blanco . biomol . uci . edu/mpex/ ) by using the scale developed by Wimley and White [58] . Monolayers of the highly permissive cell lines Huh7-Lunet [59] and Huh7 . 5 [60] were grown in Dulbecco's modified minimal essential medium ( DMEM; Life Technologies , Karlsruhe , Germany ) supplemented with 2 mM L-glutamine , nonessential amino acids , 100 U/ml of penicillin , 100 µg/ml of streptomycin , and 10% fetal calf serum . Owing to highest permissiveness for JFH-1 , Huh7 . 5 cells were used for virus production and infection assays whereas Huh7-Lunet cells and derivatives thereof were used for immunofluorescence analyses because of their superior morphology as compared to Huh7 . 5 cells . NS2-GFP fusion constructs were derived from pFK1-9605Con1 ( [7]; HCV Con1 strain ) . First , a BamHI restriction site was eliminated by introducing a silent mutation replacing the cytidine at nucleotide position 2920 by an adenosine . Sequences encoding NS2 fragments from codon 1–27 , or 27–59 , or 60–99 , or 1–99 or the complete NS2 coding region were fused to EcoRI and BamHI recognition sequences by PCR and amplified fragments were inserted via these two restriction sites into pCMV-KEB-GFP [61] , yielding constructs pCMVNS21-27-GFP , pCMVNS227-59-GFP , pCMVNS260-99-GFP , pCMVNS21-99-GFP and pCMVNS2-GFP . Unless otherwise stated all mutations were introduced into JFH1mut4-6 [10] corresponding to the JFH1 genome [9] , but containing three virus titer enhancing mutations that do not affect RNA replication ( V2153A , V2440L and V2941M ) . All nucleotide and aa numbers refer to the JFH1 genome ( GenBank accession no . AB047639 ) . Single aa substitutions and helix-swap mutations were introduced by PCR-based site-directed mutagenesis or overlap-PCR , respectively , using standard procedures . In case of the helix-swap mutations the following nucleotide sequences of JFH1 were replaced by the corresponding sequences of Con1: nucleotides 2811 - 2839 in case of pFK-JFH1-CT1 . h; nucleotides 2879 - 2911 for pFK-JFH1-CT2 . h; nucleotides 2967 - 2986 for pFK-JFH1-CT3 . h1; nucleotides 3002 - 3025 in case of pFK-JFH1-CT3 . h2; nucleotides 3047 - 3070 with pFK-JFH1-CT3 . h3 . To generate the JFH1 genome encoding a tagged NS2 protein a sequence encoding the peptide YDAPVSGDYKDDDDKGSG ( corresponding to the first 5 aa of NS2 , a 2 aa flexible linker ( SG ) containig a BspEI site , a Flag tag and a flexible GSG linker ) was inserted by overlap PCR between nucleotide 2779 and 2780 of the JFH1 genome . A silent G to A mutation at position 2794 was introduced to create a BsrGI restriction site whereas the natural BsrGI site at position 7786 was destroyed by a silent A to T mutation . In addition , a silent A to C nucleotide substitution at position 1741 was introduced to create a dam methylation site affecting the BspEI cleavage site at this position . To generate genomes with double tagged NS2 , oligonucleotides encoding the Flag- , or hexahistidine- or HA-tag fused to the GSG linker were inserted in-frame into the BspEI site . The experimental procedures used to generate in vitro transcripts from cloned HCV sequences and transfection of Huh-7 cells by electroporation have been described in detail recently [6] . For trans-complementation assays a mixture of 7 . 5 µg NS2 mutant and 5 µg helper replicon RNA was used . After electroporation , cells were immediately transferred to complete DMEM and seeded as required for the assay . Virus titers were determined as described elsewhere with slight modifications [13] . In brief , Huh7 . 5 cells were seeded into 96-well plates and fixed 3 - 4 days after infection . For immunohistochemistry we used an antibody specific for the JFH1 NS3 helicase ( 2E3 , generated in cooperation with H . Tang , Florida State University , USA ) at a dilution of 1∶100 or the 9E10 monoclonal antibody specific against NS5A protein in dilution 1∶2 , 000 ( NS5A-9E10; kindly provided by C . M . Rice , New York , USA ) . Bound antibody was detected with a peroxidase-conjugated secondary antibody specific to murine IgG ( Sigma-Aldrich ) diluted 1∶200 in PBS . Virus titers ( 50% tissue culture infective dose per ml; [TCID50/ml] ) were calculated as described recently [62] . Huh7 . 5 cells were electroporated with 10 µg in vitro transcript and culture supernatant harvested 72 h later was concentrated by ultrafiltration using an Amicon Ultra Centrifugal Filter Column ( Milipore ) . Naïve Huh7 . 5 cells were inoculated with this concentrate and continuously passaged up to 6 times . Thereafter , culture supernatants were passaged 4 times on naïve Huh7 . 5 cells and virus titers were determined by TCID50 or immunofluorescence assay . Details of the adaptation method have been described elsewhere [63] . HCV RNA present in Huh7 . 5 cells was amplified and cloned as described previously [10] , [63] . In brief , total RNA was isolated from a confluent 10 cm-diameter dish of Huh7 . 5 cells infected with the adapted virus population by using the Nucleo Spin RNAII Kit ( Macherey-Nagel , Düren , Germany ) as recommended by the manufacturer . One µg total RNA and 50 pmol of primer A9482 ( 5′-GGA ACA GTT AGC TAT GGA GTG TAC C-3′ ) were applied for cDNA synthesis by using the Expand-RT system ( Roche , Mannheim , Germany ) as recommended by the manufacturer . Two to four microliters of the reaction mixture were used to amplify the 5′ half of the HCV genome with the Expand Long Template PCR kit ( Roche ) according to the instructions of the manufacturer with primers S59-EcoRI ( 5′-TGT CTT CAC GCA GAA AGC GCC TAG-3′ ) and A4614 ( 5′-CTG AGC TGG TAT TAT GGA GAC GTC C-3′ ) . PCR products were directly sequenced ( mutants G10S , E45R , P73I and F77A ) or inserted into pFK-I389Luc-EI/NS3-3′/JFH1-dg after restriction with EcoRI and SpeI . Sequence analysis of two independent plasmid clones was performed with an appropriate set of primers . The following antisera were used in this study: rabbit polyclonal antibody specific for NS2 ( NS2-1519; [6] ) ; rabbit polyclonal antibody specific for the core protein ( C-830; [64] ) ; rabbit polyclonal antibody specific for NS3 of JFH1 ( NS3-4949; [30] ) ; mouse monoclonal antibody specific for JFH1 NS3 ( NS3-2E3; generated in co-operation with H . Tang , Florida State University , USA ) ; rabbit polyclonal antibody specific for JFH1 NS5A ( NS5A-52; [6] ) ; mouse monoclonal antibody specific for NS5A ( NS5A-9E10 , kindly provided by C . M . Rice , New York , USA ) ; rabbit polyclonal anti p7-2716 and anti p7-2717 ( kindly provided by M . Harris and S . Griffin , Leeds , UK ) ; rabbit polyclonal antibody specific for J6 E2 [6]; mouse monoclonal antibody specific for E2 protein ( AP33 , kindly provided by Arvin Patel , Glasgow , U . K . ) ; mouse monoclonal anti-Flag antibody , mouse monoclonal anti-HA antibody and mouse monoclonal anti-β-actin , all from Sigma-Aldrich ( Munich , Germany ) . Monoclonal antibody ( mAb ) 1D3 against protein disulfide isomerase ( PDI ) was purchased from StressGen ( Victoria , BC , Canada ) . Western blot analysis was performed as described previously [6] . Samples harvested 48 h after transfection were heated for 20 min at 95°C in sample buffer ( 125 mM Tris/HCl , 2% ( w/v ) SDS , 5% ( v/v ) 2-mercaptoethanol , 10% ( v/v ) glycerol , 0 . 001% ( w/v ) bromophenol blue , pH 6 . 8 ) and separated by SDS polyacrylamide gel electrophoresis . Proteins were electro-transferred to a polyvinylidene fluoride ( PVDF ) membrane ( PerkinElmer Life Sciences ) for 1 h . Membrane was blocked overnight in PBS supplemented with 0 . 5% Tween ( PBS-T ) and 5% dried milk ( PBS-M ) at 4°C prior to 1 h incubation with primary antibody diluted in 2% milk in PBS-T . Membrane was washed 3 times with PBS-T and incubated for 1 h with horseradish-peroxidase conjugated secondary antibody . Bound antibodies were detected after 3 times washing with the ECL Plus Western Blotting Detection System ( GE Healthcare Europe , Freiburg , Germany ) . Huh7 . 5 cells were mock treated or transfected with HCV RNA , and samples were harvested 72 h later by scraping into IP buffer ( 0 . 5% n-dodecyl-β-D-maltoside , 100 mM NaCl , 20 mM Tris pH 7 , 5 ) . After 60 min incubation on ice , cell debris was removed by 15 min centrifugation at 20 , 000xg . Samples were incubated with HA-specific antibody beads ( Sigma Aldrich ) over night at 4°C . After three times washing with IP buffer , samples were eluted into sample buffer and separated by electrophoresis into a 11% Tris-Tricine gel as described elsewhere [65] . Proteins were transferred onto PVDF membrane and HCV proteins were detected by Western blot as described above . U-2 OS human osteosarcoma cells grown on glass coverslips were transfected with GFP fusion constructs , fixed 24 to 48 h post transfection with 2% paraformaldehyde , and mounted in SlowFade ( Molecular Probes , Eugene , OR ) . Immunofluorescence staining was performed as described previously [66] . Bound primary antibody was revealed with Alexa-488-conjugated goat anti-mouse antibody ( Molecular Probes ) . Mounted coverslips were examined using a Leica SP5 AOBS confocal laser scanning microscope . Immunofluorescence detection of HCV proteins in Huh7-Lunet cells was conducted in the analogous way with some modifications . Cells were transfected with HCV RNA and fixed 24 , 36 , 48 and 72 h post-transfection with 4% paraformaldehyde . Bound primary antibodies were detected with a Alexa-488-conjugated goat anti-rabbit antibody or a Alexa-568-conjugated goat anti-mouse antibody ( Molecular Probes ) . LDs were stained with HCS LipidTOX™ Deep Red neutral lipid stain ( Molecular Probes ) . Coverslips were mounted in Fluoromount-G mounting medium ( Electron Microscopy Sciences , Ft . Washington , USA ) and examined with a Perkin Elmer spinning disk confocal ERS 6Line microscope . Images were deconvolved with the Huygens Essential 3 . 5 software using a theoretical point spread function . 3D reconstructed images were created using the Volocity 5 . 3 . software package . The NS2[27]-[59] and NS2[60-99] peptides representing aa segments 27–59 and 60–99 of NS2 of the Con1 strain ( AC number AJ238799 ) were synthetized by Clonestar Biotech and purified by RP-HPLC ( purity >98% ) . CD , NMR spectroscopy , NMR-derived constraints and structure calculation , and molecular modeling and structure representation were performed by standard approaches as described in materials and methods S1 . The atomic coordinates for the NMR structures of peptides NS2[27]-[59] and NS2[60-99] and the NMR restraints in 50% TFE are available in the Research Collaboratory for Structural Bioinformatics ( RCSB ) Protein Data Bank under accession number 2KWT and 2KWZ respectively . The chemical shifts of all NS2[27]-[59] and NS2[60-99] residues have been deposited in the BioMagResBank ( BMRB ) under the accession number 16886 and 16892 , respectively . | Formation of infectious virus particles ( assembly ) is a complex process by which structural proteins and the viral genome must be transferred to the same subcellular sites to allow their direct or indirect interaction . In case of the hepatitis C virus ( HCV ) , this process appears to take place in close proximity of lipid droplets ( LDs ) and requires in addition to the structural proteins core , envelope glycoprotein 1 ( E1 ) and E2 two auxiliary factors , designated p7 and nonstructural protein 2 ( NS2 ) , contributing to virion formation by unknown mechanisms . In this study we used a combination of structural , genetic and biochemical assays to study the role of NS2 in HCV assembly . By using nuclear magnetic resonance spectroscopy of NS2 peptides we established a membrane topology model of the amino-terminal membrane binding domain of NS2 . We found that this protein participates in multiple interactions with E2 , p7 , NS3 and NS5A that appear to recruit the viral proteins to sites in close proximity of LDs . In this respect , NS2 is a key organizer of the assembly of infectious HCV particles . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases",
"virology/virion",
"structure,",
"assembly,",
"and",
"egress",
"cell",
"biology/membranes",
"and",
"sorting",
"infectious",
"diseases/viral",
"infections",
"biochemistry/macromolecular",
"assemblies",
"and",
"machines"
] | 2010 | Structural and Functional Studies of Nonstructural Protein 2 of the Hepatitis C Virus Reveal Its Key Role as Organizer of Virion Assembly |
Glossina palpalis palpalis ( G . p . palpalis ) is one of the principal vectors of sleeping sickness and nagana in Africa with a geographical range stretching from Liberia in West Africa to Angola in Central Africa . It inhabits tropical rain forest but has also adapted to urban settlements . We set out to standardize a long-lasting , practical and cost-effective visually attractive device that would induce the strongest landing response by G . p . palpalis for future use as an insecticide-impregnated tool in area-wide population suppression of this fly across its range . Trials were conducted in wet and dry seasons in the Ivory Coast , Cameroon , the Democratic Republic of Congo and Angola to measure the performance of traps ( biconical , monoconical and pyramidal ) and targets of different sizes and colours , with and without chemical baits , at different population densities and under different environmental conditions . Adhesive film was used as a practical enumerator at these remote locations to compare landing efficiencies of devices . Independent of season and country , both phthalogen blue-black and blue-black-blue 1 m2 targets covered with adhesive film proved to be as good as traps in phthalogen blue or turquoise blue for capturing G . p . palpalis . Trap efficiency varied ( 8–51% ) . There was no difference between the performance of blue-black and blue-black-blue 1 m2 targets . Baiting with chemicals augmented the overall performance of targets relative to traps . Landings on smaller phthalogen blue-black 0 . 25 m2 square targets were not significantly different from either 1 m2 blue-black-blue or blue-black square targets . Three times more flies were captured per unit area on the smaller device . Blue-black 0 . 25 m2 cloth targets show promise as simple cost effective devices for management of G . p . palpalis as they can be used for both control when impregnated with insecticide and for population sampling when covered with adhesive film .
Human and Animal African Trypanosomiasis ( sleeping sickness and nagana ) are still a major constraint on the social and economic development of sub-Saharan Africa , [1] . The diseases affect the health of people and their livestock , resulting in reduced food production and increased poverty [2]–[4] . Tsetse flies ( Diptera: Glossinidae ) transmit the trypanosomes that cause these illnesses for which a vaccine has still to be discovered . The antigenic variation of the pathogen is a major constraint on the development of a vaccine [5] , [6] . Although new treatments based on Nifurtimox and Eflornithine are promising [7] , sleeping sickness is still difficult to treat , particularly in the second phase of the disease [8]–[10] . For the treatment of nagana in livestock , the initial success of trypanocides is increasingly compromised as trypanosomes continue to develop resistance across Africa [11] . G . p . palpalis is one of the principal vectors of sleeping sickness and nagana across large areas of central and West Africa . Its geographical range corresponds to the coastal belt of tropical rain forest stretching from Liberia in West Africa to Angola in Central Africa [12] , [13] . However , it can also adapt to man-modified environments , including large urban settlements [14]–[17] . Studies on microsatellite populations have shown that there is some genetic variability in this subspecies , probably related to geographical distance at a macro-geographical scale [18] and that at a micro-geographical scale the degree of variation is closely related to the extent of habitat fragmentation [19] , as is the case with G . palpalis gambiensis in Burkina Faso [20] . In the face of the continuing difficulties to treat human and animal trypanosomiasis , the reduction and eradication of the tsetse fly vector remains one of the most effective methods to control both diseases . Amongst the different control methods that have been employed , the deployment of visually attractive traps and targets impregnated with insecticide is the most widely used as it is one of the most accessible and efficient methods of control . Historically , the first trapping devices for controlling tsetse were black overalls worn by workers , coated in glue and hung up in the plantations of Sao Tome and Principe in 1910 [21] . Later , in the 1930s , Harris [22]–[24] developed a trap that was employed with great success in Zululand . A further series of trap types followed but was rarely used for controlling tsetse . After the Second World War , trapping was abandoned as a control method in favour of widespread spraying with DDT . It was only in the 1970s that trapping was seriously considered again , thanks to the development of the standard biconical trap by Challier and Laveissière [25] for trapping palpalis and fusca group tsetse . Based on this model , simpler traps , the pyramidal [26]–[28] and monoconical “Vavoua” [29] , were developed in the1980s to increase trapping efficiency and reduce manufacturing costs . Both traps are still regularly used for controlling G . p . palpalis [30] , [31] , with over 60 , 000 insecticide-impregnated pyramidal traps deployed in Angola alone since 2008 . In order to reduce control costs further , simpler two-dimensional targets were developed [32] . Green established that highest catches of G . p . palpalis are obtained on targets made of phthalogen blue cloth with its exceptionally high reflectivity in the blue part of the light spectrum [33] . The same author went on to show that two-colour targets incorporating phthalogen blue with either black or white are better at catching G . p . palpalis than single-colour ones [34] . Recent research has focused on the cost-effectiveness of using smaller targets [35] , [36] , and chemical attractants [37] . Within the Africa-wide WHO-TDR initiative to develop innovative control strategies for tsetse , we set out to standardize long-lasting , visually-attractive devices for G . p . palpalis , and to see if their efficiency and cost-effectiveness could be improved . The trials were based on existing trap/target/bait technology used at each location following a similar experimental approach throughout Africa [38] , [39] . Trials were conducted in wet and dry seasons in the Ivory Coast , Cameroon , the Democratic Republic of Congo and Angola to measure the performance of pyramidal , monoconical and biconical traps and targets in phthalogen blue cloth and various alternatives at different population densities and seasons under different environmental conditions across its continental range . A simple enumeration method ( adhesive film ) was used at these sometimes remote locations to compare trapping efficiencies of devices made of well-characterized colour-fast fabrics . The relative performance of devices was also compared with and without baits . The goal was to determine the most practical and cost effective device/material that would induce the strongest landing response in G . p . palpalis for future use in area-wide population suppression of this fly with insecticide-impregnated devices .
Studies were conducted in four countries: three in central Africa ( Angola , Cameroon and the Democratic Republic of the Congo ) and one in West Africa ( Ivory Coast; Figure 1 ) . Any study made on private land had the owner's consent . A brief description of each site is given below . Five catching devices were tested: standard biconical [25] , monoconical ( Vavoua type ) [29] and pyramidal [27] traps ( Figure 2 ) , and two target designs: a 1 m2 regular square cloth target ( equal vertical rectangles of blue and black , Figure 2 ) and a 0 . 91 m2 Ivory Coast target , 85 cm wide by 107 cm high made of two vertical strips of black cloth ( 17 . 5 cm wide ) on either side of a single blue panel [32] . In Angola , two additional target designs were evaluated in one set of trials: a square 1 m2 target of equal vertical rectangles of black-blue-black cloth and a reduced regular square target of 0 . 25 m2 with vertical rectangles of blue and black cloth . Four different blue fabrics were tested: ( 1 ) C180 phthalogen blue 100% cotton , 180 g/m2 , TDV , Laval , France ( reflectance spectral peak at 460 nm as measured with a Datacolor Check Spectrophotometer , Datacolor AG , Dietlikon , Switzerland ) and referred to here as the standard fabric; ( 2 ) S250 phthalogen blue 65% cotton/35% polyester , 250 g/m2 , TDV France ( peak at 450 nm ) ; ( 3 ) turquoise blue Q10067 65% polyester/35% viscose , 234 g/m2 , Sunflag , Nairobi , Kenya ( peak at 480 nm ) and ( 4 ) Top Notch 6660-563 blue 100% polyester , 410 g/m2 , Rochford Supply , USA ( peak at 470 nm ) . One black fabric ( Q15093 100% polyester , 225 g/m2 , Sunflag , Nairobi ) was used for all devices . To monitor the numbers of tsetse landing on targets , one-sided sticky adhesive film ( Rentokil FE45 , UK ) was attached to both sides of the targets . This film was also attached to the cloth component of traps in some experiments to enumerate flies that land on traps but may not be captured . To assess the influence of adhesive film , particularly its shininess , on landing responses , the number of flies attracted to non-sticky targets was compared to targets covered with adhesive film by using an electric grid of fine electrocuting copper wires ( spaced 8 mm apart ) mounted in front and behind the targets [40] . A potential difference of 40 KV was applied between adjacent wires and tsetse that landed on the E-target were electrocuted and fell into a tray ( 3 cm deep ) of soapy water . E-targets are assumed to be invisible to savannah tsetse [40] , [41] , but this assumption has hardly ever been tested on riverine species . Recently , Tirados et al . ( 2011 ) [36] showed for the first time that many G . p . palpalis are caught with traditional e-targets set up on their own . A 1∶4∶8 mixture of 3-n-propylphenol ( P ) , 1-octen-3-ol ( O ) , and p-cresol ( C ) was used as an attractant for experiments comparing baited devices based on general efficacy for several tsetse . The mixture was prepared at origin by the supplier ( Ubichem Research LTD , Budapest/Hungary ) with a global purity of 98% . Sachets made of 500 gauge/0 . 125 mm polyethylene containing 3 g of the mixture were placed below the catching devices , 10 cm above the ground , alongside a 250 ml bottle buried up to the shoulders containing acetone ( A ) with a 2 mm aperture in the stopper . This combination , termed the POCA bait , was made up according to the method described by Torr et al . [42] .
In the Ivory Coast the target with adhesive film consistently captured significantly more flies than the traps . The better performance of the target was less consistent in the other three countries , where on at least one occasion , the traps performed equally as well as the targets and actually outperformed the target in Cameroon ( Table 1 ) . There was no difference between the performance of the same trapping device made from different blue cloths ( P>0 . 05; Table 1 ) with the exception of the dry season experiments in Angola where the pyramidal trap in standard blue proved significantly better than equivalents in either turquoise or Top Notch blue ( P<0 . 05; Table 1 ) . Sex ratios varied between the field experiments but were not significantly different ( P>0 . 05 ) on the various devices and blue cloths in a given experiment and season . For example , in Angola ( wet season ) the male to female ratio only varied between 0 . 55 and 0 . 63 . The relative rankings of POCA-baited devices were very similar to those in the unbaited trials , but the capture rate on the target covered in adhesive film increased relative to the number of flies caught in the cages of the traps in all countries , most noticeably in the Ivory Coast and Angola ( Table 1 ) . The POCA bait did not affect the relative performance of the biconical compared to the monoconical trap in the Ivory Coast , but in Cameroon the performance of the biconical trap was improved to equal that of the pyramidal traps . As in the unbaited trials , there was no difference between the performance of the same trapping device made from different blue cloths ( P>0 . 05 ) . Sex ratios varied between the field experiments but were not significantly different ( P>0 . 05 ) on the various devices and blue cloths in a given experiment and season . Very similar numbers of flies landed on the traps and targets in Angola and the Ivory Coast and the slight differences recorded are not significant ( P>0 . 05; Figure 3 ) . In contrast , twice as many flies landed on the target compared to the pyramidal trap in the DRC ( P<0 . 01; Figure 3 ) , although in this experiment almost twice as many flies were caught in the cage of the pyramidal trap as on the cloth component of the trap covered with adhesive film ( Figure 3 ) and the pyramidal control caught twice as many flies as the pyramidal trap with adhesive film . In all three countries , a relatively large proportion of flies did not land on the cloth part of the trap but was caught in the cage of the traps with film ( 18% Angola , 33% Ivory Coast , 62% DR Congo ) . The proportion of females caught was slightly higher in the cage of the traps covered in adhesive film , compared to the cages of the controls in DR Congo and the Ivory Coast ( 12% more ) , but this difference was not significant . In Angola twice as many males were attracted to the pyramidal control , but this is based on only two replicates due to weather damage to the third replicate . In the experiment conducted in Angola , the 1 m2 targets in blue-black ( regular ) and black-blue-black ( Ivory Coast style ) equal sized vertical stripes covered with adhesive film caught very similar numbers of flies ( 14 and 11 flies/day , respectively; P>0 . 05 Figure 4 ) . There was a significant preference for landing on the black portion on both targets ( 60% and 71% on the black , respectively; P<0 . 05 ) , although actual fly numbers on the black were very similar on both target types . This experiment also served to confirm an earlier finding at the same location , namely that similar numbers of flies landed on targets as on the cloth panels of the pyramidal traps ( P>0 . 05 , Figure 4 ) . Contrary to this , the pyramidal control ( without adhesive film ) caught few flies on this occasion ( compare Figures 3 and 4 ) . The daily landing rate of flies on the smaller 0 . 25 m2 blue-black square target was 70% of the total recorded on the 1 m2 square target , despite being only a quarter of the size ( 10 and 14 flies per day , respectively; Figure 4 ) and this difference was not significant ( P>0 . 05 ) . When the landing rates are corrected to an equal target size of 1 m2 , the landing rate on the smaller target is nearly triple that on the standard target ( 40 flies/day/m2 and 14 flies/day/m2 , respectively ) . Trap efficiency , defined here as the proportion of flies caught in the cage of the unaltered trap relative to those caught in the cage and on the cloth by the same trap with adhesive film , has been estimated by dividing the mean daily catch of the unaltered pyramidal and monoconical traps by the mean daily catch of the matching traps with adhesive film on the cloth ( flies caught on the adhesive film and in the cage; Figure 3 and Table 2 ) . From these results , trap efficiency is estimated at 51% for the monoconical trap in the Ivory Coast , and at 34% for the pyramidal trap in Angola , although the pyramidal estimate is based on a reduced sample size , due to weather damage during the Angolan trials ( Table 2 ) . It was not possible to estimate trap efficiency for the pyramidal traps in the DRC as fly catches were higher in the control ( Figure 3 and Table 2 ) . Experiments with electric grids to kill flies indicate that the application of adhesive film to a 1 m2 regular square cloth target ( equal vertical rectangles of blue and black ) , reduced by over half the total number of G . p . palpalis that apparently attempted to land on the device . The detransformed catch index compared to the unmodified target is 0 . 45 ( P≤0 . 01; Table 3 ) , affecting both sexes equally . The effect of the adhesive film on fly behaviour nevertheless differed for the blue and black sections of the target . The adhesive film had little effect on numbers landing on the blue section , but in contrast , on the black section , addition of the adhesive film reduced catches by about two-thirds ( P<0 . 001; Table 3 ) . This response was recorded for both sexes .
Taken overall , the combined results from the four countries suggest that the addition of adhesive film to targets in blue and black makes them equal to or more efficient than traps at capturing G . p . palpalis , in most situations but not always . Indeed , earlier studies in the Ivory Coast by Laveissière and Penchenier ( 2000 ) [47] suggested that the monoconical ( Vavoua ) is more efficient for attracting G . p . palpalis than black-blue-black and blue-black targets . However , our results imply that G . p . palpalis attraction to targets is underestimated in the presence of adhesive film by up to 50% which would mean that the targets systematically surpass traps as landing devices . It is the landing response that underlies the principle of using insecticide-impregnated targets as control devices for tsetse . To determine whether traps impregnated with insecticide ( which has been a practice in West and Central Africa to control G . p . palpalis [26] , [47] and is still common practice in Angola ) are more or less efficient than targets at inducing a landing response , a second series of trials was conducted with both the targets and the cloth panels of the traps covered with adhesive film to enumerate the flies which land ( see below under performance of targets versus traps as landing devices below ) . As the baited and unbaited trials were sequential at each location they cannot be compared directly . Baits were used to see whether they increased trap efficiency as has been shown for other tsetse species [48] , but they appear to have had little impact on trap entry for G . p . palpalis , with the exception of an improved entry rate for the biconical trap in Cameroon . In comparison to the unbaited trials , the POCA bait improved catches on the targets relative to the traps in all countries , but most noticeably in Angola , and in the DRC ( by a factor of three and two respectively ) . This confirms observations made by Rayaisse et al . ( 2010 ) [37] who found that odours could increase visual responses to a black target in G . p . palpalis in the Ivory Coast . However , considering the efficacy of smaller targets for G . p . palpalis ( see below ) , one could ask how much effort should one invest in deploying and maintaining chemical baits in control campaigns ( some of which are toxic , e . g . phenols ) when it may be possible to compensate adequately by simply deploying more targets . The blue fabrics chosen for these experiments ( phthalogen blue cotton , polyester or cotton/polyester and turquoise blue polyester/viscose ) were manufactured with differences in fabric texture and with clear differences in blue-green colour , yet with only one exception ( Angola , dry season ) all performed equally well in capturing G . p . palpalis . These results agree with findings for the same fabrics tested in similar devices for several riverine and savannah tsetse species in East and West Africa [38] , [39] . Phthalogen blue cotton cloth has been used for about 30 years in tsetse sampling and control , and is the standard against which all other blues should be compared for attractive properties [49] . The fact that phthalogen blue cotton only remains in limited production has resulted in the ad hoc use of several alternative blue fabrics in tsetse control , some of which are less than optimal for attracting tsetse [50] . The turquoise blue fabric produced in Kenya by Sunflag for these experiments using generic dyes performed well in our studies , confirming that a deep turquoise can be used as a practical alternative to phthalogen blue [51] . Generic dyes are less colour-fast than phthalogen blue cloths , but fading was not a problem in the central African climate after twelve months exposure . However , in humid hot conditions , the cloth must be treated with an anti-mould additive to prevent discolouring due to fungal developments . In contrast , although the 100% polyester blue from Top Notch has excellent colour-fastness it is prohibitively expensive . There is clearly a need to develop a biodegradable and inexpensive replacement for phthalogen blue cotton . The adhesive film used to count flies for this comparison ( as in Rayaisse et al . ( 2012 ) and Mramba et al . ( 2013 ) [38] , [39] ) was found to reduce landings by G p . palpalis by half on the 1 m2 blue-black target , accounted for in the main by reduced landings on the black portion of the target . We assume that landings on panels of monoconical and pyramidal traps are affected to the same extent by the presence of the adhesive film . In any case , the surface area of blue and black parts of pyramidal traps and targets covered with adhesive film were the same in these field trials . The two trap types performed equally as well as the target as a landing device in both Angola and the Ivory Coast . In contrast to this , over twice as many flies landed on the target as on the cloth portion of the pyramidal trap in the DRC . This may be partially explained by the behavioural responses of G . p . palpalis as a relatively high proportion of flies were captured in the cage of the adhesive traps in the DRC ( 62% ) as well as in the similarly treated monoconical and pyramidal traps in the Ivory Coast and Angola ( 33% and 18% , respectively ) . This is in contrast to the results of identical experiments conducted on other tsetse species where very few flies flew directly into the cage ( Glossina swynnertoni: 7% in the cage of a pyramidal trap , [39] , G . tachinoides: 5% and G . morsitans submorsitans 2% in the cage of a monoconical trap [38] ) . The only exception was the closely related G . palpalis gambiensis with 20% of flies counted in the cage of a monoconical trap [38] . This indicates an apparent propensity of these two palpalis group tsetse to enter the cone of pyramidal and monoconical traps without first landing on the cloth panels . If this is the case , then the efficacy of an insecticide-impregnated pyramidal trap as a fly killing device would rely on the ability of the less physically robust trap netting as well as the cloth panels to retain insecticide over time , factors which argue against its use as control a device for G . p . palpalis . The 2012 field trial in Angola shows that alighting by G . p . palpalis was the same on the standard blue-black and Ivory Coast type black-blue-black 1 m2 targets covered with adhesive film , with a noticeable preference for landing on the black portion on both targets ( 60% and 71% , respectively ) . These results would suggest that there is little difference between the two target designs to induce landing by G . p . palpalis . In contrast , landing was equally divided between the blue and black panels on the pyramidal trap . However , the trials using electric grids in the Ivory Coast show that numbers of G . p . palpalis landing on the black portion of the targets would be three times higher on unmodified targets and similar results were recorded using the same experimental approach for the closely related G . p . gambiensis in Burkina Faso [38] . Capture rates using e-nets must be interpreted with a certain amount of caution as recent findings by Tirados et al . [36] have shown that e-nets on their own have a certain attraction for G . p . palpalis . The 2012 Angolan trial also included a 0 . 5×0 . 5 m blue-black target to test if smaller devices could prove effective for G . p . palpalis as has recently been demonstrated for this species in West Africa [35] and a range of riverine and a savannah tsetse spp . [35] , [36] , [39] , [52] , [53] . Landings by G . p . palpalis on the 0 . 25 m2 blue-black target in Angola were not significantly different to those on either the blue-black or blue-black-blue 1 m2 targets covered with adhesive film . In fact , fly catches normalised by unit area were three times higher on the smaller device . This confirms the three-fold higher attraction per unit area recorded for G . p . palpalis to 0 . 25 m2 black cloth targets over 1 m2 targets of the same colour by Tirados et al . in the Ivory Coast [36] ) . The same field study revealed that square and vertical oblong targets are equally attractive to G . p . palpalis and that 0 . 25 m2 is near the optimum target size . Such devices are also less prone to wind damage and theft because of their smaller size . It is a well-established fact that traps used for tsetse capture only a proportion of the flies that are attracted to their vicinity or that may even land on them [38] , [39] . For example , the efficacy of the biconical trap has been estimated at between 8 to 27% for G . p . palpalis [37] . The efficacy of the monoconical and pyramidal traps used in this study was also found to vary widely . In the Ivory Coast , the efficiency of the monoconical trap was up to 51% ( November 2010 experiment ) , whereas in Angola the efficiency of the pyramidal trap was estimated at 34% in the 2010 field trial , but at just 8% in the second trial at the same location in 2012 . From our results , the differences in the performance of a trap type for G . p . palpalis cannot be ascribed to known population structuring in this species across its West and Central African range [18] , [19] , [54] as inconsistencies in the performance of the same pyramidal trap were recorded in successive years at two sites in this study . The much higher catches recorded in Angola and the Ivory Coast on sticky targets indicate that the use of traps alone for monitoring can result in the underestimation of fly population densities . There is a need for reliable and inexpensive devices for population suppression and monitoring of G . p . palpalis across the diverse range of natural and man-made habitats this species occupies from West Africa to Central Africa . Targets that attract flies to land on insecticide-impregnated surfaces are most suitable for population suppression of this vector . We have found no significant difference between the performance of regular blue-black and traditional blue-black-blue 1 m2 targets in experiments performed in West and Central Africa . Furthermore , our results show that landings by G . p . palpalis on 0 . 25 m2 blue-black targets are not significantly different from those on either blue-black or blue-black-blue 1 m2 targets , with three times more flies per unit area on the smaller device . It is thus possible that a number of smaller insecticide-impregnated targets in blue and black could achieve the same result as larger targets in G . p . palpalis control campaigns across its geographical range . However , the most effective size of devices for controlling G . p . palpalis in terms of the costs of fabrication , deployment and maintenance of large targets versus a higher number of smaller targets needs to be determined through field trials . Either phthalogen or turquoise blue cloth would be suitable for these visual control devices . Effective control requires adaptive management [55] whereby tsetse populations are monitored and disease-transmission hot spots are identified for additional intervention . [56] . Pyramidal/monoconical traps could be used for initial monitoring , but our findings indicate that fly numbers caught in the cage of a pyramidal trap should be multiplied three to ten-fold to provide a more realistic estimate of the G . p . palpalis population visiting the device . However , for long-term eradication goals , the detection of very low-density residual pockets is also critical and 0 . 25 m2 targets covered with adhesive film would be a more effective tool , as already been proven in the eradication programme against G . p . gambiensis in the Loos islands ( Guinea ) ( J-B Rayaisse , pers comm . ) . | G . p . palpalis is one of the principal tsetse fly vectors of African Trypanosomiasis . Its range stretches from Liberia in West Africa to Angola in Central Africa . G . p . palpalis inhabits tropical rain forest but has also adapted to urban settlements . Reduction of tsetse populations remains one of the most effective methods to control disease transmission to man and animals , and development of visually-attractive insecticide-impregnated traps and targets for palpalis group tsetse dates from half a century ago . Here we describe field experiments made in wet and dry seasons in the Ivory Coast , Cameroon , Democratic Republic of Congo and Angola to establish the most efficient , long-lasting and practical object that induces the strongest landing response in G . p . palpalis . Independent of season and country , both phthalogen blue-black 1 m2 cloth targets covered with adhesive film proved as good as traps in phthalogen blue or turquoise blue cloth when employed as capturing and landing devices for G . p . palpalis . Pyramidal trap efficiency was inconsistent . As landings on 0 . 25 m2 square phthalogen blue-black targets were not significantly different from landings on the 1 m2 targets , these smaller targets show promise as simple cost effective devices for the management of G . p . palpalis populations . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"african",
"trypanosomiasis",
"parasitic",
"diseases",
"parasitology",
"integrated",
"control",
"pest",
"control",
"neglected",
"tropical",
"diseases",
"tsetse",
"fly",
"animal",
"management",
"zoology",
"veterinary",
"science",
"veterinary",
"medicine",
"infe... | 2014 | Standardising Visual Control Devices for Tsetse Flies: Central and West African Species Glossina palpalis palpalis |
Human Cytomegalovirus ( HCMV ) is a ubiquitous herpesvirus that currently infects a large percentage of the world population . Although usually asymptomatic in healthy individuals , HCMV infection during pregnancy may cause spontaneous abortions , premature delivery , or permanent neurological disabilities in infants infected in utero . During infection , the virus exerts control over a multitude of host signaling pathways . Wnt/β-catenin signaling , an essential pathway involved in cell cycle control , differentiation , embryonic development , placentation and metastasis , is frequently dysregulated by viruses . How HCMV infection affects this critical pathway is not currently known . In this study , we demonstrate that HCMV dysregulates Wnt/β-catenin signaling in dermal fibroblasts and human placental extravillous trophoblasts . Infection inhibits Wnt-induced transcriptional activity of β-catenin and expression of β-catenin target genes in these cells . HCMV infection leads to β-catenin protein accumulation in a discrete juxtanuclear region . Levels of β-catenin in membrane-associated and cytosolic pools , as well as nuclear β-catenin , are reduced after infection; while transcription of the β-catenin gene is unchanged , suggesting enhanced degradation . Given the critical role of Wnt/β-catenin signaling in cellular processes , these findings represent a novel and important mechanism whereby HCMV disrupts normal cellular function .
Human Cytomegalovirus ( HCMV ) is a betaherpesvirus that is ubiquitously present in the human population . Infection with HCMV is usually subclinical in healthy adults , but can cause serious disease in populations with underdeveloped or compromised immune systems [1] . HCMV is the leading viral cause of congenital birth defects in the developed world [2] , [3] . Although the majority of congenitally infected infants are asymptomatic at birth , 5 to 20 percent of the infected neonates are born with symptoms . About 5 percent of symptomatic children die during the neonatal period and the majority of those who survive develop permanent neurodevelopmental complications , such as hearing loss and mental retardation [4] , [5] . A small percentage of infants who lack symptoms may also suffer long-term neurological abnormalities later in life [6] . Opportunistic infections with HCMV also increase morbidity in immunocompromised individuals , such as organ transplant recipients and AIDS patients [1] . In addition , infection with HCMV compromises the success of bone marrow and solid organ transplantations , where it is associated with graft rejection [7] . Recent evidence also implicates HCMV infection as a contributing factor in the development of atherosclerosis and cardiovascular disease [8] . Upon entry into the cell , HCMV manipulates the host environment in order to establish productive infection and ensure progression of the viral replication cycle . HCMV controls the host cell through dysregulating a multitude of signaling , cytoskeletal and regulatory pathways that control key cellular events . HCMV is known to highjack the host replication machinery [9] , alter the cell cycle [10]–[12] , and manipulate the host innate and adaptive immune responses ( reviewed in [13] ) . One important signaling pathway that has been reported to be perturbed by herpesviruses is the canonical Wnt/β-catenin pathway ( reviewed in [14] ) . This signaling pathway is crucial for driving a large number of the molecular events that take place during embryonic development , such as cell fate determination and establishment of tissue polarity [15] as well as playing a critical role in differentiation of multiple cell types including neurons [16] , [17] and mesenchymal stem cells [18] . Recent evidence also demonstrates that activation of this pathway is indispensible for differentiation of fetal cytotrophoblasts into an invasive phenotype during placentation [19] . Apart from its role in development , canonical Wnt signaling is involved in regulating many homeostatic events in adult tissues . Among the processes regulated by Wnt signaling are cell proliferation , motility , survival , and stem cell maintenance [20]–[22] . Aberrant activation of this pathway is associated with the onset of several types of human malignancies [23] , [24] . β-catenin , the major effector protein in the canonical Wnt signaling pathway , is normally retained in the cytoplasm in an inactive state through its interaction with a large protein complex comprised of axin , adenomatous polyposis coli ( APC ) and two serine/threonine kinases: glycogen synthase kinase-3β ( GSK-3β ) and casein kinase 1 ( CK1 ) . This complex maintains low levels of β-catenin in the cell through constant proteasome-mediated degradation . Phosphorylation of serine-45 by CK1 and subsequent phosphorylation on serine-33 and -37 by GSK-3β marks β-catenin for polyubiquitination and proteolytic degradation . Pathway activation is initiated when Wnt ligands , a large family of secreted glycoproteins , bind to heterodimeric Frizzled ( FZD ) /low-density lipoprotein receptor-related protein-5/6 ( LRP-5/6 ) receptors on the surface of target cells . This initiates a cascade of events leading to phosphorylation and activation of Disheveled ( Dsh/Dvl ) , a cytoplasmic scaffolding protein that relays Wnt signaling downstream and disrupts the axin/APC/GSK-3β complex . Disruption of the complex causes accumulation of stable , hypophosphorylated β-catenin in the cytoplasm , followed by its translocation to the nucleus ( reviewed in [25] ) . Once in the nucleus , β-catenin binds T cell-specific factor ( TCF ) /lymphoid enhancer-binding factor-1 ( LEF-1 ) DNA-binding factors to activate transcription of over fifty target genes involved in cell maintenance , proliferation and survival , such as cyclin D1 , c-myc , metalloproteinase-2 ( MMP-2 ) and metalloproteinase-9 ( MMP-9 ) [26]–[28] . To date , the effect of HCMV infection on this essential signaling pathway has not been reported . In this study , we demonstrate that HCMV infection induces juxtanuclear accumulation and degradation of β-catenin resulting in a decrease in its transcriptional activity in response to Wnt ligand stimulation . Diminished activation of this important pathway introduces a novel mechanism whereby HCMV causes impaired cellular function .
The TCF/LEF-1-luciferase reporter construct TOPflash was used to determine whether HCMV infection affects β-catenin activity . Human foreskin fibroblasts ( HFFs ) were transiently transfected with TOPflash or FOPflash containing mutated TCF/LEF-1 binding sites as a control . After 12 hr , transfected cells were infected with HCMV or mock-infected . At 48 hr after infection , the cells were stimulated with the canonical ligand Wnt-3A or lithium chloride ( LiCl ) for an additional 12 hr and then analyzed for luciferase expression . LiCl is a potent activator of Wnt signaling that acts downstream of Wnt receptors by inhibiting GSK-3β thus stabilizing β-catenin and promoting nuclear translocation [29] . Stimulation of mock-infected cells with both Wnt-3A and LiCl markedly increased luciferase expression compared to the PBS-treated control ( Figure 1A ) . In contrast , TOPflash activity in HCMV-infected cells stimulated with Wnt-3A or LiCl was only slightly increased from basal levels ( Figure 1A ) , suggesting inhibited activation of the TCF/LEF-1 transcription complex . Inhibition of both Wnt-3A- and LiCl- stimulated TOPflash activity suggests that the inhibitory activity is at the level of β-catenin and not at the level of Wnt receptors . LiCl treatment of FOPflash transfected cells failed to induce luciferase expression in either mock- or HCMV-infected cells ( Figure 1B ) , demonstrating the specificity of TOPflash activity . Next , the expression of endogenous Wnt/β-catenin target genes was examined in HCMV-infected cells . Expression of cyclin D1 , c-myc and Dikkopf-1 ( DKK1 ) genes , which are transcriptionally regulated by the β-catenin/TCF/LEF-1 complex [26] , [27] , [30] , was evaluated in HCMV-infected HFFs by quantitative real-time RT-PCR ( qRT-PCR ) analyses . HCMV infection led to modest but significant downregulation of expression of all three genes , compared to mock-infected cells ( Figure 1C ) . Together , these results demonstrate that HCMV inhibits the transcriptional activity of the β-catenin/TCF/LEF-1 complex . The effect of HCMV infection on the subcellular distribution of β-catenin in HCMV-infected HFFs was investigated by immunofluorescence . HFFs that had been infected with HCMV or mock-infected for 48 hr were immunostained for β-catenin ( Figure 2A ) . In mock-infected cells , β-catenin displayed typical diffuse cytoplasmic/membranous staining . In contrast , intense β-catenin staining was observed in a distinct juxtanuclear region in HCMV-infected cells that was absent in mock-infected cells . The morphology and localization of the β-catenin aggregates observed in HCMV infected cells resembles that of aggresomes which typically form at the microtubule-organizing center ( MTOC ) in response to accumulation of misfolded proteins or when the capacity of the proteasome degradation machinery is inhibited or overwhelmed [31] . To determine if aggregation of β-catenin was a result of proteasomal impairment , protein lysates were collected from HFFs at 24 , 48 and 72 hr after infection with HCMV and analyzed for catalytic activity of the 26S proteasome using a fluorophore-labeled proteasome specific substrate Suc-LLVY-AMC . Consistent with a previous report [32] , there was a significant increase in the catalytic activity of the proteasome in HCMV-infected cells ( Figure 2B ) . This suggests that HCMV-induced aggregation of β-catenin is not a result of impaired proteasome function that prevents the normal turnover of β-catenin . To determine if HCMV infection alters β-catenin protein levels , HFFs were infected with HCMV and analyzed for total β-catenin by Western blotting at various times after infection . There was a significant ( up to 90% ) reduction in the total levels of β-catenin protein in HCMV-infected HFFs by 72 hr after infection ( Figure 3A ) . Apart from being a key component of canonical Wnt signaling , β-catenin is also an integral part of adherens junctions where it mediates contact between cadherins and the actin cytoskeleton . Cytoplasmic , nuclear and adherens junctions-associatedβ-catenin constitute distinct cellular pools of β-catenin that are tightly regulated . To determine if HCMV infection differentially affects the levels of β-catenin between subcellular compartments , membrane , cytoplasmic and nuclear fractions were isolated from mock and HCMV-infected HFFs and analyzed for β-catenin by Western blotting . Histone H4 , glyceraldehyde 3-phosphate dehydrogenase ( GAPDH ) and caveolin-1 were used as markers of the purity of the nuclear [33] , cytoplasmic [34] and membrane cell fractions [35] , respectively . HCMV infection reduced the amount of cytoplasmic , membrane and nuclear β-catenin compared to mock-infected controls . Although there was some cross contamination between fractions , the results suggest that levels of β-catenin in all three fractions were reduced in HCMV infected cells ( Figure 3B ) . β-catenin levels are normally regulated by phosphorylation-dependent degradation via the ubiquitin-proteasome system [36] . However , regulation of β-catenin at the level of transcription has been reported [37] . To determine if HCMV infection affects β-catenin gene transcription , qRT-PCR was performed . As indicated in Figure 3C , β-catenin mRNA levels were slightly higher in HCMV-infected HFFs compared to mock-infected controls , indicating that the decrease in β-catenin protein induced by HCMV is posttranscriptional . To determine whether β-catenin degradation requires de novo viral gene expression , HFFs were infected with HCMV or ultraviolet ( UV ) -inactivated HCMV . Immunofluorescence staining 24 hr after infection detected juxtanuclear accumulation of β-catenin in cells infected with HCMV virus , whereas in cells infected with UV-inactivated virions β-catenin failed to aggregate ( Figure 4A ) . Entry of UV-inactivated virions was confirmed by tegument-associated pp65 protein staining in infected cells 6 hr after infection ( Figure 4B ) . Successful inactivation of the virus was confirmed by the absence of HCMV immediate early ( IE ) 1/2 protein expression in infected cells ( Figure 4C ) . In addition , β-catenin levels were analyzed by Western blotting in cells infected with HCMV or UV-inactivated virus , at 24 and 48 hr after infection . HCMV , but not UV-inactivated virus reduced protein levels of β-catenin at each time point ( Figure 4C ) . This suggests that degradation of β-catenin is not due to HCMV binding and entry into the cell or to tegument-associated proteins , but rather requires viral gene expression . Wnt signaling plays a major role in placental development and morphogenesis . We have shown previously that HCMV inhibits differentiation of SGHPL-4 cells , an extravillous trophoblast ( EVT ) cell line derived from first trimester placenta , into a migratory and invasive phenotype [38] . To determine whether HCMV infection inhibits Wnt signaling in SGHPL-4 cells as was observed in HFFs , SGHPL-4 cells that had been infected with HCMV for 48 hr , were stimulated with Wnt-3A ligand or LiCl for 6 hr and then immunostained for β-catenin ( Figure 5 ) . Both Wnt-3A and LiCl treatments led to translocation of β-catenin to the nucleus of mock-infected cells . In contrast , β-catenin was seen to aggregate in distinct juxtanuclear region in HCMV-infected cells . Moreover , Wnt-3A- or LiCl-treatment of infected cells showed no change in β-catenin localization . Mock-infected cells showed no aggregation of β-catenin at any time . Similar to our results in HFFs , HCMV infection of SGHPL-4 cells altered the subcellular distribution of β-catenin and caused it to accumulate in a juxtanuclear location . To investigate whether HCMV infection of SGHPL-4 cells also affected the transcriptional activity of β-catenin , cells were transfected with TOPflash or FOPflash and 6 hr later infected with HCMV or mock-infected . Forty-eight hr later , cells were stimulated with Wnt-3A ligand for 12 hr and then analyzed for luciferase expression . Wnt-3A stimulation of uninfected SGHPL-4 cells resulted in ∼50 fold increase in TOPflash activity compared to PBS-treated ( Figure 6A ) , whereas FOPflash activity remained unchanged ( Figure 6B ) . Similar to our results in HFFs , HCMV-infected SGHPL-4 cells showed no response to Wnt-3A ( Figure 6A ) . Consistent with these results , HCMV infection led to modest , but significant downregulation of β-catenin regulated gene expression , including cyclin D1 , MMP-2 and MMP-9 , compared to mock-infected cells ( Figure 6C ) . To evaluate whether HCMV infection abrogates migration of EVTs upon activation of Wnt/β-catenin signaling , a transwell migration assay was performed ( Figure 7 ) . SGHPL-4 cells that had been infected with HCMV and mock-infected controls were allowed to migrate in the presence of 150 ng/ml of Wnt-3A ligand or PBS control for 12 hr . Wnt-3A treatment significantly increased migration of mock-infected controls compared to unstimulated cells . In contrast , HCMV-infected cells were unresponsive to Wnt-3A stimulation and exhibited lower basal levels of migration through the transwell inserts . Together , this suggests that Wnt/β-catenin signaling is also dysregulated in HCMV-infected cytotrophoblasts .
In this study , we report for the first time that HCMV inhibits the canonical Wnt signaling pathway in dermal fibroblasts and human extravillous cytotrophoblasts . We demonstrate that the key player in the pathway , β-catenin , becomes sequestered and degraded in infected cells . Moreover , the transcriptional activity of β-catenin is inhibited in infected cells . Among β-catenin regulated genes are: cyclin D1 and c-myc that are involved in cell cycle regulation and cell proliferation [26] , [27]; DKK-1 that is required for normal embryonic development through its negative feedback regulation of Wnt signaling [30]; as well as MMP-2 , MMP-9 , and MMP-7 that play major roles in cell proliferation , migration , differentiation , angiogenesis , apoptosis and host defense [28] , [39] . For β-catenin to drive transcription of target genes it must first translocate to the nucleus [40] . We found that β-catenin protein is prevented from translocating to the nucleus in HCMV-infected cells stimulated with Wnt-3A or LiCl . Instead , β-catenin remains sequestered in a discrete juxtanuclear location . This phenomenon is not cell type specific as we observed similar aggregation of β-catenin in HCMV-infected HFFs ( Figure 2A ) and SGHPL-4 cells ( Figure 5C ) ; nor is it a generalized response to herpesvirus infection , as HFFs infected with herpes simplex virus-1 ( HSV-1 ) failed to cause aggregation of β-catenin ( data not shown ) . The β-catenin aggregates resemble aggresomes that are juxtanuclear depositions of misfolded or damaged proteins that assemble at the MTOC when the cellular proteasome machinery is inhibited or has exceeded its capacity [31] . Accumulation of β-catenin in aggresomes has been demonstrated in hippocampal pyramidal neurons from Alzheimer's patients [41] and in multiple myeloma cells [42] . Ghanevati and Miller suggested that accumulation of β-catenin in aggresomes likely results from proteasomal impairment [41] . However , we found the catalytic activity of the proteasome in HCMV-infected cells to be functional and even increased with progression of infection ( Figure 2B ) . This result is consistent with a previous report [32] and may be explained by the requirement of host proteasome activity for HCMV DNA replication and assembly [43] , [44] . Moreover , pharmacological inhibition of the proteasome with MG132 did not lead to perinuclear aggregation of β-catenin ( data not shown ) suggesting that aggregation of β-catenin in HCMV-infected cells is not a general consequence of proteasomal impairment that hampers normal β-catenin turnover . Further characterization of the β-catenin aggregates by co-immunostaining for specific markers ( e . g . histone deacetylase 6 , γ-tubulin , vimentin ) is required to determine if they are indeed aggresomes . The mechanism ( s ) by which aggregated β-catenin is eliminated from the infected cells is not clear . Intracellular levels of β-catenin in the cell are maintained through GSK-3β mediated phosphorylation of β-catenin , which targets it for ubiquitination and degradation through the proteasome [36] . However , aggregated proteins are generally poor substrates for the proteasome and are usually eliminated through autophagy [45] . In multiple myeloma cells , β-catenin was found to accumulate in juxtanuclear aggresomes that are cleared by a mechanism involving both autophagy and the ubiquitin proteasome system [42] . We observed a significant decrease in β-catenin protein levels in HFFs infected with HCMV ( Figure 3A ) however β-catenin protein levels were only slightly decreased in HCMV-infected SGHPL-4 cells ( data not shown ) . These results may indicate that different cell types clear aggregated proteins by different mechanisms or kinetics . The HCMV viral particle consists of over forty structural and tegument proteins that are delivered to the host cell upon infection and are known to perturb multiple host signaling pathways [46] , [47] . HCMV tegument proteins have been shown to induce proteasomal degradation of a number of key cellular proteins to facilitate viral gene expression and modulate cell cycle events [48]–[50] . However , the inability of UV-inactivated HCMV to promote sequestration and degradation of β-catenin ( Figure 4 ) suggest that viral gene expression is required rather than a viral tegument protein or binding of the virion alone . More work is needed to determine which viral gene ( s ) are responsible for causing this inhibitory effect on Wnt/β-catenin signaling . The initial events taking place after HCMV infection are presumed to be essential for the establishment of virus replication and viral pathogenesis . HCMV is known to manipulate cell proliferation [10] , microtubule networks [51] and proteasome activity [32] to enhance viral replication and assembly . Wnt/β-catenin signaling either regulates or is regulated by these same cellular processes . Tumorigenic gammaherpesviruses , such as Kaposi's sarcoma associated herpesvirus ( KSHV ) and Epstein-Barr virus ( EBV ) , are well known to activate the canonical Wnt/β-catenin pathway thus promoting cell division and proliferation [52] , [53] . HCMV , to our knowledge , is the first human herpesvirus that has been reported to inhibit the Wnt/β-catenin pathway . How dysregulation of Wnt/β-catenin signaling may benefit the virus is unknown , however understanding the effect of the virus on this essential cellular pathway sheds light on HCMV pathogenesis . Wnt signaling is inherently complex . Both the ligands and receptors involved in Wnt signal transduction belong to large multi-gene families , allowing for a large number of possible ligand-receptor interactions that can elicit a variety of intracellular responses ( reviewed in [54] ) . Thus HCMV-mediated dysregulation of canonical Wnt signaling could have different effects depending on the type of cell infected and the cellular context . Dysregulation of Wnt/β-catenin signaling pathway by HCMV is likely to be of physiological significance particularly during congenital infection . It has been established that intrauterine transmission of HCMV during pregnancy is associated with abnormal placental development caused by impaired differentiation of EVTs . During gestation , these highly specialized cells differentiate into highly proliferative invasive cells that migrate through the decidua and remodel maternal spiral arteries to establish a vascular connection between the mother and the fetus [55]–[57] . HCMV infection of placental EVTs has been proposed to inhibit their ability to differentiate and adequately invade the decidua and could impair their ability to remodel the uterine spiral arteries during pregnancy resulting in shallow placentation [38] , [58]–[60] . However , the mechanisms are not well understood . Prior to differentiating into highly invasive cells , cytotrophoblasts undergo rapid proliferation in order to establish cell-anchoring columns that attach the placenta to the uterine wall . We have previously shown that HCMV markedly inhibits EVT proliferation [38] . HCMV-infected cells arrest in a pseudo G1 phase and fail to enter S phase [11] . High expression of cyclin D1 is required for initiation of S phase , after which the level of the protein drops . In this study , cyclin D1 expression that is dependent on β-catenin activation was modestly , but significantly downregulated in HCMV-infected SGHPL-4 cells and HFFs ( Figures 6C and 1C , respectively ) . In agreement with these results , downregulation of cyclin D1 protein levels has been previously reported in HCMV-infected HFFs [10] . Thus downregulation of cyclin D1 expression as a result of HCMV inhibition of β-catenin transcriptional activity likely contributes to the block in proliferation of infected EVTs . Invasion of EVTs during placental development involves production of extracellular matrix-degrading metalloproteinases and vasculogenic factors [61] , [62] . At the molecular level , these processes are mediated by the canonical Wnt signaling pathway [39] , [63] . Studies show that first trimester and term human placental tissues as well as several cytotrophoblast cell lines express high levels of Wnt signaling pathway components including ligands , FZD receptors and transcription factors from the TCF and LRP families [63] , [64] . Furthermore , canonical Wnt ligands , including Wnt-3A , have been previously reported to promote motility and invasiveness of the EVT cell line SGHPL-5 and primary EVTs purified from first-trimester placentas , through activation of the Wnt/β-catenin signaling pathway and upregulation of MMP-2 and MMP-9 [19] , [62] . In this study we demonstrate that infected SGHPL-4 cells , that are very similar to SGHPL-5 cells [65] , fail to migrate in response to Wnt-3A stimulation ( Figure 7 ) . In addition , expression of MMP-2 and MMP-9 mRNAs was significantly decreased in HCMV-infected SGHPL-4 cells ( Figure 6C ) . Consistent with these results , we previously demonstrated that the secretion and activity of MMP-2 and MMP-9 was dramatically reduced in HCMV-infected SGHPL-4 cells [38] . Therefore , inhibition of Wnt signaling by the virus may be responsible for the observed decreased MMP production and impaired invasiveness in infected EVTs . Our results do not exclude the possibility that additional components of the Wnt/β-catenin pathway or other signaling pathways that influence cytotrophoblast migration and invasion may also be affected by HCMV infection . A recent study by Rauwel et al . demonstrates that HCMV inhibits migration and invasion of EVTs through activation of peroxisome proliferator-activated receptor gamma ( PPAR-γ ) , a nuclear receptor that regulates lipogenesis and inflammation [66] . Interestingly , PPAR-γ has been shown to negatively regulate β-catenin [67] . In addition , several growth factors and cell surface proteins have been shown to stimulate cytotrophoblast migration and invasion through activating signaling molecules and pathways such as focal adhesion kinase ( FAK ) , mitogen-activated protein kinase/extracellular signal-regulated kinase ( MAPK/ERK ) , and phosphatidylinositol-3-kinase ( PI3K ) /AKT ( as reviewed in [68] ) . However , the effect of HCMV infection on the status of these pathways in cytotrophoblasts has not been established . β-catenin is not only an essential component of canonical Wnt signaling , it is also an integral constituent of adherens junctions where it mediates contact between classical cadherins , α-catenin and the actin cytoskeleton . Our data indicate that membrane-associated β-catenin is also decreased in HCMV infected cells ( Figure 3B ) . Since the E-cadherin/β-catenin complex is very important in maintaining epithelial morphology and integrity [69] , its disruption could contribute to the profound changes in cellular morphology observed in HCMV infected cells . In conclusion , we report for the first time that HCMV infection dysregulates the canonical Wnt/β-catenin signaling pathway in human dermal fibroblasts and placental EVTs . Our study establishes for the first time that HCMV inhibits canonical Wnt signaling by causing sequestration and degradation of endogenous β-catenin , thus preventing its downstream activities . Since the Wnt/β-catenin pathway is an evolutionarily conserved pathway involved in a diverse range of biological functions such as cell cycle control , cell differentiation , embryonic development , placentation and metastasis , the finding that HCMV impairs this pathway becomes globally important for understanding viral pathogenesis , particularly that related to HCMV disease .
Human foreskin fibroblasts ( HFFs ) were purchased from American Type Culture Collection ( ATCC , Manassas , VA ) and were cultured in Dulbecco's Modified Eagle Medium ( DMEM , Sigma-Aldrich , St . Louis , MO ) supplemented with 10% fetal bovine serum ( FBS ) and 1% penicillin-streptomycin-L-glutamine at 37°C in 5% CO2 . The human extravillous-like cytotrophoblast cell line SGHPL-4 cells was derived from first trimester chorionic villous tissue and exhibits features of invasive cytotrophoblasts , such as expression of HLA-G , CD9 and cytokeratin-7 [65] , [70] . SGHPL-4 cells were maintained in Ham's F10 Nutrient Mix ( Invitrogen , Carlsbad , CA ) supplemented with 10% FBS ( Atlas Biologicals , Fort Collins , CO ) , 1% penicillin-streptomycin-L-glutamine ( Invitrogen ) at 37°C in 5% CO2 . All experiments were carried out using either a laboratory strain of HCMV that expresses GFP from the UL127 promoter ( Towne-GFP ) , or the HCMV BAC-derived clinical strain TR , both kindly provided by Dr . Dan Streblow ( Oregon Health & Science University , OR ) . All viral strains were propagated in HFFs . For HCMV infections , cells were synchronized by serum starving overnight and infected with HCMV at multiplicity of infection ( MOI ) of 1 to 2 . Briefly , viral inoculum was added to the cells and allowed to adsorb for 90 min at 37°C in 5% CO2 . The virus inoculum was then removed and replaced with fresh medium containing 0 . 5% FBS . Primary antibodies: mouse anti-β-catenin ( Santa Cruz Biotechnology , Santa Cruz , CA; 1∶500 dilution ) , mouse anti-β-actin ( Abcam , Cambridge , MA; 1∶5 , 000 ) , mouse anti-CMV IE1/2 ( Millipore , Billerica , MA; 1∶200 ) , mouse anti-CMV p65 ( Santa Cruz Biotechnology; 1∶500 ) , mouse anti-γ-tubulin ( Abcam; 1∶1000 ) , rabbit anti Histone 4 ( H4 ) ( Millipore , Billerica , MA; 1∶1000 ) , rabbit anti-GAPDH ( Sigma; 1∶5000 ) , rabbit anti-caveolin-1 ( Abcam; 1∶500 ) . Secondary antibodies: AlexaFluor antibodies were purchased from Invitrogen and used at 1∶1000 dilution: AlexaFluor 488 goat anti-mouse , AlexaFluor 555 goat anti-mouse . Human recombinant Wnt-3A was purchased from R&D ( Minneapolis , MN ) . Lithium chloride ( LiCl ) was purchased from Sigma-Aldrich . For all immunofluorescence analyses , cells were seeded onto 1 . 5 mm glass coverslips coated with 0 . 2% gelatin , and infected with HCMV or mock-infected as described above . At indicated time points after infection , cells were washed with Dulbecco's phosphate-buffered saline ( DPBS , Invitrogen ) and fixed in 2% paraformaldehyde ( Ted Pella , Redding , CA ) for 20 min . Cells were washed with DPBS and incubated with 50 mM NH4Cl solution in DPBS for 20 min followed by permeablization in 0 . 1% Triton X-100 for 8 min . Prior to blocking , the cells were incubated with an Fc receptor blocker ( Innovex Biosciences , Richmond , CA ) for 30 min at room temperature . The cells were then blocked in 5% bovine serum albumin ( BSA ) for 1 hr at room temperature . After blocking , cells were incubated with primary antibodies diluted in blocking solution at the manufacturer's recommended dilution for 1 hr at room temperature . The cells were washed and incubated for 1 hr with AlexaFluor-conjugated secondary antibodies and 4′-6-Diamidino-2-phenylindole ( DAPI ) as a nuclear counterstain . Coverslips were washed and mounted with ProLong gold antifade reagent ( Invitrogen ) and imaged at 40× using a Zeiss Axioplan II microscope ( Carl Zeiss , Thornwood , NY , USA ) . Processing of the acquired images was performed using Adobe Photoshop software . To evaluate activation of the Wnt/β-catenin pathway , SGHPL-4 cells were transiently transfected with the TCF/LEF-1 reporter plasmid TOPflash , which contains multimeric TCF/LEF-1 sequences upstream of a firefly luciferase reporter gene . FOPflash plasmid containing mutated TCF/LEF-1 binding sites was used as a specificity control for TOPflash activity . Both TOPflash and FOPflash plasmids were obtained from Millipore ( Bedford , MA ) . A Renilla luciferase-expressing plasmid pRL-TK ( Promega , Madison , WI ) was co-transfected with both TOPflash and FOPflash to serve as an internal transfection control . For each transfection , SGHPL-4 cells were transfected with 1 µg of either TOPflash or FOPflash DNA and 0 . 1 µg of Renilla plasmid using the Neon transfection system ( Invitrogen ) according to the manufacturer's instructions ( 1 pulse , a pulse width of 20 ms and voltage of 1400 V ) . Each transfection was performed in triplicate . Six hr after transfection , the cells were infected with HCMV ( MOI of 1 to 2 ) or mock-infected . Forty-eight hr after infection , the cells were stimulated with 150 ng/ml Wnt-3A or PBS vehicle control for an additional 12 hr and cells were harvested . Luciferase activity was assayed with a dual-luciferase reporter assay kit ( Promega , Madison , WI ) and measured by a Lumat LB 9507 tube luminometer ( Berthold Technologies , Bad-Wildbad , Germany ) . For each sample , the firefly luminescence signal was normalized to the corresponding Renilla signal . SGHPL-4 cells were infected with HCMV or mock-infected at an MOI of 1 to 2 . Cells were trypsinized 48 hr after infection , collected in serum-free F-10 media and added to the upper side of an 8 µm FluoroBlok 24-well multiwell insert system ( BD Discovery Labware , Bedford , MA ) at a density of 5×104 cells per insert . Wnt-3A ligand ( 150 ng/ml ) or PBS control was added to both the upper and the lower chambers of the multiwell plate . Each condition was performed in triplicate . The cells were allowed to migrate through the FluoroBlok membrane for 12 hr at 37°C in 5% CO2 , after which they were fluorescently labeled with calcein AM ( Molecular Probes , Eugene , OR ) and visualized by fluorescent microscopy . The 12 hr timepoint was empirically determined to be optimal for SGHPL-4 cell migration . Three random fields from each insert were captured with a Nikon TE200 inverted fluorescent microscope ( Nikon Instruments , Melville , NY ) and the average fluorescence intensity was determined by ImageJ analysis software . HCMV was UV-inactivated by exposure to 426 mJ of 254 nm UV light in Bio-Rad GS gene linker UV Chamber ( Bio-Rad , Hercules , CA ) . UV inactivation was assured by the absence of HCMV IE1/2 expression in HFFs . Total RNA was collected from mock and HCMV-infected cells using the Qiagen RNeasy Kit ( Qiagen , Valencia , CA ) according to the manufacturer's instructions . RNA ( 500 ng ) was reverse transcribed using Bio-Rad iScript cDNA synthesis kit ( Bio-Rad ) and PCR reactions were performed using SYBR Green supermix ( Bio-Rad ) or TaqMan Universal PCR ( Applied Biosystems , Carlsbad , CA ) master mix and the iCycler Real-Time PCR detection system ( Bio-Rad ) . Oligonucleotide primers ( Integrated DNA Technologies , Coraville , IA ) used are as follows: cyclin D1 forward ( 5′-CGCCCTCGGTGTCCTACTTC-3′ ) , cyclin D1 reverse ( 5′-GACCTCCTCCTCGCACTTCTG-3′ ) ; MMP-2 forward ( 5′-ATGTCGCCCCCAAAACGGACAAAG-3′ ) , MMP-2 reverse ( 5′-CGCATGGTCTCGATGGTATTCTGG-3′ ) ; MMP-9 forward ( 5′-AGACGGGTATCCCTTCGACG-3′ ) , MMP-9 reverse ( 5′-AAACCGAGTTGGAACCACGAC-3′ ) ; β-catenin forward ( 5′-T ACAAACTGTTTTGAAAATCCA-3′ ) , β-catenin reverse ( 5′-CGAGTCATTGCATACTGTCC-3′ ) ; DKK1 forward ( 5′-TTCCAACGCTATCAACCTGC-3′ ) , DKK1 reverse ( 5′- CAAGGTGGTTCTTCTGGAATACC-3′ ) ; c-myc forward ( 5′- GCCACGTCTCCACACATCAG-3 ) , c-myc reverse ( 5′-TCTTGGCAGGAGGATAGTCCTT-3′ ) ; 36B4 forward ( 5′-TGGAGACGGATTACACCTTC-3′ ) , 36B4 reverse ( 5′-CTTCCTTGGCTTCAACCTTAG-3′ ) ; Human GAPDH mRNA was measured using a TaqMan Gene Expression Assay according to the manufacturer's protocol ( Applied Biosystems; Carlsbad , CA ) . Prior to performing real-time PCR on experimental samples , primer concentrations were optimized to provide equal priming efficiency ( ∼100% ) for each primer pair . Negative controls , including cDNA reactions without reverse transcriptase or RNA and PCR mixtures lacking cDNA were included in each PCR . Following amplification , specificity of the reaction was confirmed by melt curve analysis . Relative quantitation was determined using the comparative CT method with data normalized to GAPDH or 36B4 and calibrated to the average ΔCT of mock-infected control at the specified time point . For Western blot analyses , cells were lysed in SDS lysis buffer ( 62 . 5 mM Tris-HCl , 2% SDS , 10% glycerol , 50 mM DTT ) supplemented with protease inhibitor cocktail ( Roche Chemicals , Indianapolis , IN ) . The lysates were sonicated briefly and protein content was determined by Bradford assay ( Bio-Rad ) . Equivalent amounts of protein were separated by SDS-PAGE and transferred to nitrocellulose . The blots were blocked with 5% BSA and incubated overnight with primary antibodies diluted in blocking solution at the recommended dilutions at 4°C . Following three washes with Tris-buffered saline ( TBS ) , blots were incubated with horseradish peroxidase-conjugated anti-mouse or anti-rabbit antibodies ( Invitrogen ) diluted in 5% BSA at 1∶10 , 000 . After washing in TBS , proteins were detected using SuperSignal chemiluminescent substrate ( Pierce Biotechnology , Rockford , IL ) according to the manufacturer's instructions . For densitometric analysis of Western blot images , density profiles of the bands were measured using ImageJ software . Membrane/cytoskeletal , cytoplasmic and nuclear fractions were isolated from HFFs that had been mock and HCMV-infected for 48 hr , using the Qproteome cell compartment kit ( Qiagen ) according to the manufacturer's protocol . Resulting fractions were separated on SDS-PAGE and immunoblotted for β-catenin . Histone H4 , GAPDH and caveolin-1 were used to assess the purity of the cell fractions as previously described [33]–[35] . HFF cells ( 1 . 5×104 ) were plated in a 96-well tissue culture plate in 100 ìl DMEM supplemented with 10% FBS . The cells were infected with HCMV-TR ( MOI of 1 to 2 ) or mock-infected . Cells were analyzed at 24 , 48 and 72 hr after infection using a Proteasome Assay kit ( Cayman Chemical , Ann Arbor , MI ) that measures the catalytic activity of the 26S proteasome , according to the manufacturer's instructions . The fluorescence intensity was measured at an excitation of 360 nm and emission of 480 nm using a fluorescence plate reader . Data from HCMV-infected groups were compared to mock-infected groups and significant differences were determined by Student's t-test or one-way analysis of variance ( ANOVA ) followed by Tukey's post hoc t test using GraphPad Prism 4 software . Data are presented as the means ± standard error of the means ( SEM ) . | A large percentage of the world population is infected with HCMV . As a leading viral cause of birth defects in the developed world , HCMV represents a significant public health burden . For the first time , we report that HCMV infection dysregulates the canonical Wnt/β-catenin signaling pathway which is essential in regulating a diverse range of biological functions . We demonstrate that HCMV infection leads to sequestration and degradation of β-catenin protein , the effector transcription factor in the pathway , thus preventing its downstream signaling activities . Since this pathway is essential in regulating mammalian development and homeostasis , the finding that HCMV impairs this pathway becomes globally important for understanding viral pathogenesis , particularly that related to HCMV disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"obstetrics",
"and",
"gynecology",
"microbiology",
"cell",
"differentiation",
"developmental",
"biology",
"pregnancy",
"obstetrics",
"pregnancy",
"complications",
"signaling",
"in",
"cellular",
"processes",
"virulence",
"factors",
"and",
"mechanisms",
"biology",... | 2012 | Human Cytomegalovirus Infection Dysregulates the Canonical Wnt/β-catenin Signaling Pathway |
Proliferation of bacterial pathogens in blood represents one of the most dangerous stages of infection . Growth in blood serum depends on the ability of a pathogen to adjust metabolism to match the availability of nutrients . Although certain nutrients are scarce in blood and need to be de novo synthesized by proliferating bacteria , it is unclear which metabolic pathways are critical for bacterial growth in blood . In this study , we identified metabolic functions that are essential specifically for bacterial growth in the bloodstream . We used two principally different but complementing techniques to comprehensively identify genes that are required for the growth of Escherichia coli in human serum . A microarray-based and a dye-based mutant screening approach were independently used to screen a library of 3 , 985 single-gene deletion mutants in all non-essential genes of E . coli ( Keio collection ) . A majority of the mutants identified consistently by both approaches carried a deletion of a gene involved in either the purine or pyrimidine nucleotide biosynthetic pathway and showed a 20- to 1 , 000-fold drop in viable cell counts as compared to wild-type E . coli after 24 h of growth in human serum . This suggests that the scarcity of nucleotide precursors , but not other nutrients , is the key limitation for bacterial growth in serum . Inactivation of nucleotide biosynthesis genes in another Gram-negative pathogen , Salmonella enterica , and in the Gram-positive pathogen Bacillus anthracis , prevented their growth in human serum . The growth of the mutants could be rescued by genetic complementation or by addition of appropriate nucleotide bases to human serum . Furthermore , the virulence of the B . anthracis purE mutant , defective in purine biosynthesis , was dramatically attenuated in a murine model of bacteremia . Our data indicate that de novo nucleotide biosynthesis represents the single most critical metabolic function for bacterial growth in blood and reveal the corresponding enzymes as putative antibiotic targets for the treatment of bloodstream infections .
Bacteremia , characterized by the presence of pathogenic bacteria in the bloodstream , is a major cause of morbidity and mortality worldwide . Bacteremia often leads to sepsis and death [1] . To survive in blood , bacterial pathogens must evade a multitude of host defense mechanisms such as complement-mediated lysis , phagocytosis and antimicrobial peptide-mediated killing . The spectrum of complement resistance mechanisms of bacteria is very wide and includes different activities like antigenic variation , use of membrane proteins to block binding of complement proteins and capsule biosynthesis [2–5] . Most extracellular pathogens avoid phagocytosis by synthesizing a protective capsule that helps to evade recognition [6 , 7] . Direct degradation of antimicrobial peptides and modification of cell surface properties are the major strategies used by bacteria to resist the bactericidal activity of host antimicrobial peptides like the platelet-derived thrombocidins in blood [8 , 9] . However , these immune-evasion strategies are mostly pathogen-specific and it is difficult to use the underlying mechanisms as targets for broad-spectrum antibiotics . To proliferate in the various host niches that bacterial pathogens invade during the course of infection they need to adjust their metabolism to suit local nutrient availability . For example , the amount of free iron in human blood is 10−18 M [10] . Most pathogens need 10−6 to 10−7 M iron for growth and hence they use very complex and diverse strategies to acquire and store iron in order to grow in this iron-limiting environment of the host's blood [11] . The most common strategy of iron acquisition is the production of siderophores , high-affinity ( 1030 M−1 ) ferric iron-binding molecules that can sequester iron from the host's iron-binding proteins . Knockout of iron acquisition mechanisms has been shown to attenuate the virulence of many bacteria [12 , 13] . However , the uniqueness of blood as a niche for bacterial growth extends far beyond iron limitation: low availability of certain nutrients may define the ability of bacteria to proliferate in blood . Although the absolute abundance of various metabolites , such as amino acids and nucleotides , in human serum is known [14 , 15] , it is unclear which nutrients are freely available and sufficient and which are limiting for bacterial growth in human serum . Several previous reports described the importance of amino acid or nucleotide biosynthesis by bacteria in the cause of infection . For instance , certain auxotroph mutants of Salmonella [16–18] , Staphylococcus aureus [19] or Streptococcus pneumoniae [20] have been shown to be avirulent in murine models of infection . These reports suggest that purines and some amino acids are scarce in vivo . Also , the inactivation of a potassium transporter in Vibrio vulnificus [21] or of a manganese , zinc and iron transporter in Streptococcus pyogenes [22] have been shown to attenuate virulence of the respective pathogens , suggesting that acquisition of some metal ions is critical for growth in vivo . Notwithstanding the fact that most of these genes were identified in non-exhaustive screens , these studies only provide evidence of the limited availability of the corresponding metabolites in the host . They do not describe nutrient availability in various host compartments invaded by the pathogens during infection . It is unclear whether the reduced virulence of these mutants can be attributed to their inability to grow specifically in blood . The comprehensive identification of genes that are critical , specifically for the bloodstream growth of pathogens , has never been attempted . Hence , crucial nutrient requirements that need to be fulfilled during bacterial growth in blood are essentially unknown . Identification of the limiting nutrients and of bacterial genes that are critical for growth in blood can pinpoint biosynthesis and acquisition strategies that are crucial during the bacteremic stage of infection . Enzymes critical for survival and proliferation of pathogenic bacteria in blood can provide potential targets for treatment of bloodstream infections . To this end , we sought to identify genes required for the growth of bacteria in human blood . We screened a comprehensive gene-deletion library of the model Gram-negative organism , Escherichia coli , for mutants unable to grow in human serum . We found that de novo purine and pyrimidine biosynthesis is the key pathway critical for E . coli growth in serum , thereby revealing the limited availability of nucleotide precursors as the major limitation for bacterial growth in blood . Salmonella enterica , an important Gram-negative pathogen , exhibited a similar requirement for de novo biosynthesis of purines and pyrimidines for growth in serum . Deletion of the corresponding genes in the evolutionarily distant Gram-positive pathogen Bacillus anthracis demonstrated the universal need for these two biosynthetic pathways for bacterial growth in serum .
Our major goal was to identify genes that are critical for the survival and growth of bacteria in blood . Specifically , we were interested in identifying genes that mediate adaptation to the unique nutrient composition of blood rather than those which facilitate immune evasion . Hence , in our experiments we used human serum in which the function of the complement system was inactivated by heat-treatment . E . coli is a major cause of Gram-negative bacteremia in hospitalized patients [23] . We used E . coli as an experimental model for our initial experiments . Specifically , we used a defined library of 3985 E . coli single-gene deletion mutants ( “Keio collection” ) , where all non-essential genes of an E . coli K12 laboratory strain BW25113 have been individually replaced by a kanamycin-resistance cassette [24] . For the identification of genes required for the growth of bacteria in serum , we employed a genetic technique called MGK ( Monitoring of Gene Knockouts ) [25] . MGK is a microarray-based approach that uses the chromosomal flanks of inactivated genes as hybridization targets for custom-designed oligonucleotide microarrays . It allows simultaneous assessment of the relative abundance of thousands of mutants in a population and identification of genes whose inactivation is unfavorable for cell growth under selective conditions . To apply MGK , mutants in the Keio collection were individually grown and mixed at a similar ratio ( see Protocol S1 for details ) . The pooled library was grown in either serum or in LB for 5 h ( approximately 4 generations in serum ) ( Figure 1A ) . Mutants lacking genes critical for growth in blood are expected to be underrepresented in the resulting population of cells grown in serum . Harvested cells were allowed to re-grow in fresh LB medium in order to enrich the population for viable cells and minimize isolation of genomic DNA from dead cells . “MGK targets” corresponding to the flanks of inactivated genes were generated as described in Protocol S1 using genomic DNA isolated from cells grown in the reference ( LB ) and the selective ( serum ) conditions ( Figure 1A ) , and co-hybridized to an oligonucleotide microarray . Two independent MGK experiments ( with dye-swapping ) were performed . Twenty-two mutants with a potential growth defect in serum were identified that consistently showed at least a 2-fold reduction in the hybridization signal of the serum sample as compared to the LB sample . Strikingly , the majority of these mutants ( 15 out of 22 ) carried a deletion of a gene involved in biosynthesis of either purines or pyrimidines ( Table 1 and Figure 2 ) . This result suggested that the de novo biosynthesis of purines and pyrimidines is crucial for the growth of E . coli in human serum and that the scarcity of nucleotide precursors is the major limiting factor for bacterial growth in blood . In an MGK experiment , thousands of mutants are grown together in a mixed population and growth characteristics of each mutant can be potentially affected by metabolites secreted by other cells . In addition , during re-growth in LB media , the mutants that had growth defects in serum could potentially catch up with the rest of the cells . To exclude this scenario , we supplemented the MGK approach with an independent screen involving replica growing of the 3 , 985 individual mutants from the Keio collection , arrayed in a 96-well format , in serum and LB . The inherent turbidity of serum prevents the use of optical density as a measure of bacterial growth . Therefore , we used a dye , 3- ( 4 , 5-dimethylthiazol-2-yl ) -2 , 5-diphenyltetrazolium bromide ( MTT ) , to detect viable cells [26 , 27] . After overnight incubation of mutants in 96-well plates in LB or in human serum , MTT dye was added to the wells and the plates were incubated for another 4–5 h at 37°C . Overnight incubation of the cells in human serum before addition of the MTT dye is expected to identify mutants that are severely impaired in growth in serum rather than mutants that show only modest growth defects . Live , actively growing bacteria reduce MTT to a blue formazan precipitate resulting in a deep blue color of the live serum cultures . The lack of a blue color served as a qualitative indicator of the inability of that mutant to grow in serum ( Figure 1B ) . The MTT screen identified 21 mutants that failed to grow selectively in serum ( Table 1 and Figure S1A ) . Of these , 17 mutants ( 81% ) carried deletions of genes involved in the nucleotide biosynthesis pathway . Fifteen of these 17 mutants were also identified by the MGK analysis ( Figure 2 ) . To verify the phenotypes of E . coli mutants identified by the MGK and MTT screens , we followed their survival and growth in human serum by determining the number of colony forming units ( cfu ) . This rigorous verification showed that mutants which lacked genes belonging to pathways other than purine or pyrimidine biosynthesis ( gcvR , rpoN , lysS , ihfB and rseA identified only by the MGK analysis and nadB , panB , panC , iscS identified only by the MTT assay ) were apparently false positives . Two mutants , ydaS and ydaT , carrying deletions in genes with unknown function , found only in the MGK screen , exhibited a mild growth defect ( ∼10-fold reduction in cfu ) and were not studied further . Notably , however , all mutants carrying a deletion in one of the pur or pyr genes showed a significant ( 20- to ∼1 , 000-fold ) reduction in the viable cell number compared to wild-type E . coli after 24 h growth in serum ( Figure 3A ) . These included 15 pur or pyr mutants identified by both MGK and MTT screens and two pur mutants ( guaA and guaB ) identified only by the MTT screen ( Figure 2 ) . Importantly , after incubation in serum , the cfu counts of some mutants dropped below the initial inoculum suggesting that these mutants not only had a growth defect , but were actively dying in serum ( Figure S1B ) . All pur and pyr mutants grew well in LB medium , indicating that the growth defect was serum-specific ( Figure S1B ) . These results suggest that nucleotide biosynthesis is a critical metabolic function required for growth of E . coli in human serum and that the scarcity of nucleic acid precursors , but not other metabolites , is the key metabolic limitation for bacterial growth in human blood . Some of the E . coli nucleotide biosynthesis genes are organized in operons . Therefore , their replacement by the gene inactivation cassette could have a polar effect on the expression of downstream genes . We tested and eliminated this possibility by genetic-complementation studies . Structural regions of three operon-encoded E . coli genes carA , pyrE , and guaB ( the first genes in the respective operons ) ( Figure S1C ) , along with ∼300 bp upstream regions that include their native promoters , were cloned on a suitable vector and introduced into corresponding mutant strains . The genetically complemented E . coli strains ( carA/pCarAEC , purE/pPurEEC , guaB/pGuaBEC ) thus obtained ( Table S1 ) were tested for growth in serum . Similarly complemented two strains of gene deletions corresponding to monocistronic operons , purA and pyrE ( purA/pPurAEC and pyrE/pPyrEEC ) were also tested . All complemented E . coli mutant strains grew as well as the wild type in serum ( Figure 3B ) . This confirmed that the observed growth defect of the E . coli mutants in serum was due to the lack of the corresponding gene and not the manifestation of polar effects of gene replacement on downstream genes . In order to determine whether the observed growth defect of the E . coli mutants in serum was indeed due to a limiting supply of nucleic acid precursors , mutant growth was tested in serum supplemented with an appropriate nucleobase , either adenine or uracil , for the purine or pyrimidine biosynthesis mutants respectively . Addition of the appropriate nucleobase to serum rescued the growth of the mutants ( Figure 3C and 3D ) . This result confirmed that the purine and pyrimidine deficiency of human blood forces invading bacteria to rely on the de novo biosynthesis of these metabolites . Virulent E . coli strains are extracellular enteric pathogens . In contrast , the Gram-negative pathogen , Salmonella enterica serovar Typhimurium ( S . typhimurium ) , can also replicate intracellularly in phagocytes . Previous studies have shown that inactivation of nucleotide biosynthesis genes in S . typhimurium attenuates its virulence [16 , 28 , 29] . This effect has been attributed solely to the inability of the mutants to colonize the intracellular niche . Our finding that the inactivation of nucleotide biosynthesis genes prevents E . coli from growing in human serum prompted us to test whether Salmonella strains defective in purine or pyrimidine biosynthesis show growth defects in serum . We tested the growth in human serum of 14 different S . typhimurium mutants in which pur or pyr genes were inactivated by transposon insertions ( Table S1 ) . Most of the mutants showed a substantial growth defect with 10- to ∼100-fold or more reduction in viable cell counts after 24 h of growth in human serum as compared to the wild-type strain ( Figure 3E ) . This result demonstrated that de novo nucleotide biosynthesis is required for growth of S . typhimurium in human serum and that the attenuated virulence of such mutants may be due , at least in part , to an inability of the pathogen to multiply in the host's blood . The systemic stage of the life-threatening anthrax infection is characterized by the rapid growth of B . anthracis in the host's blood reaching up to108 bacteria/ml [30] . Few treatment options are available for the late stages of anthrax infections . Thus , it was of particular interest to investigate whether purine and pyrimidine biosynthesis is critical for the growth of Gram-positive B . anthracis in human serum , as was observed for Gram-negative E . coli and S . enterica . Mutants with deletions in either pur ( purE and purK ) or pyr ( carA and pyrC ) genes were constructed in B . anthracis Sterne 34F2 strain ( pXO1+ pXO2− ) by allelic gene replacement . In agreement with the observations made with Gram-negative pathogens , all of the B . anthracis mutants displayed a severe growth defect in human serum: a 50- to ∼1 , 000-fold decrease in viable cell counts as compared to the wild type after 24 h ( Figure 4A ) . None of these mutants showed reduced growth in LB medium ( data not shown ) . Introduction of the plasmid carrying the deleted gene or addition of appropriate nucleobases rescued growth of these mutants in serum ( Figure 4B and 4C and Table S1 ) . This result demonstrates that the limiting amounts of purines and pyrimidines in serum restrict the growth of B . anthracis mutants and shows that de novo nucleotide biosynthesis is a conserved requirement for the growth of at least three bacterial species in human serum . We hypothesized that the growth defect exhibited by B . anthracis nucleotide biosynthesis mutants in serum would manifest at the bacteremic stage of the infection resulting in attenuated virulence . We used a murine model of anthrax infection to test this hypothesis . For these experiments we employed the B . anthracis Sterne strain , which causes lethal infections in certain strains of inbred mice with pathologies similar to systemic anthrax infection in humans [31 , 32] . Murine serum contains about 30-fold more cytidine as compared to human serum ( 3 μM compared to 0 . 1 μM , respectively ) [14] . Therefore , it was not surprising that unlike in human serum , the pyrimidine biosynthesis mutants of B . anthracis did not show any strong growth defect in murine serum . However , as expected , purine biosynthesis mutants purE and purK were defective for growth in murine serum ( Figure S3 ) . The virulence of these two mutants was tested in a murine infection model in which mice were challenged by direct intravenous inoculation with increasing numbers of bacilli and observed for survival following this experimental bacteremia . The purE knockout mutant showed a dramatic decrease in virulence as evidenced by ∼3 . 5 log increase in LD50 ( p = 0 . 002 ) and an increased survival of the challenged mouse cohort ( p < 0 . 001 ) as compared to mice challenged with the wild-type Sterne strain ( Figure 4D ) . Mice challenged with the purE mutant remained healthy for the entire additional 2-week period of observation and no bacteria could be cultured from their blood or spleen . In contrast , a purE mutant strain complemented with a plasmid carrying the purE gene ( purE/pPurEBA ) , was as virulent as the wild-type Sterne strain . Unexpectedly , the purK mutant , which exhibited a similar growth defect in serum in vitro as the purE mutant was almost as virulent as the wild-type Sterne strain ( Figure 4D ) . This result showed that purE , but apparently not purK , is essential for virulence of B . anthracis in mice and thus reveals PurE as a putative antibiotic target for treatment of anthrax bacteremia .
In this paper we demonstrate that de novo nucleotide biosynthesis is critical for survival and growth of bacteria in human serum . A near exhaustive search using two independent screening strategies based on co-growth of E . coli gene knockout mutants ( MGK ) and analysis of individual mutants ( MTT assay ) applied to a comprehensive library of E . coli mutants consistently pointed to the importance of mainly pur and pyr genes for E . coli growth in human serum . The overlap of the results of both screens identified 15 pur or pyr genes that are required for successful growth of E . coli in human serum ( Figure 2 ) . This result was confirmed for two other pathogenic species of bacteria , Gram-negative S . typhimurium and Gram-positive B . anthracis . Inactivation of most of the pur and pyr genes was detrimental to bacterial growth in serum . Of the 13 non-essential genes involved in the purine biosynthetic pathway in E . coli , only two genes , purN and purT , were not identified as being critical for growth in human serum ( Figure 2 ) . This is not surprising because purN and purT both encode 5'-phosphoribosylglycinamide transformylases with partly overlapping specificities [33] , and their individual inactivation should have little effect on cell growth . Of the 9 non-essential genes involved in the pyrimidine biosynthetic pathway our screens did not identify pyrI , pyrD , and ndk mutants . Upon checking the growth of the pyrD mutant in serum , we observed a significant growth defect similar to that of the mutants identified by the two screens ( Figure S2A ) . The other two mutants did not show strong phenotype in serum ( Figure S2A ) . This result was expected because pyrI encodes the regulatory subunit of the aspartate carbamoyltransferase that is not critical for the function of the enzyme [34] , while Ndk is a nucleoside-diphosphate kinase whose function is partly redundant [35] . Iron acquisition genes are known to be important for bacterial adaptation to growth in the iron-limiting environment of blood [10] . Yet no such mutants were identified in our screens , and testing of several such individual E . coli mutants ( entA , fepE , fecA , and tonB ) showed that their growth in serum was not impaired ( Figure S2B ) . One possible explanation for this result is that heat inactivation destroys transferrin-iron complexes and releases free iron that can be used by bacteria . The virulence of B . anthracis pur mutants has been previously characterized in a murine peritoneal cavity infection model [36] . Of all the mutants tested by Ivanovics et al , only those lacking PurA or PurB activity were found to be significantly attenuated in mice . On the other hand , our study shows that the B . anthracis purE mutant is significantly attenuated in virulence in a murine bacteremia model . These results suggest that only certain enzymes in the purine biosynthetic pathway can be potential targets for antibiotics . The observed attenuation of the B . anthracis purE mutant not only reveals the PurE enzyme as a novel target for the development of anti-anthrax therapies , but also indicates that deletion of purE in a fully virulent B . anthracis strain could be a promising way to develop a live attenuated vaccine . In contrast to the purE knockout strain , the purK mutant of B . anthracis remained virulent and killed mice with an LD50 that was only slightly higher than that of the wild-type Sterne strain ( Figure 4D ) . This result is not too surprising . PurK catalyses the carboxylation of aminoimidazole ribonucleotide ( AIR ) leading to an unstable intermediate , N5-carboxyaminoimdazole ribonucleotide ( N5-CAIR ) [37 , 38] , which is then converted by PurE to carboxyaminoimidazole ribonucleotide ( CAIR ) . In vitro studies showed that a significant fraction of AIR can be non-enzymatically converted to N5-CAIR at a high concentration of bicarbonate [39] . The small amounts of N5-CAIR produced spontaneously in the purK mutant might relieve the block in purine biosynthesis and rescue the auxotrophy [40] . Our data strongly suggest that nucleotide biosynthesis is the key metabolic pathway which is critical for proliferation of bacterial pathogens in blood . It might be argued that our conclusions could be deduced empirically from a known low concentration of free nucleobases in the human blood serum . Indeed , some previous studies pointed to a relative scarcity of purines in blood [19] . However , other studies suggested that purines and pyrimidines are sufficiently abundant in this niche [41] and the importance of nucleotide prototrophy for bacterial growth in human blood has never been clearly demonstrated . Furthermore , the simple knowledge of the nutrient composition of blood serum is not sufficient to predict which biochemical pathway will be rate-limiting for bacterial growth in this niche . Thus , although threonine , lysine , histidine , aromatic amino acid and riboflavin biosynthetic functions have been shown to be important for bacterial infections , none of these pathways appears to be critical for growth of bacteria in blood [19 , 42 , 43] . Several previous studies have shown that pathogens require nucleotide biosynthesis to establish a successful infection [16 , 19 , 20 , 44 , 45] . Our study , however , is the first to our knowledge to demonstrate the importance of the de novo synthesis of purines and pyrimidines for successful proliferation of pathogens specifically in blood . Reproduction of the phenotypes associated with several identified gene knockouts in Gram-negative bacteria , E . coli and S . typhimurium , in a Gram-positive pathogen B . anthracis , suggests the universal importance of the nucleotide biosynthesis pathways for growth of bacteria in the bloodstream . Indeed , as our data shows , nucleotide biosynthesis may be the only metabolic pathway that is universally required by bacterial pathogens invading the blood . The corresponding enzymes thus appear as putative antibiotic targets for curbing bacterial growth in the bacteremia stage of infection . Of the enzymes identified in this study as essential for growth of bacterial pathogens in human blood , two , PyrC and PurE , are especially attractive as targets for antibiotics . PyrC is a dihydroorotase that catalyzes the reversible cyclization of carbamoyl aspartate to dihydroorotate [46] . The known dihydroorotases fall into two major sequence classes . Class I enzymes are conserved among fungi and most Gram-negative proteobacteria whereas higher eukaryotes use class II dihydroorotases [47] . The limited sequence conservation between the two types of PyrC ( less than 20% ) [48] makes this enzyme an attractive target for the treatment of Gram-negative and fungal infections . PurE shows a higher degree of overall conservation between bacteria and eukaryotes ( Table S2 ) . However , the catalytic mechanisms of the bacterial and eukaryotic PurE enzymes are substantially different . In bacteria , PurE utilizes N5-CAIR to make CAIR , whereas human PurE uses AIR and CO2 and does not recognize N5-CAIR as a substrate [49] , pointing to a significant difference in the structure of the catalytic centers in the bacterial and human enzymes . Given that the B . anthracis purE mutant is avirulent , apparently due to inability of bacteria to proliferate in blood , PurE emerges as an attractive target for antibiotic therapies . Nucleotides are important substrates not only for DNA synthesis but also for DNA repair . Thus inhibitors targeting the nucleotide biosynthesis functions identified in this study can also impede the repair processes following bacterial DNA damage induced by the host's reactive oxygen intermediates during infection . Detailed exploration of PyrC and PurE as well as other enzymes of the nucleotide biosynthesis pathways as potential antibiotic targets may lead to development of new therapies for treatment of bacterial bloodstream infections .
The collection of gene knockout mutants of the E . coli strain BW25113 ( Keio collection ) was obtained from Nara institute of Science and Technology , Japan [24] . Wild-type Salmonella enterica serovar Typhimurium LT2 and isogenic Tn5 transposon insertion mutants were obtained from the Salmonella Genetics Stock Center ( Alberta , Canada ) ( http://www . ucalgary . ca/~kesander/ ) . B . anthracis Sterne ( pXO1+ , pXO2− ) wild-type strain [50] was used to construct pur and pyr biosynthesis mutants . All the cloning procedures were carried out using the OneShot TOP10 chemically competent E . coli cells ( Invitrogen ) as the host . E . coli strain GM 2163 ( New England Biolabs ) was used to obtain unmethylated plasmid DNA for transformation of B . anthracis . The vector pKS1 was used to construct deletion mutants in B . anthracis [51] . Genetic complementation was carried out using recombinant plasmids based on the pBAD22 vector for E . coli or pHT304 vector [52] and pHTPSAC vector ( H . Lee , unpublished data ) for B . anthracis . Construction of the plasmids used for genetic complementation is outlined in Protocol S1 . All primers used in this study are listed in Table S3 . For all experiments involving growth of bacteria in serum , frozen serum [Sterile filtered type-AB human serum , Cat No . H4522 ( Sigma ) ] was thawed at 37°C , heat-inactivated by incubation for 30 min at 56°C and buffered at pH 7 . 0 by addition of 1 M HEPES buffer ( pH 5 . 2 ) to the final concentration of 5 mM . Control cultures were grown in Luria-Bertani medium [53] . When necessary , antibiotics were added at the following final concentrations: kanamycin 30 μg/ml for E . coli or 100 μg/ml for B . anthracis , erythromycin 200 μg/ml for E . coli or 5 μg/ml for B . anthracis and ampicillin 100 μg/ml for E . coli . The MGK selection was carried out essentially as described , by Smith et al [25] , with minor modifications specified in Protocol S1 . Mutants from the Keio collection were replicated into 96-well plates containing 100 μl/well of LB supplemented with kanamycin . Plates were incubated overnight at 37°C with shaking . Cells were pelleted , washed once with PBS and resuspended in 100 μl PBS . 1 μl culture from each well was inoculated into fresh 96-well plates containing either LB or heat-inactivated human serum and incubated overnight with shaking at 37°C . The next day , 10 μl of 0 . 5% dye solution [MTT; ( 3- ( 4 , 5-dimethylthiazol-2-yl ) -2 , 5-diphenyltetrazolium bromide ) in PBS] was added to each well and the plates were incubated for 4–5 h at 37°C . When quantitation was required , plates were scanned using a Hewlett Packard Scanjet 5300C and Adobe Photoshop was used to determine luminosity values in each well ( Figure S1A ) . B . anthracis Sterne strain , the purE and purK knockout mutants , or genetically complemented mutant purE/pPurEBA ( Table S1 ) were grown in LB media , washed three times with PBS , and resuspended in PBS at a cell density of ∼5x108 cfu/ml . Cohorts of adult ( 8–10 weeks old ) , female , NIH-Swiss mice , obtained from Frederick Cancer Research Center , were inoculated via tail vein with 0 . 2 ml of PBS containing serial 10-fold dilutions of vegetative B . anthracis cells ( 103-108 cfu/mouse; 4–6 mice per bacterial dose in the studies represented ) . Mice were observed daily for 6 days for signs of fatal outcome , using humane end points approved by the UIC institutional animal care and use committee . To confirm that the fatal infection was caused by an inoculated strain , the presence of the correct gene knockout was verified by PCR analysis of the bacterial colonies recovered from blood or spleen of diseased animals . 50% lethal dose ( LD50 ) was estimated using the Spearman-Karber method [54] . The LD50 data were converted to average latency survival ( ALS ) curves using the described transformation [55] . Data were analyzed using SigmaPlot software and log rank test of the significance of differences in survival curves and t-test analysis of the significance of differences in LD50s among animal cohorts challenged with different bacterial strains . | Bacterial growth in the bloodstream is a common manifestation of a number of bacterial infections . When growing in blood , bacteria not only have to evade the host's immune response , but also adjust their metabolism to suit availability of nutrients . Although the concentrations of various metabolites in human blood are known , it is difficult to predict which nutrients are abundant and which are scarce . To proliferate in human blood , bacteria need to synthesize metabolites that are present in the limiting concentrations . For that , they need to produce specific enzymes that are , thus , critical for the bacterial growth in the bloodstream . We carried out a comprehensive , genome-wide search for Escherichia coli genes that are essential for growth in human serum . We found that inactivation of nucleotide biosynthesis genes leads to a significant growth defect in human serum not only for E . coli but also for two other pathogens , Salmonella Typhimurium and Bacillus anthracis . The results of this study demonstrate that the limiting amounts of the nucleotide bases in human serum force invading pathogens to rely on de novo nucleotide biosynthesis . Hence , our findings reveal nucleotide biosynthesis enzymes as a possible target for the treatment of bloodstream infections . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases",
"microbiology",
"eubacteria"
] | 2008 | Nucleotide Biosynthesis Is Critical for Growth of Bacteria in Human Blood |
Pyrethroid resistance is envisioned to be a major problem for the vector control program since , at present , there are no suitable chemical substitutes for pyrethroids . Cross-resistance to knockdown agents , which are mainly used in mosquito coils and related products as spatial repellents , is the most serious concern . Since cross-resistance is a global phenomenon , we have started to monitor the distribution of mosquito resistance to pyrethroids . The first pilot study was carried out in Vietnam . We periodically drove along the national road from the north end to the Mekong Delta in Vietnam and collected mosquito larvae from used tires . Simplified susceptibility tests were performed using the fourth instar larvae of Aedes aegypti , Aedes albopictus , and Culex quinquefasciatus . Compared with the other species , Ae . aegypti demonstrated the most prominent reduction in susceptibility . For Ae . aegypti , significant increases in the susceptibility indices with a decrease in the latitude of collection points were observed , indicating that the susceptibility of Ae . aegypti against d-allethrin was lower in the southern part , including mountainous areas , as compared to that in the northern part of Vietnam . There was a significant correlation between the susceptibility indices in Ae . aegypti and the sum of annual pyrethroid use for malaria control ( 1998–2002 ) . This might explain that the use of pyrethroids as residual treatment inside houses and pyrethroid-impregnated bed nets for malaria control is attributable to low pyrethroid susceptibility in Ae . aegypti . Such insecticide treatment appeared to have been intensively administered in the interior and along the periphery of human habitation areas where , incidentally , the breeding and resting sites of Ae . aegypti are located . This might account for the strong selection pressure toward Ae . aegypti and not Ae . albopictus .
One of the most successful events in the development of pesticide chemicals was the discovery of pyrethrum and the successful synthesis of pyrethroids . For example , allethrin [1] , a classical synthetic pyrethroid , continues to be used for preventing mosquito bites without any toxicological and operational problems . There are two main groups of pyrethroids: one possessing high knockdown activity but low killing activity and the other possessing high killing activity . The pyrethroids in the former group , such as d-allethrin , are aptly labeled as knockdown agents , and those in the latter group , as killing agents . Generally , the pyrethroids belonging to the latter group exhibit high photostability that enables their outdoor use , for example , as agricultural pesticides . Nowadays , photo-stable pyrethroids are emerging as the predominant insecticides for vector control . In fact , photo-stable pyrethroids comprise 40% of the insecticides used annually on a global level for indoor residual spraying against malaria vectors and 100% of the WHO-recommended insecticides for the treatment of mosquito nets . The exception is the use of dichlorodiphenyltrichloroethane ( DDT ) in African countries [2] . Pyrethroid resistance to photo-stable pyrethroids is envisioned to be a major problem for the vector control program since , at present , there are no chemical substitutes for pyrethroids . Moreover , cross-resistance to knockdown agents is the most serious concern . Since this is a global phenomenon , we have started to monitor the distribution of mosquito resistance to pyrethroids . The first pilot study was carried out in Vietnam . In the course of the National Dengue Control Program in Vietnam , the ecological differences between two dengue vectors , namely , Aedes aegypti ( L . ) and Aedes albopictus ( Skuse ) , that are of different geographical origins have become a subject of discussion . The ecological differences in the vectors may be the reason for the differences in the epidemics occurring nationwide . In order to examine the nationwide distribution of the two species and pyrethroid resistance , used tires , one of the main breeding grounds of the two species , were targeted for larval collection . Used tires were suitable targets for our study as a standard breeding sites because they were extensively and commonly distributed along roads in Vietnam . In the present paper , distribution analysis of pyrethroid-susceptibility of the larvae of three major mosquito species , namely , Ae . aegypti , Ae . albopictus , and Culex quinquefasciatus ( Say ) , that were collected from used tires located along the national road in Vietnam was carried out . The relationships between pyrethroid susceptibility and the geographical factors and province based annual pyrethroid use are also discussed .
We periodically drove along the national road from the north end to the Mekong Delta in Vietnam ( 7–16 December 2006; 17–20 March , 15–20 May , and 1–12 July 2007; and 7–16 January 2008 ) and collected mosquito larvae from used tires ( Figure 1 ) . Whenever we encountered the used tires , most of which were found along the periphery of repair shops , while driving along a systematically determined route , the geographical position ( with a global positioning system [GPS] ) , number of tires , presence of water in the tires , existence of mosquito larvae in the tires , and environmental characteristics such as land with vegetation within a 25-m radius and distribution of houses along the road were recorded . Mosquito larvae were collected from tires with larvae ( 6–20 tires per site ) by netting ( 5 times per tire ) . The collected larvae , except those used for the susceptibility test , were placed in 1 . 5 ml plastic vials containing ethanol solution for identification at a later time . The bioassay for the detection of knockdown susceptibility was carried out on the day of collection by using the mosquito larvae obtained from the collection sites from where we could procure an adequate number of insects . The larvae collected from each collection site were briefly identified on the day of collection , and fourth instar larvae of the Stegomyia group , which occasionally comprised a mixed population of Ae . aegypti and Ae . albopictus , and Cx . quinquefasciatus individuals , were used for the susceptibility test . Each larva was placed in a glass vial with 20 ml of water . An emulsifiable concentrate of 90% d-T80-allethrin was diluted with water to obtain a 250 ppm solution . After releasing the larva , 32 or 8 microliters of the solution was added in each vial to obtain a concentration of 0 . 4 and 0 . 1 ppm , respectively . Twenty larvae from each site were used for each concentration regime . Knockdown of the larvae was observed for 30 min . Larvae that sank to the bottom of the glass vial and could not swim , float , or were paralyzed were judged as knocked down larvae; the time to knockdown was recorded for each larva . After the test , each larva was placed in a 1 . 5 ml plastic vial containing ethanol solution for identification at a later time . After identification , individual knockdown data were summarized for each mosquito species ( Ae . aegypti , Ae . albopictus , and Cx . quinquefasciatus ) . The median knockdown times ( KT50s ) , i . e . , the time required for 50% knockdown , were scored according to the 6 following categories: 1 , <5 min; 2 , 5–10 min; 3 , 10–15 min; 4 , 15–20 min; 5 , 20–30 min; and 6 , >30 min . The susceptibility index was calculated as the product of the scores at 0 . 1 and 0 . 4 ppm . Thus , mosquitoes with susceptibility index of 1 were considered to be the most susceptible and those with susceptibility index of 36 were considered to be the least susceptible to d-allethrin . In order to confirm the validity of adopting a simplified larval bioassay as a substitute method for determining adult knockdown susceptibility , we performed a knockdown test using several adult colonies of Ae . aegypti with known larval susceptibility indices . Four Ae . aegypti colonies collected in the field in Vietnam ( Ho Chi Minh city , Ben Tre , Ben Luc , Long An; all colonies were collected in 2007 ) and 1 standard colony ( transferred from Sumitomo Chemical Co . Ltd . , Hyogo , Japan ) that had been maintained in a laboratory after collection were used in the bioassay . A mosquito coil containing 0 . 3% d-T80-allethrin was cut into pieces of 0 . 5 g . A piece of the coil was ignited at both sides and burned completely in a glass chamber ( 70×70×70 cm ) that contained a small battery-operated fan for moving the air . Immediately after the mosquito coil burned out , 3 to 5 day old 20 blood-unfed female mosquitoes were released into the chamber , and their knockdown was observed for 20 min . The test was performed in triplicate . The KT50 values were calculated by Bliss' probit method [3] . Univariate analysis was performed to analyze the effects of several factors on susceptibility . Latitude at each collection points were used as the geographical factor . Province based annual pyrethroid uses for malaria control in 1998–2002 ( National Institute of Malariology , Parasitology and Entomology , Vietnam , unpublished data ) were also used for the analysis . Province-based data were transformed into a raster image using the spline function with ArcGIS 9 . 2 ( ESRI Japan Corp . ) to obtain extrapolated values at the collection points . The extrapolated values were log transformed and the susceptibility indices were categorized into two levels , ≥36 and <36 , for the statistical analysis using JMP 7 . 0 ( SAS Institute Inc . ) .
Larvae were collected from used tires along roads in Vietnam from 527 collection points throughout the collection surveillance . Used tires were extensively and commonly distributed . Collection sites were categorized into 5 areas: ( 1 ) northern mountainous area ( 172–506 m in height ) , ( 2 ) northern plain area , ( 3 ) eastern coastal area , ( 4 ) southern mountainous area ( 103–1563 m in height ) , and ( 5 ) southern wetlands . Among the 19188 tires surveyed in total during the investigation , 4757 were examined and 2468 contained water ( 51 . 9% ) ; 852 contained mosquitoes ( 34 . 5% ) , and of these 653 contained dengue vector mosquitoes ( Ae . aegypti and Ae . albopictus ) ( 26 . 5% ) . In total , 8771 Ae . aegypti , 5916 Ae . albopictus , and 11356 Cx . quinquefasciatus larvae were collected . In the northern part of Vietnam , Ae . albopictus was dominant , and this dominance gradually reduced toward the south . Ae . aegypti was dominant in the southern part . In the eastern coastal areas of the southern part , almost 100% of the Aedes larvae collected were Ae . aegypti . However , in the mountainous areas , the number of Ae . albopictus increased , suggesting that the distribution of the 2 species is determined by different environmental gradients ( Figure 2 ) [4] . Figure 3 shows the correlation between the larval susceptibility index and the adult KT50 of the same colony of Ae . aegypti for samples collected from several places in Vietnam . The adult KT50 for the standard strain , which showed normal susceptibility to insecticides was 3 . 74 min , and its larval susceptibility index was 6 . In contrast to the standard strain , the field-collected strains showed higher KT50s and larval susceptibility indices . Further , a relatively high correlation was observed between the knockdown susceptibility of larvae and that of adults , indicating that the present simplified knockdown test is a good reflection of adult susceptibility . Distribution of susceptibility indices for Ae . aegypti , Ae . albopictus , and Cx . quinquefasciatus are shown in Figure 4 . The number of points where the susceptibility tests were carried out was 67 for Ae . aegypti , 50 for Ae . albopictus , and 73 for Cx . quinquefasciatus . The average susceptibility index and the proportion of mosquitoes with susceptibility indices greater than 20 was 24 . 9 and 59 . 7% for Ae . aegypti , 8 . 48 and 8 . 0% for Ae . albopictus , and 12 . 6 and 21 . 9% for Cx . quinquefasciatus , respectively . When compared with the other species , reduction in susceptibility was most prominent in Ae . aegypti ( analysis of variance [ANOVA] , P<0 . 0001; Tukey's HSD test , P = 0 . 05 ) . In contrast , the reduction in the susceptibility of Ae . albopictus and Cx . quinquefasciatus was moderate . The susceptibility of the abovementioned species to d-allethrin was lower in the southern part ( <13°N ) than in the northern part ( >13°N ) of Vietnam . For Ae . aegypti , univariate analysis showed significant increases in the susceptibility indices with decrease in the latitude of the collection points ( χ2 = 36 . 7 , p<0 . 0001 ) . In contrast , there was no significant correlation between the susceptibility indices and latitude of collection points in Ae . albopictus ( χ2 = 2 . 57 , p = 0 . 109 ) and Cx . quinquefasciatus ( χ2 = 2 . 43 , p = 0 . 119 ) . In Ae . aegypti , significant correlation between the susceptibility indices and the sum of annual pyrethroid use for malaria control during 1998 and 2002 ( χ2 = 5 . 95 , p = 0 . 0147 ) was also observed . No significant correlations were seen between the susceptibility indices and the sum of annual pyrethroid use for malaria control in Ae . albopictus ( χ2 = 0 . 945 , p = 0 . 331 ) and Cx . quinquefasciatus ( χ2 = 0 . 224 , p = 0 . 636 ) .
The most plausible procedure for investigating the knockdown susceptibility of mosquitoes to pyrethroids is to acquire an adequate number of field-collected female adults or laboratory-reared colonies by rearing field-collected larvae and to carry out a knockdown bioassay with the adults and the actual pyrethroid formulation , such as a mosquito coil , in a laboratory . However , it was impossible for us to follow this procedure since 190 field-collected larval samples were acquired in the present study . Susceptibility tests were , therefore , performed on the day of collection according to a simplified protocol by using the fourth instar larvae . In many cases , the adult susceptibility scores correlated with the larval susceptibility scores [5] , [6] . This is , however , not always true for all cases since mosquitoes might develop different resistant mechanisms with different metabolic pathways in the larval and adult stages . In the present study , the authors focused on knockdown that might not be chiefly dependent on the enhancement of metabolic activity but may be dependent on the nervous insensitivity controlled by the kdr gene . The simple bioassay with mosquito larvae that is presented in this paper might be a convenient and cost-effective method for evaluating mosquito knockdown resistance in the field . The present paper reports some interesting points concerning the distribution of pyrethroid resistance among mosquitoes in Vietnam . The most prominent observation is that the susceptibility of Ae . aegypti to d-allethrin decreased significantly with decrease in the latitude of the collection points . Vu et al . reported similar tendency in pyrethroid susceptibility in Ae . aegypti in Vietnam [7] . The authors conducted WHO standard bioassay using adult Ae . aegypti collected in 22 places in 11 provinces and cities in four different regions of Vietnam and found that the mosquitoes were susceptible to pyrethroids in many places in the North and Centre regions but they were resistant in the South and Central Highlands in Vietnam . They concluded this discrepancy in pyrethroid susceptibility in different regions to be due to the longer and extended use of pyrethroids in malaria and dengue fever control programs and in agriculture in the Southern and Central Highlands . Actually , a lot of pyrethroids have been treated as residual treatment inside houses and pyrethroid-impregnated bed nets for malaria control ( Figure 5 ) as a part of the National Malaria Control Program [8] , [9] , [10] . The pyrethroid use for malaria control seems to be important factor in developing pyrethroid resistance in Ae . aegypti in highland region , since the DF/DHF cases are not serious [11] and forest malaria continues to be endemic [12] , [13] , [14] in highland region as compared to the other regions and consequently the amount of pyrethroid treatment for dengue vector control in highland region is lower than the other regions ( Figure 5 ) . The pyrethroid treatment for malaria vector control appears to have been intensively conducted in the interior and along the periphery of human habitation areas , where incidentally , the breeding and resting sites of Ae . aegypti are located . This might account for the strong selection pressure toward Ae . aegypti and not so much toward Ae . albopictus since Asian Ae . aegypti is generally a domestic and endophagic that has a greater preference for indoor breeding than Ae . albopictus [15] , [16] , [17] , [18] , [19] . Several papers report the pyrethroid resistance of both Asian Ae . aegypti and Ae . albopictus . Most of them report that Ae . aegypti had higher pyrethroid resistance than Ae . albopictus , indicating pyrethroid resistance was affected by ecological differences in mosquitoes [20] , [21] , [22] , [23] . In Vietnam , 24 tonnes of DDT was used for residual treatment against malaria vectors in 1993 and 1994 . However , since the abandoning of DDT sprays in 1995 , only pyrethroids ( residual spraying of λ-cyhalothrin and α-cypermethrin and occasionally deltamethrin , and permethrin-impregnated bed nets ) have been extensively used in large amounts , unlike in the other Asian countries [2] , [10] . Although details regarding the amount of insecticides used for dengue control in Vietnam have not been published , 21 , 000 liters of photo-stable pyrethroid formulations such as λ-cyhalothrin , deltamethrin , and permethrin was reported to be used for dengue control in 20 southern provinces in 2007 [24] . The extensive use of photo-stable pyrethroids , therefore , seems to have been very common in southern Vietnam . Insecticides still provide the most promising countermeasures for controlling malaria , dengue hemorrhagic fever ( DHF ) , and other arthropod-borne diseases . On an average , at the global level , 547 tones of DDT , 39 tones of organophosphates , 23 tones of carbamates , and 41 tones of pyrethroids are used as active ingredients annually for indoor residual spraying against malaria vectors [2] . The average total amount of pyrethroids used annually as active ingredients between 2003 and 2005 at the global level was 161 tones , which is 16% of the total insecticide consumption and 36% of the total insecticide consumption if the amount of DDT , which is exclusively used in African countries , is excluded . Among pyrethroids that are used for vector control , 98 . 7% comprise photo-stable pyrethroids such as α-cypermethrin , cypermethrin , bifenthrin , cyfluthrin , deltamethrin , etofenprox , λ-cyhalothrin , and permethrin . On the other hand , the most popular and long-standing formulations using pyrethroids are mosquito coils , mosquito mats , and liquid vaporizers . Pyrethroids belonging to the knockdown agent group , such as allethrin , pyrethrin , and prallethrin , are used in these formulations . In particular , d-allethrin still continues to be used in these types of formulations . The use of pyrethroids for preventing mosquito bites as a “spatial repellent” [25] , [26] , [27] , [28] , [29] , [30] , [31] , [32] , [33] , [34] is believed to be biorational since mosquitoes develop minimum physiological resistance , since it does not kill the affected insects and causes no selection pressure on insect populations . Several factors are believed to play major roles in inducing pyrethroid resistance in mosquitoes . The most serious factor is the uncontrolled use of photo-stable pyrethroids . Photo-stable pyrethroids persist on substrates such as wall and floor surfaces for long periods and hence continue to kill insects that make contact with these substrates . This induces a strong selection pressure on the insect population resulting in a population of resistant offspring . Ultra-low volume ( ULV ) space sprays are commonly applied for the emergency control of dengue vectors in urban areas . Space sprays by themselves do not induce a very strong selection pressure; however , uncontrolled treatment at high frequency facilitates the development of resistance . In the past , the use of pyrethroids in aqueous environments was impossible since pyrethroids are highly toxic to aqueous organisms . However , recently , a new pyrethroid with low fish toxicity has been commercially produced and is widely used in aqueous environments such as paddy fields . This might cause considerable selection pressure on the mosquito larvae distributed in such environments . The discovery of the phenoxybenzyl alcohol moiety accelerated the development of photo-stable pyrethroids that could be used for agricultural purposes . These “second generation” pyrethroids have been used worldwide as good vector control agents with various application techniques such as residual spraying , ULV spraying , and bed net treatment ( long-lasting insecticide-treated net [LLITN] ) . However , photo-stable and highly effective pyrethroids might accelerate the development of pyrethroid resistance in mosquito populations . The most serious problem is that resistance to a single pyrethroid causes cross-resistance to all other pyrethroids , including knockdown agents . In fact , many reports concerning pyrethroid resistance have emerged after the successful application of pyrethroids as vector control agents [35] . Therefore , the uncontrolled use of such pyrethroids might lead to the end of the golden age of pyrethroids . Humans have invented insecticides to ensure comfort and to achieve ideal conditions . Good insecticides , therefore , should be as effective as possible so that the abovementioned goals are realized . However , the development and manufacturing costs of insecticides should be as low as possible . It is , therefore , our duty to use insecticides in the most effective and prudent manner possible in order to maintain their effectiveness and sustain their use . In order to effectively manage pyrethroid resistance , the establishment of a feasible insecticide management system and a regular monitoring system of pyrethroid susceptibility will be essential . Moreover , it is expected that the use of photo-unstable knockdown agents as spatial repellents , which effectively interfere with disease transmission without causing any selection pressure to insect populations , will be reconsidered . | Pyrethroid is one of the most successful insecticidal chemicals for controlling insect pests and vectors of household and public health importance . However , extensive treatment of photo-stable and highly effective pyrethroids sometimes causes resistance to insect populations . Resistance to pyrethroids belonging to the knockdown agent groups that have been used as the spatial repellents , such as mosquito coils and related products , is the most serious concern . Since this is a global phenomenon , we have started to monitor the distribution of mosquito resistance to pyrethroids and the first pilot study was carried out in Vietnam . Our study concluded that Aedes aegypti , the most important vector of dengue and dengue haemorrhagic fever , demonstrated the prominent reduction in susceptibility against d-allethrin , one of the most popular pyrethroids for spatial repellents . The extensive treatment of photo-stable pyrethroids for malaria control seemed to be one of the attributable factors . In order to effectively manage pyrethroid resistance , the establishment of a feasible insecticide management system and a regular monitoring system of pyrethroid susceptibility will be essential . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"public",
"health",
"and",
"epidemiology"
] | 2009 | Nationwide Investigation of the Pyrethroid Susceptibility of Mosquito Larvae Collected from Used Tires in Vietnam |
Latently infected resting CD4+ T cells are a major barrier to HIV cure . Understanding how latency is established , maintained and reversed is critical to identifying novel strategies to eliminate latently infected cells . We demonstrate here that co-culture of resting CD4+ T cells and syngeneic myeloid dendritic cells ( mDC ) can dramatically increase the frequency of HIV DNA integration and latent HIV infection in non-proliferating memory , but not naïve , CD4+ T cells . Latency was eliminated when cell-to-cell contact was prevented in the mDC-T cell co-cultures and reduced when clustering was minimised in the mDC-T cell co-cultures . Supernatants from infected mDC-T cell co-cultures did not facilitate the establishment of latency , consistent with cell-cell contact and not a soluble factor being critical for mediating latent infection of resting CD4+ T cells . Gene expression in non-proliferating CD4+ T cells , enriched for latent infection , showed significant changes in the expression of genes involved in cellular activation and interferon regulated pathways , including the down-regulation of genes controlling both NF-κB and cell cycle . We conclude that mDC play a key role in the establishment of HIV latency in resting memory CD4+ T cells , which is predominantly mediated through signalling during DC-T cell contact .
Antiretroviral therapy ( ART ) for the treatment of HIV has led to a substantial reduction in morbidity and mortality; however , ART cannot cure HIV and life-long treatment is required . This is directly due to the persistence of long-lived latently infected cellular reservoirs , that include microglia , astrocytes , macrophages and naïve T cells [1]–[4] , however , resting memory CD4+ T cells [5]–[7] , are considered to be the major contributors . Latently infected resting CD4+ T cells are found in blood and tissue sites , including lymphoid tissue and the gastrointestinal tract [7]–[10] . The frequency of latently infected cells is up to ten times higher in tissue than in blood in HIV-infected patients or SIV-infected macaques on suppressive ART [8] , [10] . It is unclear how latency is established in vivo . However , in vitro , latency can be established following survival of an activated CD4+ T cell that returns to a resting state carrying integrated virus [5] , [11]–[13] . Alternatively , latency has also been established following direct infection of resting cells in the presence of chemokines or following spinoculation [14]–[19] . Dendritic cells ( DC ) are found throughout the body and interact closely with resting CD4+ T cells within lymphoid tissues . Therefore , given the high frequency of latently infected cells in lymphoid tissue , we hypothesised that latency in resting CD4+ T cells may result from interactions with DC as CD4+ T cells recirculate through lymphoid tissue . Using a novel model of resting CD4+ T cells co-cultured with primary DC , we demonstrate that myeloid DC ( mDC ) induce post-integration latency in resting memory CD4+ T cells , which required close DC-T cell contact .
Resting CD4+ T cells and syngeneic DC ( including the two major blood DC subpopulations , plasmacytoid ( pDC ) and myeloid ( mDC ) DC ) were sorted from the blood of healthy donors ( Fig . S1 ) . Seminaphtharhodafluor-1 ( SNARF ) -labelled resting CD4+ T cells were cultured either alone or co-cultured with DC at a DC: T cell ratio of 1: 10 . Following 24 hours of culture , cells were infected with a CCR5-tropic enhanced green fluorescent protein ( EGFP ) -reporter virus , NL ( AD8 ) -nef/EGFP ( multiplicity of infection , MOI 0 . 5 ) , and cultured for 5 days ( Fig . 1 A ) . Cells were then analysed for expression of EGFP by flow cytometry to quantify productive infection ( Fig . 1 B ) . In the DC-CD4+ T cell co-cultures , we detected a spreading productive infection with the number of infected cells 5 days post-infection significantly greater ( median ( IQR ) = 40 ( 31 , 150 ) EGFP+ cells/104 cells; n = 5 ) compared to CD4+ T cells cultured alone ( 1 . 5 ( 1 , 2 . 5 ) EGFP+ cells/104 cells , p = 0 . 03; Fig . 1 C ) . These results were consistent with previous work demonstrating enhanced productive infection of CD4+ T cells in the presence of DC [20] , [21] . At day 5 post-infection , non-proliferating ( SNARFhi ) CD4+ T cells that were not productively infected ( EGFP− ) were sorted ( purity was always >99% ) . Latent virus was quantified in the SNARFhiEGFP− CD4+ T cells upon stimulation with phytohaemagglutinin ( PHA ) in the presence of feeder peripheral blood mononuclear cells ( PBMC ) following a further 5 days of culture ( Fig . 1 B ) . The number of EGFP+ cells following stimulation was , therefore , a surrogate measure for the number of latently infected cells in the SNARFhiEGFP− CD4+ T cells . When SNARFhiEGFP− CD4+ T cells were sorted from cultures infected in the absence of DC , few latently infected cells were detected ( 2 ( 1 , 8 . 5 ) EGFP+ cells/104 cells; n = 5 ) . In contrast , when SNARFhiEGFP− CD4+ T cells were sorted from the DC-T cell co-cultures a significant increase in the number of latently infected cells was observed ( 41 ( 28 , 73 ) EGFP+ cells/104 cells; p = 0 . 03; n = 5; Fig . 1 D ) . Furthermore , when infections were performed in the presence of the protease inhibitor indinavir , there was no significant difference in the number of latently infected cells , as measured by EGFP expression following co-culture of SNARFhiEGFP− with PHA and feeder PBMC ( Fig . 1 E ) . This confirms that a productive , spreading infection was not required to establish latency . Together , these results demonstrate that DC facilitate latent HIV infection in non-proliferating CD4+ T cells . We next asked whether DC-T cell co-culture had activated the SNARFhi CD4+ T cells , which allowed for HIV entry . Sorted SNARFhiEGFP− CD4+ T cells that were co-cultured with DC for 5 days showed signs of early activation with increased expression of CD69 ( 1 . 5% ( 0 . 1 , 2 . 1 ) ; p = 0 . 02; n = 4; Fig . 1 F ) . However , these cells did not express either HLA-DR or Ki67 ( Fig . 1 F ) . As expected , resting CD4+ T cells that were cultured alone did not express any of the activation markers . These results confirmed that the sorted SNARFhiEGFP− cells were non-proliferating , partially activated CD4+ T cells . To determine whether mDC or pDC were facilitating latency in resting CD4+ T cells , we next co-cultured sorted mDC and pDC with SNARF-labelled resting CD4+ T cells for 24 hours prior to infection , and experiments were performed as described above . While productive infection was enhanced in both the mDC and pDC-T cell co-cultures ( Fig . 2 A ) , latent infection was only identified in the SNARFhiEGFP− CD4+ T cells that had been co-cultured with mDC ( 33 ( 19 , 51 ) EGFP+ cells/104 cells; n = 5 ) following re-stimulation with PHA and feeder PBMC ( Fig . 2 B ) or following direct activation with anti-CD3/CD28 , together with IL-7 and the integrase inhibitor L8 , which allowed the detection of post-integration latency ( Fig . 2 C ) . To further confirm that post-integration latency was established in resting CD4+ T cells co-cultured with mDC , we used a real time PCR assay to quantify integrated HIV DNA . Integrated HIV DNA was present in SNARFhiEGFP− CD4+ T cells sorted from the mDC co-cultures ( 1100 ( 686 , 4960 ) HIV DNA copies/106 cells , n = 3; Fig . 2 D ) , but not in the CD4+ T cells sorted from the pDC co-cultures or the CD4+ T cells cultured alone ( both <330 HIV DNA copies/106 cells ) . Similar results were observed with a nef competent EGFP-reporter virus ( Fig . 2 E ) , demonstrating that the establishment of mDC-induced latency was not dependent on Nef . Unlike experiments that utilised R5 EGFP HIV , when experiments were performed with an X4 EGFP reporter virus , latent infection was detected in the resting CD4+ T cells cultured alone ( 87 ( 51 , 155 ) EGFP+ cells/104 cells; n = 4; Fig . 2 F ) . However , latency was still significantly enhanced in the non-proliferating CD4+ T cells in the presence of mDC ( 468 ( 213 , 621 ) EGFP+ cells/104 cells ) when compared to the T cells cultured alone . In some experiments we added a low dose of Staphylococcus enterotoxin B ( SEB; 10 ng/mL ) to the mDC-T cell co-cultures to enhance productive infection and increase cognate interactions between mDC and T cells . In the presence of SEB we observed a significant increase in the level of productive infection; however , there was no difference in the level of latent infection ( Fig . 2 G ) . Finally , we cultured mDC and T cells together at ratios ranging from 1∶10 to 1∶100 to determine the minimum interaction required between mDC and T cells to induce latency . We found that latency could still be established at a ratio of DC: T cells as low as 1∶100 ( Fig . 2 H ) . Taken together , these results demonstrated that in vitro mDC and not pDC facilitated post-integration latency in non-proliferating CD4+ T cells . We have previously shown that memory CD4+ T cells and not naïve CD4+ T cells are susceptible to latent infection following chemokine exposure [14] . To determine whether mDC-induced T cell latency occurred in memory or naïve CD4+ T cells , we separated the SNARFhiEGFP− CD4+ T cells into CD45RO+ ( memory ) and CD45RO− ( naïve ) fractions prior to culture with feeder PBMC . In these experiments , latent infection was detected at significantly higher levels in the CD45RO+ memory CD4+ T cell fraction ( 146 ( 14 , 197 ) EGFP+ cells/104 cells; Fig . 2 I ) . A proportion of non-proliferating SNARFhiEGFP− cells sorted from DC-T cell co-cultures ( Fig . 1 F and 3 A ) expressed CD69 . Therefore , in order to exclude the possibility that we were only detecting infection of the cells showing early signs of activation , we depleted CD69+ cells from the SNARFhiEGFP− T cells at day 5 post-infection prior to co-culture with PHA and feeder PBMC . We found no significant difference in the level of latency following depletion of the CD69+ cells ( Fig . 3 B ) . CD69 expression can be transient , therefore , to confirm that we had not missed cells that expressed CD69 , which had then been down-regulated; we measured the expression of CD69 over time following co-culture with mDC and infection with HIV . We demonstrated that CD69 expression peaked at day 2 and remained elevated out to day 5 post-infection ( data not shown ) . These results demonstrate that the subpopulation of CD4+ T cells that were partially activated and expressing CD69 were not preferentially latently infected following mDC-T cell co-culture . To determine why mDC and not pDC led to the establishment of latency , we compared cytokine levels in pDC-T cell and mDC-T cell co-cultures 5 days following HIV infection using bead arrays for known DC-secreted cytokines . Supernatants collected from HIV-infected mDC-T cell co-cultures compared to the HIV-infected pDC-T cell co-cultures had significantly increased expression of IL-6 ( p = 0 . 002 ) , IL-10 ( p = 0 . 01 ) and CXCL9 ( p = 0 . 002; Fig . 4 A ) . TNF-alpha was also up-regulated in the mDC-T cell compared to pDC-T cell co-cultures but the difference was not statistically significant ( Fig . 4 A ) . As expected , IFN-alpha was detected at high levels in the pDC-T cell co-cultures but not in the mDC-T cell co-cultures ( Fig . 4 A; p = 0 . 01 ) , and latency was not established in pDC-T cell co-cultures even in the presence of neutralising antibodies to IFN-alpha ( Fig . 4 B ) . Interestingly , when equal numbers of pDC and mDC were added to resting CD4+ T cells latency was reduced when compared to T cells cultured only with mDC ( Fig . 4 C ) . This suggests that while pDC themselves do not induce latency , they are able to inhibit the establishment of latency mediated by mDC . To determine whether the soluble factors that were differentially expressed in mDC-T cell co-cultures compared to pDC-T cell co-cultures were contributing to the establishment of mDC-induced latency , neutralising antibodies ( nAb ) , to either the soluble factor or its receptor , were added to the T cells prior to co-culture with DC and the addition of HIV . Specific nAbs or an anti-IgG control were added to eFluor670 ( alternative proliferation dye to SNARF ) -labelled resting memory ( CD45RO+ ) CD4+ T cells prior to co-culture with mDC , and again following infection , and latency determined as described in Fig . 1 ( Fig . 4 D ) . When nAbs to IL-6 , the IL-10 receptor ( IL-10R ) or CXCR3 ( CXCL9 and CXCL10 receptor ) were added to the mDC-T cell co-cultures , no significant decrease in the number of latently infected CD4+ T cells was observed when compared to cultures where control anti-IgG was added . As we had previously shown that the chemokine CCL19 can condition resting CD4+ T cells allowing for enhanced entry and integration of HIV [14] , we also added anti-CCL19 , either alone or in combination with anti-CXCR3 , to the mDC-T cell co-cultures . However , we did not detect a significant decrease in the number of latently infected cells . The activity of these nAbs was confirmed by their ability to block STAT3 signalling ( aIL-6 , aIL-10R ) or chemokine induced migration ( aCXCR3 , aCCL19; Fig . S2 ) . To determine whether DC-T cell contact was required to establish latency in resting CD4+ T cells , we co-cultured mDC and resting CD4+ T cells separated by a 0 . 4 µm membrane transwell . Following 24 hours of culture , HIV was added to the mDC in the upper chamber and the CD4+ T cells in the lower chamber . Without mDC-T cell contact , the establishment of latency was significantly inhibited ( <1 EGFP+ cell/104 cells ) when compared to co-cultures without membranes ( 18 ( 10 , 85 ) EGFP+ cell/104 cells; n = 5; p = 0 . 03; Fig . 4 E ) . To further elucidate whether soluble factors other than those previously inhibited were involved in DC-induced latency , we added supernatant from infected mDC-T cell co-cultures to uninfected resting CD4+ T cells and then infected these cells with EGFP-HIV . Media changes were performed daily using supernatant from infected mDC-T cell co-cultures . Under these conditions the resting CD4+ T cells would be exposed to any soluble factors and free viral particles present in the mDC-T cell co-cultures but would not have any contact with the mDC . Latency was not detected in these cultures ( Fig . 4 F ) , providing further evidence that DC-T cell contact , and not a soluble factor , was required for the establishment of DC-induced latency . Direct DC-T cell signalling can occur following interactions between several cell surface receptors . In particular , interactions between lymphocyte function associated antigen-1 ( LFA-1; composed of CD11a and CD18 ) on T cells and intercellular adhesion molecule 1 ( ICAM-1 ) on DC are involved in DC-T cell adhesion [22] and subsequent T cell activation via formation of an immunological synapse [23] . To inhibit DC-T cell clustering , we used blocking antibodies to CD18 ( 10–20 µg/mL ) . Blocking of CD18 significantly inhibited , but did not eliminate , DC-T cell clustering , as observed by microscopy ( data not shown ) . Following incubation with anti-CD18 , there was no effect on the number of productively infected cells , however , we observed a significant decrease in the number of latently infected cells from the mDC-T cell co-cultures ( 20 ( 8 , 34 ) latently infected cells/104 cells ) , when compared to cells cultured without anti-CD18 ( 25 ( 17 , 48 ) latently infected cells/104 cells; p = 0 . 03 ) or in the presence of control anti-IgG ( 32 ( 22 , 36 ) latently infected cells/104 cells; n = 6; p = 0 . 03; Fig . 5 A ) . However , when resting CD4+ T cells were stimulated with soluble ICAM-1 and anti-IgG there was no increase in latency observed ( Fig . 5 B ) , suggesting that ICAM-LFA signalling alone does not induce latency in this model system . Furthermore , in the presence of anti-CD18 the number of latently infected cells from the mDC-T cell co-cultures remained greater than the CD4+ T cells cultured alone ( <1 latently infected cell/104 viable cells; p = 0 . 01 ) suggesting that the effect of anti-CD18 was most likely due to the partial decrease in clustering/DC-T cell contact . In order to determine whether mDC transfer of HIV was involved in the establishment of latency we performed experiments where we added virus to resting CD4+ T cells , washed off virus and added uninfected mDC to the CD4+ T cells ( Fig . 5 C ) . Under these conditions we were still able to detect latency in the non-proliferating CD4+ T cells . Together , these results indicate that cell-cell contact plays a role in DC-induced T cell latency but that the mDC were not required to be infected and then transfer HIV to the resting CD4+ T cells . To determine the effect of DC on gene transcription in latently infected resting CD4+ T cells , SNARF-labelled resting CD4+ T cells , from four independent donors , were cultured either alone or with syngeneic bulk blood DC at a 1∶10 ratio for 24 hours prior to infection with NL ( AD8 ) -nef/EGFP . In these experiments , we included IL-7 ( 10 ng/mL ) in all cultures to increase cell survival of the resting cells and infections were performed at an MOI of 5 to ensure a high frequency of latently infected cells . Mock infections were performed in parallel with media alone . Non-proliferating ( SNARFhi ) CD4+ T cells that were not productively infected ( EGFP− ) were sorted 5 days post-infection and lysed for either the detection of HIV DNA by real-time PCR or RNA for microarray studies ( Fig . 6 A ) . Infection was confirmed in the resting CD4+ T cells following co-culture with DC , in 4 independent experiments , by detection of HIV DNA ( 3×104 ( 7 . 4×103 , 5 . 7×105 ) copies/106 cells; Fig . 6 B ) . Changes in gene expression were quantified in the sorted SNARFhiEGFP− CD4+ T cells using Illumina oligonucleotide microarrays . To identify genes expressed in DC-induced latency , we compared the expression profiles of non-proliferating , latently infected CD4+ T cells ( HIV T ( +DC ) ) to mock infected CD4+ T cells ( Mock T ( +DC ) ) that had been co-cultured with DC . In order to control for the effect of virus or DC alone , we first subtracted the gene expression profiles of control cells , which were T cells that had been cultured alone that were either uninfected ( Mock T ) or exposed only to virus ( HIV T ) . A scatter plot ( Fig . 7 A ) , representing the common ( genes that fall on the diagonal ) and differentially expressed genes ( genes that fall off the diagonal ) from this comparison , highlighted the significant differences in gene expression between latently infected cells and controls ( r = 0 . 77 ) . Additionally , this plot showed that several of the genes that discriminate latently infected cells from control cells were genes downstream of type I interferons , including interferon-induced protein with tetratricopeptide repeats 1 ( IFIT-1 ) , interferon alpha-inducible protein 27 ( IFI27 ) , and 2′-5′-oligoadenylate synthetase 1 ( OAS1 ) . Heatmaps of the top 100 genes ( Fig . S3 ) confirmed the de novo induction of genes encompassing several biological and metabolic pathways in T cells exposed both to DC and virus . These included transcripts of the Interferon pathway , genes involved in the regulation of cell cycle entry and mitosis , as well as receptor and effector molecules of cell survival and apoptosis . Network analysis ( Fig . 7 B and C ) showed that two major pathways were regulated in T cells following co-culture with DC exposed to HIV . Exposure of CD4+ T cells to HIV and DC led to the up-regulation of genes downstream of type I interferon by nucleotide sensors . Figure 7 B confirms the wide-ranging impact of the up-regulation of the type I Interferon pathway , as several molecules with antiviral activity were up-regulated , including ISG15 ubiquitin-like modifier ( ISG15 ) , and DEAD box polypeptide 58 ( DDX58; also known as RIG-I ) . Genes involved in actin polymerisation , and the organisation of microtubules , were also up-regulated as a consequence of interferon pathway up-regulation . Additionally , the interferon pathway intercepted with the mammalian target of rapamycin complex 2 ( mTORC2 ) pathway , which plays an important role in autophagy and T cell survival [24] ( Fig . 7 B; Table S1 ) . Network analysis confirmed the negative impact of exposure of CD4+ T cells to DC and virus on the NF-κB pathway as well as several cellular metabolic pathways ( fatty oxidation and glucose metabolism ) regulated by peroxisome proliferator-activated receptor gamma ( PPARG; Fig . 7 C ) . Inhibition of the NF-κB transcriptional network , which plays a significant role in HIV transcription , led to the down-regulation of protein kinase C alpha ( PRKCA ) . This observation confirms the quiescence of these cells and may also be a step towards the induction of HIV latency . Triggering of a transcriptional program leading to T cell quiescence was confirmed by the increased expression of Kruppel-like factor 6 ( KLF6 ) , a gene with anti-proliferative functions [25] , [26] , as well as activating transcription factor 3 ( ATF3 ) that has been recently been shown to negatively regulate activating protein 1 ( AP-1 ) -mediated HIV transcription . Additionally , we observed a down-regulation of several molecules involved in DNA replication such as members of the minichromosome maintenance ( MCM ) complex ( MCM4 , MCM5 , and MCM10 ) and the aurora kinase ( Fig . 7 C ) . Pathways controlling pyrimidine and purine synthesis were also expressed at lower levels in cells exposed to virus and DC highlighting a reduced availability of nucleotides for cell division ( Table S1 ) . The down-regulation of NF-κB resulted in the decreased expression of several molecules that play a critical role in T cell survival , including CD27 ( TNFRSF7 ) , baculoviral IAP repeat containing 5 ( BIRC5/survivin ) and tumor necrosis factor receptor superfamily , member 6b ( TNFRSF6B/DCR3 ) , a decoy receptor that inhibits Fas ligand and LIGHT-mediated signalling [27] . In order to confirm the differential expression of genes in these different populations of cells , we used a highly quantitative PCR approach and showed a strong correlation between gene expression data measured by either gene array or PCR ( Fig . 7 D and Table S2 ) . Taken together , results of transcriptional profiling highlighted the impact of two major transcriptional nodes in the inhibition of viral replication and the induction of latency . The up-regulation of Type I Interferons and their downstream target genes could trigger several genes endowed with antiviral activities and would also impact cell proliferation , survival and metabolic processes . Concomitantly , the down-regulation of NF-κB will lead to T cell quiescence and decreased levels of activation , both of which are required for HIV replication .
The study of latently infected resting CD4+ T cells ex vivo from HIV-infected patients on ART is greatly limited by the low frequency of latently infected cells and the lack of a distinctive surface marker to distinguish latently infected from uninfected cells . Here we demonstrate that latency can be efficiently established via direct infection of non-proliferating CD4+ T cells in the presence of DC . Using primary blood DC and resting CD4+ T cells we have demonstrated that: [1] co-culture of resting memory CD4+ T cells with DC can establish latent infection; [2] mDC but not pDC mediate this effect; [3] close cell-cell proximity is required between DC and T cells; and [4] multiple cell cycle genes were altered in non-proliferating CD4+ T cells , containing latently infected cells . These novel findings provide a potential pathway for the establishment and maintenance of latent infection in resting CD4+ T cells that recapitulates the likely events within lymphoid tissues in HIV-infected patients in vivo . Previous studies have explored the ability of DC to enhance productive HIV infection within DC-CD4+ T cell co-cultures [28]–[31]; however , we are the first to present data demonstrating the ability of specific subpopulations of DC to induce latency in resting CD4+ T cells in these co-cultures . Using this model we clearly demonstrated that following co-culture of mDC with resting memory CD4+ T cells , post-integration latency was established . This was demonstrated by inducible virus ( established both in the presence and absence of indinavir ) and detectable integrated HIV DNA in T cells cultured with mDC but not those cultured alone following infection with an R5 EGFP reporter virus . While integrated R5 HIV DNA was only detected following co-culture with mDC in our model of latency , it was similar to that previously reported for resting CD4+ T cells infected in isolation with a wild type X4 NL4 . 3 virus [32] . Resting CD4+ T cells express very low levels of CCR5 in contrast to expressing very high levels of CXCR4 . Additionally , we saw significantly higher levels of latency ( ∼100 fold ) in our T cells cultured alone when we used an X4 EGFP reporter virus compared to an R5 EGFP virus . Unlike memory CD4+ T cells , we were unable to detect latency in naïve CD4+ T cells following mDC co-culture . It is possible that the differential establishment of latency in resting naïve and memory T cells was due to differences in their cortical actin density and actin dynamics as previously suggested by others [33] . While mDC induced T cell latency in this model , pDC did not . Interestingly , pDC played an active inhibitory role in establishing latency , when co-cultures were performed with equal numbers of pDC and mDC ( Fig . 4 C ) . One potential explanation for the difference between co-cultures of bulk DC and T cells ( where latency was established ) and DC-T cell co-cultures containing purified equal numbers of pDC and mDC ( where latency was inhibited ) could potentially be that the number of pDC present in the bulk DC ( roughly 1 pDC to 3 mDC ) was too low to inhibit latency . How pDC actively suppress the establishment of latency is unknown , but it does not appear to be mediated by IFN-alpha . Establishment of mDC-induced latency was not dependent on DC-T cell transfer of HIV , as latency was still detected when T cells were infected in isolation and uninfected mDC added only after virus had been washed off . Nor was it dependent on the amount of virus replication , because while only mDC were able to induce latent infection , similar levels of productive infection were observed in both the pDC and mDC co-cultured CD4+ T cells ( Fig . 2 A ) . Furthermore , we found that addition of SEB to the culture model enhanced productive infection but did not increase latent infection ( Fig . 2 G ) . Together , these data provide evidence that the establishment of latency in the non-proliferating CD4+ T cells when co-cultured with mDC was not simply due to higher viral exposure in these cultures . These results differ from a previous study that also looked at direct infection of resting CD4+ T cells , which concluded that DC had no effect on the integration levels of R5 or X4 virus in either naïve or memory CD4+ T cells [34] . However , in this study , although primary DC were used ( defined as BDCA-1+ and BDCA-4+ cells ) , total DC were present at a frequency of only 0 . 89% and therefore the frequency of mDC may have been too low to demonstrate an effect of mDC on the infection of resting CD4+ T cells . Additionally , contrary to our data , a recent study has reported the ability of monocyte derived DC ( MDDC ) to activate latent infection in T cells [35] . A key difference was that latency in this study was unusually established in proliferating CD4+ T cells and not non-proliferating T cells as in our study . Furthermore , MDDC , as opposed to primary DC , were utilised in this study . MDDC have multiple significant functional and lineage differences to primary DC as we have recently demonstrated using detailed sorting and gene expression analyses [36] . In this study we utilised total blood CD11c+ mDC , which consist of at least three different subsets , a major SLAN ( 6-sulfo LacNAc+ ) , an intermediate CD1c+ ( BDCA-1 ) and a minor CD141+ ( BDCA-3 ) population , each with different functional properties [37]–[40] . However , it is currently unclear whether one or more of these subsets is responsible for inducing latency in resting CD4+ T cells . We have previously demonstrated that multiple chemokines , including CCL19 , CXCL9 and CXCL10 , can condition resting CD4+ T cells allowing for the establishment of HIV latency [14] , [17] . However , blocking CCL19 and CXCR3 , the receptor for CXCL9 , 10 and 11 , had a minimal impact on DC-induced latency ( Fig . 4 D ) . While it is possible that there may be involvement of chemokines other than those inhibited , given that latency was not detected in resting CD4+ T cells infected in the presence of infected mDC-T cell culture supernatants , this is unlikely ( Fig . 4 F ) . Rather , our data supports an essential role for direct DC-T cell interactions or DC-T cell signalling as mDC-induced latency was prevented when the mDC were cultured in transwells above the resting CD4+ T cells ( Fig . 4 E ) . Unlike our previous work that was performed in the absence of productive infection [14] , latency following DC-T cell co-culture was established in the presence of productive infection , which may more accurately mimic the establishment of latency in acute infection in vivo . Therefore , in the presence of productive infection it is possible that there are alternative pathways that lead to the establishment and maintenance of latency in resting CD4+ T cells . Interactions between ICAM-1 , found on DC , and LFA-1 , found on T cells , strengthen DC-T cell adhesion and play a key role in the formation of the immunological synapse [41] . We have shown that clustering , facilitated by ICAM-1-LFA-1 interactions , contributed to DC-induced T cell latency , as latency was significantly reduced , but not eliminated , when blocking antibodies to CD18/LFA-1 were added to the DC-T cell co-cultures ( Fig . 5 A ) . However , interactions between ICAM-1 and LFA-1 alone were not sufficient to induce T cell latency in the absence of mDC ( Fig . 5 B ) , therefore , the reduction in latency observed in the presence of anti-CD18 was most likely due to the reduction in DC-T cell clustering rather than specific LFA-ICAM signalling events . As there are numerous other molecules involved in DC-T cell clustering , such as LFA-3 and CD2 , additional signalling pathways should be explored as potential mediators of DC-induced HIV latency . Interestingly , a recent paper has demonstrated enrichment of latency in CD2 expressing T cells from HIV-infected patients on ART [42] . Transcriptional profiling experiments served to highlight changes in cellular gene expression in resting non-proliferating CD4+ T cells that contained latently infected CD4+ T cells . We showed significant differences in gene expression between resting CD4+ T cells from HIV and mock infected DC-T cell cultures . However , it is important to note that , while all cells within our “latent” cell population were exposed to virus , only a proportion were actually infected ( median of 3% ) . It is possible that some of the observed differences in gene expression may be due to uninfected cells that were exposed to HIV but not infected . This may include the genes downstream of type I interferons as our in vitro experiments have shown that pDC , the major producers of type I interferons , were not involved in the induction of T cell latency . Therefore , it is possible that type I interferons are necessary but alone are not sufficient to induce HIV latency . We have demonstrated significant differential expression of genes involved in cell cycle , in particular those associated with cell cycle arrest ( Fig . 7 C ) . During DC-T cell interactions in the presence of HIV , differential expression of co-stimulatory and negative regulatory factors determines the fate of the interacting CD4+ T cell [43] . These interactions can result in active suppression of T cell cycle and as a result may inhibit post-integration steps in viral replication and promote the establishment of latency . Indeed , in HIV-infected patients on ART , HIV DNA is found at higher frequencies in CD4+ T cells expressing the negative regulator PD-1 [44] . Latency has also been shown to be triggered by the absence of certain transcriptional machinery in resting CD4+ T cells , such as NF-κB [45] and nuclear factor of activated T ( NFAT ) [11] , [46] . In DC-induced latently infected CD4+ T cells we observed the suppression of multiple genes associated with the activation of NF-κB ( Table S1 ) , including protein kinase C alpha , PRKCA , which also plays a role in the activation of NFAT [47] , [48] . Therefore , it is possible that the global suppression of genes associated with the activation of NF-κB and/or NFAT may also contribute to the maintenance of latency in DC- T cell co-cultures by preventing progression to productive infection in cells that contain integrated HIV . However , while this data provides insights into genes that may potentially be important for both the establishment and maintenance of latency , it will be important to conduct gene knockdown experiments within our model in order to determine the specific role of individual genes in establishing and maintaining mDC-induced T cell latency . In summary , this study has demonstrated a novel pathway for the establishment of latency in resting memory CD4+ T cells that was dependent on close proximity to mDC . Efficient infection of resting CD4+ T cells in close contact with mDC and HIV could explain the rapid early establishment of the latent HIV reservoir . Additionally , if infectious virus persists in tissues such as lymph node in patients on ART , mDC may facilitate ongoing infection of resting T cells leading to replenishment of the reservoir .
PBMC were isolated from buffy coats obtained from the Australian Red Cross Blood Service ( Melbourne , Australia ) . Resting CD4+ T cells were negatively selected using magnetic cell sorting and a cocktail of antibodies to CD8 , CD11b , CD16 , HLA-DR , CD19 and CD69 , as previously described [17] , [49] . Sorted cells were routinely negative for CD69 , CD25 and HLA-DR ( Fig . S1 A ) . In some experiments bulk resting CD4+ T cells were further sorted into CD45RA+ naïve and CD45RA− memory CD4+ T cells using phycoerythrin ( PE ) -labelled antibody to CD45RA and a FACSAria ( BD Biosciences ) . DC were isolated from blood as previously described [50] . Briefly , DC were enriched using magnetic bead depletion and antibodies to CD3 , CD11b and CD19 . Enriched cells were then sorted using a FACSAria ( BD Biosciences ) to obtain a bulk cocktail− HLA-DR+ DC population , HLA-DR+CD11c+ mDC or HLA-DR+CD123+ pDC . The purity of sorted cells was always >98% ( Fig . S1 B ) . In all experiments except where noted we used an NL4-3 virus with EGFP inserted into the nef open reading frame at amino acid position 75 at the aKpnI ( Acc651 ) site with a CCR5-tropic ( AD8 ) envelope ( NL ( AD8 ) -nef/EGFP ) , alternatively we used this virus with a CXCR4-tropic ( NL4-3 ) envelope ( NL4-3-nef/EGFP; both kindly provided by Damian Purcell , University of Melbourne , Melbourne , Australia ) . In one set of experiments we used a Nef-competent EGFP reporter virus , kindly provided by Yasuko Tsunetsugu-Yokota ( National Institute of Infectious Diseases , Tokyo , Japan ) [51] . HIV stocks were generated by FuGene ( Promega , Madison , WI ) transfection of 293T cells as previously described [49] , [50] . Cells were infected at 37°C for 2 hours at an MOI of 0 . 5 or 5 , as determined by limiting dilution using the Reed and Muench method [52] , followed by a wash step to remove unbound virus . Resting CD4+ T cells were labelled with proliferation dye , either SNARF ( 10 µM; Invitrogen ) or eFluor®670 ( 5 µM; eBiosciences , San Diego , CA ) , according to the manufacturer's instructions . SNARF/eFluor670-labelled resting CD4+ T cells were cultured in media supplemented with IL-2 ( 2 U/mL; Roche Diagnostics ) for 24 hours , with or without syngeneic bulk DC or sorted DC subsets ( DC: T cell ratio of 1∶10 ) , in the presence or absence of SEB ( 10 ng/mL; Sigma ) . Cells were then infected using an EGFP-reporter virus and cultured for a further 5 days ( Fig . 1 A ) . In some experiments , cells were cultured with and without the protease inhibitor Indinavir ( 0 . 1 µM final ) for 30 minutes at 37°C prior to infection . At day 5 post-infection , cells were analysed by flow cytometry for productive infection by detecting EGFP+ cells . Subsequently , the non-proliferating ( SNARFhi/eFluor670hi ) CD4+ T cells that were not productively infected ( EGFP− ) were sorted using a FACSAria ( BD Biosciences ) . In order to amplify any latent infection , the sorted CD4+ T cells were stimulated with PHA ( 10 ug/mL ) /IL-2 ( 10 U/mL ) in the presence of PBMC and cultured for a further 5 days . The number of EGFP+ cells following re-stimulation was used as a surrogate measure for the number of latently infected , non-proliferating CD4+ T cells in the original cultures ( Fig . 1 B ) . In some experiments , we also stimulated the sorted SNARFhi/eFluor670hiEGFP− CD4+ T cells directly with plate bound anti-CD3 ( Beckman Coulter; 5 µg/mL ) and soluble anti-CD28 ( BD; 5 µg/mL ) . Flat bottomed 96 well plates were coated with anti-CD3 ( 5 µg/mL ) for 3 hours at 37°C . Unbound antibody was then removed and 1×105 SNARFhi/eFluor670hiEGFP− CD4+ T cells were plated per well in 200 µL of media containing soluble anti-CD28 ( 5 µg/mL ) , IL-7 ( 50 ng/mL ) and the integrase inhibitor , L8 ( 1 µM final ) . The number of EGFP+ cells was determined following 72 hours of culture . As a control for the activity of the integrase inhibitor L8 , we cultured SEB-stimulated PBMC with and without L8 for 30 minutes prior to infection , and productive infection was determined at day 5 post-infection ( Fig . S2 ) . For phenotypic analysis of the CD4+ T cells before culture , we stained sorted resting CD4+ T cells with CD69-FITC , CD25-PE and HLA-DR-perCP ( BD Bioscience ) on ice for 25 minutes . To determine whether co-culture with DC had altered the activation state of the resting CD4+ T cells , in some experiments the sorted SNARFhiEGFP− CD4+ T cells were labelled with either CD69-FITC or HLA-DR-FITC ( BD Biosciences ) . Intracellular staining was also performed on the sorted SNARFhiEGFP− CD4+ T cells to detect expression of the cell cycle marker Ki67 . Cells were permeabilised with 500 µL of 1× FACS Permeabilising Buffer ( BD Biosciences ) in the dark at room temperature for 10 minutes , washed once with FACS wash and incubated with Ki67-FITC ( 5 µL/105 sorted CD4+ T cells; Dako ) for 45 minutes on ice . Following incubation , cells were washed twice with FACS wash and resuspended in 1% FACS fix . We performed analyses on a FACSCalibur ( BD Biosciences ) and results were analysed using Weasel software ( Walter and Elisa Hall Institute , Melbourne , Australia ) . Cytokine bead arrays ( eBioscience ) were used according to the manufacturer's directions to determine the concentration of IL-1-beta , IL-6 , IL-10 , IL-12p70 , TNF-alpha , CXCL9 and CXCL10 in the cell cultures . In some experiments , nAbs to CD18 ( 10 µg/mL ( prior to infection ) and 20 µg/mL ( post-infection ) ; clone 7E4; Beckman Coulter ) , CCL19 ( 25 µg/mL ) , CXCR3 ( 20 µg/ml ) , IFN-alpha ( 5 µg/mL ) or control IgG ( R&D Systems , Minneapolis , MN ) ; IL-6 or IgG1 ( 10 ug/mL; BioLegend , San Diego , CA ) ; IL-10R or IgG2 ( 10 ug/mL; Biolegend ) were used . In these experiments , both the DC and the resting CD4+ T cells were pre-incubated with nAbs for 15 minutes on ice prior to culture . The nAbs were added again to the co-cultures following infection . Neutralising activity of anti-CCL19 ( 25 µg/mL ) and anti-CXCR3 ( 20 µg/ml ) was confirmed using a chemokine-induced migration assay . Resting CD4+ T cells were added to the top chamber of a 3 µM pore transwell migration plate ( Sigma ) and either CCL19 ( 100 nM ) or CXCL10 ( 300 nM ) was added to the bottom chamber . In experiments using anti-CXCR3 , cells were treated with nAb for 15 minutes at 37°C and washed off prior to chemokine treatment . In comparison , in experiments using anti-CCL19 , nAb was added together with chemokine to the bottom chamber . Migrated cells in the bottom chamber were then counted in duplicate at 20 hours post addition of chemokine . Anti-IL-6 and anti-IL-10R were used at neutralising concentrations previously described [53] , [54] . As positive controls for these nAbs we demonstrated that 10 µg/mL of anti-IL-6 and anti-IL10R or 5 µg/mL of anti-IFN-alpha efficiently blocked IL-6 ( 100 ng/mL ) , IL-10 ( 50 ng/mL ) or IFN-alpha ( 50 ng/mL ) mediated STAT3 phosphorylation respectively ( Fig . S2 ) . In order to determine the role of ICAM-1 and LFA-1 interactions in mDC-induced latency , resting CD4+ T cells were cultured alone or with 10 ug/mL of ICAM-1fc together with 6 µg/mL of anti-IgG-fc ( both from R&D Systems ) for 24 hours prior to infection and maintained post-infection . DC were cultured with resting CD4+ T cells in the presence and absence of 0 . 4 µm cell culture inserts ( BD , Franklin Lakes , NJ ) with DC in the top chamber and resting CD4+ T cells in the lower chamber . Following 24 hours of culture , both the DC and the CD4+ T cells were infected as described above . In other experiments , we added supernatant from infected mDC-T cell co-cultures to uninfected resting CD4+ T cells and then infected these cells . Media changes were performed daily using supernatant from infected mDC-T cell co-cultures . To determine the role of DC-T cell transfer , resting CD4+ T cells were infected in the absence of mDC and uninfected mDC were added back to the T cells only after virus had been washed off . SNARFhiEGFP− CD4+ T cells cultured with DC , in the presence ( latently infected ) or absence ( mock-infected ) of HIV , were sorted 5 days following infection with NL ( AD8 ) -nef/EGFP . In these experiments , all culture media was supplemented with 10 ng/mL of IL-7 ( Sigma ) instead of IL-2 , in order to increase cell survival of resting cells , and infections were performed at an MOI of 5 to ensure high numbers of latently infected cells . Microarrays were performed as previously described [55] . Briefly , cells were lysed and RNA extracted ( Qiagen , Valencia , CA ) , amplified ( Ambion Applied Biosystems , Austin , TX ) and hybridised to an Illumina Human-Ref8 ( v3 ) BeadChip ( Illumina , San Diego , CA ) . Beadchips were scanned using an Illumina BeadStation 500GX scanner and Illumina BeadStudio ( version 3 ) software ( Ilumina ) . Illumina probe data was exported from BeadStudio as raw data and screened for quality . Samples failing chip visual inspection and control examination were removed . Gene expression data was analysed using Bioconductor ( http://bioconductor . org/ ) [56] , an open-source software library for the analyses of genomic data based on R , a language and environment for statistical computing and graphics ( www . r-project . org ) . The R software package was used for pre-processing , first to filter out genes with intensities below background in all samples , then to minimum-replace ( a surrogate-replacement policy ) values below background using the mean background value of the built-in Illumina probe controls as an alternative to background subtraction ( which may introduce negative values ) to reduce “over inflated” expression ratios determined in subsequent steps , and finally quantile-normalise the gene probes intensities . Genes were then filtered by intensity and by variance filters to allow a reduction in the number of tests and a corresponding increase in power of the differential gene expression analysis . The resulting matrix showing filtered genes as rows and samples as columns was log2 transformed and used as input for linear modelling using Bioconductor's limma package , which estimates the fold-change between two predefined groups by fitting a linear model and using an empirical Bayes method to moderate standard errors of the estimated log-fold changes for expression values from each gene . P values from the resulting comparison were adjusted for multiple testing according to the method of Benjamini and Hochberg [57] . This method controls the false discovery rate , which was set to 0 . 05 in this analysis . Microarray data is available through the National Center for Biotechnology Information Gene Expression Omnibus ( GEO ) , series accession number pending . Ingenuity Pathway Analysis ( IPA ) software ( Ingenuity Systems , www . ingenuity . com ) was used to identify canonical signalling pathways and networks associated with the expression profiles of the non-proliferating CD4+ T cells cultured with DC in the presence ( HIV T ( +DC ) ) or absence ( Mock T ( +DC ) ) of HIV . Differentially expressed Illumina Probe IDs were imported into the Ingenuity software and mapped to the Gene Symbol from Ingenuity knowledge database . The significance of the association between the dataset and the canonical pathway was measured in two ways: 1 ) A ratio of the number of genes from the dataset that map to the pathway divided by the total number of genes that map to the canonical pathway; 2 ) Over-representation Fisher's exact test was used to calculate a p-value determining the probability that the association between the genes in the dataset and the canonical pathway is explained by chance alone . The pathways were ranked by −log p-value . This score was used as the cut-off for identifying significant canonical pathways ( p value<0 . 05 ) . IPA's networks are built from direct or indirect physical , transcriptional , and enzymatic interactions between the mapped genes ( focus genes ) . Two genes are considered to be connected if there is a path in the network between them . Ingenuity's approach is based on a multi-stage , heuristic algorithm that iteratively constructs networks that greedily optimize for both interconnectivity and number of Focus Genes under the constraint of a maximal network size . Each individual IPA network has a maximum of 35 focus genes and is assigned a significance score ( based on P value ) representing the likelihood that the focus genes within the network are found there by random chance . As previously described , full length viral DNA was quantified using primers specific for the HIV-1 long terminal repeat ( LTR ) and Gag [58] , and integrated HIV-1 DNA was quantified using a nested Alu-LTR real-time PCR [15] , [59] , [60] . Results were normalised for total input DNA as determined by real-time PCR for the CCR5 gene [61] . The correlation between gene arrays and real-time PCR was performed using a Spearman correlation test . Microarray expression data were validated in two donors by reverse transcriptase real-time PCR ( RT-qPCR ) , as previously described [62] . Briefly , SNARFhiEGFP− CD4+ T cells were lysed for RNA extraction and DNAse treatment ( Qiagen , RNAeasy mini kit ) . cDNA was generated using CellsDirect qRT-PCR mix ( Invitrogen ) . After reverse transcription all target genes were pre-amplified ( 18 cycles ) using Taqman primers ( Roche Probe library ) specific for the transcripts of interest , which were also used for quantification . qPCR were performed on a Roche Light Cycler 348II and analysed according to the ΔΔct method . In all experiments , Wilcoxon signed-rank or student paired t tests ( for n<5 ) were performed for comparisons between populations using Graphpad Prism 5 . 0 software . P values of less than 0 . 05 were considered significant . Statistical analyses for microarray data were performed with program R , according to the method of Benjamini and Hochberg [57] . | Current antiretroviral drugs significantly prolong life and reduce morbidity but are unable to cure HIV . While on treatment , the virus is able to hide in resting memory T cells in a silent or “latent” form . These latently infected cells are rare and thus are hard to study using blood from HIV-infected individuals on treatment . Therefore , it is very important to have laboratory models that can closely mimic what is going on in the body . We have developed a novel model of HIV latency in the laboratory . Using this model we have shown that the presence of dendritic cells , an important type of immune cell that can regulate T cell activation , at the time of infection allows for the infection of resting T cells and the establishment of latency . We have demonstrated that this is predominantly mediated by direct cell-to-cell interactions . Further exploration of the mechanisms behind HIV latency could lead to new ways to treat and possibly eradicate HIV . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | Myeloid Dendritic Cells Induce HIV-1 Latency in Non-proliferating CD4+ T Cells |
The essential role of the lipophosphoglycan ( LPG ) of Leishmania in innate immune response has been extensively reported . However , information about the role of the LPG-related glycoinositolphospholipids ( GIPLs ) is limited , especially with respect to the New World species of Leishmania . GIPLs are low molecular weight molecules covering the parasite surface and are similar to LPG in sharing a common lipid backbone and a glycan motif containing up to 7 sugars . Critical aspects of their structure and functions are still obscure in the interaction with the vertebrate host . In this study , we evaluated the role of those molecules in two medically important South American species Leishmania infantum and L . braziliensis , causative agents of visceral ( VL ) and cutaneous Leishmaniasis ( CL ) , respectively . GIPLs derived from both species did not induce NO or TNF-α production by non-primed murine macrophages . Additionally , primed macrophages from mice ( BALB/c , C57BL/6 , TLR2−/− and TLR4−/− ) exposed to GIPLs from both species , with exception to TNF-α , did not produce any of the cytokines analyzed ( IL1-β , IL-2 , IL-4 , IL-5 , IL-10 , IL-12p40 , IFN-γ ) or p38 activation . GIPLs induced the production of TNF-α and NO by C57BL/6 mice , primarily via TLR4 . Pre incubation of macrophages with GIPLs reduced significantly the amount of NO and IL-12 in the presence of IFN-γ or lipopolysaccharide ( LPS ) , which was more pronounced with L . braziliensis GIPLs . This inhibition was reversed after PI-specific phospholipase C treatment . A structural analysis of the GIPLs showed that L . infantum has manose rich GIPLs , suggestive of type I and Hybrid GIPLs while L . braziliensis has galactose rich GIPLs , suggestive of Type II GIPLs . In conclusion , there are major differences in the structure and composition of GIPLs from L . braziliensis and L . infantum . Also , GIPLs are important inhibitory molecules during the interaction with macrophages .
In the Americas , Leishmaniases are widely distributed from the southern United States to northern parts of Argentina [1] . In Latin America , especially in Brazil , Leishmania braziliensis and Leishmania infantum are the causative agents of cutaneous ( CL ) and visceral leishmaniasis ( VL ) , respectively . The severity of the disease may range from self-healing cutaneous ulcers to potentially lethal visceral form [2] . During the life cycle , Leishmania parasites have to survive to extreme adverse conditions in both vertebrate and invertebrate hosts [3] . In the vertebrate host , inoculation of metacyclic Leishmania promastigotes by the sand fly is followed by neutrophil phagocytosis prior to intracellular differentiation into amastigotes [4] . At the early steps of infection , innate cellular microbicidal mechanisms may include the production of reactive nitrogen intermediates ( RNI ) , reactive oxygen intermediates ( ROI ) and cytokines ( IL-12 , TNF-α and IFN-γ ) [5] , [6] . This is crucial for Th1 polarization and subsequent parasite control in the mouse model . Failure in this process can lead to higher parasite burden and increase severity of disease [7] . To avoid destruction , intracellular parasites must interfere with the cytocidal signaling system of the host . In vivo and in vitro studies have demonstrated the importance of nitric oxide ( NO ) production in response to several stimuli such as bacterial lipopolysaccharide ( LPS ) , IFN-γ and TNF-α [8] . It is known that Leishmania-infected macrophages fail to activate MAPKs , become less responsive to cytokine stimulation ( IL-12 and IFN-γ ) [9] , [10] , [11] and express lower amounts of iNOS and IL-12 [12] , [13] , impairing T CD4+ cell differentiation to a TH1 phenotype . The molecular mechanisms involved in the immune system modulation by Leishmania have been the focus of many studies . GPI-anchored molecules are closely associated with cell signaling and can act as agonists and second messengers in response to cytokines and other stimuli [9] , [14] , [15] , [16] . The most studied Leishmania glycoconjugate is lipophosphoglycan ( LPG ) , whose functions include: attachment and entry into macrophages [17] , modulation of NO production [18] , inhibition of protein kinase C ( PKC ) dependent cell activation [19] , [20] , retardation of phagosome maturation [21] , disruption of NADPH oxidase assembly at the phagosome membrane [22] , induction of neutrophil extracellular traps ( NETs ) [23] , induction of protein kinase R ( PKR ) [24] , and attachment to the sand fly vector midgut [25] . In Leishmania , Toll-like receptor 2 ( TLR2 ) is the main receptor for both LPG and glycoinositolphospholipids ( GIPLs ) , the latter as a less potent agonist [26] , [27] . Besides TLR2 , in vivo studies have also demonstrated the importance of TLR4 and TLR9 during Leishmania infection [28] , [29] , [30] . Little is known about the functions of GIPLs in Leishmania biology , although they are present as the major component of the parasite surface in numbers greater than LPG [31] . The basic GIPL structure is a Manα1-4GlcN linked to an alkyl-acylglycerol through a phosphatidylinositol ( PI ) residue . Polymorphism in this family of molecules relies on the variety of fatty acid substitutions in the lipid anchor and monosaccharide substitutions in the glycan core moiety , leading to their classification into three groups ( Figure 1 ) : Type-I GIPLs are characterized by having an α1 , 6-mannose residue linked to the Manα1-4GlcN motif . This group is represented by M2 and M3 GIPLs which structures are Manα1-6Manα1-4GlcN-PI and Manα1-2 Manα1-6Manα1-4GlcN-PI . Type I GIPLs are closely related to GPI anchors of proteins with a very homogeneous lipid composition , predominantly C18∶0 fatty acids , and are found in Old World species such as L . donovani , L . tropica and L . aethiopica promastigotes [32] . Type-II GIPLs have a much more heterogeneous lipid composition with C18∶0 , C22∶0 , C24∶0 and C26∶0 fatty acids . They can be found in Old World L . major [33] , [34] and New World L . mexicana [35] , [36] and L . panamensis [36] . Type II GIPLs are characterized by having an α1 , 3-mannose residue linked to the Manα1-4GlcN motif , similarly to the glycan core of LPG . Structurally , they can range from small iM2 GIPL , Manα1-3Manα1-4GlcN-PI , to longer structures like GIPL-A , Galfβ1-3Galα1-3Galfβ1-3Manα1-3Manα1-4GlcN-PI and GIPL-3 , Galα1-6Galα1-3Galfβ1-3Manα1-3Manα1-4GlcN-PI . The third group is the Hybrid-type GIPLs , sharing common features to both Type-I and II with mannose residues located on both C-3 and C-6 positions of the Manα1-4GlcN motif ( isoM3 and isoM4 ) . There may be also other substitutions like phosphate sugars and ethanolamine residues [35] , [37] . Early studies have shown that GIPLs from L . major were highly antigenic , being recognized by sera from chronic CL patients [38] . Recent findings have demonstrated that L . braziliensis GIPLs are components of complex membrane microdomains and that these structures were crucial for parasite infectivity and survival [39] . However , little is known about the role of GIPLs in the innate immune compartment , especially in L . braziliensis and L . infantum . This work is part of a wider study on the glycobiology of New World species of Leishmania . In previous studies , we reported on the LPGs of L . braziliensis and L . infantum [40] , [41] and showed that the differences in LPG structures were relevant in the parasite biology . In this study , we expanded those findings and show the GIPL structures of the two New World Leishmanias also differentially modulate the innate immune system in mouse peritoneal macrophages .
World Health Reference strains of L . braziliensis ( MHOM/BR/1975/M2903 ) , L . infantum ( MHOM/BR/1974/PP75 ) and L . donovani ( MHOM/SD/00/1S-2D ) were used . Promastigotes were cultured in M199 medium supplemented with 10% heat-inactivated fetal bovine serum ( FBS ) , penicillin 100 units/ml , streptomycin 50 µg/ml , 12 . 5 mM glutamine , 0 . 1 M adenine , 0 . 0005% hemin , and 40 mM Hepes , pH 7 . 4 at 26°C [40] . Cells were harvested and washed in PBS twice prior to GIPLs extraction with methanol∶chloroform∶water ( 10∶10∶3 ) . This material was dried under nitrogen stream , resuspended on 0 . 1 M ammonium acetate buffer containing 5% 1-propanol and loaded onto an octyl-sepharose column ( 80 ml ) equilibrated in the same buffer . The column was subjected to a gradient of 1-propanol in 0 . 1 M ammonium acetate buffer ( 5–60% ) . Three mL fractions were collected and the presence of GIPLs in the fractions was detected by staining aliquots of the fractions on a TLC plate with orcinol∶sulfuric acid ( 100°C , 5 min ) [34] . GIPLs containing fractions were pooled , dried and resuspended in endotoxin-free water ( Sanobiol , São Paulo , Brazil ) . GIPLs concentrations determined as described elsewhere [42] . Prior to use on in vitro macrophage cultures , GIPLs were diluted in fresh RPMI . Thioglycollate-elicited peritoneal macrophages were removed from BALB/c , C57BL/6 and respective TLR2−/− and TLR4−/− knockouts by peritoneal washing with RPMI and enriched by plastic adherence for 18 h . Cells ( 3×105 cells/well ) were cultured in RPMI , 2 mM glutamine , 50 U/ml of penicillin and 50 µg/mL streptomycin in 96-well culture plates ( 37°C/5% CO2 ) . They were incubated with gamma interferon ( IFN-γ ) ( 100 IU/mL ) [43] , live stationary Leishmania parasites ( 10∶1 ) , GIPLs ( 1 , 5 , 10 and 25 µg/mL ) and lipopolysaccharide ( LPS ) ( 100 ng/mL ) . For CBA multiplex cytokine detection , cells were plated as described above for 1 h before washing with RMPI without serum . RPMI supplemented with 10% FBS was added with ( for primed macrophages ) or without ( for non-primed macrophages ) the addition of IFN-γ ( 3 IU/mL ) [44] and incubated for 18 h ( 37°C , 5% CO2 ) . GIPLs ( 25 µg/mL ) and LPS ( 100 ng/mL ) were added and incubated for 48 h . Supernatants were collected and stored at −70°C and cytokines ( IL1-β , IL-2 , IL-4 , IL-5 , IL-10 , IL-12p40 , IFN-γ and TNF-α ) were determined using the BD CBA Mouse Cytokine assay kits according to the manufacturer's specifications ( BD Biosciences , CA , USA ) . Flow cytometric measurements were performed on a FACS Calibur flow cytometer ( Becton Dickinson , Mountain View , CA ) . Cell-Quest™ software package provided by the manufacturer was used for data acquisition and the FlowJo software 7 . 6 . 4 ( Tree Star Inc . , Ashland , OR , USA ) was used for data analysis . A total of 1 , 800 events were acquired for each preparation . Results are representative of two experiments in duplicate . For inhibition studies , cell suspensions were washed with RPMI and enriched by plastic adherence for 18 h as described above without the addition of IFN-γ . Cells were pre-incubated with GIPLs ( 15 min ) prior to stimulation with LPS or IFN-γ . Supernatants were collected after 24 h for NO , TNF-α and IL-12 measurements . When used , LPS or IFN-γ were added 15 min after the addition of GIPLs . Culture supernatants were collected and nitrite concentrations determined by Griess reaction [45] and TNF-α and IL-12 concentrations were determined using ELISA ( BD ) . Results are representative of two experiments in triplicate . To evaluate whether intact GIPL structure is required for activity . Purified GIPLs were ressuspended in 150 µl CHAPS buffer ( 298 mg HEPES , 47 mg EDTA and 50 mg CHAPS in 50 ml endotoxin-free water ) and 2 U of PI-PLC ( Sigma ) ( 37°C , 16 h ) . Peritoneal macrophages were plated and stimulated with intact and PI-PLC treated GIPLs as described above . Nitrite content was measured on the supernatants by Griess reaction [45] . Stimulated cells ( 3×106/sample ) were washed with ice-cold PBS , lysed in lysis buffer ( 20 mM Tris-HCl pH 7 . 5 , 1% Triton X-100 , 1 mM sodium orthovanadate , 1 mM phenylmethylsulfonyl fluoride ( PMSF ) , 50 mM sodium fluoride , 150 mM NaCl , 5 mM ethylenediamine tetraacetic acid ( EDTA ) , 10% Glycerol ( v/v ) , 0 . 5 mM dithiothreitol ( DTT ) and protease inhibitor cocktail from Sigma® ) . Cells were harvested with a plastic scraper and centrifuged at 13 , 000× g ( 4°C , 10 min ) . Supernatants were transferred to fresh tubes and stored at −20°C until used . Cell lysates were resolved by SDS-PAGE , transferred to a nitrocellulose membrane and blocked ( 5% milk in TBS-0 . 1% Tween 20 ) for 1 h . Primary Abs ( anti dually phosphorylated ERK , dually phosphorylated p38 and Total ERK , 1∶1 , 000 ) were incubated for 16 h at 4°C . Membranes were washed ( 3×10 min ) with TBS-0 . 1% Tween 20 and incubated 1 h with anti-mouse IgG conjugated with peroxidase ( 1∶10 , 000 ) . The reaction was visualized using luminol . Purified GIPLs were delipidated by nitrous acid deamination ( 300 µl of 0 . 5 M sodium acetate and 300 µl of 0 . 5 M NaNO2 ) for 16 h at 37°C [40] . Samples were dried , resuspended in 0 . 1N HAc/01M HCl and applied to a phenyl-sepharose column ( 1 mL ) . The sugar headgroups were eluted using 0 . 1N HAc/0 . 1M HCl . After washing column with 2 volumes of water , lipids and unreacted GIPLs were eluted using Solvent E ( H2O/ethanol/diethyl ether/pyridine/NH4OH; 15∶15∶5∶1∶0 . 017 ) [46] . To desalt , deaminated GIPLs glycan headgroups were applied to Sephadex G-25 ( 1×5 cm ) columns equilibrated with 10 ml of water . Eluted deaminated glycan headgroups were collected in 0 . 5 ml fractions , checked for the presence of salt using silver nitrate and dried in Speed-Vac [40] . To obtain depolymerized neutral monosaccharides , deaminated glycan headgroups were subjected to strong acid hydrolysis ( 2N trifluoracetic acid , 3 h , 100°C ) and dried in Speed-Vac . To remove acid , 500 µl of toluene were added to samples , homogenized using vortex and evaporated twice under N2 . Samples were resuspended in 500 µl of water and desalted by ion exchanging chromatography . To remove salt from neutral monosaccharides , dried depolymerized neutral monosaccharides were diluted in 500 µl of H2O and applied onto a column containing AG1-X8 acetate form over AG50W-X12 resins . Samples were eluted with 5 mL of water and dried in a Speed-Vac instrument [47] . Intact and deaminated GIPLs were chromatographed on TLC Silica Gel 60 plates ( Merck ) . To compare rough GIPL content of L . braziliensis , L . infantum and as reference L . donovani . Intact GIPLs were chromatographed in 1-butanol∶methanol∶water ( 4∶4∶3 v/v ) for 20 h . To access Deamination by nitrous acid sensitivity , GIPLs were subjected to nitrous acid deamination as described above and resolved in chloroform∶methanol∶13M ammonium hydroxide∶1M ammonium acetate∶water ( 180∶140∶9∶9∶23 v/v ) for 20 h . Bands were visualized as described above [46] , [48] . To access the oligosaccharide composition , deaminated GIPLs headgroups were fluorescently labeled with 0 . 05 N ANTS ( 8-aminonaphthalene-1 , 3 , 6-trisulfate ) and 1 M cyanoborohydride ( 37°C , 16 h ) . To determine the monosaccharide composition of the GIPLs , depolymerized and desalted monosaccharides were fluorescently labeled with 0 . 1 M AMAC ( 2-aminoacridone ) in 5% acetic acid and 1 M cyanoborohydride . Labeled sugars were subjected to FACE and the gel was visualized under UV light . Oligoglucose ladders ( G1–G7 ) and monosaccharides ( D-galactose , D-glucose and D-mannose ) ( Sigma ) were used as standards for oligosaccharides and monosaccharide gels , respectively [47] , [49] . Desalted monosaccharides were separated using a DX-500 HPLC ( Dionex Corp . ) with ED40 electrochemical detection . Samples were run on a CarboPac PA10 column ( 4×250 mm ) in the presence of 18 mM NaOH ( flow rate 1 mL/min , 2000 psi ) . D-galactose , D-glucose and D-mannose ( 100 µg/mL ) were used as standards . For nitrite and cytokine measurements , the Shapiro–Wilk test was conducted to test the null hypothesis that data were sampled from a Gaussian distribution [50] . The P value ( P>0 . 05 ) showed that data did not deviate from Gaussian distribution . For this reason , student's “t” test and ANOVA were performed to test equality of population medians among groups and independent samples . Data were analysed using GraphPad Prism 5 . 0 software ( Graph Prism Inc . , San Diego , CA ) and P<0 . 05 was considered significant .
To determine whether GIPLs from both L . braziliensis and L . infantum are able to induce the production of nitrite , peritoneal macrophages were incubated with live promastigotes ( 10∶1 ) or treated with different concentrations of GIPLs ( 1 to 25 µg/mL ) with IFN-γ serving as positive control ( 100 IU/mL ) . Neither of the purified GIPLs could induce any detectable increase in the production of nitric oxide ( NO ) in primed BALB/c macrophages ( Figure 2 ) nor the production of the cytokines tested ( IL1-β , IL-2 , IL-4 , IL-5 , IL-10 , IL-12p40 , IFN-γ and TNF-α ) in non-primed macrophages in all other mice lineages ( data not shown ) . No NO production was detected in non-primed macrophages of BALB/c , C57BL/6 , TLR2−/− and TLR4−/− mice ( data not shown ) and in BALB/c primed macrophages ( Figure 3A ) . A higher NO production was detected on C57BL/6 IFN-γ-primed macrophages stimulated with GIPLs and live promastigostes when compared to BALB/c mice ( P<0 . 001 ) . There was a significant NO production in primed C57BL/6 and TLR2 ( −/− ) macrophages stimulated with GIPLs in comparison to TLR4 ( −/− ) ( P<0 . 01 ) ( Figure 3A ) suggesting the involvement of TLR4 in this activation . Also , a slight reduction of NO production was noticed in macrophages from TLR2 ( −/− ) mice stimulated with live promastigotes when compared to C57BL/6 ( P<0 . 04 ) . This reduction may indicate the participation of other parasite molecules that are recognized by TLR2 such as the LPG . The LPG is known to be a potent agonist of TLR2 and is capable of inducing the production of cytokines ( IL-12 , IFN-γ and TNF-α ) in macrophages and NK cells [26] , [27] . Differently from NO , TNF-α production was higher in BALB/c mice than in C57BL/6 ( P<0 . 05 ) in response to the stimulation of GIPLs from both species . Similarly this production was higher in TLR2 ( −/− ) than TLR4 ( −/− ) ( P<0 . 02 ) . This data also indicate a slight TLR4 involvement in TNF-α production . In both WT macrophages , the TNF-α production was higher after stimulation with GIPLs in comparison to live promastigotes ( Figure 3B ) ( P<0 . 01 ) . A lower TNF-α production was noticed in TLR2 ( −/− ) suggesting the involvement of TLR2 in this process . GIPLs did not induce the production of any of the cytokines tested ( IL1-β , IL-2 , IL-4 , IL-5 , IL-10 , IL-12p40 and IFN-γ ) in BALB/c , C57BL/6 , TLR2 ( −/− ) and TLR4 ( −/− ) mice ( data not shown ) . In all experiments , live parasites from both species induced cytokine production close to background levels ( Figure 3B and data not shown ) . These results suggest that GIPLs are able to activate NO in C57BL/6 mice and TNF-α in either BALB/c or C57BL/6 during the early steps of infection , and were not able to stimulate most of the cytokines assayed . Compared to LPG , GIPLs had a less potent agonistic activity to stimulate nitrite and cytokine production in previous studies [27] . To test if this pattern was due to inhibition and/or lack of activation , thioglycollate elicited peritoneal macrophages were pre-incubated with GIPLs prior to stimulation with IFN-γ or LPS . A strong inhibition ( aprox . 42% ) of NO production stimulated by IFN-γ was observed for L . infantum GIPLs and was almost completely abolished for L . braziliensis ( P<0 . 01 ) ( Figure 4A ) . A similar response was observed for LPS and this inhibition was more pronounced in L . braziliensis ( P<0 . 001 ) ( Figure 4B ) . Pre-incubation with GIPLs was also able to inhibit approximately 65% of IL-12 , but not TNF-α production ( Figures 4C and D ) . These results indicate an inhibitory role of GIPLs . Also , to test whether the intact structure of GIPLs is required for its inhibitory activity Macrophages were incubated with intact and PI-PLC treated GIPLs . As shown on Figure 5 PI-PLC treated GIPLs failed to inhibit NO production by IFN-γ stimulated cells . Since GIPLs were strong inhibitors of cytokine production , we investigated whether those molecules could modulate MAPKs activation . Mouse peritoneal macrophages were previously incubated with GIPLs and MAPK activation was detected using western blot . No significant activation of p38 and only a minimal induction of ERK were observed . Also when cells were preincubated with GIPLs prior to stimulation with LPS , there was a reduction on the phosphorylation of both ERK and p38 ( Figure 6 ) . Densitometer analysis normalized by total-ERK expression detected an 18% and 17 . 5% decrease on ERK activation for L . braziliensis and L . infantum , respectively . For p38 this inhibition was 16 . 5% and 33% , respectively . Due to the interspecific differences in the intensity of NO and IL-12 production inhibition ( Figures 4 ) and MAPKs activation ( Figure 6 ) , we examined whether those variations could be due to polymorphisms in GIPLs structure and composition . Intact GIPLs were resolved on TLC plates and the GIPL profile differed between the two species ( Figure 7A ) . Leishmania braziliensis exhibited slower migrating GIPLs compared to L . infantum , whose profile was very similar to L . donovani [32] with three main bands co-migrating with isoM2 , isoM3 and isoM4 . In L . braziliensis , the three faster bands co-migrated with bands isoM2 , isoM3 and isoM4 of L . donovani . All bands were susceptible to nitrous acid deamination , and this is consistent with the presence in the GIPLs of a non-N-substituted glucosamine residue ( Figure 7B ) , a hallmark of Leishmania GIPLs anchors [51] . To better determine sizes of the glycan portions , purified GIPLs were deaminated and desalted . The carbohydrate portions were reductively labeled with a fluorphore and then subjected to FACE . Consistent with the TLC data ( Figure 7 ) , the carbohydrate portions of the GIPLs from L . braziliensis were larger exhibiting up to 8–9 sugars while those from L . infantum and L . donovani consisted of up to 4–5 sugars ( Figure 8 ) . To access sugar composition , GIPLs were subjected to strong acid hydrolysis and the resulting monosaccharides were analysed by FACE and HPLC ( Figure 9A and B ) . Consistent with the TLC data ( Figure 7A ) , the monosaccharide composition of L . infantum GIPLs was very similar to the GIPLs from L . donovani ( Figure 9A ) . The relative amounts of galactose , glucose and mannose ( calculated by the relative peak areas on HPLC ) were determined ( Figure 9B ) . Supporting our other findings and GIPL assignments , the GIPLs from L . infantum had higher concentrations of mannose ( 82% ) , followed by galactose ( 12% ) and glucose ( 6% ) . This indicates that these are mostly Type I or hybrid GIPLs , whose structure bears a terminal mannose , but a small proportion of Type II GIPLs ( terminated in galactose ) is probably present . On the other hand , L . braziliensis GIPLs had higher galactose content ( 42% ) , followed by , mannose ( 30% ) and glucose ( 28% ) , thus suggesting a Type II GIPL structure .
Infection with protozoan parasites remains a prominent problem in different parts of the world having a major impact on public health in the developing countries . Leishmaniases are considered by World Health Organization [52] as one of the major six important infectious diseases worldwide . This class of parasitic diseases currently affects over 12 million people all around the world , up to 1 . 5 million new individuals developing the visceral and tegumentar disease respectively each year . In Brazil , most of those cases are caused by L . infantum and L . braziliensis , respectively . The question of how parasites interact with hosts cells to promote infection and survival has been the focus of interest for a long time . In order to survive in the macrophage cells , Leishmania has to prevent or inhibit a variety of intracellular mechanisms of parasite killing , one of which is dependent on ROS and RNI [53] , [54] . However , RNI alone is effective for controlling visceral Leishmaniasis [55] . Parasite surface molecules , especially the LPG , have long been known to play an important role in the host parasite interactions [17] , [27] , [56] . In this work , we focused on another class of glycoconjugates , the GIPLs in two New World species of Leishmania with different known immunopathologies . These molecules are abundantly present on the parasite surface in numbers great that 107 . Recently , they have been found associated to lipid rafts , essential for parasite infectivity and selective modulation of the host cell response [39] . In fact , there are several indications that GIPLs and other GPI-anchored molecules participate in cell signaling and are involved in the assembly of the NADPH oxidase complex , NO production [16] , [57] , [58] , [59] , [60] and inhibition of LPS and TNF-α induced c-fos gene expression by macrophages [61] . Also synthetic LPG , whose GPI anchor is structurally similar to GIPLs , can stimulate ERK activation and therefore inhibit IL-12 synthesis by macrophages [9] . Previous studies have demonstrated GIPLs antigenicity in chronic patients infected with L . major [38] , [62] . However , information concerning the biological relevance of GIPLs at early steps of infection in the innate immune compartment was still limited . Here , we demonstrated that GIPLs from both New World species were not able to activate the production of NO in non-primed macrophages , which was similar to published data from Old World species [59] , [60] . In primed macrophages an initial NO and TNF-α production was detected . Further , GIPLs differentially inhibited NO production even in the presence of IFN-γ and LPS , two major NO inducers . Previous studies indicated that LPG was a more potent agonist than GIPLs for the induction of pro-inflammatory cytokines [26] , [27] . In general , in comparison to LPS , GIPLs induced a lower production of NO and TNF-α . Also , they exhibited a strong inhibitor pattern during NO and cytokine induction , especially IL-12 . Similar strategy was demonstrated using crude extracts of the rat tapeworm Hymenolepis diminuta , although using different pathways . As shown by Johnston et al . ( 2010 ) [63] , crude extracts of this tapeworm could inhibit the production of TNF-α and IL-6 by mouse and human macrophages stimulates with TLR agonists poly ( I:C ) and Flagellin . These extracts also protected mice from experimental colitis accompanied by enhanced IL-10 and IL-4 production . In vivo studies using Old World species of Leishmania have demonstrated the importance of TLRs and other components of the innate immune system during infection . MyD88 is the most common adaptor molecule for the activation of NF-κB in most TLRs [28] . Also many studies using gene knockout have shown the importance of TLR and MyD88 adaptor molecule for cytokine production [29] , IL-1 promoter activation [64] , IFN-γ and IL-12 production [65] . NF-κB activation through TLR2 [26] , elastase dependent neutrophil control of L . amazonensis promastigotes [66] , and ultimately parasite control and lesion healing [27] , [65] , [67] . Indeed , in primed macrophages , GIPLs from both New World species were able to stimulate the production of NO , and this induction was mostly via TLR4 and to a lesser extent TLR2 ( Figure 3A ) . However , no difference was observed while stimulating with live parasites . Interestingly , in the L . braziliensis model , the TLR2 receptor plays a much more regulatory role in dendritic cells , repressing IL-12p40 and promoting IL-10 expression . This observation is correlated with sustained IFN-γ production and enhanced parasite control in TLR2 ( −/− ) mice [68] . However , in macrophages exposed to GIPLs , this difference in NO expression between TLR2 ( −/− ) and TLR4 ( −/− ) strains was not due to IL-12 , IFN-γ or IL-10 production ( Figures 3B and 4 ) . Also this induction was more pronounced in C57BL/6 than in BALB/c this was expected since C57BL/6 derived macrophages tend to be more responsive to stimuli than BALB/c macrophages [69] . These data are in accord with previous studies showing that related GIPLs from Trypanosoma cruzi are able to activate TLR4 [70] and studies with Old World species of Leishmania being able to activate TLR2 , TLR3 , TLR4 and TLR9 [28] . With exception to TNF-α , GIPLs and live parasites from L . braziliensis and L . infantum were not able to induce the other cytokines studied ( IL-1β , IL-2 , IL-4 , IL-5 , IL-10 , IL-12p40 and IFN-γ ) in primed and non-primed macrophages ( data not shown ) . Thus , we conclude that the GIPLs from these two New World species are less potent agonists or strong inhibitors for macrophages and the data presented here supports that the later might be true . When pre incubated with GIPLs , a strong inhibition of both NO and IL-12 production was observed ( Figures 4C and D ) . This inhibitory effect seems to be in specific pathways since no significant inhibition was detected for TNF-α ( Figure 4C ) . This inhibition is dependent on the intact structure of GIPLs since PI-PLC digested GIPLs that have its glycan core detached from its lipid anchor , failed to inhibit NO production by IFN-γ stimulated macrophages ( Figure 5 ) . Also , regarding TNF-α , only WT mice were able to trigger the production of this cytokine and this production was very low for TLR2 ( −/− ) and completely absent in TLR4 ( −/− ) ( Figure 3B ) . These data supports the premise that NF-κB translocation is not affected by GIPLs exposure [71] . It is noteworthy that the inhibition of IL-12 is not due to production of IL-10 , because we observed no IL-10 production either in unprimed ( data not shown ) or in primed macrophages incubated with GIPLs ( Data not shown ) . In TLR signaling , the most common adaptor molecule is MyD88 but other adaptor molecules may be involved in NF-κB translocation such as mitogen-activated protein kinases ( JNK or p38 ) [72] . Early studies showed that the Leishmania LPG can inhibit IL-12 without affecting NF-κB translocation to the nucleus [9] . For maximal downstream activation and GPI-induced gene expression , a full activation and cooperation Protein Tyrosine Kinase ( PTK ) and Protein Kinase C ( PKC ) are required . Although iM4 L . mexicana GIPL stimulated rapid PTK phosphorylation it failed in activating PKC [16] . In fact the unusual glycolipid composition ( mostly alkyl-acyl-glycerol ) of Leishmania GIPLs inhibits the activations of PKC [58] , [73] . This is in accordance with our observations that GIPLs not only fail on inducing a pro-inflammatory response in non-macrophages but also that the GIPLs inhibit the productions of IL-12 and NO . Also we tested whether GIPLs from both New World species were able modulate the phosphorylation of MAPKs . We observed that the GIPLs activate only ERK , whereas LPS activated both ERK and p38 ( Figure 6 ) . Also we observed that the GIPLs can prevent the phosphorylation of both ERK and p38 MAPKs stimulated by LPS . However , ERK activation was too low to provide evidence for any further effect on IL-12 production . It is likely that L . braziliensis and L . infantum GIPLs have a profound effect on macrophage cell signaling affecting PTKs , PKCs and MAPKs , and that GIPLs from both species use similar pathways but differ in the intensity in which they modulate NO and IL-12 production . In this work , GIPLs interacted with primed macrophages resulting only in the production of NO and TNF-α . GIPLs are abundant in the amastigote stage of Leishmania and are associated to highly specialized microdomains [39] and the participation of each kind of GIPL on the process is still under debate [74] , [75] , [76] . Also it is possible that the dependency on a particular glycolipid may vary throughout species and life cycle stage . The data presented here clearly supports the hypothesis that Leishmania GIPLs , differently from other trypanosomatids , may contribute to build a safer environment to promote infection by manipulating macrophage function and by disrupting the polarization of TH1/TH2 response , through inhibiting IL-12 production during the initial stages of infection and manipulate macrophage for parasite survival . In general , LPGs and GIPLs share similar lipid anchor moieties among the various species of Leishmania and the integrity of this portion is important for TLR2 activation [27] . To ascertain if the differences in the inhibition of NO and IL-12 production could be related to polymorphisms in GIPL structure , we analyzed the carbohydrate core of L . braziliensis and L . infantum GIPLs . Previous studies from our group showed that the phosphoglycan domains of LPGs from L . braziliensis and L . infantum differ in structure and composition [40] , [41] and differences in glycan portions of GIPLs were also observed in this study . The iM2 species of GIPLs possesses the structure Manα1-3Manα1-4GlcN-P ) similar to LPG core region , and isoM3 has a hybrid glycan in GIPLs ( substitutions on both the third and sixth carbons of the distal mannose ) with the structure of Manα1-6 ( Manα1-3Manα1-4GlcN-PI . Our structural observations indicated that the GIPLs from L . infantum are similar to the known structures in L . donovani [32] and are composed mainly of mannose residues . This data suggests that the majority of these GIPLs as Type I GIPLs and Hybrid GIPLs . On the other hand , L . braziliensis GIPLs shows a different profile of sugar composition and different bands distinguishable on TLC ( Figure 7A ) . We determined that there was a stoichiometric ratio of galactose and mannose in the glycan portion of these GIPLs . This data suggest that these GIPLs are similar to the closely related species L . panamensis [36] , which have a common Galfβ1-3Manα1-3Manα1-4GlcN-myoinositol glycan headgroup and a structurally related to LPG lipid anchor , suggestive of Type II GIPLs . Type II GIPLs can be very diverse and substitutions on the 3rd carbon of the Galf residue by Galα-1 , Galα1-3galα1 , and even longer saccharides like Manα1-PO4-6Galα1-6Galα1 can be detected in other species like L . major [31] . These substitutions can lead to large GIPLs containing up to 7 , 8 or even more hexoses [34] , [36] , which we observed from the L . braziliensis GIPLs as seen on Figure 8 . In conclusion , GIPLs from both New World species L . infantum and L . braziliensis have a strong inhibitory potential during intracellular Leishmania infection of the mammalian host . Only an initial production of NO and TNF-α was detected after stimulation by GIPLs . Due to their importance in modulating NO and cytokine production , these molecules could be possible targets to alternative immunological and chemotherapeutic control methods . The preliminary qualitative analysis of GIPLs from these two species showed that they differ in composition and structures thus , suggesting that the structural distinctions could be responsible for differential NO and IL-12 inhibition in macrophages . Also , GIPLs were also capable of affecting macrophage ability to produce NO in the presence of IFN-γ and LPS . These data , together with already published data from other groups , suggest that GIPLs may be involved in the interaction with the macrophage triggering a minimal pro-inflammatory response in the host and to the benefit of the parasite . Glycoconjugate interspecies polymorphisms , not only in the GIPLs , but also in LPG , gp63 and other GPI-anchored molecules could be important for differential establishment of infection . These polymorphisms could result in different clinical outcomes , such as those shown by L . infantum and L . braziliensis , causative agents of a visceral and tegumentary forms , respectively [77] . | Leishmania infantum ( syn . L . chagasi ) and L . braziliensis are the causative agents of VL and CL , respectively , in the New World . A vital part of the parasite's life cycle involves the circumvention of the host immune system and the infection of macrophages . This work focused on an important class of surface glycoconjugates , the glycoinositolphospholipids ( GIPLs ) , and their role in the interaction with murine macrophages . GIPLs are expressed on every stage of the parasite life cycle and are the most abundant molecules on its surface . Here we show that these molecules modulate many macrophage functions such as cytokine production , release of nitric oxide and differentially activate MAPK . Although the GIPLs of both New World species are capable of modulating the same mechanisms , they do so to different degrees requiring an examination of their glycan composition . We show that L . infantum synthesize mannose rich GIPLs whereas L . braziliensis express galactose rich GIPLs . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"biochemistry",
"infectious",
"diseases",
"immunology",
"biology"
] | 2012 | Glycoinositolphospholipids from Leishmania braziliensis and L. infantum: Modulation of Innate Immune System and Variations in Carbohydrate Structure |
Adaptive mate choice by females is an important component of sexual selection in many species . The evolutionary consequences of male mate preferences , however , have received relatively little study , especially in the context of sexual conflict , where males often harm their mates . Here , we describe a new and counterintuitive cost of sexual selection in species with both male mate preference and sexual conflict via antagonistic male persistence: male mate choice for high-fecundity females leads to a diminished rate of adaptive evolution by reducing the advantage to females of expressing beneficial genetic variation . We then use a Drosophila melanogaster model system to experimentally test the key prediction of this theoretical cost: that antagonistic male persistence is directed toward , and harms , intrinsically higher-fitness females more than it does intrinsically lower-fitness females . This asymmetry in male persistence causes the tails of the population's fitness distribution to regress towards the mean , thereby reducing the efficacy of natural selection . We conclude that adaptive male mate choice can lead to an important , yet unappreciated , cost of sex and sexual selection .
We first develop a graphical model in which a single quantitative trait is a reliable , direct indicator ( rather than an indirect indicator , like a costly ornament ) of a female's “intrinsic” fecundity ( i . e . , fecundity in the absence of costly male persistence ) . For example , in a wide diversity of taxa , variation in female fecundity is strongly correlated with body size [2] , [4] because larger females have more resources to invest in fecundity . Henceforth , we arbitrarily assume that a female's body size is the phenotypic trait correlated with fecundity , but our logic applies to other indicator traits that directly influence her fecundity , such as parasite load [25] or abdomen size in many insects [7] . Males are expected to evolve a mating preference for larger females whenever this preference increases their own lifetime reproductive success [14] . Such an adaptive male mate preference will cause larger , intrinsically high-fecundity females , to receive more antagonistic male persistence , compared to smaller , intrinsically low-fecundity females . The fitness consequences of this relationship will depend upon how female resistance to male-induced harm scales with the indicator trait ( in this case , body size ) . Assuming that a female's resistance to the harmful male persistence does not rise sufficiently fast with increasing body size , the male preference should reduce the fecundity of large females and increase that of small females , thereby reducing the standing variance in fitness ( Figure 1 ) . As a result , the selective advantage of any beneficial genetic variation that makes females more competitive for limiting resources , and hence more fecund , will experience a smaller selective advantage than if harmful male persistence was randomly applied to females throughout the population . Such nonrandom male persistence will cause adaptive evolution in females to be slowed whenever female fecundity is: ( i ) heritable , ( ii ) genetically correlated with the indicator trait , and ( iii ) a major determinant of her lifetime fitness that does not strongly trade off with her other fitness components . A reduced rate of adaptive evolution by females can also be deduced from Fisher's fundamental theorem [26] , so long as the male-induced reduction in the phenotypic variation in female fecundity also leads to a reduction in the additive genetic variation among females . Furthermore , when there is a positive genetic correlation for fitness between females and males , adaptive male mate choice is expected to reduce the rate of adaptation in both sexes . Although a counteracting effect could occur if male preference for high-fecundity females increases the variance in male fitness , or if male preferences lead to positive assortative mating for fitness , here we focus on female fitness and the potential for male mate preferences to reduce its heritable variation . The conclusion that adaptive male choice leads to a reduced rate of adaptation by females can also be deduced by focusing on mutations at a single arbitrary locus . Let the mutation rate to new beneficial mutations be UBen and the selective advantage of the mutation , expressed as a selection coefficient and averaged across the sexes , be s . Assuming approximate additivity ( i . e . , little dominance ) , the probability of the mutation becoming fixed can be calculated using the diffusion approximation [27] as 2s ( Ne/N ) , where N is the population size and Ne is the effective population size . With recurrent mutation to new beneficial mutations , the rate of advance of adaptive evolution is approximated by: ( 1 ) ( 2 ) If we partition selection between the sexes and let the selective advantage of a mutation be s♂ in males and s♀ in females then , ( 3 ) ( 4 ) Next , we assume that the expression of the beneficial mutation also increases the attractiveness of females to males ( e . g . , as a result of increasing her body size ) , so that those females expressing the beneficial mutation receive an excess of antagonistic male persistence . We express this cost with an additional selection coefficient s♀biased-persist , which is applied only to females , ( 5 ) Comparison of Equations 4 and 5 demonstrates that the rate of adaptive evolution will always be slower whenever males bias their antagonistic persistence towards fitter females and cause s♀biased-persist to be negative; i . e . , when there is adaptive male mate choice and increased male persistence is harmful to females . Increased male persistence directed towards more fecund females that express the beneficial allele reduces the selective advantage of those females and thereby reduces the variance in fitness among females in the population . The primary prediction from our models is that , in species with antagonistic male persistence , adaptive male mate preference leads to a “cost of being an attractive female . ” This cost reduces the selective advantage of females expressing more beneficial genetic variation ( and hence are larger , on average ) and increases the fitness of females expressing less of this variation ( and hence are smaller , on average ) . Put more simply , adaptive male mate preference causes the tails of the population's distribution of female lifetime fecundity to regress towards the mean ( Figure 1 ) . This prediction is contingent on four assumptions that must be met in order for our model to operate: ( i ) lifetime fecundity and net fitness are strongly genetically correlated , ( ii ) body size and fecundity are positively correlated , both phenotypically and genetically , ( iii ) more antagonistic male persistence is directed towards females with higher intrinsic fecundity ( i . e . , potential fecundity in the absence of costly male persistence ) , and ( iv ) female resistance to male-induced harm does not rise sufficiently fast with increasing body size . We tested the major prediction of the model , and its underlying assumptions , using a laboratory population of the model species D . melanogaster . In this population , assumption ( i ) is well established [28] , [29] , so here we focus on testing whether our population meets assumptions ii–iv , before experimentally assessing whether male mate preference for high-fitness females causes the tails of the distribution of lifetime fecundity to regress towards the mean .
Joint measures of female body size and lifetime fecundity in our base population ( LHM ) of D . melanogaster indicated that these two phenotypic traits are strongly correlated . As predicted from past studies of many taxa [2] , [4] , [30] , fecundity was higher in large females compared to small females ( Figure 2 ) . This result was found both when females experienced minimal exposure to males ( mean ± standard error [SE]: large females , 27 . 3±1 . 59; small females , 18 . 42±1 . 07; t-test t = 4 . 64 , df = 98 , p<0 . 0001; p-values reported throughout the manuscript are two-tailed ) and when male exposure was continuous ( large females , 16 . 5±1 . 06; small females , 12 . 62±0 . 72; t-test t = 3 . 01 , df = 98 , p<0 . 003 ) . We tested for a genetic correlation between body size and fecundity in a separate study in which two populations each were artificially selected for either large or small body size . After 83 generations of divergent selection , lifetime fecundity was significantly higher in the lines selected for large body size compared to the lines selected for small body size ( mean ± SE: large females , 31 . 38±1 . 81; small females , 15 . 25±1 . 58; t-test t = 6 . 69 , df = 2 , p = 0 . 02 ) . Since body size was the only target of artificial selection , the large divergence in fecundity between treatments demonstrates a strong positive genetic correlation between body size and fecundity , a result consistent with other research [31] . To test this assumption , we first measured how the persistence ( courtship behaviour ) of individual males was allocated between two nonvirgin females ( differing in eye colour phenotype , brown or red , for ease of individual identification ) . We performed a two-way ANOVA on the amount of persistence behaviour directed towards each female , with the body size of that “target” female ( large or small ) , the body size of the competitor female present in the test tube ( large or small ) , the eye colour of the target female ( red or brown ) , and all possible interactions as predictor variables . This analysis was significant overall ( F7 , 232 = 7 . 50 , p<0 . 0001 ) , with significant effects of the target female body size ( F1 , 232 = 38 . 37 , p<0 . 0001 ) and the body size of the competitor female ( F1 , 232 = 13 . 53 , p = 0 . 0003 ) , but no effects of eye colour ( F1 , 232 = 0 . 41 , p = 0 . 52 ) , or any of the interactions ( all p>0 . 60 ) . When individual males were housed with two nonvirgin females differing in body size , males directed more persistence towards the larger female than towards the smaller female ( paired t-tests , p≤0 . 0002 , Figure 3 ) . When males were housed with two nonvirgin females of similar body size ( both small or both large ) , the levels of male persistence directed towards the red- and brown-eyed females were not significantly different ( paired t-tests , p≥0 . 50 , Figure 3 ) . Further evidence of a male mate preference for nonvirgin females of larger body size was obtained from mating assays conducted under conditions that more closely mimicked the normal culture environment of the LHM population ( 16 males combined with 16 females during the “adult competition” phase of the life cycle [28] , [32] ) . When presented with a choice of nonvirgin females differing in body size , males mated with large-bodied females at a greater rate than with small-bodied females ( generalized linear model [GLM] with binomial error terms; pconsensus = 1 . 17×10−5 , Replicate 1: χ21 , 18 = 5 . 90 , p = 0 . 015; Replicate 2: χ21 , 74 = 22 . 87 , p<0 . 0001; Figure 3 ) . These remating results are unlikely to have arisen from large females possessing a greater receptiveness to male courtship effort because males kept under “no-choice” mating conditions ( where either only large or only small nonvirgin females were present ) mated with small females more frequently than with large females ( GLM with binomial error terms; pconsensus<1×10−6 , Replicate 1: χ21 , 28 = 14 . 79 , p = 0 . 0001; Replicate 2: χ21 , 28 = 7 . 16 , p = 0 . 0075; Figure 4 ) . To test this assumption , we compared the reduction in lifetime fecundity of large and small females when they were either minimally or continuously exposed to males ( Figure 2 ) . Continuous male exposure harmed large females more than small females ( two-way ANOVA , interaction between body size and male exposure , F1 , 196 = 4 . 75 , p = 0 . 031 ) , indicating that larger females were not more resistant to the harmful male persistence that they received . Finally , we tested the model's key prediction by comparing the mean fecundities of large- and small-bodied females used in the choice and no-choice mating assays described earlier . In the no-choice assays , where all females were either of large or small body size , male preference for large female body size could not cause them to direct their antagonistic persistence away from smaller females and towards larger females . In contrast , in the choice assays , where females of different body sizes were simultaneously present , a redirection of antagonistic male persistence towards larger females was possible . We found that the difference in the mean fecundities of large- and small-bodied females was smaller when males could direct their antagonistic persistence towards large females ( Figure 5 , pconsensus = 0 . 012 , interaction tests for each replicate: F1 , 46 = 2 . 79 , p = 0 . 1 for the smaller , first replicate; F1 , 101 = 4 . 92 , p = 0 . 03 for the larger , second replicate ) .
The results of our male mate preference tests clearly demonstrate that males have mate preferences for larger nonvirgin females—a result consistent with earlier work on virgin females [33] . Rather than displaying an “undiscriminating eagerness” [34] to mate , when given a choice between females differing in body size , male D . melanogaster preferred to court and mate with large , high-fecundity females over small , low-fecundity females . Given the significant fecundity differences associated with female body size described above , this mate preference is likely to be adaptive from the male's perspective , as mating with larger , more fecund females is likely to yield greater direct , as well as indirect [35] , benefits . It is unlikely that this male mate preference is adaptive from the female's perspective , as several studies have established that chronic male persistence in the LHM population is very harmful to females , and it is not sufficiently compensated by indirect genetic benefits [36]–[38] . Our experiments demonstrate that larger females receive more harmful male persistence but do not reveal the specific mechanism by which this harm accrues . Further work will be needed to resolve the degree to which this increased harm is due to harassment during courtship [24] , damage associated with copulation [39] , and/or the activity of products transferred in the male's seminal fluid [23] , [40] . Having experimentally ascertained that the LHM population of D . melanogaster satisfied all the assumptions necessary in which to test the key prediction our model , we were able to meaningfully assess the fitness consequences of adaptive male mate preferences . When males had the ability to bias their antagonistic persistence towards large-bodied females , we saw a decrease in the mean fecundity of these preferred females , compared to those large females that were in an experimental environment where all females were of similar size , and biases of antagonistic male persistence were not possible . In contrast , small-bodied females were , on average , able to realize relatively higher fecundities when they were housed with larger females ( which , our study indicates , were attracting more harmful male persistence ) than they were when they were housed in an environment in which males had no other choice of mates . Although our study found that males directed more courtship towards large females and also mated them more frequently , both of which can be harmful in and of themselves [24] , the observed cost to large females might also have occurred because large females were mated , on average , to more harmful males [30] , [41] than were smaller females . Irrespective of the mechanism of this cost , together these assays revealed how male mate preferences will ultimately cause the tails of the distribution of fecundity to regress towards the mean . Since adult lifetime fecundity is strongly correlated with lifetime fitness in females of the LHM population [42] , this male-driven sexual selection is expected to reduce the rate of adaptive evolution of any trait that is positively correlated with female body size . It is common for deleterious mutations to reduce body size in D . melanogaster [43] , and it is reasonable to assume that many beneficial mutations will cause their carriers to be more competitive as larvae , allowing them to garner more resources during the larval competition phase of their life cycle and become larger , more fecund , adults . As a consequence , male mate preference for larger females is expected to commonly interfere with both progressive evolution and to increase the population's mutational load by interfering with purifying selection . For example , suppose that environmental change led to selection for alleles conferring higher desiccation tolerance . If more desiccation-tolerant females had a competitive advantage such that they grew to a larger size prior to reproduction ( e . g . , [44] ) , then a male preference for these females would reduce their relative fecundity and increase that of smaller , less desiccation-tolerant females . As a result , the population may be less responsive to environmental change , become an inferior competitor species , and be at a greater risk of extinction . Collectively , our results support our model's key prediction that male mate preference for high-fitness females reduces the selective advantage of larger , more fecund females and increases that of smaller , less fecund females . This finding , obtained in a laboratory population , is likely to apply to natural populations for two reasons . First , the study was done on a large , outbred population that has been maintained in a competitive laboratory environment , at continuous large size , for over 400 generations [28] , [32] . Over this period of time , the opportunity for adaptation to the laboratory environment should have been substantial , permitting the flies to be experimentally assayed under conditions to which they are highly adapted . Second , we measured natural variation in body size , rather than inducing extreme body size variation via nutritional deprivation and/or excessive larval crowding . This was accomplished using a sieve shaker device ( developed by ADS and WRR ) , which enabled us to quickly sort thousands of adult flies based on natural variation in their body size , and obtain the largest and smallest individuals to use in our experiments . Flies from these two body size groups differed markedly in fecundity , with the larger females producing over 30% more eggs than small females under both minimal and continuous male exposure conditions . Although our assays of male mate preference support a directional preference for large-bodied females , in one assay ( Figure 2 ) , males could only choose between females of large and small body size . Thus , there is the possibility that the true male preference function favours females of intermediate size . However , in our second assay ( Figure 3 ) , males were able to choose between large or small females versus random females ( average ) , and these data support the conclusion that male preference is monotonic for larger females . Our model of adaptive male mate choice in the context of harmful male persistence has important limitations . First , we have implicitly assumed that the increased male persistence ( directed toward larger , more intrinsically fecund females ) does not cause larger females to have lower than average fecundity . Second , male condition may be more variable in nature compared to the laboratory , and condition-specific patterns of male persistence could either enhance or reduce the bias of male persistence toward larger females . Third , we have ignored complicating factors such as size-assortative mating interactions , e . g . , smaller females receiving persistence predominantly from smaller or poor-condition males . Fourth , we have assumed that male mate choice is based on a female trait that directly influences her fecundity , such as body size . Theory predicts that this type of male mate preference will lead to a monotonic preference for larger females [45] , [46] . However , when the preferred female trait is a costly indicator of fecundity , such as an energetically expensive ornament , then males can evolve to prefer intermediate trait values in females [45] , [46] , and our model would not apply . Fifth , our model may not apply to species where females obtain direct net benefits from increased mating rates , such as those with nuptial feeding [47] . Lastly , we have assumed a static male preference and female indicator trait . In many contexts , these two traits can be expected to coevolve , and this dynamic is not included in our model . Nonetheless , our empirical work suggests that the requisite conditions for the model to operate , at least transiently , can feasibly be achieved . Our finding of harmful effects of adaptive male mate choice represents a previously unappreciated cost of sexual reproduction in species with antagonistic male persistence . Rather than simply showing that male-induced harm reduces overall female fecundity , we have shown that biases in the distribution of this harm among mates reduces the selection differential between females with intrinsically high and low fecundity . This reduced efficacy of natural selection will retard a population's rate of adaptive evolution and increase both its equilibrium mutational load and its stochastic accumulation of harmful mutations . The cost of adaptive male mate choice , however , only applies when males can reliably ascertain a female's fecundity using a trait that is heritable and correlated with heritable fitness variation . In Drosophila , female body size represents such a trait since it is influenced by both genotype [48] , [49] and a number of environmental factors ( including temperature , nutrition and larval crowding conditions [50] ) , and responds rapidly to directional selection . In species with little or no heritability for body size , however , an adaptive cost of male preference for high-fecundity females would not apply . Nonetheless , given the prevalence of male mate preferences [7] , this new cost that we describe may be a widespread evolutionary phenomenon . For this reason , it should be considered in the broader context of the ongoing debates over the interfering or reinforcing role that sexual selection plays in the process of adaptation , and whether sexual selection increases or decreases the risk of extinction of populations and species [51] .
For all male–female interaction assays , we used D . melanogaster adults obtained from the wild-type LHM population [28] , [29] or from a replicate population ( LHM-bwD ) in which a dominant brown-eyed marker ( bwD ) had been introgressed through repeated backcrossing into the LHM genetic background . The LHM population is maintained on a 14-d culture cycle with a 12-h L∶12-h D diurnal cycle at 25°C in humidity-controlled incubators . Briefly , each generation begins with eggs placed in 56 “juvenile competition” vials ( 150–200 eggs per vial; each vial containing 10 ml of cornmeal/molasses medium ) . After 11 . 25 d , emerging adults are lightly anesthetized with CO2 , mixed among vials , and transferred to “adult competition” vials ( 16 pairs of males and females per vial ) , which are seeded with 6 . 4 mg ( dry weight ) of live yeast . After 2 d of adult competition , the flies are transferred to “oviposition” vials , and then discarded after laying eggs for 18 h . The eggs laid in these oviposition vials are culled to a density of 150–200 eggs per vial and become the “juvenile competition” vials of the next generation . Because only eggs from the oviposition phase of the life cycle are used to propagate the next generation , and populations have been consistently maintained under these culture conditions for over 400 generations , the number of eggs laid during the 18-h oviposition phase represents a meaningful measure of lifetime fecundity in these populations . As such , experiments were designed to mimic these culture conditions as closely as possible . Detailed culturing protocols for these large populations ( adults n>1 , 800 per generation for LHM and n>1 , 300 per generation for LHM-bwD ) can be found elsewhere [28] , [29] ) . We altered the quality of potential female mates by collecting females of differing adult body size , a phenotypic trait that is frequently positively correlated with fecundity [2] , [30] , [52] . We collected flies from the ends of the normal distribution of body sizes that are produced under typical lab culture conditions . Flies were sorted by size with the use of a sieve shaker device ( Gilson Performer III , Gilson Company ) which mechanically separates anesthetized flies on the basis of their ability to pass through a series of 20 electroformed sieves , in which the diameter of the holes in each sieve was 5% larger than the diameter of the holes of the sieve below ( diameter of top sieve holes = 1 , 685 µm; diameter of bottom sieve holes = 800 µm ) . Flies were placed into the column ( under light CO2 anaesthesia ) , and were agitated at a rate of 3 , 600 vibrations min−1 for 2 min . By using this technique , it was possible to quickly sort hundreds of flies simultaneously on the basis of their body size . For all experiments , “small” flies were defined as those that were small enough to pass through the 1 , 095-µm diameter sieve , whereas “large” flies were those that were too large to pass through the 1 , 281-µm diameter sieve . To assess the phenotypic correlation between body size and fecundity , we collected adult flies from the LHM population as they eclosed as virgins on day 9 of their life cycle . Flies were separated by sex , and on the following day , females were sorted by size using the sieve sorter protocol described above . One hundred female flies each of large and small body size were then placed individually ( under light anaesthesia ) into small test tubes that had been seeded with 0 . 4 mg of yeast ( the amount of yeast per female experienced under normal culture conditions ) . Into each of these vials , three adult males were placed for a period of 2 h , during which time all virgin females were observed to have mated once . Males were then removed randomly from half of the vials to create 50 adult competition vials with minimal male exposure and 50 with continuous male exposure for each female body size category . Maintaining flies under these two conditions allows us to confirm that there is an intrinsic difference in fecundity between females of different sizes that is independent of the negative net fitness effects of continuous male presence . Matching the normal culturing protocol of the flies , vials were returned to the incubator for an additional 2 d , at which time flies were transferred to oviposition vials containing fresh medium ( with a scored surface to encourage oviposition ) for a period of 18 h before being discarded . The number of eggs laid in each vial was counted , and mean fecundities were compared using t-tests for females differing in size in each male-exposure treatment . The complete dataset was also used to test the assumption that female resistance to male-induced harm does not rise sufficiently fast with increasing body size , by examining whether or not female flies of one size were harmed more by continuous male exposure . In order to verify that there was a genetic correlation between body size and fecundity , we assessed the fecundity of females obtained from populations of D . melanogaster that are part of an ongoing experimental evolution project ( of ADS and WRR ) in which females had been artificially selected for either large or small body size using a size-sorting procedure similar to that described above . These populations are otherwise cultured in a manner similar to the LHM population from which they were all originally derived . At the time of the assay , the artificial selection had been operating for 83 generations in each of two replicate populations per treatment , and there had been considerable divergence in body size ( mean female diameter [µm] ± SE: large treatment , 1 , 218 . 5±37 . 67; small treatment , 786 . 7±43 . 5; t-test t = 7 . 51 , df = 2 , p<0 . 01 ) . For this assay , 72 virgin females were obtained at random from each of the four experimental populations . On day 11 of their life cycle , these females were placed in adult competition vials in groups of 16 , along with 16 males taken randomly from the LHM population , for a period of 2 h , during which time all females were observed to have mated once . Males were removed from the vials , and after 2 d in the incubator , females were transferred to individual oviposition vials containing fresh medium ( with a scored surface ) for a period of 18 h before being discarded . The number of eggs laid in each vial was counted and the mean fecundity of the two replicates of each treatment was compared using a t-test ( with population as the unit of replication ) . Since the selected trait in these experimental populations was body size , any consistent change in fecundity between the two treatments must be due to a genetic correlation between the two traits . In order to test whether males have mate preferences , a series of behavioural assays were conducted . Nonvirgin flies from both the LHM and LHM-bwD populations were collected on day 11 of their life cycle , and females were sorted by size to isolate large- and small-bodied individuals . Pairs of female flies differing in eye colour ( to aid individual identification ) were placed into small , adult competition vials ( test tubes ) in all possible combinations of body size . After a 1-h anaesthesia-recovery period , a single unanaesthetized adult male fly was added to each test tube , which were then placed on their sides in a well-lit room . Over the course of 11 sessions , spaced 40 min apart , the male in each test tube was observed . Male persistence behaviour was defined as being located within 5 mm of a female and oriented towards her [53]–[55] . Data on the frequency of the male persistence behaviour was collected for each type of female in each treatment . A total of 30 replicate test tubes per treatment were scored . In these assays , nonvirgin adult female LHM flies were collected on day 11 of their life cycle and sorted by size ( see above ) to isolate large and small individuals . Females were then placed into one of two types of adult competition vials ( a vial containing fresh medium seeded with 6 . 4 mg of live yeast ) . In the first , choice experiment , either eight large or eight small red-eyed LHM females were placed into an adult competition vial along with eight randomly collected LHM-bwD females and 16 LHM-bwD males . In the second , no-choice treatment , either 16 large or 16 small red-eyed LHM adult females were placed into an adult competition vial along with 16 LHM-bwD males . These vials were kept in the incubator ( on their sides ) for 24 h , at which time males were removed . The vials , containing females only , were then returned to the incubator for an additional 24 h . Remating rates were assayed by placing all females into individual oviposition vials ( test tubes ) containing fresh , scored medium for the purpose of measuring the paternity of her offspring . Eighteen hours later , the adult flies were discarded , and the test tubes containing eggs were incubated for 11 d . At this time , the presence and number of red-eyed and brown-eyed progeny in each brood were scored to ascertain whether the female had remated . The proportion of females in each adult competition vial that produced brown-eyed offspring ( indicating a remating event ) was recorded . To examine remating rates in relation to female body size and treatment , we constructed GLMs that used a logit link function and binomial error distribution , where the number of females that remated is the dependent variable and the total number of females assayed is the binomial denominator . We tested whether male mate preferences caused the tails of the distribution of female lifetime fecundity to regress towards the mean by performing a two-way ANOVA , with body size , remating treatment , and their interaction as predictor variables . A significant interaction term ( that was associated with a smaller difference between the mean fecundity of large and small females when male preference was possible ) would indicate that the tails of the fecundity distribution had regressed toward the mean . Each type of remating assay was repeated twice . The first , choice assay was comprised of ten adult competition vials ( the unit of replication ) for each body size treatment , whereas the second replicate was comprised of 38 adult competition vials in the large body size treatment and 37 in the small body size treatment . Both replicates of the no-choice assay were comprised of 15 adult competition vials for each body size treatment . | In many species , females are frequently subject to harassing courtship from males attempting to mate with them . These persistent male behaviors can result in females incurring substantial direct fitness costs . We set out to examine how these costs may influence adaptive potential in a species that also exhibits male mate choice , i . e . , a preference by males for females exhibiting certain traits . We found that harmful courtship behaviors were directed predominantly towards females of greater reproductive potential ( and away from females of lesser potential ) , resulting in a reduction in the variation of lifetime reproductive successes among females in the population . This change in distribution of realized fitnesses represents a previously unappreciated consequence of sexual conflict–adaptive male mate preference can slow the rate of accumulation of beneficial mutations and speed the rate of accumulation of harmful mutations , thereby creating a “sexual conflict adaptive load” within a species . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"evolutionary",
"biology",
"evolutionary",
"biology/sexual",
"behavior",
"evolutionary",
"biology/animal",
"behavior"
] | 2009 | A Cost of Sexual Attractiveness to High-Fitness Females |
Promyelocytic leukemia protein ( PML ) has antiviral functions and many viruses encode gene products that disrupt PML nuclear bodies ( PML NBs ) . However , evidence of the relevance of PML NB modification for viral pathogenesis is limited and little is known about viral gene functions required for PML NB disruption in infected cells in vivo . Varicella-zoster virus ( VZV ) is a human alphaherpesvirus that causes cutaneous lesions during primary and recurrent infection . Here we show that VZV disrupts PML NBs in infected cells in human skin xenografts in SCID mice and that the disruption is achieved by open reading frame 61 ( ORF61 ) protein via its SUMO-interacting motifs ( SIMs ) . Three conserved SIMs mediated ORF61 binding to SUMO1 and were required for ORF61 association with and disruption of PML NBs . Mutation of the ORF61 SIMs in the VZV genome showed that these motifs were necessary for PML NB dispersal in VZV-infected cells in vitro . In vivo , PML NBs were highly abundant , especially in basal layer cells of uninfected skin , whereas their frequency was significantly decreased in VZV-infected cells . In contrast , mutation of the ORF61 SIMs reduced ORF61 association with PML NBs , most PML NBs remained intact and importantly , viral replication in skin was severely impaired . The ORF61 SIM mutant virus failed to cause the typical VZV lesions that penetrate across the basement membrane into the dermis and viral spread in the epidermis was limited . These experiments indicate that VZV pathogenesis in skin depends upon the ORF61-mediated disruption of PML NBs and that the ORF61 SUMO-binding function is necessary for this effect . More broadly , our study elucidates the importance of PML NBs for the innate control of a viral pathogen during infection of differentiated cells within their tissue microenvironment in vivo and the requirement for a viral protein with SUMO-binding capacity to counteract this intrinsic barrier .
Promyelocytic leukemia nuclear bodies ( PML NBs ) , also called nuclear domain 10 ( ND10 ) bodies , are spherical nuclear structures that are present in most mammalian cells [1] . PML protein is the essential major component and recruits other proteins , such as Sp100 , Daxx , SUMO1 , CBP and p53 [2] . PML NBs are associated with many cellular processes , including transcription , the DNA damage response , apoptosis and oncogenesis [1] . Under normal conditions , the number of PML NBs varies depending on cell type and differentiation status [3]-[5] . PML is also an interferon ( IFN ) -inducible protein and IFN treatment increases the number and size of PML NBs [6] . PML NBs have been shown to contribute to innate defenses against a broad range of viruses [7] . In turn , many viruses encode products that modify or eliminate PML NBs in cultured cells [7] . However , few studies have investigated the role of PML NBs in viral pathogenesis or mechanisms by which they are modified during viral infection in vivo [8] , [9] . Varicella-zoster virus ( VZV ) is the etiologic agent of varicella ( chickenpox ) and herpes zoster ( shingles ) and causes characteristic cutaneous lesions in both diseases [10] . VZV is an alphaherpesvirus closely related to herpes simplex virus ( HSV ) 1 and 2 [10] . PML is known to interfere with HSV early viral gene transcription and it is important for the antiviral effects of IFNs on HSV [11] , [12] . PML knock-down and over-expression experiments indicate that PML is also involved in restricting VZV replication [9] , [13] . In HSV-infected cells , the ICP0 protein triggers the proteasome-dependent degradation of sumoylated PML and Sp100 through ubiqutin E3 ligase activity mediated by its RING domain [14]-[17] . In contrast to HSV , VZV does not degrade PML protein [9] , [13] , although it does disrupt PML NBs , causing a reduction of approximately five-fold in PML NB frequencies in vitro [9] . Our recent work demonstrated that the PML NBs that persist in VZV infected cells in vitro and in skin and dorsal root ganglia ( DRG ) in vivo have the capacity to sequester newly formed nucleocapsids [9] . The entrapment of VZV capsids in these nuclear PML cages depended upon an interaction between PML and the ORF23 capsid protein and acted as an intrinsic antiviral host defense [9] . The VZV ortholog of HSV ICP0 is ORF61 [10] . Like ICP0 , ORF61 colocalizes with PML NBs shortly after virus entry and disperses Sp100 NBs in transfected cells if the conserved RING domain is intact [18] , [19] . The ORF61 RING domain also exhibits E3 ligase activity in vitro [19] , [20] . Besides these functions , ORF61 has been shown to act as a transactivator to regulate a number of viral and cellular promoters in transient transfection assays and contributes to the optimal expression of VZV glycoprotein E during virus replication in cultured cells [21]-[23] . ORF61 is essential for VZV replication in vitro; deleting ORF61 is not compatible with recovery of infectious virus and truncating ORF61 or severely limiting its expression by mutating promoter elements markedly impairs virus replication [22]-[24] . Although VZV is a highly human-restricted virus , its pathogenesis can be investigated using xenografts of human skin , thymus and DRG in severe combined immunodeficiency ( SCID ) mice [25] , [26] . The evaluation of VZV recombinant viruses with genetic mutations in the SCID mouse model makes it possible to analyze mechanisms of virus-host cell interactions in differentiated cells in vivo and to determine whether putative functional motifs in viral proteins or promoters contribute to the capacity of the virus to overcome intrinsic barriers . Targeted mutations in the viral genome that have little or no effect in VZV replication in cultured cells can disrupt functions that are critical for pathogenesis [27] . By evaluating ORF61 promoter mutants in the skin xenograft model , we demonstrated that ORF61 is necessary for VZV skin pathogenesis [23] , but the reason for this ORF61 requirement was not defined . In this study , our goal was to investigate the functional elements of ORF61 protein , which are required for interaction with PML NBs in differentiated cells infected in vivo and the contribution of this interaction to VZV infection in skin using the xenograft model . To identify potential ORF61 functional domains , we analyzed the ORF61 sequence and found that it has three putative small ubiquitin-like modifier ( SUMO ) -interacting motifs ( SIMs ) in addition to the conserved RING domain . SIMs have been identified in a number of proteins and have a hydrophobic core , consisting of 3-4 aliphatic residues ( V/L/I-x-V/L/I-V/L/I or V/L/I-V/L/I-x-V/L/I; ‘x’ means any amino acid ) , which are typically flanked by a stretch of negatively charged amino acids [28]-[30] . Structural studies indicate that the motif has an extended configuration and is embedded in the groove formed between the α-helix and the β-strand of SUMO [31] . PML protein contains a SIM and the binding through the SIM to sumoylated PML is considered to be the nucleation event for recruitment of other sumoylated and SIM-containing proteins [32] . In support of this model , SIMs in cellular proteins , including Daxx , RNF4 and Sizn1 and in the Kaposi's sarcoma herpesvirus ( KSHV ) LANA2 protein , are necessary for their association with or modification of PML NBs [33]-[36] . This study was designed to investigate whether the putative SIMs that we identified in ORF61 mediated ORF61 binding to SUMO and if so , whether these SIMs were important for ORF61 association with PML NBs and for the disruption of PML NB in vitro . These questions were addressed by mutagenesis of the SIMs in ORF61 plasmid constructs or in ORF61 in the context of the VZV genome . We then examined the significance of the ORF61 SUMO-binding capacity for PML NB association and dispersal in differentiated skin cells using the SCID mouse model . In summary , we have demonstrated that ORF61 has a SIM-mediated SUMO-binding capacity which is necessary for its capacity to target and disrupt PML NBs in differentiated skin cells in vivo and that this ORF61 SIM-mediated function is a critical determinant of the pathogenic potential of VZV in skin .
By sequence analysis , we observed that ORF61 has three putative SIMs; two are located in the N-terminus near the RING domain ( designated as SIM-N1 and SIM-N2 ) and a third is in the C-terminus ( designated as SIM-C ) ( Fig . 1A ) . Of interest , one or more putative SIMs are also present in the ORF61 orthologs of other alphaherpesviruses ( Table 1 ) . We first investigated whether the ORF61 SIMs had the functional capacity to mediate ORF61 binding to SUMO . In GST pull down assays , ORF61 expressed in VZV-infected cells and in transiently transfected cells bound to GST-SUMO1 and also to GST-SUMO2 , although less efficiently; non-specific binding to GST was not detected ( Fig . 1B ) . The ORF61 SIMs were then mutated by alanine substitutions in the ORF61 plasmid , resulting in mSIM-N ( both SIM-N1 and SIM-N2 disrupted ) , mSIM-C ( SIM-C disrupted ) , and mSIM-N&C ( all three sites disrupted ) . When these SIM mutants were evaluated with GST pull down assay , the SUMO1 binding capacity of mSIM-N was similar to ORF61 whereas binding of mSIM-C was reduced markedly and mSIM-N&C did not bind to SUMO1 ( Fig . 1C ) . Since ORF61 is expressed and colocalizes with PML NBs at very early times in VZV-infected cells [18] , we investigated the pattern of ORF61 association with PML NBs in transfected melanoma cells . As expected , distinct ORF61 nuclear puncta were observed in cells with low levels of expression and most colocalized with PML NBs ( Fig . 2A , panel I ) . The effect of the ORF61 SIM mutations on association with PML NBs was then evaluated . The localization of mSIM-N puncta with PML NBs was similar to ORF61 ( Fig . 2A , panel II ) ; mSIM-C nuclear puncta were found in only a few cells but also colocalized with PML NBs ( Fig . 2A , panel III ) . In contrast , mSIM-N&C nuclear puncta , also found in only a few cells , did not colocalize with PML NBs ( Fig . 2A , panel IV ) . Examination of an ORF61 mutant that lacked the previously characterized RING/E3 ligase domain [19] , [20] showed that this ΔRING mutant formed large bright nuclear puncta , most of which were associated with PML NBs ( Fig . 2A , panel V ) . Thus , in contrast to the ORF61 SIMs , the ORF61 RING domain was not required for PML NB association . Next , the capacity of ORF61 and SIM mutants to disperse PML NBs was examined . In cells with abundant ORF61 expression , punctate staining was less evident , the protein was distributed more diffusely in the nucleus and the number of PML NBs was decreased significantly compared to cells in the same monolayer that did not express ORF61 ( Fig . 2B , panel I ) . The PML NB frequency in ORF61-positive cells was 0 . 97±0 . 08 ( N = 224 ) , which was a 4 . 7-fold reduction ( p<0 . 0001 ) compared to mock-transfected cells ( 4 . 53±0 . 17 , N = 245 ) ( Fig . 2C ) . PML NBs were also obviously reduced in cells expressing abundant mSIM-N compared to untransfected cells but were better preserved in cells with abundant mSIM-C or mSIM-N&C ( Fig . 2B , panel II to IV ) . The ΔRING mutant had no obvious effect on PML NBs ( data not shown ) , which was consistent with a previous report using Sp100 as a PML NB marker [19] . As determined by quantification of PML NB frequencies in these cells , the capacity of mSIM-N to disperse PML NBs was slightly less than ORF61 , the effect of mSIM-C was reduced significantly and mSIM-N&C was completely ineffective ( Fig . 2C ) . The PML NB frequency in mSIM-N&C-expressing cells was comparable to that in cells expressing ORF61 ΔRING and mock-transfected cells ( Fig . 2C ) . Taken together , these experiments demonstrated that when expressed in the absence of other viral proteins , ORF61 targets and disrupts a majority of PML NBs and that the ORF61 SIMs were necessary and as important as the RING domain for ORF61-mediated dispersal of PML NBs . As a single motif , the C-terminal SIM was the most important for ORF61 SUMO-binding and PML NB dispersal but all three SIMs were required to achieve both the association with PML NBs and their efficient dispersal . To investigate the effect of ORF61 on PML protein levels , ORF61 was co-expressed with each of six PML isoforms in melanoma cells for 24 h and cells were then treated with the proteasome inhibitor MG132 for 6 h . As shown for PML IV , the PML protein level was unchanged between mock-treated and MG132-treated cells ( Fig . 2D , lanes 5 & 6 ) . Similar results were obtained with PML isoforms I-III , V and VI ( data not shown ) . ORF61 accumulated abundantly ( Fig . 2D ) , confirming its rapid turnover [20] . These data suggested that , unlike HSV ICP0 , ORF61 does not degrade PML protein despite its conserved RING domain and E3 ligase activity , which is consistent with the persistence of PML protein in VZV-infected cells [9] , [13] . We next investigated the potential role of the ORF61 SIMs in other ORF61 functions . ORF61 is known to regulate expression of the essential VZV glycoprotein , gE , through a RING domain-dependent mechanism [21] , [37] . However , disrupting the ORF61 SIMs did not alter activation of either the gE or the ORF61 promoters ( Fig . 3A ) . Since VZV inhibits NF-κB activation [38] , we also investigated ORF61 regulation of the NF-κB promoter . Of interest , NF-κB activation by TNF-α was suppressed by ORF61 in a dose-dependent manner and again , this function required the RING domain but not the SIMs ( Fig . 3B and 3C ) . Thus , while the RING domain is essential for all of these ORF61 functions , the SIMs are specifically needed for ORF61 association with and dispersal of PML NBs . These data also suggest that the SIM mutations did not result in aberrant folding of ORF61 protein , which would be expected interfere with these other functions . In order to assess the contributions of the ORF61 SIMs to PML NB association and dispersal in infected cells and to VZV replication , we made three VZV recombinants that contained the same ORF61 SIM mutations as in the ORF61 mutant plasmids . These mutants were designated as pOka-mSIM-N , pOka-mSIM-C and pOka-mSIM-N&C , respectively . Evaluation of the SUMO1-binding capacity of the ORF61 SIM mutant proteins produced in infected cells with the GST pull-down assay showed that interactions of mSIM-C and mSIM-N&C with SUMO1 were almost undetectable whereas mSIM-N binding was diminished only slightly ( Fig . 4A ) , confirming the SIM-mediated binding of ORF61 to SUMO1 as was observed in transfection experiments . Treating cells infected with pOka and ORF61 SIM mutants with MG132 showed that the intracellular processing of the ORF61 SIM mutant proteins was indistinguishable from ORF61 ( Fig . 4B ) , again suggesting that these mutant proteins were not likely to have major structural changes and that their reduced SUMO-binding capacity was attributable to disruption of the functional SIMs . When the effect of the SIM mutations on ORF61 association with PML NBs was examined in melanoma cells at 6 h after infection , all three ORF61 SIM mutants formed minute ORF61-positive nuclear puncta , colocalizing with PML NBs ( Fig . 5A ) . After 24 h , PML NBs were reduced substantially in cells infected with pOka-mSIM-N as well as pOka , whereas disruption of PML NBs was considerably less in pOka-mSIM-C- and pOka-mSIM-N&C-infected cells ( Fig . 5B ) . Quantitative analysis showed that cells infected with pOka-mSIM-C retained 2 . 5-fold more PML NBs compared to pOka-infected cells; the PML NB frequency in cells infected with pOka-mSIM-N&C was equivalent to that in mock-infected cells ( Fig . 5C , left panel ) . Similar patterns were observed in HELFs ( Fig . 5C , right panel ) . These experiments indicate that the capacity of ORF61 to bind to SUMO1 through its SIMs correlates with how efficiently it disrupts PML NBs in VZV-infected cells . The preservation of some colocalization of the mSIM-N&C protein with PML NBs in infected cells , which was not observed under transfection conditions , suggests that another viral factor ( s ) or a virus-induced factor ( s ) may also contribute to ORF61 association with PML NBs . Since neither pOka infection nor ORF61 expression alters the levels of endogenous PML protein ( 9 , Fig . 2D ) , it was expected that PML levels in melanoma cells and HELFs infected with the ORF61 SIM mutants would be unchanged ( Fig . 5D ) . Despite the compromised effect on PML NB dispersal , the growth kinetics of all three ORF61 SIM mutants was similar to pOka in melanoma cells , as determined by infectious focus assay ( Fig . 5E ) . Yields of ORF61 SIM mutants and pOka were also comparable in HELFs ( data not shown ) . Thus , the dispersal of PML NBs mediated through ORF61 SIMs is not required for VZV replication in permissive cells in vitro . When the ORF61 SIM mutants were evaluated in human skin xenografts in SCID mice , the infectious virus yield of pOka-mSIM-N was similar to pOka , while the replication of pOka-mSIM-C was delayed slightly , with virus yields that were lower than pOka at day 10 ( p<0 . 05 ) but had increased by day 21 ( p<0 . 05 ) ( Fig . 6A ) . Infectious virus was recovered from 5 of 6 or all 6 xenografts inoculated with pOka , pOka-mSIM-N or pOka-mSIM-C . In contrast , pOka-mSIM-N&C replication in skin was severely impaired . pOka-mSIM-N&C was recovered from only 4 of 6 inoculated xenografts at day 10 and 3 of 6 at day 21 . Titers from the xenografts that yielded pOka-mSIM-N&C were ∼30-fold lower than pOka at day 10 ( p<0 . 001 ) and ∼500-fold lower at day 21 ( p<0 . 001 ) ( Fig . 6A ) . Analysis of skin tissue sections showed that pOka formed the usual large necrotic VZV lesions that penetrate across the basal cell layer into the dermis whereas pOka-mSIM-N&C lesions were very small and restricted to the epidermis ( Fig . 6B ) . It is important to note that PML NBs were prominent in skin cells in mock-infected xenografts , with the highest frequency being found in cells of the basal layer ( Fig . 6C , upper panels ) . However , many more PML NBs were present in the uninfected cells present within skin xenografts that were infected with pOka or pOka-mSIM-N&C , compared to these mock-infected tissues ( Fig . 6C , middle and lower panels and Fig . S1 ) . Quantification of two skin xenografts that were infected with pOka and two infected with pOka-mSIM-N&C showed 3 . 8-6 . 6 fold increase in PML NB frequency in uninfected cells respectively , compared with the mock-infected xenograft ( Fig . 6D ) ; the PML NB frequencies did not differ significantly between xenografts inoculated with pOka or pOka-mSIM-N&C ( Fig . 6D ) . Since skin cells dramatically up-regulate IFN production in response to VZV infection in vivo and PML is IFN-inducible [6] , [39] , our findings suggest that the innate IFN response increases PML NB numbers in dermal and epidermal cells , reinforcing a pre-existing barrier to VZV spread . The small epidermal lesions and the failure of pOka-mSIM-N&C to create lesions that extend across the basal layer suggested that VZV spread from cell to cell in skin might depend on ORF61 SIM-mediated PML NB disruption . To investigate this possibility , we examined the pattern of ORF61/PML NB association and the PML NB frequencies in skin cells infected with pOka and the ORF61 SIM mutants . ORF61 nuclear puncta that colocalized with PML NBs were evident in the newly infected cells located at the margins of pOka skin lesions ( Fig . 7A , panel I ) . pOka-mSIM-N&C also formed distinct ORF61-positive nuclear puncta but in contrast to pOka , the pattern of their association with PML NBs was heterogenous in individual cells ( Fig . 7A , panel II-IV ) . Skin cells that had mSIM-N&C puncta could be categorized into three groups: those in which either all or some puncta colocalized with PML NBs ( Fig . 7A , panel II and III ) and those in which none colocalized with PML NBs ( Fig . 7A , panel IV ) . When these associations were quantitated in the skin cell nuclei that had ORF61 or mSIM-N&C puncta , colocalization of ORF61 with PML NBs was observed in 100% of pOka-infected cells ( N = 47 ) whereas only 35% of pOka-mSIM-N&C-infected cells ( N = 46 ) showed any colocalization of ORF61 with PML NBs ( p<0 . 0001 ) ( Fig . 7B , left panel ) . Analyzing these data based on total numbers of ORF61 positive puncta in infected cell nuclei showed that 89% of pOka ORF61 puncta ( N = 167 ) were associated with PML NBs compared to only 25% of mSIM-N&C puncta ( N = 202 ) ( p<0 . 0001 ) ( Fig . 7B , right panel ) . Next , the frequency of PML NBs in skin cells infected with pOka and the ORF61 SIM mutants was assessed . ORF23 , a VZV nucleocapsid protein , was used as a marker for infected cells as abundant ORF23 protein indicates a later stage of VZV infection [10] . Only infected cells with intact nuclear membranes were analyzed to exclude those in which VZV had induced necrosis . PML NBs in pOka-infected cells were reduced compared with uninfected cells in the same section ( Fig . 7C , left panels ) , while the frequency of PML NBs in pOka-mSIM-N&C-infected cells was similar to uninfected cells ( Fig . 7C , right panels ) . Quantitative analysis showed that more PML NBs ( 2 . 8-fold higher ) were preserved in pOka-mSIM-N&C-infected cells compared to pOka-infected cells ( pOka , 0 . 71±0 . 07 , N = 153; pOka-mSIM-N&C , 1 . 97±0 . 17 , N = 108; p<0 . 0001 ) ( Fig . 7D ) . The PML NB frequency in pOka-mSIM-C-infected cells ( N = 143 ) was equivalent to pOka-infected cells ( Fig . 7D ) . Both the detection of PML expression and the frequency of PML NBs were similar in the uninfected cells within xenografts infected either with pOka or pOka-mSIM-N&C ( Fig . 6C and 7E ) . This finding makes it unlikely that the difference in PML NB frequency in skin cells infected by these two viruses reflects a variation of PML NB frequencies between individual xenografts . Taken together , these results demonstrate that ORF61 SIMs determine the capacity of VZV to target and disrupt PML NBs in skin cells in vivo and indicate that VZV spread in skin depends on this function . Since these experiments in skin xenografts showed that pOka-mSIM-N&C had a growth deficiency in vivo which was not detectable in vitro and IFN-α , which induces PML , is present in VZV-infected xenografts in vivo but is not produced by VZV-infected cells in vitro [39] , [40] , we evaluated the replication of pOka and pOka-mSIM-N&C in cultured cells treated with IFN-α . The number of PML NBs was increased by IFN-α treatment and as expected , the PML NBs were disrupted by pOka but not by pOka-mSIM-N&C ( Fig . 8A and 8B ) . Nevertheless , IFN-α treatment had a comparable effect on reducing pOka and pOka-mSIM-N&C titers in melanoma cells and HELFs ( Fig . 8C ) . However , even with IFN treatment , PML NB frequencies in melanoma cells remained significantly lower than frequencies observed in skin cells in vivo ( Fig . 8A ) . That pOka-mSIM-N&C was not differentially inhibited in IFN-α treated cells when compared to pOka in vitro , in contrast to the severely impaired growth of pOka-mSIM-N&C in skin supports the significance of the very high frequencies of PML NBs that are present in dermal and epidermal cells for limiting the spread of VZV in vivo . Since VZV , like other viruses , replicates more efficiently in proliferating cells , another reason for the defective growth of the ORF61 SIM mutants in skin compared to cultured cells might be fewer proliferating cells in vivo . Based on the expression of Ki67 , a proliferation marker [41] , most HELFs were proliferating under standard cell culture conditions whereas only a few skin cells from mock- or VZV-infected xenografts were Ki67-positive ( Fig . 9A and 9B ) . No significant difference in Ki67 expression was observed between uninfected cells in pOka- or pOka-mSIM-N&C-infected xenografts ( Fig . 9A ) . To further investigate whether the ORF61 SIMs confer an advantage for VZV replication in non-proliferating cells , the growth kinetics of pOka and pOka-mSIM-N&C were compared in serum-starved HELFs . The percentage of Ki67-positive HELFs was reduced to a minimal level by serum starvation ( Fig . 9B ) . However , titers of pOka-mSIM-N&C in serum-starved HELFs were indistinguishable from pOka titers ( Fig . 9C ) . These results suggested that the difference in the importance of ORF61 SIMs for VZV replication in skin compared with cultured cells did not reflect an advantage for VZV replication in non-proliferating cells that depends upon the ORF61 SIMs .
This study provides the first direct evidence of the importance of PML NB disruption for the efficient replication of a viral pathogen in differentiated human cells located within their usual tissue microenvironment in vivo and of the requirement for the SUMO-binding capacity of a viral protein , VZV ORF61 , to counteract the innate antiviral control mediated by PML NBs . This function is critical in VZV skin infection because the persistence of the virus in the human population depends upon its capacity to produce cutaneous lesions that contain high concentrations of infectious virus particles for transferring to susceptible individuals [10] . Notably , we found that PML NBs were highly abundant in human skin cells , especially in the basal cell boundary between the epidermis and the dermis . In the absence of functional ORF61 SIMs , these PML NBs were preserved and VZV infection was severely impaired . Thus , our experiments indicate that VZV pathogenesis in skin requires ORF61 association with and dispersal of PML NBs and that this modification of host cell nuclear structures depends upon ORF61 SIMs . We suggest that PML NB-mediated control of VZV replication must be counteracted by ORF61 in order to produce the characteristic cutaneous lesions of varicella and herpes zoster and that ORF61 SUMO-binding capacity is necessary for this essential phase of the VZV life cycle in the human host . More generally , the finding that ORF61 has functional SIMs that modulate PML NBs provides new evidence that non-covalent SIM-SUMO interactions can alter the structure and dynamics of PML NBs [32]-[36] . Importantly , the evaluation of VZV ORF61 SIM-deficient mutants in skin xenografts shows that this SIM-dependent effect on PML NBs occurs in differentiated cells within tissues in vivo . To our knowledge , ORF61 is the first example showing that a viral RING finger protein has functional SIMs , like those that are present in cellular RING finger proteins , such as RNF4 , which is involved in PML degradation [34] . Of interest , the less conserved SIM-C motif ‘TIDL’ , which has also been identified in other SUMO-binding proteins [42] , appeared to be more critical than the other two highly conserved ORF61 SIMs . One possible explanation is the access of SIM-N1 and SIM-N2 sites to SUMO1 might be hindered since they are in a proline-rich region and proline may interrupt the β-strand conformation required for the interaction . Our analysis of ORF61 SIM-C provides additional evidence that the less conserved ‘TIDL’ motif can be a functional SIM and demonstrates its role in a viral SUMO-binding protein . In previous work , we found that VZV infection in skin causes a dramatic upregulation of IFN-α in uninfected cells surrounding VZV lesions and that this response is critical for controlling infection , as shown by enhanced VZV replication and extensive skin lesion formation when the IFN pathway is blocked [39] . We now report that PML NBs are also increased substantially from an already high baseline when skin is infected with VZV . Since PML expression is regulated by IFN , these observations are further evidence of the significant role of IFN-mediated innate immunity in skin [39] . Our data indicate that PML NBs are an important mechanism by which IFN control of VZV infection is achieved in vivo and that ORF61 SIM-mediated targeting and dispersal of PML NBs is necessary to counteract this innate response during VZV pathogenesis . These observations also help to explain why antiviral immunity was defective in a PML knockout mouse model [8] . The finding that the growth of pOka-mSIM-N&C was severely defective in skin xenografts , whereas it was indistinguishable from pOka in cultured cells , illustrates that the functional significance of the ORF61 SIMs could only be demonstrated when assessed in the context of VZV replication in differentiated host cells in vivo . The very high frequencies of PML NBs in differentiated skin cells in vivo suggests that the abundance of PML NBs determines whether the ORF61 SIM-mediated effect on PML NBs is necessary for VZV to achieve efficient replication . PML NB frequencies could not be enhanced to these levels in cultured cells even with IFN treatment in vitro . Since PML NBs are also numerous in the nuclei of both neurons and satellite cells [9] , it will be of interest to investigate the contribution of ORF61 SIM-dependent PML NB disruption in VZV neuropathogenesis . The need to define functional requirements for disrupting PML NBs in vivo was also evident from the fact that mutating SIM-C alone was sufficient to substantially affect ORF61 dispersal of PML NBs in cultured cells whereas all three SIMs were necessary for PML NB disruption in skin . Another difference between cell culture and skin cells in vivo was that the mutant ORF61 lacking all SIMs retained some capacity to associate with PML NBs in cultured cells but the association was reduced significantly in skin cells . These observations suggest another VZV protein ( s ) or VZV-activated cellular factor ( s ) may bind to ORF61 and facilitate ORF61 targeting of PML NBs in vitro but has a limited role in vivo . We speculate that the differential availability or expression of these factors in skin cells compared to cultured cells might have contributed to the differential requirement of the three ORF61 SIMs to VZV replication in vivo and in vitro . These questions warrant further investigation , particularly in the different types of human cells that are targeted during VZV pathogenesis . In HSV-infected cells , the ICP0 RING domain acts as an ubiquitin E3 ligase and triggers proteasome-dependent PML degradation and PML NB disruption [14] , [15] . Our experiments confirmed that the RING domain was also necessary for ORF61-induced PML NB dispersal but we found that it was dispensable for ORF61 association with PML NBs . In contrast , the ORF61 SIMs were required for both PML NB association and disruption , indicating that PML NB dispersal by ORF61 is a two-step process: the ORF61 SIMs first recognize sumoylated PML protein in PML NBs and the RING domain is needed to execute dispersal . As was consistent with the persistent PML protein expression in VZV-infected cells [9] , [13] , ORF61 expression as a single protein caused PML NB disruption but PML protein was not degraded; therefore , we speculate that the ORF61 RING domain has functions other than E3 ligase that are necessary for its contribution to PML NB disruption . For example , since RING domains mediate protein-protein interactions [43] , the RING domain in ORF61 may form a complex with other RING finger proteins in the nucleoplasm and thereby dislodge ORF61 SIM-bound PML proteins from PML NBs . However , during other cellular events in which ORF61 is involved , it is quite possible that the ORF61 RING domain acts as an E3 ligase and mediates its substrate degradation , since ORF61 RING domain is known to exhibit E3 ligase activity in vitro [19] , [20] . It is known that all alphaherpesviruses encode RING finger proteins related to ICP0 and ORF61 and most of them have been shown to target PML NB components [44] . Of interest , our sequence analyses indicate the presence of conserved SIM ( s ) in these RING finger proteins . Taken together , we propose that the interaction of these SIM-containing viral proteins with sumoylated PML is conserved and essential for alphaherpesviruses to target PML NB structures for dispersal whereas the RING domain in the orthologs has variable functions , and these functions determine whether , or the extent to which , PML protein is degraded . Our recent experiments in VZV-infected skin cells and neurons showed that PML protein that persists can form a novel class of PML nuclear cages with the capacity to sequester newly synthesized capsids and constitute an intrinsic anti-VZV defense [9] . Viruses like HSV have the capability to completely overcome the PML-mediated intrinsic barrier as they not only disrupt PML NBs but also efficiently degrade PML protein; therefore these viruses may have an advantage against the IFN defense that is triggered and amplified within infected tissues . We speculate that this difference in the effect on PML protein , resulting from the different functional capacities of ICP0 and ORF61 , may help to explain why HSV reactivates much more frequently than VZV in the naturally infected host [45] . Of note , the gammaherpesvirus KSHV-encoded LANA2 protein disperses PML NBs by a SIM-dependent process in cultured cells although it does not have a RING domain [36] . We propose that SIM-dependent disruption of the architecture of PML NBs might be a conserved mechanism among herpesviruses , independent of the fate of PML protein . As noted , many viruses encode gene products that have been shown to modify PML NBs in cultured cells [7] . However , to define the specific importance of PML NB modulation for viral replication and pathogenesis , viral mutants that are disabled only in this function are needed . HSV ICP0 RING domain mutant viruses do not disrupt PML NBs but the RING domain is also essential for other ICP0 functions , including viral gene expression , innate immune evasion and virion formation [46]-[48] . Therefore , it is difficult to attribute the replication defect of the HSV ICP0 RING domain mutant to a single function . Experiments with the ORF61 SIM mutants avoided these concerns because the RING domain was intact and other functions of ORF61 , including viral gene regulation and NF-κB suppression and ORF61 protein stability were preserved . These ORF61 functions , such as NF-κB regulation , are also likely to be important contributions of this multi-functional protein to VZV pathogenesis and warrant further investigation [23] . Nevertheless , since ORF61 SIMs might target many more sumoylated proteins in PML NBs or other nuclear compartments , other ORF61 SIM-mediated interactions could contribute to VZV replication in vitro and to pathogenesis in vivo in addition to their role in PML NB disruption . In summary , we have demonstrated that VZV ORF61 interacts with SUMO1 via three conserved SIMs . The interaction was essential for the association of ORF61 with PML NBs and ORF61-mediated PML NB disruption in infected skin cells in vivo . VZV replication and spread in skin tissues in vivo depended on this ORF61 SIM-mediated function . Our data support the conclusion that PML NBs provide an intrinsic host defense against VZV infection in skin , which is an essential stage in VZV pathogenesis during primary and recurrent infection of the human host and we have elucidated the ORF61 SIM-dependent mechanism that is used by VZV to counteract these antiviral nuclear structures .
The human fetal tissues for SCID xenograft studies were obtained from Advanced Bioscience Resources , Inc . ( Alameda , CA ) in accordance with state and federal regulations for tissue acquisition for biomedical research , in accordance with FDA 21 CFR Part 1271 GTP ( Good Tissue Practices ) , UAGA and NOTA . Human melanoma cells ( a tumor cell line ) and primary human embryonic lung fibroblasts ( HELF ) were derived and used as described previously [23] . Animal protocols complied with the Animal Welfare Act and were approved by the Stanford University Administrative Panel on Laboratory Animal Care . Human melanoma cells ( a tumor cell line ) and primary human embryonic lung fibroblasts ( HELF ) were derived and used as described previously [23] . The PML knock-down melanoma cell line ( siPML ) was a gift from Prof . Saul Silverstein , University of Columbia and was used as the negative control in PML western blot experiments . Proteasome inhibitor MG132 ( Calbiochem ) was diluted in DMSO and used at 10 µM; DMSO was used for mock treatment . Human tumor necrosis factor α ( TNF-α ) ( Biovision ) was used at 20 ng/mL for activation of the NF-κB pathway . Human IFN-α Hu-IFN-α2b ( PBL InterferonSource ) was used at 1000 international units ( IU ) per mL for upregulation of PML in melanoma cells and HELFs . Rabbit polyclonal antibodies against ORF61 and IE63 were kindly provided by Prof . Paul Kinchington , University of Pittsburgh; rabbit polyclonal anti-ORF23 was generated by rabbit immunization with purified protein [49] . PML antibodies were mouse monoclonal anti-PML ( PG-M3 ) and rabbit polyclonal anti-PML ( Santa Cruz Biotech ) . Other antibodies included mouse monoclonals against SUMO1 ( Zymed ) , VZV gE ( Chemicon ) , and α-Tubulin ( Sigma ) and rabbit polyclonal against Ki67 ( Abcam ) . The ORF61 expression plasmid pcDNA-ORF61 was constructed by cloning full-length ORF61 coding sequence into pcDNA3 . 1 ( + ) . ORF61 DNA was amplified by PCR with forward primer ( ORF61-F ) 5′-GATCAAGCTTATGGATACCATATTAGCGGG-3′ containing a HindIII site and reverse primer ( ORF61-R ) 5′-GCGAATTCCTAGGACTTCTTCATCTTGT-3′ containing an EcoRI site . The PCR product was digested with HindIII and EcoRI and ligated to HindIII&EcoRI-digested pcDNA3 . 1 ( + ) to generate pcDNA-ORF61 . The ORF61 ( ΔRING ) fragment , from which amino acids 1-60 were deleted , was amplified by PCR with the forward primer ( ΔRING-F ) 5′-GATCAAGCTTATGGTGCAATCCATCCTGCATAAG-3′ containing the HindIII site and the reverse primer ( ORF61-R ) , and ligated to HindIII/EcoRI-digested pcDNA3 . 1 ( + ) to generate pcDNA-ORF61 ( ΔRING ) . pcDNA constructs expressing ORF61 SIM mutants ( pcDNA-mSIM ) were constructed by substituting the four hydrophobic core residues of SIMs with alanines in the context of pcDNA-ORF61 ( SIM-N1 , IDIL to AAAA; SIM-N2 , IDLL to AAAA; SIM-C , TIDL to AAAA ) . Primers used for SIM mutagenesis are: SIM-N1: 5′-CATTTGAAGATTCCGCTGCCGCTGCACCGGGAGATG-3′ and 5′-CATCTCCCGGTGCAGCGGCAGCGGAATCTTCAAATG-3′; SIM-N2: 5′-CCGGGAGATGTCGCAGCTGCTGCGCCACCAAGCCCA-3′ and 5′-TGGGCTTGGTGGCGCAGCAGCTGCGACATCTCCCGG-3′; SIM-C: 5′-CGATGCTTAGCAGCAGCCGCGACATCTGAGTCTGA-3′ and 5′-TCAGACTCAGATGTCGCGGCTGCTGCTAAGCATCG-3′ PML mammalian expression plasmids were kind gifts from Prof . Peter Hemmerich , Leibnitz-Institute of Age Research , Germany and Prof . Annie Sittler , Universite Pierre et Marie Curie , France . The NF-κB reporter plasmid was gift from Prof . Dingxiang Liu in Institute of Molecular and Cell Biology , Singapore . The gE promoter luciferase construct was as described [50] . All constructs were confirmed by nucleotide sequencing ( Elim Biopharm , Inc . , Hayward , CA ) . ORF61-transfected cells or VZV-infected cells grown on 10 cm dishes were lysed in 100 ul high salt buffer ( 20 mM HEPES [pH 7 . 2] , 450 mM NaCl , 1 . 5 mM MgCl2 , 0 . 5% NP-40 , 20% Glycerol ) supplemented with EDTA-free protease inhibitor cocktail ( Roche ) . Insoluble proteins were removed by centrifugation . The supernantant was combined with 400 ul NaCl-free buffer ( 20 mM HEPES [pH 7 . 2] , 1 . 5 mM MgCl2 , 0 . 5% NP-40 , 20% glycerol ) and centrifuged again before GST pull down . GST or GST-SUMO1/2 proteins ( 5 µg ) ( Boston Biochem ) were added to the pre-cleared lysate and rotated at 4°C for 2 h . Glutathione Sepharose beads ( GE Healthcare ) were added subsequently and incubated at 4°C for 2 h . Beads were washed 4 times with the binding buffer ( 20 mM HEPES [pH 7 . 2] , 90 mM NaCl , 1 . 5 mM MgCl2 , 0 . 5% NP-40 , 20% glycerol ) and eluted with SDS sample buffer . The ORF61 protein that bound to the beads was analyzed by Western blot using the ORF61 antibody and the GST and GST-SUMO1/2 proteins that bound to the beads was analyzed by amido black staining . Recombinant viruses were generated using cosmids derived from pOka [51] . The entire pOka genome is covered by four overlapping cosmids designated Fsp73 ( pOka nucleotides [nt] 1 to 33128 ) , Spe14 ( pOka nt 21795 to 61868 ) , Pme2 ( pOka nt 53755 to 96035 ) , and Spe23 ( pOka nt 94055 to 125124 ) . The ORF61 coding region is located in the unique long region in the cosmid Spe23 ( pOka nt 103045-104445 ) . A 4 . 8 kb PstI-PmlI fragment containing full length ORF61 gene was subcloned into pCR4-TOPO ( Invitrogen , Inc . ) to make pCR4- ( PstI/PmlI ) , in which ORF61 SIM mutations were generated using two round PCR method . The PstI-PmlI fragments containing ORF61 SIM mutations were cloned into pLit ( ORF59-65 ) [23] . Mutant Spe23 cosmids were made by ligating the NheI-AvrII fragment from pLit ( ORF59-65 ) to Spe23 digested with NheI/AvrII . Recombinant viruses , designated pOka-mSIM-N , pOka-mSIM-C , and pOka-mSIM-N&C , were isolated by transfection of melanoma cells with the mutated Spe23 cosmid and the other three intact cosmids , Fsp73 , Spe14 , and Pme2 . Genomic DNA was extracted from virus-infected melanoma cells or HELFs with DNAzol reagent ( Invitrogen , Carlsbad , CA ) . A PCR fragment covering the mutated region was amplified from the genomic DNA using Taq polymerase ( Invitrogen ) and sequenced ( Elim Biopharm , Inc . , Hayward , CA ) to confirm mutations . Melanoma cells ( 106/well ) were seeded in a 6-well plate , infected with 1×103 PFU/well at day 0 , and cultured for 5 days . On each day , cells from one well were trypsinized , centrifuged , and resuspended in 1 mL of culture medium . The infected cells were serially diluted 10-fold , and 0 . 1 mL was added to melanoma cells in 24-well plates in triplicate . Cells were fixed in 4% paraformaldehyde and stained with polyclonal anti-VZV human immune serum and secondary anti-human biotin ( Vector Lab , Burlingame , CA ) . The staining was developed with the Fast Red substrate ( Sigma ) . Statistical analysis of growth kinetics was done by the Student's t test . Skin xenografts were made in homozygous CB-17scid/scid mice , using human fetal tissue supplied by Advanced Bioscience Resources ( ABR , Alameda , CA ) according to federal and state regulations; the methods used to engraft and infect the skin xenografts were as described previously ( 25 ) . Animal use was in accordance with the Animal Welfare Act and approved by the Stanford University Administrative Panel on Laboratory Animal Care . pOka and ORF61 SIM mutant viruses were passed three times in primary HELF and titered before inoculation . Skin xenografts were harvested at day 10 and 21 and titers were determined by infectious focus assay . DNA was extracted from skin tissues with proteinase K and phenol chloroform ( Invitrogen , Carlsbad , CA ) . PCR and sequencing were performed to confirm the expected mutations . Cells seeded on 12 mm glass cover slips were fixed with 4% paraformaldehyde for 15 min and permeablized with 0 . 5% Triton-X100 for 10 min . For dual staining of ORF61 and PML proteins , cells were incubated with ORF61 rabbit polyclonal antibody ( 1∶200 dilution ) and PML mouse monoclonal antibody PG-M3 ( 1∶50 dilution ) in blocking buffer ( PBS with 5% fetal bovine serum ) at room temperature ( RT ) for 1 h , followed by incubation with Alexflour 488 donkey anti-mouse immunoglobulin ( Invitrogen ) and Texas Red-conjugated donkey anti-rabbit immunoglobulin ( Jackson ImmunoResearch ) for 30 min . Cell nuclei were stained with 2 µg/mL Hoechst33342 for 10 min after secondary antibody incubation . All images were obtained with a SP2 Leica confocal microscope . Images were taken in a fixed setting with the 63x objective in PML NB number quantification experiments , and with the 100x objective to quantify the association between ORF61 puncta and PML NBs in infected skin cells . For skin experiments , sections ( 5 µm ) were made from formalin-fixed , paraffin-embedded skin tissues . After deparaffinization and rehydration , sections were treated with citrate-based antigen unmasking solution ( Vector labs ) and stained with specific antibodies ( anti-PML [PG-M3] , 1∶10; anti-ORF61 , 1∶25; anti-ORF23 , 1∶100 ) . Skin sections were examined for cell proliferation using Ki67 antibody ( Abcam ) . Skin sections were stained with VZV gE antibody ( Chemicon , 1∶2000 ) and IHC immunoperoxidase secondary detection system ( Chemicon ) . Staining was developed with Vector VIP substrate kit and methyl green counterstain ( Vector labs ) . Melanoma cells or HELFs were seeded at 5×105 cells/well in 6-well plates . Cells in each well were mock-treated or treated with 1000 IU of Hu-IFN-α2b ( PBL InterferonSource ) for 24 h prior to inoculation . Cells were inoculated with 500 PFU/well of either pOka or pOka-mSIM-N&C . The media was aspirated at 2 h post-infection to remove the inoculum and replaced with media with or without 1000 IU of IFN-α2b . Virus titers from 2 h to 36 h post-inoculation were determined by infectious focus assay on melanoma cells . HELFs were seeded at 5×105 cells/well in 6-well plates and at 2×105 cells/well in 2-well chamber slides . Cells were starved in serum-free culture medium for 48 hours and then inoculated with 50 PFU/cm2 of pOka or mSIM-N&C . At 2 hours post infection , the inoculum was removed and the medium was replaced with normal medium or serum-free medium . Virus titers in cells recovered from the 6-well plates at intervals of 2 h to 36 h post-inoculation were determined by infectious focus assay on melanoma cells . Cells on chamber slides were fixed with paraformaldehyde and stained with VZV gE ( Chemicon ) and Ki67 ( Abcam ) antibodies . All transfections were performed with Lipofectamine 2000 ( Invitrogen , Carlsbad , CA ) following the manufacturer's instructions . The transfected cells for immunofluorescence were fixed at 24 h post-transfection . To prepare heavily infected cells for Western blot analysis of PML proteins , melanoma cells and HELFs were inoculated with pOka or ORF61 SIM mutants at a ratio of 1 infected cell:100 uninfected cells . At 24 h post-inoculation , each cell monolayer was resuspended by trypsinization and re-plated on the same dish and left for another 24 h . The percentage of infected cells was examined by immunofluorescence with ORF23 antibody and was >90% at the time of harvest . To prepare infected cells for immunofluorescence microscopy , cells growing on glass coverslips were inoculated with pOka or ORF61 SIM mutants at a ratio of 1 infected cell: 1000 uninfected cells and infected for 24 h . Transfected or infected cells growing on 6-well plates were lysed in high salt buffer as described above . Proteins were separated on SDS-PAGE gels and transferred to polyvinylidene difluoride ( PVDF ) membrane ( Millipore , Bedford , MA ) with a semidry transfer cell ( Bio-Rad ) . The membranes were blocked for 1 h in 5% non-fat milk in PBST ( 1x PBS plus 0 . 1% Tween-20 ) , incubated with primary antibody at RT for 2 h , washed three times with PBST , incubated with horseradish peroxidase-conjugated rabbit or mouse immunoglobulin ( Amersham ) at RT for 1 h , and washed three times with PBST . Proteins were detected using the enhanced chemiluminescence plus detection system ( Amersham ) . Melanoma cells were seeded in 24-well plates one day before transfection . Three independent transfections were preformed for each experiment . In the gE and ORF61 promoter assay , 900 ng of the promoter construct was transfected to melanoma cells with 100 ng of pcDNA-ORF61 or pcDNA-mSIM or empty vector for 24 h . In NF-κB experiments , 500 ng of NF-κB reporter plasmid was cotransfected with 250 ng of pcDNA-ORF61 or pcDNA-mSIM or empty vector; at 24 h post-transfection , cells were treated with 20 ng/mL TNF-α for 6 h before cell lysis and luciferase assay . In the NF-κB dose-dependent experiment , increasing amounts of pcDNA-ORF61 or pcDNA-ORF61 ( ΔRING ) ( 10 ng , 50 ng , and 250 ng ) were used and the total DNA was brought to 850 ng with empty vector . In all transfections , 0 . 07 ng of the plasmid pRL-TK ( – ) in which TK promoter has been removed was included to normalize the transfection efficiency . Luciferase assays were performed using the dual luciferase kit ( Promega ) according to the manufacturer's recommendations . VZV ORF61: NP_040183; VZV ORF23: AAY57709 . 1; VZV IE63: Q77NN7; VZV gE: Q9J3M8 . | PML nuclear bodies ( PML NBs ) are spherical nuclear structures that are present in most human and animal cells . These bodies contribute to anti-viral defense and therefore many viruses have developed strategies to disrupt them . This interaction has been demonstrated for a number of viruses in cultured cells but little is known about these processes in differentiated cells within human tissues . Varicella-zoster virus ( VZV ) is a human alphaherpesvirus that causes chicken pox and shingle lesions in skin . Here we show that VZV disrupts PML NBs in epidermal and dermal cells in skin tissues implanted subcutaneously in immunodeficient mice . We found that PML NB dispersal is mediated by VZV ORF61 protein and is required for VZV cell to cell spread and lesion formation in skin . The ability of ORF61 to disrupt PML NBs depends on its capacity to bind to SUMO1 protein , which is conjugated to PML and other proteins within PML NBs . To our knowledge , our study provides the first evidence of PML NB modification through the SUMO-binding function of a viral protein , VZV ORF61 , and the importance of this molecular mechanism for virus-induced PML NB disruption in differentiated cells infected within their tissue microenvironment in vivo . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biology"
] | 2011 | Disruption of PML Nuclear Bodies Is Mediated by ORF61 SUMO-Interacting Motifs and Required for Varicella-Zoster Virus Pathogenesis in Skin |
Nrf2 , a transcriptional activator of cell protection genes , is an attractive therapeutic target for the prevention of neurodegenerative diseases , including Alzheimer’s disease ( AD ) . Current Nrf2 activators , however , may exert toxicity and pathway over-activation can induce detrimental effects . An understanding of the mechanisms mediating Nrf2 inhibition in neurodegenerative conditions may therefore direct the design of drugs targeted for the prevention of these diseases with minimal side-effects . Our study provides the first in vivo evidence that specific inhibition of Keap1 , a negative regulator of Nrf2 , can prevent neuronal toxicity in response to the AD-initiating Aβ42 peptide , in correlation with Nrf2 activation . Comparatively , lithium , an inhibitor of the Nrf2 suppressor GSK-3 , prevented Aβ42 toxicity by mechanisms independent of Nrf2 . A new direct inhibitor of the Keap1-Nrf2 binding domain also prevented synaptotoxicity mediated by naturally-derived Aβ oligomers in mouse cortical neurons . Overall , our findings highlight Keap1 specifically as an efficient target for the re-activation of Nrf2 in AD , and support the further investigation of direct Keap1 inhibitors for the prevention of neurodegeneration in vivo .
The transcription factor Nrf2 ( nuclear factor E2-related factor 2 ) targets cellular defence genes containing antioxidant response elements ( ARE ) , which include antioxidant enzymes ( glutamate cysteine ligase; GCL ) , drug metabolising enzymes ( cytochrome P450s , glutathione S-transferases; GSTs ) , molecular chaperones , DNA repair enzymes and proteasome subunits[1] . Activation of these protective genes in response to Nrf2 enables the cell to maintain redox balance and to remove damaged proteins under conditions of oxidative and xenobiotic stress . Such cellular stress is a key feature of several neurodegenerative diseases . Markers of oxidative damage are increased in the brains of Alzheimer’s disease ( AD ) [2 , 3] , Parkinson’s disease ( PD ) [4–6] , Huntington’s disease ( HD ) [7] and , in CSF , of Amyotrophic Lateral Sclerosis ( ALS ) patients[8] . Some evidence also suggests that AD , PD and ALS patients have reduced xenobiotic metabolism [9] and that the AD-causing Aβ42 peptide may act as a xenobiotic[10] . As Nrf2 is inhibited in several neurodegenerative diseases[1 , 11] , including AD[11] and ALS[12] , as well as in APP/PS1 mutant mouse models of AD[13] , deficits in this important cellular protection pathway may , in part , explain the cellular damage associated with these conditions . Conversely , Nrf2 over-expression protects against toxicity induced by the Aβ42 peptide in mammalian cells[13 , 14] , and prevents neuronal pathology in mouse models of ALS[15] , PD[16] and AD[17] . Activation of Nrf2 is , therefore , increasingly implicated as an attractive target for the prevention of several neurodegenerative conditions . Potent activators of Nrf2 have been developed[18] , and confer protection against chemically-induced neurotoxic insults[19] and improve memory deficits in mouse models of AD[20] . Many of these Nrf2 activators , however , have been reported to exert toxicity due to off-target effects[21][22] . Additionally , unregulated activation of Nrf2 can have detrimental consequences , with prolonged , ubiquitous activation shortening lifespan in Drosophila[23] , and mutations in the Nrf2 inhibitor Keap1 ( kelch-like ECH-associating protein 1 ) causing cancer in humans[24] . Hence , a better understanding of the mechanisms by which Nrf2 function is inhibited in neurodegenerative disease may enable the design of drugs targeted for the prevention of these conditions with minimal side-effects . Nrf2 activity is tightly regulated by two main inhibitors , Keap1 and GSK-3 ( glycogen synthase kinase-3 ) [25] ( S1 Fig ) . GSK-3 plays a well-established role in the pathogenesis of AD[26] , and GSK-3 inhibitors , including lithium , prevent pathology in animal models of AD[26 , 27] , ALS[28] and PD[29] . Emerging evidence also suggests that inhibiting Keap1 can ameliorate neuronal degeneration , with Keap1 RNA interference ( RNAi ) protecting against Aβ42[30] and MPTP[31] toxicity in cells , and heterozygous loss of Keap1 protecting against neuronal pathology in Drosophila models of PD[32 , 33] . Both GSK-3 and Keap1 may , therefore , serve as valid candidates for mediating the inhibition of Nrf2 in neurodegenerative diseases , and their inhibition may enable the prevention of Nrf2 deficits , neuronal stress and degeneration specifically in these conditions . A comparative analysis of the efficiency of their inhibition in rescuing Nrf2 deficits in neurodegenerative diseases specifically , however , is required . We aimed therefore to investigate the role of Nrf2 in mediating the protective effects of GSK-3 and Keap1 inhibition against Aβ42 toxicity . Using an inducible Drosophila model of AD[34] , we confirmed that Aβ42 inhibits activity of the fly homolog of Nrf2 ( cap-n-collar isoform C , cncC[35] ) in neurons , consistent with previous findings in mice[13] . Both inhibition of GSK-3 , using lithium , and loss-of-function mutations in Keap1 protected against Aβ42 toxicity in this model . We found , however , that neuronal protection in response to Keap1 inhibition correlates with the rescue of Aβ42-induced Nrf2 defects , whereas lithium treatment appears to exert neuro-protection independently of Nrf2 . Consistent with Nrf2 activation , Keap1 inhibition prevented the enhanced sensitivity of Aβ42-expressing flies to xenobiotic stress , but exerted minimal protection against oxidative damage in comparison to lithium treatment . Combined modulation of Keap1 and lithium additively protected against Aβ42 toxicity , in comparison with either treatment alone , but did not improve their respective effects on xenobiotic and oxidative damage . This further supports the divergent beneficial effects of these manipulations . Down-regulation of Keap1 alone additionally protected against Aβ42 toxicity by mechanisms correlating with enhanced degradation of Aβ42 peptide . Overall our data highlight Keap1 as an efficient target for the amelioration of Nrf2 deficits and protection against neuronal damage in AD . Finally , we show for the first time that a newly-described direct inhibitor of the Keap1-Nrf2 protein-protein interaction[22] can indeed protect against the synapto-toxicity of naturally-derived Aβ oligomers in primary mouse neuronal cultures . As current Nrf2 activators may exert toxicity due to off-target effects , our data suggest that blocking the specific interaction of Nrf2 with Keap1 may provide an exciting new avenue for the discovery of disease-modifying treatments for AD , and potentially other neurodegenerative conditions .
Over-expression of APP in mice inhibits expression of Nrf2 target genes[13] . Comparing expression , at equivalent levels[36] , of various Aβ species in adult fly neurons , we have shown that the presence of aggregating Aβ42 peptides may specifically mediate this effect . Using Nrf2/cncC reporter flies ( gstD1-GFP; [35] ) , RU486 induction of a single , site-directed , copy of Arctic mutant Aβ42 ( ArcAβ42 ) , but not WT Aβ40 or Aβ42 peptides , significantly reduced Nrf2/cncC activity compared to un-induced gstD1-GFP-expressing controls ( Fig 1A and 1B ) . Only ArcAβ42 flies develop pathological phenotypes under these expression conditions . Moreover , inducing high levels of WT Aβ42 using an independent random-insertion line , which does cause neuronal decline [37] , also significantly reduced Nrf2/cncC activity ( Fig 1C ) . This suggests that the inhibition of Nrf2 may be related to the concentration-dependent ability of Aβ42 to aggregate and exert toxicity . Consistent with a specific inhibition of neuronal Nrf2/cncC activity in response to Aβ42 expression , ArcAβ42 reduced GFP levels in the heads , but not bodies ( S2A Fig ) , of RU486-induced flies . Microarray analyses further revealed several pathways and processes that might mediate toxicity in response to ArcAβ42 expression in fly neurons ( Fig 1D , Arc Aβ42 ) . Comparing this data-set to a previously published microarray analysis using flies ubiquitously over-expressing Nrf2/cncC ( Fig 1D , cncC[38] ) , we found that Aβ42 expression induced reciprocal effects on gene ontology ( GO ) categories normally regulated by the Nrf2 pathway . Aβ42 activated processes that are down-regulated , and inhibited processes that are up-regulated , by cncC ( Fig 1D & S2B Fig ) , further confirming its suppressive effect on Nrf2 signalling . Overall these findings suggest that the inhibition of Nrf-2 in AD is robustly conserved in Drosophila , and that this deficiency is caused by the Aβ42 peptide directly . Moreover , as we also observed a suppression of Nrf2/cncC in flies over-expressing human 0N3R tau or the ALS-related C9orf72 ( GR ) 100 di-peptide repeat protein ( DPR ) [40] in adult neurons ( S3A and S3B Fig ) , our findings suggest that the accumulation of toxic proteins may lead to the generalised defect in Nrf2 signalling observed in neurodegenerative diseases . Drosophila thus provide an excellent context for further analysis of the mechanisms by which Aβ42 regulates Nrf2 in vivo . As cncC , MafS and Keap1 mRNA transcripts were unaltered by Aβ42 in our flies ( S3C Fig ) , we hypothesised that Aβ42 modulates Nrf2 activity by altering its biochemical interaction with upstream regulators . Keap1 is a well-known negative regulator of Nrf2 activity in response to oxidative and xenobiotic stressors[41] , but its role in mediating damage in response to neurotoxic proteins has not been widely studied in vivo . Genetically reducing Keap1 , alone , extended lifespan ( Fig 2A ) and rescued neuronal-specific motor defects in both WT ( S4 Fig ) and ArcAβ42-expressing flies ( Fig 2B ) using two independent alleles , Keap1del [42] and Keap1EY5[35] . GSK-3 has more recently been shown to inhibit Nrf-2[43] , independently of Keap1 , but has long been implicated in the pathogenesis of several neurodegenerative diseases including AD[26] . We have previously shown that lithium , a well-described GSK-3 inhibitor[44–46] , prevents Aβ42 toxicity in our fly model [34] . Thus both GSK-3 and Keap1 may serve as effective targets for the prevention of neurodegeneration in AD . We next compared the effects of manipulating both Keap1 and GSK-3 regulatory pathways on Aβ42 toxicity and Nrf2/cncC activity in a parallel study . Using a dose of lithium ( 25 mM ) that prevented Aβ42-induced lifespan-shortening to a similar extent as heterozygous loss-of-function mutations in Keap1 ( Fig 2C , P = 0 . 739 comparing +RU , + LiCl 25 mM to +RU , Keap1 del ) , this comparative analysis revealed that both manipulations rescued lifespan-shortening in Aβ42-expressing flies in an additive manner ( Fig 2C , P<0 . 001 comparing +RU + LiCl 25 mM to +RU + LiCl 25 mM , Keap1 del ) , consistent with an independent mechanism of action . At these therapeutic levels , genetically reducing Keap1 , alone , significantly rescued the decline in Nrf2/cncC reporter expression observed in Aβ42-expressing flies ( Fig 2D , p<0 . 05 comparing +RU vs +RU , Keap1 del ) . By contrast , although 25 mM lithium was sufficient to rescue Aβ42 toxicity , a dose of 50–100 mM lithium , which may exert toxic effects [47] , was required to rescue Nrf2/cncC activity as measured using both microarray ( S5A Fig ) and gstD2 expression ( S5B Fig ) analyses . A heterozygous loss-of-function mutation in cncC ( cncCK6[48]; S5C Fig ) did not alter the ability of lithium to protect against Aβ42-mediated lifespan-shortening , further suggesting that Nrf2 is not required for lithium to exert its protective effects in vivo . Consistent with our previous finding that lithium reduces Aβ42 levels , and prevents toxicity , by blocking translation[47] , early intervention concurrent with RU486 induction was required for lithium to protect against Aβ42-induced climbing defects ( S5D Fig ) and was correlated with reduced Aβ42 levels from the point of induction ( S5E Fig ) . This suggests that , at therapeutically active concentrations , the protective effect of lithium against Aβ42 toxicity is not predominantly mediated by activating Nrf2 , but rather by blocking Aβ42 peptide accumulation through inhibition of translation . Our data provide the first evidence that specific Keap1 inhibition can protect against Aβ42 toxicity in vivo , and they highlight Keap1 as a more efficient target for the prevention of Aβ42 peptide-induced Nrf2 inhibition in comparison with lithium treatment . As Nrf2 plays an important role in protecting against oxidative and xenobiotic stress , we investigated the nature of the Aβ42-induced molecular damage that was ameliorated by Keap1 inhibition ( Fig 3 ) . UAS-ArcAβ42>elavGS control and heterozygous Keap1 mutant flies were induced with RU486 for 14 days before measuring sensitivity to xenobiotic ( DDT; dichlorodiphenyltrichloroethane ) or oxidative ( hyperoxia or paraquat ) stressors ( see methods ) . Reducing Keap1 alone protected against Aβ42-induced sensitivity to DDT ( Fig 3A ) and paraquat ( Fig 3C ) , but not hyperoxia ( Fig 3B; +RU vs +RU , Keap1 del ) . Lithium treatment , on the other hand , protected against sensitivity of Aβ42 flies to DDT to a similar extent as Keap1 inhibition ( Fig 3A ) , but more significantly protected against sensitivity to both hyperoxia ( Fig 3B ) and paraquat ( Fig 3C ) -induced oxidative damage . Moreover , when combined , lithium did not add significantly to the ability of reduced Keap1 to protect against the sensitivity of Aβ42 flies to DDT ( Fig 3A , +RU , +LiCl 25 mM Keap1 del vs +RU , Keap1 del ) and , conversely , reduced Keap1 did not add to the ability of lithium to protect against paraquat ( Fig 3B ) or hyperoxia ( Fig 3C , +RU , +LiCl 25 mM Keap1 del vs +RU , +LiCl 25 mM ) . Consistent with the apparently different mechanisms of neuronal protection by Keap1 and lithium , this finding suggests that Keap1 inhibition acts mainly through protection against xenobiotic stress and lithium predominantly by limiting oxidative damage . In addition to increasing Nrf2/cncC activity , the protective effect of reducing Keap1 on Aβ42 toxicity correlated with a reduction in Aβ42 peptide levels ( Fig 4A ) . Although Aβ42 mRNA was also slightly reduced by down-regulation of Keap1 ( Fig 4B ) , the peptide levels were not altered until 14 days of age ( Fig 4C ) , suggesting that reducing Keap1 may clear Aβ42 peptide by activating protein degradation mechanisms . We have shown previously that Aβ42 peptide is stable for several weeks following induction in our inducible fly model[49] . To assess whether the peptide is indeed degraded as a consequence of reducing Keap1 , we induced Aβ42 expression in Keap1del and control flies for one week with RU486 then measured Aβ42 levels by ELISA at several time-points following transfer to non-RU486-containing medium ( Fig 4D ) . Aβ42 levels were equivalent between heterozygous Keap1del and control flies at the end of the induction period ( 0d since switch-off; age 7d ) , but then declined only in flies with reduced Keap1 , starting from 3 days following switch-off ( age 10d ) . This finding confirms that inhibition of Keap1 does indeed lead to degradation of the Aβ42 peptide . Moreover , analysis of insoluble proteins one week following RU486 induction for 10 days ( Fig 4E ) , revealed that inhibition of Keap1 reduced the level of aggregated Aβ42 . Keap1-Nrf-2 signalling has been implicated in both autophagy and proteasomal degradation . p62 , a selective autophagy substrate , competes with Nrf-2 for binding to Keap1[50][51] and Keap1 may activate autophagy directly by binding to p62[52] . However , we did not observe any alteration in autophagy activity upon reduction of Keap1 in the Aβ42-expressing flies , as measured by western blotting using an antibody specific for Drosophila ATG8 ( Fig 5A ) . Proteasome subunits are transcriptional targets of Nrf2/cncC[53] and over-expression of cncC in flies , either directly or by reducing Keap1 levels , increases proteasome expression and activity[23] . Accordingly , we detected an increase in proteasome activity in the heads of Aβ42-expressing flies upon reduction of Keap1 ( Fig 5B ) , at a time-point coinciding with reduced Aβ42 levels ( age 14d ) . Pharmacological inhibition of this enhanced proteasome activity ( S6A Fig ) , however , did not prevent the degradation of Aβ42 in response to Keap1 inhibition ( S6B Fig ) , suggesting that this may not be the mechanism by which loss of Keap1 induces Aβ42 degradation . Further studies are therefore required to elucidate the precise clearance mechanisms mediating Aβ42 degradation following Keap1 inhibition . Our data highlight Keap1 as a valid in vivo therapeutic target for AD . Current activators of Nrf2 are electrophilic compounds , which commonly act by modifying cysteine residues on Keap1 and thus disrupt its interaction with , and inhibition of , Nrf2[54] . Of these , the synthetic triterpenoids , such as Bardoxolone methyl ( CDDO-Me ) , are potent Nrf2 activators and have exhibited therapeutic potential against several diseases , including neurodegeneration , in animal models and clinical trials[55] . One problem with the mechanism of action of these electrophilic Nrf2 activators , however , is the modification of cysteine residues on other targets[21 , 22] , which may lead to side-effects in humans[22 , 55] . To improve specificity , compounds have more recently been designed to directly disrupt Keap1-Nrf2 binding , and these enhance Nrf2 activity in vitro [56] and in cells[22 , 57] . Activity of these compounds against disease models , however , remains to be examined . We tested the ability of a recently published direct Keap1-Nrf2 inhibitor , 22h ( Fig 6A[22] ) , to protect against exogenous amyloid toxicity in cultured cells . At their EC50 concentrations for Nrf2 activation ( S7 Fig ) , we first performed a comparative analysis of 22h , CDDO-Me and the GSK-3 inhibitor , TDZD-8 , for protection against natural Aβ oligomer-induced toxicity[58] in SH-SY5Y cells ( Fig 6B ) . Interestingly , Nrf2 and Keap1 protein levels were increased following 24 h Aβ oligomer treatment ( S8 Fig ) , suggesting that the Keap1-Nrf2 pathway is indeed dysregulated by acute Aβ exposure . TDZD-8 was a poor activator of Nrf2 , inducing its target gene NQO1 only at a single concentration of 1 μM ( S7C Fig ) , and exerted toxicity in control SH-SY5Y cells ( Fig 6B ) . Although CDDO-Me was a more potent activator of Nrf2 ( S7D Fig ) , 22h significantly protected against Aβ toxicity , in comparison with both CDDO and TDZD-8 , in this experimental paradigm ( Fig 6B ) . This comparative analysis supports our suggestion that Keap1 may serve as a more efficient target , than inhibition of GSK-3 , for the rescue of Nrf2-dependent effects of Aβ42 toxicity . Moreover , as TDZD-8 has previously been shown to protect against Aβ toxicity in primary neuronal cultures , our data suggest that the threshold for this protective effect of GSK-3 inhibition lies below that for its effects on Nrf2 activity . These findings indicate that direct Keap1-Nrf2 disruptors may protect against Aβ42 toxicity more effectively than established Nrf2 activators . To further test the effects of Keap1-Nrf2 disruption on neuronal function in response to Aβ oligomers , we measured the effects of 22h in primary mouse neurons ( Fig 6C ) . Conditioned medium obtained from Tg2576 mouse neurons ( Tg2576CM ) has previously been shown to contain oligomeric Aβ species at concentrations similar to those in human CSF , and to reduce spine density of GFP-transfected WT mouse cortical neurons[59] . We pre-treated GFP-transfected WT mouse neurons of 12 d in vitro ( DIV ) , with 1 or 10 μM 22h for 16 h prior to administration of either wt or Tg2576 conditioned media ( wtCM or Tg2576CM ) , with continued 22h treatment , for a further 24 h before examining spine density ( see Materials & Methods ) . As previously reported[59] , Tg2576CM reduced total spine density of cortical neurons compared to wtCM ( Fig 6C and 6D ) . Strikingly , spine density was rescued by treatment with compound 22h ( Tg2576CM , 0 . 01% DMSO vs Tg2576CM , 10 μM 22h ) at non-toxic doses ( wtCM , 0 . 01% DMSO vs wtCM , 10 μM 22h ) . Aβ42 levels in conditioned media were unchanged following treatment with 10 μM 22h ( Fig 6E ) which further correlated with increased expression of Nrf2 target genes[60] ( Fig 6F ) . These findings suggest that directly blocking the Keap1-Nrf2 interaction can protect neurons downstream of extracellular amyloid toxicity .
Our study firstly confirms , in Drosophila , previous reports that Nrf2 target genes are down-regulated in AD brain[11] , and in mouse models of AD[13 , 65 , 66] . Some studies suggest that this inhibition may be mediated by Aβ42 directly , with exogenous , synthetic , Aβ42 peptide reducing Nrf2 target gene expression in primary mouse neurons[13] , and blocking Nrf2 nuclear translocation following injection into the hippocampus of rat[65] and mouse[66] brain . As Aβ42 peptide is expressed in our fly model independently of APP processing , our study further confirms that this direct effect on Nrf2 activity occurs in response to naturally-derived Aβ42 peptide conformations in vivo ( Fig 7B ) . Importantly , other disease-related proteins , including tau and C9orf72 DPRs , also suppressed Nrf2/cncC signalling , but the non-toxic Aβ40 peptide did not . This suggests that inhibition of the Nrf2 pathway is a generalised response to the accumulation of aberrant proteotoxic proteins . The mechanisms mediating the effect of Aβ42 on Nrf2 activity remains to be established . Although a recent report , using sweAPP-expressing cells , has suggested that Nrf2 transcripts can be replenished by altering DNA methylation [67] , Aβ42 did not alter mRNA expression of Nrf2/cncC , or of its co-transcription factor MafS , in our Drosophila model . This suggests that in vivo Nrf2 de-regulation in AD may be post-transcriptional . We have shown that Aβ42 oligomers increased Nrf2 and Keap1 proteins in SH-SY5Y cells after 24 h treatment , suggesting that Nrf2 is initially stabilised in response to acute amyloid exposure . This may represent a protective response to the initial toxic insult , as similar effects on Nrf2 have been observed at early time-points following transient focal ischaemia and correlate with preservation of peri-infarct regions of the brain under these conditions [68] . Keap1 is also an Nrf2 target gene [35] , however , and may subsequently be upregulated to control Nrf2 activity following the initial exposure to Aβ42 in our study . If sustained , this increase in Keap1 levels could provide a potential mechanism for the inhibition of Nrf2 observed in AD brain and other chronic neurodegenerative conditions [11] . This hypothesis is supported by observations that Nrf2 protein is downregulated and Keap1 upregulated in mouse brain following 15 days exposure to synthetic Aβ42 [66] . Further work is required , however , to investigate the detailed timing of these events following chronic in vivo exposure to natural Aβ oligomers . Since Aβ42 did not directly affect Nrf2/cncC gene expression , we investigated the role of its upstream inhibitors , GSK-3 and Keap1 , on toxicity . GSK-3 has a well-documented role in Alzheimer’s , and is suggested to provide a pivotal connection between the characteristic amyloid plaque and neurofibrillary tangle pathologies of this disease[26] . Activation of GSK-3 has been proposed to exert neuronal toxicity by many mechanisms , including increasing apoptosis and inflammation , and impairing axonal transport , synaptic function , cell cycle regulation and adult neurogenesis [26] . Conversely , several GSK-3 inhibitors , including lithium , protect against AD-pathology in mice[69–71] and some have been tested in clinical trials for AD[72] . Although GSK-3 inhibition increases Nrf2 activity in AD models[73] , however , only one study has shown epistatically that Nrf2 mediates the neuroprotective effect of lithium treatment against paraquat toxicity in cells[74] . The role of Nrf2 in mediating the protective effect of GSK-3 inhibition in AD has also not been empirically investigated . Our current study confirms that lithium treatment increases transcription of Nrf2 target genes in a dose-dependent manner . Concentrations of lithium required to activate Nrf2 ( ≥50 mM-100 mM LiCl ) , however , have previously been shown to exert toxicity in Drosophila[42 , 47] . Lower doses of lithium ( 25 mM ) were sufficient to protect against Aβ42 toxicity , to a level comparable with reducing Keap1 , but did not significantly activate Nrf2/cncC . Moreover , genetically reducing Nrf2/cncC function did not prevent the lifespan-extending effects of lithium , in Aβ42 flies , even at a maximizing concentration ( 50 mM ) . Our data , therefore , suggest that lithium mediates neuroprotection against Aβ42 independently of its effects on Nrf2 ( Fig 7B ) . Consistent with our previous observations that lithium ( 10–100 mM ) reduces Aβ42 peptide levels by inhibiting translation[47] , early lithium administration reduced Aβ42 peptide , from the point of induction , in our current study and this was required for lithium to exert its protective effects . These findings are further supported by our observation that the specific GSK-3 inhibitor TDZD-8 has a narrow window for Nrf2 activation in cultured cells , inducing its target gene NQO1 only at a single concentration of 1 μM ( Fig 6B ) . At concentrations ≤ 1 μM TDZD-8 has been described to prevent Aβ-induced reductions in spine density in mouse neurons[59] , but doses ≥ 1 μM exerted toxicity ( Fig 6B and [59] ) . Together these data suggest that the threshold for protection against Aβ toxicity by GSK-3 inhibitors lies below the concentration required for activation of Nrf2 in neuron . This supports the hypothesis that blocking GSK-3 exerts its protective effects independently of Nrf2 . Contrary to our findings with lithium , we provide the first in vivo evidence that Keap1 inhibition can exert neuroprotective effects in response to Aβ42 by ameliorating deficits in Nrf2 activity ( Fig 7C ) . As with Drosophila models of PD[32 , 33] , heterozygous loss of Keap1 protected against Aβ42-induced lifespan-shortening and climbing defects in the fly . Moreover , we show that this protective effect of reducing Keap1 correlates significantly with a rescue of Nrf2/cncC activity in response to Aβ42 . Only one previous study has addressed the role of Keap1 inhibition , specifically , in AD . Keap1 RNA interference increased expression of Nrf2 target genes , protected against synthetic Aβ42-induced cytotoxicity and oxidative damage to proteins and lipids , and enhanced autophagy in cultured cells[30] . Our study adds to these in vitro findings , demonstrating that Keap1 inhibition rescues Aβ42-induced Nrf2 deficits , and protects against neuronal decline in vivo . These effects occur in correlation with prevention of xenobiotic and , to a lesser extent , oxidative damage as well as the degradation of endogenous , aggregated , Aβ42 peptide . Using primary cortical neurons , our findings further confirm that direct pharmacological inhibition of Keap1-Nrf2 binding can protect against neuronal damage downstream of extracellular Aβ42 oligomers . The mechanisms by which reducing Keap1 enhances degradation of Aβ in vivo , however , requires further investigation . Autophagy , as measured by cleavage of ATG8 , was unaltered in response to Keap1 inhibition in our Aβ42 expressing flies . Consistent with the established effect of Nrf2 on transcription of proteasome subunits[23] , loss of Keap1 enhanced proteasome activity in Aβ42 flies , but blocking this increase , using the proteasome inhibitor Bortezomib , did not prevent the reduction in Aβ42 levels . This suggests that the reduction of aggregated Aβ42 in response to Keap1 inhibition may not be directly mediated via enhanced autophagy or proteasomal degradation . Nrf2 has also been implicated to play a role in the unfolded protein response ( UPR ) , serving as a target for PERK[75] and activating transcription of several components of the ER associated degradation ( ERAD ) pathway[76] , including chaperones and ubiquitin-conjugating enzymes , in addition to proteasome subunits and autophagy . Future studies will therefore be required to investigate the functional role of such protein quality control pathways , potentially upstream of the proteasome and autophagy , in the clearance of Aβ42 following Keap1 inhibition in vivo . Overall , our data point to Keap1 , in comparison with GSK-3 , as an efficient target for Nrf2 activation in response to Aβ42 toxicity in vivo . Recent advances in the development of direct inhibitors of the Keap1-Nrf2 binding domain may , therefore , enable the prevention of Nrf2 deficits in neurodegenerative diseases with minimal side-effects . Our study shows for the first time that a direct , small molecule inhibitor of Keap1-Nrf2 binding , 22h[22] , can ameliorate synaptotoxicity in response to naturally-derived Aβ oligomers in mouse cortical neurons . As synaptic loss correlates well with cognitive decline[77] , our work suggests that these compounds show promise as therapeutic agents for AD . Moreover , as this is the first demonstration that these compounds can prevent neuronal toxicity , our findings warrant their investigation in other neurodegenerative conditions . Both GSK-3 inhibitors[71] and Nrf2 activators[55] can exert toxicity due to off-target effects and thus it is important that modifiers of these pathways achieve therapeutic benefits at low doses . More specific inhibitors of each of these targets are currently being developed[22 , 71] . It has also been postulated that combined inhibition of both GSK-3 and Keap1 may activate Nrf2 , and confer protection , at lower doses than either intervention alone , hence limiting their detrimental effects[25] . We present the first in vivo study showing that combined Keap1 and GSK-3 inhibition , by lithium , confers additive protective effects against Aβ42 toxicity . Keap1 deletion and lithium treatment extended lifespan and improved climbing ability of Aβ42-expressing flies to a greater extent than either manipulation alone . Additionally , reducing Keap1 limits the dose of lithium required to reach maximal levels of protection since heterozygous Keap1del combined with low dose ( 25 mM ) lithium extended lifespan of Aβ42 flies to a similar extent as high dose ( 50 mM ) lithium alone . Our data , however , do not predict that these additive protective effects are due to mechanisms converging on Nrf2 . Rather , low dose lithium treatment exerts protective effects independently of Nrf2 activation , whereas the neuroprotective effects of Keap1 inhibition correlate strongly with increasing Nrf2 . Hence , the additive nature of combined lithium and Keap1 inhibition appears to be mediated through divergent , complementary protective mechanisms . Oxidative[2 , 3] and xenobiotic damage[9] are key features of AD brain and may potentially be explained by the down-regulation of Nrf2 , which is also observed in several neurodegenerative diseases . Consistent with this idea , Keap1 inhibition and , therefore , Nrf2/cncC activation , correlated strongly with protection against Aβ42-induced sensitivity to the xenobiotic DDT in our fly model . Conversely , Keap1 inhibition exerted minimal protection against oxidative damage in comparison with lithium treatment , which strongly protected against Aβ42-induced sensitivity to both paraquat and hyperoxia-induced damage . As therapeutic concentrations of lithium , which can prevent oxidative damage , did not strongly activate Nrf2/cncC in our flies , this suggests that rescue of Nrf2 activity is not required to protect against oxidative damage in response to Aβ42 . Although pharmacological activation of Nrf2 , has been shown to protect against oxidative damage , in response to Aβ42 peptide , in cells[78 , 79] and in animal models of AD[65 , 66] , conflicting reports have also described protective effects against such damage that are mediated independently of Nrf2[80] . Moreover , studies showing protection against Aβ42-induced oxidative damage in response to GSK-3 inhibition , using antisense oligonucleotides[73] and in response to lithium[81] , have not demonstrated a causal role of increasing Nrf2 activity . Our direct comparison of Keap1 and GSK-3 pathways , however , reveals that Nrf2 activity and prevention of Aβ42-induced oxidative damage do not strongly correlate . Finally , combining lithium treatment with manipulation of Keap1 did not confer additional protection against oxidative and xenobiotic damage respectively . Overall this suggests that Keap1 and lithium treatment combine to maximise protection against AD-phenotypes , through divergent effects on Nrf2 and by additively protecting against Aβ42-induced oxidative and xenobiotic damage ( Fig 7D ) .
Our findings provide compelling support for the use of direct Keap1-Nrf2 inhibitors for the treatment of neurodegenerative diseases , particularly AD . Future work is warranted to develop these compounds further for in vivo use , and to investigate their effects in combination with other established therapeutic targets for AD , such as specific GSK-3 inhibitors .
Stocks were maintained at 25°C on a 12:12-h light:dark cycle at constant humidity on a standard sugar-yeast ( SY ) medium ( 15gl-1 agar , 50 gl-1 sugar , 100 gl-1 autolysed yeast , 100gl-1 nipagin and 3ml l-1 propionic acid ) . Adult-onset neuronal-specific expression of Arctic mutant Aβ42 peptide was achieved by using the elav GeneSwitch ( elavGS ) -UAS system ( GAL4-dependant upstream activator sequence ) and treatment with 200 μM mifepristone ( RU486 ) , as previously described[34] . ElavGS was derived from the original elavGS 301 . 2 line and obtained as a generous gift from Dr H . Tricoire ( CNRS , France ) . Aβ lines used in Fig 1 were: UAS-attP Aβ lines , as previously published [36] , and UAS-WT Aβ42x2 , obtained from Prof . P . Fernandez-Funez ( University of Florida , USA ) [37] . Random insertion UAS-ArcAβ42 was obtained from Dr D . Crowther ( University of Cambridge , UK ) and was used in all other experiments . UAS-0N3R tau line was obtained from Guy Tear ( Kings College London , UK ) . UAS-C9orf72 ( GR ) 100 flies are published[40] . Keap1del was generated by P-element mediated male recombination using the P-element insertion line , Keap1[EY02632][42] , Keap1EY5 and gstD1-GFP lines were obtained from Dr D . Bohmann ( University of Rochester , USA ) , and cncCK6 mutant flies , originally generated in the laboratory of William McGinnis ( University of California , San Diego , USA ) , from Dr A . Whitworth ( University of Cambridge , UK ) . All fly lines were backcrossed six times into the w1118 genetic background . LiCl ( Sigma ) was dissolved in ddH20 at a concentration of 5 M before diluting to the indicated final concentrations in SYA medium . Bortezomib ( New England Biolabs ) was dissolved in ethanol at a stock concentration of 10 mM , and stored frozen at– 20°C , before diluting to the indicated final concentrations in SYA medium . Flies were raised at a standard density on SY medium in 200 mL bottles . Two days after eclosion once-mated females were transferred to experimental vials containing SY medium with or without RU486 at a density of 10 flies per vial . Deaths were scored and flies were transferred to fresh food 3 times per week . Data are presented as cumulative survival curves , and survival rates were compared using log-rank tests . Climbing assays were performed using methods adapted from Sofola O et al . , 2010 [34] . Briefly , 15 adult flies were placed in a vertical glass column ( SciLabware ) , tapped to the bottom , and their ability to climb subsequently analysed . Flies reaching the top ( above 10 cm ) and flies remaining at the bottom ( below 3 cm ) of the column after a 30 sec period were counted . Scores recorded , from three trials per biological repeat , were the mean number of flies at the top ( ntop ) , the mean number of flies at the bottom ( nbottom ) and the total number of flies assessed ( ntot ) . A performance index ( PI ) defined as ½ ( ntot + ntop—nbottom ) / ntot ) was calculated . Data are presented as the mean PI ± SEM obtained in three independent repeats for each group . cncC activity was measured by crossing UAS-ArcAβ42;elavGS flies to gstD1-GFP reporter flies , expressing green fluorescent protein ( GFP ) under the control of a 2 . 7kb genomic sequence upstream of the cncC target gene gstD1 [35] , and analyzing GFP levels by western blotting . Fly heads were homogenized in 2× laemmli sample buffer containing 100mM DTT . Proteins were separated by SDS-PAGE at 150V for 1h using 10% Bis-Tris gels and Mes-SDS running buffer ( Invitrogen ) . Gels were then transferred to nitrocellulose membrane , incubated in a blocking solution containing 5% milk proteins in TBST for 1h at room temperature , then probed with GFP ( Cell Signaling , 2955S , 1:1000 ) , ATG8 ( custom made anti-rabbit polyclonal against peptide EP113385 ( Eurogentec ) [42] , 1:1000 , ) or actin ( Santa Cruz , sc-47778 , 1:5 , 000 ) primary antibodies overnight at 4°C . Anti-horseradish peroxidase ( HRP ) -conjugated secondary antibodies ( Abcam , 1:12 , 000 ) were used and blots were developed using the enhanced chemiluminescence method ( ECL ) according to the manufacturers’ instructions ( Luminata Crescendo; Millipore ) . Proteins were visualized using a luminescent image analyzer ( ImageQuant LAS 4000; GE Healthcare ) and relative intensities measured using ImageQuant software . Proteins of interest were expressed as a ratio relative to actin levels in each sample . RNA extraction , cDNA synthesis and quantitative PCR ( qPCR ) reactions were performed as previously published[34 , 49] . For gene expression in Drosophila , primers , 5’ to 3’ , were as follow: Aβ forward GATCCTTCTCCTGCTAACC and reverse CACCATCAAGCCAATAATCG; cncC forward GAGGTGGAAATCGGAGATGA and reverse CTGCTTGTAGAGCACCTCAGC; MafS forward AGATCGTTCGGATGAAGCAG and reverse GTCTCCAGCTCGTCCTTCTG; gstD2 forward CATCGCCGTCTATCTGGTGGA and reverse GGCATTGTCGTACCACCTGG; and eIF-1A reference gene forward ATCAGCTCCGAGGATGACGC and reverse GCCGAGACAGACGTTCCAGA[82] . Genes of interest were expressed as a ratio relative to eIF-1A . For gene expression in primary mouse neuronal cultures , primers 5’ to 3’ were: Hmox1 forward AGCACAGGGTGACAGAAGAG and reverse GGAGCGGTGTCTGGGATG , Srnx1 forward GACGTCCTCTGGATCAAAG and reverse GCAGGAATGGTCTCTCTCTG , and xCT forward ATACTCCAGAACACGGGCAG and reverse AGTTCCACCCAGACTCGAAC , as previously published [60] . Reference gene primers were to mouse actin forward AACCGTGAAAAGATGACCCAGA and reverse CACAGCCTGGATGGCTACGTA . Genes of interest were expressed as a ratio relative to actin . Total Aβ42 peptide , from fly heads , was extracted into guanidinium HCl ( GnHCl ) buffer based as previously described[34 , 83] . Alternatively , soluble and insoluble Aβ pools were isolated by differential centrifugation followed by formic acid extraction , as previously described [49] . Aβ42 levels were then measured using the High Sensitivity Human Amyloid Aβ42 ELISA kit ( Millipore ) . Samples were diluted 1:100 , for total Aβ42 , or 1:10 , for insoluble Aβ42 , in dilution buffer and ELISA performed according to the manufacturers’ instructions . Protein extracts were quantified using the Bradford protein assay ( Bio-Rad laboratories Ltd , UK ) and the amount of Aβ42 in each sample expressed as a ratio of the total protein content ( pmoles/g total protein ) . For cell culture , conditioned media was removed from cells , following compound treatment , and diluted 1:2 in dilution buffer before measurement of Aβ42 levels by ELISA as described above . Fly heads were homogenized in 25 mM Tris , pH 7 . 5 and protein content determined by Bradford assay . Chymotrypsin-like peptidase activity of the proteasome was assayed using the fluorogenic peptide substrate Succinyl-Leu-Leu-Val-Tyr-amidomethylcoumarin ( LLVY-AMC ) , as previously described[49] . Proteasome activity was determined as the slope of AMC accumulation over time per mg of total protein ( pmoles/min/mg ) . Flies were prepared as for lifespan experiments then aged to the indicated time-points before measuring stress resistance . For xenobiotic stress , flies were exposed to DDT vapour , 1 mg/mL diluted in acetone , in glass vials in the absence of food . For oxidative stress , flies were subjected to hyperoxia ( 95% oxygen; PROOX model 110 , Biospherix , USA ) or injected with paraquat ( 75 nLs of 1 mg/ml in Ringers buffer ) and maintained on SYA media containing RU486 ± LiCl as indicated . As survival times were short in these experiments flies were not transferred to fresh food . Survival under each stress condition was then monitored by recording the number of deaths every 2 hours from the start of decline . Data are presented as cumulative survival curves , and survival rates were compared using log-rank tests . The raw microarray data generated in this study are deposited in ArrayExpress ( http://www . ebi . ac . uk/arrayexpress ) with identifier E-MTAB-4611 . UAS-ArcAβ42/+;elavGS/+ flies were treated , for 17 days , with standard SY medium alone ( -RU ) or medium containing RU486 in the absence ( +RU ) or presence of 100 mM Lithium Chloride ( +RU , +LiCl ) . Flies used for microarray analyses were snap-frozen in liquid nitrogen and , for each array , RNA extracted from 200 heads using RLT buffer + 0 . 01% β-mercaptoethanol and purified with RNeasy columns ( Qiagen , Valencia , CA , USA ) following the manufacturer's instructions . The quality and concentration of RNA was confirmed using an Agilent Bioanalyzer 2100 ( Agilent Technologies , Santa Clara , CA , USA ) , and further procedures followed the standard Affymetrix protocol . All samples were hybridized to the Drosophila Genome 2 . 0 Genechip . In total , 4–5 biological replicates of each condition ( -RU , +RU and +RU , +LiCl ) were performed . Raw data ( cel files ) were processed to correct for probe-sequence biases , by using bioconductor's package gcrma ( http://www . bioconductor . org ) in R ( http://www . r-project . org ) . Affymetrix's MicroArray Suite 5 . 0 ( bioconductor's package affy ) was used to determine present target transcripts[84] . Transcripts were deemed present if the p-value was <0 . 111 and absent otherwise . The raw data were summarized and normalized by using Robust Multichip Average ( rma function , part of bioconductor's package affy [85] . In order to identify differentially expressed genes a linear model was fitted and differential expression was assessed using the empirical Bayes moderated t-statistic as implemented in R's limma package [86] . P-values were adjusted for multiple hypothesis testing by applying the Benjamini and Hochberg correction for false discovery rate . Summarized probe-sets were mapped to transcripts using R's package "drosophila2 . db" . Transcripts not mapping to any known or predicted genes were excluded from further analysis . The following freely available gene expression microarray datasets were used: control vs . hsp70-CncC ( E-GEOD-30087 ) . The Wilcoxon rank sum test , as implemented in Catmap[87] , was used to perform functional analysis , that is significant enrichment of Gene Ontology categories . FlyBase ( http://flybase . org ) gene identifiers were mapped to Gene Ontology identifiers ( FlyBase version FB2014_01 ) . Ranks of genes were based on the p-value derived from the Bayes t-statistic for differential expression . To account for multiple hypothesis testing , an enrichment of GO terms was deemed statistically significant if the p-value derived from the wilcoxon rank sum test was ≤1 . 0x10-05 . Gene lists were sorted by log-fold change and p-value . For all microarray experiments two sets of lists were derived; a gene list comprising most differentially up-regulated ( log-fold change > 0 ) genes at the top of the list and most differentially down-regulated genes ( log-fold change < 0 ) at the bottom of the list ( termed up-to-down ) and vice versa ( termed down-to-up ) . If a GO category was found to be statistically significant in the up-to-down list , this GO was referred to as up-regulated , meaning that a large enough proportion of the genes in this GO category were found to be up-regulated or at the top of the list . If a GO category was found to be statistically significant in the down-to-up list , this GO was referred to as down-regulated , meaning that a large enough proportion of the genes in this GO category were found to be down-regulated or at the top of the list . Statistical significance of overlaps of GOs in two microarray experiments was determined using fisher's exact test . To account for multiple hypothesis testing a p-value cut-off of ≤1 . 0x10-05 was used . SH-SY5Y cells were incubated in a humidified atmosphere at 37°C , 5% CO2 in Dulbecco’s modified medium ( DMEM ) supplemented with 4 . 5g/L glucose , 10% FBS , 1% penicillin- streptomycin , and differentiated with 10 μM retinoic acid . Cells were seeded at an appropriate density in 96-well plates prior to subsequent analyses . SH-SY5Y cells were seeded at a density of 2 x 104 cells per well in a 96-well plate and cultured for 2 days before treatment for 24 hours with compound or vehicle ( 0 . 1% DMSO final concentration ) . Cells were then lysed in 50 μl/well lysis buffer ( 0 . 1% Tween20 in 2 mM EDTA [pH 7 . 5] ) for 15 mins at room temperature . 200 μl enzyme reaction mixture ( 25 mM Tris buffer [pH 7 . 5] containing BSA [0 . 067%] , Tween20 ( 0 . 01% ) , FAD ( 5 μM ) , Glucose 6 Phosphate ( G6P ) ( 1 mM ) , NADP ( 30 μM ) , G6P dehydrogenase ( 40 units ) , MTT ( 0 . 03% ) , and menadione ( 50 μM ) ) was added to each well . After 5 min at room temperature , 40 μl/well stop solution ( 10% SDS ) was added and the absorbance at 595 nm measured . The background optical density was measured using wells containing lysis buffer , enzyme and stop solutions without SH-SY5Y cells . The optical density values at 595 nm were averaged and the background corrected ratio of optical densities ( compound treated/control ) was calculated . SH-SY5Y cells were seeded at a density of 2 x 104 cells per well in a 96-well plate . Cells were pre-treated with compound or vehicle overnight before either Aβ-conditioned Chinese Hamster Ovary ( CHO ) cell media ( 7PA2CM ) or WT-conditioned CHO media ( wtCM ) was added to wells at a dilution of 50% for 24 hours . Resazurin ( final concentration 20 μM ) was added to wells and incubated for 4 hours at 37°C , 5% CO2 . The resulting fluorescence intensity was measured at 590 nm . The relative fluorescence values were averaged and normalised to wtCM , DMSO-treated control intensities . SH-SY5Y cells were seeded at a density of 6 , 000 cells per well , in a 96-well plate then , after 24 hours , treated with 50% 7PA2CM or wtCM for a further 24 hours . Plates were fixed with 4% PFA ( v/v ) for 15 minutes , washed three times with PBS , permeabilized with 0 . 5% Triton-X before staining with primary antibodies against Keap1 ( 1:50; ab150654 , Abcam ) and Nrf2 ( 1:100; ab62352 , Abcam ) in blocking solution containing 10% goat serum and 3% BSA overnight at 4 C . After 3 washes in PBS secondary antibodies ( α rabbit Alexa 594 , ab150080 , α mouse FITC , ab6785 , Abcam ) were applied . Nuclei were stained in a PBS solution of 10 μM Hoechst 33342 before high-resolution digital imaging using a Zeiss LSM700 confocal microscope . For quantification of Nrf2 and Keap1 staining , and cell morphology , 250 cells per well were imaged using the IN Cell Analyzer 6000 automated laser-based imaging platform with confocal modality ( GE Healthcare ) at 40X magnification . Automated image analysis was conducted with the IN Cell Developer software using custom-developed analysis protocols . Cell nuclei were identified by acquiring images in the DAPI channel ( 405 nm excitation , 455/50 emission ) . Whole cells were identified using images of Keap1 from the FITC channel ( 488 nm excitation , 524/48 emission ) . Images were also acquired of Nrf2 in the dsRed channel ( 561 nm excitation , 605/52 nm emission ) . The whole-cell intensity of Keap1 FITC and Nrf2 Alexa 594 was measured and expressed as average fluorescence units per cell for each well . N = 5 wells per condition . To obtain transgenic conditioned medium ( TgCM ) enriched in Aβ , neurons from Tg2576 mice were maintained for 14 days in vitro ( DIV ) without changing the medium[88] . Medium from wild type neurons from littermates , wild type conditioned medium ( wtCM ) , was used as a control . The genotype of the animals was determined by polymerase chain reaction on DNA obtained from fibroblasts . Wild type primary neurons were obtained from cerebral cortex of CD1 mouse embryos , at embryonic day 15 , as previously described [89] . Neurons were plated to a density 6×105 viable cells/35-mm2 on glass-bottomed dishes ( MatTek Corporation , Ashland , MA , USA ) previously coated with poly-D-lysine ( 10 μg/ml ) for at least 1 h at 37°C . Cultures were maintained at 37°C with 5% CO2 , supplemented with neurobasal medium with 2% B27 nutrient , 2 mM L-glutamine , penicillin ( 100 units/ml ) and streptomycin ( 100 μg/ml ) . At 12 days in vitro ( DIV ) neurons , transfected on day 7 with the plasmid peGFP-N1 ( Clontech , Mountain View , CA ) using lipofectamine 2000 ( Invitrogen ) , were pre-treated with the Keap1 inhibitor , 22h ( 10 μM[22] ) , or 0 . 1% DMSO , for 16 h . Conditioned medium ( CM ) from Tg2576 transgenic , Aβ-enriched , or wild type mouse neurons , with and without 22h , was then added for a further 24h before analysis . Neuronal morphology was assessed at 14 DIV by high-resolution digital imaging using a Zeiss LSM700 confocal microscope and analysis using NeuronStudio software ( CNIC , Mount Sinai School of Medicine ) . Spine density was defined as the number of spines per micrometer of dendrite length according to previously published protocols[59] . Dendritic spine densities were calculated from 6–17 neurons per condition . Data are presented as means ± SEM obtained in at least three independent biological samples . Log-rank , analysis of variances ( ANOVA ) and Tukey’s HSD ( honestly significant difference ) post-hoc analyses were performed using JMP ( version 11 . 0 ) software ( SAS Institute , Cary , NC , USA ) . Animals were maintained and treated in accordance with the Animals ( Scientific Procedures ) Act , 1986 , following approval by the Animal Welfare and Ethical Review Body ( AWERB ) , KCL , and the Home Office Inspectorate , and performed in accordance with the European Communities Council Directive of 24 November 1986 ( 86/609/EEC ) . | As our population ages the incidence of neurodegenerative diseases , including Alzheimer’s disease ( AD ) , is predicted to increase dramatically . Despite providing important symptomatic relief , existing treatments for such conditions do not slow-down disease progression , and this will cause an overwhelming future burden on our healthcare system and immense suffering for many more patients and their families . Nrf2 is a gene that normally protects cells from stressful conditions . Although we don’t know why , Nrf2 is reduced in the brains of AD patients and this may explain the increased susceptibility of neurons to damage in neurodegenerative diseases . Our research , using a fruit fly model , identifies Keap1 , a negative regulator of Nrf2 , as a valid target for the rescue of AD-related Nrf2 defects and the subsequent prevention of neuronal degeneration . Moreover , we show that a new compound , which directly blocks the binding between Nrf2 and Keap1 , can prevent toxicity of the AD-initiating Aβ peptide in mouse neurons . Hence , our study provides strong evidence that direct Keap1-Nrf2 disruptors can specifically target the defects in Nrf2 activity observed in neurodegenerative diseases , and supports the further development of such compounds as potential new drugs to prevent neuronal decline AD and other neurodegenerative conditions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Conclusions",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"neurodegenerative",
"diseases",
"neuroscience",
"toxicology",
"toxicity",
"lithium",
"genome",
"analysis",
"alzheimer",
"disease",
"genomics",
"animal",
"cells",
"proteins",
"chemistry... | 2017 | Direct Keap1-Nrf2 disruption as a potential therapeutic target for Alzheimer’s disease |
Protein function is mediated by different amino acid residues , both their positions and types , in a protein sequence . Some amino acids are responsible for the stability or overall shape of the protein , playing an indirect role in protein function . Others play a functionally important role as part of active or binding sites of the protein . For a given protein sequence , the residues and their degree of functional importance can be thought of as a signature representing the function of the protein . We have developed a combination of knowledge- and biophysics-based function prediction approaches to elucidate the relationships between the structural and the functional roles of individual residues and positions . Such a meta-functional signature ( MFS ) , which is a collection of continuous values representing the functional significance of each residue in a protein , may be used to study proteins of known function in greater detail and to aid in experimental characterization of proteins of unknown function . We demonstrate the superior performance of MFS in predicting protein functional sites and also present four real-world examples to apply MFS in a wide range of settings to elucidate protein sequence–structure–function relationships . Our results indicate that the MFS approach , which can combine multiple sources of information and also give biological interpretation to each component , greatly facilitates the understanding and characterization of protein function .
Vast amounts of sequence and structural data are being generated by high-throughput technologies . Functional annotations of the uncharacterized sequences and structures are significantly lagging . The time and cost of experimental techniques required to probe the function of all uncharacterized proteins are prohibitive . Therefore , computational means have been increasingly useful and popular in predicting and annotating functions for the huge amount of sequence and structure data [1] , [2] . However , protein function prediction is itself a difficult problem to formulate , since it is difficult to define function [2] , [3] . Various functional definition schemes ( such as the Enzyme Commission [4] , the Gene Ontology [5] , and the SCOP superfamily [6] ) have been developed over the years and have addressed various aspects of protein function . Instead of adopting an existing functional definition scheme , we proposed to probe the role of individual amino acid residues in protein function , regardless of the functional definition schemes that are used . In such cases , the protein function can be represented simply as a series of quantitative values , each of which indicates the functional importance of the corresponding amino acid residue in the protein sequence or structure . To calculate the quantitative values for each residue , we used a combined approach , the meta-functional signature ( MFS ) , which takes into account the individual scores from various function prediction algorithms and generates a composite score for each amino acid residue in a given protein . Currently our signature generation protocol consists of the following four types of scores for four different types of information: ( 1 ) sequence conservation , ( 2 ) evolutionary conservation , ( 3 ) structural stability , and ( 4 ) amino acid type . All these scores are generated via conceptually simple and easily implementable algorithms ( described below ) , and their combined use outperforms sophisticated algorithms that use only one source of information . Sequence conservation is one of the most utilized methods for measuring the functional importance of individual amino acids . Amino acid residues with more conservative variation patterns are usually more important for the preservation of protein function . This concept is often used to identify the functional regions of proteins by building multiple alignments between the target sequence and all its sequence homologues , and then analyzing the degree of sequence conservation among each alignment site . Various measures of sequence conservation have been proposed over the years , with differing complexity and sophistication [7] . The simplest measures of sequence conservation are the entropy score and its variants [8]–[13] . More complicated measures [14]–[16] incorporate other information , such as amino acid pairwise similarity , physicochemical properties , and theoretical sequence profiles , into the scoring schemes . The AL2CO program package incorporates nine different scoring schemes , but these scores tend to correlate with each other [17] . Recently it was also shown that a Jensen-Shannon divergence measure improves predicting functionally important residues , and that considering conservation in sequentially neighboring sites further improves accuracy [18] . We previously demonstrated that a relative entropy measure which incorporates amino acid background frequencies , can better predict functional sites than simple entropy measures [19] . Furthermore , we found that incorporating the amino acid frequencies as estimated by the hidden Markov Models ( HMMs ) further improves the performance of the relative entropy measure [19] . In the current study , we use a sequence conservation measure derived from HMMs ( HMM_rel_ent ) as one component of our meta-functional signature generation protocol . In addition to sequence conservation , we also incorporate evolutionary conservation information in the meta-functional signature . Many studies have shown that the use of phylogenetic relationships among a group of evolutionarily related sequences help accurate prediction of functional sites . The Evolutionary Trace method , one of the first and the most successful of such methods , analyzes residue variation patterns within and between protein subfamilies from multiple alignments , maps important residues to protein structure , and quantitatively ranks residue importance [20] , [21] . A further development of the Evolutionary Trace method allows quantitative ranking of residue importance , by combining the use of evolutionary information and the entropy measures [22] , [23] . Similarly , the ConSurf method constructs phylogenetic relationships from a group of similar sequences , calculates the conservation score by a Bayesian or a maximum likelihood method , and maps the conservation information to the protein surface [24] , [25] . Further , a study by Soyer et al . used site-specific evolutionary models that assumed a different substitution matrix for each site , for detecting protein functional sites [26] . La et al . used evolutionary relationships among sequence fragments ( phylogenetic motifs ) to infer protein functional sites [27] . del Sol Mesa et al . presented several automated methods that divide a given protein family into subfamilies and search for residues that determine specificity [28] . The commonality among all these methods is that sequence relationships are analyzed based on the topology of an evolutionary tree , thus providing an additional level of information instead of relying on multiple sequence alignments alone . Here , we propose a novel method , called the state to step ratio score ( SSR ) , for measuring evolutionary conservation . Based on given multiple alignments , we construct a maximum parsimony tree , and analyze the variation patterns from the root of the tree ( theoretical ancestral sequence ) to the leaf of the tree ( sequences in multiple alignments ) to create a score for each amino acid residue . The SSR score is a simple yet effective way of measuring evolutionary conservation . Functional signature scores can also be derived from biophysics-based methods , using experimentally determined or computationally predicted protein structures . For example , a recent study demonstrated that destabilizing regions in protein structures can often be used to provide valuable information for functional inference and functional site identification [29] . For a given structure and a given position , we propose that we can mutate the wild-type residue to 19 other amino acids and calculate their structural stability scores , which can in turn be used to assign a score to each residue in a protein . Hence , these scores can also serve as a component of protein function prediction . We previously developed a residue-specific all-atom probability discriminatory function ( RAPDF ) [30] that compiles statistics from a database of experimental structures to score and pick “decoy” structures that are more likely to be similar to experimentally derived structures . The RAPDF has been optimized and enhanced in recent years for protein structure prediction [31]–[33] . Here , we further expanded the RAPDF to score residue mutations on a per-residue basis . Each residue in a given protein was mutated to one of the 19 alternative amino acids , producing new structures that were further optimized for topology ( via side chain rearrangement ) and maximized for stability ( via global conformation perturbation ) . In our current MFS generation protocol , we used two RAPDF based scoring functions ( RAPDF_spread and RAPDF_dif ) , to measure how all mutated structures deviate from each other and how the experimentally determined structure differs from mutated structures , which represent the potential impact on stability for the position and for the naturally occurring residue , respectively . These scores separate residues conserved for structure versus function . An additional component of the meta-functional signature is information on the type of amino acids , such as histidine and cysteine , which are more likely to be located in functional sites than other amino acids . However , such “prior probability” for a functional site is not explicitly modeled and incorporated by most current functional site prediction algorithms . In our MFS generation protocol , we used 19 binary variables ( all except Alanine ) to represent the amino acid identity for each position in a given protein . We also examined whether the explicit use of amino acid information ( for example , AAType ) , as opposed to the implicit use ( for example , via relative entropy calculation ) , could provide additional information and better performance . Given the complexity of defining and identifying protein functional sites , clearly no single method will always work to capture all protein functional site information . Therefore , several groups have begun to incorporate information from various sources , especially structure-derived information , to give more accurate predictions . Work by Chelliah et al . has shown that distinguishing the structural and functional constraints for amino acid residues leads to better prediction of protein interaction sites [34] . We have shown that by considering both structural and functional constraints on protein evolution , we can better identify functional sites and signatures [35] , [36] . Recently , Petrova et al . showed that integration of seven selected sequence and structure features into a support vector machine ( SVM ) framework can improve identification of catalytic sites [37] . Furthermore , Fischer et al . integrated sequence conservation , amino acid distribution , predicted secondary structure and relative solvent accessibility into a probability density framework , and showed that at 20% sensitivity the integrated method leads to a 10% increase in precision over non-integrated methods for predicting catalytic residues from the Catalytic Site Atlas and PDB SITE records [38] . Youn et al . investigated the various features for discriminating catalytic from noncatalytic residues in novel structural folds , and showed that a measure of sequence conservation , a measure of structural conservation , a degree of uniqueness of a residue's structural environment , solvent accessibility , and residue hydrophobicity are the best predictors of catalytic sites [39] . Other similar studies also incorporated dozens to hundreds of features into a machine-learning framework for catalytic site identification [40] , [41] . Altogether , the previous work suggests great value in using several complementary sequence and structure components for scoring catalytic sites . Unlike these approaches that were largely based on machine-learning algorithms , in the current study , we aim to combine several sources of information regarding the sequence , structure , evolution , and type of amino acids together via a simple logistic regression model for function prediction , including both catalytic sites and binding sites . The major advantage of the regression model is that each component can be associated with a biologically meaningful interpretation , and that individual scores for a protein can be manually studied to gain additional insights into different aspects of protein function , which are not available when many components are thrown into a sophisticated machine-learning framework . We compare the MFS approach with several other functional site prediction algorithms , propose enhancements to our approach , exemplify the wide definition of function assessed by MFS , and discuss how different components of MFS can be used to understand biological function via four real-world examples .
We used the Thornton dataset [50] and the Lovell dataset [34] to evaluate the performance of MFS and its variants in identifying functional sites from protein structures . The Thornton dataset contains 1 , 546 enzyme active sites from 508 proteins , and the Lovell dataset contains 1 , 137 functional sites from 243 proteins . We evaluated the performance of functional site identification by two criteria that were used in previous studies [19] . The first criterion is the ROC score , which evaluates how the quantitative predictions on functional importance correlate with the binary assignments of whether the site is functional . This score is calculated as the area-under-the-curve by plotting the false positive rate against the true positive rate across a range of threshold values . The second criterion is the top-10 hits scores , which counts how many of the top-10 scoring residues in a given protein are also active site residues . For a given dataset , the sum of the top-10 hits scores for all proteins are used for evaluating the performance of different algorithms . In addition , we also calculated the specificity and the false positive rates for each protein , when 20% sensitivity is achieved . Assuming that TP , TN , FP , and FN represent true positive , true negative , false positive and false negative predictions , respectively , the sensitivity refers to TP/ ( TP+FN ) , precision refers to TP/ ( FP+TP ) and the false positive rate refers to FP/ ( FP+TN ) . For the MFS and SeqonlyMFS methods , we applied five-fold cross-validation experiments to evaluate their performance: the entire dataset was divided into five parts , and during each cross validation , 80% of the proteins were used for training the model , which was then tested on the remaining 20% of the proteins . We evaluated the performance of the MFS method by comparison to two widely used functional site identification programs for protein structures: the Evolutionary Trace server ( http://mammoth . bcm . tmc . edu/report_maker ) and the ConSurf server ( http://consurf . tau . ac . il ) . We used the PDB identifier to query the Thornton and Lovell datasets using both servers with all default parameters and collected the output ZIP files from the ET server and the output “amino acid conservation score” files from the ConSurf server . Some proteins generated error messages or cannot be handled by either one of the servers and therefore were omitted from our analysis . We then used the “rho ET score” value from the ET scoring file and the conservation value from the ConSurf scoring file to evaluate the performance of these methods by the ROC and top-10 hits scores . The ET server generates many equal-valued scores ( usually much more than 10 ) for the highest-scoring residues; therefore , the top-10 hits score was not used for ET in our comparative analysis . For each method , we also generated modified PDB structure files in which the temperature field was replaced by the predicted functional importance scores . These structures were then visualized using the UCSF chimera software [51] so that the color of each residue represents the functional importance score value . Visual inspection of the generated structures helps to understand how and why each method worked or failed . We implemented the MFS generation protocol as a web server , available at http://protinfo . compbio . washington . edu/mfs . The input for this server is either a single chain sequence or structure in FASTA or PDB format , respectively , and the output is the predicted MFS score for each residue in the structure . In addition , when an input structure is provided , a new structure file with the temperature factor field replaced by the MFS scores is created to enable visual inspection of functionally important regions using molecular graphics software . If the structure file contains many chain breaks in the ATOM records , the user can additionally submit the complete sequence so that more accurate sequence alignments can be generated for the query protein . If users only submit amino acid sequence information , then the SeqonlyMFS generation protocol will be used to predict functional sites . For an average sized protein with 200 residues , the computation for SeqonlyMFS can be performed within one hour , while the computation for structure-based MFS can be performed within one day , when the processing queue is not busy . This server will be continuously updated when our MFS generation protocol is refined and improved . The standalone source code used for the MFS generation can also be downloaded at the same URL .
Evaluating the performance of our meta-functional signature ( MFS ) protocols required us to use a “gold standard” functional site dataset of proteins with known structures . We did not use the “SITE” records in PDB files or “ACT_SITE” records in Swiss-Prot files because these annotations are generally not well-defined and contain high error and low coverage rates [50] . Instead , we used the Thornton dataset [50] and the Lovell dataset [34] , which have been used in previous experiments [19] , [36] . The Thornton dataset contains hand-annotated enzyme active sites extracted from the primary literature; the Lovell dataset contains manually compiled ligand binding sites based on literature . We used the ROC score and the top-10 hits score to evaluate performance , as previously described [19] . To investigate the added value of each component of the meta-functional signatures , we compared the performances of the incremental components of MFS: sequence conservation ( HMM_rel_ent ) , evolutionary conservation ( SSR ) , amino acid type ( AAType ) , position structural stability ( RAPDF_spread ) , and residue structural stability ( RAPDF_dif ) ( Figure 1 ) . Sequential incorporation of each component improves performance . The MFS using the maximum number of components has the best performance in predicting functional sites . High correlations between components ( independent variables ) in a linear model will tend to destabilize the model parameters and give erroneous statistical significance . To investigate whether our MFS models have such problems , we checked the variance inflation factor ( VIF ) . The VIF is a measure for each independent variable to estimate how collinearity among variables affects the precision of parameter estimation . VIF scores higher than 10 generally indicate problematic models . We found that all VIF scores for the parameters in MFS models when applied to both datasets are less than 4 , indicating that our models do not suffer from collinearity problems . In addition , we calculated the pairwise correlation coefficients between the HMM_rel_ent score , the SSR score , the RAPDF_spread score , and the RAPDF_dif score for both datasets ( Table 1 ) . We found that the highest absolute value of correlation coefficient is 0 . 45 between the HMM_rel_ent and SSR scores . Therefore , each component of the MFS protocol provides additional and predominantly orthogonal information , and they can be used individually to assess the different aspects of function . Several web servers have been established that assign quantitative scores to functionally important amino acid residues , and map these scores to protein structures for identifying the spatial clusters of important residues . We compared the performance of MFS with two such web servers , the Evolutionary Trace ( ET ) server and the ConSurf server . The ET server implements a method that combines evolutionary and entropic information to rank each residue by its functional importance [23] , while the ConSurf method uses phylogenetic information to measure residue conservation [24] . Although both the ET and the ConSurf methods map the scores to protein structures , these methods do not use structural information explicitly in their calculation of functional importance . Therefore , for comparison purposes , we also used the SeqonlyMFS method , which does not use structural information . We used the same datasets and performance measures described in the previous section to compare these methods . However , since the ET server and the ConSurf server produced error messages or could not handle some proteins , we focused our analysis on the 453/508 proteins in Thornton dataset and the 226/243 proteins in Lovell dataset for which both servers generated outputs ( Figure 2 ) . In addition , we did not calculate top-10 hits scores for the ET server , because for any given protein this server typically generates many more than 10 equal scores tied at first place . We found that MFS and SeqonlyMFS outperform both servers when their ROC measures were compared: for the SeqonlyMFS and ET comparison , the sign test P-values were 1 . 2e-25 and 4 . 4e-15 for the Thornton and Lovell datasets , respectively; for the SeqonlyMFS and ConSurf comparison , the P-values were 1 . 4e-39 and 1 . 3e-16 , respectively . In addition , the SeqonlyMFS and MFS generated significantly more top-10 hits than the ConSurf server for both datasets . We note that in real-world applications , it is more important to evaluate the performance when only the most confident predictions are given; therefore , we also compared the precision measure and the false positive rate when 20% sensitivity is achieved for each protein . For both measures , MFS still has the best performance among all the methods ( Figure 2 ) . Finally , since each protein may have a variable number of functional sites , the sum of top-10 hits for all proteins may not be an optimal measure of the expected performance for a given protein . We therefore calculated the sensitivity of each method for each protein . For the Thornton dataset , the average sensitivity values for all proteins are 67 . 0% , 62 . 5% , and 33 . 7% for MFS , SeqonlyMFS , and ConSurf , respectively . For the Lovell dataset , the average sensitivity values are 70 . 0% , 66 . 9% , and 40 . 8% , respectively . Altogether , compared with methods that use only one source of information , the MFS approach that combines multiple sources of information can give improved performance in predicting functionally important residues . The MFS method can be regarded as a tool to define protein function as a series of quantitative values . Alternatively , when considering each component , MFS can also be treated as several vectors with equal dimensions . In previous sections we have demonstrated the application of MFS in functional site identification . Here we also demonstrate the use of MFS in other types of computational biology problems using four examples .
In this work we describe a meta-functional signature ( MFS ) generation protocol that combines multiple sources of information for protein functional site prediction . We also demonstrate the ability of this protocol to characterize protein function on a per-residue basis using four real-world examples . The key ideas presented in this study include the separation of structural and functional contributions , the use of pseudo-energy functions for mutated structures to determine their effects on protein function , and the combination of knowledge- and biophysics-based approaches to comprehensively annotate the functional importance of residues in a protein sequence . Most of the components of our approach are not unique: other function prediction algorithms use multiple sequence alignments , database information , and experimental and predicted protein structures . One unique aspect of our approach is in the integration of all the components into one unified knowledge- and structure-based framework that can achieve more accurate and more comprehensive predictions , yet each component can also provide different aspects of biological insight into the interpretation of protein function . Since two different datasets ( the Thornton set and the Lovell set ) from different sources have been used in our study , we wish to compare and discuss the model parameters for different datasets here . This analysis may help us understand the relative contribution of the different scoring components in the two datasets . To account for the different magnitude of the predictor variables , we calculated the slope of the regression coefficient when transforming all predictors to Z-scores . For the Thornton dataset , the slope for the normalized HMM_rel_ent , SSR , RAPDF_spread , and RAPDF_dif are 1 . 1 , 0 . 25 , 0 . 52 , and 0 . 23 , respectively; for the Lovell dataset , the corresponding values are 1 . 1 , 0 . 28 , 0 . 45 , and 0 . 19 , respectively . Therefore , for the Thornton dataset that contains catalytic sites , the model contains slightly more contribution from structure-based scores , indicating that structure information is relatively more important in inferring catalytic sites than binding interfaces . In addition , we also compared the relative contribution from the 20 amino acids to the model . For the Thornton dataset , the five amino acids with the strongest contributions are Glu , Lys , Asp , Arg , and Ser , respectively , with normalized coefficients ranging from 0 . 55 to 0 . 83 . For the Lovell dataset , the five amino acids with the strongest contributions are also Glu , Lys , Asp , Arg , and Ser , respectively , with normalized coefficients ranging from 0 . 66 to 0 . 84 . Therefore , the amino acid identity seems to play equally important roles in these two datasets . We note that “functional residues” in the context of this study represent both catalytic sites and binding sites , yet due to the limitations of the data sources , each test dataset only contains part of the true functional sites , so some true positive hits may be mistreated as non-functional sites in each dataset . Besides comparison of two datasets , to evaluate the stability of the regression models , we have also performed similar analysis by comparing the five sets of models used in cross-validation experiments , and found that the model parameters are mostly identical between cross validations ( data not shown ) . Although we have presented MFS as an ensemble of scoring components integrated by a simple logistic regression model , an alternative way to integrate information is to use a sophisticated machine-learning approach , for example , via SVM based algorithms . We investigated this issue but decided to use the regression model due to several reasons: First , although SVM is well known to perform well on binary classification problems , it suffers from a lack of “biological” interpretation . For example , Petrova et al evaluated 26 different algorithms/classifiers in the WEKA software package , and presented the best combination of components as a set of seven ( out of 24 ) residue properties for predicting catalytic residues [37] . Furthermore , Youn et al tested SVM on 314 different features , demonstrated that the combined use of multiple features improves performance , and presented the most highly ranked features [39] . Pugalenthi et al . tested 278 different features for catalytic site prediction and investigated the performance when a subset of 50–250 features are used [40] . Although these machine-learning approaches usually lead to improved performance , it is difficult to decode these “black box” methods and use an individual component ( out of dozens or hundreds ) to interpret different aspects of biological function , as we have done with MFS on four real-world examples . Therefore , in these cases , a simple logistic regression model is a conceptually better choice , where the regression parameters are easily intelligible . Second , functional importance may be efficiently captured by several largely independent features in a simple linear model , without resorting to testing many more complicated models and selecting the best performing model . For example , in Figure 1 of Petrova et al , although SVM ranks higher than logistic regression when comparing many different algorithms , the performance of these two methods is indeed highly similar . Therefore , we relied on a simple logistic regression model as the best approach to present and integrate an ensemble of knowledge- and biophysics-based methods in MFS . More than just another functional site prediction algorithm , MFS can be used as a way to define protein function via a series of quantitative values that captures the functional importance of the protein . By abstracting protein function into a vector ( or several vectors if each individual component is considered separately ) , more sophisticated algorithms can be applied to use this information more efficiently . Traditionally , two proteins can be aligned together based on their sequence similarity , structure similarity , or sequence-structure compatibility . However , the introduction of the MFS concept makes it possible to generate functional alignments between the two proteins . For example , we have demonstrated that by comparing the MFS scores for two proteins , we can potentially improve alignment accuracy using functional signatures in a manual manner . However , an automatic algorithm for aligning two variable-length matrices is non-trivial . Algorithmic advancements are needed to find an optimal solution to perform automated functional alignments for two proteins . We are actively pursuing approximate solutions to this problem . Besides the functional site identification methods used in the paper , we realize that many other different types of methods exist to identify important residues from protein sequence or structure . Many of the methods are based on a continuous stretch of amino acid patterns , for example , the PROSITE pattern [62] and the BLOCKs pattern [63] . All residues in a given protein that match particular motifs are regarded as functionally important and the properties of the motifs may also suggest specific functional roles for the protein . However , these methods usually result in a significant over-prediction of “functional site” residues; for example , some PROSITE patterns are composed of 3-residue motifs that match multiple sites in multiple proteins . Therefore , while these methods are useful for confirming whether a pattern corresponding to a biological function exists , or for hypothesis generation to predict the possible functional category , these methods are usually too general for defining functional importance on a per-residue level . We regard our method and the motif-scanning methods as ideologically different methodologies to solve similar problems . Together they may help users gain complementary biological insights for protein characterization . The MFS generation protocol can be enhanced in several ways . One advantage of the MFS concept is that it is composed of several independent modules , so each module can be updated and improved , without disrupting functionality of other modules . We are improving the performance of MFS from multiple aspects . First , while many other web servers ( such as SIFT ) use the entire NR or the entire TrEMBL sequence collection , we used only the Uniref90 data , thus allowing us to speed up BLAST searches . However , the Uniref90 dataset is not of high-quality . Many extremely short sequences exist and can be easily incorporated into the alignments and many unknown amino acids are annotated as long stretches of “X” . In addition , we used the PSI-BLAST program to scan the sequence database and generate multiple alignments , which are in fact simply the pile-up version of multiple pairwise alignments . The generation of more accurate multiple alignments will help sequence-based conservation estimations and phylogeny inferences . Furthermore , the RAPDF calculation for mutated structures can also be optimized . An optional step after side chain replacement is to minimize energy by global perturbation of the structure . This step can be implemented by the ENCAD protocol [48] . Since this procedure significantly increases execution time we made it an optional step . A faster generation of more accurate structural stability scores for mutated structures would improve MFS performance . Further development and optimization of the current protocol will greatly improve the functional annotation of sequence and structure space . Besides improving the performance of protein functional site prediction , MFS scores treated as vectors may be used to discern functional categories for a given protein ( for example , assignment of SCOP superfamily [35] , [64] or a GO node in the GO hierarchy ) . MFS analysis also elucidates functional importance on a per-residue level , which enables the design of rational mutagenesis and biochemical experiments . Finally the MFS method may be used to modify protein function , resulting in application to protein design and drug discovery . The application of MFS protocols to many areas of computational biology and bioinformatics , as shown by examples in the paper , may significantly advance our understanding of protein sequence-structure-function relationships and guide experimental characterization of protein function . | Proteins are the main building blocks and functional molecules of the cell . Function is mediated by specific amino acid residues in a protein sequence , in a manner dependent on both their positions and types . Proteins are traditionally described as a sequence of amino acids and , when known , the experimentally determined coordinates of this covalently linked chain . Here we propose to expand the description of a protein to include a quantitative measure of the functional importance for each constituent amino acid . The resulting signature for a protein sequence or structure is referred to as its meta-functional signature ( MFS ) . We present an ensemble of knowledge- and biophysics-based methods , which exploit different types of evidence for functional importance , as an automated publicly available tool to build such an MFS . We use two benchmark datasets to show that MFS can be used to identify functionally important residues from protein structure or sequence alone . Finally , we assess four diverse real-world biological questions to demonstrate the ability of MFS to give insight into the structural and functional roles of individual residues and positions , by exploiting protein sequence–structure–function relationships . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"computational",
"biology/protein",
"structure",
"prediction",
"genetics",
"and",
"genomics/bioinformatics",
"computational",
"biology/genomics",
"genetics",
"and",
"genomics/functional",
"genomics"
] | 2008 | Protein Meta-Functional Signatures from Combining Sequence, Structure, Evolution, and Amino Acid Property Information |
When making judgments in a group , individuals often revise their initial beliefs about the best judgment to make given what others believe . Despite the ubiquity of this phenomenon , we know little about how the brain updates beliefs when integrating personal judgments ( individual information ) with those of others ( social information ) . Here , we investigated the neurocomputational mechanisms of how we adapt our judgments to those made by groups of different sizes , in the context of jury decisions for a criminal . By testing different theoretical models , we showed that a social Bayesian inference model captured changes in judgments better than 2 other models . Our results showed that participants updated their beliefs by appropriately weighting individual and social sources of information according to their respective credibility . When investigating 2 fundamental computations of Bayesian inference , belief updates and credibility estimates of social information , we found that the dorsal anterior cingulate cortex ( dACC ) computed the level of belief updates , while the bilateral frontopolar cortex ( FPC ) was more engaged in individuals who assigned a greater credibility to the judgments of a larger group . Moreover , increased functional connectivity between these 2 brain regions reflected a greater influence of group size on the relative credibility of social information . These results provide a mechanistic understanding of the computational roles of the FPC-dACC network in steering judgment adaptation to a group’s opinion . Taken together , these findings provide a computational account of how the human brain integrates individual and social information for decision-making in groups .
When making decisions in a group , individuals can adapt their initial beliefs according to the social influence produced by the opinions of other individuals in the group . In modern society , this type of process is widespread and can be seen in settings ranging from work meetings to the courtroom . In these instances , individuals are asked to come together to produce a group decision while considering the opinions of others [1 , 2] . An essential component of this process is the need for individuals to adapt their opinions to others to reach a collective decision [3] . Determining the factors that underlie such judgment adaptation is , therefore , critical to understanding collective decisions . A key factor identified in social psychology that is known to influence whether subjects change their decisions to conform to those of others is group size [4 , 5] . Specifically , the larger the group size , the more people conform to the group’s opinion , but only up to a certain point [6 , 7] . To benefit from the opinion of others , the brain needs to track the likelihood that the group is making the most appropriate judgment [8] by estimating the credibility of social information [9–11] . One potential way to make such judgment adaptation to the group’s opinion possible is to assign greater credibility to a larger group’s opinion than to a smaller group’s opinion . Evidence for this mechanism come from studies comparing 6- and 12-person juries , showing that a larger jury is more likely to overcome its biases [12 , 13] and to obtain a result that represents the population mean more accurately than a smaller jury [14 , 15] . A statistical phenomenon that is known as the “wisdom of the crowd” also explains how the aggregated opinion of a group of individuals can be even more accurate than the estimates of experts [16–18] . Herding behavior in purchasing decisions [19 , 20] and collective animal behavior demonstrate a similar effect of group size on conformity [21 , 22] . These examples indicate the importance of group size for social decision-making . Assigning credibility to a more reliable source of information and reducing uncertainty are not only critical for learning appropriately but also to adapt to a group’s opinion [9 , 23] . However , little is known about how the brain estimates the credibility of aggregate opinions of groups of different sizes . A mechanistic account of how the human brain integrates individual and social information for making decisions in groups is still unclear . Previous studies have established that people track and use social information based on its credibility [24 , 25] . However , so far , studies on social conformity have only accounted for the fact that changes in judgments within a group are driven by the motive to decrease social conflict [26–28] . In contrast , here , we considered that the changes in judgments in a group could be driven by belief updates that consider the credibility of each source of information rather than by social conformity . Such belief updates are distinct from social conformity , which exclusively concerns the difference between individual and social judgments . Moreover , previous studies did not take into account the confidence that people have in their own judgments and the influence of this confidence on judgment adaptation [11 , 29] . Yet , confidence in one’s decision plays a key role in revising one’s previous decision [30 , 31] , and this is also likely to be true when integrating social information . Indeed , people are more likely to follow social information when they lack confidence in their own judgment [9] . Thus , we hypothesized that both confidence in one’s initial decision and group size are critical to account for integration of individual and social information during group decision-making . Here , we examined the neural mechanisms enabling the consideration of the credibility of both individual information and social information . We first investigated whether the credibility of individual information and the credibility of social information are respectively modulated by one’s own confidence level and group size . Second , we investigated how the brain adapts to the judgments of others according to the credibility of each source of information . To address these questions , we tested different computational models that accounted for the changes in one’s judgments within a group as an adaptive decision-making process . Moreover , to characterize the neurocomputational mechanisms engaged in judgment adaptation , we used model-based functional MRI ( fMRI ) and a new paradigm in which participants were asked to make a series of punitive judgments on murder cases as part of a jury ( Fig 1A ) . First , they made a judgment on the appropriate punishment ( in prison years ) for a criminal ( first individual judgment [J1]; default was 15 years ) along with an estimation of their level of confidence ( C ) in each judgment . They were then given the opportunity to reconsider and to update their first judgment ( J1 ) to make a second judgment ( J2 ) with the given knowledge of the average judgment of the other members of the jury ( judgment of a social group [JS] where “S” stands for social information ) . Crucially , the size of the juries varied , being either large ( 20 jury members ) or small ( 5 jury members ) groups . Thus , the credibility of social information available to jurors could be estimated as either high or low , respectively ( Fig 1B ) . We developed and tested a Bayesian model of decision making under social influence . Neuroimaging studies of perceptual decision-making provide evidence that an observer has a mental model to conduct Bayesian inferences to infer probable states of the world from their observations [32 , 33] . In the field of social decision-making , although recent theoretical models proposed that such a Bayesian framework may be applied to understand decision-making under social influence [23 , 34] , empirical evidence for characterizing the neural signals of such social decisions is still lacking . Our Bayesian model assumes that people process their own judgments and the judgments of others as distinct probability distributions considering the likelihoods of the most appropriate judgment . In this framework , the level of precision of each distribution represents the credibility that an individual assigns to each type of information [8 , 35 , 36] . We tested whether participants use , on one hand , the confidence levels in their own judgments to infer the credibility of their individual information and , on the other hand , the size of the group to estimate the credibility of social information . In such a Bayesian account , individuals integrate different sources of information and generate a novel belief [37] , suggesting that the brain assigns different weights to individual and social information according to their respective credibility . First , we compared our Bayesian model with alternative models . These alternative models had access to the same information about the choice options but assumed different computations when participants were presented with the judgments of others . We found that the Bayesian decision-making model explained the magnitudes of judgment changes of jurors better than other alternative models . That is , jurors integrated their individual levels of confidence and estimated levels of credibility of social information to construct a novel belief about what would be the most proper judgment and adjusted their judgments accordingly . Second , we investigated the neural implementations of 2 fundamental computations of Bayesian inference during collective decisions , belief updates and credibility estimates of social information . We found that activity of the dorsal anterior cingulate cortex ( dACC ) reflected updates of one’s beliefs after integrating social information , rather than alternative computations from the other 2 models . For Bayesian computation , the belief update signals should monitor the cognitive process tracking changes in the credibility of social information . We thus further investigated which brain regions estimated the credibility of social information during judgment adaptation . To this end , we reasoned that if a region is engaged in computing the credibility of social information: ( i ) it should show greater activity when participants were given more-credible judgments by the large group than when given less-credible information by the small group; ( ii ) this region should provide greater inputs to the dACC to modulate one’s belief with increases in the credibility of social information , and this may be reflected by greater functional connectivity between this region and the dACC at the time of judgment adaptation; and ( iii ) both the activity and the functional connectivity of this region may represent individual differences in the willingness to assign greater credibility to judgments of the large group compared to judgments of the small group .
The mean punishment magnitude at the stage of the initial judgment ( J1 ) was 14 . 44 ± 0 . 25 years ( standard error of the mean [SEM]; Fig 1C , left graph ) . For trials in which the jury group was small , the mean judgment magnitude was 14 . 17 ± 0 . 35 years , and , for trials in which the jury group was large , the mean judgment magnitude was 14 . 71 ± 0 . 36 years . Magnitudes of initial punishment judgment did not differ between trials of different jury group size ( P = 0 . 28 ) . Participants changed their initial judgment , J1 , to conform to the group opinion ( JS ) in 27 . 58 ± 1 . 47 trials on average . In the rest of the trials ( 24 . 42 ± 1 . 47 ) , participants kept their initial judgments . We then investigated how the opinion of others influenced the second judgment ( J2 ) by separating the cases in which the initial judgment was either lower or higher than the judgment of others . The range of the difference between the initial judgment and social information ( D = JS − J1 ) was , by design , divided into 2 intervals of punishment years , and the group never had the same level of punishment judgments as individual participants ( J1 ≠ JS; see Materials and methods ) . In addition , the mean confidence rating was 0 . 69 ± 0 . 22 , being normalized on a scale from 0 , no confidence at all , to 1 , high confidence . We confirmed that all participants chose the lowest and the highest confidence rating at least once . Confidence did not differ between the size of jury group ( F = 4 . 91 , P = 0 . 27 ) or between blocks ( F = 3 . 81 , P = 0 . 34 ) . We observed that individuals tended to increase the magnitude of their punishments ( J2 = 18 . 93 ± 0 . 33 years , P < 0 . 001 ) when the judgments of group were more severe than those of the juror ( JS > J1 ) and tended to decrease the magnitude of their punishments ( J2 = 10 . 82 ± 0 . 30 years , P < 0 . 001 ) when the judgments of group were more lenient than those of the juror ( JS < J1 ) . To investigate whether changes in judgments ( J2 − J1 ) were driven by the perceived difference in judgments ( J1 − JS ) , we computed the level of conformity ( LC; Eq 1 ) . Based on this measure , conformity trials were defined as LC > 0 and nonconformity trials as LC ≤ 0 . The average level of conformity was strictly positive ( LC = 0 . 24 ± 0 . 01 ) , suggesting that the effect of social influence was significant . Next , we investigated how different levels of confidence assigned to the initial judgments ( J1 ) influenced the level of changes in judgments . After median splits of all trials based on each individual’s confidence ratings , we observed that confidence had a similar effect on changes in judgments ( from J1 to J2 ) . That is , on average , people tended to conform more to the group opinion when they had low confidence in their initial judgment . When one’s initial judgment was more lenient than the judgment of the group , individuals tended to increase the magnitude of their punishments more when they were less confident of their initial judgment ( J2L ( low confidence ) = 19 . 88 ± 0 . 36 years , P < 0 . 001 ) compared to when they were more confident of their initial judgment ( J2H ( high confidence ) = 16 . 64 ± 0 . 46 , P < 0 . 001 , Fig 1C , middle graphs ) . When the initial judgment was more severe than the judgment of the group , individuals tended to decrease the magnitude of their punishments more when they were less confident of this initial judgment ( J2L = 9 . 88 ± 0 . 41 , P < 0 . 001 ) compared to when they were more confident of this initial judgment ( J2H = 12 . 21 ± 0 . 42 , P < 0 . 001; Fig 1C , right graph ) . To clarify how the factors manipulated in the task influence changes in judgments ( J2 − J1 ) , we performed a statistical analysis using the linear mixed-effects modeling analysis ( LMEM ) . We found that changes in judgments significantly depended on the difference between judgments ( D , F = 31 . 79 , P < 0 . 01 ) , their confidence levels ( C , F = 2 . 67 , P = 0 . 003 ) , their interaction ( D × C , F = 2 . 45 , P < 0 . 01 ) , and the interaction between the difference in judgments and group size ( D × G , F = 2 . 05 , P = 0 . 02 ) . These changes did not depend on the group size ( G , F = 3 . 49 , P = 0 . 12 ) , the types of scenario ( S , F = 1 . 05 , P = 0 . 39 ) , the regression-to-the-mean effect of the initial judgments ( |J1| , F = 3 . 12 , P = 0 . 12; see Materials and methods for details ) , and the interaction between the difference between judgments and types of scenario D × S ( F = 1 . 19 , P = 0 . 28 ) . To illustrate this effect more clearly , we performed additional factorial analyses . By splitting all trials with high- and low-confident judgments using each individual’s median value of confidence rating , we found that participants were more likely to conform to others when confidence in their judgments was low ( F = 61 . 64 , P < 0 . 001 ) , which provided additional evidence supporting the effects of confidence level ( C ) on decisions to conform to the judgments of others ( LC ) . We further examined whether some of these experimental factors are correlated with each other . The level of collinearity between experimental factors is shown in Fig 1D . As illustrated , we only found a relationship between initial judgments ( |J1| ) and level of confidence ( C ) . That is , participants tended to make more extreme judgments ( close to 0 or 30 years ) when they had greater confidence in their judgments ( R2 = 0 . 43 , P < 0 . 01 ) . This correlation might be caused by the “status quo”—the tendency of decision-makers to stay close to the given default when having low confidence in their decision [38 , 39] . To examine how the brain computes social updating during group decision-making and to know whether and how much the brain decides to change initial judgments , we proposed and compared 3 computational models . The “linear model” ( Eq 7 ) predicts changes in judgments by a linear regression model that takes into account one’s initial judgments ( J1 ) , given deviation in the group judgments ( D = JS−J1 ) , the level of confidence that one had for their judgments ( C ) , and 3 interaction effects considering group size ( G ) and the types of scenarios ( S ) ( D × C , D × G and D × S ) . In doing so , the “linear model” tests whether individuals are motivated to reduce the perceived deviation in the group judgments by adjusting their prior judgments to be more consistent with that of the group . The “surprise model” ( Eq 8 ) assumes that participants change their judgments more when they are more likely to observe unexpected group judgments , given their prior belief about the most preferable judgments ( J ) . Thus , surprising events have greater impacts to drive changes in judgments . According to information theory [40] , surprise evoked by the observation of unpredicted event α is computed as the inverse entropy of the predictability of the event , participants having their own prior belief ( −log p ( α|Prior ) ) . The level of precision of the prior belief , thus , was computed as the variance of a normal distribution , which depends upon a function of confidence ratings , ( p ( J1|Prior ) ~ N ( μ = J1 , σ2 = f ( C ) −1 ) . Moreover , we assumed that participants were more surprised when the deviation ( D ) was given from the judgments of a larger group ( D × G ) and when the group proposed a more severe ( or more lenient ) level of punishment than their own when making judgments for the sympathetic ( or nonsympathetic ) scenarios ( D × S ) . The “Bayesian model” ( Eq 6 ) assumes that participants tried to estimate the most preferable judgments ( J ) . The punishment—the years of prison that the defendant deserved—is dependent on this estimated judgment . Initial judgments ( J1 ) were based on individuals’ private estimates ( jIndividual ) after reading the scenario of a crime case . The distribution , p ( jIndividual|J ) , represented the probability that their private estimate was the most preferable . The credibility of individual information , thus , was represented as the level of precision of a normal distribution that depends on a function of confidence ratings , p ( jIndividual|J ) ~ N ( μ = J1 , σ2 = f ( C ) −1 ) . When social information is presented , participants may assign different levels of credibility as a function of changes in group size , p ( jSocial|J ) ~ N ( μ = JS , σ2 = f ( G ) −1 ) . More importantly , to explain that changes in judgments are driven by changes in one’s belief about what would be morally correct and what level of punishment would be the right amount for the crime , the “Bayesian model” integrates available information while considering their level of credibility . Therefore , this model also assumes that the brain computes the extent of belief updates . The extent of belief updates was well captured with Kullback–Leibler divergence ( DKL ) , which shows the dissimilarity between 2 probability distributions of the individual estimate , p ( jIndividual|J ) , and final estimate , p ( J|jIndividual , jSocial ) , of individuals ( Fig 2 ) . The parameters that maximize the likelihood of the model based prediction of the actual changes in judgments were estimated for each of the 3 computational models . The parameter estimates in the “linear model” suggest that changes in judgments were significantly driven by the difference in judgments ( D ) , the level of confidence in the initial judgments ( C ) , and their interaction ( D × C ) ( P < 0 . 05 , 1-sample t test , Fig 3A ) . Notably , both the “surprise model” and the “Bayesian model” assume that the confidence level in one’s own belief serves to estimate the level of precision of the prior belief . In contrast to the “Bayesian model , ” in which greater credibility is assigned to more confident judgments , the “surprise model” assumes that participants are more likely to be surprised by the unexpected judgments of others , especially when they had a high degree of confidence in their own judgment [33 , 40] . However , the parameter estimates showed that the “surprise model” could not capture the different levels of judgment adaptation when confidence varied ( Fig 3B ) . The “Bayesian model” not only took into account the punishment magnitude of individual judgments , J1 , and the judgments of others , JS , but also the beliefs that the participant had about the credibility of individual information and social information , respectively . This framework allowed us to find the credibility that participants assigned to their individual information as a function of confidence rating and its updates after integrating social information according to its credibility . The relationship between confidence reports and the credibility of individual information was measured by a parameter , ωC , across participants ( Eq 2 ) . The mean parameter estimate of ωC was 1 . 13 ± 0 . 47 , which was significantly greater than 0 ( P < 0 . 001 , 1-sample test ) , suggesting its significant effect . That is , participants put more credibility on their own judgments when they had higher confidence ( F = 15 . 63; P < 0 . 001; LMEM; Fig 3C , left graph ) . Moreover , we tested whether participants took into account changes in group size when estimating the credibility of social information through the integration of social information with their prior individual information . The “Bayesian model” provided evidence of significant effects of group size on the credibility of social information . Again , we observed that participants assigned more credibility to the judgments of a larger group ( F = 121 . 77; P < 0 . 001; 1-way ANOVA; Fig 3C right graph ) . Among the models , the Bayesian model offered a better fit than the others ( comparing the likelihood of each model , P < 0 . 001 , 1-way ANOVA; Eq 10 ) . The mean –2log likelihood of the Bayesian model was 8 . 55 ± 0 . 16 , while those of the “linear model” and the “surprise model” were 14 . 12 ± 0 . 25 and 20 . 62 ± 0 . 44 , respectively . To avoid potential overfitting , we predicted each of the levels of judgment adaptation across trials of all participants using an iterative leave-one-out-cross-validation ( LOOCV ) procedure and estimated the likelihood of the prediction and actual changes in judgments . Furthermore , we penalized the likelihood and computed the Bayesian information criteria ( BIC ) that took into account the number of free parameters in each model ( P < 0 . 001 , 1-way ANOVA , Eq 11; Fig 3D ) . We found that dACC activation ( peak voxel x , y , z = −3 , 14 , 44 ) increased when participants reconsidered their initial judgment and adapted it towards the judgment of the group ( P < 0 . 05 , family-wise error [FWE] corrected within small volume correction [SVC] , T = 4 . 45 , Fig 4A ) . We also found dACC activation in the same cluster ( peak voxel x , y , z = 0 , 12 , 39 ) when an alternative general linear model ( GLM ) with a fixed boxcar duration independent of reactions times ( = 4 s ) was applied ( P < 0 . 05 , FWE corrected within SVC , T = 4 . 75; S1 Fig ) . This alternative model allowed us to exclude alternative hypotheses regarding dACC functions as reflecting speed of response or invigoration rather than judgment adaptation . Having established that the Bayesian model better accounted for the behavioral data than other models , we tested whether the changes in judgment predicted by the Bayesian model more accurately explain the conformity-related brain signal in dACC than those predicted by the other models . To address this , we used an additional GLM ( GLM3 ) in which the brain response was modeled by 4 parametric regressors without serial orthogonalization . We then extracted the time courses of beta parameters from the dACC ROI . The goodness-of-fits of those time courses were measured with the likelihood of predicting the dACC time courses that represented the changes in judgment . The mean value of −2log likelihood and BIC ( in parenthesis ) are as follows: 2 . 37 ± 0 . 14 ( 6 . 32 ± 0 . 14 ) for the Bayesian model , 4 . 68 ± 0 . 12 ( 8 . 63 ± 0 . 12 ) for the “linear model , ” and 5 . 23 ± 0 . 19 ( 9 . 18 ± 0 . 19 ) for the “surprise model . ” In addition , we computed BIC to penalize likelihoods by the numbers of free parameters used by each model . We found that the time courses extracted from the “Bayesian model” explained conformity decision–related dACC activity more robustly than time courses extracted from the other models ( P < 0 . 001 , 1-way ANOVA , Fig 4B ) . We also performed the same analysis with an alternative GLM using a fixed boxcar duration of 4 s . The time courses of beta parameters were extracted from the same predefined dACC ROI . The goodness of fit of those time courses of model predictions was measured with the likelihood and BIC . The mean value of −2log likelihood and BIC ( in parenthesis ) are as follows: 2 . 69 ± 0 . 20 ( 6 . 64 ± 0 . 20 ) for the Bayesian model , 5 . 16 ± 0 . 11 ( 9 . 11 ± 0 . 11 ) for the “linear model , ” and 5 . 80 ± 0 . 17 ( 9 . 75 ± 0 . 17 ) for the “surprise mode . ” We found that the Bayesian model was significantly lower than those of other models ( P < 0 . 001 , 1-way ANOVA ) . To ensure that dACC activation cannot be explained by response selection difficulty , as suggested previously by Shenhav and colleagues [41–43] , we performed a similar analysis to the one from O’Reilly and Kolling [33 , 44 , 45] by including , in the same GLM , the update term regressor DKL as well as the level of surprise and reaction times ( GLM4 ) . This allowed the belief update regressor to compete with these measures of difficulty to explain variance in dACC activity . The results of this new GLM confirmed the robustness of our dACC activity as reflecting a Bayesian update signal of one’s beliefs after integrating social information ( peak voxel x , y , z = 6 , 8 , 49; P < 0 . 05 , FWE corrected within SVC; S2 Fig ) . This result demonstrates that this dACC signal can be dissociated from tracking choice difficulty . Computationally , tracking changes in the credibility of social information is required to update one’s belief . If dACC activity represents the level of belief updates as evidenced by the results of the “Bayesian model , ” then this brain region would require feedbacks when the credibility of group judgments changes . Moreover , if the credibility of group judgment is processed separately by other brain areas , functional connectivity to dACC should increase in order to guide subsequent judgment adaptation . We thus tested for brain regions implicated in social credibility during the integration of social information . First , we found that activity in the right lateral frontopolar cortex ( FPC; [x , y , z] = [42 , 44 , 19] ) , in the precuneus ( x , y , z ) = ( 18 , −58 , 31 ) , and in the bilateral inferior parietal lobule ( [x , y , z] = [60 , −34 . 43] for right inferior parietal lobule [iPL] and [−54 , −49 , 52] for left iPL ) was greater when participants were presented with the judgments of large group compared to when presented with the judgments of small group ( Fig 5A; PFWE < 0 . 05 , whole-brain corrected at cluster level; GLM2 ) . Second , to investigate which brain regions show individual differences in relative credibility as a function of group size , we performed a whole-brain regression analysis with Δσ ( Eq 4 ) . We found that the bilateral FPC was the only area representing this parameter: left FPC peak voxel ( x , y , z ) = ( −30 , 58 , 15 ) ; right FPC peak voxel ( x , y , z ) = ( 27 , 57 , 10 ) ( Fig 5B; PFWE < 0 . 05; whole-brain corrected at the cluster level; GLM2 ) . This result indicates that participants who had greater activation in the FPC showed higher differences in the credibility that they assigned to the judgments of a larger group relative to those of a smaller group . By extracting the parameter estimates from FPC activations , we showed a linear relationship between individual differences in FPC activation and sensitivity to group sizes for assigning credibility to social information . Moreover , these effects were significant when we applied bootstrapping sampling: the effect size in right FPC was 3 . 53 ± 0 . 18 ( SEM ) with a 95% confidence interval between 2 . 16 and 5 . 53; the effect size in left FPC was 3 . 11 ± 0 . 24 ( SEM ) with a 95% confidence interval between 1 . 22 and 5 . 68 ( P < 0 . 001 ) . Moreover , to further check this effect , we also tested whether any brain activity represented the scrambled individual parameter ( Δσ ) , in which we randomly assigned each participant’s Δσ to another participant . We found no significant neural correlates of the scrambled Δσ even at a liberal statistical threshold ( P < 0 . 005 , uncorrected ) . Third , given that results indicate that only the right lateral FPC satisfies both criteria , we further tested whether functional connectivity between the dACC and right lateral FPC was modulated by the size of the group while making decisions of judgment adaptation . To do this , we performed psychophysiological interaction ( PPI ) analyses . In the PPI analysis , we took as a seed the right lateral FPC and tested whether its functional connectivity to the dACC was modulated during judgment adaptation . The ROI was defined as a 10-mm diameter spherical ROI at the peak voxel in the right lateral FPC , ( x , y , z ) = ( 42 , 44 , 19 ) , which was identified as being modulated by group size ( large group trials > small group trials; GLM2 ) . We computed PPI maps to identify the brain areas for which functional connectivity with the right FPC increased in trials in which participants were confronted with judgments of the large group compared with trials in which they were confronted with judgments of the small group . We found that changes in group size modulated connectivity between dACC and right lateral FPC ( PSVC < 0 . 05 , FWE corrected within the small volume cluster in the dACC ROI , peak [x , y , z] = [9 , 11 , 43]; Fig 6A ) . We also confirmed that the dACC region for which functional connectivity to the lateral FPC was modulated by group size is located in the same dACC cluster as that observed for conformity decisions ( Conformity > No-conformity; black contour in Fig 6A ) . We further found that individual variability in the strength of functional connectivity between large group ( n = 20 ) trials and small group ( n = 5 ) trials correlated with Δσ , the tendency of individuals to assign greater credibility to the group’s opinion when the group size was large ( Fig 6B ) . These results showed that connectivity between the FPC and the dACC increased in those individuals with higher sensitivity to group size . Thus , both FPC activity and the strength of FPC-dACC coupling influenced the degree to which social information was integrated into the posterior belief of individual participants .
Our findings show that a Bayesian model provides a good account of observed behavior when a judgment based on a private estimate ( individual information ) is confronted by the aggregated opinion of fellow members of a group ( social information ) . According to our Bayesian model , participants used both the confidence in one’s own judgments and group size of social information to estimate the credibility of each type of information . Participants thus weighed their initial judgment and the judgment of others by their respective level of credibility , integrated them into a new belief , and changed their punitive judgments accordingly . Our Bayesian model explains judgment-adaptation behavior better than other models . The Bayesian model differs from alternative models in an important way . It predicts that judgment adaptation should be sensitive to the credibility of both individual and social information , whereas other models predict exclusive sensitivity to the level of social conflict . When individuals made a decision to change their initial judgments to fit in with that of the group , we observed that participants did not simply conform more to a larger group than to a smaller group , nor did they have a higher level of social conflicts with the judgments of a larger than a smaller group . Instead , participants tended to attribute more credibility to a larger than a smaller group . Only the Bayesian model captures this effect . These results explain a long-standing debate about the relationship between group size and social conformity [6] . While some studies have observed that increasing group size does not influence conformity to the group beyond a minimal number [4 , 5] , others have reported that the larger the size of the group , the larger the effect [46] . These studies suggest that the relationship between conformity and group size cannot be described by a simple function but instead varies systematically with factors that impact social-influence processes . At a mechanistic level , our Bayesian model describes the neurocomputational mechanisms underlying judgment adaptation in a group . A number of frontal cortex regions have previously been shown to be engaged during social decision-making , such as the dACC and the FPC . Here , we show that specific signals integrating individual and social information are computed in these 2 regions and shared between them . In particular , we investigated the neural implementations of 2 fundamental computations of Bayesian inference during collective decisions: belief updates and credibility estimates of social information . The dACC computed the belief updates that were necessary to adapt judgments , while the FPC computations reflected the credibility that people assigned to social information . Furthermore , an increase in functional connectivity between these 2 regions predicted individual differences in credibility assigned to the judgments of a larger group , compared to the one assigned to the judgments of a smaller group . The estimate of credibility of social information , computed in the FPC , was critical for efficient Bayesian computation of belief updates , processed in the dACC . By monitoring the changes in the credibility of social information , the FPC may test the validity of individual information and modulate the integration of social information . Our results suggest a general functional role for the dACC , which is to update one’s belief by integrating different types of information according to their respective credibility . Several accounts have proposed that the dACC is engaged during conflict monitoring [45 , 47] , and this general function has been recently extended to the social domain , possibly reflecting computation of the difference between one’s own judgment and those of others ( “social conflict” hypothesis ) [26 , 48 , 49] . However , if conformity is only understood as a resolution of social conflicts , the changes in judgments should exclusively depend on perceived differences in judgment ( D = JS−J1 ) but not on the changes in credibility of each source of information . Thus , the effects of changes in confidence ( C ) , group size ( G ) , and their interaction effects with the perceived differences in judgments ( D × C and D × G ) should be independent from the changes in judgments . Here , by incorporating the changes in credibility of each source information , however , we provide evidence that the Bayesian model captures the variances of judgment adaptation across trials . Specifically , in some instances , individuals maintain their initial judgments , while in other instances , the same individuals change their judgments to conform to the group , even though they are faced with the same level of social conflict . We found that participants with high confidence in their judgments tended to assign greater credibility to individual information . Moreover , the Bayesian model shows that the dACC encodes the updates of the beliefs about the validity of one’s judgment ( DKL ) after integrating the judgments of others . This Bayesian model allows us to explain intra-individual variability in judgment adaptation and to account for dACC activity better than other psychological conformity models . A previous study using a saccadic planning task demonstrated that the dorsal anterior cingulate cortex ( dACC ) was activated when updating internal models about the probabilistic state of uncertain environments while integrating perceptual information [33] . In the current study , we found that belief updates about the best judgment to make also rely upon dACC computations . That is , the activity of the dACC reflected updates of one’s beliefs after integrating individual information with social information , rather than alternative computations from the other 3 models . By providing evidence that the dACC contributes to belief updates in the context of our group decision-making task , our findings generalize this computational role of dACC to the domain of social decision-making . However , we do not claim that this region is engaged in processing social information per se . In fact , other regions of the ACC appear to have much more specialized roles in social cognition [50–52] . In particular , the anterior ACC gyrus may be more engaged in tracking the intention of others or in computing the costs and benefits of acting in social contexts , whereas the ACC sulcus may be more involved in monitoring the value of one’s own action [53 , 54] . It could be argued that dACC activation reflects increases in decision difficulty rather than model updating [41 , 43] . According to this interpretation , dACC signals the need for control when overriding a default belief . In the Bayesian model , the optimal decision of when to update one’s belief depends on when the value of a conformity decision is equal to the value of keeping the previous decision . In many decision-making situations , decision difficulty and belief updates are confounded . Critically , in our study , decision difficulty can be distinguished from belief updates when participants had low levels of confidence in their individual information and were confronted with large differences in social information . In such cases , participants needed to update their belief by a large amount , but at the same time , the decision was easy . If dACC activity represented decision difficulty rather than belief updates , the Bayesian model would not be able to explain the neural correlates of judgment changes better than other models . Indeed , we found that the dACC activity predicting conformity decisions is explained by Bayesian belief updates measures ( referred to as DKL ) , even in a GLM allowing belief updates to compete with measures of difficulty ( level of surprise and reaction times ) to explain variance in dACC activity [33 , 44 , 45] . In addition , if greater surprise caused the larger demands for control ( corresponding to decision difficulty ) , then dACC activity would have been better explained by the surprise model , which was not the case . In addition , our findings demonstrate that the FPC computes the credibility that people assign to social information . That is , those individuals who assign a greater credibility to social information with larger group sizes show higher activity in FPC when integrating the social information of a larger group compared to that of a smaller group . Moreover , our neuroimaging results indicate that FPC activity , and also FPC-dACC connectivity , play an important role in regulating the degree to which group size influenced credibility estimates of social information . Indeed , both FPC activity and the strength of connectivity between the dACC and FPC predicted inter-individual variances in the credibility assignment to the judgments of a group as a function of changes in group size . For adaptive decision-making in ever-changing environments , the human FPC has been reported to serve probabilistic inferences about the credibility of available information to make optimal use for decision-making [55–57] . Moreover , the lateral FPC is involved in monitoring alternative behavioral strategies and in deciding to switch to alternative courses of actions when one alternative strategy becomes more credible in comparison with the ongoing one [25 , 58 , 60] . Activity in this region also reflects individual differences in the extent to which learning is driven by the prediction errors of one’s belief about the best judgments when new evidence is given [59] . Consistent with this role , our results provide evidence that the FPC processes the credibility of alternative opinions . Thus , a key role of the FPC in social decision-making may be to monitor the credibility of social information when individual information is uncertain . This is an essential capacity to promote flexible behavior in environments in which the current decision strategies become unreliable . It is worth noting that the FPC may have evolved to manage our unusually complex social systems [60 , 61] . In nonhuman primates , gray matter density of the FPC has been reported to increase with social network size [62] . The FPC , which develops late both from a phylogenetic and ontogenetic perspective in humans , may serve demands requiring interactions with larger social groups . Such cognitive demands of sociality could place a constraint on the number of individuals with whom we can interact and maintain contact with [60 , 62] . The current findings differ from previous studies on social influence and conformity [49 , 63] by allowing us to dissociate the credibility of social information from changes in confidence/uncertainty in one’s belief . These previous studies reported ventromedial prefrontal cortex ( vmPFC ) activation of choice options that vary with the choices of others [49] , as well as with the level of confidence in the choices of others [63] . Such vmPFC activity may signal increases in the confidence of individuals’ decisions once social information has been integrated with individual information , reflecting the possibility that participants gain reassurance in their choice from the choices of others who are confident in their decision . This interpretation is supported by the recent findings showing that vmPFC activity encodes decision confidence by reflecting the amount of accumulated evidence favoring one option over the others [38 , 64] . Future work will need to measure the level of confidence not only after making the initial judgment but also after the judgment adaptation to clarify how vmPFC , dACC , and lateral FPC interact to flexibly exchange information at these different stages of social decision-making . While most previous research has focused on ACC or FPC functions in isolation [45 , 65] , dACC-FPC interactions have been relatively unexplored . Our results show that a higher degree of credibility to a larger group was reflected by an increase in the connectivity between the dACC and FPC . This change in the strength of the functional connectivity between the dACC and FPC may reflect a readout function of the dACC from the FPC to compute the Bayesian inference , since credibility estimates of social information are required to update one’s belief in an optimal fashion . Thus , our findings suggest the dACC-FPC network plays a key role in the context of group decision-making and judgment’s adaptation in the context of social decision-making . To develop neurocomputational models of social decision-making , we specified which variables are computed for judgment adaptation within a group and how they are implemented in specific brain regions . A key novel aspect of the current study is to determine the behavioral algorithm for belief updates of individuals about the most preferable judgment and examine the neural correlates of this process . This is distinct from past studies asking participants to rate their preferences for goods or attractiveness of faces [26–28] , in which the credibility of different sources of information was not relevant to make a decision of social conformity . The crucial novel contributions of the present study are also as follows: ( i ) to determine how the credibility of individual and social information are computed for updating one’s judgment , taking into consideration one’s confidence and group size , which was ignored in previous preference-based rating paradigms [26 , 27 , 66]; and ( ii ) to test a Bayesian model against plausible alternative models , which were matched with regards to access to information about the choice options . Together , our findings transform the current thinking about the neural basis of conformity and collective decision-making by proposing a neurocomputational understanding about how individuals adapt their judgments by integrating social influences of other individuals in a group . Our study specifies how the human brain benefits from the wisdom of a larger group while preserving confidence in one’s initial judgment [10] . It also delineates the neurocomputational mechanisms at the source of inter-individual differences in assigning credibility to the opinion of groups with different sizes . By identifing the brain region tracking the credibility of social information , our findings also provide a mechanistic account of the computational mechanisms underlying judgment adaptation during collective decision-making .
This study was approved by the Institutional Review Board of the local ethics committee ( Lyon , France , IRB n°A13-37030 ) , and all participants gave their informed written consent . Participants were 25 healthy French volunteers ( age range 20–26 years , 13 males ) . Data from 2 participants were discarded because of excessive movements during scanning . Therefore , data from 23 subjects ( 11 males; mean age 21 . 22 ± 0 . 463 years; error indicates SEM ) were included in the final analysis . During this experiment , participants were asked to make a series of punitive judgments on murder cases as part of a jury while undergoing conditional blocks of fMRI . At the beginning of each block , we displayed the size of the jury for 3 s ( either 5- or 20-person juries ) to inform participants of how many individuals were also making a judgment in the current murder case along with them . In each trial ( 60 trials in total ) , subjects first read the scenario of a murder case for 15 s . They then made 3 successive decisions for each murder case . First , participants were asked to make a judgment about how many years of prison the defendant deserved ( J1 ) . They reported it by moving a continuous numerical sliding cursor from the initial sentence of 15 years to the final value of prison years ranging from 0 to 30 for 8 s . We also informed participants that 15 years is the average prison sentence for murder case trials in France . The trials for which the participants failed to respond within 8 s were excluded from the further analyses . The position of the cursor was marked on the screen with the corresponding number of years in prison to make sure that the participants were aware of their judgment . Second , participants had 6 s to rate the level of subjective confidence ( C ) that they had while making their J1 , using a Likert scale of 10 items from −5 ( low confidence ) to +5 ( high confidence ) by moving a similar cursor on the screen . Third , participants were given a chance to reconsider their first judgment ( J1 ) with the knowledge of the average sentence given by the other members of the jury ( JS ) . Social information was revealed by a green bar on the slider with its corresponding numerical value in prison years ( bottom of the screen ) . During 8 s , the participants could review J1 by moving the cursor from the value they had chosen during their first assessment to one corresponding to their desirable reconsidered judgment ( J2 ) . To encourage fully sincere judgments of participants who may moderate their J1 by considering the future chance of reviewing their judgment ( J2 ) with the knowledge of JS , we instructed them that the chance to review would be given only in some trials . The participants did not have a chance to revise for 8 randomly arranged trials among 60 trials . We analyzed 52 valid trials per participant . Critically , we instructed the participants that they did not need to change their J1 , but that they could if they felt that the changes would be more preferable judgments . A fixation cross was shown for the inter-stimuli intervals ( ISI ) after the scenario presentation , and between judgments ( J1 , C , and J2 ) , and for the inter-trial interval ( ITI ) . Both ISI and ITI durations were randomized from 4 s to 7 s . We presented scenarios in pseudo-random order across participants . Specifically , the order of presentation of conditions of the different size jury group and its combination with sets of scenarios were counterbalanced across subjects . We instructed the participants that the given social information , JS , is the average judgments of some of the previous participants . Specifically , we informed them that the computer randomly selected the judgments of 4 or 19 previous participants according to the condition . Moreover , participants were told that only those individuals who had assigned a higher level of confidence to their initial judgments ( J1 ) than the average level of confidence of all participants were selected . Likewise , participants were informed that their judgments would be presented to the next participants when they reported a high level of confidence . Using this design , participants were explicitly informed that the other jury members would be different in each trial . JS was a computer-generated value . We manipulated the judgments of others to make sure that all participants reconsidered their judgments under influences of all ranges of differences in judgments ( |D| ) . Notably , by design , the group never had the same level of punishment judgments as individual participants ( J1 ≠ JS ) . Simultaneously , we ensured that the judgments of the group were not too different from those of participants’ in order to make participants believe that those were made by other previous participants . To do this , JS was established for each trial by the computer so that the difference ( D = JS − J1 ) between JS and J1 was within the range of 4 years to 10 years ( 4 ≤ |D| ≤ 10 ) . Moreover , to control for the possibility that participants might be able to learn the consistency of the group judgments ( social information ) over the trials , the direction of differences in judgments ( whether social information was severer or milder than one’s initial judgments , J1 ) was pseudo-randomly determined across trials . That is , we ensured that every participant was given judgments of others that were more severe than theirs during one half of trials and that judgments of others were more lenient than theirs during the other half . On average , the participants perceived 6 . 98 ± 0 . 18 ( SEM ) years difference from their initial judgments when confronted with the judgments of others . This was also true within participants while they were making judgments within different jury group sizes ( G ) ( |D¯|= 7 . 10 ± 0 . 25 ( when n = 20; |D¯|= 6 . 85 ± 0 . 27 when n = 5 ) . A crime scenario was composed of the facts ( plain explanation of who did what ) and the circumstances ( contexts and reasons why the defendant committed the murder ) . One hundred murder-case scenarios were initially produced for this study , which included 32 cases collected from a previous study [67] and 68 additional scenarios inspired by real stories taken from the news with the same structure . The length ranged from 50 to 60 words ( mean length = 54 . 95 ± 0 . 36 [SEM] ) . All scenarios were written in 3 sentences . To minimize the bias , we informed participants before the experiment that we changed the name of the defendants to either “Jean” or “Marie . ” Half of the scenarios included the circumstances that were expected to induce sympathetic emotion for the defendant ( sympathetic cases ) , and the other half did not ( nonsympathetic cases ) . Sixty among 100 scenarios were selected based on the responses of elicited sympathetic emotions that were acquired from the 20 different healthy subjects ( 10 males , mean age = 21 . 43 ± 0 . 46 [SEM] years ) using the same scale of sympathetic emotion ratings . The average rating for selected scenarios was −0 . 48 ± 0 . 21 ( SEM ) . They were significantly split into 2 groups according to their rating ( t = −39 . 31 , P < 0 . 001 , 2-sample t test ) . The duration of scenario presentation in the main experiments ( 15 s ) was also decided based on the sample group’s speed of reading . The impact of social information , modulating individual judgments , was measured at the level of conformity ( LC ) . We also compared them under the impacts of different levels of subjective confidence ( C , individual median split ) and different group sizes ( G ) . All the comparisons were performed by LMEM . The Bayesian decision-making model assumed that participants were trying to estimate the most preferable punishment judgment ( J ) for each scenario . There were 2 cues that participants could rely on: ( 1 ) their initial reading of the scenario , which led to a private estimation of individuals ( jIndividual ) of J , and ( 2 ) the estimation made by the social group ( jSocial ) of J . Participants combined these 2 cues to produce a final estimate of J where the value of J maximized the probability distribution , p ( J|jIndividual , jSocial ) . Applying Bayes rule , we evaluated the probability distribution as p ( J|jIndividual , jSocial ) =p ( jIndividual , jSocial|J ) ×p ( J ) /p ( jIndividual , jSocial ) Note that in the above equation , p ( J ) was the prior belief about J ( i . e . , the belief about J even before receiving any information ) . In the current study , this prior belief was considered to have uniform distribution ( no biases ) . By considering the social and the individual estimates to be independent in this study , we could estimate the p ( jIndividual|J ) as below . The p ( jIndividual|J ) was assumed to be Gaussian , ~ Norm ( J1 , τ2 ) , with mean , J1 , and variance , τ2 . The credibility of the individual estimate was thus 1/τ2 , which was dependent on the level of confidence reports . Similarly , we considered p ( jSocial|J ) to be Gaussian , ~ Norm ( JS , σ2 ) , with mean , JS , and variance , σ2 . The credibility of the social information p ( jSocial|J ) was thus 1/σ2 , which was dependent on the group size . Specifically , when a participant was confronted with the judgments of groups of different sizes , she might assign a different level of credibility to social information ( e . g . , the credibility 1/σ202 was assigned to jSocial when the group was large; the credibility 1/σ52 was assigned to jSocial when the group was small ) . Inter-individual differences in estimated credibility of social information were measured by their ratio , called “relative credibility” ( Δσ ) . This parameter Δσ indicated , therefore , the individual variability of the sensitivity to the changes in group size when participants assigned credibility to social information when confronted with the judgments of others ( JS ) . When the prior belief , p ( J ) , was given as a uniform distribution , the final distribution , p ( J|jIndividual , jSocial ) , also followed a normal distribution with its mean being the weighted average of these 2 cues . Each cue was weighted by its credibility , and the credibility of the combined belief was the sum of these 2 levels of credibility . Eventually , participants who made the initial judgments , J1 , based on their individual estimates of J ( p ( jIndividual|J ) ) changed to J2 based on their final estimates of J ( p ( J|jIndividual , jSocial ) ) . According to that , the Bayesian model predicted the judgments , J2^ . Specifically , changes in judgments , J2^−J1 , were predicted by the Kullback–Leibler divergence ( DKL ) between 2 probability distributions—the individual estimate of J , p ( jPrivate|J ) , and the final estimate of J , p ( J|jIndividual , jSocial ) . The value of DKL was computed as [40 , 68] J2^−J1=DKL ( p ( jIndividual|J ) ||p ( J|jIndividual , jSocial ) ) =∑i=030p ( i|jIndividual ) ×logp ( i|jIndividual ) p ( i|jIndividual , jSocial ) ( 6 ) where p ( i|jIndividual ) was the probability that the punishment magnitude of i years would be made when the juror had an individual estimate of J , p ( jIndividual|J ) as the preferable punishment for the defendant in the scenario , and p ( i|jIndividual , jSocial ) was the same quantity , given the final estimate of J , p ( J|jIndividual , jSocial ) after integrating social information p ( jSocial|J ) . We also predicted changes in judgments ( J2^ ) under social influences with 2 alternative models . Together with the prediction of the “Bayesian model” ( Eq 6 ) , these predictions ( J2^−J1 ) of alternative decision-making models—the “linear model” ( Eq 7 ) and “surprise model” ( Eq 8 ) —were compared with actual behavioral changes ( J2 − J1 ) . The “linear model” ( Eq 7 ) predicts the changes in judgments ( J2^−J1 ) by a linear regression model that takes into account the given deviation in the group judgments ( D = JS − J1 ) , the level of confidence that one had for their judgments ( C = [−1:+1] ) , the group size ( G = [0 , small group trials; 1 , large group trials] ) , and 2-way interaction effects ( D × C , and D × G ) . We also assumed an interaction effect between the deviation in judgments and the types of scenarios ( D × S where S = [−1 , sympathetic cases; 1 , nonsympathetic cases] ) . That is , participants were expected to change their judgments more when the group made milder level of punishment than themselves ( JS < J1 ) in sympathetic scenarios and when groups made severer punishment than themselves ( JS > J1 ) in nonsympathetic scenarios . We also tested the effect of initial judgments ( J1 ) to test its potential effect on regression-to-the-mean: the behavioral tendency that participants were more ( or less ) likely to conform to JS when J1 was close to extreme ( 0 years or 30 years ) . To test this effect , we included the absolute value of the normalized J1 ( −1 to 1 range; |J1 | ) in the regression analysis . Therefore , |J1 | indicated the scale of judgments that participants moved from the default ( 15 years ) while making the initial judgment . In this model , βL indicates the constant . Taken together , the “linear model” tested whether individuals were motivated to reduce the perceived deviation in the group judgments by adjusting their prior judgments . In particular , this model also tested whether this behavioral tendency was stronger when they had high confidence in their judgments , when they were confronted by judgments of the large group , and when they were confronted by judgments of the group that were more extreme than theirs . The “surprise model” ( Eq 8 ) predicts the changes in judgments ( J2^−J1 ) by a linear regression model that also takes into account the interaction effects with group size and scenario types ( D × G , and D × S ) , and the initial judgments ( |J1 | ) , which were defined the same as those in the “linear model . ” One difference of the “surprise model” is that it considers confidence as a level of precision of one’s own belief about the right amount of punishment for the criminal . Therefore , this model assumes that participants change their judgments based on the surprise of how much the judgment of others differs from their own belief . According to information theory [40] , the surprise ( U ) evoked by such unpredicted observation is the entropy of the unpredictability of an event , α , given the belief of the participant U=−log p ( α|Prior ) ) =−log p ( Js|Norm ( J1 , τ2 ) ) ( 9 ) where τ follows a function of one’s reported confidence as in Eq 2 in the “Bayesian model , ” which includes 2 free parameters ( βC and ωC ) . This model predicts that changes in judgments are proportional to how unlikely others made their judgment , JS , given the prior beliefs of participants , J1 , with the subjective confidence , C . The model predictions of the changes in judgments ( J2^−J1 ) were fitted with subjects’ changes in behavioral performance ( J2 − J1 ) . We estimated the parameters that maximized the predictability of each of the 3 models . The goodness-of-fit of each model was measured with its log likelihood . If the t × 1 vector of observation is denoted by X ( t indicates the number of trials ) and its prediction is estimated as the vector X^ ( same length as X ) , the relationship between the sample and its prediction is defined as: X=ηX^+ε The vector of the error term , ε , has a multivariate normal distribution conditional on η . We assumed that the mean of the error distribution was equal to 0 and that the covariance was equal to ϵ2 . The log likelihood of each model is equal to the logarithm of the product of the likelihoods of each change in judgment . Therefore , the −2log likelihood function is computed as follows: −2lnL ( X , X^ ) =t ln ( 2πϵ2 ) +1ϵ2∑i=1t ( X^i−ηXi ) 2 ( 10 ) To account for over-fitting , we trained each of the models using an in-sample optimization procedure using an iterative LOOCV procedure , in which data from all but 1 trial ( 51 trials ) were used to perform an out-sample prediction against the left-out trial data of each subject . This procedure was repeated 52 times by omitting a different trial each time . Therefore , data from each of the 52 trials was used exactly once as validation data . The series of log likelihoods were averaged across trials and across participants to produce a single estimation . By comparing the prediction power of our models , we examined which model could best explain the underlying mechanisms of the process of judgment adaptation . Model parameters were fitted using the multivariate constrained minimization function in MATLAB 2015 ( MathWorks , MA , USA ) . To select the best model among the 3 , we compared the log likelihood that introduced a penalty according to the number of free parameters . The BIC is computed based on the log likelihood where k is the number of estimated parameters in the model . We assumed that the model parameters were fixed throughout execution of the task , because subjects had been instructed that each murder case was independent and also because participants were told that the other members of the jury differed on each trial . The number of free parameters of each model was as follows: ( 1 ) 7 for the “linear model”—ωD , ωG , ωS , ωC , ωDC , ωJ , and βL; ( 2 ) 7 for the “surprise model”—ωU , ωG , ωS , ωJ , ωC , βC , and βS , ; and ( 3 ) 4 for the “Bayesian model”—βC , ωC , σn = 5 , and σn = 20 . We also compared the neural correlates of the model predicting changes in judgment with the corresponding brain responses encoding changes in judgments . We ran a separate GLM using parametric regressors of changes in judgement and the model predictions of these changes ( GLM3 ) . We extracted the parameter estimates representing the neural correlates of each model prediction . We then compared them with brain responses that correlate with behavioral changes in judgment . Parameter estimates were extracted from the dACC ROI . Consistent with neural model comparisons from previous studies , we compared the likelihood of each model [69 , 70] . When the variance of errors of each model prediction is ϵ2 , and t is the number of time series , the −2 log likelihood is computed as described by Eq 10 . All models had 1 parameter , the coefficient of the single regressor . Functional images were acquired with a 1 . 5 T Siemens Magnetom Sonata Maestro Class MRI System ( Siemens , Munich , Germany ) at the CERMEP in the Groupement Hospitalier Est , Lyon , France . A higher order shimming procedure was completed covering the whole brain of each participant . The imaging parameters for the EPI T2*-weighted sequence were as follows: repetition time ( TR ) , 2500 ms; echo time ( TE ) , 60 ms; flip angle , 90°; FOV , 220 mm , acquisition matrix , 64 × 64 , slice thickness = 4 mm . Contiguous slices were acquired in interleaved order . To acquire whole-brain images , the magnetic field was tilted with minus 20° from the anterior to posterior commissure line ( AC-PC ) of each participant . The imaging parameters for the T1-weighted anatomical scan were as follows: TR , 1970 ms; TE , 3 . 93 ms; FOV , 256 mm; matrix 256 x 256; slice thickness , 1 mm . The stimuli were presented with a screen resolution of 1024 × 768 pixels , displayed at a visual angle of 24 × 18° , centered on a 500 × 500 pixel array , and surrounded by a black background . The participants were asked to use their index and middle fingers of both hands to answer by pressing a 4-button controller . Stimuli were presented , and the responses to the stimuli were collected using the software Presentation ( Neurobehavioral Systems , CA , USA ) . Image preprocessing was performed using SPM8 ( Wellcome Trust Centre for Neuroimaging , UCL , UK ) . Time-series images were registered in a 3-dimensional space to minimize any effect that could result from the motion of the participants’ heads . Functional scans were realigned to the last volume , corrected for slice timing , and unwarped to correct for geometric distortions . Inhomogeneities , distortions related to correction maps , were created using the phase of nonEPI gradient echo images measured at 2 echo times ( 5 . 19 ms for the first echo and 9 . 95 ms for the second ) . These were coregistered with structural maps , spatially normalized into the standard Montreal Neurological Institute ( MNI ) atlas space , and then spatially smoothed with an 8 mm isotropic full-width at half-maximum ( FWHM ) Gaussian kernel using standard procedures in SPM8 . We constructed 3 separate GLMs . For the first GLM ( GLM1 ) , we ran a first-level analysis , modeling brain responses related to revising judgments while confronted with the judgments of others . The conformity trials ( LC > 0 ) and nonconformity trials ( LC ≤ 0 ) were modeled separately . They were modeled as a boxcar function time-locked to the onset of social information ( JS ) with the duration of a response time in each trial to make a judgment ( J2 ) . Brain responses related to making punishment judgments after reading a crime scenario ( J1 ) were modeled separately . These were modeled as a boxcar function time-locked to the onset of decision-making with duration of reaction times in each trial . In addition , the 6 motion parameters produced for head movement and the 2 motor parameters produced for buttons pressing with the right and the left hands were also entered as additional regressors of no interest to account for motion-related artifacts . All these regressors were convolved with the canonical hemodynamic response function . Contrast images were calculated and entered into a second-level group analysis . In the GLM1 , brain regions recruited by conformity decisions were first identified using the contrasts “conformity trials > nonconformity trials . ” We also tested another version of GLM1 with a fixed boxcar duration of 4 s starting at the onset of social information ( JS ) with all other settings kept identical to GLM1 . By fixing the boxcar duration , we tested whether the estimated brain activity in GLM1 was still robust when disregarding inter-trial differences in responses times that could reflect different levels of difficulty in decision-making . To do this , we compared the effects sizes of GLM1 to those of an alternative model . If the effect size estimated from the GLM1 does not decrease in this alternative model , this supports the idea that brain activity observed in GLM1 is involved in judgment adaptation rather than processing the level of difficulty . The second GLM ( GLM2 ) was the same as GLM1 except that the blood-oxygen-level dependent ( BOLD ) response related to revising judgments was separately modeled by group size , instead of conformity decisions . In detail , BOLD responses related to revising judgments when confronted by the judgments of a larger group ( 20 juries ) were separately modeled from those related to revising judgments when confronted with the judgments of a smaller group ( 5 juries ) . They were further modulated by parametric regressors accounting for a perceived difference in judgments ( D = JS − J1 ) . Contrast images were calculated based on the parameter estimates output by the GLM and were then entered into a second-level group analysis . First , we found the group-size–related brain activity ( large group trials > small group trials ) . Second , we found the brain area where the group-size–related activity further correlated with individual differences of Δσ , which was defined as the ratio between the estimated credibility of social information in the large group and the credibility of social information in the small group . To this end , we performed a regression analysis in the second-level analysis by entering Δσ as a covariate across participants . The third GLM ( GLM3 ) was designed to identify the brain regions in which activity parametrically encoded the model-based predictions of changes in judgments that were estimated by the 3 different computational models ( J2^−J1 ) and the actual behavioral of changes in judgments ( J2 − J1 ) . For this GLM , each event was treated as a regressor to extract the time series of beta parameters . This approach has previously been used for multivariate and functional connectivity analyses [71] . In GLM3 , the BOLD responses related to revising judgments were modeled as a boxcar function including reaction times and modulated by parametric regressors accounting for the 3 different model predictions . Again , because differences in reaction times may reflect inter-trial differences in the level of difficulty in decision-making , we also tested an alternative model of GLM3 by fixing the boxcar duration to 4 s starting at the onset of social information ( JS ) . All the other settings were kept identical with the GLM3 . We compared the parameter estimates related with the model-based predictions of changes in judgments to find which of them closely represents the parameter related to actual changes in judgments . BOLD responses to the initial judgments ( J1 ) and 8 types of motion regressors were also included in the GLMs and convolved with the canonical hemodynamic response function ( HRF ) . Modulation of brain activity by the changes in judgments and the model-based predictions of changes in judgments were calculated and entered into a second-level analysis . We further examined how accurately the behavioral models explained BOLD responses related to changes in judgments . For this analysis , we extracted the time courses of dACC activity from the first-level contrasts of each participant modulated by model prediction of changes in judgments . They were then compared with the time course of dACC activity modulated by the actual changes in judgments . All time courses were extracted from the dACC ROI that we defined as a 10-mm diameter sphere , centered on the dACC ROI at x , y , z = 8 , 18 , 46 , based on a previous meta-analysis study [28] . The accuracy level of BOLD activity was estimated by the −2 log likelihood of all participants ( Eq 10 ) and the BIC , considering the number of free parameters of each model were used for the model comparison . To investigate whether dACC activity reflects Bayesian updates rather than choice difficulty , we performed an additional GLM ( GLM4 ) . With this model , we addressed the alternative interpretation that the variance in decision-related dACC activity reflects how difficult the decision was and how long it took . As in GLM3 , the BOLD responses related to revising judgments were modeled with a 4-s boxcar duration , starting at the onset of social information ( JS ) . We included the following parametric regressors: the Bayesian updates measures ( DKL in Eq 6 ) , the level of unexpected surprise ( U in Eq 9 ) , as a measure of difficulty , and the reaction times . In doing so , we examined whether the Bayesian update regressors could still explain dACC activity when allowed to compete with other regressors that possibly explain variance in the dACC signals . In the model specification process , the serial orthogonalization of parametric modulators was turned off . A SVC was performed within the dACC cluster activated for the conformity decision ( PFWE < 0 . 05; on the basis of an initial uncorrected threshold at P < 0 . 001 ) . We report results corrected for FWE with multiple comparisons ( PFWE < 0 . 05 ) . This approach assesses the strength of activations defined by an initial uncorrected threshold , which we take as P < 0 . 001 for all analyses [72] . Three ROIs were defined independently to test and to support our findings . They include dACC and right and left lateral FPC . For each analysis , a single , predefined ROI was used . For the ROI analyses , a SVC was performed ( PFWE < 0 . 05; on the basis of an initial uncorrected threshold at P < 0 . 001 [72] ) . We used an a priori anatomically defined region of the dACC , which was defined by a 10-mm diameter spherical ROI centered at x , y , z = 8 , 18 , 46 ( MNI coordinates ) . This dACC ROI was adopted based on the results of an ALE meta-analysis concerning the neural substrates of conformity behavior [28] . To correct the results of the conformity-decision–related activation for multiple comparisons ( GLM1 ) , we used a SVC . The same ROI was also used to extract beta parameter estimates to compare computational models to the dACC activity representing the changes in judgments ( GLM3 ) . Parameter estimates were extracted from all voxels in the ROI and averaged using MarsBaR 0 . 43 ( http://marsbar . sourceforge . net ) . Finally , this ROI was used to test whether its functional connectivity to the right lateral FPC was modulated by the changes in group sizes . We defined 2 ROIs in the bilateral FPC to extract parameter estimates . The beta parameters were extracted from 2 FPC ROIs , defined as 10-mm diameter spheres centered at each peak voxel of the clusters in the bilateral FPC ( x , y , z ) = ( 27 , 57 , 10 ) and ( −30 , 58 , 15 ) in MNI coordinates . To further check that the recorded pattern of FPC activity reflected individual differences , we used 2 additional analyses [73 , 74] . First , the changes in the parameter estimate ( large group trials—small group trials ) served to perform a bootstrapping sampling analysis ( 10 , 000 times of iteration ) . Notably , parameter estimates and bootstrapping were performed separately from the ROIs in the right and the left FPC . This procedure allowed us to infer the effect size of the relationship between the changes in FPC activation and sensitivity to group size when assigning credibility to social information ( Δσ ) across individuals in a larger dataset . The relationship between the vector of the predicted parameter and the same length vector of the extracted parameter estimates were tested with a linear regression , which enabled us to estimate mean predictability and standard deviation . Based on an assumption that the predictability distribution is normal , we further estimated the 95% confidence interval . Second , we performed a scrambling analysis [75 , 76] . That is , we scrambled Δσ across participants and assigned each participant’s Δσ to another participant . In doing so , we tested whether any brain activity represents this randomly assigned Δσ . To assess changes in functional connectivity during the presence of social information as a function of changes in group size ( large , 20-person jury versus small , 5-person jury conditions ) , we performed a PPI analysis . The PPI allowed us to identify the brain areas where activity can be explained by the interaction between activity in a seed region and the subsequent process involved in the decisions to conform to social information . For the PPI , we defined the seed region in right FPC on the peak voxel , ( x , y , z ) = ( 41 , 44 , 19 ) . We used the Generalized PPI toolbox from SPM ( gPPI ) [77] , which allowed us to create a new GLM in which the deconvolved activity of the seed region is assigned to the context-dependent regressors and reconvolved with HRF . Average time courses were extracted from all voxels in a predefined ROI surrounding the peak voxel in the right FPC clusters . The time courses were extracted from the activities modulated by group size ( GLM2 at the contrast of “large-group trials > small-group trials” ) . The main effects of trials in different group sizes , the seed-region time course , and motion parameters were included as regressors of no interest . The PPI contrast compares large-group trials × right FPC ( +1 ) with small-group trials × right FPC ( −1 ) . | In collective decisions , both the size of groups and the confidence that each member has in their own judgment determine how much a given individual will adapt to the judgment of the group . Here , we show that judgment adaptation during collective decisions—a fundamental brain mechanism needed for fluid functioning of social organizations—can be accounted for by Bayesian inference computations . At the time of judgment adaptation , individuals trade off the credibility inferred from their own confidence levels against the credibility of social information . The dorsal anterior cingulate cortex ( dACC ) represented belief updates , while the lateral frontopolar cortex ( FPC ) monitored the changes in credibility assigned to social information . These results provide a neurocomputational understanding of how individuals benefit both from the wisdom of larger groups and from their own confidence . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"social",
"influence",
"cingulate",
"cortex",
"criminal",
"justice",
"system",
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"functional",
"magnetic",
"resonance",
"imaging",
"criminal",
"punishment",
"decision",
"making",
"reaction",
"time",
"brai... | 2017 | Integration of individual and social information for decision-making in groups of different sizes |
Campylobacter jejuni is the leading cause of bacterial gastro-enteritis in the developed world . It is thought to infect 2–3 million people a year in the US alone , at a cost to the economy in excess of US $4 billion . C . jejuni is a widespread zoonotic pathogen that is carried by animals farmed for meat and poultry . A connection with contaminated food is recognized , but C . jejuni is also commonly found in wild animals and water sources . Phylogenetic studies have suggested that genotypes pathogenic to humans bear greatest resemblance to non-livestock isolates . Moreover , seasonal variation in campylobacteriosis bears the hallmarks of water-borne disease , and certain outbreaks have been attributed to contamination of drinking water . As a result , the relative importance of these reservoirs to human disease is controversial . We use multilocus sequence typing to genotype 1 , 231 cases of C . jejuni isolated from patients in Lancashire , England . By modeling the DNA sequence evolution and zoonotic transmission of C . jejuni between host species and the environment , we assign human cases probabilistically to source populations . Our novel population genetics approach reveals that the vast majority ( 97% ) of sporadic disease can be attributed to animals farmed for meat and poultry . Chicken and cattle are the principal sources of C . jejuni pathogenic to humans , whereas wild animal and environmental sources are responsible for just 3% of disease . Our results imply that the primary transmission route is through the food chain , and suggest that incidence could be dramatically reduced by enhanced on-farm biosecurity or preventing food-borne transmission .
Campylobacter is the most commonly identified cause of bacterial gastro-enteritis in the developed world [1] , [2] , [3] . Infection can lead to serious sequelae such as Guillain-Barré syndrome and reactive arthritis [4] . Of the species pathogenic to humans , 90% of disease is caused by C . jejuni and most of the rest by C . coli [5] . Both species are zoonotic pathogens with wide host ranges including farm animals ( cattle , sheep , poultry , pigs ) and wild animals ( birds and mammals ) [1] , [6] , [7] . The bacterium thrives at 37–42C in the mammalian and avian gut , but survives longest ex vivo in cold , dark , moist environments . Campylobacter is routinely isolated from fresh and marine water sources , and sewage [8] . Epidemiological studies have demonstrated a link with exposure to contaminated food . Handling and eating raw and undercooked poultry have consistently been shown to be important risk factors . Case-control studies show that red meat and seafood are risk factors , as are eating at restaurants and barbecues , and drinking raw milk [9] , [10] . However , food is not the only danger , and some studies have shown that regularly eating poultry and red meat in the home actually has a protective effect [10] . Water , particularly when untreated , can present a threat . Incidence of campylobacteriosis is typically sporadic , but outbreaks do occur that can often be traced to contamination of the water supply [11]–[13] . Some authors have suggested that the strong seasonal variation in sporadic disease , which rises sharply in spring and peaks in summer , bears the hallmark of water-borne diseases such as cryptosporidiosis [8] , [9] , [14] . DNA-based methods of typing C . jejuni have the potential to resolve the controversy surrounding the origin of infection , but have thus far failed to do so . C . jejuni isolated from humans show considerable genetic overlap with meat and poultry isolates [15] , [16] , [17] . However , a model-based approach that includes disparate sources is needed . Although C . jejuni genotypes do show some host association , the population is not strongly structured into differentiated clusters , so predicting host from genotype is challenging [6] . Phylogenetic approaches to tracing the source of infection have suggested that human isolates are more closely related to C . jejuni found in non-livestock than livestock [18] . But recombination is frequent in C . jejuni [19] , which means that a single phylogenetic tree is not an appropriate way to represent the ancestral history of a collection of C . jejuni genomes [20] . Here we report a systematic study of 1 , 231 cases of C . jejuni infection in Lancashire , England , which we have DNA-sequenced using multi-locus sequence typing [21] ( MLST ) . We infer the source of infection of each patient by comparison to 1 , 145 animal and environmental C . jejuni isolates collated from previous studies in livestock , wild animals and the environment [7] , [16] , [21]–[28] , using a novel population genetics approach that models DNA sequence evolution and zoonotic transmission . We treat the animal and environmental reservoirs of C . jejuni as populations between which there may be gene flow ( migration ) . Within these populations the bacteria evolve through de novo mutation and horizontal gene transfer ( recombination ) . We estimate the amount of mutation , migration and recombination , and use these estimates to assign probabilistically each human case to one of the source populations . From these population assignments we estimate the total amount of human disease attributable to each source .
We observed 256 distinct genotypes ( or sequence types , STs ) in the 1 , 231 human isolates . The frequency of genotypes was highly skewed , with 20% of STs accounting for 80% of cases . The two most frequent genotypes ( STs 21 and 257 ) made up a quarter of cases alone , while 182 genotypes were observed once only . There were 375 distinct STs in the 1 , 145 animal and environmental isolates , and overlap with the human genotypes was extensive . Six STs featured in both the human and non-human lists of ten most common genotypes ( STs 19 , 21 , 45 , 50 , 53 , 61 ) . However , nearly a quarter of human cases ( 278 ) exhibited genotypes exclusive to humans ( 189 STs ) , most of those at low frequency . The most abundant human-specific genotypes were ST 572 ( 14 cases ) and ST 584 ( 19 cases ) . Over a third of non-human isolates ( 440 ) possessed genotypes absent from our human sample ( 308 STs ) . Certain genotypes common in non-human isolates were host-restricted to varying degrees . For example , ST 403 was the most prevalent in pigs ( 5/30 isolates ) , but absent from other non-human groups . ST 61 is common in ruminants ( cattle and sheep ) but rare in other groups , while ST 45 was frequent in all the non-human reservoirs except pig and sand . At the level of individual loci , many alleles were frequently observed in a range of animal and environmental sources . Because of the large overlap in genetic variation between C . jejuni reservoirs , our method utilizes differences in gene frequency , rather than allele presence or absence per se . By pooling samples of C . jejuni from similar sources ( e . g . chicks and chicken meat/offal ) and several studies , we intended to improve inference by increasing sample size . See Table S1 for details . However , by combining samples this way , we implicitly assumed that within each group ( chicken , cattle , sheep , pig , bird , rabbit , sand and water ) gene frequencies are consistent across sources and across studies . To test this assumption , and to quantify genetic differentiation between groups , we used analyses of molecular variance ( AMOVA [29] ) . AMOVA quantifies genetic differentiation within and between groups using Φ-statistics , which measure the correlation in gene frequencies within sub-populations relative to the total population . A smaller value of Φ indicates lower genetic differentiation between the populations . Table 1 shows ΦSG , the genetic differentiation within each group ( e . g . chicken ) between isolates of a different sub-group: i . e . source type ( e . g . chick vs chicken meat/offal ) or published study . Except for the sand group , there was significant heterogeneity within the groups that comprised more than one source type or study . Genetic differentiation between sub-groups ranged from 2 . 4% ( cattle ) to 23 . 2% ( pig ) . This suggests that gene frequencies vary significantly between similar sources and between different studies of the same source . In order to assign human cases to source populations with any degree of accuracy , there must be genetic differentiation between the groups , over and above within-group heterogeneity . We estimated that quantity , ΦGT , using nested AMOVA between pairs of groups . Table 1 shows the results . For each pair of groups , ΦGT is displayed below the diagonal and the associated p-value above the diagonal . All groups were significantly differentiated from at least one other group in this way , with ΦGT ranging from 4 . 4% ( chicken vs cattle ) to 26 . 2% ( sheep vs pig ) . However , there were some major groups that were not significantly differentiated . Notably , ΦGT was 0 . 1% for cattle vs sheep ( p = 0 . 182 ) , which suggested it would be difficult to distinguish human cases attributable to these two groups . The preliminary analyses of the animal and environmental C . jejuni isolates presented several potential concerns . There was significant variation in gene frequencies within groups , probably caused by the heterogeneous nature of the studies from which the non-human isolates were drawn , and the inherently stochastic nature of the epidemic process . This could distort the gene frequency information upon which source assignment relies , and cause higher than expected linkage disequilibrium between loci . AMOVA also showed that genetic differentiation between groups was weak in some cases . Within-group heterogeneity could therefore obscure or potentially distort the signal of differentiation between groups . Another concern was that the large differences in sample size between non-human groups , which reflect a tendency among researchers to preferentially sample certain hosts , could bias the source assignment . To investigate the sensitivity of our method to these effects , and to test its robustness to violating the assumption of homogeneous mixing within groups , we performed empirical cross-validation . Traditional methods that can assign large numbers of individuals to populations based on their genotype tend to assume that loci provide independent sources of information [30] , [31] . In other words , they assume that gene frequencies between loci are uncorrelated in the source populations . While this simplifying assumption is computationally convenient , it may not be appropriate for C . jejuni because of appreciable linkage disequilibrium between MLST loci [19] . Therefore we developed two models , one in which loci were assumed to be unlinked ( i . e . independent , or in linkage equilibrium ) and another in which loci were linked ( i . e . in linkage disequilibrium ) ; see Materials and Methods . We used empirical cross-validation to scrutinize both . In each of 100 simulations , we removed the source information from half the non-human isolates , chosen at random . These we termed the pseudo-human cases . We used our unlinked and linked models to assign the source of the pseudo-human cases using the other non-human isolates . Table 2 shows that the two models differed considerably in performance . On average , the unlinked model correctly assigned 52% of the pseudo-human cases ( using the rule that each case is assigned to its most probable source ) , whereas the linked model correctly assigned 64% . The linked model was well-calibrated in the sense that its estimated success rate was 64% on average , whereas the unlinked model grossly over-estimated its success rate ( 82% on average ) . We used a number of performance indicators to measure the ability of each model to correctly estimate the total proportion of pseudo-human cases attributable to a given source ( see Table 2 ) . The parameter estimates obtained by using the linked model generally exhibited lower bias and smaller variance ( measured by root mean squared error , RMSE ) than those obtained using the unlinked model . The linked model also out-performed the unlinked model in coverage , which we defined as the number of simulations , out of 100 , in which the 95% credible interval for the proportion of cases attributable to a given source enveloped the true value . For seven out of eight groups , the linked model obtained the target coverage of 95 or above . Coverage was 93 for chicken; the small negative bias suggests this may have been caused by slightly under-estimating the proportion of pseudo-human cases attributable to chicken . In the empirical cross-validation the linked model performs well despite the potential concerns due to heterogeneity within the animal and environmental groups , and differences in sample size . Most importantly , it is well-calibrated in assigning isolates to source populations , and estimating the overall proportion of cases attributable to each source . In contrast , the unlinked model assigns fewer isolates to source populations correctly , and is very poorly calibrated . This underlines the importance of adequately modeling recombination in the study of pathogen evolution . Clearly the computational efficiency gains made by assuming independent inheritance among loci in the unlinked model are out-weighed by its poor performance . Therefore we use the linked model for our analysis proper . We applied our novel method to the 1 , 231 newly-sequenced human isolates from Lancashire , England . For every case , the assignment probability was calculated for each source population ( chicken , cattle , sheep , pig , bird , rabbit , sand , water ) , and the total proportion of cases attributable to each source was estimated . We found that the vast majority ( 96 . 6% ) of human cases are attributable to populations of C . jejuni carried by livestock ( 95% credible interval 92 . 7–98 . 8% ) as opposed to wild animals ( 2 . 3% ) or environmental isolates ( 1 . 1% ) . Figure 1 shows a breakdown of attribution by source; errors bars indicate the 95% credible intervals . We estimated that chicken is the source of infection in the majority ( 56 . 5% ) of cases ( 95% C . I . 51 . 1–61 . 8% , see Table S2 ) , followed by cattle ( 35 . 0% ) and sheep ( 4 . 3% ) . The 95% credible intervals were wider for cattle ( 20 . 8–43 . 2% ) and sheep ( 0 . 1–17 . 5% ) than other groups , which reflects the greater difficulty in distinguishing these populations of C . jejuni from one another . We found that pig is unlikely to be the source of C . jejuni infection in humans ( 0 . 8% of cases ) . Of the two groups of wild animals we studied , bird and rabbits , there was somewhat more support for a wild bird origin of human C . jejuni ( 1 . 7% ) than rabbit ( 0 . 6% ) , although the credible intervals ( 0 . 1–5 . 5% and 0 . 0–3 . 7% respectively ) were largely overlapping . There was very little support for an environmental origin of human infections . Even so , the results suggested that infection with C . jejuni found in environmental water sources was more likely ( 0 . 9% ) than infection with C . jejuni isolated from bathing beaches ( 0 . 2% ) , which was the least likely of all sources . Overall , the analysis reported that with 98 . 3% probability , chicken is the primary , and cattle the secondary source of human infections in our study . The posterior probability of source of infection was estimated for each patient in our study; Figure 2 illustrates the results . The source populations are color-coded as in Figure 1 . Cases are arranged horizontally , and the vertical column space occupied by each color represents the posterior probability of infection from that source . The dominant color in any column indicates the most likely source for a particular case . The principal distinction in human cases is between those attributed to chicken versus ruminants ( cattle and sheep ) . Most cases lie on a continuum between assignment to ruminants and to chicken . The existence of this continuum , as opposed to a clear separation , emphasizes the overlap in genotypes between these source populations , and the advantage of a probabilistic approach to assignment . Some common genotypes were strongly assigned to ruminants ( e . g . ST 48 , 86 cases , posterior probability [Pr] = 0 . 91 ) and others to chicken ( ST 104 , 64 cases , Pr = 0 . 93 ) . But within ruminants , it is harder to distinguish cattle from sheep sources . This is borne out by the strong correlation among cases between cattle and sheep assignment probabilities ( ρ = 0 . 80 ) . In some cases , there is moderate or strong support for a source that is generally found to be rare . For example , there were six cases of ST 403 , with a moderately high assignment probability to pig of 0 . 37 . Except for the human isolates , we observed ST 403 only in the pig population . However , because the evidence overall suggests that pig is an unlikely source of infection for humans , and because of the genetic similarity to cattle genotypes ( e . g . ST 933 ) , it is marginally more likely under the model that cattle is the source of these cases ( Pr = 0 . 46 ) . Although it is most probable , on a case-by-case basis , that the source of infection was cattle , when considered together we would expect the source of infection to have been cattle in 2 . 7 of those cases , pig in 2 . 2 cases and chicken in 0 . 6 cases . Another example of this phenomenon is found in birds . There are 28 cases , of which ST 508 was the most common genotype , with an assignment probability to birds greater than 10% , but a larger assignment probability to another source , usually chicken . On an individual basis none of these cases would be assigned to birds , but taken together we estimate that the source of infection was birds in 5 . 6 of them , chicken in 10 . 6 , cattle in 5 . 5 and water in 3 . 8 . Overall , the source probabilities in Figure 1 and Table S2 suggest that of the 1 , 231 human cases , the source of infection was chicken in 696 . 6 cases , cattle in 432 . 1 , sheep in 53 . 5 , bird in 20 . 5 , water in 10 . 9 , pig in 10 . 3 , rabbit in 7 . 9 and sand in 2 . 2 . Sometimes it is useful to assign a case to a single source , in which case the optimal strategy is to attribute it to the source with highest assignment probability a posteriori . We estimate that 76 . 5% of human cases would be correctly assigned by this procedure . Earlier we showed that this quantity , which is the average maximum source attribution probability per case , was well-calibrated during empirical cross-validation . When cases are assigned to sources in this fashion , most are assigned to chicken ( 722 ) or cattle ( 503 ) . None are assigned to sheep , because ruminant-associated isolates are assigned preferentially to cattle . A small number are assigned to pig , bird and water ( three in each case ) . For example , STs 1286 , 1927 and 2973 were the genotypes most strongly assigned to environmental water , pig and wild bird respectively ( Pr = 0 . 58 , 0 . 65 , 0 . 87 ) . Interestingly , all three genotypes were human-specific , and each was found in a single patient only . In the case of ST 1286 , there was also considerable support for a wild bird origin ( Pr = 0 . 35 ) , an observation that may reflect the low genetic differentiation detected between these sources ( Table 1 ) . Table S3 gives a detailed breakdown of source attribution probabilities by sequence type . Our collection of animal and environmental isolates which we collated from previously-published studies [7] , [16] , [21]–[28] were non-ideal in several respects . AMOVA revealed significant variation between isolates from the same group that originated in different sub-groups – i . e . different source types or studies . Such genetic structuring will cause higher than expected linkage disequilibrium within groups , and may distort the gene frequencies upon which source attribution relies . Although empirical cross-validation showed that the linked model was robust to these effects , the full extent of the difficulty caused by within-group heterogeneity may have been masked because individual isolates were assigned to the pseudo-human class independently , and without reference to their sub-group . Therefore we performed additional simulations in which whole sub-groups of isolates were removed , and the human isolates re-analyzed based on the reduced set of animal and environmental isolates . In each simulation , we removed at least 20% of the animal and environmental isolates , 24 . 5% on average . Figure S2 illustrates the simulation scheme and contrasts it to the simulations used in empirical cross-validation . Our main conclusions are robust to genetic heterogeneity within the source populations . Figure S3 summarizes the analysis of robustness by plotting the point estimate and 95% credible interval of various parameters based on the 100 simulations and the full data . In all of the 100 simulated datasets analyzed , chicken was found to be the primary source of human infections . Figure S3A shows that in the majority of simulations , chicken accounted for more than 50% of human disease . The conclusion that ruminants are the second most important source of human infection was also supported by the analysis ( Figure S3B ) . Despite the low genetic differentiation between cattle and sheep , as witnessed by the AMOVA results , the finding that cattle account for considerably more disease than sheep is surprisingly robust to re-sampling the non-human isolates . In Figure S3C , the posterior median ( rather than the mean ) is used to illustrate that in 87 simulations , a greater proportion of human cases were attributed to cattle than to sheep . The greatest effect of the re-sampling of non-human isolates was seen in the proportion of human cases attributed to the bird group . Figure S3D shows that in a minority of simulations ( 16 out of 100 ) , the proportion of cases attributed to birds leapt ten-fold to around 20% , promoting it to the second or third most important source , compared to fourth in the analysis of the full data . Of these 16 simulations , there was a significantly lower number of bird isolates ( p = 0 . 005 ) and a significantly higher number of chicken isolates ( p = 0 . 040 ) compared to the other simulations , the relevance of which is that the chicken and bird groups were shown by AMOVA to exhibit extremely low genetic differentiation ( Table 1 ) . While these results demonstrate that more intense sampling of the smaller groups , particularly birds , is highly desirable , our main conclusions are supported by the vast majority of re-sampled datasets , indicating a satisfactory level of robustness to within-group heterogeneity . A tacit assumption in our study , and in the ongoing sampling of C . jejuni populations , is that the major reservoirs have been identified . However , if a major source of human disease were undiscovered , we would expect to see an excess of genotypes unique to humans . In our study we observed 189 genotypes unique to humans . Of the 1 , 231 human cases , 278 possessed genotypes absent in the non-human isolates , but most of these ( 238 cases ) were re-assortments of alleles or allele fragments that were present in the non-human isolates . In 254 cases , they differed at three loci or fewer to a non-human isolate . Out of 531 single nucleotide polymorphisms in humans , 40 were absent from the non-human samples . Of those , all were rare except an adenosine at nucleotide 448 in the glnA locus ( 12 copies ) , and a cytosine at nucleotide 93 in the tkt locus ( 13 copies ) . Two human-specific STs ( 572 and 584 ) had appreciable frequency ( 14 and 19 cases respectively ) . It is difficult to quantify exactly what would constitute an excess of genotypes unique to humans . We employed a re-sampling procedure to compare the number of unique genotypes in human isolates compared to other groups , controlling for sample size . When sets of human isolates were drawn , equal in size to the number of chicken isolates ( 515 ) , we observed fewer unique genotypes on average among human isolates ( 104 . 4 ) than among chicken isolates ( 153 ) , where uniqueness was determined by reference to the “pool” of other non-human and non-chicken isolates . The same pattern was observed when comparing humans to cattle and birds , but not sheep ( Figure S4A ) . Sheep isolates are genetically similar to cattle , which may explain why humans exhibit no more unique genotypes than do sheep . The observation that cattle isolates appear to exhibit relatively more unique genotypes than sheep suggests there might be an effect of sample size , or that sheep isolates are a subset of cattle isolates . A second re-sampling procedure was designed to emulate the status of humans as a sample of isolates drawn from the putative source populations . Taking each non-human group in turn , half the isolates were removed , leaving the other half in the pool , and the number of genotypes unique to the removed isolates was calculated . A set of human isolates was drawn of equal number , and the number of unique genotypes calculated relative to the same pool . The whole procedure was repeated 100 times . If major source populations remained to be discovered , or if humans acted as a reservoir of C . jejuni rather than a terminus in the transmission chain , then an excess of genotypes unique to humans would be expected . However , in these simulations the distribution of the number of genotypes unique to humans and the non-human groups overlapped to a great extent ( Figure S4B ) . Therefore while the more abundant STs and SNPs unique to humans deserve further attention , on the whole there is little indication that another major , genetically distinct , reservoir of human infection remains undiscovered .
Our results show that livestock are the principal source of C . jejuni infection in Lancashire , England . The vast majority of those human infections can be attributed to populations of C . jejuni found in chicken and cattle . These findings immediately lend weight to the suggestion that the incidence of campylobacteriosis in humans could be significantly reduced by intervention strategies targeted at livestock [32] , [33] , chiefly the strict enforcement of on-farm biosecurity measures including disinfecting farm premises and water supplies , restricting access to livestock to essential personnel , minimizing the use of invasive practices such as thinning in chickens , securing premises from wild birds and mammals , and protecting food supplies from bacterial contamination . Moreover , our results are informative about the likely mode of transmission of C . jejuni to patients in our study . The genetic analysis identifies the source of infection , rather than the transmission route . The importance of livestock as a reservoir for human disease is consistent with food-borne transmission , but alternative pathways , such as ingestion of animal feces or contamination of water by human or animal waste , must also be considered . Our findings show that , while we can detect cases of human infection with isolates of an environmental or wild animal origin , such cases are rare , and this is surprising if pathways other than food-borne transmission are important . Therefore the dual observations that ( i ) livestock are a frequent source of human disease isolates and ( ii ) wild animals and the environment are not , strongly support the notion that preparation or consumption of infected meat and poultry is the dominant transmission route . Transmission through the food chain can be controlled in a number of ways . Preventing cross-contamination of carcasses during processing is an effective measure [34] that can be achieved , for example , by minimizing meat contamination with animal feces , treating carcasses with antimicrobial agents , sterilizing equipment , and careful management of animals or flocks known to be infected . Meat products can be treated directly , for example by freezing or irradiation [35] . Promoting better standards of food hygiene during preparation and cooking is also an effective measure [32] , [34] . Our results pertain to sporadic disease; we know that contamination of drinking water occasionally causes outbreaks [11]–[13] . The lack of evidence for pigs as a source of C . jejuni infection is consistent with their greater susceptibility to C . coli [36] . Since C . coli causes less than 10% of sporadic campylobacteriosis , pigs must be a less important source of infection than chicken and cattle . We found considerable variation in the genetic make-up of C . jejuni populations sampled from similar sources ( e . g . chicks vs chicken meat/offal ) and between different populations from the same source type . This variation may reflect functional differences between C . jejuni even from closely related sources , or it may reflect stochastic differences in gene frequency over time or space . The epidemic process may increase variation in gene frequencies because hosts sampled locally are infected from the same source , causing non-independence within samples . We found our method was robust to this heterogeneity , but it is reasonable to think that inference would be improved by sensibly modeling the phenomenon . How to do so is unclear: one option is to split heterogeneous groups into further sub-categories , but that increases the number of parameters in the model and may reduce statistical efficiency or lead to over-fitting . Comprehensive sampling of putative source populations in parallel to human sampling is most desirable , and such studies are on-going by groups in Scotland , New Zealand and the USA . Assigning the source of human isolates based on genotype has been attempted before in C . jejuni . Our results are in contrast to those of Champion et al . [18] who , using a Bayesian phylogenetic approach applied to comparative genomic hybridization data , found that C . jejuni isolates can be divided into livestock and non-livestock clades , with 55 . 7% of human isolates falling into the non-livestock clade . The existence of these clades was supported by high posterior probabilities , close to Pr = 1 . The implications of such findings would be dramatic , however there are difficulties with the approach . The principal problem is that C . jejuni is known to be highly recombining which means that different genes , or even different parts of the same gene , will have different phylogenetic histories . Inferring a single phylogenetic tree for the whole genome is therefore a case of gross model mis-specification , and the resulting phylogeny is difficult to interpret in any meaningful way [37] . Many pathogens exist as weakly differentiated , genetically overlapping populations or strains between which there is frequent gene flow and within which there is frequent recombination . Such strains may be epidemiologically relevant , but it will be difficult to find stable , well-differentiated genetic markers , the standard tools of molecular epidemiology , to type them unambiguously . In this paper the method we developed addressed the problem in C . jejuni by assigning isolates to source populations probabilistically . We used a simple epidemiological model , in which we inferred the probability of infection with each source , to efficiently combine information over cases . That model could be readily extended in the general linear model framework to employ covariates , such as age , sex or host genotype , were they available . In conclusion , we have used a novel population genetics approach to identify the source of infection of the zoonotic pathogen Campylobacter jejuni . We found that cases of human infection in our study were overwhelmingly attributable to bacteria characteristic of those colonizing animals farmed for meat and poultry , based on genetic similarity . We hope that demonstrating the importance of livestock as reservoirs of Campylobacter infectious to humans will add impetus to initiatives aimed at controlling food-borne pathogens .
Stool samples were collected from 1 , 549 patients diagnosed with campylobacteriosis and notified through general practitioners and hospitals to the Preston Microbiology Services Laboratory in the Preston postcode district between January 1st 2000 and December 31st 2002 . This covers an area of 968 km2 , comprising 403 , 000 people at the 2001 census , consisting of both urban ( Preston , Leyland , Chorley , Garstang ) and rural ( Ribble estuary and Ribble valley ) districts . As is the norm with campylobacteriosis , the cases we studied were sporadic in nature; there was no evidence for outbreaks . We followed previously published methods for multilocus sequence typing C . jejuni [21] , [38] . We obtained culturable , uncontaminated isolates of Campylobacter species from 1 , 353 samples , of which we identified 1 , 255 C . jejuni , 86 C . coli and 11 other species . One isolate tested positive for both C . jejuni and C . coli using the hippurate hydrolysis test and PCR . We fully sequenced all seven MLST loci ( 3 , 309 nucleotides in total ) in 1 , 231 C . jejuni isolates , a sequencing success rate of 98% . We collated 1 , 145 C . jejuni isolates of animal and environmental origin from ten previously published studies [7] , [16] , [21]–[28] . Where the sampling date was available , we excluded isolates sampled prior to 1990 . We grouped the isolates by host or environmental origin as follows: chicken ( 515 isolates ) , cattle ( 282 ) , sheep ( 160 ) , pig ( 30 ) , wild bird ( 44 ) , wild rabbit ( 20 ) , bathing beach ( 71 ) and environmental water sources ( 23 ) . Table S1 gives a detailed breakdown of groups by source type and publication . To analyze the genetic heterogeneity within each group , we estimated Φ-statistics using analyses of molecular variance ( AMOVA [29] ) . Genetic distance between a pair of isolates was defined as the number of loci , out of seven , at which they differed . We defined sub-groups using detailed sampling information from each publication ( Table S1 ) . E . g . we defined isolates sampled from calves versus cows milk as separate sub-groups within the cattle group . Isolates sampled from the same source type in different studies were also defined as separate sub-groups . Significance was assessed by permutation test , using 999 permutations . To analyze genetic differentiation between groups , over and above within-group differentiation , we performed pairwise nested AMOVA . Significance was assessed in the same fashion . The parameter of primary interest was the proportion , Fj , of human cases attributable to source population j ( j = 1…ng ) where ng = 8 was the number of putative source populations , and . If we knew the source of each case , we could estimate F directly using the multinomial likelihoodwhere N = 1 , 231 was the number of cases and Gi was the source of origin for case i . Our approach was Bayesian , so the posterior probability distribution for F , upon which inference is based , would bewhere p ( F ) is a prior probability distribution on the source attribution probabilities . We used a symmetric Dirichlet ( 1 ) prior on F in which all sources are considered equally likely a priori . Of course we did not know G , so we used a genetic model of DNA sequence evolution to co-estimate the probable source of human isolates based on their genotypes , H , as follows , where p ( H|G ) is the likelihood of the source assignments G under our evolutionary model . In our evolutionary model , we envisage the population of C . jejuni as a number of discrete islands: each source corresponds to a different island . Within each island the population is homogeneously mixing , and between islands there is migration . Migration rates may be higher between some islands than others , resulting in different levels of genetic differentiation . This is known as the migration matrix model [39] , a generalization of Wright's island model . We modeled the generation of new alleles within each MLST locus using the infinite alleles model [40] , and investigated two models of recombination between loci . In the first , the loci were assumed to be unlinked ( inherited independently , or in linkage equilibrium ) , which is a computationally convenient but biologically unrealistic assumption . In the second , the loci were treated as linked ( in linkage disequilibrium ) using a model of recombination suitable for bacteria [41] . Human isolates were treated as a direct draw from one of the source populations . Therefore we assumed that the genotype of a human isolate would be representative of genotypes in the source population from which it was acquired . As a consequence , source attribution relies on the calculation of sampling probabilities; the likelihood that human isolate i was sampled from source population j . Unfortunately the complexity of the evolutionary model , in particular the linked model , renders direct calculation of the joint sampling probabilities p ( H|G ) impracticable , so we developed an approximation; full details of the approximation and the Markov Chain Monte Carlo sampler are provided in the Supplementary Methods ( Text S1 ) . To summarize , we used the animal and environmental isolates to estimate mutation , recombination and migration parameters . We then used these estimates together with all the genetic data ( human and non-human genotypes ) to jointly estimate the source attribution probabilities F and the source of human cases G . Except where stated otherwise , we used the mean of the posterior distribution for point estimates , and the ( 2 . 5% , 97 . 5% ) quantiles of the posterior distribution for 95% credible intervals . We employed empirical cross-validation to assess various of aspects of our approach: ( i ) the adequacy of the approximations made in order to perform inference ( ii ) the robustness to violations of the modeling assumptions , such as genetic heterogeneity within groups , and ( iii ) the sensitivity to sample size differences between groups . During each iteration of the empirical cross-validation , we artificially removed the population of origin of half the 1 , 145 animal and environmental isolates at random . We then used the other half to infer their origin , and evaluated the performance of the two models . This procedure was repeated over 100 iterations . We calculated several indicators of performance . The predicted proportion of isolates correctly assigned was calculated aswhere there were M = 100 simulation , N = 572 pseudo-human cases , ng = 8 putative source populations , and was the posterior probability that population k is the source of pseudo-human case j in simulation i . The bias in the estimate of the proportion of pseudo-human cases attributable to population j was calculated aswhere was the actual proportion attributable to population j in simulation i , and was the point estimate , i . e . the mean of the posterior distribution . The root mean squared error , which measures the variance of the point estimate , was calculated as As an additional test of robustness to the potentially confounding effects of genetic heterogeneity within the putative source populations , we repeated the source attribution analysis using subsets of the animal and environmental isolates , using the linked model only . The idea was to study the effect of removing whole sub-groups of isolates that were derived from the same source type or publication , as defined in Table S1 . We conducted 100 simulations in order to generate samples of the non-human isolates in which 20% or more of the isolates were excluded , a whole sub-group at a time . The simulations were conducted as follows: On average , this procedure generated subsets in which 24 . 5% of isolates were excluded . Each of the 100 simulated subsets of the non-human isolates was used to infer the proportion of human cases attributable to each source . Figure S2 illustrates the difference in the simulation schemes between the empirical cross-validation and the analysis of robustness . We performed two re-sampling procedures to compare the number of unique genotypes in human isolates to the number in other groups . The aim was to scrutinize two modeling assumptions: ( i ) that human isolates are merely a sample of C . jejuni isolates found in the putative source populations , and ( ii ) that the major source populations have been identified . In the first procedure , we removed one non-human group , e . g . chicken , from the “pool” of non-human isolates and calculated the number of unique genotypes by reference to the pool . We sampled a subset of human isolates , equal in size to the number of chicken isolates , and calculated the number of unique genotypes by reference to the same pool . We repeated the sampling of human isolates 100 times to generate a distribution for the number of genotypes unique to humans , which we compared to the number of genotypes unique to chicken . Because of assumption ( i ) we expect humans to exhibit fewer unique genotypes . In the second re-sampling procedure , we removed half of the isolates belonging to a non-human group , e . g . chicken , leaving the rest in the pool in order to emulate the status of human isolates , which we assumed are merely a sample of isolates found in the non-human source populations . We sampled a subset of human isolates equal in number , and calculated the number of genotypes unique to chicken and humans , by reference to the same pool . We repeated the procedure 100 times to generate a distribution of the number of genotypes unique to chickens and humans . Violation of assumptions ( i ) or ( ii ) could lead to an excess of genotypes unique to humans . All newly-sequenced multi-locus sequence types are available for download from pubMLST . org/campylobacter . | C . jejuni is a bacterium commonly found in the guts of birds and mammals . In humans , it is responsible for causing more gastro-enteritis than any other identified bacterial species . Humans may contract campylobacter from a variety of sources . Eating raw or undercooked meat or poultry , and poor food hygiene that leads to cross-contamination of uncooked food , can cause human disease . However , humans may be exposed to the feces of infected wild animals , and campylobacter can survive in water . Contamination of drinking water can lead to outbreaks , and previous genetic studies have suggested that livestock are not the principal source of human infection . We extracted campylobacter DNA from patients and compared it to campylobacter DNA found in livestock , wild animals , and the environment . We developed a new evolutionary model to identify the most probable source populations . In 97% of cases , we identified chicken , cattle , or sheep as the source of infection . Very few cases were attributable to campylobacter found in wild animals or the environment . Our results imply that the primary transmission route is the food chain and also add new impetus to measures that reduce infection in livestock and prevent food-borne transmission . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"computational",
"biology/population",
"genetics",
"genetics",
"and",
"genomics/microbial",
"evolution",
"and",
"genomics",
"evolutionary",
"biology/microbial",
"evolution",
"and",
"genomics",
"molecular",
"biology/molecular",
"evolution",
"microbiology/microbial",
"evolution",
... | 2008 | Tracing the Source of Campylobacteriosis |
Although there is growing interest in measuring integrated information in computational and cognitive systems , current methods for doing so in practice are computationally unfeasible . Existing and novel integration measures are investigated and classified by various desirable properties . A simple taxonomy of Φ-measures is presented where they are each characterized by their choice of factorization method ( 5 options ) , choice of probability distributions to compare ( 3 × 4 options ) and choice of measure for comparing probability distributions ( 7 options ) . When requiring the Φ-measures to satisfy a minimum of attractive properties , these hundreds of options reduce to a mere handful , some of which turn out to be identical . Useful exact and approximate formulas are derived that can be applied to real-world data from laboratory experiments without posing unreasonable computational demands .
What makes an information-processing system conscious in the sense of having a subjective experience ? Although many scientists used to view this topic as beyond the reach of science , the study of Neural Correlates of Consciousness ( NCCs ) has become quite mainstream in the neuroscience community in recent years—see , e . g . , [1 , 2] . To move beyond correlation to causation [3] , neuroscientists have begun searching for a theory of consciousness that can predict what physical phenomena cause consciousness ( defined as subjective experience [3] ) to occur . Dehaene [4] reviews a number of candidate theories currently under active discussion , including the Nonlinear Ignition model ( NI ) [5 , 6] , the Global Neuronal Workspace ( GNW ) model [7–9] and Integrated Information Theory ( IIT ) [10 , 11] . Rapid progress in artificial intelligence is further fueling interest in such theories and how they can be generalized to apply not only to biological systems , but also to engineered systems such as computers and robots and ultimately arbitrary arrangements of elementary particles [12] . Although there is still no consensus on necessary and sufficient conditions for a physical system to be conscious , there is broad agreement that it needs to be able to store and process information in a way that is somehow integrated , not consisting of nearly independent parts . As emphasized by Tononi [10] , it must be impossible to decompose a conscious system into nearly independent parts—otherwise these parts would feel like two separate conscious entities . While integration as a necessary condition for consciousness is rather uncontroversial , IIT goes further and makes the bold and controversial claim that it is also a sufficient condition for consciousness , using an elaborate mathematical integration definition [11] . As neuroscience data improves in quantity and quality , it is timely to resolve this controversy by testing the many experimental predictions that IIT makes [11] with state-of-the-art laboratory measurements . Unfortunately , such tests have been hampered by the fact that the integration measure proposed by IIT is computationally infeasible to evaluate for large systems , growing super-exponentially with the system’s information content . This has lead to the development of various alternative integration measures that are simpler to compute or have other desirable properties . For example , Barrett & Seth [13] proposed an attractive integration measure that is easier to compute from neuroscience data , but whose interpretation is complicated by the fact that it can be negative in some cases [14 , 15] . [16] used an integration measure inspired by complexity theory to successfully predict who was conscious in a sample including patients who were awake , in deep sleep , dreaming , sedated and with locked-in syndrome . [17] suggest that state transition entropy correlates with consciousness . Griffith & Koch have proposed defining integration of a system as the synergistic information that its parts have about the future , which appears promising although there does not yet exist a unique formula for it [18] . Even the team behind IIT has updated their integration measure twice through successive refinements of their theory [10 , 11] . Despite these definitional and computational challenges , interest in measuring integration is growing , not only in neuroscience but also in other fields , ranging from physics [12] and evolution [19] to the study of collective intelligence in social networks [20] . It is therefore interesting and timely to do a comprehensive investigation of existing and novel integration measures , classifying them by various desirable properties . This is the goal of the present paper , as summarized in Tables 1 and 2 . In the Methods section , we investigate general integration measures and their properties . In the Results section , we first present our taxonomy of integration measures , then derive useful formulas for many of these measures that can be applied to the sort of time-series data that is typically measured in laboratory experiments with continuous variables , and finally explore further algorithmic speedups and approximations . We summarize our conclusions in the Discussion session .
Consider two random vectors x0 and x1 whose joint probability distribution is p ( x0 , x1 ) . We will interpret them as the state of a time-dependent system x ( t ) at two separate times t0 and t1 . For example , if these are two vectors of 5 bits each , then p is a table of 210 numbers giving the probability of each possible bit string , while if these are two vectors in 3D space , then p is a function of 6 real continuous variables . We obtain the marginal distribution p ( n ) ( xn ) for the nth vector , where n = 0 or n = 1 , by summing/integrating p over the other vector . Below we will often find it convenient to denote these vectors as single indices i = x0 and j = x1 . For example , this allows us to write the marginal distribution p0 ( x0 ) as ∑j pij , where the sum over j is to be interpreted as summation/integration over all allowed values of x1 . We also adopt the notation where replacing an index by a dot means that this index is to be summed/integrated over . This lets us write the marginal distributions p ( 0 ) ( x0 ) and p ( 1 ) ( x1 ) as p i ( 0 ) = p i · and p j ( 1 ) = p · j ( 1 ) As illustrated in Fig 1 , it is always possible to model this relation between x0 and x1 as resulting from a Markov process , where x1 is causally determined by a combination of x0 and random effects . If we write the marginal distributions from eq ( 1 ) as vectors p ( 0 ) and p ( 1 ) , this Markov process is defined by p ( 1 ) = M p ( 0 ) , ( 2 ) where the Markov matrix Mji specifies the probability that a state i transitions to a state j , and satisfies the conditions Mji ≥ 0 ( non-negative transition probabilities ) and M⋅i = 1 ( unit column sums , guaranteeing probability conservation ) . The standard rule for conditional probabilities gives p i j = P ( x 0 = i & x 1 = j ) = = P ( x 0 = i ) P ( x 1 = j | x 0 = i ) = p i ( 0 ) M j i , ( 3 ) which uniquely determines the Markov matrix as M j i = p i j p i ( 0 ) = p i j p i · , ( 4 ) which is seen to satisfy the Markov requirements Mji ≥ 0 and M⋅i = 1 . Note that any system obeying the laws of classical physics can be accurately modeled as a Markov process as long as the time step Δt ≡ t1 − t0 is sufficiently short ( defining x ( t ) as the position in phase space ) . If the process has “memory” such that the next state depends not only on the current state but also on some finite number of past states , it can reformulated as a standard memoryless Markov process by simply expanding the definition of the state x to include elements of the past . Also note that although full knowledge of the Markov matrix M completely specifies the dynamics of the system , a person wishing to compute its integration may not know M exactly . If M is not known from having built the system or having examined its inner workings , then passively observing it in action ( without active interventions ) may not provide enough information to fully reconstruct M [21] . The n → ∞ limit of continuous variables describes a convenient class of systems where M is relatively easy to determine in practice . We will now see that this Markov process interpretation allows us create a simple taxonomy of integration measures ϕ that quantify the interaction between two subsystems . The idea is to approximate the Markov process by a separable Markov process that does not mix information between subsystems , and to define the integration as a measure of how bad the best such approximation is . Consider the system x as being composed of two subsystems xA and xB , so that the elements of the vector x are simply the union of the elements of xA and xB , and let us define the probability distribution p i i ′ j j ′ ≡ P ( x 0 A = i & x 0 B = i ′ & x 1 A = j & x 1 B = j ′ ) . ( 5 ) ( For brevity , we will sometimes refer to this distribution pii′ jj′ as simply p below , suppressing the indices , and we will sometimes write x without indices to refer to the full state at both times . ) The Markov matrix of eq ( 4 ) then takes the form M j j ′ i i ′ = p i i ′ j j ′ p i i ′ · · . ( 6 ) The Markov process of eq ( 2 ) is separable if the Markov matrix M is a tensor product MA ⊗ MB , i . e . , if M j j ′ i i ′ = M j i A M j ′ i ′ B ( 7 ) for Markov matrices MA and MB that determine the evolution of xA and xB . If our system is integrated so that M cannot be factored as in eq ( 7 ) , we can nonetheless choose to approximate M by a matrix of the factorizable form MA ⊗ MB . If we retain the initial probability distribution pii′⋅⋅ for x0 but replace the correct Markov matrix M by the separable approximation MA ⊗ MB , then eq ( 6 ) shows that the probability distribution p i i ′ j j ′ = M j j ′ i i ′ p i i ′ · · ( 8 ) gets replaced by the probability distribution qii′ jj′ given by q i i ′ j j ′ = M j i A M j ′ i ′ B p i i ′ · · ( 9 ) which is an approximation of pii′ jj′ . If M is factorizable ( meaning that there is no integration ) , we can factor M such that the two probability distributions qii′ jj′ and pii′ jj′ are equal and , conversely , if the two probability distributions are different , we can use how different they are as an integration measure ϕ . To define an integration measure ϕ in this spirit , we thus need to make four different choices , which collectively specify it fully and determine where the ϕ-measure belongs in our taxonomy: These four options are described in Tables 3 , 4 and 5 , and we will now explore them in detail . Table 3 lists five factoring options which all have attractive features , and we will now describe each in turn . Table 4 lists four options for which probability distributions p and q to compare . Arguably the most natural option is to simply compare the full distributions pii′jj′ and qii′jj′ that describe our knowledge of the system at both times ( the present state and the future state ) . Another obvious option is to merely compare the predictions , i . e . , the probability distributions p⋅⋅jj′ and q⋅⋅jj′ for the future state . A third interesting option is to compare merely the predictions for one of the two subsystems ( which we without loss of generality can take to be subsystem A ) , thus comparing p⋅⋅j⋅ and q⋅⋅j⋅ . Generally , the less we compare , the easier it is to get a low ϕ-value . To see this , consider a system where A affects B but B has no effect on A . We could , for example , consider A to be photoreceptor cells in your retina and B to be the rest of your brain . Then the second comparison option ( “f” ) in Table 4 would give ϕ > 0 because we predict the future of your brain worse if we ignore the information flow from your retina , while the third comparison option ( “a” ) in the table would give ϕ = 0 because the rest of your brain does not help predict the future of your retina . In other words , comparison option “a” makes ϕ vanish for afferent pathways , where information flows only inward toward the rest of the system . IIT argues that any good ϕ-measure indeed should vanish for afferent pathways , because a system can only be conscious if it can have effects on itself—other systems that it is affected by without affecting will act merely as parts of its unconscious outside world [10] . Analogously , IIT argues that any good ϕ-measure should vanish also for efferent pathways , where information flows only outward away from the rest of the system . The argument is that other systems that the conscious system affects without being affected by will again be unconscious , acting merely as unconscious parts of the outside world as far as the conscious system is concerned . Option “p” in Table 4 has this property of ϕ vanishing for efferent pathways . It is simply the time-reverse of option “a” , quantifying the ability of x 1 A to determine its past cause x 0 A instead of quantifying the ability of x 0 A to determine its future effect x 1 A . To formalize this , consider that there is nothing in the probability distribution pii′jj′ that breaks time-reversal symmetry and says that we must interpret causation as going from t0 to t1 rather than vice versa . In complete analogy with our formalism above , we can therefore define a time-reversed Markov process M ˜ whereby the future determines the past according to the time-reverse of eq ( 2 ) : p ( 0 ) = M ˜ p ( 1 ) , ( 15 ) where eqs ( 6 ) and ( 9 ) get replaced by M ˜ i i ′ j j ′ = p i i ′ j j ′ p · · j j ′ ( 16 ) and q ˜ i i ′ j j ′ = M ˜ i j A M ˜ i ′ j ′ B p j j ′ ( 1 ) . ( 17 ) This time reversal symmetry doubles the number of q-options we could list in Table 4 to six in total , augmenting qii′jj′ , q⋅⋅jj′ and q⋅⋅j⋅ by q ˜ i i ′ j j ′ , q ˜ · · j j ′ and q ˜ · · j · . In the interest of brevity , we have chosen to only list q ˜ · · j · , because of its ability to kill ϕ for efferent pathways—the formulas for the two omitted options are trivially analogous to those listed . Above we listed options for which probabilities p and q to compare to compute ϕ . To complete our specification of these probabilities , we need to choose between various options for our knowledge of the present state; the three rightmost columns of Table 4 correspond to three interesting choices . The first option is where the state is unknown , described simply by the probability distribution we have used above: p i j ( 0 ) = p i i ′ · · ( 18 ) This corresponds to us knowing M , the mechanism by which the state evolves , but not knowing its current state x0 . Note that a generic Markov process eventually converges to a unique stationary state p = p ( 0 ) = p ( 1 ) which , since it satisfies Mp = p , can be computed directly from M as the unique eigenvector whose eigenvalue is unity ( the only Markov processes that do not converge to a unique steady state are ones where M has more than one eigenvalue equal to unity; these form a set of measure zero on the set of all Markov processes ) . This means that if we consider a system that has been evolving for a significantly long time , its full two-time distribution pii′jj′ is determined by M alone; conversely , pii′jj′ determines M through eq ( 6 ) . Alternatively , if pii′jj′ is measured empirically from a time-series xt which is then used to compute M , we can use eq ( 18 ) to describe our knowledge of the state at a random time . A second option is to assume that we know the initial probability distributions for x 0 A and x 0 B , but know nothing about any correlations between them . This corresponds to replacing eq ( 18 ) by the separable distribution p i j ( 0 ) = p i · · · p · i ′ · · , ( 19 ) and can be advantageous for ϕ-measures that would conflate integration with initial correlations between the subsystems . A third option , advocated by IIT [10] , is to treat the current state as known: p i j ( 0 ) = δ i k δ i ′ k ′ , ( 20 ) i . e . , we know with certainty that the current state x0 = kk′ for some constants k and k′ . IIT argues that this is the correct option from the vantage point of a conscious system which , by definition , knows its own state . A natural fourth option is a more extreme version of the first: treating the state not merely as unknown , with p ( 0 ) given by its ensemble distribution , but completely unknown , with a uniform distribution: p i j ( 0 ) = constant . ( 21 ) Although straightforward enough to use in our formulas , we have chosen not to include this option in Table 4 because it is rather inappropriate for most physical systems . For continuous variables such as voltages , it becomes undefined . For brains , such maximum-entropy states never occur: they would have typical neurons firing about half the time , corresponding to much more extreme “on” behavior than during an epileptic seizure . The related option of consistently treating x 0 A as known but x 0 B as unknown when predicting x 1 A ( and vice versa when predicting x 1 B ) corresponds to the noising factorization options described above . For further discussion of this , including so-called “noising at the connection” , see [11 , 14 , 22] . Finally , please note that if we choose to determine the past rather than the future ( the “p”-option from the previous section and Table 4 ) , then all the choices we have described should be applied to p i j ( 1 ) rather than p i j ( 0 ) . The options in the past three sections uniquely specify two probability distributions p and q , and we want the integration ϕ to quantify how different they are from one another: ϕ ≡ d ( p , q ) ( 22 ) for some distance measure d that is larger the worse q approximates p . There are a number of properties that we may consider desirable for d to quantify integration: Any distance measure d meets the mathematical requirements of being a metric on the space of probability distributions if it obeys positivity , symmetry and the triangle inequality d ( p , q ) ≤ d ( p , r ) + d ( r , q ) . Table 5 lists seven interesting probability distribution distance measures d ( p , q ) from the literature together with their definitions and properties . All these measures are seen to have the positivity and monotonicity , and all except the first are also symmetric and true metrics . We will now discuss them one by one in greater detail . The distance dKL is the Kullback-Leibler divergence , and measures how many bits of information are lost when q is used to approximate p , in the sense that if you developed an optimal data compression algorithm to compress data drawn from a probability distribution q , it would on average require dKL ( p , q ) more bits to compress data drawn from a probability distribution p than if the algorithm had been optimized for p [23] . This has been argued to be the be the best measure because of its desirable properties related to information geometry [24 , 25] . d1 and d2 measure the distance between the vectors p and q using the L1-norm and L2-norm , respectively . The former is particularly natural for probability distributions p , since they all have L1 norm of unity: d1 ( 0 , p ) = p . = 1 . It is easy to see that 0 ≤ d1 ( p , q ) ≤ 2 and 0 ≤ d 2 ( p , q ) ≤ 2 . The measure dH is the Hilbert-space distance: if , for each probability distribution , we define a corresponding wavefunction ψ i ≡ p i 1 / 2 , then all wavefunctions lie on a unit hypersphere since they all have unit length: 〈ψ|ψ〉 = p . = 1 . The distance dH is simply the angle between two wavefunctions , i . e . , the distance along the great circle on the hypersphere that connects the two , so dH ( p , q ) ≤ π/2 . It is also the geodesic distance of the Fisher metric , hence a natural “coordinate free” distance measure on the manifold of all probability distributions . The measure dSJ is the Shannon-Jensen distance , whose square is defined as the average of the KL-divergences of the two distributions to their average: d S J ( p , q ) 2 ≡ d K L ( p , [ p + q ] / 2 ) + d K L ( q , [ p + q ] / 2 ) 2 = S [ ( p + q ) / 2 ] - ( S [ p ] + S [ q ] ) / 2 . ( 23 ) It is bounded by 0 ≤ dSJ ≤ 1 , satisfies the triangle inequality and is information-theoretically motivated [26] . The measure dEM is the Earth-Movers distance [27] . If we imagine piles of earth scattered across the space x , with p ( x ) specifying the fraction of the earth that is in each location , then dEM is the average distance that you need to move earth to turn the distribution p ( x ) into q ( x ) . The quantity dij in the definition in Table 5 specifies the distance between points i and j in this space . For example , if x is a 3D Euclidean space , this may be chosen to be simply the Euclidean metric , while if x is a bit string , dij may be chosen to be the L1 “Manhattan distance” , i . e . , the number of bit flips required to transform one bit string into another . IIT 3 . 0 argues that the earth mover’s distance dEM is the most appropriate measure d on conceptual grounds ( whereas IIT 2 . 0 was still implemented using dKL ) . Unfortunately , dEM rates poorly on the tractability criterion . It’s definition involves a linear programming problem which needs to be solved numerically , and even with the fastest algorithms currently available , the computation grows faster than quadratically with the number of system states—which in turn grows exponentially with the number of bits . For continuous variables x , the number of states and hence the computational time is formally infinite . The measure dMD is based on “mismatched decoding” as advocated by [15] . The distance measure dMD is defined not for all probability distributions , but for all distributions over two variables , which we can write with two indices as pij: d M D ( p , q ) ≡ I ( p ) - max β I * ( p , q , β ) , ( 24 ) where I ( p ) ≡ - ∑ j p · j log p · j + ∑ i j p i j log p j | i , I * ( p , q , β ) ≡ - ∑ j p · j log ∑ i q j | i β p i · + ∑ i j p i j log q j | i β , ( 25 ) and the conditional distribution qj|i ≡ qij/qi⋅ . Here I ( p ) is simply the mutual information between the two variables , since combining eq ( 34 ) with the conditional entropy definition from eq ( 37 ) gives the well-known equivalent expression for mutual information I ( A , B ) = S ( A ) - S ( A | B ) . ( 26 ) I* ( p , q , β ) can be interpreted as the amount of information that one variable predicts about the other if the correct conditional distribution pj|i is replaced by a possibly incorrect one q j | i β ( renormalized to sum to unity ) when making the prediction [28] . This renormalization is strictly speaking unnecessary , because it cancels out between the two terms in I* ( p , q , β ) . Raising probabilities to positive powers β has the effect of concentrating them ( decreasing entropy ) if β > 1 and spreading them more evenly ( increasing entropy ) if β < 1 . It can be shown that I* ( p , q , β ) ≤ I ( p ) with equality for q = p and β = 1 , and that I* ( p , q , β ) ≥ 0 , so one always has 0 ≤ dMD ( p , q ) ≤ I ( p ) [28] . Mismatched decoding can presumably be further generalized by replacing the maximization over powers pβ by maximization over arbitrary monotonically increasing functions f ( p ) that map the unit interval onto itself . The integration measures of IIT3 . 0 have a more complex probability comparison that cannot be fully cast in the form of a simple function of d ( p , q ) : it makes the metric choice d ( p , q ) = dEM ( p , q ) , but considers not only probability distributions for the whole system and a bipartition , but also for all possible subsets , providing an elaborate interpretation of the results in terms of “conceptual structures” [11] .
Our taxonomy of integration measures is determined by four choices: of factorization , variable selection , conditioning and distance measure . Although we have now explored these four choices one at a time , there are important interplays between them that we must examine . First of all , the three optimal factorization options in Table 3 depend on what is being optimized , so let us now explore which of these optimizations are feasible and interesting to perform in practice and let us find out what the corresponding factorizations and ϕ-measures are . The mathematics problem we wish to solve is ϕ ≡ min M A , M B d ( p , q ) ( 27 ) i . e . , minimizing d ( p , q ) over MA and MB given the constraints that MA and MB are markov Matrices: M · j A = 1 , M · j ′ B = 1 , M i j A ≥ 0 and M i ′ j ′ B ≥ 0 . Table 4 specifies the options for how p and q are computed and how q depends on MA and MB , while Table 5 specifies the options for computing the distance measure d . We enforce the column sum constraints using Lagrange multipliers , minimizing L ≡ d ( p , q ) - ∑ i λ i ( M · i A - 1 ) - ∑ i ′ μ i ′ ( M · i ′ B - 1 ) , ( 28 ) and need to check afterwards that all elements of MA and MB come out to be non-negative ( we will see that this is indeed the case ) . As mentioned , numerical tractability is a key issue for integration measures . This means that it is valuable if the Lagrange minimization can be rapidly solved analytically rather than slowly by numerical means , since this needs to be done separately for large numbers of possible system partitions . There is only one d-option out of the above-mentioned five for which I have been able to solve the optimization over M-factorizations analytically: the KL-divergence dKL . The runner-up for tractability is d2 , for which everything can be easily solved analytically except for a final column normalization step , but the resulting formulas are cumbersome and unilluminating , falling foul of the interpretability criterion . Although dKL lacks the symmetry property , it has the above-mentioned positivity , monotonicity and interpretability properties , and we will now show that it also has the tractability property . Let us begin with the q-options in the upper left corner of Table 4 , i . e . , comparing the two-time distributions treating the present state as unknown . Substituting eq ( 9 ) into the definition of dKL from Table 5 gives d K L ( p , q ) = ∑ i i ′ j j ′ p i i ′ j j ′ log p i i ′ j j ′ p i i ′ · · M j i A M j ′ i ′ B = S ( x 0 ) - S ( x ) - ∑ i j p i · j · log M j i A - ∑ i ′ j ′ p · i ′ · j ′ log M j ′ i ′ B , ( 29 ) where the entropy for a random variable x with probability distribution p is given by Shannon’s formula [29] S ( x ) = - ∑ i p i log p i . ( 30 ) To avoid a profusion of notation , we will often write as the argument of S a random variable rather than its probability distribution . For convenience , we will take all logarithms to be in base 2 for discrete distributions ( so that entropies are measured in units of bits ) and in base e for continuous Gaussian distributions ( so that equations get simpler ) . In the latter case , where the entropy is based on the natural logarithm , entropy is measured in “nits” or “nats” which equal 1/ln2 ≈ 1 . 44 bits . Substituting eq ( 29 ) into eq ( 28 ) and requiring vanishing derivatives with respect to M i j A , M i ′ j ′ B , λj and μj′ shows that the solution to our minimization problem is M j i A = p i · j · p i · · · , M j ′ i ′ B = p · i ′ · j ′ p · i ′ · · . ( 31 ) We recognize these equations as simply the Markov matrix estimator from eq ( 4 ) applied separately to subsystems A and B after marginalizing over the other system . Substituting this back into eq ( 9 ) gives q i i ′ j j ′ = p i i ′ · · p i · j · p · i ′ · j ′ p i · · · p · i ′ · · . ( 32 ) Although the full probability distributions q and p typically differ , eq ( 32 ) implies that three marginal distributions are identical: qi⋅j⋅ = pi⋅j⋅ , q⋅i′⋅j′ = p⋅i′⋅j′ and qij′⋅⋅ = pij′⋅⋅ . Substituting eq ( 32 ) back into the definition of dKL gives the extremely simple result that the integration is ϕ o t u k ( p ) = ∑ i i ′ j j ′ p i i ′ j j ′ log p i i ′ j j ′ p i · · · p · i ′ · · p i i ′ · · p i · j · p · i ′ · j ′ = I ( x A , x B ) - I ( x 0 A , x 0 B ) , ( 33 ) where the mutual information between two random variables is given in terms of entropies by the standard definition I ( x A , x B ) ≡ S ( x A ) + S ( x B ) - S ( x ) . ( 34 ) Since we will be deriving a large number of different ϕ-measures that we do not wish to conflate with one another , we superscript each one with four code letters denoting the four taxonomical choices that define it . These letter codes are and are defined in Tables 3 , 4 and 5 . For example , the integration measure ϕotuk from eq ( 33 ) denotes optimized ( o ) factorization comparing the two-time ( t ) probability distributions with the current state unknown ( u ) and KL-divergence ( k ) . Almost all measures discussed below will involve the k-measure ( KL-divergence ) , so when this is the case we will typically drop this last index k to avoid a confusing profusion of indices , for example writing ϕotuk = ϕotu . For brevity , we will also define ϕM ≡ ϕotu , since we will be referring to this “Markov measure” ϕotu many times below . Although we derived this optimal factorization by comparing the two-time distribution ( option t ) for an unknown state ( option u ) , an analogous calculation leads to the exact same optimal factorization for the options a+u , s+f and a+s . The option t+s is undefined and the option f+u gives messy equations I have been unable to solve analytically . It is therefore reasonable to view eq ( 31 ) as the optimal factorization when the state is unknown ( option o ) , and for the remainder of this paper , we will simply define the o-option as using the factorization given by eq ( 31 ) . Note that our result in eq ( 33 ) involves a time-asymmetry , singling out t0 rather than t1 in the second term . This is because we chose to interpret our Markov process as operating forward in time , determining the state at t1 from the state at t0 . As we discussed in the previous section , we could equally well have done the opposite , using the Markov process M ˜ operating backward in time , which would have yielded the alternative integration measure ϕ o t ˜ u ( p ) = I ( x A , x B ) - I ( x 1 A , x 1 B ) . ( 35 ) In practice , one usually estimates all statistical properties from a time-series that is assumed to be stationary . This means that I ( x 0 A , x 0 B ) = I ( x 1 A , x 1 B ) , so that the these two integration measures become identical . In the paper [13] where Barrett & Seth proposed their easier-to-compute integration measure ϕB ( see below ) , they also mentioned an alternative measure that they termed ϕ ˜ E , defined by ϕ ˜ E ≡ S ( x 0 A | x 1 A ) + S ( x 0 B | x 1 B ) - S ( x 0 | x 1 ) , ( 36 ) where the conditional entropy of two variables A and B is defined by S ( A | B ) ≡ S ( A , B ) - S ( B ) . ( 37 ) This measure had been introduced earlier by Ay [30 , 31] in a context unrelated to IIT , under the name “stochastic interaction” , and was further discussed in [15 , 32] . Applying eqs ( 37 ) and ( 34 ) to eq ( 36 ) shows that ϕ ˜ E = S ( x A ) - S ( x 1 A ) + S ( x B ) - S ( x 1 B ) - S ( x ) + S ( x 1 ) = I ( x A , x B ) - I ( x 1 A , x 1 B ) = ϕ o t ˜ u , ( 38 ) i . e . , that ϕ ˜ E is identical to the time-reversed Markov measure ϕ o t ˜ u . This equivalence provides another convenient interpretation of ϕ o t ˜ u: as the average KL-divergence between ( i ) the probability distribution of the past state x0 given the present state x1 and ( ii ) the product of these conditional distributions for the two subsystems . It is also interesting to compare our result in eq ( 33 ) with the popular integration measure ϕ B ( p ) = I ( x 0 , x 1 ) - I ( x 0 A , x 1 A ) - I ( x 0 B , x 1 B ) ( 39 ) proposed by Barrett & Seth [13] . The intuition behind this definition is to take the amount of information that a system predicts about its future and subtract of the information predicted by both of its subsystems . Unfortunately , the result can sometimes go negative [14 , 15] , violating the desirable positivity property and making the ϕB difficult to interpret . Consider the simple example of two independent bits that never change . If they start out perfectly correlated , then they will remain perfectly correlated , giving I ( x 0 , x 1 ) = I ( x 0 A , x 1 A ) = I ( x 0 B , x 1 B ) = 1 and integrated information ϕB ( p ) = −1 . By substituting eq ( 34 ) into eqs ( 33 ) and ( 39 ) , we find that ϕ B ( p ) = ϕ M ( p ) - I ( x 1 A , x 1 B ) . ( 40 ) In other words , we can make the Barrett-Seth measure non-negative by adding back any final mutual information between the two subsystems . When this is done , it becomes the integration measure we derived , therefore having a simple information-theoretic interpretation: it is the KL-divergence between the actual probability distribution p and the best separable approximation , which is guaranteed to be non-negative . The measure ϕM is also closely related to the mismatched decoding measure ϕMD introduced in [15] . ϕMD makes the same taxonomical choices “otu” as ϕM for the first three options: optimal factorization ( o ) , comparing full two-time distributions ( t ) , and treating the past state as unknown ( u ) . However , it uses probability distance measure “m” ( mismatched decoding dMD ) instead of KL-divergence . We can therefore write this measure in our notation as ϕotum = dMD ( p , q ) , where q is the optimal factorization given by eq ( 32 ) . Whether this factorization is also optimal in the sense of minimizing dMD ( p , q ) is not obvious . The measure ϕM ( or more specifically its time-reverse ϕ o t ˜ u k ) has been criticized in [15 , 25] for being able to exceed the mutual information I ( x0 , x1 ) between the past and present: for example , if a two-bit system evolves from “00” to either “00” or “11” with equal probability , then ϕ M = I ( x A , x B ) - I ( x 0 A , x 0 B ) = 1 - 0 = 1 bit , even though I ( x0 , x1 ) = 0 . This means that ϕM counts as a contribution to integration also correlated random noise added to both subsystems . It is debatable whether this should count as integration: the “con” argument is that no information flows between the subsystems , while the “pro” argument is that the two subsystems get linked by shared information flowing into both of them . Both ϕM and ϕMD have intuitive bounds: 0 ≤ ϕM ≤ I ( xA , xB ) and 0 ≤ ϕMD ≤ I ( x0 , x1 ) ; these upper bounds correspond to the total mutual information across space and time , respectively . Let us now turn to factorization option “x” , optimized knowing the current state . Consider some conscious observer ( perhaps the system itself ) who knows nothing about the system except its dynamics ( encoded in M ) and its state at the present instant , encoded in x0 = kk′ . What can this observer say about the system state at earlier and later times ? How integrated will this observer feel that the system is ? To answer this question , we simply want to find the best approximate factorization of the conditional future state Mjj′kk′ ( or the past state Mkk′ii′ ) , where k and k′ are known constants . To gain intuition for this , let us temporarily write this conditional distribution as pii′ , suppressing the known parameters kk′ for simplicity . Given an arbitrary bivariate probability distribution pii′ , what is best separarable approximation qii′ ≡ ai bi′ in the sense that it minimizes dKL ( p , q ) ? By minimizing dKL ( p , q ) using Lagrange multipliers , one easily obtains the long-known result that ai = pi . , bi′ = p . i′ and dKL ( p , q ) = I , the mutual information of p . In other words , even if we had never heard of marginal distributions or mutual information , we could derive them all from dKL: the best factorization simply uses the marginal distributions , and the mutual information of a bivariate distribution is simply the KL-measure of how non-separable it is . This means that the optimal factorization given k and k′ is simply the one giving the marginal conditional distributions M j i A = p k k ′ j · p k k ′ · · , M j ′ i ′ B = p k k ′ · j ′ p k k ′ · · , ( 41 ) and the corresponding integration is simply ϕ x f k k = I ( x 1 A , x 1 B | x 0 ) . ( 42 ) ϕxtkk is identical . We can alternatively obtain this result directly from eq ( 33 ) by noting that the I ( x 0 A , x 0 B ) -term vanishes now that the state x0 is known . This result highlights a striking and arguably undesirable feature of measures based on the x-factorization option: they vanish for any deterministic system ! If the system is deterministic and the present state x0 is known , then the future state x1 is also known , so all entropies in eq ( 42 ) vanish and we obtain ϕ = 0 . With ϕ-measures based on x-factorization , the only source of integration is therefore correlated noise generated by the system . Let us now turn to our final factorization option , “a” , where we pick the state-independent factorization that minimizes integration on average . Given the present state x0 = kk′ , let us compare the exact and approximate future probability distributions p j j ′ = P ( x 1 = j j ′ | x 0 = k k ′ ) = M k k ′ j j ′ , q j j ′ = P ( x 1 A = j | x 0 A = k ) P ( x 1 B = j ′ | x 0 B = k ′ ) = M k j A M k ′ j ′ B ( 43 ) by computing their KL-divergence ϕ = dKL ( p , q ) . The answer clearly depends on the present state kk′ , and we saw in the previous section what happens when we minimize separately for each state kk′ . Let us now instead average dKL ( p , q ) over all current states and find the state-independent factorization that minimizes this average: 〈 d K L ( p , q ) 〉 = ∑ k k ′ P ( x 0 = k k ′ ) d K L ( p , q ) | x 0 = k k ′ = ∑ k k ′ p k k ′ · · ∑ j j ′ M k k ′ j j ′ log M k k ′ j j ′ M k j A M k ′ j ′ B . ( 44 ) Substituting eq ( 6 ) shows that this expression is identical to that from eq ( 29 ) , so minimizing it gives the exact same optimal factors MA and MB and the exact same minimum ϕ . The comparison option “t” gives the same result as well , so in conclusion , although they appear quite different from their definitions , the factorization options “o” and “a” are in fact identical . Now that we have derived the explicit form of all our factorization options , we can complete our integration measure classification . Our taxonomy is determined by four choices: of factorization ( n/m/o/x/a ) , variable selection ( t/f/a/p ) , conditioning ( u/s/k ) and distance measure ( k/1/2/h/s/e/m ) . Although this nominally gives 5 × 4 × 3 × 7 = 420 different integration measures , most of these options turn out to be zero , undefined or identical to other options . For noising factorizations ( factorization options n and m ) , subsystem B is randomized , so the only well-defined options are ϕnas* , ϕnak* , ϕnps* , ϕnpk* , ϕmas* , ϕmak* , ϕmps* and ϕmpk* , where * denotes any option for the distance measure . For o-factorization , we find that ϕoau* = ϕoas* = ϕopu* = ϕops* = 0 and ϕotk* = ϕofk* . For x-factorization , ϕxt** is undefined and one easily shows that ϕxak* = ϕxpk* = 0 , ϕxau* = ϕxas* and ϕxpu* = ϕxps* . We interpret k-conditioning as x0 being known for o-factorization and as x 0 A being known for noising factorizations , since the reverse options vanish and are undefined , respectively . Whereas there are strong interactions between the factorization , variable selection and conditioning , we can freely choose any of the 7 distance measures independently of the other choices without changing whether ϕ vanishes or is well-defined . We consider the option k ( KL-divergence ) by default below since it results in the simplest and most intuitive formulas; the formulas for the other options are straightforward to derive by combining Tables 3 , 4 and 5 . This leaves us with only the 21 separate options shown in Table 2 to consider . To provide intuition for these formulas , let us recapitulate key definitions in words: Table 1 summarizes the desirable and undesirable traits for each of these integration measures , showing that merely a handful lack any major drawbacks . Let us now rate the various options in more detail . For the choice of probability distance measure ( k/1/2/h/s/e/m ) , option “e” ( the Earth-Mover’s distance dEM used in ϕ3 . 0 [11] ) remains an attractive candidate for discrete distributions with small number of bits , but is otherwise computationally unfeasible as we discussed above . All options in Table 1 except ϕ3 . 0 and ϕMD therefore use option “k” ( the KL-divergence ) . Note that whether it is an advantage for the probability distance measure to be symmetric ( as advocated in [11] ) depends on the interpretational context . For example , there is nothing asymmetric about the mutual information that ends up defining ϕM in Table 2 . For the choice of factorization ( n/m/o/x/a ) , we can quickly dispense with option “a” ( for being identical to “o” ) and option “x” ( because it has the highly undesirable property of always vanishing for deterministic systems ) . Which of the remaining options ( n/m/o ) is preferable depends on other choices . If one wishes to use a distance measure other than the KL-divergence , then the noising options “n” or “m” are computationally preferable , since the optimal factorization “o” can no longer be found analytically . Otherwise , “m” is arguably inferior to “o” because it is no simpler to evaluate and can overestimate the integration as described above . If one has a philosophical preference for the factorization depending only on the mechanism M and not on any other information about state probabilities , then “n” is the only choice . If one wishes to consider continuous systems , on the other hand , “n” is undefined . In summary , the best factorizations are therefore “o” and “n” , depending ones preferences . In practice , numerical experiments show that “n” , “m” and “o” usually give quite similar ϕ-values for a wide range of M-matrices and probability distributions , so the choice between the three is a relatively minor one . Turning now to the choice variable selection and conditioning , Table 1 shows that many of the otherwise well-defined integration measures from Table 2 have serious flaws . Neither ϕots and ϕofs are guaranteed to vanish for separable systems , which means that we cannot in good conscience interpret them as measures of integration . Numerical experiments show that ϕnas , ϕnps , ϕmas and ϕnps tend to be extremely small in practice ( ϕmas is plotted in Fig 2 ) . This is because they differ little from the corresponding measures using optimal factorization ( ϕoas and ϕops ) , which always vanish . In other words , they are not really measures of integration , merely measures of how suboptimal the factorizations “n” and “m” are . For brevity , we have included merely three of these six flawed measures in Table 1 . Fig 2 shows that ϕofu also tends to be much smaller than some other integration measures . We can intuitively understand this by recalling that ϕoau = 0 , which means that optimal factorization lets us predict the future marginal distributions for A and B perfectly . Since ϕofu quantifies the inability of optimal factorization to predict the full future distribution , we expect that it will at most be of the order of I ( x 1 A , x 1 B ) , the extent to which this distribution is not separable ( determined by its marginal distributions ) . For randomly generated probability distributions generated as in Fig 2 ) , one can show that I ( x 1 A , x 1 B ) → 1 - 1 / 2 ln 2 ≈ 0 . 28 bits in the limit where n → ∞ , and numerical experiments indicate that ϕofu is never much larger than this value for any p . Dispensing with flawed/problematic ϕ-measures narrows our list of remaining top candidates to merely nine: ϕotu , ϕotum , ϕofk , ϕoak , ϕopk , ϕnak , ϕnpk , ϕmak and ϕmpk . Morover , the last six can be elegantly combined into merely three even better ones . As we discussed above , they have the advantage that they vanish for either afferent or efferent systems . By following the prescription of [10] and taking the minimum of two such complementary measures , we can construct an even better one that vanishes for both afferent and efferent systems . All three of these improved measures are listed in Table 2 . The first is ϕ2 . 5 ≡ min{ϕnak , ϕnpk} . We denote it “2 . 5” because it combines attractive features of both IIT2 . 0 and IIT3 . 0: it starts with the ϕnpk , which is precisely the IIT2 . 0 measure , and improves it by taking the minimum of cause/effect integration in the spirit of IIT3 . 0 ( but retaining the KL-divergence of IIT2 . 0 instead of the harder-to-compute Earth-mover’s distance of IIT3 . 0 ) . The second is ϕ 2 . 5 ′ ≡ min { ϕ mak , ϕ mpk } , which has the advantage of remaining defined even for continuous variables . The third is ϕ 2 . 5 ′ ′ ≡ min { ϕ oak , ϕ opk } , which uses the optimal factorization . In summary , our taxonomy of ϕ-measures produces merely a handful of truly attractive options: ϕ2 . 5 , ϕ 2 . 5 ′ , ϕ 2 . 5 ′ ′ , ϕ3 . 0 , ϕMD , ϕM and ϕ k k ′ M . Fig 2 shows examples of what they evaluate to numerically . The lower panel shows that for randomly generated probability distributions , none of them exceed 1 − 1/2ln2 ≈ 0 . 28 bits on average , which as mentioned above is the mutual information in a random bivariate distribution . However , ϕ2 . 5 , ϕ 2 . 5 ′ , ϕ 2 . 5 ′ ′ , ϕM , ϕMD and ϕ k k ′ M can get arbitrarily large for some systems , as illustrated in the top panel , growing logarithmically with the size n of the subsystems A and B . In other words , the maximum integration is of the order of the number of subsystem bits . For the example shown where the dynamics merely swaps the two subsystems , we obtain ϕ2 . 5 = log2 n , because noising gives MA = 1/n , q = 1/n2 and p is a Kronecker δ . ϕM , ϕMD and ϕ k k ′ M are seen to give about twice the integration for this example . Note that although this dynamics M that merely swaps the subsystems has such a large ϕ-value only for this particular cut that separates the systems being swapped . Consider , for example , a system of four bits labeled 1 , 2 , 3 and 4 , where the dynamics swaps 1 with 3 and 2 with 4 . There is a different cut where ϕ = 0: simply define the new subsystems A’ and B’ to be the first and second halves of the A and B-systems , i . e . , A′ = 1 , 3 and B′ = 2 , 4 . The swapping is now carried out internally within A’ and B’ , revealing that there is no integration and upper-case Φ = 0 . However , there are plenty of systems for which even the true integration Φ grows like the number of subsystem bits , log2 n . A simple example accomplishing this ( in the spirit of the random coding example in [12] ) is when the n4 probabilities pii′jj′ are all set to zero except for a randomly selected subset of n2 of them that are set to 1/n2 . Now ϕM ∼ log2 n even when minimized over all bipartitions of the 2log2 n bits in the system . For this example , we have S ( x ) = log2 n2 = 2 log2 n . The marginal distributions for xA , xB , x 0 A and x 0 B are all rather uniform , with entropy on average less than a bit from the value for a uniform distribution , giving S ( xA ) ∼ S ( xB ) ∼ log2 n2 , S ( x 0 A ) ∼ S ( x 0 B ) ∼ log 2 n , I ( xA , xB ) = S ( xA ) + S ( xB ) − S ( x ) ∼ 2 log2 n , I ( x 0 A , x 0 B ) = S ( x 0 A ) + S ( x 0 B ) - S ( x 0 ) ∼ 0 and therefore ϕ M = I ( x A , x B ) - I ( x 0 A , x 0 B ) ∼ 2 log 2 n ∼ log 2 n . Fig 3 shows that the measures ϕM and ϕMD can sometimes be quite similar: they give numerically similar values for the 3 , 000 random examples shown . Moreover , they appear to satisfy the inequality ϕofum ≤ ϕotuk . Further examination shows that for these these random examples , the β-complication in eq ( 24 ) makes essentially no perceptible difference in practice , in the sense that the computation of ϕMD can be accurately accelerated by setting β = 1 rather than minimizing over it . However , [15] shows that there are real-world cases where β is far from unity and also where ϕMD ≪ ϕM , particularly when noise correlations dominate over causal correlations . To understand this , consider the extreme case of two perfectly correlated bits that are independently randomized by both time 0 and time 1 , so that x 0 A = x 0 B and x 1 A = x 1 B , with no correlation between the two times . Then ϕMD = 0 whereas ϕ M = I ( x A , x B ) - I ( x 0 A , x 0 B ) = 2 - 1 = 1 , which is arguably undesirable . All our previous results are fully general , applying regardless of whether the variables are discrete ( such as bits that equal zero or one ) or continuous ( such as voltages or other variables measured in fMRI , EEG , MEG or electrophysiology studies ) . We can view the latter as the n → ∞ limit of the former , since a single real number can be represented as an infinite string of bits . In this section , we will focus on the continuous case and see how our previous formulas can be greatly simplified by assuming Gaussianity . We therefore replace i , i′ , j and j′ in all our formulas by x 0 A , x 0 B , x 1 A and x 1 A , respectively , and replace all sums by integrals .
As described in detail in Results , six Φ-measures stand out from the taxonomy of hundreds of measures as particularly attractive: ΦM , Φ k k ′ M , Φ3 . 0 , Φ2 . 5 , Φ 2 . 5 ′ and Φ 2 . 5 ′ ′ . ΦM retains all the attractive features of the Barrett/Seth measure ΦB and adds further improvements: it is guaranteed to vanish for separable systems and to never be negative . If state-dependence is viewed as desirable , then its cousin Φ k k ′ M adds that feature too . Φ3 . 0 is the measure advocated by IIT3 . 0 and has the many attractive features described in [11] . It has the drawback of being the slowest of all the measures to evaluate numerically: its definition involves a linear programming problem which needs to be solved numerically , and even with the fastest algorithms currently available , the computation for a given bipartition grows faster than quadratically with the number of system states—which in turn grows exponentially with the number of bits , and is infinite for continuous variables . The remaining three top measures , Φ2 . 5 , Φ 2 . 5 ′ and Φ 2 . 5 ′ ′ , share with Φ3 . 0 the arguably desirable feature of vanishing for afferent and efferent systems , but are much quicker to compute . Φ2 . 5 combines core ideas from IIT3 . 0 with the computational speed of IIT2 . 0 [10 , 11] and elegantly depends only on the system’s dynamics and present state , not on any assumptions about which states are more probable . Its drawback of being infinite for continuous variables is overcome by its cousin Φ 2 . 5 ′ . A potential philosophical objection to both Φ2 . 5 and Φ 2 . 5 ′ is that they are arguably not measures of integration , but measures of how suboptimal the factorizations “n” and “m” are , since they would both vanish if an optimal factorization were used—the measure Φ 2 . 5 ′ ′ eliminates this concern . Although the results in this paper will hopefully prove useful , there is ample worthwhile work left to do on integration measures . One major open question is how to best handle asymmetric partitions . We deliberately sidestepped this challenge in the present paper , since it is independent of our results , which is why the subtle normalization issue raised by [10 , 11 , 13 , 22] never entered . The crux is that if we apply any of the measures in our taxonomy with an asymmetric bipartition , the resulting ϕ-value will tend to get small when any of the two subsystems is very small , so simply defining Φ as the minimum of ϕ over all bipartitions ( symmetric or not ) makes no sense . IIT3 . 0 makes an interesting proposal [11] for how to handle asymmetric partitions , and it is worthwhile exploring whether there are other atttractive options as well . Another foundational question is whether our taxonomy can be placed on a firmer logical footing . Although our classification based on factorization , comparison , conditioning and measure may seem sensible and exhaustive , it is interesting to consider whether one or several Φ-measures can be rigorously derived from a small set of attractive axioms alone , in the same spirit as Claude Shannon derived his famous entropy formula , eq ( 30 ) . Yet another foundational question is whether integration maximization can be placed on a firmer physical footing , as advocated by [33 , 34] in the context of continuous physical fields and by [12] in the context of quantum systems . The formulas in our taxonomy take information , measured in bits , as a starting point . But when I view a brain or computer through my physicist eyes , as myriad moving particles , then what physical properties of the system should be interpreted as logical bits of information ? I interpret as a “bit” both the position of certain electrons in my computer’s RAM memory ( determining whether the micro-capacitor is charged ) and the position of certain sodium ions in your brain ( determining whether a neuron is firing ) , but on the basis of what principle ? Surely there should be some way of identifying consciousness from the particle motions alone , or from the quantum state evolution , even without this information interpretation ? If so , what aspects of the behavior of particles corresponds to conscious integrated information ? In other words , how can we generalize the quest for neural correlates of consciousness to physical correlates of consciousness ? IIT argues that the consciousness occurs at precisely the level of course-graining in space and time that maximizes Φ [10] , which is a prediction that should be tested . A more practical question involves exploring ways of generalizing and further improving our graph-theory-based approximation for exponential speedup . One obvious generalization would involve taking advantage of the structure of ∑ ( which our method ignored ) and the effect of x ( for those Φ-measures that are state-dependent ) . Another interesting opportunity is to generalize from continuous Gaussian systems to arbitrary discrete systems . For example , if the system consists of b different bits coupled by a nonlinear network of gates , one can apply a similar graph-theory approach by defining a b × b coupling matrix Aij that in some way quantifies how strongly flipping the jth bit would affect the ith bit at the next timestep . As an example , consider defining Aij as the probability that flipping the jth bit will flip the ith bit at the next timestep . If we have six bits evolving according to x 1 = ( a 1 b 1 c 1 d 1 e 1 f 1 ) = f ( x 0 ) = ( a 0 NOT a 0 RANDOM c 0 XOR d 0 c 0 AND d 0 c 0 AND d 0 AND e 0 ) , then the coupling matrix is A = ( 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 p d p c 0 0 0 0 p d e p c e p c d 0 ) , were pc denotes the probability that c0 = 1 , pde denotes the probability that d0 = 1 and e0 = 1 , etc . This coupling matrix is block-diagonal , showing that the bits a , b are completely independent of the others . For a state-independent Φ-measure , these probabilities can be computed as time-averages , otherwise they are each zero or one depending on the state . In either case , some elements of the A-matrix can be small but non-zero ( making the graph-theory approximation useful ) if the system involves noisy gates or other randomness . As regards practical challenges , it is important to note that there are many other issues besides speed that deserve further work because they have hindered the practical computation of integration Φ-measures from real brain data , including non-stationarity , statistical issues with estimating large numbers of parameters from short data windows without overfitting , possibilities of statistical bias , numerical instabilities , etc . Last but not least , a veritable goldmine of data is becoming available in neuroscience and other fields , and it will be fascinating to measure Φ for these emerging data sets . In particular , the exponentially faster Φ-measures we have proposed will hopefully facilitate quantitative tests of theories of consciousness . | How can one determine whether an unresponsive patient is conscious or not ? Of all the information processing in your brain that can be measured with modern sensors , which corresponds to information that you are subjectively aware of and which is unconscious ? A theory that has garnered much recent attention proposes that the answer involves measuring a quantity called integration that quantifies the extent to which information is interconnected into a unified whole rather than split into disconnected parts . Unfortunately , proposed measures of integration are too slow to compute in practice from patient data . In this paper , I explore and classify existing and novel integration measures by various desirable properties , and derive useful exact and approximate formulas that can be applied to real-world data from laboratory experiments without posing unreasonable computational demands . This improves the prospects of making fascinating questions and theories about consciousness experimentally testable . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"distance",
"measurement",
"taxonomy",
"engineering",
"and",
"technology",
"random",
"variables",
"neuroscience",
"covariance",
"data",
"management",
"cognitive",
"neuroscience",
"probability",
"distribution",
"mathematics",
"statistics",
"(mathematics)",
"thermodynamics",
"c... | 2016 | Improved Measures of Integrated Information |
Dogs are considered sentinels in areas of Trypanosoma cruzi transmission risk to humans . ELISA is generally the method of choice for diagnosing T . cruzi exposure in dogs , but its performance substantially depends on the antigenic matrix employed . In previous studies , our group has developed four chimeric antigens ( IBMP-8 . 1 , 8 . 2 , 8 . 3 , and 8 . 4 ) and evaluated their potential for diagnosing T . cruzi exposure in humans . For human sera , these chimeric antigens presented superior diagnostic performances as compared to commercial tests available in Brazil , Spain , and Argentina . Therefore , in this study we have evaluated the potential of these antigenic proteins for detection of anti-T . cruzi IgG antibodies in dog sera . The IBMP-ELISA assays were optimized by checkerboard titration . Subsequently , the diagnostic potential was validated through analysis of ROC curves and the performance of the tests was determined using double entry tables . Cross-reactivity was also evaluated for babesiosis , ehrlichiosis , dirofilariosis , anaplasmosis , and visceral leishmaniasis . Best performance was shown by IBMP-8 . 3 and IBMP-8 . 4 , although all four antigens demonstrated a high diagnostic performance with 46 positive and 149 negative samples tested . IBMP-8 . 3 demonstrated 100% sensitivity , followed by IBMP-8 . 4 ( 96 . 7–100% ) , IBMP-8 . 2 ( 73 . 3–87 . 5% ) , and IBMP-8 . 1 ( 50–100% ) . The highest specificities were achieved with IBMP-8 . 2 ( 100% ) and IBMP-8 . 4 ( 100% ) , followed by IBMP-8 . 3 ( 96 . 7–97 . 5% ) and IBMP 8 . 1 ( 89 . 1–100% ) . The use of chimeric antigenic matrices in immunoassays for anti-T . cruzi IgG antibody detection in sera of infected dogs was shown to be a promising tool for veterinary diagnosis and epidemiological studies . The chimeric antigens used in this work allowed also to overcome the common hurdles related to serodiagnosis of T . cruzi infection , especially regarding variation of efficiency parameters according to different strains and cross-reactivity with other infectious diseases .
Chagas disease ( CD ) is a vector-borne , neglected parasitic illness caused by the hemoflagellate protozoan Trypanosoma cruzi . According to recent estimates , approximately 6 million people are affected by CD in 21 Latin American endemic countries with ~14 , 000 deaths per year being attributed to this disease [1] . Increased international migration of infected individuals has spread CD toward non-endemic settings , including North American , European , Asian , and Oceanian countries [2 , 3] . Trypanosoma cruzi transmission involves complex networks of interactions of wild and domestic mammalian hosts , and triatomine vectors . Dogs , cats , pigs , and goats are the main domestic mammalian species investigated for T . cruzi infection in endemic areas . Dogs and cats represent the first domestic T . cruzi hosts studied by Carlos Chagas: a cat in Lassance ( Minas Gerais state , Brazil ) was the first mammalian host in which he found trypomastigote forms of the parasite in the blood , whereas dogs were among the first experimental models used in his research . Since then , several studies have shown that dogs and cats can be competent T . cruzi reservoirs , but as described for the other mammals , their importance in the transmission cycle varies according to the geographic regions and local characteristics . Particularly , domestic dogs have been commonly implicated as blood meal sources for triatomine vectors [4 , 5] . In the Argentinean Gran Chaco , both dogs and cats are epidemiologically important and described as highly infective to triatomine vectors [4 , 5] . Active transmission , which includes symptomatic dogs , was also observed in the southern United States [6 , 7] and throughout the Americas [8] . A different scenario is seen in Brazil . Despite being exposed to parasite ( as evidenced by the high seropositive rates by IFAT and ELISA ) , T . cruzi isolation from dogs , whether by hemoculture or xenodiagnoses , is rarely documented [6–8] . However , dogs may act as efficient sentinel animals . Xavier et al . [7] observed positive association between serologically positive dogs and: ( 1 ) lower diversity of small mammal fauna and ( 2 ) high rates of small mammal fauna with high infective competence as expressed by positive hemocultures . The presence of T . cruzi-infected dogs in households is associated with higher risk of human infection; as such , they can be used as sentinels [9] . Furthermore , dogs may eventually develop American trypanosomiasis ( AT ) by T . cruzi , presenting morphofunctional cardiac lesions or sudden death similar to those seen in humans [10 , 11] . Hence , AT may represent a health risk for dogs [12] . The pivotal importance is that seropositive dogs reflect exposure to T . cruzi and indicate the presence of this parasite in areas where these animals roam . The knowledge of T . cruzi infection in these hosts may direct epidemiological measures to risk settings even before the occurrence of human cases . Despite the importance of dogs in the epidemiological scenario of T . cruzi transmission , there are no commercial serological tests to diagnose canine T . cruzi-infection . Several studies employed conventional in-house ELISA to diagnose T . cruzi exposure in dogs , which uses either fractionated T . cruzi lysates or whole-cell epimastigote homogenates as antigenic matrices [10 , 13–15] . This complex antigenic mixture of the variable component is highly sensitive and some drawbacks have been already described , such as difficult standardization , low specificity , and cross-reactivity with Leishmania spp . , and other trypanosomatids [16 , 17] . Therefore , accurate serological diagnostic tests are needed to fill this gap . Multi-epitope recombinant proteins , composed of several antigenically distinct amino acid sequences , have been proposed to improve the diagnostic performance of human chronic CD diagnosis [18–23] . Four multi-epitope proteins , namely IBMP-8 . 1 , IBMP-8 . 2 , IBMP-8 . 3 , and IBMP-8 . 4 , have been expressed by our team and their diagnostic potential has previously been evaluated on several technical platforms to diagnose human CD [18–20] . These antigens presented high levels of sensitivity , specificity , and accuracy for samples from both endemic and non-endemic areas for several geographic regions [19 , 24 , 25] . Cross-reactivity has already been evaluated and only a small number of samples were classified as reagent for various infectious diseases of clinical interest , including leishmaniasis [19 , 26] . The results obtained for human samples are very promising . Therefore , in this study , we aimed to assess the efficiency of these multi-epitope proteins as an antigenic matrix in serological assays for anti-T . cruzi IgG antibody detection in dogs with chronic T . cruzi infection .
This investigation was approved by the Animal Ethical Committee from Gonçalo Moniz Institute , Oswaldo Cruz Foundation , Salvador , Bahia , with the number 002/2017 . The samples were provided by the biorepositories of the Trypanosomatid Biology Laboratory ( Fiocruz , Rio de Janeiro ) , Immunopathology Laboratory of the Nucleus of Research in Biological Sciences ( NUPEB—Federal University of Ouro Preto–UFOP ) , Host-Parasite Interaction and Epidemiology Laboratory ( Fiocruz-BA ) , and Immunoparasitology Laboratory ( Fiocruz-Pernambuco ) . The multi-epitope antigens used in the indirect ELISA were obtained according to Santos et al . [18] . Briefly , the synthetic genes were subcloned into the pET28a vector and expressed in Escherichia coli BL21-Star ( ThermoFisher Scientific ) . Expression was induced with 0 . 5 μM of IPTG ( isopropyl β-D-1-thiogalactopyranoside ) and the soluble proteins purified by both ion exchange and affinity chromatography . Finally , the purified muti-epitope antigens were quantified by Qubit fluorometric assay ( ThermoFisher Scientific ) . Fig 1 illustrates the SDS-PAGE of the antigens after purification . Details about IBMP composition has been seen in previous studies of our group [18 , 19 , 23] . Three convenience canine serum panels from dogs of different origins were used in this investigation ( Fig 2 ) . The first panel ( panel 1 ) was composed of either non-infected ( n = 31 ) or experimentally infected mongrel dogs with the Colombian ( n = 12 ) , Y ( n = 8 ) and Berenice ( n = 10 ) strains of T . cruzi . These samples were obtained from animals born and maintained at UFOP’s Animal Science Center kennel , Ouro Preto , Minas Gerais , Brazil . Serum samples from experimentally infected dogs were collected 2–15 months after T . cruzi inoculation . The second panel ( panel 2 ) , composed of T . cruzi-negative ( n = 119 ) and positive ( n = 16 ) samples , was provided by the biorepository of the Trypanosomatid Biology Laboratory ( Fiocruz-RJ ) or the Host-Parasite Interaction and Epidemiology Laboratory ( Fiocruz-BA ) . These samples were obtained from previous investigations and were collected in several T . cruzi endemic settings from Brazil [27–29] . Combined analysis of panel 1 and panel 2 was denominated as “Merged . ” In addition to T . cruzi-positive and negative sera , 51 samples from animals with unrelated pathogens infection ( panel 3 ) were combined into the sample set to assess cross-reactivity . The unrelated pathogens infection evaluated included anaplasmosis ( n = 6 ) , babesiosis ( n = 17 ) , dirofilariosis ( n = 8 ) , and ehrlichiosis ( n = 13 ) . These samples were from the biorepository of the Immunoparasitology Laboratory ( Fiocruz-PE ) and previously characterized by molecular and serological methods [27] . Additionally , sera from mongrel dogs experimentally infected with Leishmania infantum ( n = 7 ) were provided by the Immunopathology Laboratory ( UFOP ) and assessed herein [30] . All samples were initially re-assayed for IgG antibodies against T . cruzi using two in-house ELISA . In the first test , 100 μl of fractionated lysates of T . cruzi at the epimastigote stage ( 2 . 4 μg/ml ) in carbonate-bicarbonate buffer ( 50 mM , pH 9 . 6 ) was used to sensitize the microplates for 60 min at 37°C . Following washing steps with PBS 0 . 05% Tween 20 ( PBS-T ) , the microplates were blocked with phosphate-buffered saline ( PBS ) supplemented with 2% milk lecithin for 30 min at 37°C . After washing , 100 μl of anti-dog IgG-HRP conjugated ( Sigma , St . Louis , USA ) , loaded at 1:40 . 000 in PBS , were added and the microplates incubated for 30 min at 37°C . Following incubation and washing cycles , 100 μl of TMB substrate ( tetramethyl-benzidine; Sigma , St . Louis , USA ) were added , and the microplates were incubated in the dark at room temperature ( RT ) for 30 min . Then , the enzymatic reactions were stopped by adding 50 μl of 0 . 5 M H2SO4 , and the optical density ( OD ) at 450 nm was read in a MultiskanFC microplate spectrophotometer ( Thermo Scientific , Finland ) . The other test refers to a modified Gold ELISA Chagas kit ( Rem Indústria e Comércio , São Paulo , Brazil ) . In this test , dilutions of anti-dog IgG-HRP conjugate tested were 1:20 , 000 , 1:40 , 000 , and 1:80 , 000 ( Bio-Manguinhos , Rio de Janeiro , Brazil ) . Similarly , T . cruzi-positive and negative serum samples were assayed at dilutions 1:100 , 1:200 , 1:400 , and 1:800 . The best conditions to separate negative and positive samples ( delta median—Δ ) were conjugated antibody diluted at 1:40 , 000 and serum dilution at 1:800 . The cut-off ( CO ) was established by using the mean optical absorbance of negative sera plus three standard deviations . If a sample’s optical density ( OD ) value fell within ± 10% of the CO value , it was considered as an indeterminate result ( or in the grey zone ) . Samples with repeatedly discrepant results between both tests or inconclusive in one of them ( or within the gray zone ) were excluded . These two in-house ELISA were used as reference tests to determine the presence of IgG anti-T . cruzi antibodies in the investigated samples . Each sample was given an identifier code in the laboratory to ensure a blinded analysis . The optimal dilutions of the antigen coating , as well as the dilutions of the antibody-enzyme conjugate ( HRP ) and serum concentrations , were determined by cross-titration . The selected conditions were established considering the largest difference in the average optical density ( OD ) value between positive and negative samples . The conditions were considered satisfactory when negative samples’ OD averaged below or around 0 . 25 and positive samples above or next to 1 . 00 . The selected conditions for each chimeric antigen were established considering the highest difference between the median OD for positive and negative T . cruzi samples ( delta median—Δ ) . Flat bottom , high-binding , transparent “Maxisorp” 96-well microplates ( Nunc , Roskilde , Denmark ) were coated with IBMP antigens ( 12 . 5 ng , 25 ng , and 50 ng ) in carbonate-bicarbonate buffer ( 50 mM , pH 9 . 6 ) . Following the blocking step with Well Champion reagent ( Kem-En-Tec , Taastrup , Denmark ) , 100 μl of a serial dilution of each serum sample ( 1:100 and 1:200 ) diluted in phosphate-buffered saline ( pH 7 . 4 ) was added to the selected well and the microplate incubated at 37 °C for 60 min . After washing with phosphate-buffered saline-0 . 05% Tween 20 , 100 μL of HRP-conjugated goat anti-dog IgG ( Fiocruz , Rio de Janeiro , Brazil ) , diluted at 1:20 , 000 , 1:40 , 000 , and 1:80 , 000 ratios in phosphate-buffered saline , were added to the wells and the microplate incubated at 37°C for 30 min . After another washing cycle , 100 μl of TMB substrate ( tetramethyl-benzidine; Kem-En-Tec , Taastrup , Denmark ) were added to each well and the microplates incubated in the dark at RT for 10 min . The reaction was interrupted by adding 50 μl of 0 . 3 M H2SO4 to each well . The OD was measured using a 450nm filter ( SPECTRAmax 340PC , USA ) . CO values were established under ROC curve analysis . The results were normalized as a reactivity index ( RI ) that denotes the ratio between the OD of the samples and the CO . Samples that resulted in RI > 1 . 0 were considered positive . If a sample’s RI value was within ± 10% of 1 . 0 , it was classified as inconclusive ( or in the grey zone ) . Each sample was given an identifier code in the laboratory to ensure a blinded analysis . Data were analyzed using a scatter plot graphing software ( GraphPad Prism version 7 , San Diego , CA , USA ) . Continuous variables were presented as geometric mean ± standard deviation ( SD ) . The Shapiro-Wilk test was used to test data normality . When the assumed homogeneity was confirmed , Student’s t-test was used . If not , Wilcoxon’s signed-rank test was employed . All analyses were two-tailed , and p values under 5% were considered significant ( p < 0 . 05 ) . Areas under the ROC curve ( AUC ) were calculated to assess the global accuracy for each IBMP antigen , which can be classified as outstanding ( 1 . 0 ) , elevated ( 0 . 82–0 . 99 ) , moderate ( 0 . 62–0 . 81 ) , or low ( 0 . 51–0 . 61 ) [31] . IBMP-ELISA performance parameters were determined using a dichotomous approach and compared regarding sensitivity ( Se ) , specificity ( Sp ) , and accuracy ( Ac ) . A 95% confidence interval ( 95% CI ) was calculated to address precision of the proportion estimates . The agreement strength between the reference standard tests and IBMP-ELISA was established by Cohen’s kappa ( κ ) analysis , which was interpreted as follows: 1 . 0 ≤ κ ≥ 0 . 81 ( almost perfect agreement ) , 0 . 80 ≤ κ ≥ 0 . 61 ( substantial agreement ) , 0 . 60 ≤ κ ≥ 0 . 41 ( moderate agreement ) , 0 . 40 ≤ κ ≥ 0 . 21 ( fair agreement ) , 0 . 20 ≤ κ ≥ 0 ( slight agreement ) , and k = 0 ( poor agreement ) [32] . A checklist ( S1 Table ) and flowchart ( Fig 2 ) are provided according to the Standards for the Reporting of Diagnostic accuracy studies ( STARD ) guidelines .
The optimal dilutions of sera , antigens and antibody-enzyme conjugate were assessed by checkerboard titration . The best condition was chosen by considering the higher Δ median between T . cruzi-positive and negative samples . The pre-established criteria ( OD < 0 . 25 for negative samples and OD > 1 . 00 for T . cruzi-positive samples ) conditions were classified as satisfactory with the antibody-enzyme conjugate at a dilution of 1:20 . 000 for IBMP-8 . 3 and 1:40 . 000 for IBMP-8 . 1 , IBMP-8 . 2 , and IBMP-8 . 4 . With respect to serum dilution , all tests presented higher Δ median when diluted at 1:100 compared to 1:200 ) . Conversely , the best quantity of antigen to sensitize each well varied from 25 ng for IBMP-8 . 1 , IBMP-8 . 2 , and IBMP-8 . 4 to 50 ng for the IBMP-8 . 3 . The phase I study was performed using two distinct serological panels . Panel 1 was composed of sera from dogs experimentally infected with three known strains of T . cruzi , whereas panel 2 was formed by sera from dogs naturally infected with unknown strains of T . cruzi . The merged analysis was also performed considering the samples from both panels 1 and 2 ( Fig 3; individual data points are available in the S2 Table ) . Based on AUC values , all IBMP chimeric antigens were classified with either high or outstanding diagnostic potential , regardless of serum panel assayed . However , AUC values for IBMP-8 . 3 and IBMP-8 . 4 were statistically higher than those for IBMP-8 . 1 and IBMP-8 . 2 . For T . cruzi-positive samples , IBMP-8 . 3 produced the highest RI value for merged analysis . No significative difference was observed between IBMP-8 . 3 and IBMP-8 . 4 , considering overlap of 95% CI values . The lowest RI value was seen for IBMP-8 . 1 . However , no differences were shown between IBMP-8 . 1 and IBMP-8 . 3 or IBMP-8 . 2 considering only panel 2 . With respect to T . cruzi-negative samples , RI values were below 0 . 45 for all four chimeric antigens in all investigated panels . IBMP-8 . 3 presented a sensitivity of 100% for all panels , followed by IBMP-8 . 4 ( ranging from 96 . 7% to 100% ) , IBMP-8 . 2 ( ranging from 73 . 3% to 87 . 5% ) , and IBMP-8 . 1 ( ranging from 50% to 100% ) . Considering the merged panel , no statistical differences in sensitivity were observed for IBMP-8 . 3 and IBMP-8 . 4 . Conversely , IBMP-8 . 1 showed the lowest sensitivity , mainly for T . cruzi-positive samples from panel 1 ( Sen 50% ) . False-negative results , produced by this antigen , were due to six dogs infected with the Colombian strain , six with the Y strain , and three with Berenice strain , in which the serum collection occurred two months ( nine animals: four infected with Colombian strain , two infected with Y strain , and three infected with Berenice strain ) , six months ( three animals: one and two infected with Colombian and Y strain , respectively ) , and 15 months ( three animals: one infected with Colombian strain and two infected with Y strain ) postinfection . IBMP-8 . 2 also produced a high number of false-negative results , which was due to eight Colombian and one Y strain infected dogs , in which the serum collection occurred two months ( three animals infected with Colombian strain ) , six months ( two animals infected with Colombian strain ) , and 15 months ( three animals: two infected with Colombian strain and one with Y strain ) postinfection . The performance of IBMP-ELISA evaluated by accuracy showed values of 99% for IBMP-8 . 4 , 98% for IBMP-8 . 3 , and 94% for IBMP-8 . 2 , without significant difference among them . Conversely , IBMP-8 . 1 was 85 . 6% accurate , with statistical difference compared with the other antigens ( lack of 95% CI values overlapped ) . By adopting an inconclusive zone of 1 . 0 ± 10% , a small number of samples fell inside the grey zone using IBMP-8 . 2 , IBMP-8 . 3 , and IBMP-8 . 4 . However , the RI values of six ( 2% ) T . cruzi-negative samples fell in the grey zone when assayed with IBMP-8 . 1 . All T . cruzi-negative samples fell inside the conclusive space when tested using IBMP-8 . 2 and IBMP-8 . 4 . Two T . cruzi-negative samples presented inconclusive result when evaluated with IBMP-8 . 3 . With respect to T . cruzi-positive samples , one sample was inconclusive when assayed with IBMP-8 . 3 or IBMP-8 . 4; and two samples tested by IBMP-8 . 2 . Overall analysis showed that 0 . 33% of the samples assayed using IBMP-8 . 4 , 1% using IBMP-8 . 2 or IBMP-8 . 3 , and 2% using IBMP-8 . 1 presented RI values falling within the inconclusive result threshold . In order to evaluate the heterogeneity of recognition of IBMP chimeric antigens by IgG anti-T . cruzi specific antibodies due to the expected genetic variability of parasite strains , RI and sensitivity values were compared using samples from dogs infected with T . cruzi Berenice ( n = 10 ) , Colombian ( n = 12 ) , and Y ( n = 8 ) strains . Consistent with the results described above , the highest RI values were found when T . cruzi-positive samples were assayed using IBMP-8 . 3 , followed by IBMP-8 . 4 . However , RI values for T . cruzi Berenice and Y strains were statistically higher for IBMP-8 . 3 compared to IBMP-8 . 4 ( Fig 4; individual data points are available in the S2 Table ) . Contrarily , no statistical difference was found when T . cruzi Colombian strain was assayed with IBMP-8 . 3 and IBMP-8 . 4 . The lowest RI values were found for IBMP-8 . 1 . Only two out of eight samples were correctly classified as T . cruzi-positive , producing an RI value of 0 . 9 . Similar RI value was observed when T . cruzi Colombian strain samples were tested using IBMP-8 . 2 antigen . As shown in Fig 4 , IBMP-8 . 1 showed the lowest sensitivity value ( 25% ) . IBMP-8 . 2 also showed low sensitivity ( 41 . 7% ) . Higher sensitivity values were observed for IBMP-8 . 3 ( 100% ) and IBMP-8 . 4 ( 87 . 5–100% ) . IBMP-ELISA tests were performed to evaluate the antigenic cross-reactivity against antibodies of unrelated diseases ( RI ≥ 1 . 0 ) using a panel of 51 serum samples . As shown in Fig 5 , no cross-reaction was observed when serum samples were assayed using IBMP-8 . 2 and IBMP-8 . 4 ( individual RI values are given in the S3 Table ) . The incidence of cross-reactivity using IBMP-8 . 3 was negligible; only one L . infantum-positive sample produced a false-positive result . Conversely , 13 . 5% of the samples assayed using IBMP-8 . 1 presented positive result . At least one sample from each unrelated pathogen cross-reacted with this chimeric antigen , producing a cross-reaction incidence of 16 . 7% for anaplasmosis , 17 . 7% for babesiosis , 12 . 5% for dirofilariosis , and 15 . 4% for ehrlichiosis . No cross-reaction was observed when L . infantum-seropositive samples were assayed using IBMP-8 . 1 .
In the present study , we assessed the diagnostic performance of four recombinant chimeric antigens for detection of specific IgG anti-T . cruzi antibodies in sera from T . cruzi-positive dogs . The main objectives for surveillance of T . cruzi infection in domestic animals are to identify mammalian species that can act as amplifiers of parasite populations and to determine the mammalian species that can act as bioindicators ( sentinels ) of T . cruzi transmission risk to humans . It is known that , similar to what occurs in Leishmania spp . infection , in areas that present a high prevalence of T . cruzi infection in wild mammalian and triatomine hosts , domestic and peridomestic ( synanthropic ) mammalian species are exposed to infection , a scenario that favors establishment of human CD in that same area [7] . The antigens assayed in this work exhibited a high diagnostic efficiency . Indeed , the AUC values were higher than 90% for all antigens , indicating an optimal discriminative power between T . cruzi-positive and negative canine sera . IBMP-8 . 3 and IBMP-8 . 4 are of particular interest since they presented AUC values of 99 . 7% and 100% , respectively . These data are similar to previous results obtained by our group when these T . cruzi antigens were assessed to diagnose CD in humans [18–20] . Currently , human chronic CD diagnosis is troublesome due to the lack of a gold-standard serological test and the only gold standard tests based on T . cruzi DNA detection are solely applicable in the brief initial stage of the infection ( acute phase ) . However , there is currently a significant difference in performance among the commercial tests due to the high genetic variability of the parasite and the employed antigenic matrices used to capture specific antibodies [18 , 33] . Hence , the World Health Organization recommends the concomitant use of two antigenically distinct commercial tests to diagnose T . cruzi infection in humans [34] . The diagnosis of the T . cruzi infection in dogs is even more difficult , owing to the absence of validated tests . Most studies usually employ either fractionated T . cruzi lysates or whole-cell epimastigote homogenates as antigenic matrices [10 , 13–15] , which can lead to difficulties in standardizing , low specificity , and cross-reactivity with antibodies against Leishmania spp . , amongst other trypanosomatids [16 , 17] . So , a strategy to address these limitations can be proposed by adopting synthetic recombinant chimeric antigens , composed of conserved amino acid sequences of several antigenic T . cruzi proteins [21 , 23 , 35 , 36] . To the best of our knowledge , this is the first study using recombinant chimeric proteins to diagnose T . cruzi infection in experimentally infected animals with known strains . The diagnostic sensitivity was higher for IBMP-8 . 3 and IBMP-8 . 4 compared to the other chimeric proteins . Indeed , IBMP-8 . 3 correctly diagnosed all T . cruzi positive samples , regardless of the panel or strain , whereas IBMP-8 . 4 misidentified as negative only one positive sample from a dog infected with the Y strain . Conversely , IBMP-8 . 1 and IBMP-8 . 2 presented lower sensitivity values , probably due to their amino acid composition and the short period between inoculation and sampling ( acute phase ) . We observed that the majority of false-negative results produced by IBMP-8 . 1 and IBMP-8 . 2 was associated with samples from panel 1 . This panel was composed of dogs submitted to well-established infection protocols in which known strains were used . The sera sampling occurred two , six or 15 months postinfection . We observed that almost 50% of the false-negative results occurred in animals whose samples were collected two months postinfection . Thus , the quantity of specific IgG anti-T . cruzi antibodies did not appear sufficient to be detected by the IBMP-8 . 1 and IBMP-8 . 2 immunoassays performed here . Furthermore , the limited repertoire of antigenic epitopes in IBMP-8 . 1 and IBMP-8 . 2 compared to IBMP-8 . 3 and IBMP-8 . 4 could not be wide enough to identify the specific antibodies . Considering the cross-reactivity assessment with sera from dogs carrying unrelated pathogens , the small number of samples that cross-reacted was statistically insignificant for the assays with IBMP-8 . 2 , IBMP-8 . 3 , and IBMP-8 . 4 . IBMP-8 . 3 cross-reacted with one Leishmania infantum-positive serum , which was negative for other chimeric antigens . Conversely , IBMP-8 . 1 recognized at least one serum for anaplasmosis , babesiosis , dirofilariosis and ehrlichiosis as T . cruzi-positive samples , producing an overall cross-reaction rate of 13 . 7% . This was not expected owing to the low similarity of IBMP antigen sequences of T . cruzi to those deposited in the NCBI’s Genbank for other species . Furthermore , cross-reacting samples also presented a low RI value . Inconclusive results using this panel were statistically insignificant , except to IBMP-8 . 1 . These findings are similar to previous results obtained when our group assessed the cross-reactivity in human samples for pathogens of medical interest , such as dengue , B and C hepatitis , HIV , HTLV , visceral and cutaneous leishmaniasis , leptospirosis , rubella , measles , schistosomiasis , and syphilis using both ELISA and liquid microarray [19 , 20 , 26] . The low number of cross-reaction suggests that IBMP antigens , specially IBMP-8 . 3 , can be safely used to diagnose canine T . cruzi infection in co-endemicity areas with other infectious parasites . The main limitation of the study was the lack of a validated standard test to pre-classify the sera to be used to evaluate the efficiency of the antigens . To overcome this limitation we employed two in-house ELISAs as reference tests . So , the diagnostic performance of the present IBMP chimeric antigens could be biased due to shortcomings in the accuracy of the reference test . Another limitation was the number of samples with unrelated pathogens . Despite these restrictions , we conclude that IBMP-8 . 3 and IBMP-8 . 4 can be used for anti-T . cruzi IgG antibodies detection in dogs . These antigens could be potentially employed in a test to evaluate parasite’s transmission cycle of T . cruzi in endemic settings and for veterinary purposes . | Despite dogs being considered T . cruzi’s most important domestic sentinel/reservoir and also suffering from the outcomes of the infection , there has never been a commercially available test to diagnose T . cruzi infection in dogs . As such , our group’s objective was to develop a state-of-the-art serological diagnostic test utilizing four chimeric antigens ( IBMP-8 . 1 , 8 . 2 , 8 . 3 , and 8 . 4 ) in an ELISA platform to accurately identify canine anti-T . cruzi IgG antibodies . The IBMP-ELISA assays were optimized , evaluated for cross-reactivity towards multiple canine parasite’s , including Leishmania spp . , its diagnostic potential was validated and the test’s performance was determined using double entry tables . The IBMP-8 . 3 antigen demonstrated 100% sensitivity , followed by IBMP-8 . 4 , whereas the highest specificities were achieved with IBMP-8 . 2 ( 100% ) and IBMP-8 . 4 ( 100% ) . Therefore , we’ve concluded that the serodiagnosis through anti-T . cruzi IgG detection in dogs , utilizing chimeric antigenic matrices in immunoassays is a promising tool for veterinary diagnosis and epidemiological surveillance . Furthermore , the use of chimeric antigens efficiently addressed common hurdles related to T . cruzi serodiagnosis , especially regarding efficiency variation in response to different strains and cross-reactivity . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"immunology",
"vertebrates",
"parasitic",
"diseases",
"dogs",
"mammals",
"parasitic",
"protozoans",
"animals",
"protozoans",
"veterinary",
"diagnostics",
... | 2019 | Performance of recombinant chimeric proteins in the serological diagnosis of Trypanosoma cruzi infection in dogs |
The influenza virus is an important human pathogen , with a rapid rate of evolution in the human population . The rate of homologous recombination within genes of influenza is essentially zero . As such , where two alleles within the same gene are in linkage disequilibrium , interference between alleles will occur , whereby selection acting upon one allele has an influence upon the frequency of the other . We here measured the relative importance of selection and interference effects upon the evolution of influenza . We considered time-resolved allele frequency data from the global evolutionary history of the haemagglutinin gene of human influenza A/H3N2 , conducting an in-depth analysis of sequences collected since 1996 . Using a model that accounts for selection-caused interference between alleles in linkage disequilibrium , we estimated the inherent selective benefit of individual polymorphisms in the viral population . These inherent selection coefficients were in turn used to calculate the total selective effect of interference acting upon each polymorphism , considering the effect of the initial background upon which a mutation arose , and the subsequent effect of interference from other alleles that were under selection . Viewing events in retrospect , we estimated the influence of each of these components in determining whether a mutant allele eventually fixed or died in the global viral population . Our inherent selection coefficients , when combined across different regions of the protein , were consistent with previous measurements of dN/dS for the same system . Alleles going on to fix in the global population tended to be under more positive selection , to arise on more beneficial backgrounds , and to avoid strong negative interference from other alleles under selection . However , on average , the fate of a polymorphism was determined more by the combined influence of interference effects than by its inherent selection coefficient .
The influenza A virus is an important human pathogen , infecting between 5 and 15% of the global population each year [1] . Infection by influenza leads to strain-specific immunity in the human population , driving a rapid process of adaptation in the virus [2] , particularly in the viral surface proteins haemagglutinin ( HA ) and neuraminidase ( NA ) [3] . Influenza evolves according to two distinct processes . Firstly , variation within gene sequences arises through a high mutation rate [4] , the resulting mutations being subject to selection pressure . For HA and NA , the host immune system is a key driver behind selection [5]–[8] , but selection can also arise from interactions with other viral genes [9] , stability of the protein structure [10] , adaptation to new species [11] , drug resistance [12] , and , in HA , glycosylation [13] . In a second process , known as reassortment , genes from viruses co-infecting a host combine to produce novel viral strains [14] . Many previous studies have measured the importance of selection in the evolution of influenza . Constructing a dendrogram from viral sequences , and evaluating occurrences of synonymous and non-synonymous substitutions , the ratio dN/dS has been used to identify positive selection , both at given loci , and in the HA gene as a whole [15]–[17] . Other attempts to measure selection have looked for changes in frequencies , or in the distribution of frequencies , of nucleotides or amino acids , or rates of substitution between them , in individual loci , or gene-wide [18]–[22] . Each of these methods has produced estimates of selection based upon sets of multiple substitutionary events , either within single loci , or across a set of loci within a gene , rather than by considering single polymorphisms one by one . Homologous recombination in influenza is extremely rare , if not entirely non-existent [23] , [24] . As such , within the viral population , interference effects , by which selection on one allele affects the frequency of another , are likely to occur [25]–[28] . On the basis that mutations observed at antigenic sites are more likely to be fixed than those at non-antigenic sites , it has been argued that the effects of hitchhiking in influenza are small [29] . However , evidence supporting the existence of clonal interference has been identified from the timings of fixation events in the HA gene , and from the inferred number of co-existent strongly beneficial mutations [30] . Here , we consider the evolution of the HA gene of the human influenza A/H3N2 strain since its emergence in 1968 . H3N2 evolves more rapidly than H1N1 [3] , having acquired multiple changes in amino acids over time . While reassortment between genes has occurred in this strain [31] , the HA sequences represent a consistent strand of evolution , described by a large number of sequences [32] . We converted sequences of the HA gene into time-resolved measurements of allele frequencies at polymorphic loci , using these data to infer the strength of selection acting upon each allele . Inference of selection coefficients was carried out using adapted forms of two methods , described in a previous publication [33] , and illustrated in Figure 1 . Collected frequencies of an allele , from its first observation until its death or fixation , are referred to as the allele trajectory . Firstly , an “unlinked” method , representing a minimal model for inferring selection , was applied . The selection acting upon an allele was measured in terms of a selection coefficient , equal to the difference in Malthusian fitness between it , and the wild-type allele at the same locus . For each trajectory , a single selection coefficient , , was learnt , along with a single frequency , , for some time-point within the trajectory ( see Methods ) , so as to maximise the fit between the model , and the observed trajectory frequencies . Secondly , a “linked” model was applied , which used observed two-locus haplotype frequencies to account for interference effects between trajectories . In the linked model , for each trajectory , two selection coefficients were considered . The inherent selection coefficient , , again denotes the inherent advantage or disadvantage conveyed by the allele , relative to the wild-type allele , to individuals which possess it . The effective selection coefficient , , by contrast , denotes the mean total advantage or disadvantage experienced by individuals with that allele , when selection effects acting upon alleles at all polymorphic loci are considered . This effective selection coefficient is time-dependent , and is written , where is the time-point of the trajectory . Given a set of selection coefficients for all trajectories , , our method calculates values of for each trajectory . Under the model , allele frequencies at each locus change deterministically according to these effective selection coefficients . Selection coefficients , , and allele frequencies , , were optimised to give the maximum likelihood fit between the model , and the observed allele frequencies . Our “linked” model represents a minimal model of selection accounting for interference effects; while it is more complex than the unlinked model , it has no additional complexity , in the sense of additional parameters to be learnt . Rather , the additional component of the model derives from the underlying sequence data . Two-locus haplotype frequencies are measured at each time-point . Next , if some allele is under selection , and is in linkage disequilibrium with another allele , selection acting on changes the frequency of the allele in accordance with that linkage disequilibrium . Due to its minimal nature , our linked model has some limitations . For example , the inherent selection coefficients , , do not change with time . Polymorphic alleles were assumed to interact according to a model of additive fitness . Further , to simplify the calculation , synonymous trajectories were assumed to have zero inherent selection , moving only via linkage disequilibrium with non-synonymous trajectories . Limitations of the model are discussed later in the text . Application of our model gave an insight into the processes driving the evolution of influenza . Recent theoretical studies of the evolution of asexual populations have underlined the importance of the background upon which polymorphisms arise , and of interference from other polymorphisms under selection [34] , [35] . Experimental work has shown that , in a non-recombinant population , the background upon which a mutation lands can have a significant impact on its eventual fate [36] . On the basis of our inferences , we measured the importance of inherent selection , the initial genetic background , and subsequent interference effects , for the eventual fate of a polymorphism . Our results suggest that background effects , subsequent interference , and inherent selection are of similar importance for determining the fate of polymorphisms in the HA gene of human influenza A/H3N2 . The combined effect of interference outweighs that of inherent selection .
A total of 3327 complete sequences of the HA gene of human influenza A/H3N2 were obtained from the NCBI influenza virus resource [32] . Trajectories reaching a frequency of at least 2 . 5% were included in the analysis ( for discussion of this choice see Supporting Information ) , giving a total of 1638 trajectories ( 655 non-synonymous , 983 synonymous ) , spanning 256 time-points . In the unlinked method , a selection coefficient and a single allele frequency were learnt for each trajectory , so as to obtain the maximum likelihood fit between the model , and the observed allele frequencies . In the linked method , synonymous trajectories were assumed to evolve under zero inherent selection , so for these , only an allele frequency was inferred . In the application of the linked method , pairwise interactions between all simultaneously polymorphic trajectories were accounted for; a mean of 4939 such pairs existed at any one given time-point . One limitation in the inference arose from a lack of available sequences at early time points ( see Supporting Information ) . As a result only polymorphisms which arose from 1996 onwards were included in the final analysis ( spanning 84 time-points ) . The statistics that follow encompass 622 trajectories , of which 238 were non-synonymous , representing polymorphisms across 442 nucleotide positions ( at 339 amino acid residues ) in the HA gene . For these trajectories each observed allele frequency was calculated from a sample of mean size 202 sequences ( range 57 to 449 ) . Including the effects of linkage disequilibrium between polymorphisms allowed the linked model to capture much of the complex behaviour of the observed trajectories . Relative to the unlinked approach , the model gave a substantially better fit to the observed trajectories ( Figure 2 gives one example ) . In the unlinked model , the inference of selection coefficients for synonymous trajectories required , for the trajectories from 1996 onwards , an additional 384 parameters to be learnt in the optimisation . Nevertheless , the unlinked model gave a substantially worse fit to the data . Measured across more than 7000 observed frequencies , the mean absolute difference between inferred and observed allele frequencies was larger , at 0 . 09 , for the unlinked model , than the equivalent value of 0 . 05 from the linked model . Where linkage disequilibrium is ignored , inferred frequencies are restricted to change according to a model of constant selection , giving a monotonically increasing or decreasing trajectory . Many trajectories encompassed both rising and falling allele frequencies , a phenomenon which , in a deterministic framework , requires a model of variable selection . Plots of all trajectories , together with model fits , are given in Supporting Information . We now discuss results obtained using the linked method . Our model suggested interference effects to be of substantial importance in determining the eventual fixation or death of mutations in the influenza virus . For each trajectory , a measure , of the total selection acting upon the mutant allele ( including both inherent effects and those arising from linkage disequilibrium ) was calculated , as the mean value over time of the effective selection coefficient: ( 1 ) where is the length of trajectory . Considering sets of polymorphic alleles at given time-points , substantial differences between and were seen ( see for example Figure 3A ) . Both clonal interference and genetic hitchhiking were evident , with mutations under strong positive inherent selection , but negative interference , dying out , and mutations with lesser inherent selection , but experiencing positive interference , reaching fixation . Both in the examples shown , and across all trajectories ( Figure 3B ) , the statistic , inferred after the event , discriminated strongly between mutations that fixed and those which died out . Derivative components of selection , associated with interference effects , were calculated for each of the trajectories . Firstly , we derived a measure of the mean effect , over time , of interference acting upon each trajectory . We note that the mean effective selection coefficient , , calculated above , combines the inherent selection acting on a trajectory , , plus interference effects from other trajectories . We calculated a measure , , of the mean interference acting on a trajectory by taking the difference of the mean effective and inherent selective effects: ( 2 ) This interference term can be divided into two parts , representing the effects of interference in the initial and later parts of the history of a polymorphism . Considering these in turn , any given mutation observed in the population is first seen in the context of a specific sequence background . Within the global viral population at any one time , non-synonymous polymorphisms at frequencies considered by our method ( reaching at least ) exist at multiple loci . The sequence ( s ) upon which a mutation arises contain specific alleles at these loci , this background conferring an initial fitness effect . We estimated this background fitness for a trajectory , denoted , as the mean interference over the period before the trajectory reached a frequency of 2 . 5%: ( 3 ) where denotes the first time at which the allele of trajectory was observed at a frequency of 2 . 5% or greater . Subsequently , across the lifetime of a trajectory , its evolution is affected by the arrival of new mutations under selection . Changes also occur in the effect of its background; as the mean fitness of the population increases , any given background becomes relatively less beneficial . We measured the combined effect of these factors , terming it the post-emergence interference , . This statistic was calculated by subtracting the initial background effect from the total interference: ( 4 ) We note that each of the above components of selection are derived from the inferred selection coefficients , rather than being learnt independently from the data . Across trajectories , the magnitude of the effect of interference was generally greater than that of inherent selection . Measured for non-synonymous trajectories , the mean absolute effect of interference acting on a polymorphism at a given time was larger , at , than the mean absolute inherent selection of ( selection is described throughout in units of ) . Simple characterisation of interference in terms of a few large effects was difficult; at any given time , a mean of 15 other trajectories made an absolute contribution of more than to the effective selection coefficient of a polymorphism , from a total of 29 trajectories contributing more than . To examine selective effects in more detail , we calculated distributions of each statistic for trajectories that eventually fixed , and for those that died out ( Figure 4 shows data for non-synonymous trajectories ) . In each case , a bias can be seen , trajectories proceeding to fixation having generally more positive coefficients of selection and interference . Different components of selection biased the fates of trajectories in different ways . For example , the initial background for observed trajectories was generally positive , being more so for fixing trajectories , while the mean post-emergence interference was predominantly negative . Fixing mutations tended to arise in good backgrounds , and were subsequently fortunate in avoiding strong negative interference . In order to quantify the importance of each of the components of selection for the fate of polymorphisms , we measured the accuracy with which each component , considered in retrospect , separated trajectories which reached fixation from those that died out . Across non-synonymous trajectories , the inherent selection coefficient showed some ability to perform this task , with a calculated accuracy of 0 . 70 ( Figure 5 ) . Both the initial background fitness , and the post-emergence interference , performed similarly well , with measured accuracies of 0 . 64 and 0 . 70 . When interference effects were combined , the accuracy achieved was greater than that for inherent selection , returning a value of 0 . 82 . Across synonymous trajectories , the inherent selection coefficient was set by default to zero , such that the effect of any inherent selection could not be measured . However , for these trajectories , the accuracies of the initial background and post-emergence interference were again roughly equivalent , at 0 . 75 and 0 . 71 respectively . The evolution of synonymous trajectories was well-explained purely in terms of interference from non-synonymous polymorphisms under selection; absolute differences between observed and inferred frequencies were no larger than those for non-synonymous trajectories . Measuring the relative strengths of selection acting upon different parts of the HA protein , we obtained results equivalent to those previously obtained from measurements of dN/dS , the ratio between rates of non-synonymous and synonymous mutations . We derived a region-wide measure of selection by calculating the mean of the inherent selection coefficients inferred for non-synonymous trajectories at loci in that region . This measure was then calculated for trajectories in the HA1 and HA2 segments of the protein , and for both epitope and non-epitope loci . Substitutions in the influenza phylogenetic tree can be divided into those occurring on trunk and side branches [37]; to parallel this , we calculated separate measurements for trajectories which fixed in the global population ( c . f . trunk substitutions ) , and for those which died out ( c . f . side-branch substitutions ) . A previous calculation performed on substitutions in a tree of sequences from the human influenza A/H3N2 strain between 1994 and 2005 found dN/dS ratios fitting the order epitope HA1HA2 , with ratios in each case higher for substitutions in trunk branches than in non-trunk branches [17] . Our measurement replicated this finding across corresponding sets of trajectories ( Table 1 ) . The levels of significance underlying our results were calculated by applying a Kolmogorov-Smirnov test to to the inferred distributions of selection coefficients; differences in inferred selection coefficients between epitope and non-epitope trajectories , and between the subclasses of fixed epitope and fixed non-epitope trajectories were each significant ( p = 0 . 014 and p = 0 . 021 respectively ) . A significant difference was seen between selection coefficients at epitope and HA2 trajectories ( p = 0 . 030 ) , though differences between trajectories in HA1 and HA2 were not significant at the 95% level . In our calculation , epitope loci were identified according to the alignment of Wolf et al . [17] . The difference between epitope and non-epitope loci was also seen via measurements of fitness flux , a statistic quantifying the total amount of adaptation at a locus [38] . Across trajectories reaching fixation or death , the total fitness flux for a trajectory is given simply by if the trajectory fixed , and if the trajectory died . On the basis of our inferred selection coefficients , the total fitness flux across trajectories at epitope loci was approximately 8 . 6 times larger than the flux across trajectories at non-epitope loci . While both beneficial and deleterious mutations were observed in all parts of the gene , this result implies that events at epitope loci are of substantial importance for the adaptation of the virus . Regions in which beneficial mutations were identified correlated with those under lesser evolutionary constraint . The extent of evolutionary constraint was measured by examining the number of non-synonymous polymorphisms observed across different sets of multiple loci . Supposing a constant mutation rate across the gene , and lack of any constraint , two regions each comprising a similar proportion of the gene would be expected to harbour a similar proportion of non-synonymous mutations . In regions with an increased constraint on mutations , as might arise from requirements of protein function or structural stability , a greater number of these mutations would be strongly deleterious , leading to a relative under-representation of non-synonymous trajectories from that region . In our data , the HA2 region , and , to a lesser extent , the non-epitope region of HA1 , were each underrepresented for observed polymorphisms and trajectories , while the HA1 region as a whole , and in particular , epitope loci , were overrepresented . This pattern was exaggerated for trajectories identified to be under positive selection ( ) , and further so for trajectories with ( Table 2 ) , suggesting that sites experiencing greater numbers of beneficial mutations , were also sites of lesser evolutionary constraint . This result corresponds to the HA protein structure , in which epitope residues tend to be in regions more exposed regions of the protein [39] .
As noted above , our linked model represents a minimal approach for inferring selection while accounting for interference effects . While our model provides strong evidence of the importance of interference effects , it does not account for the full biological complexity of influenza , making a number of assumptions and approximations . Firstly , synonymous mutations were assumed to have zero inherent selection coefficient , synonymous trajectories changing their frequency purely by the effect of linkage disequilibrium with non-synonymous trajectories . Such an assumption is not unique to our method , reflecting similar assumptions made , for instance , in the measurement of dN/dS . However , the identification in HA of some selection for RNA secondary structure [41] and of long-term changes in codon preference [19] each suggest the possibility that some synonymous changes are non-neutral . This shortcoming in our method is unlikely to affect our central result , of the importance of interference effects . While individual synonymous mutations may be under significant selection , the magnitude of selection , considered across groups of trajectories , is likely to be substantially lower for synonymous than non-synonymous mutations . As such , any errors in our statistics of selection arising from this assumption are likely to be small when considered across the set of all trajectories . Secondly , in our use of a deterministic model for the evolution of trajectories , we do not take into account genetic drift , a process , independent of selection , in which the transmission of alleles from one generation to the next introduces noise into the allele frequencies . ( Within our framework , the frequency of an allele evolving under zero selection would remain constant . ) We believe that the effect of this omission is also likely to be small . Considering alleles at intermediate frequencies , we note that genetic drift is a slow process . In influenza it has been estimated that the expected time for an allele to fix under drift alone is of the order of 200 years , far longer than the typical length of a trajectory [29] , implying that frequency changes are driven much more by selection than by drift . As evidenced by antigenic drift , strongly selected alleles arise frequently [6] , and , due to the lack of recombination , are likely to remain in linkage disequilibrium with other alleles at polymorphic loci . Interference arising from these strongly selected alleles would mean that changes in the frequencies of neutral alleles would also be dominated by selection [34] . Considering alleles at lower frequencies , the influence of selection on an allele is overtaken by that of drift below a frequency of 1/N , where is the population size and the magnitude of selection acting on the allele [42] . In recent work , an estimate has been made that in H3N2 influenza this threshold is close to 1% [30] . Across our observed trajectories , one out of 12 observed frequencies were below this threshold . If this threshold estimate is correct , our exclusion of trajectories with minimal frequencies less than 2 . 5% ensures that the majority of our trajectories are influenced more by selection than drift for the majority of their lifetimes . Thirdly , the inferences we make are drawn on the basis of limited data . While we filter trajectories to consider only those for which we had enough data to derive consistent results , studies of simulated systems suggest that a larger sample would lead to more accurate inference . Consideration of lower frequency events may affect our estimate of the relative importance of the initial background and subsequent interference on a polymorphism; measuring at lower frequencies , more deleterious mutations , and mutations on more deleterious backgrounds , would likely be seen . We further note that , while detecting selection , our method does not infer reasons for this selection . The molecular basis for the evolution of influenza is complex , involving multiple host-virus interaction processes , and the biophysical properties of the viral proteins . Each of these processes may result in changes in allele frequencies , and their mean effect can potentially be recaptured via inference . However , assignment of selection to specific physical effects would require substantial further work . Our approach is not well-suited to identifying the magnitude of selection arising from evolutionary constraints , such as protein structural stability; where an allele is under strong purifying selection , by its nature it is not likely to be observed at frequencies high enough for selection to be inferred . Extensions to the linked model could be made to include the possibility of time-dependent inherent selection acting on a given allele , or of non-additive models of interference between trajectories . Either of these would add to the complexity of the model , and would require care in implementation . Time-dependent inherent selection in influenza might arise through the acquisition of strain-specific immunity among the human population to an antigenic change in the virus [43] , [44] . A mutation causing a novel antigenic type might initially be at a selective advantage within the viral population . However , as people encountered the new antigenic type , and acquired immunity , this selective advantage would decrease . Variable selection corresponding to antigenic type has been included in models of influenza evolution by mapping changes in genetic sequence to changes in antigenic type [2] , [45] , [46] . Time-dependent selection could be incorporated into the model presented here either with the incorporation of antigenic data , or simply by allowing inherent selection coefficients to vary over time . However , we note , first of all , that time-dependent selection would not remove the role of interference . Linkage disequilibrium between alleles implies that the evolution of one allele is affected by any non-neutral selection acting on the other , whether or not that selection is time-dependent . Secondly , any adding of parameters to the model requires care to avoid over-fitting; in the limit case , allowing the inherent selection on each allele to vary between each time point would result in a perfect fit to the observations by default . The inherent selection coefficients inferred here are best interpreted as mean values over time . Epistasis ( non-additive selective effects ) between alleles has been detected in the HA and NA genes of influenza by considering patterns of ordered substitution events [47] . Full inclusion of epistasis into a model , however , is difficult , due to the rapid inflation of the number of parameters required . With the addition of a substantial number of parameters , epistatic effects between pairs of mutations could potentially be modelled . However , there is no guarantee that epistasis operates purely on this level; interactions between three , four , or more mutations may also play a role in viral fitness . Our model includes epistatic effects where all but one of the mutations in question are fixed; repeated instances of identical mutations are assigned independent selection coefficients . We note that , under certain circumstances , epistasis may be irrelevant . Where two alleles are seen only in isolation in the population , and never together in combination , epistatic effects between them do not come into play . Beyond this , incorporation of epistasis into the model would be challenging , and was not attempted here , the assumption of additive fitness effects representing a first approximation . Using a minimal model accounting for interference , we have here characterised selection in the HA gene of influenza at the level of individual polymorphisms . Our approach takes into account the fact that the gene evolves as part of a haploid , asexual virus with high mutation rate , and essentially no homologous recombination , so that linkage disequilibrium between polymorphisms has a substantial impact on its evolution . Examining statistics of selection inferred for historic data , our model suggests that the combined effect of interference is more important than the effect of inherent selection for the fixation or death of a polymorphism . While many methods for inferring selection exist , we believe that it is by accounting for the specific genetic properties of systems such as influenza that progress in understanding their evolution will most rapidly be made .
Complete sequences for the HA gene of the subtype H3N2 of human influenza A were obtained from the NCBI influenza virus resource [32] , including all sequences sampled from 1968 until 7th Feb 2011 , a total of 3327 sequences . Sequence alignment of protein sequences was carried out using MUSCLE [48] , corresponding RNA sequences being matched to this alignment using in-house software . Beginning at 29th June 1968 , samples of sequences were taken every 60 days , each sample comprising all sequences recorded within 180 days of the sample point . Sampling in this manner grouped viral sequences taken from all seasons in both hemispheres , accounting for geographical and seasonal differences . Ambiguously dated sequences were assigned to the central day of their specified year or month of sampling . The frequency of a given allele was measured as the fraction of sequences within a sample having that allele at the time of measurement . Trajectories were defined as sets of observations of codon frequencies differing from the wild-type codon at their respective position . The wild-type codon for each position was initially defined according to the consensus codon from the first sequence sample; subsequent to this , a codon which reached fixation was defined as the new wild-type for that position . The identification of the point at which a trajectory has fixed or died is non-trivial . Within a single patient , the high mutation rate and large viral population size of influenza make it likely that every viable single-mutant of an infecting virus will likely exist , such that concepts of fixation and death require a qualified understanding . We regarded a trajectory as having fixed if , in our sample of the global population , we observed it at a frequency of greater than 97 . 5% for eight consecutive time points , and to have died if we observed a frequency of less than 2 . 5% for eight consecutive time points . The wild-type codon at a position was used to classify new mutations as synonymous or non-synonymous . Synonymous polymorphisms were set to have inherent selection coefficient equal to zero , changes in their frequency being inferred to arise through linkage disequilibrium with non-synonymous polymorphisms . In order to reduce the influence of noise from sampling error , and to avoid difficulties in inferring selection from very sparse data , trajectories with very low maximal frequencies ( 2 . 5% ) were excluded from the optimisation ( see Supporting Information ) . In an earlier publication we described two methods for using time-resolved data for inferring the selection acting upon a polymorphic allele [33] . For notation , we describe a set of observations of the frequency of a polymorphic allele collectively as an allele trajectory . We consider a trajectory , describing the evolution of a mutant allele at a two-allele locus , at which wild-type and mutant loci are denoted and respectively ( separate annotation for trajectories and loci , while superfluous here , becomes necessary in the multi-allele case ) . The observed frequency of the mutant allele at locus at time is denoted . We assume that the mutant allele has a constant , trajectory-specific selection coefficient , , defined as the difference in Malthusian fitnesses of the two alleles at the given locus . Our first model , referred to above as the “unlinked” model , assumes that alleles at different loci evolve independently of each other . Using the above notation , the expected frequency of the mutant allele at locus evolves according to the equation ( 5 ) This equation describes a family of trajectories , specified by the allele frequency at time : ( 6 ) In practise , the allele frequency at any time point may be used to specify curves; the central point of the trajectory , denoted , was used in our calculations . The model which best fits the observed trajectory was identified by optimising the selection coefficient and allele frequency ( here and ) using a binomial model to fit inferred frequencies to the observed frequencies . Our second model , referred to above as the “linked” model , is essentially the same as this , but for the constraint that , if two alleles are in linkage disequilibrium , a change in the frequency of one allele , caused by selection , implies a change in the frequency of the other . At each time-point , two-locus haplotype frequencies , denoted for loci and , were observed and used to calculate the effect of selection at one locus on the frequency of another . Given a trajectory representing the evolution of the allele at locus , a time-dependent selection coefficient was calculated . This term describes the selection acting on the mutant allele of the trajectory given its own , or inherent selection coefficient , , and the selection acting as a result of linkage disequilibrium with alleles at other loci , which were also under selection , at the time . Under a model of additive selection , this can be approximated by ( 7 ) where ( 8 ) The values were assumed to remain constant between sampling points . Under the model , these time-dependent selection coefficents were fed into Eq . ( 6 ) , to describe the evolution of the trajectory; again a fit was performed between the inferred and observed frequencies . For simplicity , we have here described the case in which there are two alleles per locus . In the adaptation of our previously published method to influenza data , changes were made to account for the process via which trajectories were sampled , and to allow independent trajectories at the same locus to have differing selection coefficients . Further changes were made to the manner in which the fixation or death of trajectories was handled , to the treatment of very low frequency trajectories and of non-polymorphic observations in the middle and ends of trajectories , to the precise use of effective selection coefficients in inferring trajectories , and to the fitting of inferred to observed trajectory frequencies . Fuller details of the method are given in Supporting information . | Success in life is the product of many factors . Inherent ability often underlies great achievement . But other factors may play their part . The circumstances a child is born into may help or hinder his or her progress . Later events also have their effect; a life may be influenced by a lucky break , or an unforeseen disaster . In this work , we examine the factors underlying success for mutations in the HA gene of human influenza virus A/H3N2 , defining success as the attainment of a high frequency in the global population . We examined the history of the gene from 1968 until 2010 . For each observed mutation , a mathematical model was used to estimate the inherent benefit or disadvantage it conferred to the virus . We calculated the advantageousness or otherwise of the background upon which it arose , and the subsequent effect of interference from other mutations under selection . We found that successful mutations tended to have an advantageous background , and were subsequently fortunate in avoiding negative events throughout their lifetime . Beneficial mutations were more likely to be successful . But a mutation's chances of success were influenced more by circumstances of birth and subsequent events , than by its inherent effect on the virus . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genomic",
"evolution",
"natural",
"selection",
"virology",
"genetics",
"population",
"genetics",
"biology",
"computational",
"biology",
"evolutionary",
"biology",
"microbiology",
"genetics",
"and",
"genomics",
"viral",
"evolution",
"evolutionary",
"theory"
] | 2012 | Components of Selection in the Evolution of the Influenza Virus: Linkage Effects Beat Inherent Selection |
Nutrient homeostasis—the maintenance of relatively constant internal nutrient concentrations in fluctuating external environments—is essential to the survival of most organisms . Transcriptional regulation of plasma membrane transporters by internal nutrient concentrations is typically assumed to be the main mechanism by which homeostasis is achieved . While this mechanism is homeostatic we show that it does not achieve global perfect homeostasis—a condition where internal nutrient concentrations are completely independent of external nutrient concentrations for all external nutrient concentrations . We show that the criterion for global perfect homeostasis is that transporter levels must be inversely proportional to net nutrient flux into the cell and that downregulation of active transporters ( activity-dependent regulation ) is a simple and biologically plausible mechanism that meets this criterion . Activity-dependent transporter regulation creates a trade-off between robustness and efficiency , i . e . , the system's ability to withstand perturbation in external nutrients and the transporter production rate needed to maintain homeostasis . Additionally , we show that a system that utilizes both activity-dependent transporter downregulation and regulation of transporter synthesis by internal nutrient levels can create a system that mitigates the shortcomings of each of the individual mechanisms . This analysis highlights the utility of activity-dependent regulation in achieving homeostasis and calls for a re-examination of the mechanisms of regulation of other homeostatic systems .
Cells must maintain relatively constant internal concentrations of nutrients even though the supply of nutrients from the environment can fluctuate wildly , a process called nutrient homeostasis [1 , 2] . In microorganisms , errors in nutrient homeostasis can have dramatic effects on growth , since low internal nutrient concentrations limit growth , while excessive internal nutrient concentrations can be toxic [3 , 4] . In mammalian cells , nutrient uptake , cell growth , and proliferation are controlled by the overlapping signaling pathways [5 , 6] and defects in nutrient regulation play a role in the pathogenesis of diseases such as cancer and diabetes [1 , 7–9] . Nutrient homeostasis is a major determinant of both organismal and cellular fitness . There are two axes that are important for homeostasis . The first axis is the robustness of homeostasis: the more robust the homeostasis , the smaller the change in internal nutrient concentration for a given change in external nutrient concentration . The limit of robust homeostasis is when the internal nutrient concentration is completely insensitive to the external nutrient concentration , a condition we refer to as 'perfect' homeostasis . The second axis is the range of homeostasis . The wider the range of homeostasis , the large the range of external nutrient concentrations over which the system achieves a given robustness of homeostasis . Global homeostasis occurs when the system is homeostatic regardless of the external nutrient concentration . In this work , we solve for conditions that achieve global perfect homeostasis . There are many examples of biological systems that exhibit homeostasis [10–14] . In general , homeostasis can arise from fine tuning kinetic parameters or from structural properties of the regulating network [1 , 2 , 12 , 14] . Achieving nutrient homeostasis requires cellular circuitry that is able to sense nutrient levels and then regulate uptake and/or usage accordingly . All nutrient homeostatic systems need a plasma membrane transporter that allows passage of the nutrient through the plasma membrane . The majority of nutrient homeostatic systems share a common architecture where the synthesis of this plasma membrane transporter is under the regulation of nutrient levels ( Fig 1 ) [3 , 4 , 11] . This regulation is a negative feedback system such that when nutrient concentrations are low , transporter synthesis is increased and when nutrient synthesis is high transporters , synthesis is decreased [1 , 5 , 6 , 11] . In eukaryotes , this type of regulation has been demonstrated for metal ions [10 , 15–18] , sugars [1 , 2 , 19 , 20] , phosphate [3 , 4 , 21] , and amino acid transport [6 , 22–24] . While the mechanistic details of this design can vary , e . g . regulation of synthesis through transcription [1 , 7–9 , 25] or trafficking [10 , 26] , regulation of transporter synthesis is typically assumed to be the critical factor in nutrient homeostasis . It has been shown that this negative feedback regulation makes nutrient homeostasis more robust and this robustness depends on the sensitivity of the transporter synthesis rate to nutrient levels [1 , 2 , 12–14] . A second and equally widespread motif in homeostatic systems is post-translational downregulation of transporters [28–32] . As with regulation of synthesis , the mechanistic details of this design can vary; e . g . , plasma membrane transporters can be inactivated by modification , sequestrations , or degradation . In yeast , this mode of regulation has been demonstrated for many different homeostatic systems including ion transporters such as zinc , copper , and iron [33–35]; sugar transporters such as glucose and maltose transporters [25 , 36]; a phosphate transporter [37]; and amino acid transporters [25 , 38] . This mechanism is usually considered a stress response to an extreme change in nutrients [39] . Indeed , the response to extreme changes is a homeostatic process . Transporter downregulation has been less well studied than transcriptional regulation of transporter synthesis . In the best-characterized systems , transporter downregulation is mediated by ubiquitination [25 , 40–42] but the theoretical cost and benefits of using transporter downregulation to achieve homeostasis are not understood . Mathematically , regulation by transporter synthesis and transporter downregulation are interconvertible at steady-state; i . e . , adding a term to the transporter synthesis flux ( thin green arrow in Fig 1 ) and adding 1 over that term to the transporter downregulation flux ( thin red arrow in Fig 1 ) yields the same steady state solution . Yet , biological realizations of some mathematical terms are not easily achieved . Hence , these two forms of regulation may not be biologically interconvertible . We determined that if transporter levels are inversely proportional to flux through the plasma membrane transporters , global perfect homeostasis is achieved . Flux sensing is distinct from the internal and external nutrient sensing that is typically considered to regulate nutrient homeostatic systems . Work from the Heinemann lab [43 , 44] has shown that flux sensing is used for the regulation of some intracellular metabolism . We showed that flux sensing based homeostasis can be easily achieved if active transporters are downregulated . We will refer to this as activity-dependent downregulation of transporters . Given that both transporter synthesis and downregulation are regulated in most nutrient homeostatic systems , we sought to determine the potential trade-offs of each architecture [45–48] . The combination of activity-dependent downregulation and regulation through synthesis makes a more efficient system than either mechanism alone .
To define homeostasis we consider a dynamical system that has an output , e . g . the internal nutrient concentration ( Sint ) , which depends on other variables such as external nutrient concentration ( Sext ) and transporter concentration ( T ) . The output of the system achieves global perfect homeostasis with respect to a specific variable if the stationary concentration of the output is invariant to perturbations in this variable . Formally , for the example of nutrient transport , Sint would achieve global perfect homeostasis with respect to Sext if Sint has a steady state such , Sintss , such that ∂Sintss ( Sext , Tss ( Sext ) ) ∂Sext=0 for every value of Sext . Some systems may exhibit perfect homeostasis for a range of Sext values; we refer to this as local homeostasis . Additionally , to quantify the dependence of the SintSS on Sext , we defined a unitless parameter that is the steady state value of Sint for a given Sext normalized by the maximal steady state value of Sint over the range of relevant Sext , r ( Sext ) =Sintss ( Sext , Tss ( Sext ) ) maxSext=[0 , ∞ ) Sintss ( Sext , Tss ( Sext ) ) . We will refer to , r , as the robustness of homeostasis . A system achieves global perfect homeostasis when r = 1 for every Sext . When the system is not perfectly homeostatic , r can be any value between zero and one and this value can change as a function of Sext . Biological systems that approach this limit of perfect homeostasis , i . e . r is close to 1 , are often still considered homeostatic although there is no standard value for r which segments between whether or not a system is considered homeostatic [10] . In this work , we look for the biological circuitries that achieve global perfect homeostasis . Note that homeostasis as we define it is a property of the steady state concentration; during transitions between different steady states , the homeostatic output can transiently change . We first we sought to determine all conditions that could lead to global perfect homeostasis in a general uptake system ( Fig 1 ) . Our system is composed of an external nutrient ( Sext ) , an internal nutrient ( Sint ) , and a plasma membrane transporter ( T ) ( Fig 1 ) . Transporters allow nutrient to pass into the cell through the plasma membrane and can be both synthesized and destroyed . While there are more molecular players , e . g . , mRNA and translocation machinery , this system encapsulates the key biological variables while subsuming the rest of the players into the parameters . We use the following notation convention: Greek symbols denote fluxes ( with units of concentration/time ) , the symbol k denotes rate constants ( with units of 1/time or 1/concentration/time depending on the reaction order ) , and u denotes nutrient flux per transporter . This system ( Fig 1 ) can be described by two ordinary differential equations , S˙int=T⋅u ( Sext , Sint ) −γS ( Sext , Sint ) T˙=αT ( Sext , Sint , T ) −kγT ( Sext , Sint ) ⋅T , ( 1 ) where u is the rate of nutrient uptake per transporter , γs is the nutrient usage flux , αT is the transporter synthesis flux , and kγT is the transporter downregulation rate . We made the standard simplifying assumption that transporter downregulation flux is linearly proportional to transporter levels , γT=kγT ( Sext , Sint ) ⋅T [49–51] . In theory , the system could be further generalized by making u , γs , and kγT arbitrary functions of T , but this is not supported by the biology of any of the commonly studied nutrient uptake systems . Simplified versions of this system , such as Eq ( 1 ) , have been used to show that internal nutrient-dependent regulation of transporter synthesis can be homeostatic and thereby has provided a rationale for the ubiquity of this architecture [10] . To define the necessary and sufficient conditions for global perfect homeostasis we applied the method of Steuer et al . [14] . The conditions that are necessary and sufficient for homeostasis of this system are ( S1 Appendix , sections I , II ) : ∂Sext ( kγT ( Sext , Sint ) /αT ( Sext , Sint , T ) ) kγT ( Sext , Sint ) /αT ( Sext , Sint , T ) ⋅ ( 1−T⋅∂TαT ( Sext , Sint , T ) αT ( Sext , Sint , T ) ) =∂Sextu ( Sext , Sint ) u ( Sext , Sint ) −∂Sextγs ( Sext , Sint ) γs ( Sext , Sint ) . ( 2 ) This relationship is complicated and it is hard to imagine the regulatory interactions that would allow a biological instantiation of this general system . But Eq ( 2 ) does yield the insight that global perfect homeostasis is achieved by regulating transporters levels ( T ) such that they compensate for the change in usage rate ( γs ) or uptake per transporter ( u ) . In the following sections , we will constrain this general system; this will reduce Eq ( 2 ) to a condition that has a clear biological interpretation .
We sought to determine special cases of the general uptake system described in Eq ( 1 ) where the resulting homeostatic criterion is achievable by biologically plausible mechanisms . We started with the following biologically reasonable and standard assumptions: 1 ) there is little or no evidence for transporter levels directly affecting transporter synthesis , αT ( Sext , Sint ) ; 2 ) Sext negligibly affect nutrient usage , γS ( Sint ) ; and 3 ) internal nutrients negligibly affect nutrient uptake per transporter , u ( Sext ) . Under these simplifying assumptions , Eq ( 1 ) reduces to: S˙int=T⋅u ( Sext ) −γS ( Sint ) T˙=αT ( Sext , Sint ) −kγT ( Sext , Sint ) ⋅T . ( 3 ) The criterion for homeostasis , Eq ( 2 ) ( S1 Appendix , section II ) , reduces to: kγT ( Sext , Sint ) αT ( Sext , Sint ) ∝u ( Sext ) ⋅func ( Sint ) ( 4 ) where func ( Sint ) is a general function that depends solely on Sint ( and could also be constant ) ( Fig 2 ) . This condition states that any system in which the regulation of the transporter does not explicitly depend on Sext cannot provide global perfect homeostasis . Moreover , as transporter levels at steady state are given by αT/kγT , this criterion is satisfied when the transporter level , at steady state , is inversely proportional to the nutrient uptake rate . When this condition is met , the solution for the steady state value of Sint is given by solving 0=func ( Sint ) −γs ( Sint ) . ( 5 ) While , this condition does not guarantee a steady state solution for Sint , when a solution exists , it is independent of Sext . Is there a biological mechanism that can satisfy Eq ( 4 ) by making transporter downregulation proportional to nutrient uptake per transporter , i . e . kγT ( Sext , Sint ) ∝u ( Sext ) ? This condition corresponds to the requirement that only transporters that are actively transporting have the potential to be downregulated . We postulated that a simple biological scheme could achieve activity-dependent downregulation and thereby satisfy Eq ( 4 ) . This scheme is composed of a standard transporter cycle , where: 1 ) external nutrient binds to a transporter , 2 ) the transporter undergoes a conformational change allowing the nutrient to be released on the opposite side of the membrane , and 3 ) the transporter returns to its original conformation . In addition to this core system , we added an enzyme that recognizes and modifies only the nutrient bound conformation of the transporter ( Fig 3 ) . The modified transporter is downregulated . As long as the process is irreversible , direct and indirect inactivation are equivalent . Indeed , this system almost trivially couples the uptake and downregulation rate making them directly proportional ( Fig 3 and S1 Appendix , section IV ) . Two studies directly link uptake and downregulation [52 , 53] . In both cases , nutrient uptake leads to a conformational change in the transporter that is then ubiquitinylated in a manner analogous to our scheme in Fig 3 . This basic scheme is also supported by a series of crystal structures that show the conformational changes that occur upon nutrient binding [54 , 55] . While we only know of two proven examples for transporters , there are many examples of activity-dependent downregulated receptors . For example , ligand-dependent modification and downregulation is a core feature of G protein-coupled receptors and is required for ligand-mediated desensitization [56–62] . In bacteria , methyl-accepting chemotaxis receptors are modified in a ligand-dependent manner and this modification affects their sensing [63 , 64] . This activity-dependent methylation was demonstrated to be critical for robust adaptation [65] a process that is similar to homeostasis . While these examples involve receptors , transports share many features with receptors . Nutrient sensors and transporters are high related [66–71] . Multiple transporters have both transporting and signaling functions [72–78] and point mutations can interconvert receptors and transporters [79] . In some cases , receptor-mediated endocytosis [80–82] can even be considered a hybrid of transport and downregulation . Furthermore , transporters are modified and internalized in a manner that is very similar to receptors . Many transporters undergo internalization and degradation in a ubiquitin-dependent manner . Together , these examples argue that the proposed activity-dependent mechanism is not just mathematically possible but likely ubiquitous . The paucity of examples is likely due to a lack of experiments that have been performed in a manner such that activity-dependent regulation would have been observed . Indeed , changing the external nutrient concentration can stimulate transporter degradation [21 , 28 , 30 , 32–34] or inactivation [30 , 37 , 52 , 83 , 84] consistent with a ubiquitous role of activity-dependent regulation . When activity-dependent downregulation is the only form of transporter downregulation , global perfect homeostasis can be achieved . But , in real systems , dilution from cell growth and basal protein degradation will always contribute to transporter downregulation . To isolate the impact of dilution and protein degradation on global perfect homeostasis we considered the following minimal system: S˙int=kcat⋅T⋅SextSext+Kext−kγS⋅SintT˙=αT− ( kγTc+kγTa⋅SextSext+Kext ) ⋅T . ( 6 ) In this case , αT is constant , and the uptake per transporter has a standard Michaelian form , u=kcatSextSext+Kext . kγTa is maximal activity-dependent downregulation rate constant and kγTc is the combined rate constant for all other downregulation processes ( Fig 4A ) . We additionally assumed that nutrient uptake is Michaelian , while this was not essential , it is the standard assumption for nutrient uptake kinetics . Given that basal degradation and dilution are not homeostatic , the robustness of homeostasis , as quantified by our robustness metric , r , depends on the relative magnitude of kγTa and kγTc ( Fig 4B ) . As kγTa/kγTc increases , the system , Eq ( 6 ) , becomes more robust to changes in Sext ( Fig 4B and 4C ) . In the limit where activity-dependent downregulation dominates , kγTa >> kγTc , homeostasis is achieved ( S1 Appendix , section V ) ; in the limit where the degradation and dilution dominates , kγTa << kγTc , Sint tracks Sext . While Sint is robust to changes in Sext when kγTa >> kγTc , transporter levels are not . Instead , T tracks Sext and serves as the latent variable that allows the system to be robust; T adapts to keep the uptake rate , kcat⋅T⋅SextSext+Kext , constant ( Fig 4D ) . The decrease in robustness when basal degradation dominates is mirrored by a decrease in the sensitivity of the T to Sext ( Fig 4D ) . We wished to explore whether other common forms of regulation could achieve global perfect homeostasis . The criterion of Eq ( 4 ) could also be satisfied by a sensor that directly measures the nutrient flux ( Fig 5A ) or a sensor with identical binding kinetics as the transporter ( Fig 5B ) . This sensor could then regulate transporter synthesis or downregulation . Recent theoretical and experimental works from Kotte et al . and Kochanowski et al . have described the existence and role that flux sensing can play in metabolic regulation [43 , 44] and some nutrient systems contain external nutrient sensors [67] . If either of these sensors led to the downregulation of nutrient transporters it would be functionally equivalent to activity-dependent downregulation . In fact , the enzyme that modifies the nutrient bound transporter in Fig 3 is effectively acting as a flux sensor . But , both of these other mechanisms would require an extra level of regulation to normalize for the number of sensor molecules . In the case of an internal flux sensor , the activity of the sensor depends on the total nutrient flux , T⋅u ( Sext ) . Simple molecular interactions between the transporter and sensor would lead the downregulation rate of the transporter to depend on the activity of the transporters multiply the number of transporters , i . e . square of the transporter concentration , u ( Sext ) ⋅T2 . In the case of an external sensor , the activity of the sensor depends on binding the external nutrients , so the transporter downregulation will have the form kγT ( Sext , Sint ) ∝TSextSext+Ksensor . Since uptake has the form of kγT ( Sext , Sint ) ∝u ( Sext ) , this mode of regulation will only achieve homeostasis when the sensor's Michaelis constant is close to transporter's Michaelis constant ( Fig 5C; S1 Appendix , section VI ) . As these constants deviate , the system loses the ability to be robust to changes in external nutrient concentration ( Fig 5C ) . Therefore , while all three mechanisms are biologically feasible , we believe activity-dependent downregulation will be the most common .
Above we described how activity-dependent downregulation can achieve global perfect homeostasis . This system is distinct from transcriptional regulation of transporter synthesis by high gain feedback of internal nutrient concentrations ( e . g . negative feedback with high cooperativity ) which is considered to be a homeostatic mechanism [10] . In this system , it is typically assumed that the synthesis rate of the transporter is decreasing as a function of the internal nutrient concentration , e . g . αT ( Sext , Sint ) =kαT ( KMnKMn+Sintn ) and that the transporter downregulation rate is constant , kγT ( Sext , Sint ) =γT . It is easy to see that in this case the condition in Eq ( 4 ) is not met , i . e . kγT/αT does not depend on Sext , and thus this system cannot achieve global perfect homeostasis for any finite n ( Fig 6; S1 Appendix , section VII ) . Under real biological conditions , neither internal nutrient sensing nor activity-dependent sensing can achieve global perfect homeostasis . The two regulatory systems can be compared based on the parameters required to achieve the same robustness , r ( Fig 7 ) . Internal nutrient sensing approaches global perfect homeostasis when n is large . But , high cooperativity is mechanistically difficult to achieve . Activity-dependent downregulation approaches global perfect homeostasis when kγTa >> kγTc . High levels of activity-dependent downregulation are easy to achieve mechanistically but might come at a high cost due to increased protein turnover . To explore the potential trade-off due to cost that comes with activity-dependent regulation , we used two metrics: efficiency and robustness . We defined efficiency as the total nutrient uptake of a single transporter over its average lifetime . Robustness is defined above . When there is no activity-dependent downregulation , robustness takes on its minimal value and efficiency its maximal value ( Fig 8A; S1 Appendix , section VIII ) . As kγTa/kγTc is increased , robustness increases but efficiency goes down . As most nutrient homeostatic systems likely utilize both internal nutrient and activity-dependent regulation , we tested whether the combination improved the robustness-efficiency trade-off . We first asked whether the combined system is still able to achieve the global perfect homeostasis that activity-dependent downregulation is able to achieve alone . Indeed , the combined system still satisfies Eq ( 4 ) and thus can still achieve global perfect homeostasis ( Fig 2 ) . But this combined system might have additional desirable properties over a system with just activity-dependent regulation . Indeed , transcriptional repression by internal nutrient levels improves the efficiency of the system for any given robustness ( Fig 8B ) . As n is increased , the minimal robustness of the system is also increased ( Fig 8B ) . In addition to the minimal robustness increasing , the trade-off between robustness and efficiency also has a higher slope such that for the same efficiency , higher robustness is achieved and vice versa ( Fig 8B; S1 Appendix , section VIII ) . This result is intuitive—when nutrient levels are high , instead of just degrading transporter , the combined systems allows fewer transporters to be made .
We analyzed a general nutrient uptake system and derived the conditions for global perfect homeostasis . For a large family of scenarios , the internal nutrient can be made independent of the external nutrient if the ratio of transporter synthesis and depletion rates explicitly depends on the net uptake rate of the nutrient from the environment . A simple way to achieve this regulation is for active transporters to be preferentially downregulated , that is , have activity-dependent downregulation . While the transcriptional regulation of transporter gene synthesis has been relatively well characterized , the mechanisms that lead to the downregulation of plasma membrane transporters are less clear . Interestingly , a recent study of the yeast uracil transporter , Fur4p , suggests it exhibits properties that cause its degradation to depend on its activity . Specifically , there is evidence that binding of nutrients to the transporter causes it to adopt a conformation that marks it for ubiquitinylation , followed by endocytosis and degradation [52] . While this detailed analysis has not been performed on many transporters , it’s known that many transporters are endocytosed when nutrient levels increase . While previous work had suggested that this likely serves to protect cells from toxicity during acute increases in nutrient levels [39] , we propose that this mechanism could also play a crucial role in homeostasis at all external nutrient concentrations . Homeostasis can be regulated by internal and external nutrient concentrations . Both forms of regulation have limitations that affect their utility over all external concentrations , but both are useful in a range of biologically relevant regimes . Systems that integrate both internal and external nutrient concentrations can be robust over a wider range of concentrations ( with lower total energy input ) than systems that sense only internal or only external nutrient concentration . This complementarity may explain why these two modes of regulation are common in homeostatic systems . While out of the scope of this work , in future work it will be interesting to expand on this simple system in several ways . While here we focused on steady-state behaviors , kinetic differences between different modes of regulation are likely important . For example , different mechanisms of action can accentuate kinetic difference , e . g . transcriptional versus post-translation control . In addition , many homeostatic systems utilize internal nutrient storage or nutrient recycling which in principal can affect the homeostatic response . Furthermore , analysis of addition constraints that can be placed on Eq ( 1 ) could identify other biologically achievable systems that satisfy the homeostatic relationship in Eq ( 2 ) . The results we derived here are relevant for any multi-compartment biological system that implements homeostasis , where there is flux among the compartments . This work thus should be useful as a guide in studying homeostasis in any biological system and well as in the design of synthetic ones .
Detailed derivations can be found in S1 Appendix . To determine the requirements for homeostasis for the general uptake system described in Eq ( 1 ) , we used the method developed in [14] . In brief , we required the rule rank ( P|I ) =rank ( I ) ( 7 ) where P is a column of elements pi=∂log ( σi ) ∂log ( Sext ) with σi = αS , γS , αT , γT , which are the uptake and usage fluxes of the internal nutrient and uptake and usage fluxes for transporter respectively; each of those could be a function of Sext , Sint , and T . I is the matrix representation of largest parameter-independent subspace spanned by the columns of ( M | K ) where M is a column of elements mi=∂log ( σi ) ∂log ( T ) with σi = αS , γS , αT , γT , and K= ( αS00γS000αT00γT0 ) where αS0 , γS0 , αT0 , γT0 are the steady state solutions of fluxes . We constructed P|I and I matrixes and analytically found all the conditions that satisfy Eq ( 7 ) . The homeostasis condition was analytically derived using the described above criterion . Criterion necessary for global perfect homeostasis for special cases of Eq ( 1 ) were analytically derived by substituting into the resulting general solution with simplified forms of αS , γS , αT , γT . We considered several simplified conditions . If transporter production is independent of transporter concentration the following homeostasis condition was obtained γT ( Sext , Sint , T ) αT ( Sext , Sint ) ∝αS ( Sext , Sint , T ) γS ( Sext , Sint ) ⋅func ( Sint ) . ( 8 ) where func ( Sint ) is any function of internal nutrient concentration . To evaluate homeostasis we used two metrics: efficiency: e = αS / γT - the total nutrient uptake of a single transporter over its average lifetime and robustness: r = Sint / Sint , max where Sint , max is the maximum Sint across all Sext . To compare activity-dependent transporter downregulation with the internal nutrient based transcriptional regulation we compared the following two systems: S˙int=kcat⋅T⋅SextSext+Kext−kγS⋅SintT˙=αTKsynnKsynn+Sintn−kγTcT ( 9 ) and S˙int=kcat⋅T⋅SextSext+Kext−kγS⋅SintT˙=αT−kγTSintnKdegn+Sintn⋅T ( 10 ) where Ksyn and Kdeg are saturation constants and n is the Hill coefficient . To compare a system with internal nutrient based regulation of synthesis to a system with activity-dependent transporter downregulation we analytically calculated the parameters required to achieve a threshold robustness of r = 0 . 95 and r = 0 . 8 ( Fig 7 ) . In addition , we analytically compared the efficiency-robustness dependence in a system with both internal nutrient based regulation of synthesis and activity-dependent transporter downregulation ( n = 0 , 1 in Fig 8; n = 0 , 1 , 3 , 10 in S2 Fig ) . All analytical solutions were found in Mathematica version 10 . 2 ( Wolfram Research ) and were visualized in MATLAB R2015a ( MathWorks ) . | Homeostasis , the ability to maintain relatively constant internal conditions in the face of fluctuating environments , is fundamental to many biological processes . In nutrient homeostasis , a model homeostatic system , homeostasis is typically thought to be achieved through negative feedback regulation of the plasma membrane transporters synthesis by intracellular nutrient levels . Here , we first derive the general conditions that can achieve global perfect homeostasis in a simple uptake system . We found that this condition can be satisfied by the ubiquitous but less studied mechanism of activity-dependent transporter downregulation . If transporter downregulation is dependent on nutrient uptake rates , i . e . , activity-dependent downregulation , the system in principle can achieve homeostasis in any external environment . Activity-dependent and internal regulation can synergize to achieve homeostasis across a wide set of conditions at minimal energetic cost . Activity-dependent downregulation is likely to play a role in many diverse homeostatic systems . | [
"Abstract",
"Introduction",
"The",
"general",
"criterion",
"for",
"global",
"perfect",
"homeostasis",
"Flux-dependent",
"regulation",
"can",
"achieve",
"global",
"perfect",
"homeostasis",
"Comparison",
"and",
"combination",
"of",
"activity-dependent",
"downregulation",
"a... | [
"medicine",
"and",
"health",
"sciences",
"chemical",
"compounds",
"mechanisms",
"of",
"signal",
"transduction",
"phosphates",
"gene",
"regulation",
"cell",
"processes",
"homeostatic",
"mechanisms",
"physiological",
"processes",
"homeostasis",
"cell",
"growth",
"cellular",... | 2017 | Achieving global perfect homeostasis through transporter regulation |
Trichothecenes are a family of terpenoid toxins produced by multiple genera of fungi , including plant and insect pathogens . Some trichothecenes produced by the fungus Fusarium are among the mycotoxins of greatest concern to food and feed safety because of their toxicity and frequent occurrence in cereal crops , and trichothecene production contributes to pathogenesis of some Fusarium species on plants . Collectively , fungi produce over 150 trichothecene analogs: i . e . , molecules that share the same core structure but differ in patterns of substituents attached to the core structure . Here , we carried out genomic , phylogenetic , gene-function , and analytical chemistry studies of strains from nine fungal genera to identify genetic variation responsible for trichothecene structural diversity and to gain insight into evolutionary processes that have contributed to the variation . The results indicate that structural diversity has resulted from gain , loss , and functional changes of trichothecene biosynthetic ( TRI ) genes . The results also indicate that the presence of some substituents has arisen independently in different fungi by gain of different genes with the same function . Variation in TRI gene duplication and number of TRI loci was also observed among the fungi examined , but there was no evidence that such genetic differences have contributed to trichothecene structural variation . We also inferred ancestral states of the TRI cluster and trichothecene biosynthetic pathway , and proposed scenarios for changes in trichothecene structures during divergence of TRI cluster homologs . Together , our findings provide insight into evolutionary processes responsible for structural diversification of toxins produced by pathogenic fungi .
Secondary metabolites ( SMs ) are low-molecular-weight metabolites that are not required for growth or development , but instead provide ecological advantages under certain environmental conditions . Microbial SMs are diverse in chemical structure and biological activity; some are toxins , plant hormones , pigments , or antibiotics , and some have pharmaceutical properties . Many SMs contribute to host-pathogen interactions . Despite their structural diversity , most microbial SMs are derived from one of three classes of parent compounds: non-ribosomal peptides , polyketides , and terpenes [1] . SM structural diversity results from functional variation in enzymes that synthesize the parent compounds ( i . e . , non-ribosomal peptide synthetases , and polyketide and terpene synthases ) as well as enzymes that catalyze modifications of the parent compound . The latter enzymes include acyltransferases , amino transferases , dehydrogenases , reductases , dioxygenases , methyltransferases , monooxygenases , and prenyltransferases . In fungi , genes encoding enzymes required for synthesis of the same SM are typically located adjacent to one another in a biosynthetic gene cluster [2] . Such clusters can also encode transport proteins that export SMs from cells , and transcription factors that activate expression of cluster genes . SMs often consist of families of analogs that share a core structure , but vary in the presence of substituents ( functional groups ) attached to the core structure . Structural variation among analogs of the same SM family typically results from the presence , absence , or differences in function of genes encoding modifying enzymes [2 , 3] . Trichothecenes are a family of toxic SMs produced by some , but not all , species in multiple fungal genera , including Fusarium , Isaria , Microcyclospora , Myrothecium , Peltaster , Spicellum , Stachybotrys , Trichoderma , and Trichothecium [4–8] . Most known trichothecene-producing fungi are plant pathogens , and one , Isaria tenuipes , is an insect pathogen [5] . In Fusarium , trichothecene production contributes to pathogenesis on multiple crop plants [9–11] , and some Fusarium trichothecenes are among the mycotoxins of greatest concern to food and feed safety [12] . In addition , Stachybotrys trichothecenes have been implicated in negative health effects of mold growth in damp buildings [13] . In contrast , trichothecene production by Trichoderma arundinaceum contributes to its biological control activity against some plant pathogenic fungi [14] . The core structure of trichothecenes consists of a three-ring molecule known as 12 , 13-epoxytrichothec-9-ene ( EPT; Fig 1 ) , and analogs of trichothecenes differ from one another in the patterns of substituents attached to EPT ( Fig 2 ) . One type of structural variation has resulted in classification of trichothecenes into two groups [15] . Analogs in the first group , macrocyclic trichothecenes , have a macrolide ring resulting from a 12 or 14-atom chain esterified via hydroxyl groups at carbon atoms 4 and 15 ( C4 and C15 ) of EPT . Analogs in the second group , simple trichothecenes , lack a macrolide ring . The genetics and biochemistry of trichothecene biosynthesis have been studied most extensively in Fusarium , and biosynthetic pathways for Fusarium trichothecenes that significantly impact agriculture have been elucidated ( e . g . , deoxynivalenol , nivalenol and T-2 toxin; Fig 1 ) [3 , 16] . Additional studies indicate that at least the initial steps in trichothecene biosynthesis are similar in Fusarium , Myrothecium and Trichoderma [17–20] . Trichothecene biosynthesis begins with the cyclization of the primary metabolite farnesyl diphosphate to form the terpene trichodiene . This reaction is catalyzed by a terpene synthase ( trichodiene synthase ) . Subsequently , a cytochrome P450 monooxygenase ( trichodiene oxygenase ) catalyzes oxygenation of trichodiene at three or four positions to yield isotrichodiol or isotrichotriol , which can cyclize nonenzymatically to form EPT or 3-hydroxy EPT ( isotrichodermol ) , respectively . These latter molecules undergo one or more additional oxygenations , acylations and sometimes other modifications to form all trichothecene analogs [4] . The trichothecene biosynthetic gene ( TRI ) cluster is one of the most studied SM gene clusters in fungi . Homologs of the TRI cluster have been identified in Fusarium , Stachybotrys , and Trichoderma [3 , 16 , 17 , 21] . In addition , sequence analysis of a Myrothecium roridum cosmid clone identified three adjacent TRI genes presumed to be part of a larger cluster [22] , and RNAseq analysis of the fungus has identified homologs of six TRI genes [23] . The number of TRI genes per cluster varies among Fusarium , Stachybotrys , and Trichoderma and in some cases among species of the same genus . The Fusarium and Stachybotrys cluster homologs include the trichodiene synthase gene ( TRI5 ) , the trichodiene oxygenase gene ( TRI4 ) , two regulatory genes ( TRI6 and TRI10 ) , and other genes encoding enzymes that catalyze addition of substituents to the core EPT structure ( Table 1 ) . The Trichoderma TRI cluster differs in that it lacks TRI5 , which is located elsewhere in the genome [17] . Fusarium and Stachybotrys also have TRI genes at loci other than the TRI cluster . In some Fusarium species , monooxygenase ( TRI1 ) and acyltransferase ( TRI16 ) genes are at a second locus , and an acetyltransferase gene ( TRI101 ) is at a third locus [3] . In other Fusarium species , however , TRI1 and TRI101 are located in the cluster [24] . In Stachybotrys , TRI genes at loci other than the cluster are paralogs of genes in the cluster [21] . Functional analyses of TRI genes have elucidated the genetic bases for much of the structural diversity of trichothecene analogs produced by Fusarium [3 , 16 , 26] . However , the genetic bases for most of the structural diversity of trichothecenes produced by other fungi are not known . For example , T . arundinaceum produces harzianum A , a trichothecene with a polyketide-derived side chain [6 , 27] . The macrolide ring of macrocyclic trichothecenes is thought to be composed of both polyketide- and isoprenoid-derived moieties [15] . Although the genes responsible for formation of these substituents have not been identified , a polyketide synthase ( PKS ) gene is located in the TRI cluster of Stachybotrys species that produce macrocyclic trichothecenes [21] . The objective of the current study was to investigate variation of TRI genes among selected fungi in order to identify evolutionary processes that have likely contributed to structural diversity of trichothecene analogs produced by different fungi . To this end , we used genome sequencing to compare the gene content , arrangement , and sequences of TRI loci in selected species of nine genera . We also conducted additional functional analyses of selected TRI genes . The results indicate that gain , loss , and changes in function of genes are major contributors to structural diversity of trichothecenes . We used the results to infer an ancestral trichothecene biosynthetic pathway and to propose scenarios for gain and loss of trichothecene substituents during divergence of TRI cluster homologs . Together , our findings and inferences provide insights into the evolutionary processes that have given rise to biochemical diversity in plant pathogenic , entomopathogenic , and other fungi .
We used genome sequence data to examine the content and arrangement of TRI genes in 20 fungal strains that included 14 species from nine genera ( Table 2 ) . Genome sequence data for 12 strains were generated during the current study , while data for eight strains were generated in previous studies . The strains represented fungi with a range of lifestyles , including saprophytism , endophytism , plant pathogenicity , and entomopathogenicity . The two entomopathogenic fungi , Beauveria bassiana and Cordyceps confragosa ( Lecanicillium lecanii ) , have TRI genes but have not been reported to produce trichothecenes as far as we are aware . To assess TRI gene content in the fungi , we used coding region sequences of the 18 previously described TRI genes ( Table 1 ) as queries in BLASTn and BLASTx analyses against genome sequence databases of the 20 fungi . There was considerable variation in the presence and absence of TRI genes among the fungi examined ( Fig 3 , Table 3 ) . The number of TRI genes per genome varied from six in B . bassiana and C . confragosa to 15 in Fusarium sporotrichioides and Stachybotrys chartarum strain 40293 . The number of TRI genes varied within some genera and species as well . For example , S . chartarum had from nine to 15 TRI genes , because some TRI genes were duplicated in two strains [21] . TRI3 , TRI5 and TRI14 were the only TRI genes that were present in all 20 fungi examined , while TRI4 was present in all the fungi except the two Spicellum strains ( Table 3 ) . In some cases , TRI-gene counts per genome included two or three paralogs of the same gene ( Table 3 ) . We identified paralogs of the structural genes TRI3 , TRI4 , TRI5 , TRI17 and TRI18 and the regulatory genes TRI6 and TRI10 ( Table 3 ) . TRI6 had the largest number of paralogs; two each in Myrothecium , Spicellum , and S . chartarum 40293 , and three in each Trichothecium strain ( Table 3 ) . S . chartarum 40293 had the largest number of TRI paralogs , with over half of the TRI genes in this strain being paralogs . The sequence data also indicated that TRI genes can occur at one to as many as five distinct genomic locations ( loci ) . In some fungi with multiple TRI loci , genes located at different loci were paralogs . For example , the Myrothecium and Stachybotrys TRI clusters include nine and ten known TRI genes , respectively , but the TRI genes at other loci were paralogs of genes in the cluster . In Fusarium and Trichoderma , by contrast , TRI genes occurred at two or three loci , but the gene at the same or different loci in these fungi were not paralogous . In the Spicellum and Trichothecium strains examined , TRI genes were dispersed over five or six contigs ( S2 Fig ) . Although this dispersion of TRI genes on different contigs was likely an artifact of the genome sequence assembly in some cases , in other cases it was not artefactual . In S . roseum , for example , TRI12 was near the middle of 243-kb contig and TRI3 , TRI5 , TRI6a , TRI10 , and TRI14 were located adjacent to one another and near the middle of a 169-kb contig ( S2 Fig ) . In both of these contigs , the TRI genes were flanked by multiple genes that were unlikely to be involved in trichothecene biosynthesis based on their predicted functions . Thus , like Fusarium , Myrothecium , Stachybotrys , and Trichoderma , TRI genes in Spicellum and Trichothecium occur at two or more loci ( S2 Fig ) . The apparent absence of TRI4 in the Spicellum genome sequences was unexpected , because TRI4 is required for essential steps that occur early in trichothecene biosynthesis in other fungi [3 , 16] . We used three approaches to determine whether the absence of TRI4 was a sequencing or assembly artifact: 1 ) generation of genome sequence data for both strains of Spicellum using two or three Methods ( MiSeq , TruSeq , and Ion Torrent ) ; 2 ) RNAseq analysis of S . roseum ( strain 209012 ) grown under conditions that induced expression of other TRI genes; and 3 ) PCR analysis of the Spicellum strains using multiple primer pairs that amplified TRI4 fragments from Fusarium , Myrothecium and Trichothecium strains . None of these methods yielded evidence for a full-length TRI4 homolog in either Spicellum strain . However , BLASTx analysis of the TRI3-TRI6a intergenic region in S . roseum 209012 revealed a 558-nucleotide sequence that is likely a remnant of TRI4 ( S3 Fig ) . The absence of TRI4 in both Spicellum strains led us to predict that neither strain would produce trichothecenes . In Gas chromatography-mass spectrometry ( GC-MS ) analysis , we did not detect trichothecenes in culture extracts of S . ovalisporum , but we did detect them in culture extracts of S . roseum strain 209012 ( S4 Fig ) . Consistent with a previous study [32] , the most abundant trichothecene analog produced was 8-deoxy-trichothecin ( 4-O-butenoyl EPT ) . The absence of a TRI4 homolog in S . roseum suggests that it must have a gene ( s ) that encodes another trichodiene oxygenase . Attempts to identify such a gene by RNAseq analysis were not successful . That is , we did not find evidence for an oxygenase gene in S . roseum 209012 that exhibited a pattern of expression similar to those of known TRI genes . To gain insight into the variation of TRI gene homologs , we generated phylogenetic trees for individual TRI genes and for concatenated sequences of three of the genes . In preliminary analyses of individual TRI genes , we employed outgroup sequences of non-TRI genes from fungi that do not produce trichothecene and do not have a TRI cluster . Although distantly related , these outgroup sequences aligned to TRI sequences . In trees inferred from the resulting alignments , Microcyclospora homologs were consistently either the most or among the most basal lineages of TRI genes ( S7 Fig ) . Given this and its distant relationships to the other trichothecene-producing fungi examined in this study , Microcyclospora homologs were used as the root in subsequent TRI gene trees that excluded a non-TRI-gene outgroup . In trees inferred from homologs of individual TRI genes , relationships among more closely related homologs were generally well resolved ( bootstrap values >70% ) , whereas relationships among more distantly related homologs were generally not well resolved ( Figs 7 and 8 , S8 Fig ) . In most single-TRI-gene trees , Myrothecium and Stachybotrys formed a well-supported clade , and with the exception of TRI22 , Spicellum and Trichothecium formed a well-supported clade . In addition , Beauveria , Cordyceps and Fusarium also formed a well-supported clade in which Beauveria and Cordyceps had a sister relationship . Although branch conflicts were observed in comparisons of some TRI gene trees , most of the conflicts were not statistically supported by bootstrap analysis . We also performed Shimodaira-Hasegawa ( SH ) [35] and the Approximately Unbiased ( AU ) [36] tests to assess the significance of conflicting branches with bootstrap values > 70 . According to the results of these tests , the conflicts were not significant , with one exception; in the TRI22 tree , Spicellum homologs grouped in a well-supported clade with Myrothecium and Stachybotrys rather than with Trichothecium ( Fig 7 ) . This result suggests that the evolutionary history of TRI22 differs from other TRI genes in Spicellum , a phenomenon that has been previously reported for some Fusarium TRI genes [37 , 38] . We surmised that trees inferred from multiple TRI genes would better reflect relationships of TRI-cluster homologs than single-TRI-gene trees . Therefore , we inferred a phylogenetic tree from concatenated sequences of TRI3 , TRI5 and TRI14 , the only three TRI genes that were common to all fungi that were the focus of this study ( Table 3 ) . Trees inferred from these three genes individually did not have any well-supported branches that conflicted and were not significantly different from one another according to the SH and AU tests . Although results of a partition homogeneity test indicated inclusion of F . graminearum sequences resulted in significant heterogeneity in the data , relationships between genera did not differ in the concatenated gene trees with or without inclusion of F . graminearum sequences . Therefore , F . graminearum sequences were included in the final concatenated dataset . Some relationships among TRI gene homologs that were evident in single-TRI-gene trees were also evident in the concatenated gene tree . For example , Myrothecium and Stachybotrys formed a well-supported clade , as did Beauveria , Cordyceps , and Fusarium in the concatenated and most single-TRI-gene trees ( Figs 7 and 9 ) . The concatenated-TRI-gene tree had high bootstrap support for almost all branches , and therefore provided information for relationships of more distantly related TRI clusters . Based on the clades resolved in the concatenated-gene tree , we divided the cluster homologs into four lineages: lineage A consisted of the outgroup , M . tardicrescens; lineage B was the next most basal clade and consisted of Trichoderma sequences; lineage C consisted of Spicellum and Trichothecium sequences; and lineage D consisted of Beauveria , Cordyceps , Fusarium , Myrothecium and Stachybotrys sequences ( Figs 3 and 9 ) . Although there were no consistent differences in gene content of the different cluster lineages , lineages A-C occurred in fungi that produce less complex trichothecenes ( i . e . , with a hydroxyl or ester at C4 of EPT , a carbonyl at C8 in some cases , and a hydroxyl at C7 in one case ) , whereas lineage D cluster homologs occurred in fungi that can produce more complex trichothecenes ( i . e . , with carbonyl , hydroxyl or ester groups at up to five positions of EPT ) ( Fig 2 ) [7] . Visual inspection indicated that there was one or more well-supported branches ( bootstrap value > 70 ) in the TRI10 , TRI18 , TRI22 and TRI101 trees that conflicted with branches in the concatenated-TRI3-5-14-gene tree . Results of SH and AU tests indicated that the conflicts for TRI22 and TRI101 were significant , but those for TRI10 and TRI18 were not ( S2 File ) . To compare phylogenetic relationships of TRI cluster homologs to the relationships of the fungi in which the homologs occur , we inferred a species tree from the concatenated sequences of 20 housekeeping genes . We analyzed trees inferred from each housekeeping gene individually and assessed whether conflicts between the single-gene trees affected the species tree inferred from all 20 genes . The results of these analyses are shown in S3 File and suggest that the 20-housekeeping-gene tree provides a reasonable estimate of the species phylogeny . The high bootstrap values for almost all branches in the species tree provided evidence for the hierarchical relationships of most of the genera examined ( Fig 9 ) . There were two notable conflicts in the topologies of the species tree and the concatenated-TRI-gene tree . First , the sister relationship of Beauveria-Cordyceps and Fusarium observed in the TRI tree did not exist in the housekeeping gene tree; and second , the sister relationship of Beauveria-Cordyceps and Trichoderma observed in the housekeeping gene tree did not exist in the TRI tree . We used SH and AU tests to assess the significance of the conflicts between the trees overall and between the branches noted above . In a first set of tests , the TRI tree was constrained to conform to the housekeeping-gene tree , and housekeeping-gene tree were constrained to conform to the TRI tree . In these reciprocal assessments , the constrained trees were significantly worse than the unconstrained trees ( p < 0 . 05 ) . In a second set of tests , the TRI tree was constrained to include a sister relationship of Beauveria-Cordyceps and Trichoderma , and the housekeeping-gene tree was constrained to include the sister relationship of Beauveria-Cordyceps and Fusarium . Both tests indicated that the conflicts were significant ( p < 0 . 05 ) . In addition , none of the trees inferred from individual housekeeping genes included a well-supported Beauveria-Cordyceps-Fusarium clade ( S3 File ) , and none of the single-TRI-gene trees included a well-supported Beauveria-Cordyceps-Trichoderma clade . Homologs of TRI101 have been identified in trichothecene-producing and nonproducing species of Fusarium and other fungal genera [39 , 40] . In the current study , BLAST analyses indicated that the Beauveria and Cordyceps TRI101 homologs were more similar to TRI101 homologs in some other genera of trichothecene-nonproducing fungi than they were to homologs in Fusarium . To further investigate sequence differences of TRI101 homologs from Beauveria , Cordyceps , and Fusarium , we conducted a phylogenetic analysis with TRI101 homologs from diverse Ascomycetes . The resulting tree suggests that TRI101 homologs from Beauveria and Cordyceps are more closely related to a homolog from the trichothecene-nonproducing fungus Torrubiella hemipterigena than to homologs from trichothecene-producing fusaria ( Fig 8 ) . The tree also suggests that TRI101 homologs in trichothecene-producing fusaria are more closely related to homologs in trichothecene-nonproducing species of Fusarium , Cylindrosporum , Ilyonectria , and Neonectria than they are to the homologs in Beauveria-Cordyceps . The relatively distant relationships of TRI101 homologs in Beauveria-Cordyceps and trichothecene-producing species of Fusarium were unexpected given the close relationships of other TRI genes in these fungi .
We consider gain of a TRI gene to be the addition of a gene to trichothecene biosynthesis that was not previously involved in the process . Gain is suggested by the absence of a TRI gene in the genomes of multiple trichothecene-producing fungi , particularly those with a basal TRI cluster ( lineage A and B clusters ) , and the presence of the gene in the genome ( s ) of only one or a few fungi . TRI1 , TRI7 , TRI8 , TRI11 , and TRI13 are examples of such genes , because they were absent in all the fungi examined except Fusarium . Multiple mechanisms , including neofunctionalization , horizontal gene transfer ( HGT ) , and horizontal chromosome transfer , have the potential to contribute to gain of a SM biosynthetic gene in fungi . The presence of TRI gene paralogs in several fungi suggested gain of some TRI genes might have resulted from neofunctionalization ( i . e . , the process of gene duplication and subsequent divergence in function of a resulting paralog ) . However , with the exception of the paralogs , known TRI genes are more closely related to non-TRI genes than they are to other TRI genes [21 , 24 , 37] . This suggests that neofunctionalization of TRI genes has not contributed to TRI-gene gain , but instead neofunctionalization of closely related non-TRI genes has contributed to gain . TRI-gene gain may have also resulted when non-TRI genes changed function due to selection or other evolutionary processes to become involved in trichothecene biosynthesis . In fungi , SM biosynthetic gene clusters can degenerate such that some genes are pseudogenized or deleted and others remain intact [42–45] . If a TRI gene were gained by adaptation of a gene once involved in another process , the gene could have originated in a degenerating cluster . In our search for outgroups for phylogenetic analyses of individual TRI genes , BLASTx analysis of the fungal protein database in GenBank indicated that distantly related homologs of TRI genes occur in other fungi ( Fig 8 , S7 Fig ) . Furthermore , F . graminearum and F . sporotrichioides have genes that can partially compensate for the absence of TRI genes in deletion mutants [46 , 47] . Such genes encode enzymes that can modify trichothecene structures , and suggest another possible origin of gained TRI genes . The discussion above indicates that multiple mechanisms could have contributed to gain of TRI genes , but that the mechanisms responsible for gain of specific genes are not evident from our analyses . The trichothecene C3 acetylation gene , TRI101 , is a possible exception . Homologs of TRI101 are present in some trichothecene-producing fungi and in many trichothecene-nonproducing fungi [39 , 48] . In fact , all trichothecene-nonproducing species of Fusarium that have been examined have a TRI101 homolog , which is often designated as TRI201 [49] . All of the fungi examined here with a lineage-A–C TRI cluster and some fungi with a lineage-D cluster ( i . e . , Myrothecium and Stachybotrys ) produce trichothecenes that lack a C3 acetyl group , and therefore do not require TRI101 for trichothecene biosynthesis . But , some trichothecene-producing fungi that do not require TRI101 for production have a TRI101 homolog ( Fig 8 ) that is not located near other TRI genes . The presence of TRI101 in some trichothecene-nonproducing fungi and trichothecene-producing fungi that do not require C3-acetylation activity indicates that some TRI101 homologs have a function ( s ) other than trichothecene biosynthesis . In most trichothecene-producing fusaria that have been examined , TRI101 is not in the TRI cluster , but instead is located in the same genomic context as the homolog in some trichothecene-nonproducing species [40 , 49] . In addition , there is evidence that TRI101 has translocated into the TRI cluster rather than out of it during the evolutionary history of the Fusarium incarnatum-equiseti species complex ( FIESC ) [24] . These observations plus the knowledge that TRI101 functions in trichothecene C3 acetylation in both Beauveria and Fusarium suggest that TRI101 has become incorporated into trichothecene biosynthesis ( i . e . , gained ) in Beauveria-Cordyceps and Fusarium . The presence of TRI101 in trichothecene-nonproducing fungi further suggests that TRI101 gain has involved its adaptation from another function . It has been proposed that in Fusarium TRI101 and TRI201 are paralogs [49] . If this is the case , gain of TRI101 could be a result of neofunctionalization , whereby the ancestral gene was and the TRI201 paralog is involved in a process other than trichothecene biosynthesis , and the TRI101 paralog diverged to function in trichothecene biosynthesis . Results of the phylogenetic analysis of Tri101 homologs from trichothecene-producing and nonproducing fungi ( Fig 8 ) suggests that the gain of TRI101 and , therefore , the evolution of the C3 acetylation in the trichothecene biosynthetic pathway occurred independently in Fusarium and in Beauveria-Cordyceps . Two other trichothecene structural modifications , C8 and C15 oxygenation , appear to have also evolved independently in different fungi . Some Fusarium , Microcyclospora , and Trichothecium trichothecenes have a C8 oxygen atom ( Fig 2 ) . Functional analyses of F . graminearum and F . sporotrichioides indicate that Tri1 catalyzes trichothecene C8 oxygenation in Fusarium [50–53] . The absence of TRI1 in the Microcyclospora and Trichothecium genome sequences indicates that an enzyme other than Tri1 catalyzes C8 oxygenation in these fungi , which in turn indicates that C8 oxygenation in Fusarium evolved independently of its evolution in Microcyclospora and Trichothecium . ( Fig 10A ) . It is not known whether C8 oxygenation arose independently in Microcyclospora and Trichothecium . Among the fungi examined in this study , only Fusarium , Myrothecium and Stachybotrys are reported to produce trichothecenes with a C15 oxygen . In Fusarium , Tri11 catalyzes trichothecene C15 oxygenation [54] . The absence of a TRI11 homolog in the Myrothecium and Stachybotrys genome sequences ( Table 3 ) indicates that a gene other than TRI11 is required for C15 oxygenation in these fungi , and therefore that trichothecene C15 oxygenation in Myrothecium and Stachybotrys arose independently of its evolution in Fusarium ( Fig 10B ) . On the other hand , the presence of TRI11 homologs in Beauveria , Cordyceps and Fusarium suggests that gain of TRI11 occurred prior to divergence of the TRI cluster homologs in these fungi ( Fig 10B ) . To our knowledge and with the exception of TRI17 , genes required for synthesis of the macrolide rings of macrocyclic trichothecene have yet to be identified . Production of macrocyclic trichothecenes only by fungi with lineage-D TRI clusters suggests that the formation macrolide rings of these trichothecenes resulted from gain of genes in the Myrothecium-Stachybotrys lineage of trichothecene-producing fungi ( Fig 10C ) . We consider that loss of a TRI gene results from pseudogenization or complete deletion of the gene such that a functional version of it is no longer present in a genome . Evidence for loss is the occurrence of a gene in multiple fungi with a more basal ( i . e . , lineage A and B ) TRI cluster , but absence of the gene in one or more other fungi . TRI4 , TRI6 , TRI10 , TRI12 , TRI13 , TRI17 , and TRI22 are examples of genes that have likely been lost ( Table 3 ) . Gene loss is reported to contribute to structural variation of multiple fungal SMs [42 , 55 , 56] . Given the structural diversity of trichothecenes , TRI gene loss was expected to contribute to variation in gene content among the fungi examined , and indeed was previously reported from analyses of Fusarium and Stachybotrys [21 , 50 , 55] . However , absence of TRI4 in the Spicellum genomes was unexpected , because Tri4 catalyzes multiple reactions that are essential for formation of the EPT structure common to all trichothecenes ( Fig 2 ) [18–20] . Production of trichothecenes by S . roseum 209012 ( S4 Fig ) indicates that the fungus has a gene ( s ) that compensates for the absence of TRI4 . Furthermore , production of low levels of trichothecenes by T . arundinaceum tri4 deletion mutants indicates the existence of a gene that can partially compensate for the absence of TRI4 in this fungus [14] . The absence of TRI4 in Spicellum strains raises a question: what caused the loss of an enzyme that catalyzes multiple reactions essential for trichothecene biosynthesis and its replacement with another enzyme ? The absence of TRI6 and TRI10 in the B . bassiana and C . confragosa genomes was also unexpected given that these genes regulate expression of TRI genes in Fusarium [57–59] . Assuming B . bassiana and C . confragosa produce trichothecenes under some conditions , the absence of TRI6 and TRI10 indicates the existence of two fundamentally different regulatory systems for trichothecene biosynthesis in fungi . Among the fungi examined , B . bassiana and C . confragosa are the only insect pathogens . This raises a question: does the apparent change in regulation of TRI gene expression in B . bassiana and C . confragosa reflect an adaptation of trichothecene production for a lifestyle that includes insect pathogenesis ? The absence of the MFS transporter gene TRI12 was previously noted in analyses of the FIESC [24] and Stachybotrys species [21] . As a result , the absence of TRI12 in the B . bassiana and C . confragosa genomes was not surprising , but instead contributes to evidence that TRI12 is not essential for trichothecene production in fungi [60] . Presumably , another transporter ( s ) can compensate for the absence of Tri12 in trichothecene-producing fungi that lack TRI12 . The presence of TRI12 in all fungi with a lineages A–C TRI cluster and its absence in some fungi with a lineage-D cluster ( Fig 3 ) suggests that TRI12 was present in the ancestral TRI cluster . Further , the variable presence of TRI12 in lineage-D TRI clusters suggests three independent losses of the gene: once in Stachybotrys after divergence from Myrothecium; once in the Beauveria-Cordyceps clade after divergence from Fusarium; and once in FIESC after divergence from other fusaria . Except for F . graminearum , all known trichothecene-producing fungi examined here can produce trichothecenes that have a hydroxyl or ester group at C4 ( Fig 2 ) . This suggests that C4 oxygenation arose early in the evolutionary history of trichothecene biosynthesis , and therefore , that the common ancestor of extant TRI clusters likely encoded an enzyme that catalyzed this reaction . The results of the current and previous studies indicate that C4 hydroxylation is catalyzed by Tri22 in T . arundinaceum [17] and B . bassiana ( Fig 6E and 6F ) but by Tri13 in Fusarium [55] . Together , the presence of TRI22 in all the fungi examined here except Fusarium and the presence of TRI13 only in Fusarium ( Table 3 ) suggest that TRI22 is the ancestral C4 hydroxylase gene , and that TRI13 was acquired after the Fusarium TRI cluster diverged from the cluster in other genera ( Fig 10D ) . With the exception of the position of Spicellum and absence of Fusarium , the topology of the TRI22 tree is similar to the topology of the combined TRI3-TRI5-TRI14 tree , suggesting that the evolutionary history of TRI22 mirrors that of the TRI cluster to some extent . This , in turn , is consistent with the hypothesis that TRI22 is the ancestral C4 hydroxylase gene . If this hypothesis is correct , TRI22 would have been lost from and TRI13 would have been gained during the evolutionary divergence of the Fusarium cluster . Within Fusarium , production of trichothecenes that lack a C4 oxygen ( e . g . , deoxynivalenol ) results from pseudogenization of TRI13 [55 , 61] . This observation suggests a possible scenario to explain how the gene conferring trichothecene C4 hydroxylation changed from TRI22 to TRI13 . The scenario is based on the idea that if some extant fusaria produce trichothecenes that lack a C4 oxygen , ancestral trichothecene-producing fusaria could have produced trichothecenes that lack a C4 oxygen as well . In the scenario , TRI22 conferred C4 hydroxylation in ancestral trichothecene-producing fungi . Subsequently , during early divergence of the Fusarium TRI cluster , selection for production of trichothecenes with a C4 oxygen was relaxed and , as a result , TRI22 was lost . This gave rise to production of trichothecenes that lack a C4 oxygen . Subsequent to TRI22 loss , selection for production of trichothecene with a C4 oxygen was restored and , as a result , TRI13 was gained . Because all trichothecene-producing fusaria that have been examined to date have a functional or pseudogenized TRI13 , we surmise that gain of TRI13 occurred early in divergence of the Fusarium TRI cluster [16 , 24 , 62] . After gain of TRI13 , the gene was pseudogenized in one lineage of Fusarium , resulting in a mixed population in which some individuals produced trichothecenes that lacked C4 oxygen and others produced trichothecenes that have a C4 oxygen , a situation that still occurs in some lineages of Fusarium [37 , 63] . Thus , according to this scenario , trichothecene C4 hydroxylation has undergone a cycle whereby it existed in the ancestral trichothecene-producing fungus , was lost , then reacquired , and subsequently lost again . The results of the current and previous studies indicate that some trichothecene structural diversity has resulted from changes in function of TRI3 ( Fig 4 ) and TRI4 ( Fig 6A and 6B ) [18–20 , 25] . Tri3 catalyzes C4 acylation in T . arundinaceum and C15 acylation ( acetylation ) in Fusarium . We propose that the C4 acylation activity is ancestral and C15 acylation is derived , because trichothecenes produced by fungi with a lineage A or B TRI cluster ( i . e . , Microcyclospora and Trichoderma ) have an acyl group at C4 but not at C15 ( Fig 10E ) . The proposed ancestral and derived activities of Tri3 are consistent with the acquisition of C15 oxygenation in fungi with a lineage C and D TRI cluster ( Fig 10B ) , because a hydroxyl group at C15 is a prerequisite for C15 acetylation catalyzed by the Fusarium Tri3 [33 , 64] . If the proposed ancestral and derived activities of Tri3 are correct , the low level of C4 acetylation activity reported for recombinant F . graminearum Tri3 [64] indicates that some of the ancestral activity has been retained in this fungus . The results of this and previous studies indicate Myrothecium and Trichoderma Tri4 homologs catalyze oxygenation at three positions of trichodiene , whereas Beauveria ( Fig 6A and 6B ) and Fusarium Tri4 homologs catalyze oxygenation at four positions [18–20 , 25] . The three oxygenations result in formation of trichothecenes that lack a C3 hydroxyl , while the four oxygenations result in formation of trichothecenes that have a C3 hydroxyl . Further , trichothecenes that lack a C3 hydroxyl are produced by all the fungi with a lineage-A–C TRI cluster and some fungi with a lineage-D cluster , whereas trichothecenes that have a C3 hydroxyl are produced only by some fungi with a lineage-D cluster ( i . e . , Fusarium and presumably Beauveria and Cordyceps ) . Based on this information , we propose that the ability to catalyze three oxygenations is the ancestral condition of Tri4 , and the ability to catalyze four reactions is derived ( Fig 10F ) . This hypothesis is consistent with the idea that C3 acetylation is also a derived condition , because an oxygen atom at C3 is a prerequisite for C3 acetylation catalyzed by Tri101 [3 , 16] . There is evidence for changes in functions of other TRI genes/enzymes as well , because Tri1 and Tri8 are reported to differ in function within and/or among Fusarium species [26 , 50–52] . Functional analyses of the T . arundinaceum and M . roridum TRI17 homologs indicate that Tri17 catalyzes synthesis of the polyketide precursors of the substituents esterified to C4 of some trichothecene analogs ( Fig 5 , S5 Fig ) . The polyketide-derived substituents of harzianum A and macrocyclic trichothecenes are made up of linear molecules that are either six ( hexa-2 , 4-dienedioate ) or eight ( octatrienedioate and 6 , 7-dihydroxy-2 , 4-octadienoate ) carbon atoms long ( Fig 11A ) [15 , 65] . The presence of TRI17 in Spicellum and Trichothecium suggests that the Tri17 homologs in these fungi catalyze synthesis of the four-carbon chain ( 2-butenoyl ) that is esterified to the C4 oxygen of trichothecenes produced by these fungi ( Fig 11A ) . The variable lengths of the polyketide-derived substituents in trichothecenes produced by Myrothecium , Stachybotrys , Spicellum , Trichoderma , and Trichothecium suggest that collectively , Tri17 homologs can catalyze synthesis of four , six or eight-carbon polyketides . Furthermore , single species of Myrothecium and Stachybotrys and even single isolates of some species can produce macrocyclic trichothecenes that have six- or eight-carbon polyketide-derived substituents [66 , 67] . Thus , it is likely that in some species , a single Tri17 homolog can catalyze synthesis of both six and eight-carbon polyketides . The Tri17 protein is predicted to include an enoyl reductase ( ER ) domain [21] . During polyketide biosynthesis , an ER domain catalyzes reduction of carbon-carbon double bonds to carbon-carbon single bonds [68] . PKSs that have the appropriate combination of other functional domains but lack a functional enoyl reductase domain catalyze synthesis of polyketides that have alternating double and single bonds . The polyketide-derived substituents in Trichoderma , Myrothecium , and Stachybotrys trichothecenes have such alternating double and single bonds . Thus , the Tri17 enoyl reductase domain is almost certainly nonfunctional . In polyketide biosynthesis , carbon-chain length is controlled by the PKS enzyme [69] . In trichothecene biosynthesis , therefore , differences in polyketide-chain length ( i . e . , four carbons vs . six or eight carbons ) likely result from differences in amino acid sequence of Tri17 homologs . In phylogenetic trees inferred from predicted amino acid sequences of Tri17 and related PKSs , Spicellum and Trichothecium Tri17 homologs , which likely catalyze synthesis of a four-carbon polyketide , form a clade basal to the Myrothecium , Stachybotrys and Trichoderma homologs , which likely catalyze synthesis of six- and eight-carbon polyketides ( Fig 8 ) . These phylogenetic relationships suggest that Tri17 homologs that catalyze synthesis of a four-carbon polyketide represent the ancestral Tri17 condition , whereas homologs that catalyze synthesis of six and/or eight-carbon polyketides represent a derived condition . Given the predicted Tri17 functional domains and the structures of polyketide-derived substituents of harzianum A and macrocyclic trichothecenes , the polyketides precursors of these substituents are likely modified after release from Tri17 . For example , the polyketide precursor of octatrienedioate is likely a linear , eight-carbon polyketide with alternating double and single carbon-carbon bonds , and one carboxylic acid group ( Fig 11B ) . Because octatrienedioate has two carboxylic acid groups , its polyketide precursor likely undergoes modifications to form a second carboxylic acid group ( Fig 11B ) . Likewise , the 6 , 7-dihydroxy-2 , 4-octadienoate substituent in some macrocyclic trichothecenes contains two adjacent hydroxyl groups . Because the polyketide precursor of this substituent is likely the same as the octatrienedioate precursor , formation of 6 , 7-dihydroxy-2 , 4-octadienoate would also require modifications of its polyketide precursor ( Fig 11B ) . The structural diversity of trichothecene analogs produced by the fungi examined in this study combined with information on the distribution , phylogenetic relationships and functions of TRI genes allow for inference of ancestral states of the TRI cluster and trichothecene biosynthetic pathway ( Fig 12 ) . The inferred ancestral cluster included the structural genes TRI3 , TRI4 , TRI5 , TRI17 , TRI18 and TRI22 , the regulatory genes TRI6 and TRI10 , the transporter gene TRI12 , and TRI14 ( Fig 12A ) . The presence of these 10 genes in the ancestral cluster is consistent with their presence in all fungi examined and/or their presence in all lineages of the TRI cluster ( Fig 3 ) . Based on observations discussed above , TRI4 in this ancestral cluster conferred the ability to catalyze three oxygenation reactions to yield EPT [18 , 19] , and TRI3 conferred the ability to catalyze C4 rather than C15 acylation ( Fig 12B ) . In addition to rationales described above , this role for TRI3 in the ancestral cluster is consistent with the observation that during trichothecene biosynthesis in T . arundinaceum , Tri22 and Tri3 function in tandem; Tri22 catalyzes C4 hydroxylation , and Tri3 catalyzes acylation of the resulting C4 hydroxyl ( Fig 4 ) [17] . This tandem function of Tri22 and Tri3 is also consistent with the presence of both TRI22 and TRI3 in fungi that have a C4 but not C15 ester . As noted above , the ancestral Tri17 likely catalyzed synthesis of a four-carbon polyketide . Thus , the product of the inferred ancestral pathway would be 8-deoxy trichothecin ( 4-O-butenoyl EPT ) ( Fig 12B ) . We have included TRI18 in the inferred ancestral cluster because of its widespread distribution among trichothecene-producing fungi . However , we do not know its function in these fungi nor its likely function in the ancestral trichothecene biosynthetic pathway . In all fungi examined , when TRI17 is present , TRI18 is located adjacent to or near it ( Fig 3 ) . This consistent physical linkage of TRI17 and TRI18 suggests that the two genes could function together in biosynthesis . Given that TRI18 is predicted to encode an acyltransferase ( Table 1 ) , the most obvious possibility is that Tri18 catalyzes C4 esterification of the polyketide product of Tri17 to the C4 hydroxyl; i . e . , Tri18 could have the same function as the proposed ancestral function of Tri3 . Consistent with this idea is the evidence that there is a gene in T . arundinaceum that can partially compensate for TRI3 in tri3 deletion mutants of the fungus ( Fig 4 ) . Thus , one possibility is that in the ancestral trichothecene pathway , both Tri3 and Tri18 catalyzed esterification of the Tri17 product ( butenoyl ) to the hydroxyl group at C4 . Another possibility is that both enzymes catalyzed trichothecene C4 esterification , but esterified different molecules to the C4 hydroxyl ( e . g . , acetyl and butenoyl ) . Such a difference in function of Tri3 and Tri18 would have caused a branch at the end of the ancestral pathway , with one branch leading to 8-deoxy trichothecin ( 4-O-butenoyl EPT ) and the other leading to trichodermin ( 4-O-acetyl EPT ) ( Fig 12C ) . It is noteworthy that S . roseum can produce both of these metabolites ( S4 Fig ) , and has homologs of both TRI3 and TRI18 . Thus , the trichothecene products of the ancestral TRI cluster could be the same as those produced by an extant species of Spicellum . We have not included C8 oxygenation in the proposed ancestral pathway . However , the presence and absence of a C8 oxygenation step in the ancestral pathway are both consistent with currently available data . Our rational for not including a C8 oxygenation step in the pathway was based on: 1 ) the hypothesis that the ancestral pathway would have been simpler than most extant pathways; and 2 ) the trichothecene-C8-oxygenase gene ( s ) in Microcyclospora and Trichothecium has not been identified , and therefore its distribution is not known . Given this , it is not possible to say whether the gene ( s ) was likely to have been present in the ancestral TRI cluster . If the ancestral pathway included a C8 oxygenation step , this ability would have to have been lost four or more times and re-acquired at least once to account for the distribution of the C8-oxygenation ability among trichothecene-producing fungi examined in this study . On the other hand , if C8 oxygenation was absent in the ancestral pathway , three independent gain events could account for the current distribution of the ability to produce C8-oxygenated trichothecenes ( Fig 10A ) . Identification of gene ( s ) required for C8 oxygenation in Microcyclospora and Trichothecium should provide insight into the whether the ancestral pathway included this reaction . The sampling of fungi in the current study represents a majority of fungal genera in which trichothecene production has been reported [4–8] . The fungi are also phylogenetically diverse; M . tardicrescens is a member of the class Dothideomycetes , while the other fungi are members of five lineages within the class Sordariomycetes ( order Hypocreales ) . Thus , the TRI cluster and trichothecene production are uncommon and discontinuously distributed among ascomycetous fungi . How the current distribution of the TRI cluster arose is not clear , but comparison of TRI-gene and species trees inferred in the current study suggest some possibilities ( Fig 9 ) . The trees suggest that Fusarium TRI genes are more closely related than expected to those of Beauveria and Cordyceps , whereas Trichoderma TRI genes are more distantly related than expected to those of Beauveria and Cordyceps . In other studies , similar conflicts among phylogenetic trees have been attributed to lineage sorting and HGT [37 , 38 , 41] . If the conflicts were caused by lineage sorting , the Beauveria-Cordyceps , Fusarium , and Trichoderma TRI clusters could represent ancestral alleles or ancient paralogs that have been differentially inherited by these fungi such that Beauveria-Cordyceps and Fusarium inherited one allele ( or paralog ) , and Trichoderma inherited another allele ( or paralog ) . On the other hand , the close relationship of TRI cluster homologs in Beauveria-Cordyceps and Fusarium could have resulted from a HGT event between these two fungal lineages . The distant relationship of Trichoderma and Beauveria-Cordyceps TRI clusters could have also resulted from HGT events in which these two fungal lineages were recipients of distantly related TRI clusters . It is also possible that the distribution and phylogenetic relationships of the TRI cluster among the fungi examined here are a product of lineage sorting in some cases and HGT in others . Future studies aimed at identification and analysis of additional TRI cluster homologs in phylogenetically more diverse fungi could provide more definitive evidence for processes that have contributed to the distribution of the cluster . The conflict between the TRI22 tree ( Fig 7 ) and the concatenated TRI-gene tree ( Fig 9 ) with respect to the position of Spicellum suggests that the evolutionary history of TRI22 homologs has differed from other TRI genes in this fungal genus . This conflict could be attributed to lineage sorting or HGT , but either process would have involved TRI22 but not other extant TRI genes in Spicellum . Similar conflicts among TRI gene trees in closely related species of Fusarium were attributed to sorting of ancestral TRI-cluster alleles , which provide a mechanism to maintain production of two acetylated forms of deoxynivalenol [37] . It is unclear how a similar scenario could apply to TRI22 given that it confers C4 hydroxylation , and therefore , alleles of TRI22 are not likely to contribute to trichothecene structural diversity . It is noteworthy that the Spicellum strains were unusual among the fungi examined with respect to their position in the TRI22 tree as well as the absence of TRI4 in their genomes . Analyses of additional Spicellum species and their close relatives may provide insight into whether these two unusual features of Spicellum TRI genes resulted from related or independent events . Together , the results of the current and previous studies provide insights into evolutionary processes that have given rise to trichothecene structural diversity in fungi . The findings of the current study have facilitated inference of an ancestral trichothecene biosynthetic pathway ( Fig 12 ) that is consistent with extant pathways that collectively yield over 150 structurally diverse trichothecene analogs . Knowledge obtained from functional analyses of TRI genes in Fusarium and Trichoderma has contributed significantly to insights of the evolutionary history of trichothecene biosynthesis . However , it is likely that functional analyses of TRI homologs in other fungi will provide evidence that refines or disproves evolutionary scenarios proposed in this study . Functional studies in other fungi should also lead to identification of additional TRI genes responsible for structural diversity of trichothecene analogs , including genes responsible for: i ) C8 oxygenation in Microcyclospora and Trichothecium; ii ) structural modification of the polyketide precursors of macrocyclic trichothecenes and harzianum A; and iii ) formation of macrolide rings of macrocyclic trichothecenes . Future studies that sample numerous strains of the same species should provide evidence for whether gains and losses of genes are consistent across species . Trichothecene structural diversity appears to have arisen largely from gain , loss , and changes in function of TRI genes , evolutionary processes that have also attributed to structural diversity of ergot alkaloids produced by fungi of the family Clavicipitaceae [56] . Our results also indicate that the presence of some substituents of trichothecenes have evolved independently in different lineages of fungi through gain of different genes with the same function . In addition , at least one trichothecene modification ( C4 oxygenation ) appears to have been lost , reacquired , and subsequently lost again during divergence of the Fusarium TRI cluster . Structural diversity of trichothecene analogs likely reflects differences in selection experienced by fungi that produce the analogs [37] . Thus , the cycle of loss , reacquisition , and subsequent loss of C4-oxygenated trichothecenes likely reflects changes in selection for biological activity conferred by the analogs . Trichothecene production contributes to pathogenesis of some fusaria on some hosts [10] , and there is evidence that trichothecenes structural diversity in one lineage of Fusarium has been maintained by balancing selection [37] . It is not clear whether trichothecene production contributes to the pathogenicity of other fungi as well , but adaptation to pathogenesis on different plants and insects could provide selection pressure that has driven structural diversification of trichothecenes .
Genome sequences of B . bassiana [29] , C . confragosa [30] , F . graminearum [31] and Stachybotrys species [21] have been reported previously , and were downloaded from the National Center for Biotechnology Information ( NCBI ) database . Genome sequences for all other fungi were generated as part of the current study , primarily with a MiSeq Illumina platform ( Illumina , Inc . ) . In the initial genome sequence assemblies for the Spicellum and Trichothecium strains , almost every TRI gene was on a different contig , most of which were less than 5 kb in length . We partially overcame this limitation by generating a single genome sequence assembly from sequence reads generated with MiSeq , TruSeq ( Illumina , Inc . ) , and an Ion Torrent Ion Proton Sequencer ( Thermo Fisher Scientific Inc . ) . In the resulting assemblies , TRI genes were present on only five or six contigs ( Fig 3 , S2 Fig ) . To prepare DNA for genome sequencing , fungal strains were grown in YEPD medium ( 0 . 1% yeast extract , 0 . 1% peptone , 2% glucose ) for 2 days at room temperature with shaking at 200 rpm . The exception to this was M . tardicrescens HJS 1936 , which was grown in liquid YMG medium ( 0 . 4% yeast extract , 1% malt extract , 0 . 4% glucose ) for 10 days as previously described [7] . Mycelia were harvested by filtration , lyophilized , ground to a powder , and genomic DNA was extracted using the ZR Fungal/Bacterial DNA MiniPrep kit ( Zymo Research ) or the chloroform-phenol method as previously described [70] . DNA sequencing libraries were prepared as follows . For the MiSeq platform , the Nextera XT DNA library Preparation Kit was used as specified by the manufacturer ( Illumina ) . For the TruSeq platform , genomic DNA was first sonicated for four cycles with Diagenode Bioruptor system as specified by the manufacture to obtain 500-bp fragments ( Diagenode ) . Sequencing libraries were prepared from the fragmented DNA with the TruSeq DNA LT Library Preparation Kit as specified by the manufacturer ( Illumina ) . For the Ion Torrent Ion Proton Sequencer , the NEBNext Fast DNA Fragmentation & Library Prep Set for Ion Torrent was used as specified by the manufacturer ( New England BioLabs ) . Sequence reads obtained from each platform were processed and assembled using CLC Genomics Workbench ( Qiagen Inc . ) . Gene predictions were done using the program Augustus [71] and FGENESH ( Softberry , Inc . , Mount Kisco , New York ) . All TRI and housekeeping genes obtained from the genome sequences and used in phylogenetic analyses were also manually annotated . The 18 TRI genes used as queries in BLAST analyses have been described in Fusarium [55] , Trichoderma [17] and/or Stachybotrys [21] . BLASTx and BLASTn analyses were done against our in-house genome sequence database using CLC Genomics Workbench . Once contigs with TRI genes were identified in genome sequence assemblies , a portion of or the entire contig ( depending on contig length ) were subjected to gene prediction via FGENESH , and the resulting coding regions , as well as genomic sequences , were subjected to BLASTx analysis against the NCBI Non-redundant Protein Sequences database at NCBI to confirm the presence of TRI genes and to identify putative functions of other genes in the same region . For RNAseq analysis , Spicellum and Trichothecium strains were grown in liquid YEPD medium for one , two , and three days , after which mycelia were harvested by filtration and lyophilized . RNA was isolated with the RNeasy method ( Qiagen ) , and cDNA libraries were prepared with the MinElute Reaction Cleanup Kit ( Qiagen ) . cDNA libraries were then sonicated for five cycles with Diagenode Bioruptor system ( Diagenode ) as specified by the manufacturer to obtain 100- to 300-bp fragments . Sequencing libraries were prepared from the sonicated DNA using a NEBNext Fast DNA Library Prep Set for Ion Torrent ( New England BioLabs ) . The resulting library was then sequenced using Ion Torrent Ion Proton Sequencer platform ( Thermo Fisher Scientific ) . The resulting sequence reads were analyzed using the RNA-Seq Analysis function in CLC Genomics Workbench . To confirm sequences of the three TRI6 paralogs in Trichothecium , we amplified each paralog from three strains of T . roseum by PCR and sequenced the amplicons via Sanger Sequencing . PCR primers used for these amplifications are shown in S1 Table . The DNA polymerase GoTaq was used for amplification , and the conditions were those recommended by the manufacturer ( Promega ) . Amplicons were purified using standard agarose gel electrophoresis and the UltraClean protocol ( Mo Bio Laboratories ) . Amplicons were sequenced using BigDye Terminator version 3 . 1 and BigDye Xterminator Purification reagents ( Thermo Fisher Scientific ) , and sequences were determined with a 3739 DNA Analyzer ( Thermo Fisher Scientific ) . Sequences were viewed and edited using Sequencher ( Gene Codes Corporation ) . For deletion and complementation of T . arundinaceum TRI genes , previously described plasmid and protoplast-mediated transformation methods were used [17] . Deletion mutants and complemented deletion mutants were examined for their ability to produce harzianum A and other trichothecene analogs using the previously described two-step culture procedure [17] . TRI3 deletion was accomplished by transformation with plasmid pΔtri3 that had been linearized with ApaI prior to transformation ( Fig A in S1 File ) . Transformants were selected using 100 μg hygromycin B per mL of selection medium as previously described [17 , 72] . Transformants were analyzed by PCR with oligonucleotides Tarun-TrpC3/Tarun-compT35F ( S1 Table ) for the presence of a fragment expected to result from replacement of the TRI3 coding region with the hygromycin resistance cassette ( hygB ) . Transformants that yielded a PCR product were then analyzed by PCR to confirm the absence of TRI3 . TRI3 deletion was confirmed by Southern blot analysis using four hybridization probes ( Fig A in S1 File ) . Based on these analyses , we concluded that transformants tri3 . 1 , tri3 . 30 , tri3 . 33 and tri3 . 48 were tri3 deletion mutants ( Fig A in S1 File ) . tri3 deletion mutant tri3 . 1 was complemented by transformation with plasmid pTCtri3-ble linearized with EcoRI ( Fig B in S1 File ) . Transformants were selected using 75 μg phleomycin per mL of selection medium . Five transformants were analyzed by PCR for the presence of TRI3 and the phleomycin resistance cassette ( bleR ) with oligonucleotides Tarun-T3int3/Tarun-T3int5 and Tarun-Phleo-3/Tarun-Phleo-4 , respectively ( S1 Table ) . Three transformants that yielded amplicons from both primer pairs were analyzed by Southern blot analysis with two hybridization probes to confirm the presence of TRI3 and bleR ( Fig B in S1 File ) . TRI17 deletion was accomplished by transformation with plasmid pΔtri17 linearized with XhoI prior to transformation ( Fig C in S1 File ) . Transformants were selected with hygromycin as described above and analyzed by PCR with oligonucleotides Tarun-Db741/Tarun-Db742 to detect hygB , and with oligonucleotides Tarun-pks-F/Tarun-pks-R to test for the absence of TRI17 ( S1 Table ) . Transformants that yielded the appropriate amplicons were analyzed further by PCR with oligonucleotides Tarun-5-disrT/Tarun-TtrpC-disrT ( S1 Table ) to test for a fragment expected to result from replacement of the TRI17 coding region with hygB . Selected transformants were also analyzed by Southern blot analysis to confirm deletion of TRI17 ( Fig C in S1 File ) . Based on these analyses , we concluded that transformants tri17 . 96 , tri17 . 109 and tri17 . 139 were tri17 deletion mutants . tri17 deletion mutant tri17 . 139 was complemented by transformation with plasmid pTCtri17-ble linearized with NotI ( Fig D in S1 File ) . Transformants were selected with 100 μg of phleomycin per mL of selection medium . Transformants were analyzed by PCR for the presence of TRI17 and ble with oligonucleotides Tarun-pks-F/Tarun-pks-R and Tarun-Phleo-3/Tarun-Phleo-4 , respectively ( S1 Table ) . A subset of six transformants that yielded amplicons from both primer pairs were subjected to Southern blot analysis with three hybridization probes to confirm the presence of TRI17 and bleR ( Fig D in S1 File ) . Tri17 deletion mutant tri17 . 139 was also complemented with a plasmid , pTCMrtri17-ble , carrying M . roridum TRI17 and that had been linearized with EcoRI ( Fig E in S1 File ) . Transformants were selected as indicated above , and analyzed by PCR for the M . roridum TRI17 and bleR genes ( Fig E in S1 File ) . Based in the PCR results , transformants tri17 . MrT17 . C3 , . C4 , . C5 , and . C13 were selected for further studies . To express TRI4 and TRI22 from B . bassiana in F . verticillioides , the coding region ( with intron sequences intact ) plus ~500-bp of the 3’ flanking region of each gene was fused to the promoter sequence of the translation elongation factor 1α ( TEF1 ) gene from Aureobasidium pullulans . This was done using a previously described PCR-based fusion method [26] . The primers used for the PCR are shown in S1 Table , and DNA polymerases used were iProof High Fidelity DNA polymerase ( Bio-Rad Laboratories ) for TRI4 and Platinum Taq DNA Polymerase High Fidelity ( Thermo Fisher Scientific ) for TRI22 . For the fusion PCR , the A . pullulans TEF1 promoter was amplified from plasmid pTEFEGFP [73] with a reverse primer that consisted of approximately 22 nucleotides of the 3’ end of the TEF1 promoter and approximately 22 nucleotides of the 5’ end of the B . bassiana TRI4 ( or TRI22 ) coding region . Likewise , the B . bassiana TRI4 ( or TRI22 ) coding region and 3’ flanking region were amplified from genomic DNA of B . bassiana strain ARSEF 2860 using a forward primer that was essentially the reverse complement of the primer described above; that is , the primer included sequences of both the 3’ end of the A . pullulans TEF1 promoter and 5’ end of the B . bassiana TRI4 ( or TRI22 ) coding region . The two amplicons were then annealed and amplified by PCR without primers to generate the chimeric TEF1 promoter::TRI4 ( or TRI22 ) construct , which was then further amplified with nested primers as previously described [26] . The fusion amplicons were cloned into the PCR cloning vector pCR-XL1 TOPO ( Thermo Fisher Scientific ) and sequenced to confirm that PCR amplification did not introduce nucleotide errors . The geneticin resistance gene ( genR ) was then introduced into the resulting vector via NotI digestion using standard molecular biology protocols as described previously [26] . The TEF1 promoter::TRI4 ( or TRI22 ) -genR vector was then introduced into F . verticillioides strain M-3125 via a protoplast-mediated transformation protocol as previously described [72] . The presence of the construct in F . verticillioides transformants was confirmed by PCR with primer combinations used to generate the TEF1 promoter::TRI constructs . For precursor feeding experiments , F . verticillioides transformants were inoculated into liquid YEPD medium . Trichodiene or isotrichodermin ( 3-O-acetyl EPT ) , obtained from previous studies [74 , 75] , were dissolved in acetone and then added to cultures at a final concentration of 250 μM . The final concentration of acetone was less than 1% . The cultures were incubated in the dark at 28 °C with shaking ( 200 rpm ) . After six days , cultures were extracted with ethyl acetate , and the extracts were analyzed by GC-MC as described below . To express the B . bassiana TRI101 in S . cerevisiae , the TRI101 coding region was amplified from B . bassiana ARSEF 2860 genomic DNA using primers indicated in S1 Table and iProof High Fidelity DNA polymerase ( Bio-Rad Laboratories ) following the manufacturer’s recommendations . The PCR product was gel purified using the QIAEX II Gel Extraction Kit ( Invitrogen ) , treated with 1000 units of Taq DNA polymerase ( Qiagen ) to add A overhangs to the amplicon 3’ ends , and then cloned into pYES2 . 1 using the TOPO TA Yeast Expression Kit ( Invitrogen ) . In the resulting plasmid , the TRI101 coding region was fused to the GAL1 promoter and termination sequence . The cloned TRI101 was sequenced to confirm that PCR amplification did not introduce errors , and then the plasmid was introduced into an ayt1 mutant of S . cerevisiae ( strain YLL063C; GE Healthcare Dharmacon ) using the TRAFO protocol [76] . For feeding studies , yeast transformants were grown in the dark at 28 °C with shaking ( 200 rpm ) on minimal medium [77] , supplemented with leucine ( 1g/L ) , methionine ( 200 mg/L ) and histidine ( 200 mg g/L . After 3 days , cultures were centrifuged and the pellet was re-suspended in YGal medium ( 1% yeast extract , 2% peptone , 2% galactose ) to induce TRI101 expression . Isotrichodermol ( 3-hydroxy EPT ) was added to the cultures at a final concentration of 250 μM . After an additional four-day incubation in the dark at 28 °C with shaking ( 200 rpm ) , cultures were extracted with ethyl acetate and analyzed with GC-MS as described below . HPLC , LC-MS/MS and GC-MS were used to monitor trichothecenes and other metabolites produced by fungal strains . Harzianum A ( HA ) , which is not detectable by GC-MS analysis , was detectable and quantified by HPLC and/or LC-MS/MS analysis . For HPLC analysis of harzianum A , T . arundinaceum cultures were extracted with an equal volume of ethyl acetate . The upper phase was recovered and evaporated to dryness in a rotary evaporator at room temperature , and then redissolved in acetonitrile at 10% of the original volume . After a final 1/5 dilution , a 20 μL aliquot of the resulting sample was used for HPLC analysis . The HPLC system consisted of a Waters 600 HPLC connected to a 996 Photodiode Array Detector ( Waters Corporation ) [17] . The column was a Waters YMC analytical column ( 150 mm length , 4 . 6 mm internal diameter ) . The initial mobile phase was 40:60 acetonitrile:water with 0 . 1% trifluoroacetic acid , and had a flow rate of 1 mL/min . After 30 min , the mobile phase was adjusted to 100% acetonitrile over 10 min , held for 5 min , and then returned to the initial condition [17] . For both GC-MS and LC-MS/MS analyses , 5-mL aliquots of liquid cultures ( fungal biomass and medium ) were combined with 2 mL ethyl acetate , and mixed vigorously for 5 min . The ethyl acetate phase was recovered and used directly in GC-MS and LC-MS/MS analyses . The GC-MS system consisted of a Hewlett Packard 6890 gas chromatograph fitted with a HP-5MS column ( 30 m length , 0 . 25 mm film thickness ) and a 5973 mass detector ( Hewlett Packard ) . The carrier gas was helium with a 20:1 split ratio and a 20 mL/min split flow . The column was held at 120 °C at injection , heated to 210 °C at 15 °C/min and held for 1 min , then heated to 260 °C at 5 °C/min and held for 8 min . The LC-MS/MS system consisted of a ThermoDionex Ultimate 3000 UPLC fitted with a Phenomenex Kinetex F5 column ( 150 mm length , 2 . 1 mm diameter , 1 . 7 μm particle size ) connected to the ionspray interface of an ABSciex Qtrap 3200 mass spectrometer operated in negative mode . The chromatographic separation was done with a 200 μL/min gradient flow of water and acetonitrile with 0 . 3% ammonium acetate . The separation utilized a gradient from 35 to 95% aqueous acetonitrile over 10 min . The column was held at 50 °C for the entire analysis . In the analytical systems described above , identification of trichothecene analogs was confirmed by comparing their chromatographic retention times and , in MS analyses , molecular masses and mass spectra to those of standards . Novel trichothecenes were purified and their structures were determined by 1H and 13C NMR spectrometry as previously described [17 , 78] . For phylogenetic analyses of TRI and housekeeping genes , coding region sequences were aligned via the computer program MUSCLE as implemented in MEGA7 [79] . In general , nucleotide sequences were translated to amino acid sequences , aligned , and then converted back to nucleotide sequences before further analysis . Aligned sequences were subjected to maximum likelihood analysis using the computer program IQ-Tree version 1 . 5 . 5 [80] . In analyses with IQ-Tree , the best substitution model was determined by the program prior to tree building . Some alignments ( e . g . , TRI6 ) were also subjected to maximum parsimony analysis using MEGA7 , and for some genes ( e . g . , TRI17 and TRI101 ) alignments of deduced amino acid sequences were analyzed in addition to nucleotide sequences . For TRI genes , potential conflicts within sets of concatenated gene sequences were initially assessed using the Homogeneity Partition test as implemented in PAUP version 4 . 0a [81] . Bootstrap [80] , SH [35] , and AU [36] analyses , as implemented in IQ-Tree version 1 . 5 . 5 , were also used to assess whether trees inferred from different sequences were significantly different . In preliminary analyses of individual TRI genes , it was not clear which strain/species should be used to root trees . To assess the most appropriate TRI-gene root in each tree , we conducted BLASTx analyses against the GenBank non-redundant database to identify genes that were distantly related to but would still align to TRI genes [82] . Four to six of the best BLAST hits were then used for alignment and tree building with the corresponding TRI gene , and thereby obtain information on which TRI gene homolog ( s ) was most basal . | Toxins produced by pathogens can contribute to infection and/or colonization of hosts . Some toxins consist of a family of metabolites with similar but distinct chemical structures . This structural variation can affect biological activity , which in turn likely contributes to adaptation to different environments , including to different hosts . Trichothecene toxins consist of over 150 structurally distinct molecules produced by certain fungi , including some plant and insect pathogens . In multiple systems that have been examined , trichothecenes contribute to pathogenesis on plants . To elucidate the evolutionary processes that have given rise to trichothecene structural variation , we conducted comparative analyses of nine fungal genera , most of which produce different trichothecene structures . Using genomic , molecular biology , phylogenetic , and analytical chemistry approaches , we obtained evidence that trichothecene structural variation has arisen primarily from gain , loss , and functional changes of trichothecene biosynthetic genes . Our results also indicate that some structural changes have arisen independently in different fungi . Our findings provide insight into genetic and biochemical changes that can occur in toxin biosynthetic pathways as fungi with the pathways adapt to different environmental conditions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"taxonomy",
"medicine",
"and",
"health",
"sciences",
"fusarium",
"pathology",
"and",
"laboratory",
"medicine",
"fungal",
"genetics",
"pathogens",
"microbiology",
"fungal",
"structure",
"fungi",
"phylogenetics",
"data",
"management",
"plant",
"science",
"phylogenetic",
"... | 2018 | Evolution of structural diversity of trichothecenes, a family of toxins produced by plant pathogenic and entomopathogenic fungi |
Memory CD8 T cells can provide protection from re-infection by respiratory viruses such as influenza and SARS . However , the relative contribution of memory CD8 T cells in providing protection against respiratory syncytial virus ( RSV ) infection is currently unclear . To address this knowledge gap , we utilized a prime-boost immunization approach to induce robust memory CD8 T cell responses in the absence of RSV-specific CD4 T cells and antibodies . Unexpectedly , RSV infection of mice with pre-existing CD8 T cell memory led to exacerbated weight loss , pulmonary disease , and lethal immunopathology . The exacerbated disease in immunized mice was not epitope-dependent and occurred despite a significant reduction in RSV viral titers . In addition , the lethal immunopathology was unique to the context of an RSV infection as mice were protected from a normally lethal challenge with a recombinant influenza virus expressing an RSV epitope . Memory CD8 T cells rapidly produced IFN-γ following RSV infection resulting in elevated protein levels in the lung and periphery . Neutralization of IFN-γ in the respiratory tract reduced morbidity and prevented mortality . These results demonstrate that in contrast to other respiratory viruses , RSV-specific memory CD8 T cells can induce lethal immunopathology despite mediating enhanced viral clearance .
Respiratory syncytial virus ( RSV ) is a major cause of severe disease in young children , the elderly , and immunocompromised populations [1–6] . Furthermore , RSV is the leading cause of infant hospitalizations creating an immense healthcare burden for treatment and prevention [1 , 2 , 7–11] . There is currently no licensed vaccine for RSV . During a primary RSV infection , the CD8 T cell response is crucial for mediating viral clearance [12 , 13] . Depletion of CD8 T cells in mice prior to RSV challenge leads to elevated viral loads , but also ameliorates morbidity [12] . Thus , CD8 T cells contribute to both viral clearance and immunopathology following an acute RSV infection . RSV-specific memory CD8 T cells also contribute to protection from a secondary infection [12] . Antibody-mediated depletion of memory CD8 T cells in RSV-immune mice impairs viral clearance following re-infection as compared to non-treated controls [12] . Thus , vaccines that elicit robust memory CD8 T cell responses may help promote long-lived immunity against RSV . The induction of neutralizing antibodies remains a primary goal of most RSV vaccines due to their clearly established capacity to reduce the severity of RSV-induced disease [14–17] . In contrast , studies have demonstrated that robust memory CD4 T cell responses can mediate vaccine-enhanced disease following RSV infection [18 , 19] . Adoptive transfer of activated effector RSV-specific CD8 T cells , in vitro stimulated T cell lines , or in vitro propagated T cell clones leads to enhanced RSV clearance from the lung following RSV challenge . These effector CD8 T cell transfers were also associated with increased weight loss , indicating that infusion of effector CD8 T cells can induce increased systemic disease [20–23] . However , the role of memory CD8 T cells in providing protection against RSV infection remains unclear . Evaluating the capacity of memory CD8 T cells to mediate protection against RSV infection is critically important because high neutralizing antibody titers alone are insufficient to prevent RSV-induced disease in every individual [14 , 24] . Herein , we evaluated the protective capacity of memory CD8 T cells against RSV infection in the absence of RSV-specific CD4 T cell memory and antibodies . We employed a dendritic cell-Listeria monocytogenes ( DC-LM ) prime-boost immunization regimen to induce high magnitude , RSV epitope-specific CD8 T cell responses in naive mice . A similar prime-boost immunization strategy has been shown to elicit protection against other respiratory viruses including influenza A virus ( IAV ) and severe acute respiratory syndrome coronavirus ( SARS-CoV ) [25 , 26] . DC-LM immunization induced robust memory CD8 T cell responses that reduced viral titers following RSV challenge . However , despite enhanced viral clearance , immunized mice experienced increased pulmonary disease , weight loss , and mortality . Exacerbated disease and mortality was unique to the context of an RSV infection as immunized mice were protected against challenge with a lethal dose of a recombinant IAV expressing an RSV-derived CD8 T cell epitope . The lethal immunopathology observed in immunized mice was caused by rapid and excessive IFN-γ production by memory CD8 T cells in the airways . Our studies reveal that memory CD8 T cells enhance RSV clearance similar to other viral infections , but are unique in that they mediate severe immunopathology caused by the overproduction of IFN-γ .
Peptide-coated , mature DCs can be utilized to prime a CD8 T cell response that allows for robust secondary expansion following a booster immunization in mice [27] . To induce RSV-specific CD8 T cell memory in the absence of virus-specific CD4 T cells and antibodies , naive mice were immunized with matured splenic DCs loaded with M282-90 ( M282 ) peptide and boosted 7 days later via infection with an attenuated , recombinant LM strain expressing the M282 epitope . A control group without RSV-specific CD8 T cell memory was generated by immunizing mice with DCs not exposed to peptide followed by a boost with an LM that did not express an RSV-derived epitope . DC-LM immunization led to a significant ( p<0 . 001 ) increase in M282-specific CD8 T cell frequencies following the LM booster inoculation within the peripheral blood leukocytes ( PBL ) compartment as compared to the control group ( Fig 1A ) . Approximately 20% of all CD8 T cells in the PBL were M282-specific at day 42 post-boost ( Fig 1A ) . Immunized mice challenged with RSV exhibited a significant ( p<0 . 001 ) reduction in lung viral titers by day 4 post-infection ( p . i . ) compared to the control group undergoing a primary RSV infection ( Fig 1B ) . As expected , immunized mice exhibited an increased number of total and M282-specific CD8 T cells in the lung by day 5 p . i . as compared to non-infected M282-immunized controls ( Fig 1C and 1D ) . In addition , there was a greater frequency ( p<0 . 001 ) of IFN-γ+ and IFN-γ+TNF+ CD8 T cells following M282 peptide re-stimulation at days 4 and 5 p . i . as compared to control groups ( Fig 1E ) . Thus , DC-LM immunization elicited robust memory CD8 T cell responses that mediated enhanced viral clearance following RSV challenge . Pre-existing memory CD8 T cells altered the response of specific cell types following RSV infection ( S1 Fig ) . Both conventional CD4 and regulatory CD4 T cell ( Treg ) numbers were significantly reduced ( p<0 . 05 ) at days 4 and 5 p . i . in the lung compared to the acute infection control group . Furthermore , the number of monocytes were significantly ( p<0 . 05 ) increased at day 5 p . i . as compared to control mice . These changes were in contrast to eosinophil , neutrophil , and natural killer ( NK ) cell responses , which remained similar to non-infected M282-immunized mice . Overall , DC-LM immunization induced memory CD8 T cells that altered the magnitude of the subsequent cellular infiltrate and enhanced viral clearance following RSV infection . Since DC-LM immunized mice exhibited decreased viral titers , we next determined if they also experienced reduced disease severity . We evaluated weight loss and airway obstruction , both key disease manifestations that can be assessed following RSV infection in mice [18 , 28 , 29] . Despite enhanced viral control , M282-immunized mice exhibited a significant ( p<0 . 01 ) decrease in survival ( Fig 2A ) . Approximately 40% of fatalities were due to M282-immunized mice naturally succumbing to RSV infection , while 60% were euthanized upon reaching a humane weight loss endpoint . This outcome was both unexpected and unusual since an acute RSV infection is rarely fatal in adult BALB/c mice . M282-immunized mice also exhibited significantly ( p<0 . 05 ) increased weight loss ( Fig 2B ) and reduced pulmonary function ( Fig 2C and 2D ) . Additionally , we evaluated lungs by histology for evidence of diffuse alveolar damage ( DAD ) , an acute form of lung injury [30 , 31] . If severe and extensive enough , DAD is the foundational lesion in the clinical syndrome known as acute respiratory distress syndrome . M282-immunized mice revealed increased ( p<0 . 001 ) histopathological evidence of characteristics associated with early stages of DAD including cellular sloughing and necrosis , alveolar hemorrhage , early cellular infiltrates , and hyaline membrane formation ( Fig 2E and 2F and S2 and S3 Figs ) . Previous work has demonstrated that the M282-specific CD8 T cell response contributes to the immunopathology associated with an acute RSV infection [32] . Thus , it was unclear if the increased disease severity observed in M282-immunized mice was unique to the M282-90 epitope . To address this possibility , we evaluated mice immunized against the F85-93 ( F85 ) CD8 T cell epitope following RSV infection [33] . Similar to M282-immunized mice , F85 DC-LM immunization induced a high frequency of RSV F85-specific memory CD8 T cells that mediated a decrease in lung viral titers at day 4 following RSV challenge ( S4A and S4B Fig ) . In addition , F85-immunized mice exhibited increased mortality , weight loss , and pulmonary dysfunction as compared to controls ( S4C–S4F Fig ) . Thus , the severe immunopathology induced by memory CD8 T cells was not specific to either a particular epitope or an RSV protein . We next determined if RSV-specific memory CD8 T cells would also cause increased disease in C57BL/6 mice , as an acute RSV infection in this mouse strain typically causes only mild disease [34] . Therefore , we immunized C57BL/6 mice against the immunodominant M187-195 ( M187 ) CD8 T cell epitope [35] . DC-LM immunization targeting the M187 epitope resulted in approximately 33% M187-specific CD8 T cells in the PBL by day 28 post-boost ( S5A Fig ) . Similar to M82-immunized BALB/c mice , M187-immunized C57BL/6 mice exhibited significantly reduced lung virus titers ( p<0 . 001 ) , decreased pulmonary function , and increased weight loss following RSV infection ( S5B–S5E Fig ) . However , in contrast to M282-immunized BALB/c mice , all of the M187-immunized C57BL/6 mice survived following RSV infection . This data indicates that M187-specific CD8 T cells also contribute to immunopathogenic responses in the C57BL/6 genetic background . The challenge virus utilized in our study is the RSV A2 strain , which primarily induces a Th1-biased immune response [36] . However , other RSV strains can induce more heterogeneous Th responses . To determine if memory CD8 T cells generated by DC-LM vaccination induce immunopathology independently of the RSV challenge strain , we infected M282-immunized mice with the recombinant RSV A2-line19F strain , which promotes a more Th2-biased immune response than RSV A2 [36] . M282-immunized mice challenged with A2-line19F exhibited mortality , weight loss , and pulmonary dysfunction that was identical to mice challenged with A2 ( S6 Fig ) . These results indicate that M282-specific memory CD8 T cells induce immunopathology independently of the RSV strain utilized for challenge . To determine if M282-specific memory CD8 T cells would be pathogenic outside of RSV infection , we challenged M282-immunized mice with a recombinant IAV expressing the M282 epitope ( IAV-M282 ) . Slutter et al . recently used a similar DC-LM prime-boost strategy to elicit IAV-specific memory CD8 T cells and demonstrated protection against a subsequent IAV infection [25] . Similar to RSV challenge , immunized mice infected with IAV-M282 exhibited a significant increase in the number of total and M282-specific CD8 T cells in the lung by day 5 p . i . ( S7 Fig ) . IAV-M282 infection also resulted in a significant ( p<0 . 05 ) increase in monocytes and decrease in eosinophils compared to non-infected M282-immunized controls ( S7 Fig ) . CD4 T cell , Treg , neutrophil , and NK cell numbers were not significantly altered in IAV-M282 challenged mice ( S7 Fig ) . As anticipated , all control mice , which do not have virus-specific memory CD8 T cells , succumbed to a lethal IAV-M282 challenge . However , in contrast to the significant ( p<0 . 001 ) mortality observed in M282-immunized mice following RSV challenge , M282-immunized mice were protected against a lethal IAV-M282 infection ( Fig 3A ) . Pre-existing M282-specific memory CD8 T cells also prevented the prolonged weight loss and pulmonary dysfunction observed in control mice following IAV-M282 infection ( Fig 3B–3D ) . Furthermore , weight loss and pulmonary dysfunction following RSV challenge were significantly ( p<0 . 05 ) increased at early timepoints compared to IAV-M282 infected mice ( Fig 3B–3D ) . In addition , M282-specific memory CD8 T cells induced a significant ( p<0 . 001 ) reduction in IAV titers at both day 4 and day 6 p . i . ( Fig 3E ) . These data indicate that pre-existing RSV M282-specific memory CD8 T cells mediate enhanced IAV-M282 clearance and promote increased survival of mice by preventing prolonged disease . To confirm that memory CD8 T cells would also provide protection without extensive immunopathology to a lower viral inoculum , we challenged M282-immunized mice with a sublethal dose of IAV-M282 . Similar to a lethal dose of IAV-M282 , M282-immunized mice experienced significantly ( p<0 . 05 ) improved pulmonary function and less weight loss as compared to the control group following a sublethal infection ( S8 Fig ) . Thus , memory CD8 T cells do not promote distinct patterns of disease severity between sublethal and lethal IAV infections . These results indicate that lethal immunopathology associated with a high-magnitude memory CD8 T cell response is unique to the context of an RSV infection . The DC-LM immunization strategy utilized in our study induced high frequency systemic CD8 T cell memory but resulted in fatal immunopathology following RSV challenge . We hypothesized that a local immunization would promote a large population of antigen-specific resident memory CD8 T cells within the lung that would prevent the immunopathology observed after RSV infection . To evaluate the effects of local immunization , DC-M282-primed mice were either not boosted , given a systemic LM-M282 boost intravenously ( i . v . ) , or given a local IAV-M282 boost intransally ( i . n . ) . Administration of an IAV-M282 boost induced a significant ( p<0 . 001 ) increase in the total number of M282-specific CD8 T cells in the lung prior to RSV challenge as compared to a systemic LM-M282 boost ( S9A and S9B Fig ) . We next utilized intravascular staining to determine the localization of cells within the lung following immunization [37 , 38] . The majority of M282-specific CD8 T cells in the lung of DC-M282-primed mice that were either not boosted or boosted systemically with LM-M282 were located within the pulmonary vasculature ( S9C Fig ) . In contrast , greater than 85% of M282-specific CD8 T cells were localized within the lung tissue in IAV-M282-boosted mice , resulting in a significant ( p<0 . 001 ) increase in total number as compared to LM-M282-boosted mice ( S9C and S9D Fig ) . In addition , local boost in the lung with IAV-M282 induced a large frequency of RSV-specific CD8 T cells within the lung tissue expressing both CD69 and CD103 , which represent the canonical markers of resident memory CD8 T cells . In contrast , systemic LM-M282 immunization failed to elicit resident memory CD8 T cells within the lung tissue ( S9E and S9F Fig ) . To determine whether the RSV-specific resident memory CD8 T cells generated by local immunization induce less severe immunopathology than their systemically induced counterparts , DC-M282-primed mice that were either not boosted or boosted with either LM-M282 i . v . or IAV-M282 i . n . were challenged with RSV and monitored for morbidity and mortality . In contrast to a systemic LM-M282 boost , a local boost in the lung with IAV-M282 resulted in 100% survival following RSV challenge ( S10A Fig ) . The IAV-M282 boost also resulted in significantly ( p<0 . 05 ) reduced weight loss and pulmonary dysfunction following RSV infection compared to the LM-M282 boost ( S10B–S10D Fig ) . These results suggest that local prime-boost immunization generates RSV-specific resident memory CD8 T cells that prevent fatal immunopathology and ameliorate disease following RSV challenge . We next evaluated the primary antiviral effector functions of memory CD8 T cells to determine their contribution to the immunopathology . The primary pathway of CD8 T cell-mediated cytolysis is through the release of perforin and granzymes [39–41] . Therefore , we evaluated survival and disease severity following RSV infection of perforin-deficient M282-immunized mice . Mice deficient in perforin exhibited accelerated mortality as compared to wild-type ( WT ) controls ( S11A Fig ) . The kinetics of weight loss and pulmonary dysfunction were similar between WT and perforin-deficient M282-immunized mice following RSV challenge ( S11B–S11D Fig ) . Therefore , perforin is not required to mediate exacerbated disease , but may be necessary to prevent additional mortality . We also evaluated the role of TNF in immunized mice given its previously identified contribution to immunopathology associated with an acute RSV infection [42] . Antibody-mediated neutralization of TNF in the airways at the time of RSV challenge led to survival of all M282-immunized mice ( S12A Fig ) . Neutralization of TNF significantly ( p<0 . 05 ) reduced both weight loss and pulmonary dysfunction ( S12B–S12D Fig ) . These data illustrate that similar to an acute RSV infection [42] , TNF also contributes to the immunopathology associated with memory CD8 T cell responses . Furthermore , TNF neutralization had no significant impact on viral titers at day 4 p . i . ( S12E Fig ) . Assessment of TNF levels at day 2 following RSV challenge revealed similar levels between control- and M282-immunized mice ( S12F Fig ) . TNF levels in the lung were significantly ( p<0 . 01 ) increased in M282-immunized mice at day 4 p . i . as compared to the control group , but no increase was observed in the serum ( S12F Fig ) . However , the overall amount of TNF in both the lung and the serum had decreased by day 4 p . i . , the time when immunized mice began to succumb to the RSV infection . Therefore , these data demonstrate that TNF contributes to general inflammation in the lung during both a primary and recall response to RSV infection , but that TNF is necessary for the lethal immunopathology to occur . Due to the accelerated memory CD8 T cell response in DC-LM immunized mice , we speculated that the early inflammatory cytokine milieu would be distinct from an acute RSV infection . We initially evaluated IFN-γ levels , as it is a common pro-inflammatory cytokine produced by CD8 T cells following viral infection . IFN-γ protein levels were significantly ( p<0 . 001 ) increased in both the lung and serum at day 2 p . i . of M282-immunized mice as compared to the control group ( Fig 4A and 4B ) . IFN-γ levels in M282-immunized mice remained elevated above controls at day 4 p . i . , but the overall amount was reduced as compared to day 2 following infection . Interestingly , challenge with IAV-M282 resulted in significantly ( p<0 . 001 ) greater IFN-γ protein levels in the lung at day 2 and day 4 p . i . as compared to challenge with RSV ( S13A Fig ) . In contrast , serum IFN-γ and lung TNF levels in IAV-M282-infected mice were reduced ( p<0 . 001 ) at day 2 but increased ( p<0 . 001 ) at day 4 p . i . when compared to the levels observed following RSV challenge ( S13B and S13C Fig ) . To determine the in vivo source of IFN-γ in RSV-infected immunized mice , we treated mice with brefeldin A ( BFA ) to capture cells producing IFN-γ via intracellular staining and flow cytometry [43] . Leukocytes producing IFN-γ in vivo were readily identified using this previously established method ( S14A Fig ) . Upon evaluation of the primary leukocyte populations present in the lung following RSV infection , only lymphocytes had produced IFN-γ in immunized mice at day 2 p . i . ( Fig 4C ) . Only a small frequency of CD4 T cells and NK cells secreting IFN-γ were observed at day 2 p . i . , whereas approximately 45% of CD8 T cells were producing IFN-γ ( Fig 4C ) . These data also correlate with the rapid and transient increase in the amount of IFN-γ protein we observed in the lung and serum at day 2 p . i . , as virtually no IFN-γ-producing cells were recovered on day 5 p . i . ( S14B Fig ) . The IFN-γ secreting CD8 T cells were largely CD11ahi , indicating that the majority were antigen-experienced T cells ( Fig 4D ) [44] . When comparing M282-specific T lymphocytes to all other CD8 T cells in the lung at day 2 p . i . , half of M282-specific CD8 T cells produced IFN-γ , whereas almost 30% of the remaining CD8 T cells were also producing IFN-γ ( Fig 4E ) . These results indicate that both M282-specific and bystander antigen-experienced CD8 T cells secrete IFN-γ in immunized mice early following RSV infection . Lastly , we assessed IFN-γ production by CD8 T cells in the respiratory tract ( lung and BAL ) , mediastinal lymph node ( mLN ) , and periphery ( spleen and PBL ) . The majority of CD8 T cells secreting IFN-γ were localized to the lung and BAL ( Fig 4F ) . In contrast , a relatively low frequency of CD8 T cells produced IFN-γ in the spleen and PBL , and virtually no IFN-γ production was observed in the mLN ( Fig 4F ) . Taken together , these data demonstrate that antigen-experienced CD8 T cells in the airways of immunized mice secrete IFN-γ early following RSV infection leading to increased IFN-γ levels in both the lung and the periphery . Due to the increased systemic IFN-γ levels largely produced by antigen-experienced CD8 T cells in the respiratory tract , we next determined if IFN-γ was necessary to mediate the severe immunopathology in immunized mice . To evaluate the role of IFN-γ , we treated M282-immunized mice with either control IgG or anti-IFN-γ neutralizing antibody administered i . n at the time of RSV challenge . While high mortality was observed in the IgG-treated group , neutralization of IFN-γ led to the survival of all immunized mice ( Fig 5A ) . In addition , weight loss and respiratory dysfunction were significantly ( p<0 . 05 ) reduced in immunized mice following IFN-γ neutralization as compared to the IgG-treated mice ( Fig 5B–5D ) . Neutralization of IFN-γ did not significantly impact virus titers in the lung at day 4 p . i . , suggesting that IFN-γ does not contribute to pathogen clearance in this prime-boost immunization model ( Fig 5E ) . Neutralization of IFN-γ resulted in significantly ( p<0 . 001 ) decreased TNF levels in the lung at day 2 p . i . compared to IgG-treated controls ( Fig 5F ) . Overall , our results suggest that antigen-experienced CD8 T cells rapidly secrete IFN-γ , which mediates lethal immunopathology following RSV infection in DC-LM-immunized mice by promoting the production of TNF by other cell populations .
Current RSV vaccine development and assessment is focused upon induction of a strong humoral immune response [45] . In contrast , the capacity of cellular immunity to provide protection against an RSV infection has received less attention . Here we evaluated the capacity of memory CD8 T lymphocytes to protect against an RSV infection . Our results demonstrate that memory CD8 T cells , in the absence of RSV-specific CD4 T cell memory or antibodies , promote enhanced viral clearance following RSV challenge . However , pre-existing RSV-specific memory CD8 T cells also mediate exacerbated disease severity and lethal immunopathology . The CD8 T cell response has been previously shown to contribute to weight loss and illness following an acute RSV infection [12] . Therefore , the CD8 T cell response plays a crucial role in both viral clearance and immunopathology during primary and secondary responses . A number of previous reports have shown that the adoptive transfer of activated effector RSV-specific CD8 T cells , in vitro stimulated T cell lines , or in vitro propagated T cell clones results in enhanced RSV clearance from the lung following RSV challenge . These effector CD8 T cell transfers were also associated with increased weight loss , indicating that infusion of effector CD8 T cells leads to the induction of increased systemic disease [20–23] . Thus , our studies contrast substantially with these previous studies as we have examined the protective capacity of in vivo generated RSV-specific memory CD8 T cells that have not undergone in vitro activation or restimulation . Vallbracht et al . reported that mutation of the M282 epitope sequence within the RSV genome results in reduced T cell-mediated immunopathology following an acute infection [32] . However , the enhanced disease severity associated with pre-existing RSV-specific memory CD8 T cells we observe here was not limited to either a specific epitope or RSV protein . DC-LM immunization targeting either the M282 or F85 epitope resulted in exacerbated disease and high mortality following RSV challenge . Furthermore , immunization against the immunodominant M187 epitope in C57BL/6 mice also caused increased weight loss and pulmonary dysfunction , but no mortality following RSV infection . The lack of mortality in the immunized C57BL/6 mice may be due to M187-specific CD8 T lymphocytes having superior cytolytic function with limited immunopathology as compared to M282-specific CD8 T cells [46] . Interestingly , the lethal immunopathology associated with pre-existing memory CD8 T cells was unique to the context of an RSV infection . Memory CD8 T cells are protective and do not exacerbate disease severity with other respiratory viral infections such as IAV or SARS-CoV [25 , 26] . Consistent with memory CD8 T cells being able to provide protection against IAV infection [25] , we show that RSV M282-specific memory CD8 T cells mediate protection following a lethal IAV-M282 infection . A systemic increase of IFN-γ was seen in immunized mice early 2 days following RSV infection . Antigen-experienced CD8 T cells in the airways were the primary source of IFN-γ . The presence of IFN-γ-producing CD8 T cells only in the airways suggests that production was in response to antigen stimulation given the strong tropism of RSV to the epithelium and alveolar cells of the respiratory tract [47–49] . Interestingly , peak IFN-γ production occurred prior to the significant accumulation and/or expansion of virus-specific CD8 T cells within the lung by day 5 p . i . This observation is similar to work by Liu et al . showing IFN-γ production by CD8 T cells prior to expansion of the CD8 T cell response following lymphocytic choriomeningitis virus ( LCMV ) infection [43] . Neutralization of IFN-γ prevented all mortality and reduced disease severity in immunized mice without impacting RSV titers . These data indicate that IFN-γ promotes the severe immunopathology , but does not contribute to viral clearance . A robust IFN-γ response may explain why memory CD8 T cells protect against a lethal IAV-M282 and not an RSV infection . Antibody-mediated neutralization of IFN-γ does not impact RSV viral titers during an acute infection indicating a minor role in viral clearance [42] . Induction of a Th1-biased immune response is important to prevent the pathology associated with a Th2-skewed response [42 , 50–52] . Nonetheless , IFN-γ still contributes to minimal immunopathology associated with an acute RSV infection [42 , 52] . However , treatment of mice with recombinant IFN-γ at early timepoints has been shown to protect against lethal IAV infection with no difference in viral clearance [53] . This result is consistent with our study showing that DC-LM immunized mice receiving a lethal IAV infection have significantly increased IFN-γ protein levels in the lung at early time points as compared to RSV challenged mice . Together , these results suggest that supplemental IFN-γ limits the severe pathology associated with a lethal IAV infection without contributing to viral clearance . Therefore , IFN-γ can play distinct roles by either contributing to immunopathology or ameliorating disease dependent upon the respiratory virus infection . Other CD8 T cell functions were not required to mediate the exacerbated disease observed in DC-LM-immunized mice . Immunized mice deficient in perforin exhibited similar disease severity as WT controls following RSV challenge . In addition , all perforin-deficient mice eventually succumbed to infection whereas typically 10–20% of WT immunized mice survive . It has been previously demonstrated that IFN-γ and TNF levels in perforin-deficient mice are elevated following an acute RSV infection , which likely contributes to the accelerated mortality we observed in M282-immunized perforin-deficient mice [42] . A role for regulating CD8 T cell expansion and cytokine production by perforin to prevent mortality has also been demonstrated for secondary CD8 T cell responses against LCMV [54 , 55] . Lastly , the neutralization of TNF also improved survival and ameliorated disease in immunized mice following RSV infection . The cytokine TNF is known to contribute to both weight loss and inflammation during acute RSV infection [42] . Therefore , TNF plays a critical role in mediating immunopathology during both primary and secondary RSV infections . Our results are in contrast to work by Lee et al . showing enhanced RSV clearance with reduced disease severity mediated by vaccine-elicited memory CD8 T cells [56] . The DC-LM immunization utilized in our studies resulted in dramatically increased numbers of RSV-specific memory T cells compared to the immunization strategy employed by Lee et al . [56] . This disparity in RSV-specific memory CD8 T cell numbers prior to RSV challenge may account for the observed difference in outcome between our studies . It remains controversial whether a strong CD8 T cell response is desirable during RSV infection in humans . Experimental RSV infection of adult humans revealed that greater frequencies of virus-specific CD8 T cells in the BAL correlated with reduced clinical disease symptoms [57] . In contrast , a greater ratio of CD8 to CD4 T cells in the airways is associated with acute lung injury during common respiratory tract infections such as RSV [58] . RSV-specific CD8 T cell responses have also been examined in humans following RSV infection . Consistent with their critical role in clearing an acute infection , a study by Weilliver et al . found that children with a fatal primary RSV infection had fewer CD8 T cells in their lung tissue than normal controls suggesting that patients with severe lower respiratory tract illness may have an insufficient cell-mediated immune response [59] . In contrast , a report by Heidema et al . examining CD8 T cells in the airways of infants with a severe primary RSV infection was able to readily identify RSV-specific CD8 T cells in the airways with no significant difference in the number recovered between infants with more versus less severe disease [60] . Thus , it is currently unclear if CD8 T cells contribute to immunopathology during an acute RSV infection in infants . The role of RSV-specific memory CD8 T cells in humans has been difficult to address due to the difficulty in obtaining CD8 T cells from the airways . A recent report by Jozwik et al . using the human RSV challenge model has sought to address this issue . In adult volunteers experimentally infected with RSV , Jozwik et al . found that a higher baseline frequency of RSV-specific CD8 T cells in the airways correlated with a lower cumulative symptom score following RSV challenge [57] . Our results obtained by administration of a local IAV-M282 boost are consistent with this notion . Our data suggest that systemic RSV-specific memory CD8 T cells are more prone to causing immunopathology , possibly in part due to their delay in reaching the lungs . However , the use of the human RSV challenge model cannot evaluate the protective capacity of memory CD8 T cells in the absence of RSV-specific memory CD4 T cells and antibodies as we have done here using our animal model . Thus , we believe our results are applicable to humans in cases where vaccination of an RSV seronegative individual would primarily elicit a CD8 T cell response . Our study defines a clear role for memory CD8 T cells following RSV infection . Pre-existing CD8 T cell memory contributes to enhanced viral clearance upon RSV challenge , but also mediates severe immunopathology in contrast to many other viral infections . The outcome is compelling given the high rate of mortality , which is unusual in this RSV infection model . These data highlight how complex and unique the RSV-induced immune response is in contrast to other respiratory viral infections . Our results indicate that epitope-based cellular vaccines against RSV may have detrimental consequences . Our data also support that caution must be exercised during evaluation of any RSV vaccine candidate , particularly when robust memory T cell responses are involved in order to prevent the induction of immunopathology .
Female BALB/cAnNCr mice between 6–8 wk old were purchased from the National Cancer Institute ( Frederick , MD ) . Female H-2d perforin-deficient mice were provided by Dr . John Harty ( University of Iowa , Iowa City , IA ) [61 , 62] . All experimental procedures utilizing mice were approved by the University of Iowa Animal Care and Use Committee under Animal Protocols #4101196 and #7041999 . The experiments were performed under strict accordance to the Office of Laboratory Animal Welfare guidelines and the PHS Policy on Humane Care and Use of Laboratory Animals . Memory CD8 T cells were induced using a DC-LM , prime-boost immunization regimen . BALB/c mice were injected intraperitoneally ( i . p . ) with 5 x 106 B16 melanoma cells that express fms-related tyrosine kinase 3 ligand ( B16-FLT3L ) . After 14 days , mice were injected i . v . with 1–2 μg lipopolysaccharide ( LPS ) to mature DCs . 24 hrs later , spleens were harvested and digested in HBSS containing 60 U/mL DNase I ( Sigma-Aldrich ) and 125 U/mL collagenase ( Invitrogen ) while gently shaking for 20 mins at 37°C . Spleens were made into single-cell suspensions and incubated with a 2 μM concentration of either M282-90 or F85-93 peptide for 2 hrs at 37°C while rocking . DCs were isolated using anti-CD11c microbeads ( Miltenyi Biotec ) and sorted via positive selection on an autoMACS separator ( Miltenyi Biotec ) . Mice were primed with 5 x 105 peptide-pulsed DCs . DC-immunized mice were boosted with 5 x 106 actA-deficient LM that express either M282 or F85 administered i . v . 7 days later . 28–42 days following the LM boost , mice were infected with either RSV or IAV-M282 . Control mice were primed with DCs incubated without peptide and boosted with an actA-deficient LM that does not express any RSV-derived epitopes [63] . The recombinant LM strains were created using pPL2 integration vector [64] . Target DNA was inserted at digested BamH1 and PstI sites and ligated in Escherichia coli XL1-Blue cells . Recombinant chloramphenicol-resistant plasmids were conjugated in E . coli SM10 cells along with the 10403S strain of LM that is resistant to streptomycin [65] on brain heart infusion agar plates . Growth from previous step were streaked out on selective brain heart infusion agar plates to select chloramphenicol- and streptomycin-resistant colonies that contain pPL2 integrated into the 10403S LM strain . Recombinant LM were grown in tryptic soy broth ( 35 . 6g/L ) containing 50 mg/mL streptomycin . The A2 strain of RSV was a gift from Dr . Barney Graham ( National Institutes of Health , Bethesda , MD ) . The A2-line19F strain was a gift from Dr . Martin Moore ( Emory University , Atlanta , GA ) . RSV strains were propagated in HEp-2 cells ( ATCC ) . Mice were infected i . n . with 1 . 0–1 . 7 x 106 PFU of purified RSV . For RSV purification , 50% polyethylene glycol was added to crude RSV for a final dilution of 1:5 . The RSV preparation was mixed at 4°C for 2 hrs and centrifuged at 7300 g for 30 mins in a swing bucket rotor . Pellets were resuspended in 20% sucrose solution and placed on top of 60% and 35% sucrose layers and centrifuged at 170 , 000 g for 1 hr . Purified RSV at the interface between the 35% and 60% sucrose layers was collected and stored at -80°C . All solutions were created in a buffer containing 0 . 15 M NaCl , 0 . 05M Tris-HCl , and 0 . 001M EDTA . For mock infections , mice were administered an equivalent volume of sterile PBS . Recombinant IAV-M282 was kindly provided by Dr . Ryan Langlois ( University of Minnesota , Minneapolis , MN ) . The virus was created using standard reverse genetics as previously described [66] , rescued , and grown in 10 day-old embryonated chicken eggs ( Charles River ) . M282 epitope was inserted into the mRNA nucleotide position 186 encoding the neuraminidase stalk region , which is known to tolerate such insertions [67] . For lethal heterologous IAV infections , mice were challenged i . n . with a 5 LD50 dose representing 1 x 103 PFU of recombinant IAV-M282 virus . For sublethal IAV infections , mice were infected i . n . with a 0 . 1 LD50 dose representing 20 PFU of IAV-M282 . In certain experiments , mice were boosted with IAV-M282 and given a 0 . 1 LD50 dose i . n . 7 days following the DC-M282 prime i . v . Whole lungs were harvested from mice , weighed , mechanically homogenized , and supernatant was stored at -80°C until further use . 1:10 serial dilutions of supernatants were performed and incubated on Vero cells ( ATCC ) in 6-well plates for 90 mins at 37°C . Plates were overlaid with a 1:1 mixture of 2X Eagle minimum essential medium ( 2X EMEM , Lonza , Walkersville , MD ) and 1% SeaKem ME agarose ( Cambrex , North Brunswick , NJ ) . Following 5 days of incubation at 37°C , 5% CO2 , plates were stained with a 1:1 mixture of 2X EMEM and 1% agarose containing 0 . 015% neutral red ( Sigma-Aldrich ) . Plaques were counted after 24–48 hrs . For determination of IAV titers , lungs were processed in the same manner as for RSV plaque assay . MDCK ( ATCC ) cells in 6-well plates were washed 3 times with room temperature sterile PBS adding 1 mL of sterile Dulbecco’s modified Eagle’s medium afterwards . Plates were infected with 100 μl of serially diluted IAV-infected lung samples ( 10-fold dilutions ) for 1 hr at 37°C . Plates were washed twice with sterile room temperature PBS . Wells were overlaid with 2 mL of a 1:1 mixture of 2X EMEM and 1 . 6% agarose containing 1 mg/mL TPCK-trypsin and incubated at 37°C , 5% CO2 for 3 days . Agarose plugs were carefully removed , and monolayers were fixed with 2 mL 70% ethanol for 20 mins at room temperature . Monolayers were stained with 1 mL of 1% crystal violet in methanol for 10 mins at room temperature , and plates were washed in a pool of warm water . Plates were allowed to dry overnight , and plaques were counted the next morning . Pulmonary function of mice was evaluated using unrestrained whole-body plethysmography . Enhanced pause ( Penh ) and respiratory minute volume ( MVb ) were measured using a whole-body plethysmograph ( Buxco Electronics , Wilmington , NC ) and averaged over a 5 min period . Weight loss was tracked daily following RSV or IAV infection of mice . Mice that were at or below 70% of their starting weight were euthanized . For IFN-γ and TNF neutralization , mice were treated i . n . with 200 μg of anti-IFN-γ ( clone XMG1 . 2 ) or anti-TNF ( clone MP6-XT22 ) antibody during RSV challenge . For controls , mice were administered a matching dose of control isotype IgG antibody . Whole lungs were harvested on day 5 following RSV challenge and fixed in 10% neutral buffered formalin ( Fisher Scientific ) . Lungs were processed as previously described [68] and stained with H&E for routine evaluation . Representative images of lung sections were taken at 20X , 200X , and 400X magnification for each immunization regimen . Tissues were examined and scored in a manner masked to experiment groups [69] . Each sample was assessed for evidence of DAD . Histopathologically , early stages of DAD include alveolar septal injury , such as cellular sloughing , necrosis , hyaline membrane formation , hemorrhage , and early cellular infiltrates . DAD scores were assigned as follows: 1—absence of cellular sloughing and necrosis; 2—Uncommon solitary cell sloughing and necrosis; 3—Multifocal cellular sloughing and necrosis with uncommon septal wall hyalinization; 4—Multifocal cellular sloughing and necrosis with common and prominent hyaline membranes . Serum was collected and whole lungs were harvested on days 0 , 2 , and 4 p . i . Lungs were disrupted using a tissue homogenizer ( Ultra-Turrax T25; IKA Works , Inc . , Wilmington , NC ) in Cell Lysis Buffer ( eBioscience ) . Lung homogenates were centrifuged at 2000 rpm for 10 mins , and supernatants were collected . The protein levels of 20 different cytokines and chemokines in the lung and serum were determined using a ProcartaPlex Multiplex Immunoassay kit ( eBioscience ) according to the manufacturer’s instructions . The assay was run on a BioPlex instrument ( Bio-Rad , Hercules , CA ) . Lung and serum IFN-γ levels were determined by ELISA as previously described ( eBioscience ) [68] . Lung TNF levels were determined using a mouse TNF ELISA kit ( Invitrogen ) according to manufacturer’s instructions . Lung and BAL were harvested from mice as previously described [70 , 71] . Spleens and mLN were gently dissociated between the frosted ends of microscope slides . Cells from the lung , BAL , spleen , mLN , and PBL were stained for extracellular surface molecules with antibodies specific to CD11c ( clone N418 ) , Siglec F ( BD Biosciences , clone E50-2440 ) , F4/80 ( clone BM8 ) , Ly6c ( clone HK1 . 4 ) , Ly6g ( clone 1A8 ) , CD49b ( clone DX5 ) , NKp46 ( clone 29A1 . 4 ) , CD11a ( clone M17/4 ) , CD90 . 2 ( clone 53–2 . 1 ) , CD3ε ( clone 145-2C11 ) , CD4 ( clone GK1 . 5 ) , CD8 ( clone 53–6 . 7 ) , CD69 ( clone H1 . 2F3 ) , and CD103 ( clone 2E7 ) for 30 mins at 4°C and fixed with fix/lyse solution ( eBioscience ) for 10 mins at room temperature . After extracellular staining , cells were stained for FoxP3 ( eBioscience clone FJK-16s ) with transcription factor staining buffer set ( eBioscience ) according to manufacturer’s instructions . For intracellular cytokine staining , cells were stimulated for 5 hrs at 37°C with 2 μM M282-90 peptide in 10% FCS-supplemented RPMI . Stimulated cells were stained for surface markers as indicated above and then stained intracellularly with antibodies specific to IFN-γ ( clone XMG1 . 2 ) and TNF ( clone MP6-XT22 ) in FACS buffer containing 0 . 5% saponin ( Sigma-Aldrich ) for 30 mins at 4°C . Total numbers of cytokine producing cells were calculated after subtraction of background staining from BFA-only controls . All monoclonal antibodies were purchased from BioLegend unless otherwise stated . Stained cells were run on LSRFortessa and analyzed with FlowJo ( Tree Star , Ashland , OR ) software . Cell types were phenotyped as follows: CD8 T cells ( CD90 . 2+CD8+ ) , CD4 T cells ( CD90 . 2+CD4+ ) , Tregs ( CD90 . 2+CD4+FoxP3+ ) , NK cells ( CD3ε-CD49b+NKp46+ ) , monocytes ( CD11c+F4/80+ ) , eosinophils ( SiglecF+CD11clo ) , and neutrophils ( Siglec F-CD11c-Ly6c+Ly6g+ ) . Mice were injected i . v . with 1 μg CD45-FITC ( CD45 labeled with fluorescein isothiocyanate ) ( clone 30-F11 ) antibody 3 mins prior to euthanasia . Cells from the lung were processed as previously described [37] . Analysis of IFN-γ-producing cells was performed using in vivo BFA Administration [43] . Mice were injected i . v . with 250 μg BFA ( 0 . 5 mg/mL; Sigma ) in 500 μl PBS , and lungs , BAL , spleen , mLN , and PBL were harvested 6 hrs later . Leukocytes were stained as indicated above . All statistical analyses are described in each figure legend and were performed using Prism software ( GraphPad Software , San Diego , CA ) . Data were evaluated using unpaired , two-tailed Student’s t tests between two groups or one-way ANOVA with Tukey-Kramer post-test analyses for more than two groups to determine if there was a statistical significance of at least α = 0 . 05 . Asterisks or pound signs are used to define a difference of statistical significance between the indicated group and its respective control group unless otherwise indicated by a line or stated in the figure legend . | Memory CD8 T cells have been shown to provide protection against many respiratory viruses . However , the ability of memory CD8 T cells to provide protection against RSV has not been extensively examined . Unexpectedly , mice with pre-existing CD8 T cell memory , in the absence of memory CD4 T cells and antibodies , exhibited exacerbated morbidity and mortality following RSV infection . We demonstrate that the immunopathology is the result of early and excessive production of IFN-γ by memory CD8 T cells in the lung . Our research provides important new insight into the mechanisms of how memory T cells induce immunopathology . In addition , our findings serve as an important cautionary tale against the use of epitope-based T cell vaccines against RSV . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"respiratory",
"infections",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"microbiology",
"cloning",
"orthomyxoviruses",
"pulmonology",
"viruses",
"clinical",
"medici... | 2018 | Memory CD8 T cells mediate severe immunopathology following respiratory syncytial virus infection |
Study of human executive function focuses on our ability to represent cognitive rules independently of stimulus or response modality . However , recent findings suggest that executive functions cannot be modularized separately from perceptual and motor systems , and that they instead scaffold on top of motor action selection . Here we investigate whether patterns of motor demands influence how participants choose to implement abstract rule structures . In a learning task that requires integrating two stimulus dimensions for determining appropriate responses , subjects typically structure the problem hierarchically , using one dimension to cue the task-set and the other to cue the response given the task-set . However , the choice of which dimension to use at each level can be arbitrary . We hypothesized that the specific structure subjects adopt would be constrained by the motor patterns afforded within each rule . Across four independent data-sets , we show that subjects create rule structures that afford motor clustering , preferring structures in which adjacent motor actions are valid within each task-set . In a fifth data-set using instructed rules , this bias was strong enough to counteract the well-known task switch-cost when instructions were incongruent with motor clustering . Computational simulations confirm that observed biases can be explained by leveraging overlap in cortical motor representations to improve outcome prediction and hence infer the structure to be learned . These results highlight the importance of sensorimotor constraints in abstract rule formation and shed light on why humans have strong biases to invent structure even when it does not exist .
Making decisions in the complex environment of daily life often requires cognitive control to flexibly adjust our behavior: the appropriate reaction to different sensory events often depends on the current context , our goals , etc . For example , while driving , if you see a red light on your street , you will stop; if it turns green , you will go–but the opposite actions apply when the red/green light is in the intersecting street . The cognitive control literature has shown that humans use contextual cues to select abstract sets of rules ( or task-sets ) that are behaviorally relevant [1 , 2] . Specifically , cognitive control relies on a cascading hierarchical structure , where at a more abstract level , subjects choose task-sets appropriate to the context , which then constrain our action choices in response to lower-level , less abstract stimuli [3 , 4] . Thus , we use the term “context” to refer to those features that cue abstract task-set , and the term “stimulus” to refer to features that cue the appropriate response conditioned on the selected task-set . In the study of executive functions , researchers tend to focus on discretized aspects of decision making–often assuming that perception systems have transformed complex , multi-dimensional sensory signals into reduced , discrete stimuli ( e . g . , a red circle ) , and that given this percept , the executive system selects among a few discrete options ( e . g . , left vs . right ) , which are then implemented by the motor system . However , the field of embodied or grounded cognition [5–7] offers strong hints that this model is over-simplified , emphasizing instead that executive functions evolved for the control of action in continuous time [5] , and are thus scaffolded on existing sensorimotor processing systems . One hint that executive functions are grounded in sensorimotor processing [8 , 9] , is the inability to isolate functionally distinct neural systems for executive functions from motor execution , such that for example , cerebellum , mostly thought of as a motor system , is strongly involved in cognitive control [10 , 11] . Another hint comes from the organization of cortico-basal ganglia loops , where prefrontal-striatal connectivity parallels that of premotor-cortex for action selection , with hierarchical influence of more anterior loops representing cognitive rules over more posterior loops involved in selecting motor actions [12–15] . Through learning , the consequences of motor actions are leveraged not only to improve future motor choices , but also to improve selection of higher order rules [12–16] again implying a scaffolding of more abstract “cognitive” action selection onto motor action selection . Similarly , striatal dopamine manipulations have analogous effects on reinforcement learning and cognitive action selection [17 , 18] . However , to date , we know of no direct evidence to indicate that constraints within the motor system influence the way our brain represents the more abstract variables needed for cognitive control , such as task-sets . Here , we propose that an intrinsic constraint of motor action representations strongly influences how we create representations of abstract hierarchical rule structures , both during learning and while applying instructed rules . Recent research has shown that humans can acquire abstract rule representations through reinforcement learning , using only reward feedback . Structured rules are discovered when they are present in the task and helpful for learning and generalization [19–21] , and there is evidence that subjects look for and create such structure , even when it is not beneficial [15 , 16 , 22] . However , creating structure is potentially effortful and costly: a learner needs not only to discover the appropriate set of rules given a defined set of “higher-level” contexts that cue these rules and “lower-level” stimuli that cue motor actions , but also how to organize the complex multidimensional world into structured components ( contexts , stimuli ) in the first place . Indeed , more than one hierarchical structure representation of the environment is possible and it is not always evident which features should be contexts cuing task-sets and which should be considered lower level stimuli . Often , one hierarchical structure is more useful than others because it can afford greater generalization to novel contexts ( e . g . , if multiple contexts cue the same task-set , they can be clustered together [15] , Collins & Frank , submitted ) . For example , the rule to apply for red and green traffic lights is the same whether the intersecting street comes from the left or the right , such that using the position for context allows generalization , and thus simplifies the problem . Here , we investigated whether motor patterns provide an additional constraint on imposing hierarchical cognitive rule structures . To this effect , we take advantage of motor representational structure in motor cortex . Finger movement representations in M1 , while globally somatotopic , are widely overlapping for neighboring fingers [23 , 24] , reflecting the natural statistics of hand movement structure [24] . This representational constraint may lead to pre-activation of neural networks representing frequently co-activated fingers ( motor synergies [25 , 26] ) , thus facilitating further choices with those fingers . For example , cues that allow for motor preparation of adjacent , rather than segregated , finger presses facilitate action [27–29] . Therefore , we investigated whether this motor constraint could influence the nature of hierarchical rule structures when actions are selected via finger presses on a computer keyboard . Specifically , we hypothesized that subjects might naturally cluster together sets of stimulus-action associations that were more similar in motor space , for example , that involved adjacent fingers , within each task-set . This would then lead to a natural constraint on the creation of abstract rule structure: subjects should treat those feature dimensions that afford motor clustering as lower order stimulus ( allowing clustering within rules ) , and those that afford less motor clustering as higher order context . To test our hypothesis , we tested subjects in a structure-learning task , in which we previously showed that subjects create one of two possible hierarchical structures , even when they could learn the task as well without imposing structure at all . We show in two new data sets that the potential for motor clustering within a task-set influences the nature of the structure created . We further confirm and replicate this finding in a novel re-analysis of two published data-sets [15 , 22] . We then use computational modeling to investigate how these biases may emerge . Simulations show that assuming that low level motor representational biases simply influence action selection is not sufficient to account for structure learning biases observed . Rather , we implement a simple computational mechanism based on the motor scaffolding hypothesis that assumes that motor representational constraints affect outcome predictions , and thus inference of structure . We show that this mechanism can account for subject’s biases in structure learning . Finally , we show that this bias may be strong enough to induce subjects to restructure rules , even if they do not need to learn them , but know them through practice and instructions , when such rules are set up such that they conflict with motor clustering constraints .
In a first experiment , 22 subjects performed 6 independent blocks of a reinforcement learning task shown previously to induce creation of hierarchical rule structures [15 , 22] . Specifically , they used truthful correct/incorrect feedback to learn to select the correct one out of four possible actions ( button presses with four fingers of the dominant hand ) for each of four distinct visual input patterns ( colored shapes , see Figs 1A and 2 ) , presented in pseudo-randomized order . Although such problems could be learned by standard learning algorithms assuming that each input pattern is a distinct state that simply combines color and shape , we showed in previous research [15 , 22 , 30] that subjects impose structure onto such learning problems using hierarchical rules , such that one dimension ( e . g . , color in Fig 2B ) serves as a context that cues a task-set , and the other ( e . g . , shape in Fig 2B ) as a stimulus that directs action selection under the constraint of the current task-set . Imposing such structure facilitates transfer of learned task-sets to novel contexts [15 , 22 , 30] and speeds learning when multiple contexts cue the same task-set . Fig 3 shows that two structures are possible for a single learning problem ( e . g . with colors as the contexts cueing task-sets , or shapes as the contexts cueing task-sets ) . Depending on the specific mapping of actions onto four fingers of the right hand , a given structure may elicit more or less efficient motor clustering patterns; two contrasting examples are given in Fig 2 , each when color is used as context . We define a measure based on the motor representation biases introduced above: the adjacency bonus indicated whether a task-set cue clustered motor action ( if actions for both stimuli require key presses from adjacent fingers within each task-set; see Fig 2 and Methods ) . For a fixed assignment of actions to fingers , contrasting the two potential hierarchical structures ( color vs . shape structure ) leads to different motor patterns: P1 vs . P2 , or P3 vs . P4 ( Fig 3 ) . We test here the theory that subjects tend to create the structure associated with greatest affordance for motor clustering , and specifically hypothesized that subjects would select structures P1 over P2 ( a = 2 vs . 0 ) , and P3 over P4 ( a = 1 vs . 0 ) . To identify which structure was created , we relied on reaction-time task-switch costs , one of the most reliable findings in cognitive psychology [1] . Due to the additional cognitive demands associated with updating task-sets ( thought to relate to processing within prefrontal cortex ) , response times are slower when the task-set changes from the previous trial , compared to when it repeats , and this slowing is greater than the corresponding cost associated with changes in stimuli within a task-set [1 , 15]; reaction time switch-cost measures this relative slowing . This measure allows us to differentiate the structures: those subjects adopting color-on-top structures should exhibit greater RT switch costs for switches in color than shape , and vice versa . We thus use the difference between reaction time switch-cost of each dimension as evidence in favor of either structure . We have previously validated this measure in that it was independently predictive of subjects’ ability to transfer the inferred structure to novel contexts that involve the same structure , and was also predictive of neural signals reflecting the same hierarchical structure [15 , 22] . To evaluate whether the selected structure was affected by potential for motor clustering , we define the motor clustering switch-cost as the difference between the switch costs defined by the structure affording greater motor clustering and that affording less clustering , e . g . , switch cost for P1 or P3 minus the switch-cost for P2 or P4 . Thus , our prediction that subjects would be more likely to create the hierarchical rule structure that affords motor clustering translated into a prediction that the motor clustering switch-cost should be positive . Results confirmed our predictions: we found that Motor clustering RT switch cost was significantly greater than 0 ( Fig 4 left , p = 0 . 0007 , t ( 21 ) = 4 . 14 ) . This measure did not differ between configuration types P1/2 and P3/4 ( t = 0 . 36 , ns ) . Furthermore , individual learning problems were significantly more likely to be identified as subjects having created the structure affording more motor clustering ( P1-3 ) than the one affording less ( P2-4 , binomial test p = 0 . 0012; 66 out of 99 problems; Fig 4 right ) . We sought to replicate this result in three further data-sets , using nearly identical experimental designs ( two of which were previously published but where motor patterns were randomized and not assessed [15 , 22] ) , with only one learning block per subject . Identical analysis performed over all three datasets confirmed the previous results: reaction time-switch cost was significantly higher for the structure with higher adjacency bonus , measured by a positive motor-clustering switch-cost ( t ( 91 ) = 3 . 96; p = 0 . 00015; S4 Fig ) . There was no effect of experiment on motor-clustering switch-cost ( F = 0 . 34 , p = 0 . 71 ) , nor was there an effect of configuration type ( P1/3 vs . P2/4; p = 0 . 13; S4 Fig ) . Furthermore , 61 ( vs . 31 ) out of 92 subjects had a positive motor-clustering switch-cost , labeling them as having created the higher adjacency bonus structure ( S4 Fig , right ) . This is significantly more than expected by a balanced random distribution ( p = 0 . 0023 , binomial test ) , supporting our previous finding . We thus observed across four independent data sets that subjects were more likely to create hierarchical task structure that lead to grouped task sets , highlighting an influence of motor choices on abstract representations known to be crucial for executive functions [1 , 31] . Note that this bias has no bearing on the subjects’ actual motor choices: independently of how they implement the task , the sequence of valid motor actions to take is the same . Thus , we believe that this phenomenon emerges from strong biases highlighting constraints in how our brain represents rules . We next investigate the mechanisms that could lead to those biases . We previously proposed a computational model of hierarchical structure learning [15] . The model simulates creation of task-set structure , allowing for clustering of multiple contexts that link to task-sets and which guide lower order stimulus-action-outcome associations . Moreover , it can infer the type of structure , e . g . which features are indicative of contexts that cue latent task sets and which should be considered low order stimuli . It does so by considering which structure is best able to predict the observed outcomes ( here , reinforcement feedback ) and assumes that structure describes the current environment most efficiently , and allowing for transfer of structures to novel contexts . Recall that by design , the experimental paradigm used here can be equally well described by either structure , with either color or shape acting as context cueing task-sets , and as such our previous model did not distinguish between them during learning [15] . Here we augmented the model to investigate the potential influence of motor clustering mechanisms ( e . g . , the overlap in encoding of motor actions by distributed neuronal populations [23 , 24] ) that could explain the structure patterns observed empirically . Specifically , the model assumes that the hierarchical selection of a task-set constrains pre-activation of action representations available within this task-set , allowing the model to account for specific error patterns and hierarchical neural signals that are explained by such processing [15 , 22] . We now further assume that this activation spreads to adjacent motor actions , given their representational overlap ( Fig 5A ) , and can bias outcome prediction within the selected task-set . Note that this proposed mechanism is local , in that it relies on the overlap of neuronal populations encoding adjacent fingers; but it requires that this overlap is used such that it can influence selection of the higher level task-set: implying a motor constraint on cognitive rule structure . We simulated two versions of models including this bias: in one case , the bias was used only to influence action selection ( comprising a pure motor bias during choice ) ; in the other case the motor overlap was also used to predict outcomes of selected actions ( see supplement for details ) . Simulation results of the former showed that it could not account for any structure preference: inferred weights were in average equal for both structures ( wC = wS = 0 . 5 ) . Indeed , a pure motor bias only affects which keys are pressed and does not lead to inference that one structure is better than the other , and hence predicts no asymmetry in switch costs , unlike the empirical data . In contrast , when we allowed the overlap in motor representations to also influence predictions ( Fig 5 –right ) this simple local mechanism allowed our model to infer that the best structure was that which afforded greater motor clustering ( P1>P2 , P3>P4 ) . Indeed , it allowed those task-sets involving adjacent motor actions to better predict outcomes and hence the model identified that structure as better fitting the environment . Overall , these simulations confirm that scaffolding of higher-level abstract rule learning on low level motor representation can lead to biases in abstract structure representation , as observed empirically . The above findings showed that subjects favor hierarchical rule structures that lead to heuristically simple motor patterns within each rule , in accordance with patterns of motor cortical representations . We next asked whether this representational bias is strong enough that it would influence subjects’ task representation even when learning was unnecessary and task structures are instructed . To this end , a separate set of subjects performed an instructed task-switching experiment in which they were told which dimension indicates the task-set and which constitutes the stimulus . Subjects were instructed the specific rules corresponding to one of each motor configuration pattern structure ( groups for P1 , P2 , P3 and P4 ) , and practiced those rules in a way that shaped the instructed structure ( see Methods , Fig 6A ) . We were interested whether subjects that were instructed a rule that did not afford motor clustering ( groups P2 and P4 ) would show signs of restructuring those rules to represent them instead within the corresponding motor clustering structure . If subjects followed the instructed representations , we expected that they should exhibit the standard reaction-time switch-cost corresponding to that instructed structure . In contrast , if the representational bias was strong enough , we expected that they would no longer show the classical instructed task-set switch cost because it would be offset ( or even reversed ) by the motor biases . Results showed that subjects in the P1/P3 group ( allowing for motor clustering ) had a significant instructed switch-cost ( p = 0 . 002 , t ( 11 ) = 3 . 99 ) . In contrast , subjects in the P2/P4 group which did not afford motor clustering , did not show any instructed switch-cost ( t ( 12 ) = -1 . 14 ) , and this was significantly different from the other group ( p = 0 . 004 , t = -3 . 19 , Fig 6B ) . Thus the classical task-switch cost was abolished when motor clustering favored the alternative structure . Indeed , over the whole group , instructed switch-cost was not significant ( t ( 24 ) = 1 . 05 ) . Instead , the whole group showed a motor clustering switch-cost ( t ( 24 ) = 3 . 08 , p = 0 . 005 ) with no significant difference between the instructed groups ( Fig 6C t = -1 . 36 , p = 0 . 19 ) , lending support to the possibility that rather than performing the task with the instructed representations , subjects might have restructured their representation of the task according to the same motor biases observed in the learning experiment .
When learning to make choices , subjects search for structure and create representations that rely on hierarchical decision trees: at a higher level , they select abstract task-set rules based on contextual cues , which constrain how they select low-level motor actions in response to stimuli . When such a structure matches the statistical regularities of the world , subjects can discover it and leverage that structure to simplify the problem , speeding learning and facilitating transfer [15 , 19 , 32] . However , it is not always evident which structure is best . The results we presented here highlight a reversed hierarchical role in rule structure , where low-level features of the motor choices influence the nature of high-level task-set creation . We showed that a representational constraint of motor actions–adjacency overlap–influences the creation of rule structures: this bias manifests itself as a tendency to create rule structures that afford selection of adjacent motor actions for different stimuli within a same task-set . Indeed , we found across four independent data-sets that subjects were more likely to create a structure that afforded rules that primed adjacent motor actions . We showed in the last experiment that this bias was strong enough to offset classic reaction-time switch costs even in instructed task-switching experiments when the opposite structure afforded greater motor clustering . These findings shed light on internal processes of how abstract rules are created and facilitate choice , even when such structure has no bearing on which motor responses are actually executed . Indeed , no matter what structure is used , the sequence of visual inputs and required motor actions is identical . The only difference elicited by distinct sorts of structure is observable in terms of how switching from one perceived ‘task’ to another elicits a cost , which in turn is related to the ability to transfer this task to novel contexts [15 , 22] . Our predominant hypothesis , indicating that executive functions are rooted in action selection , and that that implementations of decision making are scaffolded on motor circuits such that motor constraints affect cognitive processing , is inspired from a popular subfield of the embodied cognition literature [5–7 , 33] ( though see [34] for other perspectives on what is required to be considered embodied cognition ) . Our neural network model of structure learning [15] hypothesized that task-set selection depended on hierarchical cortico-basal ganglia loops , with a prefrontal loop exerting control over a more posterior motor loop , but feedback from motor choices reciprocally reinforcing task-set selection in higher order loops . Here , we augmented an algorithmic version of this model to represent known features of motor cortex . Our model simulation showed that solely assuming facilitated action preparation could not account for preference in created rule-structures . Rather , our model accounts for the findings by assuming that overlap in motor patterns influences outcome predictions , such that structures involving motor clustering were better at predicting the observed outcomes within task-sets and hence inferred to be valid . Another theory that could potentially account for the observed biases relates to spatial representation–rather than motor selection–biases . Indeed , it is possible that subjects represent the task-sets in a mental one-dimensional space ( much as we represent it in the figures with a row of four squares ) , where a bias for adjacency in motor actions for a task-set would correspond to a spatial grouping bias ( both actions on the left , in the middle , or on the right ) . Previous research in simpler decision-making tasks have shown that both motor and sensory biases could be important [28 , 29 , 35] , particularly in conjunction , as proposed by Adam’s grouping model [28] . Disentangling both contributions here would require either systematically decorrelating the association between motor action selection and spatial action representation , or using neuroimaging to highlight the role motor cortex and overlap in motor representation play here . Both are beyond the scope of this study , but important targets for future research . However , it is important to note that such an interpretation could correspond to a visuo-motor process , involving for example eye movements , as suggested by research in other domains of high-level cognition ( e . g . mental number line [36] ) , indicating that spatial biases do not rule out an embodied cognition interpretation . We have focused here on motor adjacency as a low-level , sensory motor factor that influences abstract high level rule creation . Adjacency is well supported by potential mechanisms and neural data , and thus serves as a first intuitive example to establish this influence , inspired by the embodied cognition literature . However , we do not claim that it is the only factor; indeed other low level factors may also influence rule representation . As a preliminary example , we show in supplementary analysis that beyond similarity in motor space ( summarized here by adjacency ) , similarity in the task-set space , as measured by a parallel or symmetric left-right association between stimuli and fingers , may also provide a bias on rule creation and can be captured by the same model ( where symmetry influences prediction and inference ) . Furthermore , beyond such biases , statistics of the task itself can sometimes constrain which structures are more useful than others , e . g . when they facilitate better generalization ( Collins & Frank submitted , [14] ) . But it is tempting to ask why , when the environment does not constrain abstract rule structure learning , low-level sensorimotor biases may do so . It is possible that this is a pure incidental byproduct of the architecture in which prefrontal-cortex learning and choice scaffolds on motor-based cortico-basal ganglia loops . However , it is also possible that this reflects the result of adaptive pressure: one might imagine that sets of rules used together in the same context tend to require more similar actions , and thus that it would be a priori useful to assume this kind of prior when learning novel structure . While this hypothesis resonates with embodied cognition literature , it remains speculative within the frame of this study . Research in the field of cognitive control tends to think in abstract terms–of stimuli , choices , and values–disembodied from what they are–pictures or sounds , binary key-presses or joystick movements , points or food , assuming that this is processed independently upstream for inputs , and downstream for motor selection . The results presented here show that a very “low-level” feature–exactly with which finger presses choices were made–systematically influences how a very high-level , abstract representation is created , to the point of overwriting trained instructions . This highlights the need to pay careful attention to how sensory-motor factors may bias observed results , and supports the motor scaffolding theories of embodied cognition . It emphasizes how abstract representations that we build for high level-cognition and reasoning likely emerge from the constraints of interacting with connectivity to its input and output regions .
In all tasks , for a given trial , subjects were presented with a single two-dimensional visual pattern on a black screen . There were two possible features on each dimension ( e . g . Green and blue Color , triangle and circle Shapes ) , combining to form four distinct visual input patterns ( Figs 1–3 ) . Subjects could make one of four choices , but only one lead to correct feedback , while the others lead to incorrect feedback . The correct choice was different for each input pattern ( Fig 2A ) . In learning experiments , input-pattern order presentation across trials was pseudo-randomized to ensure equal presentation , and equal frequency of first-order transitions . Subjects were instructed to use the four fingers of the right hand to select an action , except in the R-BL experiment ( see below ) , where they used middle and index fingers of left and right hand to press four contiguous keys . This experiment included twelve independent learning blocks , with non-overlapping sets of stimuli for each block . 6 of those blocks , which are analyzed here , corresponded to a different structure learning problem as defined in “Learning task rules” section , each with different non-overlapping stimuli; they included 80 trials each . 22 subjects performed this experiment , with 40 blocks of configuration P1/2 , 59 of configuration P3/4 and 33 blocks of configuration P5/6 ( not analyzed here ) . 18 subjects had at least one P1/2 block , and all subjects had at least one P3/4 . For switch-cost difference analysis , results were first averaged across blocks within subjects per configuration , before group analysis . For binary structure assignment analysis , we included all 99 blocks of P1/2 and P3/4 configurations independently . To replicate findings from this learning task , we analyzed three further data-sets with similar structure , but only a single block iteration of rule learning: Our model includes two separate “experts” , each of which considers one of the two possible structures ( e . g . color as context expert , vs . shape as context expert , see Fig 5A ) . It then infers the most likely expert that best describes the data as a function of how well each expert predicts observed outcomes over time . Each expert rapidly learns to associate a given context to an abstract latent variable that represents the associated task-set and learns to predict outcomes ( here reinforcement feedback ) , contingent on the current stimulus , chosen action and inferred task-set ( rather than context ) . We first reiterate the model and then describe how we expand it to accommodate motor clustering effects . Specifically , at each trial t , we label Ct and St the observed color and shape , with the chosen action at , and rt the obtained reward outcome . The color-structure expert learns to associate a color context C to an abstract , latent task-set variable Z , by keeping track of the probability of a given task-set given the color , P ( Z|C ) . For a new context Cnew , the prior probability of a given task set is initialized following a Chinese Restaurant Process with concentration parameter α such that: Thus novel contexts are assumed to be most likely linked to existing task-sets that have been most popular across variable contexts , and with some possibility of creating a novel task-set . Following observation of an outcome , the posterior probability P ( Z|C ) is updated according to Bayes rule , using learned likelihood p ( rt|St , at , Zt ) . At each trial , Zt is inferred using maximum a priori . The inferred Zt constrains both the policy that is used for choice during the current trial , and learning of S-a-r contingencies: The shape-structure expert is identical to the color-structure expert , with the roles of color and shape reversed . When both experts are included in the model , they both learn and select actions independently . However , the final action choice proceeds from the mixture of each expert’s policies: π ( a ) = wC ( t ) πC ( a ) + ( 1 − wC ( t ) ) πS ( a ) , where the mixture weight wC ( t ) is the inferred reliability of each expert . It is initialized at wC ( 0 ) = 0 . 5 , and updated via Bayes rule: wC ( t+1 ) =wC ( t ) p ( rt|St , at , ColorStructure ) wC ( t ) p ( rt|St , at , ColorStructure ) + ( 1−wC ( t ) ) p ( rt|St , at , ShapeStructure ) In words , the expert that best predicts observed outcomes is assumed to be best for representing the current environment . To model effects of spatial motor patterns on structure learning , we introduce biases in the outcome prediction of each expert . Specifically , we implement an adjacency bias by assuming that task-set selection may more broadly pre-activate action representations relevant to this task-set , independent of the specific current stimulus , including spreading to adjacent motor actions ( see Fig 5 , left ) . In brief , given that a task-set has been selected , the action-outcome contingencies learned are generalized to neighboring actions such that they become predictive of corresponding outcomes , even for stimuli that have not yet been encountered . This translates into a bias that modifies expected outcomes following: bias ( a|St , Zt ) =∑iπ ( ai|others ( St ) , Zt ) neighbor ( a|ai ) where neighbor ( a|ai ) =1#neighbors ( ai ) if a is a neighbor of ai , 0 otherwise; and others ( St ) indicates the set of stimuli different from St . This bias may be used in a mixture for policy selection and outcome prediction . We show in main text that a model using bias for only policy selection does not account for the empirical results , while a model using it for both does . Biases are mixed to the normal prediction with mixture weight fi ( fi = 0 . 1 ) in simulations . Other model parameters are: We simulate the model 1000 times and report asymptotic preference for either expert . Although we assumed that motor overlap influences predictions , it is also possible that a similar effect results via imperfect credit assignment during learning . That is , when a given action is reinforced , the same overlap in motor representations could elicit partial reinforcement of adjacent motor representations . As such , the network would not only be more likely to select adjacent actions for the same task-set , but also would be better able to learn to predict accurate outcomes for a clustered task-set , leading to both better performance and greater reliability of an adjacent structure . Indeed , model simulations show that this mechanism produces similar results to the one exposed here . Our current findings cannot separate out these hypotheses , though we suspect that both could occur simultaneously ( and indeed both theories rely on motor overlap ) . | Humans’ ability to create abstract rule structures contributes greatly to intelligence and higher cognitive functions as it affords flexible re-use across various sensory-motor transformations . However , how such rule structures develop through learning is poorly understood . Models of this process imply that cognitive rule learning scaffolds on top of mechanisms that support motor action selection learning . Here , we show that the form of motor demands across multiple stimulus-action-outcome influences the form of the abstract rule structures that are created . Our research highlights the strong reciprocal influences of low and high-level cognition , offers further insight into how we learn hierarchical structures , and reminds us that how we act constrains how we think . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"learning",
"medicine",
"and",
"health",
"sciences",
"decision",
"making",
"nervous",
"system",
"motor",
"system",
"fingers",
"experimental",
"design",
"limbs",
"(anatomy)",
"social",
"sciences",
"neuroscience",
"learning",
"and",
"memory",
"motor",
"neurons",
"resear... | 2016 | Motor Demands Constrain Cognitive Rule Structures |
Changes in gene regulation may be important in evolution . However , the evolutionary properties of regulatory mutations are currently poorly understood . This is partly the result of an incomplete annotation of functional regulatory DNA in many species . For example , transcription factor binding sites ( TFBSs ) , a major component of eukaryotic regulatory architecture , are typically short , degenerate , and therefore difficult to differentiate from randomly occurring , nonfunctional sequences . Furthermore , although sites such as TFBSs can be computationally predicted using evolutionary conservation as a criterion , estimates of the true level of selective constraint ( defined as the fraction of strongly deleterious mutations occurring at a locus ) in regulatory regions will , by definition , be upwardly biased in datasets that are a priori evolutionarily conserved . Here we investigate the fitness effects of regulatory mutations using two complementary datasets of human TFBSs that are likely to be relatively free of ascertainment bias with respect to evolutionary conservation but , importantly , are supported by experimental data . The first is a collection of almost >2 , 100 human TFBSs drawn from the literature in the TRANSFAC database , and the second is derived from several recent high-throughput chromatin immunoprecipitation coupled with genomic microarray ( ChIP-chip ) analyses . We also define a set of putative cis-regulatory modules ( pCRMs ) by spatially clustering multiple TFBSs that regulate the same gene . We find that a relatively high proportion ( ∼37% ) of mutations at TFBSs are strongly deleterious , similar to that at a 2-fold degenerate protein-coding site . However , constraint is significantly reduced in human and chimpanzee pCRMS and ChIP-chip sequences , relative to macaques . We estimate that the fraction of regulatory mutations that have been driven to fixation by positive selection in humans is not significantly different from zero . We also find that the level of selective constraint in our TFBSs , pCRMs , and ChIP-chip sequences is negatively correlated with the expression breadth of the regulated gene , whereas the opposite relationship holds at that gene's nonsynonymous and synonymous sites . Finally , we find that the rate of protein evolution in a transcription factor appears to be positively correlated with the breadth of expression of the gene it regulates . Our study suggests that strongly deleterious regulatory mutations are considerably more likely ( 1 . 6-fold ) to occur in tissue-specific than in housekeeping genes , implying that there is a fitness cost to increasing “complexity” of gene expression .
Changes in gene regulation are likely to play an important role in evolution [1] , [2] . However , compared to protein-coding sequences , the fitness effects of regulatory mutations remain poorly understood . Furthermore , the relationship between changes in gene regulatory regions and the expression phenotype of the regulated gene are unclear . Both of these issues are partly a result of poor annotation of the sites that control gene regulation , the vast majority of which are likely to be noncoding . For example , transcription factor binding sites ( TFBSs ) , a major component of regulatory architecture , are small ( 6–15 bp ) , laborious to identify experimentally and potentially degenerate . Furthermore , due to their small size , genuine TFBS are difficult to differentiate from similar , randomly-occurring sequences that are present in large numbers in mammalian genomes . In an attempt to address the problem of annotation , evolutionary conservation has become popular as a metric for identifying putative regulatory regions [3] . However estimates of the true level of selective constraint ( defined as the proportion of mutations which are strongly deleterious ) in regulatory DNA will , by definition , be biased upwards in datasets predicted using evolutionary conservation as a criterion . One way to address this problem is to focus solely on regulatory regions which have been defined primarily by experimental rather than evolutionary criteria . In this study , we estimated levels of selective constraint in mammalian regulatory noncoding DNA using two complementary datasets , both of which draw upon experimental data . The first was derived from the literature collected in the TRANSFAC database [4] , such that every TFBS is supported by at least a single refereed publication . The advantages of this dataset are twofold . First , our dataset consists of individual TFBSs for which experimental support exists and which , according to an analysis by publication date ( see Discussion ) , appear to be subject to relatively little ascertainment bias with respect to evolutionary conservation . Second , the literature in TRANSFAC also provides substantial information on the gene regulated and the transcription factor bound for each TFBS . Thus , the TFBSs in our dataset can be assigned to a specific gene reliably , and we can also determine at least some of the transcription factors ( TFs ) which regulate a specific gene's expression . Our second dataset comprises sequences which have been identified as potentially transcription-factor-binding using chromatin immunoprecipitation combined with genomic microarray ( ChIP-chip ) analyses . Specifically , we combine the sequences annotated in refs 5–11 . While the resolution at which regulatory sites are identified is undoubtedly lower in the ChIP-chip dataset than in our TFBS dataset , our ChIP-chip sequences will still be highly enriched for functional regulatory DNA . Using these two datasets , we addressed the following questions: ( i ) what fraction of regulatory mutations in primates are strongly deleterious , ( ii ) does the fraction of strongly deleterious mutations at TFBSs vary between primates , ( iii ) what fraction of substitutions in human regulatory regions have been driven to fixation by positive selection , ( iv ) how does the selective constraint of human regulatory noncoding regions relate to the expression profile of the gene they regulate and ( v ) does the rate of protein evolution of a TFs also relate to the expression profile of the regulated gene ?
We next estimated the level of selective constraint at TFBSs , pCRMs and in ChIP-chip sequences ( Figure 4 ) . TFBSs appear to be reasonably highly constrained , approximately equivalent to a 2-fold degenerate synonymous site . This result is in good agreement with previous studies which have suggested that a reasonable proportion of TRANSFAC binding sites are conserved between human and a variety of mammalian species [12]–[16] . Estimates of constraint in our putative cis-regulatory modules and sequences annotated by ChIP-chip experiments are somewhat similar ( 0 . 14 and 0 . 11 , respectively ) suggesting that ChIP-chip studies can serve as reliable guides to functional regulatory regions in humans when compared with more traditional methods of identification . It has been suggested that regulatory DNA in primates is under relaxed selective constraint relative to rodents [17] . This has been attributed to the reduction of effective population size in primates facilitating the fixation of slightly deleterious mutations in gene control regions . Effective population size is likely to vary between humans , chimpanzees and macaques and we therefore investigated whether any significant difference in constraint of regulatory noncoding regions existed between humans and their close relatives . It is clear from Figure 5 that selective constraints in regulatory noncoding DNA vary significantly between primate species ( 1-way ANOVA , P<10−16 ) , and that this is primarily a result of a reduction in constraint in hominins ( post-hoc Tukey test , P<10−5 human vs rhesus , chimp vs rhesus ) when compared with rhesus macaques . Summing over both coding and noncoding sites , we also find that mean selective constraint is also reduced somewhat in humans , compared with chimpanzees ( 0 . 254 versus 0 . 279 ) , although this difference is only marginally significant ( Bootstrap t-test , P<0 . 07 ) . Two possible explanations for the reduction in constraint are that humans and chimpanzees have accumulated substantially greater numbers of deleterious mutations or are experiencing higher rates of adaptive evolution in their regulatory regions . In order to investigate whether the reduced constraints we observed in human regulatory noncoding DNA were the result of adaptive evolution we estimated the proportion of substitutions ( α ) which were driven to fixation by positive selection in our pCRMs and ChIP-chip sequences using the McDonald-Kreitman framework [18] , [19] . We were able to map 232 of our pCRMs and ChIP-chip sequences onto regions sequenced by the NIEHS Environmental Genome Project ( EGP; http://egp . gs . washington . edu ) . Polymorphism data was taken from the EGP as this dataset is free of ascertainment bias , relative to other large polymorphism datasets , such as HapMap [20] . McDonald-Kreitman analyses assumes that all mutations can be divided into strongly selected ( positively or negatively ) or strictly neutral classes . One problem with this is that a non-negligible fraction of new mutations in species with small effective population sizes , such as primates , may be weakly negatively selected . To account for this possibility we estimated α using both all segregating sites and excluding those sites where the minor allele frequency MAF ranged from 0 . 01 to 0 . 30 , many of which are likely to be slightly deleterious [21] . We find no evidence of adaptive evolution in human regulatory regions and our estimate of α is not significantly different from zero across the entire range of excluded , low frequency polymorphisms ( Figure 6 ) . We next investigated whether constraint in our regulatory sites covaried with the expression breadth of the gene regulated . Expression profile of the genes inferred to be regulated by our pCRMs and ChIP-chip sequences was estimated from the human microarray data of Su et al [22] . A gene was defined as expressed in a specific tissue based on the Affymetrix MAS5 presence/absence calls . Our results were qualitatively unchanged when gene expression was designated using a cutoff probe intensity value ( data not shown ) . For comparison , we also estimated constraint at the nonsynonymous ( 0-fold degenerate ) and synonymous sites of the genes adjacent to our regulatory regions . The results of this analysis are presented in Figure 7 . There is a clear relationship between selective constraint in regulatory regions and breadth of expression of the regulated gene . pCRM and ChIP-chip selective constraint is significantly negatively correlated with the number of tissues in which a gene is expressed ( P<0 . 005 , P<5 . 07×10−7 , respectively ) . This is not a function of the number of annotated TFBSs in our pCRMs , which is uncorrelated ( Pearson r = 0 . 015;P<0 . 738 ) with pCRM constraint . A similar relationship appears to exist between TFBS selective constraint and expression breadth . In particular , the TFBSs of tissue-specific genes are more highly constrained than those of intermediate and broadly expressed genes ( 2-sided t-test; P<0 . 012 ) . However , the equivalent regression is not significant at least in part due to the high error involved in estimating selective constraint from a small number of sites between closely related species . The relationship between constraint and expression profile is reversed in protein-coding sequence where constraint increases with increasing expression breadth ( P<1 . 11×10−15 and P<1 . 70×10−6 , nonsynonymous and 2-fold degenerate , respectively ) , a result supported by previous work [23] , [24] , [25] . Interestingly , constraint at 4-fold degenerate synonymous sites is also positively correlated with expression breadth ( P<2 . 60×10−6 ) , suggesting that constraints on mRNA stability and/or splicing efficiency reflect those on protein structure , with respect to expression breadth . These results are not a product of different rates of nucleotide substitution in the intronic control regions of genes with differing expression breadth; we find that divergence in all controls used in our study was uncorrelated with breadth of expression of the gene in which they reside ( Pearson r = −0 . 004 P>0 . 83; Figure S2 ) . It has previously been shown that mammalian promoters can be divided into two classes , CpG-rich and CpG-poor , based on the distribution of %CpG in human promoter regions [26] and these two classes of promoter region are associated with expression breadth . Following ref 26 , we divided our pCRM and ChIP-chip sequences into CpG-rich and CpG-poor classes , to investigate whether this could explain the relationship we find between expression breadth and conservation . The majority ( 95% ) of our pCRMs and ChIP-chip sequences are CpG-rich by the definition in ref 26 i . e . they have a normalized CpG content of >0 . 35 . Within this CpG-rich class , constraint of regulatory regions is still significantly negatively correlated with expression breadth ( Pearson r = −0 . 103 , P<3 . 68×10−8 ) . We also tested the influence of CpG content by regressing pCRM and ChIP-chip constraint on their %CpG . The slope of this regression is negative and significantly different from zero ( simple linear regression b = −0 . 058 , P<0 . 028 ) . However , the residuals of this regression are still negatively correlated with expression breadth ( Pearson r = −0 . 086 , P<5 . 99×10−7 ) . These results suggest that , while CpG content is indeed correlated with constraint of regulatory DNA this does not explain the majority of the relationship we see between regulatory constraint and expression profile . One advantage of our TFBS dataset is that we can identify which TF ( s ) control the expression of a specific gene , and that this relationship is also supported by experimental evidence . We therefore investigated whether the rate of protein evolution ( estimated as Dn/Ds , the ratio of nonsynonymous to synonymous substitution ) in a TF bore any relationship to the expression breadth of the regulated gene . Dn/Ds was estimated summing over all sites of all TFs which were known to regulate a specific gene . We obtained Dn/Ds estimates for 185 TFs which regulate 349 genes . The results of this analysis are presented in Figure 8 . Interestingly we find that TF Dn/Ds ratio is significantly positively correlated with gene expression breadth ( Pearson r = 0 . 15;P<0 . 005 ) . We tested whether this result was an artifact of summing across multiple TFs by restricting our analysis to the 99 genes which were regulated by a single TF . Despite this reduced dataset TF Dn/Ds is still marginally significantly correlated with gene expression breadth ( P<0 . 076 ) . One major factor which influences the rate of protein evolution is their structure . We therefore tested whether the relationship between transcription factor Dn/Ds and gene expression profile was influenced by the transcription factor structural class . We divided the regulating TFs into four protein “superclasses” based on the transcription factor protein classification tree in TRANSFAC . Of our 185 TFs we were able to assign 141 to either leucine zipper factors ( LZ; 27 TFs ) , zinc-coordinating DNA-binding domains ( ZC; 50 TFs ) , helix-turn-helix proteins ( HTH; 42 TFs ) and β-scaffold factors with minor groove contacts ( BSF; 22 TFs ) . As expected , we find clear differences in the Dn/Ds ratio of each the four classes ( 1-way ANOVA P≪1 . 69×10−5; Figure S3 ) . However , we find no relationship between protein structural class and expression breadth of the regulated gene ( 1-way ANOVA P<0 . 464; Figure S4 ) . This suggests that the relationship we observe between the rate of protein evolution in a TF and the expression breadth profile of the regulated gene are independent of the protein structure of the TF .
We have presented a study of the fitness effects of mutations in primate regulatory noncoding DNA . The regulatory regions included in our study are supported by a variety of experimental sources , both based on the extensive experimental biology literature , and inferred from more recent , high-throughput studies . Our study confirms that experimentally validated regulatory noncoding regions are selectively constrained , a result supported by other previous studies of datasets of TRANSFAC TFBSs in mammals [12]–[16] . Our estimates imply that ∼37% of new spontaneous mutations in primate TFBSs have a strongly deleterious effect and are removed by purifying selection . We find that the proportion of strongly deleterious noncoding regulatory mutations varies significantly even between closely-related primate species , reflecting a similar trend in coding DNA . We find no evidence for adaptive evolution in human regulatory regions , suggesting that these differences in selective constraint between primate taxa are likely to primarily reflect variations in effective population size . Our study also clearly shows that the level of selective constraint in primate regulatory DNA depends upon the expression profile of the gene regulated . Intriguingly , we also find higher constraint in the regulatory regions of tissue-specific genes is reflected in the rate of protein evolution of the TFs that interact with them . Our study suggests that at least some fraction of human regulatory DNA is accumulating slightly deleterious mutations at an accelerated rate relative to other , closely-related primate species . We find no evidence of adaptive evolution in our regulatory regions . Nonetheless , a number of recent reports have suggested accelerated evolution in human noncoding DNA [27] , [28] , [29] . There may be a number of reasons that we do not observe such an effect . Firstly , we restrict our analysis to experimentally-supported regulatory noncoding DNA and exclude CpG prone sites entirely from our analysis and may therefore lack sufficient power to detect all but very strong selection . Secondly our analysis is based upon the McDonald-Kreitman test which assumes that all adaptive mutations are strongly selected . However , recent work has suggested that at least some fraction of adaptive mutations may be weakly selected [30] . Although our degree of confidence in our estimates of α is small , the increasing numbers of high quality ChIP-chip datasets combined with larger resequencing studies will improve the accuracy of estimates of this important parameter . The results we have presented also shed light on the relationship between gene expression and selective constraint of both the TF and TFBSs which ultimately control this expression . A straightforward interpretation of our results is that selective constraint of regulatory DNA parallels the “complexity” of expression of the gene it regulates i . e . genes that are required to be “switched on” ubiquitously have a simpler , more degenerate regulatory architecture than those genes which require delicate control of the location and timing of expression . This interpretation is supported by a recent study of human-mouse promoter regions [31] . Furthermore , this hypothesis is intuitively appealing when we consider that tissue-specific genes may require regulatory sites both to up-regulate expression in the correct tissue , but also to suppress expression in an inappropriate tissue , a function that is presumably absent from the regulatory region of a broadly-expressed gene . Taken together with estimates of constraint in protein-coding sequence our study suggests the following: broadly expressed genes produce a protein whose structure is tightly maintained by purifying selection but whose regulatory architecture is degenerate . Tissue-specific genes on the other hand require a more elaborate and specific regulatory apparatus , but the protein produced by such genes is less rigorously maintained by selection . It has been suggested that mutations affecting the regulation of tissue-specific genes are less likely to be strongly deleterious than those in broadly-expressed genes , given that they are expressed in a subset of tissues [32] . However , our results support the opposite interpretation . Although the correlations we observe between regulatory constraint and expression breadth are weak , we note that the experimental methods of annotation of regulatory sites are imperfect , and the numbers of sites which we have used in this study are relatively small , by genomic standards . In addition , estimates of selective constraint are essentially a ratio of ratios , making them inherently noisy . In the light of this , the strength of our correlations is perhaps less surprising . It is also likely that our results to a certain extent reflect the variation in constraint of the noncoding DNA surrounding different functional “classes” of genes , as demonstrated previously ( e . g . ref 33 ) . We note , however , that the relationship between gene expression profile and gene functional class as assigned by ontological classification is uncertain . In addition , without a complete annotation of functional noncoding sites , we cannot distinguish whether between-gene variation in constraint of surrounding noncoding regions reflects variation in the number of constrained sites or in the intensity of purifying selection at these sites . One advantage of the approach we have employed here is that we can at least partially disentangle these two factors; our results suggest that the intensity of purifying selection at primate TFBSs is indeed greater in tissue-specific genes ( Figure 7 ) . We note that our estimates of constraint may also be biased upwards for two reasons . The UCSC whole genome alignments are assembled with reference to the human genome and it is therefore possible that their use in our study could exclude weakly conserved unalignable TFBSs . This could potentially lead to an overestimate of the true level of constraint . However , we suggest that the impact of this is likely to be small given that such a bias will affect our control regions also , and thus will cancel in the estimation of constraint . In addition , although ascertainment bias in the TRANSFAC annotations is reduced , compared to some computationally-predicted regions , it is unlikely to be zero , as phylogenetic footprinting has become more frequently used over time as a means of selecting candidate regulatory regions for experimental testing . Unfortunately it is difficult to quantify this bias . However , if phylogenetic footprinting has had a significant effect upon our TFBS dataset we might predict that , on average , those TFBSs that were annotated relatively recently would be less diverged than those annotated in the more distant past , given the dramatic increase in the use of comparative genomics in recent years . We find that divergence is not significantly correlated with year of appearence of the supporting publication ( Figure S5 ) . We do find that TFBSs published before 1996 ( the median age of publication of human TRANSFAC TFBSs ) are marginally ( ∼8% ) more diverged than those published during or after 1996 , although this difference is not significant ( Bootstrap t-test , P<0 . 19 ) . Thus , although our estimates of TFBS constraint may be upwardly biased , this bias is likely to be small . One straightforward implication of our results is that deleterious regulatory mutations are more likely to disrupt genes with tissue-specific expression , as a result of higher levels of constraint in both their regulatory regions and the protein-coding sequence of the TFs that bind to these regions . We estimate that deleterious mutations will occur on average 1 . 6-fold more often in regulatory regions of tissue-specific ( ≤3 tissues ) than housekeeping genes ( >35 tissues ) . This conclusion has interesting implications when we consider recent evidence suggesting that there are substantially more tissue-specific genes in primates compared with rodents [34] . Our data imply that the penalty for an increase in expression “complexity” is a concurrent increase in the genomic deleterious mutation rate . This penalty may , however , be offset by a corresponding decrease in the proportion of deleterious protein-coding mutations .
The data used in this study were collected from two sources . We first used the literature in TRANSFAC release 10 . 2 [4] to compile a dataset of known , experimentally-supported TFBSs . For those TFBSs which were linked to a specific EMBL accession , we BLASTed the binding site and up to 400 bp of flanking sequence against the human genome ( assembly 18 ) . Query sequences which matched a single , unique region in the human genome with a BLAST e-value of <10−5 were accepted . Those regions which matched more than a single location were resolved manually by comparison with any existing annotation in TRANSFAC , or excluded . For those TFBSs which were not linked to an existing EMBL record , we BLASTed the binding site sequence against the transcript of the RefSeq gene regulated , as recorded in TRANSFAC , with 20 kb flanking sequence . We accepted any binding site which matched a single unique location in this sequence , with <99% identity , for the full length of the binding site . We hereafter refer to these data as “TFBS” sequences . All binding sites were checked to be in the appropriate chromosomal location with respect to the gene they regulate . Our second dataset was derived from DNA sequences bound by a variety of TFs in 7 recent chromatin immunoprecipitation-coupled DNA microarray ( ChIP-on-chip ) analyses [5]–[11] . The locations of these sequences were extracted from the “fragment” table of TRANSFAC 10 . 2 and updated to the latest assembly of the human genome . We hereafter refer to these data as “ChIP-chip” sequences . To estimate the level of selective constraint , we needed to compare substitution rates in our TFBS and ChIP-chip datasets with those in an appropriate neutrally-evolving control region , which has a mutation rate equal to that of the region of interest . Previous analyses [33] , [35] have suggested that , in mammals , intronic regions outside the first intron and the splice sites are the fastest evolving in the genome and among the best candidates for neutrally-evolving sequence . Because sites in both datasets were highly nonrandomly distributed across the genome , we sought to define a single control region for a “case” region of binding sites , rather than for each individual annotated sequence . A “case” was defined as a group of TFBSs or ChIP-chip sequences in which the maximum distance between each cluster member and its nearest neighbour was 100 kb . A control region for each “case” was defined as the window which extended up to 250 kb either side of the midpoint of cluster . All non-first intronic sequence , excluding the first and last 100 bp , within this 500 kb window were denoted as control regions for the “case” region . Given that mutation rates in mammals appear to vary across megabase scales [36] , [37] it is likely that the mutation rate in our control sites will not differ significantly from that in our “case” sites . In a minority of cases ( <5% of TFBSs and ChIP-chip sequences ) , suitable intronic controls were unavailable . In this case , we used nearby intergenic sequence which was greater than 1 kb from an annotated coding sequence . All exon locations were taken from RefSeq annotations . The selection of an arbitrary between-site distance of 100 kb allowed us to define 473 unique , nonoverlapping binding site “cases” , each with a unique set of intronic controls . Likewise , we defined 6712 ChIP-chip “cases” from our 10104 unique ChIP-chip regions . For all TFBSs , ChIP-chip sequences and their corresponding control regions , aligned sequence data from the human , chimpanzee ( assembly 2 ) and macaque ( assembly 2 ) genomes was extracted from the 28-way vertebrate alignments available in the UCSC genome browser database [38] . In order to minimize the effects of poor sequence quality in the chimp and macaque genomes we masked all sites which were assigned a base quality of less than 20 in either species . Lineage-specific substitution rates were estimated using parsimony . Estimates of substitution rates were not corrected for multiple hits , given that this will make little difference between closely related species . In all cases , selective constraint , C , was estimated as:where O is the number of substitutions observed in the TFBS or other region of interest and E is the number of substitutions expected under neutral evolution:where n is the length of the TFBS or other region of interest and K is the substitution rate estimated from the control region . Unless stated otherwise , all confidence intervals were estimated by bootstrapping the data by binding site “case” , 1000 times . The method of estimation of selective constraint employed here explicitly accounts for local mutational variation . Previous studies of experimentally validated mammalian regulatory DNA ( e . g . refs 14–16 ) , have not accounted for such variation . This is particularly important in our study for two reasons . Firstly , substantial within-genome mutational variation is known to occur in mammals [37] , [39] meaning that the expectation of conservation under neutrality will vary from one genomic region to the next . This can substantially impact estimates of conservation between very closely related species , such as humans and chimpanzees . Secondly , regulatory regions frequently reside in CpG islands , where the level of CpG hypermutability is known to differ from other , more heavily methylated regions of the genome . Given that CpG mutations make up a disproportionately large number of all mutations in mammals , it is important to correct for variations in the level of CpG hypermutability to avoid overestimating constraint in regions of lowered CpG hypermutability such as CpG islands . Here , we account for variation in the frequency and mutability of CpG dinucleotides by excluding non CpG-prone sites ( not preceded by ‘C’ or followed by ‘G’ ) . TF Dn/Ds ratios were estimated from human-macaque alignments in the Cornell orthologues dataset using PAML [40] . In the case where multiple factors were known to regulate a gene x , Dn/Ds ( ωx ) was estimated summing over all TFs , Tx = t1 , … , tn as:where KA ( ti ) and Ks ( ti ) are the number of pairwise nonsynonymous and synonymous substitutions in TFi , and NA ( ti ) and Ns ( ti ) are the number of pairwise nonsynonymous and synonymous sites in TFi , respectively . | Changes in gene expression have been suggested to play a major role in mammalian evolution . In eukaryotes , gene expression is primarily controlled by sites , such as transcription factor binding sites ( TFBSs ) , located in the noncoding region of the genome . The majority of these TFBSs remain unannotated , however , because they are typically short , degenerate , and laborious to identify experimentally . As a result , the effects of mutations in TFBSs on organism fitness remain poorly understood . We collected a dataset of TFBSs derived from the experimental biology literature and recent high-throughput studies to estimate the proportions of new mutations in TFBSs that have strongly deleterious and strongly beneficial effects upon organism fitness . We find that a relatively high proportion of new mutations in TFBSs are strongly deleterious , although it appears that relatively few are adaptive . We also demonstrate that the fraction of strongly deleterious regulatory mutations is correlated with the breadth of expression of the regulated gene . Thus , ubiquitously expressed genes are likely to experience fewer deleterious regulatory mutations than those expressed in a small number of tissues . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/gene",
"expression",
"evolutionary",
"biology/human",
"evolution",
"evolutionary",
"biology/evolutionary",
"and",
"comparative",
"genetics",
"evolutionary",
"biology",
"genetics",
"and",
"genomics/population",
"genetics"
] | 2008 | Selective Constraints in Experimentally Defined Primate Regulatory Regions |
Parasitoid wasps are important natural enemies of arthropod hosts in natural and agricultural ecosystems and are often associated with viruses or virion-like particles . Here , we report a novel negative-stranded RNA virus from a parasitoid wasp ( Pteromalus puparum ) . The complete viral genome is 12 , 230 nucleotides in length , containing five non-overlapping , linearly arranged open reading frames . Phylogenetically , the virus clusters with and is a novel member of the mononegaviral family Nyamiviridae , here designated as Pteromalus puparum negative-strand RNA virus 1 ( PpNSRV-1 ) . PpNSRV-1 is present in various tissues and life stages of the parasitoid wasp , and is transmitted vertically through infected females and males . Virus infections in field populations of P . puparum wasps ranged from 16 . 7 to 37 . 5% , without linearly correlating with temperature . PpNSRV-1 increased adult longevity and impaired several fitness parameters of the wasp , but had no influence on successful parasitism . Strikingly , PpNSRV-1 mediated the offspring sex ratio by decreasing female offspring numbers . RNA interference knockdown of virus open reading frame I eliminated these PpNSRV-1-induced effects . Thus , we infer that PpNSRV-1 has complex effects on its insect host including sex ratio distortion towards males , as well as possible mutualistic benefits through increasing wasp longevity .
Parasitoid wasps ( Hymenoptera: Aprocrita ) are important natural enemies of arthropods ( Ecdysozoa: Arthropoda ) and are widely used as bio-control agents against insect ( Arthropoda: Hexapoda: Insecta ) pests in agro-ecosystems . They are frequently associated with viruses or virion-like particles ( VLPs ) . To date , viruses of seven families have been identified in parasitoid wasps: double-stranded DNA ( dsDNA ) viruses ( Ascoviridae , Polydnaviridae , Poxviridae ) , positive-sense , single-stranded RNA [ ( + ) ssRNA] viruses ( Coronaviridae , Iflaviridae ) , and segmented double-stranded RNA [dsRNA] viruses ( Reoviridae ) [1] . Among these viruses , polydnaviruses are most often detected in parasitoids wasps [2] and they are the best known example of an insect/virus symbiosis . Each polydnavirus from a wasp of a given species persists as an integrated provirus in the germ line and somatic cells of the wasp’s body . Polydnaviruses are therefore endogenous virus elements ( EVEs ) that have become genetically fixed in different wasp lineages . When polydnavirions are injected into the body of the wasp’s lepidopterous host , viral DNAs can be discharged into the cell nuclei and integrated into the genome of infected host cells [3] . The expression products of polydnaviral genes are exclusively beneficial to parasitoid wasps for successful survival and emergence from their hosts by suppressing the host immune system and modifying host growth , development , and metabolism [4] . Similarly , Diachasmimorpha longicaudata entomopoxvirus ( DlEPV ) infects host hemocytes and induces cytopathic effects that disable host encapsulation to promote the wasp offspring survival [5] . In contrast , Diadromus pulchellus ascovirus 4a ( DpAV4 ) , vectored by Itoplectis tunetana wasps , prevents the development of parasitoid eggs and larvae by quickly triggering cell lysis [6] . In addition , VLPs devoid of DNA or RNA have been observed in ovarian tissue or the venom apparatus of parasitoid wasps , and their functions are well studied [7–11] . For example , the unclassified filamentous Leptopilina boulardi filamentous virus ( LbFV ) alters superparasitism behavior and parasitism rates of the infected parasitoid wasp ( Leptopilina boulardi ) by increasing its tendency to lay supernumerary eggs in parasitized hosts [12–15] . The transmission of LbFV is wasp density-dependent , which may influence the coexistence of distinct leptopilines . For instance , wasps of the species L . boulardi can quickly outcompete wasps of the species L . heterotoma in the absence of LbFV , whereas wasps of the species L . heterotoma can maintain or even eliminate those of the species L . heterotoma in the presence of LbFV [16] . Although a few RNA viruses have been discovered in parasitoid wasps , their effects on the wasps have rarely been determined [1] . With the recent adoption of new next generation sequencing ( NGS ) technologies , several new ( + ) ssRNA viruses have been discovered in parasitoid wasps . For example , three novel ( + ) ssRNA viruses could be identified in cDNA libraries of Nasonia vitripennis wasps . Two of these ( + ) ssRNA viruses , ( Nasonia vitripennis viruses 1 and 2 [NvitV-1/2] ) probably belong to the family Iflaviridae in the order Picornavirales; whereas the third virus ( NvitV-3 ) is most similar to the unclassified Picornavirales member Nora virus , which infects fruit flies [17] . However , the effect of these viruses on their hosts remains unknown . Recently , a new iflavirus named Dinocampus coccinellae paralysis virus ( DcPV ) has been found in Dinocampus coccinellae wasps , which parasitize the lady beetle ( Coleomegilla maculate ) . The replication of DcPV in lady beetle cerebral ganglia most likely induces changes in lady beetle behavior , such as tremors , gait disturbance , and limitations in movement [18] . Nonsegmented negative-sense single-stranded ( - ) ssRNA viruses , i . e . viruses of the order Mononegavirales , are of considerable importance and include notorious human , animal , and plant viral pathogens , such as filoviruses , paramyxoviruses , and rhabdoviruses [19] . Most mononegaviruses have been found in vertebrates and plants whereas only a few mononegaviruses have been detected in invertebrates ( S1 Table ) : soybean cyst nematode virus 1 ( SbCNV-1 , Nyamiviridae: Socyvirus ) was found in soybean cyst nematodes , and is phylogenetically closely related to tick-borne nyaviruses ( Nyamiviridae: Nyavirus ) [20] . In insects , rhabdoviral sigmaviruses were discovered in drosophilids [21] . Flies infected with sigmaviruses become paralyzed or die after exposure to high concentrations of CO2 , whereas uninfected flies can recover . Similar symptoms can also be found in parasitized mosquitos [22] and aphids [23] . Only one ( - ) ssRNA virus , Diachasmimorpha longicaudata rhabdovirus ( DlRhV ) , has been found in parasitoid wasps ( Diachasmimorpha longicaudata ) [24] . However , no evidence is available to confirm that the DlRhV genome is indeed related to rhabdoviruses , and the role of this virus in the successful development of its host wasp has not been clearly illuminated [1] . Members of the species Pteromalus puparum are predominantly pupal parasitoid wasps that prey on butterflies of several species including the small white ( Pieris rapae ) , with the highest parasitism rate over 90% reported in fields of cruciferous vegetables in China [25] . P . puparum wasps are gregarious and synovigenic parasitoids . The vitellogenesis of P . puparum wasps is initiated after the pupal stage and completed at the adult stage 48 h after eclosion [26] . Adult nutrition significantly affects ovarian development [27] . Unlike parasitoid wasps carrying polydnaviruses , this parasitoid wasp has evolved a different means of using venom to suppress the host immune system and modify host development [28 , 29] . The composition and functions of the venom from this parasitoid have been well studied by our laboratory [30–36] . To better understand the relationship between this parasitoid wasp and its hosts , we recently sequenced the wasp’s transcriptome and discovered a novel virus . We determined the complete viral genome and confirmed that this virus , named Pteromalus puparum negative-strand RNA virus 1 ( PpNSRV-1 ) , belongs to the mononegaviral family Nyamiviridae , but is unrelated to viruses of the existing nyamiviral genera Nyavirus and Socyvirus . We further investigated the tissue and developmental expression profile , field prevalence , transmission strategy , and biological characteristics of PpNSRV-1 . The most striking discoveries are that PpNSRV-1 is vertically transmitted by both infected female and male wasps , and that the virus modifies the secondary sex ratio of the parasitoid host by reducing female offspring number . This is the first report of an unambiguous mononegavirus in parasitoid wasps .
Analysis of the transcriptome of the parasitoid wasp Pteromalus puparum uncovered a large contig ( 12 , 141 bp ) containing regions similar to viral RNA-dependent RNA polymerases ( RdRps ) . Consequently , we sequenced the associated viral genome , including the 5′ and 3′ genome termini . The complete genome of PpNSRV-1 is 12 , 230 nucleotides in length ( Genbank #KX431032 ) . G+C pairs comprise 46 . 09% of the nucleotides . The PpNSRV-1 genome contains five large open reading frames ( ORFs I–V ) located at nucleotide ( nt ) positions 153 to 2015 , 2109 to 2564 , 2608 to 3816 , 3842 to 5566 , and 5612 to 12112 , respectively . The leader and trailer regions of the PpNSRV-1 genome are 152 and 118 nt in length , respectively ( Fig 1A ) . Their terminal nucleotides are not complementary . By comparing the 5′ and 3′ untranslated regions and intergenic regions of the PpNSRV-1 genome , we pinpointed the putative transcription initiation and termination signals for each of the five ORFs . A conserved transcription initiation motif of 3′- ( G/U ) UUCUAUAUUU ( C/U ) UU-5′ was identified upstream of each putative ORF at 4–44 nt before the start codon ( Fig 1B ) . Similarly , the transcription termination motif of 3′-CUAUAUUUCUUUUG-5′ was detected downstream of every ORF at 1–104 nt downstream of the stop codon ( Fig 1C ) . No intergenic , non-transcribed regions were noted between transcription initiation and termination motifs . To further elucidate the transcriptional strategy of PpNSRV-1 , primers were designed for rapid amplification of cDNA ends ( RACE ) of each ORF . RACE results indicated that all five ORFs can be transcribed independently as shown in the transcript map ( Fig 1D ) . The virus lacks a poly- ( A ) tail at 3′ terminus . We confirmed transcription of the five ORFs of the PpNSRV-1 genome by northern blot analysis ( S1 Fig ) . General properties of all ORFs are listed in Table 1 . The predicted translation product of PpNSRV-1 ORF V was found to be similar in amino-acid sequence to that of the Midway virus RNA-dependent RNA polymerases . Predicted products of the other four ORFs are not similar to any deposited protein sequence . Computational sequence analyses were performed to identify potential functional motifs in the predicted PpNSRV-1 proteins . Only ORF IV protein was predicted to be a putative type I transmembrane protein and to possess a cleavable signal peptide . Moreover , ORF IV protein contains seven potential O-linked and four potential N-linked glycosylation sites and 29 potential phosphorylation sites , indicating that ORF IV encodes the viral glycoprotein ( mononegaviral G ortholog/analog ) . Based on the conserved genomic orientation of mononegaviruses ( 3′-N-P-M-G-L-5′ ) , PpNSRV-1 ORF II , containing nine phosphorylation sites , most likely encodes the phosphoprotein ( P ) . ORF III , containing 37 potential O-linked glycosylation sites , may encode a glycosylated matrix protein ( M ) and ORF I consequently would encode the nucleoprotein ( N ) . The most closely related known sequence to PpNSRV-1 ORF V is the RNA-dependent RNA polymerase ( L ) of Midway virus ( Mononegavirales: Nyamiviridae: Midway nyavirus ) . Therefore we conducted a maximum likelihood phylogenetic analysis of the amino acid core sequences of the PpNSRV-1 ORF V protein to determine the relationship of PpNSRV-1 to other mononegaviruses . Representatives of each official genus of the seven official mononegaviral families [19] were included in the phylogenetic tree . Notably , PpNSRV-1 clusters as a distinct lineage in family Nyamiviridae ( Fig 1E , brown ) , indicating the need for a novel virus species in a novel nyamiviral genus . Multiple alignments of the predicted core RdRp motifs of PpNSRV-1 ORF V protein with those of other mononegaviruses revealed the mononegavirus-typical four highly distinct and conserved motifs ( A to D ) ( Fig 1F ) . Pairwise Sequence Comparison ( PASC ) analysis of the PpNSRV-1 genome revealed that it is 13 . 8% identical to the genome of soybean cyst nematode virus 1 ( SbCNV-1; nyamiviral genus Socyvirus ) , 12 . 3% identical to that of Nyamanini virus ( NYMV; nyamiviral genus Nyavirus ) , 12 . 1% identical to that of Midway virus ( MIDWV; nyamiviral genus Nyavirus ) , and 10 . 9% identical to that of Sierra Nevada virus ( SNVV; nyamiviral genus Nyavirus ) , with those viruses being the closest known relatives of PpNSRV-1 . This value ( 13 . 8% identity between the PpNSRV-1 and SbCNV-1 genome ) is lower than the 17% identity measured between nyavirus and socyvirus genomes , indicating the need for a novel nyamiviral genus . The PpNSRV-1 titers in the head , thorax , and abdomen of infected P . puparum wasps as analyzed by qPCR were not significantly different ( F = 1 . 088 , df = 2 , p = 0 . 3680 , Tukey’s multiple comparison test , Fig 2A ) . However , the female body segments contained more viral RNA compared to male body segments ( F = 18 . 123 , df = 1 , p = 0 . 0011 , Fig 2A ) . Tissue specific distribution analysis revealed significant differences among ovaries , testes , digestive tracts and female or male body remnant ( F [5 , 12] = 13 . 588 , p = 0 . 0001 ) . The highest viral titer was found in ovaries , followed by digestive tract and body remnant . Testes had the lowest viral titer ( Fig 2B ) . PpNSRV-1 was detected in the eggs , larvae , pupae and adults of P . puparum wasps by developmental expression analysis . The viral titer increased significantly from the larval to adult stages ( F ( 8 , 18 ) = 21 . 935 , p <0 . 0001 ) and peaked in adults ( Fig 2C ) . PpNSRV-1 genome copy number was the highest at days 4 post-eclosion ( Fig 2D ) ( Male: [F [6 , 14] = 2 . 251 , p = 0 . 0990] and Female: [F [6 , 14] = 2 . 497 , p = 0 . 0743] ) . At higher environmental temperatures ( 35°C versus 25°C ) , viral titers in both female and male wasps decreased marginally but not significantly ( Female: [F = 0 . 263 , df = 1 , p = 0 . 6151] and Male: [F = 0 . 284 , df = 1 , p = 0 . 6313] ) ( S2A and S2B Fig ) . Immunohistochemistry revealed PpNSRV-1 antigen in ovaries ( Fig 3A2–3A4 ) , eggs ( Fig 3B2 and 3B3 ) , midgut ( Fig 3C2 and 3C3 ) , testes and seminal vesicles ( Fig 3D2–3D4 ) of the PpNSRV-1-infected ( PpNSRV-1 ( + ) ) wasps . In the ovarioles , the viral signals were much higher in oocytes , follicle cells , and the intracellular space compared to nurse cells ( Fig 3A3 and 3A4 ) . PpNSRV-1 antigen could not be detected in tissues sampled from uninfected PpNSRV-1 ( PpNSRV-1 ( - ) ) wasps ( Fig 3A1 , 3B1 , 3C1 , and 3D1 ) . To study PpNSRV-1 particle morphology , we prepared sections of the digestive tract and ovaries from PpNSRV-1 ( + ) wasps and of purified particles and examined by transmission electron microscopy ( TEM ) . Spherically shaped VLPs ( 62 . 5–100 nm in diameter ) were present in follicular cells of the ovaries ( Fig 3E1 and 3E2 and S3 Fig ) . Similarly , VLPs ( 70–100 nm in diameter ) stacked in intracellular vesicles were observed in cells of the digestive tract ( Fig 3F1 and 3F2 ) . VLPs were not found in PpNSRV-1 ( - ) wasps . The spherical PpNSRV-1 particle shape is similar to particles produced by nyaviruses , i . e . MIDWV and NYMV . PpNSRV-1 is vertically transmitted from both PpNSRV-1 ( + ) females and males to their offspring , although the patterns differ between the sexes ( Fig 4A and S4B–S4E Fig ) . Notably , transmission was 100% from infected females to both their male and female progeny , but transmission from infected males when the mate is uninfected occurred only to female offspring ( 58% to female progeny but 0% to male progeny , Fig 4A and S4E Fig ) . Field P . puparum wasps were collected from five different locations in China to investigate the prevalence of PpNSRV-1 in different geographic locations . The virus was detected in four wasp populations sampled from Ningbo , Hangzhou , Jiande ( all Zhejiang Province ) , and Hefei ( Anhui Province ) with the PpNSRV-1-positive rate ranging from 16 . 7% to 37 . 5% . PpNSRV-1 prevalence decreased with dropping latitude , with no virus detection in samples collected from the lowest latitude location in Nanchang ( Jiangxi Province ) . However , the PpNSRV-1-positive rate was neither significantly correlated with temperature ( Fig 4C , R2 = 0 . 199 , F = 1 . 493 , p = 0 . 2676 , Tukey’s multiple comparison test ) , nor with latitude in October ( R2 = 0 . 740 , F = 8 . 536 , p = 0 . 0614 , Tukey’s multiple comparison test ) . In contrast , PpNSRV-1 could not be detected in unparasitized butterfly pupae collected from all five locations ( S3 Table ) . Comparison of the PpNSRV-1 ( + ) and PpNSRV-1 ( - ) wasp colonies revealed no significant difference in the successful parasitism rate ( t = 2 . 013 , df = 5 , p = 0 . 0789 , t-test , Fig 5A ) and the total oocyte number ( t = 1 . 591 , df = 28 , p = 0 . 1229 , S5 Fig ) . However , offspring sex ratio ( t = 3 . 959 , df = 131 , p = 0 . 0001 , Fig 5B ) , female offspring number emerging per parasitized pupa ( t = 2 . 682 , df = 131 , p = 0 . 0083 , Fig 5C ) , female offspring number laid by per female wasp ( t = 5 . 833 , df = 58 , p <0 . 0001 , Fig 5D ) , and adult emergence proportion ( t = 3 . 434 , df = 151 , p = 0 . 0008 , Fig 5E ) were significantly lower in the PpNSRV-1 ( + ) colony compared to the PpNSRV-1 ( - ) colony . Interestingly , the longevity of both female and male wasps in the PpNSRV-1 ( + ) colony was significantly longer than compared to that of the PpNSRV-1 ( - ) colony ( female: t = 7 . 936 , df = 60 , p <0 . 0001; male: t = 4 . 239 , df = 57 , p = 0 . 0001 , Fig 5F ) . Similarly , under higher environmental temperature ( 35°C ) , the longevity of both female and male wasps in the PpNSRV-1 ( + ) colony was still significantly longer compared to that of the PpNSRV-1 ( - ) colony ( female: t = 2 . 192 , df = 73 , p = 0 . 0316; male: t = 2 . 781 , df = 76 , p = 0 . 0068 , ( S2C Fig ) . Additionally , the female offspring number laid by each female wasp per day for the PpNSRV-1 ( + ) colony was significantly less compared to that of the PpNSRV-1 ( - ) colony in most cases ( S6D Fig ) . However , a significantly less female offspring number per parasitized pupa per day , which was parasitized by a female , was found in the PpNSRV-1 ( + ) colony compared to that observed in PpNSRV-1 ( - ) control colony on the first batch of parasitism . Female offspring from the two colonies from subsequent batches of parasitism was not significantly different ( S6A Fig ) . To eliminate the effect of different genetic backgrounds and to confirm effects of PpNSRV-1 on parasitoid wasp biology , we injected the virus into yellow pupae from the PpNSRV-1 ( - ) wasps ( PpNSRV-1 ( + ) group ) , and injected mock virus into another group as the control ( PpNSRV-1 ( - ) group ) . Similar results were obtained for both groups ( Fig 5 and S6C and S6G Fig ) . Comparison of the injected ( + ) and injected ( - ) wasp colonies revealed no significant difference in the successful parasitism rate ( t = 1 . 9931 , df = 8 , p = 0 . 0814 , t-test , Fig 5A ) . Conversely , offspring sex ratio ( t = 4 . 976 , df = 112 , p < 0 . 0001 , Fig 5B ) , female offspring number per parasitized pupa ( t = 2 . 515 , df = 112 , p = 0 . 0133 , Fig 5C ) , female offspring number laid by per female wasp ( t = 4 . 163 , df = 58 , p = 0 . 0001 , Fig 5D ) , and adult emergence proportion ( t = 3 . 060 , df = 142 , p = 0 . 0026 , Fig 5E ) were significantly lower in the injected ( + ) colony compared to the injected ( - ) colony . The longevity of both female and male wasps in the injected ( + ) colony was significantly longer than that of the injected ( - ) colony ( female: t = 6 . 0387 , df = 67 , p <0 . 0001; male: t = 3 . 867 , df = 54 , p = 0 . 0003 , Fig 5F ) . PpNSRV-1 titer increased steadily with wasp development after virus injection ( F ( 5 , 12 ) = 38 . 041 , p <0 . 0001 ) ( S7 Fig ) . On the 3rd day after adult eclosion , the viral load in the injected females ( 8 . 911 × 104 copies/ng of total RNA ) and males ( 8 . 417 × 104 copies/ng of RNA , S7 Fig ) was similar to the viral load in the PpNSRV-1 ( + ) wasps ( 1 . 202 × 105 copies/ng in females and 1 . 010 × 105 copies/ng in males ) ( Fig 2D ) . In a previous study , dsRNA was injected into Microplitis demolitor wasp pupae to determine the effect on bracovirus degradation and specificity of dsRNA [37] . We followed this strategy using an RNA interference ( RNAi ) assay . We injected dsRNAs targeting PpNSRV-1 ORF I into the yellow P . puparum pupae from PpNSRV-1 ( + ) wasps to investigate viral load , successful parasitism rate , adult emergence proportion , and offspring number . Virus load in the wasps on day 4 post-eclosion was quantified . For example , injection of 10 ng to 1 μg of “ds-ORF1-1” dsRNA yielded a similar level of knockdown ( Fig 6A ) ( F ( 3 , 8 ) = 221 . 216 , p < 0 . 0001 , Tukey’s multiple comparison test ) . PpNSRV-1 genome copies in ds-ORF1-1- and ds-ORF1-2-dsRNA injected wasps declined compared to the ds-eGFP dsRNA control in both sexes ( Fig 6B ) ( female: F ( 2 , 6 ) = 1216 . 426 , p < 0 . 0001; male: F ( 2 , 6 ) = 22 . 439 , p = 0 . 0016 ) . The successful parasitism rate had no significant difference among the dsRNA treated colonies ( Fig 6C ) ( F ( 3 , 16 ) = 0 . 378 , p = 0 . 7701 ) . Offspring sex ratio ( ds-ORF1-1 ( + ) : [t = 2 . 7661 , df = 111 , p = 0 . 0066] and ds-ORF1-2 ( + ) : [t = 2 . 5290 , df = 103 , p = 0 . 0130] , Fig 6D ) , female offspring number per parasitized pupa ( ds-ORF1-1 ( + ) : [t = 1 . 6558 , df = 111 , p = 0 . 1006] and ds-ORF1-2 ( + ) : [t = 2 . 8628 , df = 103 , p = 0 . 0051] , Fig 6E ) , female offspring number laid by per female wasp ( ds-ORF1-1 ( + ) : [t = 2 . 3281 , df = 58 , p = 0 . 0234] and ds-ORF1-2 ( + ) : [t = 2 . 3416 , df = 58 , p = 0 . 0227] , Fig 6F ) , were significantly higher in the PpNSRV-1 dsRNA treated colonies compared to the ds-eGFP ( + ) colony . Adult emergence proportion in the ds-eGFP ( - ) colony was higher than that of ds-eGFP ( + ) colony ( Fig 6G ) ( t = 1 . 9638 , df = 185 , p = 0 . 0511 ) . The longevity of both female and male wasps in the PpNSRV-1 dsRNA treated colonies was shorter than that of the ds-eGFP ( + ) colony , but longer than that of the ds-eGFP ( - ) colony ( female: t = 2 . 9338 , df = 58 , p = 0 . 0048 and male: t = 2 . 5293 , df = 58 , p = 0 . 0142 , Fig 6H ) .
We detected and sequenced the genome of a novel ( - ) ssRNA virus , PpNSRV-1 , from parasitoid wasps ( P . puparum ) and verified the presence of PpNSRV-1 particles in various wasp tissues . In addition , we characterized the PpNSRV-1 genome organization , its phylogenetic placement , transmission strategy , and biological impacts on its host . Our results represent the first unambiguous detection of a ( - ) ssRNA virus in insect parasitoids . Intriguingly , we revealed that PpNSRV-1 mediates the secondary sex ratio of its host wasp by decreasing female offspring numbers . The arrangement of the PpNSRV-1 genome follows the typical basic five-ORF pattern ( 3′-N-P-M-G-L-5′ ) of mononegaviral genomes [38] . Indeed , ORF sequence analysis demonstrated that PpNSRV-1 ORF IV and V encoded a glycoprotein ( G ) and an RNA-dependent RNA polymerase ( L ) , respectively . The functions of the other three PpNSRV-1 ORFs remain unclear at this point in time , but it is highly likely that they encode mononegaviral N , P , and M orthologs or analogs . The spherical shape of PpNSRV-1 particles found in infected parasitoid wasp is similar to that of nyaviruses , in particular MIDWV and NYMV [39] . Phylogenetic analysis revealed that PpNSRV-1 clustered with viruses in the mononegaviral family of Nyamiviridae . This is a relatively novel family , which currently includes two genera , Nyavirus ( Midway virus [MIDWV] , Nyamanini virus [NYMV] , Sierra Nevada virus [SNVV] ) and Socyvirus ( soybean cyst nematode virus 1 [SbCNV-1] ) [19 , 40] . All four known nyamiviruses have been found in ecdysozoans: MIDWV was discovered in 1966 in ixodid Ornithodoros ( Alectorobius ) capensis ticks collected in bird nests sampled on the US Midway and Kure Atolls in the Pacific Ocean and has since been found in ticks of the same species in Japan and Ornithodoros ( Alectorobius ) denmarki in Hawaii [40 , 41]; NYMV was first discovered in 1957 in tick-infected cattle egrets ( Bubulcus ibis ibis ) and their fowl tampans ( Argas walkerae ) in Nyamanini Pan in the Ndumu Game Reserve , northern Natal , South Africa . NYMV has since been isolated from Argas ticks of various species collected in Egypt , India , Nepal , Nigeria , and Thailand [40 , 42–44]; SNVV was isolated in 1975 from argasid ticks ( Ornithodoros coriaceus ) [45]; and SbCNV-1 was discovered in the US in 2010 in an inbred laboratory culture of soybean cyst nematodes ( Heterodera glycines ) [20] . PpNSRV-1 is therefore the first nyamivirus that infects insects . Our phylogenetic analysis and pairwise genome comparisons using PASC indicated that while PpNSRV-1 could be considered a nyamivirus , PpNSRV-1 is not closely related enough to the four known nyamiviruses to allow classification in the genera Nyavirus or Socyvirus . After consultation , the International Committee on Taxonomy of Viruses ( ICTV ) Nyamiviridae Study Group agreed with our assessment . We have therefore submitted an official taxonomic proposal ( TaxoProp ) together with the Study Group to the ICTV , proposing a novel nyamiviral genus ( Peropuvirus ) including a single species ( Pteromalus puparum peropuvirus ) for PpNSRV-1 ( TaxoProp 2016 . 015a-dM . A . v1 . Peropuvirus; https://talk . ictvonline . org ) . This proposal has by now been accepted by the ICTV Executive Committee and awaits ratification . The relatively low bootstrap values ( Fig 1E ) suggest that this classification may have to be corrected in the future once additional nyami-like and/or borna-like virus genomes become available for analysis . Indeed , the discovery of PpNSRV-1 indicated that the nyamiviral clade of mononegaviruses is in all likelihood highly diverse , given that the now five members of the family have been found in very different geographic areas ( Africa , Asia , North America , and the Pacific ) and distinct ecdysozoan clades ( Chelicerata/Arachnida , Hexapoda/Insecta , and Nematoda ) . We found PpNSRV-1 in all tissues and life stages of females and males of the PpNSRV-1 host , the P . puparum wasp . At this point , we consider the expression of PpNSRV-1 to be constitutive and not specific for any wasp tissue or developmental stage . In contrast , DlRhV was found in the venom apparatus and the subchorionic space of oviposited eggs , but not in the oviduct or pre-vitellogenic or chorionated vitellogenic ova of female DlRhV hosts ( Diachasmimorpha longicaudata parasitoid wasps ) [46] . In addition , we did not detect distinct pathologies in any of the tested tissues of PpNSRV-1-infected wasps . Therefore , PpNSRV-1 could be considered as a nonpathogenic commensal virus of P . puparum wasps . Other ( - ) ssRNA viruses infecting insects , such as sigmaviruses , can be vertically transmitted through both eggs and sperm to the host offspring [47] . A vertical transmission pattern was also found in other parasitic wasp-associated RNA viruses that infect the tissues of the female parasitoid wasp host reproductive system [1] . The results from crossing experiments between PpNSRV-1 ( + ) and PpNSRV-1 ( - ) P . puparum wasps clearly supported the existence of vertical transmission route of PpNSRV-1 to the offspring via the infected wasp reproductive tissues . The vertical transmission pathway was also supported by detection of PpNSRV-1 antigen and visualization of PpNSRV-1 particles in wasp ovarioles , eggs , testes , and seminal vesicles . The vertical transmission efficiency of PpNSRV-1 differs from that of two sigmaviruses discovered in Drosophila affinis and Drosophila obscura fruit flies . In D . affinis flies , infected females transmitted Drosophila affinis sigmavirus ( DAffSV ) to 98% of their offspring , whereas males transmitted the virus to only 45% of their offspring . In D . obscura flies , infected females transmitted Drosophila obscura sigmavirus ( DObsSV ) to 92% of their offspring , whereas males transmitted the virus to 88% of their offspring [48] . But in P . puparum wasps , infected females transmitted PpNSRV-1 to 100% of their female and male offspring , whereas males transmitted the PpNSRV-1 to 58% of their female offspring and 0% of their male offspring . The vertical transmission mode of PpNSRV-1 may explain why four out of five field populations of P . puparum wasps have a ≈16 . 7–37 . 5% infection rate , and why our PpNSRV-1 ( + ) colony can be stably reared in the laboratory from one generation to the next . Especially noteworthy is that males transmit PpNSRV-1 exclusively to female offspring , as only female offspring is derived from fertilized eggs in haplodiploid insects such as P . puparum wasps . This result is intriguing , and suggests that the virus is transmitted within sperm or is tightly associated with sperm , because in haplodiploid species , sperm-fertilized eggs develop into female progeny while unfertilized eggs develop into male progeny [49] . An alternative explanation is that the virus is also integrated into the wasp’s nuclear genome and transmitted vertically through chromosomes . Further experimentation is underway to resolve these alternatives . Recently , a double-stranded RNA virus found in L . boulardi wasps ( Leptopilina boulardi toti-like virus , LbTV ) has a male vertical transmission pattern through the maternal lineage even if frequent paternal transmission also occurs [50] . This result suggests that the transmission pattern we observed with PpNSRV-1 may be more widespread in nature than previously appreciated . How P . puparum wasps are infected with PpNSRV-1 remains unclear and should be investigated in future . In our studies , PpNSRV-1 prevalence in field populations was positively correlated with latitude , but was not significantly correlated with temperature . The female sex ratio of overwintering P . puparum wasp field generations was less than 50% , which we consider to be related to superparasitism or higher offspring number in a host [25 , 51 , 52] , but could also be due to a seasonal increase in frequency of the virus . By comparing biological parameters between PpNSRV ( + ) and PpNSRV ( - ) colonies of P . puparum wasps , we found that PpNSRV-1 can affect the offspring sex ratio of the host parasitoids by decreasing the total female offspring number . The underlying mechanism of PpNSRV-1-mediated secondary sex ratio alteration remains unclear . In arthropods , a diverse array of vertically-transmitted endosymbiotic microorganisms have been discovered that alter sex ratio or sex determination , such as bacteria of the genera Cardinium , Arsenophonus , Spiroplasma , Rickettsia , and Wolbachia , as well as microsporidian fungi [53 , 54] . The best example are wolbachiae , which can induce different reproductive changes in arthropods , including feminization , parthenogenesis , male killing and sperm-egg incompatibility [55] . Examples that viruses change host sex ratio are a masculinizing virus in the common pill-bug ( Arthropoda: Isopoda: Armadillidium vulgare ) [56] and a male-killing virus-like RNA sequences in the oriental tea tortrix moth ( Homona magnanima ) [57] . The PpNSRV-1-mediated pattern is consistent with a female-killing phenotype . Although male-killing sex ratio distorters are common in nature [53 , 58] , female-killing is rare and would seem to be maladaptive for the virus , as it is transmitted vertically through females . However , unlike most other sex-ratio distorting elements , the virus is also vertically transmitted through males . It is possible that male transmission is more effective ( e . g . because male mates multiply , thus increasing transmission to other families ) , and transmission through males can be further enhanced if uninfected females produce a secondary sex-ratio bias towards daughters , which is common in P . puparum . In this regard , spread of the virus could be enhanced in the same manner of the male-biasing supernumary chromosome psr in the parasitoid was N . vitripennis [59] , and the masculinaizing virus in the isopod A . vulgare [56] . In both these cases , the male-distorting element is enhanced by female biased sex ratios in populations . P . puparum wasps generally produce a female-biased sex ratio[25 , 51 , 52 , 60 , 61] , which is consistent with Hamilton’s local mate competition theory [62] and is a phenomenon found in many other parasitoid wasps [63 , 64] . PpNSRV-1 also increases longevity of infected males and females , suggesting that the virus may provide mutualistic benefits to the wasps . Clearly , more research needs to be performed to substantiate this and other hypotheses about the effects of PpNSRV-1 on its hosts .
The laboratory colonies of Pteromalus . puparum wasps and its butterfly hosts , the small white ( Pieris rapae ) , were initially collected from cabbage fields in the experimental farmland of Zhejiang University , Hangzhou , China in 2012 and maintained as described previously [28] . The wasps that had successfully parasitized host pupae were individually tested for the presence of PpNSRV-1 using the methods described below . Offspring from a single PpNSRV-1 ( - ) or PpNSRV-1 ( + ) breeding pair was reared to produce the PpNSRV-1 ( - ) or PpNSRV-1 ( + ) colony . Primers for viral genome sequencing were designed based on the virus genome-like contig discovered during transcriptional profiling of P . puparum wasps ( S2 Table ) . Total RNA from adult female wasps was extracted using TRIzol ( Invitrogen , California , USA ) . RNA concentrations of each sample were measured using Nanodrop 2000 ( Thermo Scientific , Wilmington , DE ) . Single-strand cDNA was synthesized from the RNA using the TransScript One-Step gDNA Removal and cDNA Synthesis SuperMix Kit ( TransGen Biotech , Beijing , China ) . cDNA was used as a template for PCR . The terminal sequences of the viral genome were confirmed by 5′ or 3′ RACE according to the instructions of the SMART RACE cDNA amplification Kit ( Clontech , California , USA ) . All amplified PCR products were cloned into pGEM-T Easy vectors ( Promega ( Beijing ) Biotech Co . , Ltd . Beijing , China ) and sequenced . Nucleotide sequence analysis and assembly were performed using DNAStar software version 5 . 02 ( Madison , WI , USA ) . ORFs of the PpNSRV-1 genome were predicted using open-source NCBI ORF finder ( https://www . ncbi . nlm . nih . gov/orffinder/ ) . For each ORF , signal peptides and transmembrane regions were predicted using open-source Phobius web server [65] . Molecular weight and isoelectric point ( pI ) of the predicted ORFs were calculated via open-source ProtParam ( http://web . expasy . org/protparam/ ) . Potential phosphorylation sites were determined by open-source NetPhos 2 . 0 [66] , and glycosylation sites were determined by NetOGlyc 4 . 0 [67] and NetNGlyc 1 . 0 ( http://www . cbs . dtu . dk/services/NetNGlyc/ ) . To identify putative conserved transcription termination and initiation sequences of the PpNSRV-1 genome , we analyzed noncoding viral genome regions using the open-source MEME suite of motif-based sequence analysis tools [68] . Northern blot analysis was performed to confirm transcription of predicted PpNSRV-1 ORFs . Total RNA from PpNSRV-1 ( - ) or PpNSRV-1 ( + ) wasps were extracted as described above . Separation of RNA was performed using formaldehyde gel electrophoresis , followed by transfer to Hybond N+ nylon membranes ( GE Healthcare ) by upward capillary transfer in 20 × saline sodium citrate buffer and cross-linking by UV illumination . DNA fragments were amplified by primer pairs ( S2 Table ) based on the each ORF sequence and were labelled using the PCR DIG Probe Synthesis Kit ( Roche , Mannheim , Germany ) . Hybridization was performed at 42°C and the detection was carried out using the DIG-High Prime DNA Labelling and Detection Starter Kit II according to the manufacturer’s instructions ( Roche ) . Multiple alignments of amino acid sequences were performed using open-source Clustal Omega ( http://www . ebi . ac . uk/Tools/msa/clustalo/ ) and edited by open-source GeneDoc ( http://www . softpedia . com/get/Science-CAD/GeneDoc . shtml ) . The NCBI PASC classification tool was used for pairwise comparisons of viral genomes [69] . The phylogenetic tree was constructed using the maximum likelihood method with 1000-fold bootstrap resampling using MEGA 5 . 05 software [70] . Accession numbers of analyzed mononegaviral genomes are listed in S1 Table . To detect PpNSRV-1 in P . puparum wasps , two pairs of primers amplifying 506 bp and 523 bp fragments , VDA-1/VDS-1 and VDA-2/VDS-2 , were designed according to the genomic sequence of PpNSRV-1 ( S2 Table ) . Total RNA from each sample was used to synthesize cDNA as described above . PCR was run as follows: 30 s at 94°C , 30 s at 55°C , and 45 s at 72°C for 35 cycles . Amplifications were visualized by 1% agarose gel electrophoresis and ethidium bromide staining . qPCR was used to quantify the viral load of PpNSRV-1 . An absolute standard curve was constructed from a plasmid clone of the corresponding PpNSRV-1 genome region using specific primers ABVA/ABVS ( S2 Table ) . PCR products were cloned into pGEM-T Easy vectors and then sequenced . Standard curves were generated by determination of copy numbers ( 103–108 copies ) of standard plasmid . qPCR was performed using the Bio-Rad CFX 96 Real-Time Detection System ( Bio-Rad , Hercules , CA , USA ) with SYBR Premix Ex Taq II ( Tli RNaseH Plus ) ( Takara Bio , Otsu , Japan ) . Thermal cycling conditions were: 94°C for 30 s , 40 cycles of 95°C for 5 s , and 60°C for 30 s . Three replicates of samples of each group were performed . The equation of y = -0 . 3174x +12 . 43 ( y = the logarithm of plasmid copy number to base 10 , x = Ct value , R2 = 0 . 9937 ) was used to calculate the copy number of PpNSRV-1 genomes . Different tissues or groups of tissues from adult female and male P . puparum wasps were dissected to evaluate tissue tropism of PpNSRV-1 . Female and male wasps from virus-infected colonies that were fed on 20% ( v/v ) honey for 5 days were placed into the glass tubes . The tubes were put on the ice , and the wasps were chilled for a few seconds . The head , thorax and abdomen of each wasp was separated using a dissecting microscope ( Leica , Wetzlar , Germany ) and directly removed into TRIzol for RNA extraction . The abdomina from female or male adult wasps were further dissected to gain digestive tracts , ovaries or testes . For each tissue , the PpNSRV-1 genome copy number was measured by qPCR . A pool of fifteen wasps was used for RNA extraction for each sample as one replicate , which was repeated three times . The extraction of total RNA , the synthesis of single-strand cDNA , and qPCR were conducted for each sample as described above . To define the expression profile of PpNSRV-1 in P . puparum wasps at different developmental stages , eggs , larvae , and pupae from both female and male adults ( with the age from 1 day to 7 days post-eclosion ) were collected , and then placed directly into TRIzoL . To evaluate whether P . puparum wasps could clear PpNSRV-1 infection at high environmental temperature , both female and male adults ( aged 1 , 3 , 5 , or 7 days post-eclosion ) were fed at 25°C or 35°C and collected to determine viral titers . For qPCR , total RNA of each developmental sample was extracted and single-strand cDNA synthesized as described above . A pool of five wasps was used for each replicate , and the experiment was repeated three times . A cDNA fragment ( 1 . 9 kb ) of PpNSRV-1-ORF I was amplified using primers VORF1A/VORF1S ( S2 Table ) . Purified PCR products were cloned into vector pET-28a ( + ) and confirmed by DNA sequencing . The construct was used to transform Escherichia coli BL21 ( DE3 ) using standard procedures . Bacterial cells were collected by centrifugation and disrupted by sonication . The insoluble recombinant His-tagged protein ( 72 . 7 kDa ) was purified using Ni-chelating affinity columns ( TransGen Biotech , Beijing , China ) under denaturing conditions . To confirm the identity of the recombinant protein , proteins were separated by SDS-PAGE , transferred to polyvinylidene difluoride ( PVDF ) membranes ( Sigma , St . Louis , MO ) by a semi-dry electrophoretic transfer system ( Bio-Rad ) , and detected with an anti-His monoclonal antibody conjugated to horseradish peroxidase ( HRP ) ( HuaAn Biotechnology , Hangzhou , China ) as described previously [71] . Signals were visualized with an enhanced chemiluminescence detection system ( Super Signal West Pico Chemiluminescent Substrate; Pierce , Rockford , IL ) . The purified protein was submitted to HuaAn Biotechnology Company as an antigen for immunization in male New Zealand rabbits . The obtained polyclonal antibody to PpNSRV-1 ORF I protein was purified from antiserum by the company . The company reported an antibody titer of 1:1000 . For immunohistochemistry , ovaries , eggs , digestive tracts and male reproductive organs from PpNSRV-1 ( - ) or PpNSRV-1 ( + ) wasps were dissected , washed , and handled as described previously [72] . The primary antibody was rabbit anti-PpNSRV-1-ORF I ( diluted 1:100 in phosphate-buffered saline ( PBS ) containing 5% goat serum ) . The secondary antibody was DyLight 549-conjugated goat anti-rabbit ( diluted 1:200 in PBS , Abbkine , Redlands , CA , USA ) . The actin cytoskeleton was stained by FITC-phalloidin ( green; diluted 1:500 , Good Biotech . , Wuhan , China ) . The nucleus was stained with 1 μg/ml 4′-6-diamidino-2-phenylindole ( DAPI [blue] , Good Biotech . , Wuhan , China ) . Samples were analyzed and images recorded using a Zeiss LSM 780 confocal microscope ( Carl Zeiss SAS , Jena , Germany ) . A stack of consecutive confocal optical sections ( Z-stacks ) were recorded at 8 bit . Figures were merged and scale bars were added using Zeiss LSM ZEN 2010 software ( Carl Zeiss SAS , Jena , Germany ) . Adobe Photoshop CC ( Adobe Systems Inc . , San Jose , CA ) was used for image grouping . All samples were analyzed with the same microscope and software settings . Ovaries and digestive tracts of P . puparum wasps were dissected from 5-day-old adult females using a dissecting microscope ( Leica , Wetzlar , Germany ) . Samples of ovaries and digestive tracts were pre-fixed overnight at 4°C with 4% glutaraldehyde in phosphate buffered saline ( PBS; 0 . 01 M , pH 7 . 4 ) . Samples were washed in PBS three times every 15 min and post-fixed with 2% osmium tetroxide ( OsO4 ) in PBS for 1–2 h , and washed again . Samples were dehydrated by a graded series of ethanol ( 30% , 50% , 70% , 80% , 90% , 95% , and 100% ) for about 15 to 20 min at each step and transferred to absolute acetone for 20 min . Cells were infiltrated with a 1:1 mixture of absolute acetone and the final Spurr resin mixture for 1 h at room temperature , transferred to 1:3 mixture of absolute acetone and the final resin mixture for 3 h , and to a final Spurr resin mixture overnight . Finally , each specimen was placed in an Eppendorf tube containing Spurr resin , incubated at 70°C for more than 9 h , and then sectioned using an ultramicrotome ( LEICA EM UC7 , Wetzlar , Germany ) . Ultrathin sections were double-stained by uranyl acetate for 5 min and alkaline lead citrate for 10 min and then were observed using a transmission electron microscope ( Hitachi Model H-7650 TEM , Tokyo , Japan ) at an accelerating voltage of 80 kV . To confirm virus particle morphology , we purified virus particles from PpNSRV-1 ( + ) wasps . Two thousand adult wasps were ground , and the virus particles were purified as described previously [73] . The ground samples were dissolved in 0 . 1 M phosphate buffer ( pH 7 . 0 ) . After removing impurities by centrifugation at 12 , 000×g for 20 min at 4°C , a mixture of equal volumes of chloroform and butanol was added to the supernatant and mixed for 20 min at room temperature until the stratification disappeared . The mixture was centrifuged at 12 , 000×g for 20 min at 4°C . Virus particles resided in the upper ( butanol ) phase ) . Virus particles were precipitated by ultracentrifugation ( Beckman Coulter Life Sciences , Indianapolis , IN , USA ) at 33 , 000×g for 4 h at 4°C . The resulting precipitate was resuspended in 0 . 01 M of PBS ( pH 7 ) and then centrifuged to remove impurities ( 12 , 000×g for 20 min at 4°C ) . The resulting virus sample was frozen at -70°C for further investigation . Purified particles were stained by phosphotungstic acid and observed using a transmission electron microscope . The newly-emerged wasps from PpNSRV-1 ( - ) and PpNSRV-1 ( + ) colonies of P . puparum wasps were individually kept in sterilized glass tubes and fed with 20% honey solution for 24 h . For each crossing experiment , 10 pairs of female and male wasps were allowed to mate for 24 h . Both female and male wasps were then individually moved to new sterilized glass tube containers ( 18 × 82 mm ) and reared by providing one newly-pupated small white butterflies for the female wasp to parasitize . Before emergence , female and male wasp pupae were individually dissected from parasitized butterfly pupae to avoid offspring mating . Both dissected female and male wasp pupae were individually kept in sterilized glass tubes until emergence . Once emerged , all the female or male wasps from the same parasitized host pupa were collected for virus detection separately using methods described above . Four crossed experiments were designed: female ( + ) × male ( + ) , female ( - ) × male ( - ) , female ( + ) × male ( - ) , and female ( - ) × male ( + ) . Viral transmission efficiency was calculated based on the wasps emerged from 10 parasitized host pupae as described previously [74] . To investigate the prevalence of PpNSRV-1 in field populations , we collected some parasitized pupae from five different cities of eastern China including Ningbo ( 121 . 56°E , 29 . 86°N , 108 pupae ) , Hangzhou ( 120 . 19°E , 30 . 26°N , 75 pupae ) , Jiande ( 119 . 27°E , 29 . 49°N , 90 pupae ) , Hefei ( 117 . 16°E , 31 . 31°N , 96 pupae ) , and Nanchang ( 115 . 53°E , 28 . 41°N , 84 pupae ) in October , 2012 , respectively . To determine the overall annual PpNSRV-1 , we collected some parasitized pupae from Hangzhou ( 90 pupae ) , Jiande ( 96 pupae ) , and Hefei ( 80 pupae ) in June , 2016 . In the vertical transmission assay , we found that the female parents must be virus-free if the male offspring did not harbor the virus . So we detected the pool individuals to estimate the virus prevalence in the female parents . Once emerged , 5 female and 3 male wasps dissected from the same pupa were collected for virus detection using methods described above . Detection of virus in female or male wasps from the parasitized pupa was considered indicative of whole-colony infection . Prevalence was calculated as described previously [74] . Butterfly pupae not parasitized by P . puparum wasps were also collected from the five locations as mentioned above and individually tested for the presence of PpNSRV-1 using the same method described above to ensure they were uninfected . Major biological parameters including the successful parasitism rate , adult emergence proportion , offspring number and sex ratio , and longevity of P . puparum wasps were compared between the PpNSRV-1 ( - ) and PpNSRV-1 ( + ) colonies and calculated as described previously [60] . Total oocytes number in ovaries was calculated to compare the fecundity between PpNSRV-1 ( - ) and PpNSRV-1 ( + ) wasps . Using a dissecting microscope , 15 female wasps from the PpNSRV-1 ( - ) and PpNSRV-1 ( + ) colony were dissected to count oocytes in each ovary . 30 female wasps from the PpNSRV-1 ( - ) and PpNSRV-1 ( + ) colony were paired with males , allowed to mate for 48 h , and then placed individually in glass tube containers with one newly pupated butterfly . During the following 5 days , each pupa encountered by the female wasp was individually transferred into a new glass tube daily , and the newly-pupated pupa and cotton ball containing honey solution ( 20% ) was replaced . After the wasp offspring emerged , both female and male offspring from each pupa were collected , counted , individually moved into different glass tubes , and fed with 20% honey solution in cotton balls . The number of dead female or male adult wasps was counted daily , and the life span ( from emergence to death ) of each adult was calculated . Meanwhile , wasps that failed to emerge from parasitized pupa were also counted after peeling off the pupae . To check whether PpNSRV-1 could kill wasps at environmental temperature higher than 25°C , some newly emerged wasps ( both PpNSRV-1 ( - ) wasps and PpNSRV-1 ( + ) wasps ) were individually placed into tubes and reared at 35°C . The biological parameters were calculated as the following: successful parasitism rate ( % ) = number of host pupae parasitized by female wasps / total number of host pupae exposed to female wasps × 100; offspring number per parasitized pupa = number of emerged female or male wasps from the same parasitized pupa; offspring number per female wasp = number of emerged female or male wasps from all pupae parasitized by one female wasp; adult emergence proportion ( % ) = number of emerged adult wasps from a parasitized pupa / total number of emerged and unemerged adult wasps × 100; and offspring sex ratio ( % ) = number of female adult wasps emerged from a parasitized pupa / total number of female and male adult wasps emerged from a parasitized pupa × 100 . Crude virus liquid was prepared from adult wasps of a PpNSRV-1 ( + ) colony as follows . Firstly , thirty ovaries from female wasps were dissected and placed into a 2-ml tube with 60 μl of PBS buffer ( 0 . 01 M , pH 7 . 4 ) . Then a steel bead ( 2 . 5 mm in diameter ) was placed into the tube , and the tube was shaken for 1 min at the speed of 25/s on a TissueLyser II ( Qiagen , Maryland , USA ) . The homogenate was centrifuged at 12 , 000×g for 10 min at 4°C , and the resulting supernatant was subsequently filtered with a 0 . 22-μm sterile filter ( EMD Millpore , Billerica , MA , USA ) . Virus-free crude liquid was prepared from the adult wasps of the PpNSRV-1 ( - ) colony using the same method . The viral load of the liquid was quantified by qPCR . All samples were stored at -70°C . To further confirm the effect of PpNSRV-1 infection on P . puparum wasps , we injected 50 nl of virus crude liquid ( 105 copies/μl ) into each yellow wasp pupa from the PpNSRV-1 ( - ) colonies . The control was injected with an equal volume of virus-free crude liquid . The quantification of PpNSRV-1 in the wasps 3 days post-eclosion from injected wasp pupae was performed by qPCR . Then , we selected 30 injected female wasps and allowed them to have 48 h opportunity to mate to investigate the successful parasitism rate , adult emergence proportion , and offspring number . Newly emerged wasps [female: 34 PpNSRV-1 ( - ) , 35 PpNSRV-1 ( + ) ; male: 30 PpNSRV-1 ( - ) , 26 PpNSRV-1 ( + ) ] from injected wasp pupae were individually placed into tubes and reared as described above for determination of longevity . To evaluate the effect of PpNSRV-1 on P . puparum wasps , we utilized RNAi knockdown of PpNSRV-1 ORF I to reduce the abundance of PpNSRV-1 , and then evaluated effect of this knockdown on wasp fitness parameters . The ORF I-specific primers and primers targeting enhanced green fluorescent protein gene ( eGFP; negative control ) were designed with added T7 promoter adaptors ( S2 Table ) . All amplified PCR products ( 400–600 bp ) were cloned into pGEM-T easy vectors ( Promega [Beijing] Biotech ) and sequenced . The correct PCR products were used as templates for dsRNA synthesis with the MEGAscript T7 Transcription Kit ( Ambion , Austin , TX ) , according to the manufacturer’s instructions . Synthesized dsRNA was purified by phenol/chloroform extraction and isopropanol precipitation , dissolved in diethylpyrocarbonate-treated water , and quantified using a NanoDrop 2000 Spectrophotometer ( Nano-drop Technologies , Wilmington , DE ) at 260 nm . We injected 50 nl of dsRNA ( 2 × 103 ng/nl ) into each yellow wasp pupa from the PpNSRV-1 ( + ) or PpNSRV-1 ( - ) colony . The PpNSRV-1 titer in the wasps emerging from injected wasp pupae was quantified by qPCR . Then , 30 injected female wasps that had mated with injected male wasps for 48 h were selected to investigate the successful parasitism rate , adult emergence proportion , and offspring number . Newly emerged wasps were collected to measure longevity as described above . All values were expressed as mean ± standard error ( S . E . M . ) . For qPCR and RNAi results , data were analyzed using a one-way ANOVA or two-way ANOVA followed by Tukey’s multiple comparison test ( p <0 . 05 ) . The means on each same measured biological parameter were compared between two colonies with and without carrying PpNSRV-1 using Student’s t-test ( **p <0 . 01 , *p <0 . 05 ) . Linear regression analysis was constructed between viral detection rate and mean temperature from the five different cities in October , 2012 , and the three various cities in June , 2016 . Temperature was calculated based upon data from the Weather Underground website ( https://www . wunderground . com ) . All statistical calculations were performed using Data Processing System ( DPS ) software ( version 14 . 50 ) [75] . | Although a few viruses with RNA genomes have been discovered in parasitoid wasps , their numbers are limited and their effects on the wasps have rarely been determined . Unambiguous negative-sense , single-stranded RNA [ ( - ) ssRNA] viruses in parasitoids wasps have not been described . In this report , we identify a novel ( - ) ssRNA virus from a parasitoid wasp , verify the presence of virion-like particles in various wasp tissues , and characterize the viral genomic structure , transcription strategy , phylogenetic relationship , transmission strategy , and biological impacts on its host . Most importantly , we reveal that this novel virus mediates secondary sex ratio of its host by decreasing female offspring numbers , and can be transmitted both by male and female to offspring . Decrease in female offspring could be due to increased female mortality or to alterations in the primary sex ratio . The reduced female offspring number per female could therefore be a fecundity cost or a tradeoff for increased longevity of the wasp . However , it remains to be determined whether increased longevity under laboratory conditions translates into increased lifetime fecundity for the wasp in nature . Increased longevity could be beneficial to the virus by promoting its transmission to more hosts . The secondary sex-ratio distortion induced by the virus may also increase its transmission in host populations , because males transmit the virus to their offspring and can mate with multiple females , thus further spreading the virus . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"ecology",
"and",
"environmental",
"sciences",
"reproductive",
"system",
"pathology",
"and",
"laboratory",
"medicine",
"parasitic",
"diseases",
"animals",
"developmental",
"biology",
"pupae",
"genome",
"analysis",
... | 2017 | A novel negative-stranded RNA virus mediates sex ratio in its parasitoid host |
Human movements show several prominent features; movement duration is nearly independent of movement size ( the isochrony principle ) , instantaneous speed depends on movement curvature ( captured by the 2/3 power law ) , and complex movements are composed of simpler elements ( movement compositionality ) . No existing theory can successfully account for all of these features , and the nature of the underlying motion primitives is still unknown . Also unknown is how the brain selects movement duration . Here we present a new theory of movement timing based on geometrical invariance . We propose that movement duration and compositionality arise from cooperation among Euclidian , equi-affine and full affine geometries . Each geometry posses a canonical measure of distance along curves , an invariant arc-length parameter . We suggest that for continuous movements , the actual movement duration reflects a particular tensorial mixture of these canonical parameters . Near geometrical singularities , specific combinations are selected to compensate for time expansion or compression in individual parameters . The theory was mathematically formulated using Cartan's moving frame method . Its predictions were tested on three data sets: drawings of elliptical curves , locomotion and drawing trajectories of complex figural forms ( cloverleaves , lemniscates and limaçons , with varying ratios between the sizes of the large versus the small loops ) . Our theory accounted well for the kinematic and temporal features of these movements , in most cases better than the constrained Minimum Jerk model , even when taking into account the number of estimated free parameters . During both drawing and locomotion equi-affine geometry was the most dominant geometry , with affine geometry second most important during drawing; Euclidian geometry was second most important during locomotion . We further discuss the implications of this theory: the origin of the dominance of equi-affine geometry , the possibility that the brain uses different mixtures of these geometries to encode movement duration and speed , and the ontogeny of such representations .
As a first approximation , perceived physical space is assumed to be Euclidian . Yet , psychophysical studies of visual perception , drawing movements and locomotion indicate important departures from Euclidian geometry [1]–[4] . In these cases , space and movements seem to be perceived in terms of affine geometrical properties [2] , [5]–[8] . Affine geometry is the geometry which retains from Euclidian geometry only the existence of points , lines and planes with their geometrical properties of incidence ( i . e . , the existence of only one straight line between two points ) and parallelism ( i . e . the existence of a unique line parallel to a given line that passes through a given point , Thales' theorem , etc . ) . The study of affine geometry can also be based on displacement of points by vectors , ( see section A in Text S1 ) . There is no preferred absolute distance in affine geometry . Most important in affine geometry is the set of affine transformations , which are transformations of space or of a plane transforming straight lines into straight lines and parallel lines into parallel lines . Concretely , affine transformations are obtained by composing together translations and linear mappings which include rotations , stretching and dilatations . A property of a geometrical shape is said to be affine invariant when it is preserved under all possible affine transformations . For instance , being a closed curve is an affine invariant property but enclosing an area equal to is not . Being an ellipse is an affine property , but being a circle is not , since any given circle and any elongated ellipse can be transformed one into the other using at least one affine transformation . Since we are interested in motion timing , it is important to understand the concept of invariant duration with respect to a given set of transformations . For instance , a timing rule for a given set of curve segments is affine invariant if the duration spent moving along any arc of any one curve is equal to that spent moving along the image of this segment obtained by using any affine transformation . Thus , if timing were a totally affine invariant for all possible planar movements , all elliptical trajectories , for example , would have had the same total duration , generating a complete isochrony . We will show how such an invariance follows from a particular dependence of motor timing on the curvature of trajectories . The influence of path curvature on movement velocity is well known [9]–[11] . Originally , Viviani and Terzuolo [1] claimed that movement velocity is roughly proportional to the radius of curvature of a curved movement , i . e . , and that movement segmentation is determined by the presence of inflection points . Modifying this earlier suggestion and based on empirical observations , Lacquaniti et al . [12] formulated the two-thirds power law , stating that the instantaneous angular velocity is proportional to the instantaneous curvature raised to the power 2/3rd . Equivalently , since , an alternative formulation of this law is: , where is the ordinary tangential velocity and is the radius of curvature . The coefficient was termed the velocity gain factor and was shown to be piecewise constant . Examining the dependence of on the perimeter during periodic drawings of different figural forms ( circles , ellipses , figure eights , double ellipses etc . ) , Viviani and McCollum [13] obtained a relationship consisting of multiplying two power laws , such that where is the Euclidean perimeter of the figural form and and are empirically determined exponents . These observations , where the value of the velocity gain factor depends on the perimeter of the curve being drawn , were thought to account for the isochrony principle . This principle captures the empirical observation that the durations of movements involved in the generation of motion paths with similar geometrical forms but with different lengths are nearly equal [4] , [9] , [13]–[15] . The 2/3 power law was extended to human locomotion [16] , [17] with a somewhat different set of exponents and was also linked to the perception of visual motion [18] , [19] . Pollick and Sapiro [5] and Handzel and Flash [6] , [20] have further suggested that the 2/3 power law is equivalent to movements being performed at a “constant equi-affine speed” , defined as the time derivative of , the equi-affine arc-length . Using the Euclidian radius of curvature and arc-length , and , respectively , the equi-affine arc-length is defined as: ( 1 ) According to this definition , the equi-affine length corresponds to the integral of the infinitesimal regular Euclidian arc-length weighted by the Euclidian curvature raised to the power of 1/3 . Thus , among equally long segments , those with greater curvature have longer equi-affine length . Based on this approach , Flash and Handzel [8] further developed a framework using group theory to describe and analyze human movements . However , a significant empirical observation which was not accounted for by the equi-affine description , nor by any other model , is the observed tendency towards global isochrony of human movement , mentioned above . Moreover , neither the equi-affine description , nor any other model has made any explicit suggestion as to how the values of the velocity gain factor are selected for in any movement segment . As part of our new approach , we treat movement generation as being based on full affine geometry , without making specific choices of units of length or area . This allows us to compare the influences of affine , equi-affine , and Euclidian geometries on the temporal properties of the movements . We suggest that in each of these geometries time is proportional to a specific “measure of distance along the curve” in that particular geometry . In addition , we deduce the variation of the velocity gain factor from the need for full-affine invariance . This results in a local form of isochrony . For instance , for movements along ellipses , a full affine invariance predicts the same 2/3 principle as equi-affine invariance . But , in addition , it predicts that through appropriate adaptation of the “velocity gain factor” movement duration will be the same for all ellipses in the plane . However , total isochrony may sometimes lead to paradoxical behavior , and we therefore hypothesize that the brain takes advantage of the existence of several possible geometries rather than using a single geometry . Hence , we suggest that movement timing is continuously prescribed and realized according to an equilibrium between affine and Euclidian geometries with Equi-affine transformations , which are the area-preserving affine transformations , playing an essential , if not dominant , role [5] , [6] , [8] . Combining geometries is a totally new approach; it has not been previously considered in mathematics nor in biology , let alone in motor control or vision research . Here , using this new approach to the timing of motion , we derive new guidelines for motion segmentation and for the identification of motion primitives , while treating both hand trajectories and locomotion within the same framework . The idea that geometric invariance is of great importance in prescribing the principles underlying perception and action is quite old [21]–[23] . Since that time many psychophysical studies have discussed the importance of invariance theory for perception [24]–[27] . Summarizing informally ( see [26] ) , an invariant entity came to mean “anything which is left unaltered by selected transformations” [28] . However , the concept of invariance has benefited from mathematical formulation as initiated by Galois [29] ( see [23] , [30]–[32] ) , and it is this concept of invariance that serves as a conceptual basis for our theory . Galois has stated that in solving any given equation , it is more important to understand the structure of the ambiguity among all the possible solutions of this equation , rather than trying to directly derive them . Moreover , this roundabout approach frequently offers better means for computing such solutions . Different levels of analysis of the given equation are characterized by the sets of transformations of the solutions which are equivalent at those particular levels of analysis . It is natural to propose that , in the same way , the brain uses several levels of representation and processing in planning any particular motion . At each level the computation is organized by respecting certain symmetries . Approaching the level of motor execution fewer and fewer possibilities are allowed , thus reducing the initial larger group of symmetry of all possible movements into smaller groups . This hierarchy of decisions in motion planning and execution is reflected in the representation of space through the performed movement . In the following mathematical section , which discusses full affine , equi-affine and Euclidian geometries , we use Cartan's moving frame method [33]–[35] , whose main theme is the relation between a specific curve and the action of a group of transformations of frames defined along that curve . For instance , given a specific group of transformations we look for a parameterization of the curve which is invariant under such transformations . When curve segments are similar under transformations belonging to that group , the parameterization of these segments will also be similar . In particular , we show how Cartan's moving frame method is well suited for our problem of trajectory planning and segmentation . From the work of Galois [29] , Cayley , Jordan , Lie , [36] , [37] , Poincare [22] and Klein ( cf . [38] , [39] ) , we see that a particular geometry is captured by a particular group of transformations of the points of a space or of a plane , such that every point or every direction in can be transformed by an element of to every other point or direction . Euclidian geometry corresponds to being the group of rigid displacements consisting of translations and rotations but we can also choose to be the group of all possible affine transformations , generated by translations , rotations , reflections , but also by dilatations , stretching and shearing . Or we can choose the special subgroup of the full-affine group , namely the area-preserving equi-affine group , which includes all the above transformations except for dilatations . The last two groups correspond to the full affine and to the equi-affine geometries , respectively . These three geometries - Euclidean , full affine and equi-affine - are the most important geometries for the two-dimensional ( 2D ) plane , in our present investigation . A priori the largest possible group of invariance to be considered is that containing all continuous smooth transformations of geometrical objects in the plane . This group , however , does not provide sufficient constraints , since all possible trajectories are equivalent under such transformations and are expected to have the same duration . Another possibility is using the group of projective transformations ( consisting of compositions of several pairs of perspective projections and describing changes in the perceived positions of observed objects when the point of view of the observer changes ) . However , some projective transformations paradoxically send points at a finite distance to infinity . The affine group is formed by projective transformations for which this does not happen . Thus , it describes the largest possible reasonable invariance . At the other extreme is the group of isometries , i . e . the Euclidian group of length-preserving rigid transformations . The sub-group of equi-affine transformations lies between and . Curves can be analyzed differently in different geometries ( see Monge [39] , Lie and Cartan [33] ) . Cartan's method generalizes the moving frame method originally developed by Darboux [39] . Note that , for the case , any frame is formed by a point and by two attached basis vectors ( see section A in Text S1 ) . The essence of Cartan's moving frame method is that it creates a correspondence between the different orders of description of trajectories and the possible coordinate frames on the plane . This method specifies which geometrical transformations of frames allow identification of the trajectories that are indistinguishable at a given order ( see section A in Text S1 ) . At each point in time , locations within the plane are represented by coordinates in a moving frame . The motion along any curve is described by the equations representing the new infinitesimal coordinate frame ( the new location and the new basis vectors ) within the instantaneous current frame . When the moving frame is the canonical moving frame , the only remaining varying coefficient is the instantaneous curvature . This is an invariant quantity of the geometrical curve in the geometry defined by the group of transformation . Thus , when using the moving frame description all along the curve , the representation of the infinitesimal next frame uses only invariant quantities and , in this sense , is the simplest possible one . The choice of parametrization of the curve necessary for deriving the canonical form of the moving frame gives a unique parameter , which is also the only parameter invariant under any transformation belonging to the group of the chosen geometry . In full affine geometry , this unique parameter is called the full affine arc-length and it is denoted by ( see section A . 2 in Text S1 ) . With the same kind of analysis in Euclidian geometry , we obtain a canonical parametrization by using the ordinary Euclidian distance ( arc-length ) instead of , while in equi-affine geometry we obtain canonical parametrization by using the equi-affine arc length . To connect this description with known kinematic notions , choosing time to be proportional to the Euclidian arc length gives rise to an ordinary uniform Euclidian motion , i . e . , to a motion with a constant tangential Euclidian velocity . Setting time to be proportional to gives rise to a motion with a constant equi-affine speed . This is a motion whose tangential velocity obeys , where is the Euclidian radius of curvature and which is a constant , is the so called velocity gain factor defined by Lacquaniti et al . ( 1983 ) . Given a point on a curve , there exists a unique equi-affine frame , centered at with coordinates , whereby the curve near the point takes the following simple form , called the reduced equation of the curve: ( 2 ) Cf . [33] . This frame is the equi-affine canonical frame . The coefficient , which depends on the point appearing in this equations , is the so called equi-affine curvature . The equi-affine moving frame equations are the only infinitesimal equations for which the motion is expressed as follows: ( 3 ) where denote the basis unit vectors of the equi-affine canonical moving frame ( see section A . 2 in Text S1 ) . The mathematical expression for the equi-affine curvature when expressed as a function of Euclidian radius of curvature is: ( 4 ) where are , respectively , the first and second order derivatives of with respect to the Euclidian arc-length . Equi-affine curvature is the quantity defined on a planar curve which remains invariant ( unchanged ) under equi-affine transformations . The curves of constant equi-affine curvature are all plane conics . Those with are ellipses , those with are parabolas , and those with are hyperbolas . Focusing now on the canonical full affine moving frame , the derivative of the full affine arc-length with respect to the Euclidian arc-length is expressed by: ( 5 ) When time is set proportional to the full affine arc-length , the velocity gain factor varies according to the equation: . The canonical full-affine moving frame is simply obtained by scaling the vectors in ( 3 ) : it is formed by and by ( see section A . 2 in Text S1 ) . Another kind of curvature appears for the full affine frame . This full-affine curvature remains unchanged under all full affine transformations . It determines the relative variation of the velocity gain factor with the full affine arc-length : ( 6 ) Two kinds of special points generically arise along a path immersed within the affine plane: ordinary inflection points ( at which the Euclidean curvature equals zero ) and ordinary parabolic points ( at which the equi-affine curvature equals zero ) . Near ordinary inflection points , when the distance measured from the inflection point , tends to zero , diverges like and shrinks like ( see section A . 3 in Text S1 ) . Thus , as neither nor are defined at ordinary inflection points , neither affine nor equi-affine parameterizations can be used at or near such points . Similarly , near a parabolic point , degenerates like , and only or can be used . However , as shown in section A in Text S1 , by attributing a weight of 1/4 to the full affine arc-length and a weight of 3/4 to the equi-affine arc-length we get a parameter such that , which offers a unique , convenient strategy for moving through an inflection point while keeping equi-affine invariance of time . Summarizing the main results: at each point along a parameterized curve , locations within the plane are represented by coordinates in a moving frame . The motion along the curve is then expressed by equations representing the new location and the new basis vectors within the present frame . Given a group of plane transformations , there is a canonical moving frame along any curve . For this canonical moving frame , with its intrinsic arc-length parameter , the movement equations have the simplest possible form . The only changing coefficient is the instantaneous curvature , which is the only geometrical invariant of the curve in the geometry defined by . This results in the unique parametrization of the curves which is invariant under . This parametrization is the only one for which the motion of the frames is of minimal complexity , in the sense that changing variables from the current to the next frame is represented by an invariant quantity .
We propose here that the brain selects movement timing and duration according to a principle of geometrical invariance . A few principles are necessary to derive timing and kinematics of motion from geometry . As mentioned above , all the geometries considered here define canonical coordinates along curves: , and are the invariant arc-length parameters of the affine , equi-affine and Euclidian geometries , respectively . Using these parameters and assuming some specific relation between time and the corresponding parameter , the principle of the invariance of time becomes concrete . By definition , we will call monotonic any movement or any part of a movement during which duration is proportional to one of these invariant parameters . That is , in the case of planar movements , the affine invariance selects time such that Δt = C0Δσ . For movements with equi-affine invariance , and for movements with Euclidian invariance . The constants fix the scales of the corresponding durations but may also depend on the context , the subject and his/her intention to move quickly or slowly , etc . Let us remark that monotony necessarily neglects the fact that the motions start and end at rest with zero velocities , accelerations , and higher order derivatives of position ( i . e . , that we also make discrete movements ) . Because of the presence of such boundary conditions , the model must be generalized by considering the canonical parameters , or in the corresponding geometries ( Euclidian , equi-affine or full affine , respectively ) to be polynomials of some order of . The first or second derivatives of these polynomials are zero at the movement end-points . This gives perfectly invariant timing for discrete trajectories which start and end at rest . Thus , any rule such that is affine invariant in the sense that if one applies an affine transformation to a given curve , then the time for the transformed curve follows the same rule . Similarly , a function gives timing which is equi-affine invariant , and a function timing which is Euclidian invariant . This defines a clear general notion of “geometrical time” for all the three geometries , in which monotony is only the simplest possible case . However , in our initial presentation , all our tests of the validity of the new theory will be limited to periodic movements . In many cases , only one geometrical parameter is insufficient for deriving movement kinematics and it is both necessary and useful to use a combination of several geometrical parameters . Even for periodic motions the presence of singularities implies that monotony cannot be obeyed . For instance , the full affine parameter expands at an ordinary inflection point . Thus , if is the chosen time parametrization , it would take an infinite amount of time to reach an inflection point . In contrast , at an ordinary inflection point shrinks , requiring an infinite speed of movement when passing though such a point . The latter phenomenon also holds for parabolic points for the parameter , which shrinks near such points . A priori it would have been natural to expect that Euclidian geometry dominates near inflection points , because the curvature is zero at these points . However , our preliminary observations have indicated that this is not the case . To the contrary , it seems that at inflection points , Euclidian velocity is eliminated and a stereotypical mixture of affine and equi-affine velocities are used . This was the first case where we saw the advantage of having a mixture of several geometries . This conforms with mathematical analysis: as recalled above in the section sec∶math_pre , we showed ( in section A in Text S1 ) that a unique combination of affine and equi-affine parameters offers a convenient strategy to passing through inflection points . Combining several geometrical timing parameters offers greater flexibility and adaptivity to the motion planning strategy . Consequently , our general hypothesis is that movement duration results from the combined use of several geometries . That is , during some portion of any given movement the velocity will be more affine , while during another portion it will be more equi-affine or more Euclidian . This is formulated by expressing as a multiplication of some power functions of the canonical geometrical differential parameters . For simplicity we assume that there are intervals of time during which these combinations are stationary . Between these special intervals the time parameter is chosen by smooth interpolation . The consequence of this hypothesis is the existence of a succession of segments belonging more or less to different fixed combinations of geometries . It is natural to expect that the existence of singular points such as inflection and parabolic points implies the presence of extended segments in their vicinity , during which a mixture of geometries are employed . Note that the global shape of a figural form often forces the presence of such singularities , thus we predict that the global shape of a trajectory will influence its local kinematics . When moving from one movement segment to the next , the transition between such segments should be smooth . Hence , all of the above additional guidelines can be summarized as giving rise to a tendency for expanding singularities , motion segmentation and smoothness . To quantify the combination of geometries in selecting movement timing and total duration , we need to understand the different and modifiable weights attributed by the motor system to the various possible purely geometrical rules . For this purpose we propose an equation having an exponential form . Let denote the expected Euclidian velocity under constant affine , equi-affine and Euclidian velocities , respectively , where time is proportional to the respective geometric parameters . If is proportional to , the Euclidian velocity is proportional to and we mark it by . If is proportional to , the Euclidian velocity is proportional to and we mark it by and if is proportional to , the Euclidian velocity is a constant , , and we mark it by . Hence the Euclidian velocities corresponding to these three different choices of time are: ( 7a ) ( 7b ) ( 7c ) We examined the recorded tangential Euclidian velocity by assuming that the actually realized Euclidian velocity is prescribed according to the following product equation: ( 8 ) where and are weight functions defined along the trajectory with values lying within the range [0 , 1] . The above equation ( 10 ) can be rephrased using the following tensorial equation for movement duration: ( 9 ) A multiplicative form of the mixtures of velocities is more natural than an additive form because the parameters belong to different dimensions . In particular the treatment of the constants is technically easier when using a multiplicative form . Another reason became apparent during our mathematical analysis: as mentioned in the mathematical preliminaries section , it is possible to bypass inflection points by combining the logarithmic functions of the affine and equi-affine velocities using the weights of 1/4 and 3/4 , respectively ( see section A . 3 in Text S1 ) . Finally , the subjectively perceived velocity is related to the physical one by a nonlinear law , well approximated by a power law [40] , [41] , so the logarithmic function is able to associate the different velocities with their subjective perception . Observe that , although affine transformations introduce additional complexities to the computations , the actual hypothesis to be tested is contained in equations ( 0 ) , ( 10 ) above , which are easy to understand . Note that all functions used here were determined based on the experimental data , except at inflection points where we have chosen them to be and at parabolic points where we selected . ( See sections Experimental tests and Methods for the precise process . ) Hence , at the present stage the theory is descriptive rather than predictive . We have also tested less stringent consequences of using mixtures of invariance . In particular , we tested the possibility that during specific segments one geometry becomes more dominant than the others . A vivid manifestation of the existence of a pure geometric velocity can be obtained by comparing the times spent on two arcs such that there is a planar affine transformation with : Suppose that the time spent on a segment of a curve is affine invariant , even if affine velocity is not constant we have ( 10 ) Suppose now that the law for duration is purely equi-affine . If denote the total areas under respective arcs of curves and the corresponding secants ( i . e . , the total areas between the respective arcs of the curves ( ) and the corresponding straight segments joining the extremities of those arcs in the plane ) , then ( 11 ) Hence duration varies according to area . When Euclidian geometry is the dominant kinematic law , if we denote the lengths of the arcs traveled by , then ( 12 ) Now suppose that a curve is a union of curved segments ( not necessarily connected ) , , each being dominated by one of the three geometries , where each geometry is marked by one of three indexes ( i . e . 0 for affine , 1 for equi-affine and 2 for Euclidian ) , and suppose that movement velocity varies continuously along . It can easily be shown from the continuity of the speeds at the boundaries between adjacent segments ( see section E . 1 in Text S1 ) that the ratios of times spent on different segments , , are invariant under any similarity transformation . In other words , they are invariant under the scaling of Euclidian length with a scale factor . A consequence of this assumption is the following ( see section E . 1 in Text S1 ) : Suppose is cut into two parts , then there are three non-negative constants , depending only on which are invariant under similarity transformations of , such that , and ( 13 ) where mark the times spent on , respectively , and is the ratio of similarity between and . The data used to test our hypotheses were derived from recorded hand movements and locomotion generated along a priori prescribed curves . We aimed to test the compatibility of the temporal properties of the movements with respect to two main principles: 1 ) geometrical invariance determines movement duration , 2 ) The mixing of different geometries can account for movement segmentation . Three different tests were conducted . The first , using elliptical hand trajectories , tested whether an alternation between Euclidian and affine geometries better explains the experimentally observed relation between Euclidian curvature and velocity than describing the whole elliptical movement as obeying a single power law with constant exponent . This test also examined the limitations of the validity of the isochrony principle by investigating the relation between total duration and perimeter and the relation between the enclosed area and gain factor . The second test used trajectories generated by human subjects while tracking geometrically prescribed complex figural forms - cloverleaves , lemniscates and limaçons - during both drawing and locomotion . This experiment tested whether the proposed tensorial formulae ( 8 ) , ( 9 ) can successfully account for the experimentally observed movement durations . We also examined whether it is possible to distinguish between movement duration during drawing and locomotion based on the different degrees of influence of the different geometries , i . e . whether both tasks are based on similar principles of invariance but arise from different mixtures of geometrical invariance . The third test , applied to the same data as the second test , examined the validity of equation ( 13 ) with respect to the ratios between the durations of the movements along the large versus the small loops of the lemniscates and limaçons . It also aimed at confirming the differences between drawing and locomotion identified by the second test with respect to the influence of the different geometries on the durations of movement along large versus small segments .
We present a new theory explaining how the timing of voluntary movements changes according to path geometry . Our model proposes that the velocity along the path is a specific composition of different canonical velocities , a composition that may vary among different segments of the same movement . Both geometrical invariance and movement segmentation are consequences of this principle . The canonical velocities to be combined depend only on path geometry and are defined within three major 2D-spaces: equi-affine , Euclidean and affine spaces . Equi-affine geometry is associated with a measure of area , Euclidian geometry with a measure of distance and affine geometry with the notion of parallelism . The above notions are illustrated in our analysis of elliptical hand drawings . The trajectories contained two types of curved segments , each displaying different relationships between instantaneous velocity and Euclidian curvature and corresponding either to affine or Euclidian geometries . The Euclidian segments were those portions of the trajectory during which the Euclidian curvature was rather low - below some threshold - while the affine segments corresponded to the more curved portions . Such a description of segments accounts for the observed behavior better than a description based on a single power law ( the probability of providing a better model than a single power law model was always higher than 0 . 64 and up to 0 . 97 for small ellipses ) . The observed behavior is a compromise between constant ordinary Euclidian speed and the isochrony principle , which reflects the effect of full affine geometry on motor timing . For drawing the three figural forms studied here ( cloverleaves , lemniscates and limaçons ) , the comparison of the three canonical velocities with the corresponding experimentally recorded ones strongly supports our new theory . The predicted purely equi-affine and the experimentally recorded velocities were very close for 70% of the time . The disagreement during the remaining 30% , could be systematically explained: here the velocity departed from entirely equi-affine and varied in a direction indicated by full affine or Euclidian velocities , as shown by the velocity profiles ( see Figures S4 and S5 ) . For instance , for drawing limaçons the difference between the actual and equi-affine velocities showed a tendency towards the full affine velocity , while during locomotion along similar curves , the discrepancy between the actual and equi-affine velocities suggested an influence of the constant Euclidian velocity ( see Figures S4E and S5E ) . To quantify these relations , we analyzed the experimentally recorded trajectories of human drawing or walking along the prescribed figural forms mentioned above . More than 90% of the velocity variance of drawing movements and 60% of the velocity variance for walking ( based on the measure ) was explained by the combination of several geometries . We also showed that for locomotion our model provides more information on motor timing than the constrained minimum-jerk model . For drawing , our model is only slightly better than the minimum-jerk model and both models are excellent . Given that the minimum-jerk model has no adapted parameters and that , in contrast , our model of geometrical mixture involves the selection of up to 30 parameters for each trajectory , it was necessary to compare these models using a standard penalty score such as the AIC . Our model generally remained more successful in accounting for the data than the minimum-jerk model ( see Figures 3 and 4 ) . At first sight , the flexibility offered by three different geometries seems so large that one could imagine that such a model would produce a good fit for any possible movement data set . It was therefore important for us to ask in what way are our results non-trivial ? Firstly , the prediction that speed is a weighted product of all three canonical velocities is non-trivial since a priori the observed speed could have exceeded the envelope corresponding to a linear combination of the logarithmic functions of these velocities , obtained while the total sum of their weights is precisely equal to 1 . 0 . Secondly , we verified that the existence of several segments , during which constant combinations of the canonical velocities could account for the observed velocity , is unlikely due only to chance . Thirdly , we have shown that considering the number of free parameters and accordingly using the AIC scores , which appropriately quantify adapted measures of goodness of fit , our model successfully accounted for the observed data . Hence , based on these arguments , our principal result can be formulated as follows . The tensorial combination of canonical invariant parameters gives rise to statistically non-trivial predictions which were not rejected by the data against which they were examined . Moreover , this is a relatively simple model , which is grounded on a general point of view about the brain's mode of functioning . The new point of view provided by the geometrical combination of velocities permits us to demonstrate several characteristics of motor timing . First , we demonstrated that the global shape and size of the trajectory essentially influence motion timing ( Figure 8 ) . For instance , when drawing the three limaçons , subjects used affine geometry ( responsible for isochrony ) more than when drawing cloverleaves and even more so than when drawing lemniscates . However , on average , the influence of the ratios of the sizes of the large versus the small loops on the full affine weight was negligible . In contrast , during locomotion , it is remarkable that the Euclidian weight grew linearly with the ratios of the size of the large versus the small loops . Since the total perimeter remained constant , a decrease in the size of the smaller loop resulted in increasing the size of the larger loop . Thus , a possible explanation of the growth of is that low Euclidian curvature without a change in convexity makes Euclidian geometry dominant for locomotion ( as we verified directly for ellipses during drawing ) . Note that for lemniscates , based on the theoretical study of singularities , we imposed exactly at the inflection points . The good agreement achieved with the experimental data confirms this hypothesis . Second , we discovered that the main difference between drawing and locomotion was the opposite degree of influence of full affine versus Euclidian geometries . For drawing , was important and varied between 0 . 1 and 0 . 6 , while for locomotion it was that varied between 0 . 2 and 0 . 6 and hence was more important than . Possible reasons for these differences are differences in the control strategies used , or the existence of different biomechanical constraints . We also applied a more restricted idea of segmentation by studying the effect of alternating between different dominant geometries . As a first approximation we assumed that segments with a constant velocity only within one specified geometry can successfully account for the observed ratios of time spent moving along the large versus the small loops of the complex figural forms as a function of their respective sizes . The observed ratios of movement durations have also corroborated and provided further evidence for our finding that the net balance between Euclidian and affine geometries is totally reversed for drawing versus locomotion . All these results confirmed our expectation that affine geometry is significant in a theory of movement timing . The canonical velocity of full affine geometry yields the same total time spent on a curve and on the curve obtained through any dilatation . That is , the dominance of affine geometry here corresponds to isochrony , even though it is imperfect . However , we found experimentally that the pure affine geometrical arc-length is generally only a secondary component in determining movement timing , although it is always present during drawing movements . The notable exception is hand drawing of limaçons , where full affine canonical velocity alone explains the entire timing pattern very well . These results point to the important role of equi-affine geometry in motor timing . The equivalence of the 2/3rd power to moving at a constant equi-affine speed [8] expresses the dominance of equi-affine geometry in trajectory planning . It has recently been found that 3D human drawing movements also tend to be generated at a constant equi-affine speed [46] , [47] . How does this dominance of equi-affine geometry arise ? A first possibility is that the equi-affine invariant parameter may be computationally simpler . This invariant parameter is of third order , i . e . , the order of the variation of acceleration , namely jerk . It could be coded by proprioceptive or vestibular information especially during locomotion [48] , [49] . In contrast , full affine invariant parametrization is of the fifth order . Moreover , the equi-affine subgroup of transformations is uniquely defined by the full affine group , even without fixing a unit of area . Thus , affine invariance can be canonically broken into equi-affine invariance , thus simplifying the necessary computations . A second possibility is the probable importance of area perception during motion , and we know that equi-affine transformations preserve areas . The amplitude of a piece of motion can be judged from the area enclosed by the corresponding segment of the trajectory and by the straight segment joining its initial and final positions . As we have seen in sec∶quantitative laws ( see also E in Text S1 ) , for monotonic equi-affine trajectories , the ratio of the total times spent along trajectory segments is a function of the ratios between the above enclosed areas . A third explanation may correspond to the link between equi-affine invariance , the optimization of smoothness and the minimum jerk principle ( cf . [4] , [50] ) or the minimum variance principle , both giving similar results . Todorov and Jordan [44] observed that the 2/3 law is equivalent to nullifying the normal component of the instantaneous jerk . Related to this is the attractive power of parabolic points and parabolic segments ( see [51] , [52] ) , because parabolic segments , for which equi-affine geometry is the only possibility , minimize jerk , obey the 2/3 power law and are invariant under equi-affine transformations . This link between the 2/3 power law and the minimum-jerk model may also be the root for the agreement between these principles in explaining motion timing from path geometry . The fourth explanation for the dominance of equi-affine geometry is based on dimensional analysis which provides a completely different direction of support for the 2/3 law . Let us suppose that during motion , the total variation of energy over each unit of time and mass is equal to a constant for all the segments of the trajectory . This constant has the dimension , where , and mark the units of mass ( kg ) , length ( meters ) and time ( seconds ) , respectively . From the constancy of , the dimension of time then becomes equivalent to and the dimension of the resulting velocity is . For turbulent fluid flows this is the well-known relation of Kolmogorov and Oboukhov between the length scale and the mean variation of velocity ( see [53] , [54] ) . Note that the 2/3 law for movement duration only refers to the radius of curvature , similarly to the radius of a vortex in the Kolmogorov-Oboukhov law . All the above explanations for the dominance of equi-affine geometry in movement timing arise from some sort of invariance . However , in the framework of our theory , it is natural to propose that the main reason for the dominance of the equi-affine geometry ( and consequently the 2/3 law ) is that it offers an excellent compromise between full affine invariance and the reduction of computational complexity . We propose that movement duration is determined by invariance and computation . In the present framework , movement duration is related to space . This agrees well with Piaget's [55] suggestions about the development of children's conception of time: “The psychological interpretation of temporal notions … is that time forms a coordination of movements of different speeds . ” The production of time jointly with geometrical form also agrees well with hypotheses on the neural basis of temporal processing , see e . g . Mauk and Buanomano [56] , stating that “… given the intricate link between temporal and spatial information in most sensory and motor tasks , timing and spatial processing are intrinsic properties of neural function , and specialized mechanisms are not required . Rather temporal processing may rely on state-dependent changes in the network dynamics . ” Our suggestions also fit those of Bernstein [57]: “There exist in the higher levels of the central nervous system projections of space , and not projections of joints and muscles . ” The present study should be understood as presenting a new repertoire of organizing principles that operate at higher levels of the motor system and may be considered as a possible source for the definition of kinematic primitives . Several other recent studies have also reported strong departures from the two-thirds power law [58] , [59]; for locomotion , see also [16] , [17] . These studies either employed different experimental paradigms from those used here or offered alternative explanations for the observed violations of the power law . Schaal and Sternad [58] studied the movement patterns of a human arm ( with seven degrees of freedom ) during the generation of elliptical trajectories . The magnitude of the deviations from the power law depended on the perimeters of the trajectories and on their orientations with respect to the subject's body . To account for these observations , Schaal and Sternad suggested that subjects realize the required elliptical patterns by employing smooth oscillatory pattern generators at the joint level and that the values of the exponent in the power law depends on the geometrical transformations from joint to hand coordinates . We suggest that the geometric combinations we show are also affected by the geometrical transformations from joint to hand trajectories . Hence , our model , though considering movement duration only from the point of view of hand trajectories , must be further developed to consider Schaal and Sternad's approach and results . We believe that the motor system has evolved to make simplifications in motion planning compatible with the biomechanical characteristics of the musculoskeletal system . Examining the generation of different patterns of complex figural forms in various tasks and conditions ( tracking , drawing from memory , tracing ) Flanders et al . [59] also showed significant differences in the values of the exponent of the power law depending on the size and orientation of the trajectory . In particular the authors suggested that strategic or cognitive factors affect the relation between hand velocity and curvature . This points to many possible extensions of our study . In fact , even if a combination of geometries accounts for the link between geometry and movement duration , we suspect that the rules dictating the mixture of geometrical timing parameters chosen by the brain may depend on external or cognitive factors . It will be particularly interesting to examine in what ways cognitive factors and learning [60] contribute to the effects that global shapes of trajectories have on their local kinematics . Our present study is limited to 2D motion . Future work should deal with 3D Motion , as well as with movements performed in different orientations , as in [58] , [59] , and at different depths with respect to the body . This will certainly require considering additional geometries and may reveal certain failures of invariance due to the influence of biomechanical factors . One limitation of the present study is that we only tested periodic arm movements and locomotion trajectories . However , as suggested above , our principles may also be applied to discrete movements that start and end at rest , or to trajectories containing reversals of movement direction . In such cases all the different geometries are expected to be combined and one may need models that no longer assume constant canonical velocities . Thus new kinds of segments are expected to emerge which depend on the particular velocities combined and on the values of the different geometrical curvatures associated with these velocities . The principal limitation of the present study is that the tests of the theory have not dealt with the question of which geometrical paths the trajectories should follow . We have only dealt with the question of which velocity is chosen along a prescribed path , as a function of the geometrical form being followed . It is probably not difficult to propose which special paths should be selected , depending on the degree of geometric invariance and symmetry they offer . For instance , as suggested in [51] , [52] , parabolic arcs are selected in affine geometry , because they remain invariant under several families of transformation . Thus their group of symmetry is rather large . Some optimization principles predict the complete actual trajectories . Using only via points and end-points , the minimum jerk principle successfully accounted for both the trajectory path and velocity of curved and drawing movements [4] , [61] . Similarly , when the path is fully prescribed , the constrained minimum jerk model successfully predicts the velocities along such paths [44] , [50] . The minimum variance principle [62] , the optimal feedback control model [63] , and the minimum time principle in the presence of signal-dependent noise [64] do predict movement duration , but so far only for point-to-point movements . The minimum variance principle is grounded on Fitts [65] and Schmidt's laws [66] based on the dependence of average movement duration on movement amplitude and error tolerance , achieved through either a logarithmic function of the ratio ( Fitt's law ) or a linear relationship ( Schmidt's law ) of this ratio . This ratio is invariant under affine transformations , since only ratios between lengths of parallel segments and not absolute values of length appear in these laws . Fitt's and Schmidt's law are therefore a priori compatible with affine geometry . In many cases , the predictions of the minimum jerk or minimum variance principles are almost compatible with geometrical invariance . The use of invariance or mixtures of invariance as proposed here is only a constraint . To realize the actual movements , subjects must apply tools other than , but compatible with , invariance . For example , the geometrical invariance principles can be used together with optimization principles to solve redundancy problems at the task level . Even more importantly , the anticipation of singular points before and during movement generation can help particularly in determining where the motion should be segmented or the precise combination of the canonical geometrical parameterizations to be used . More generally , geometry may indicate in what parametric space or coordinate frames motor commands for movement generation should be planned . Our suggestion does not contradict the need for online optimization of ongoing movements . When needing to anticipate or to respond optimally to trajectory perturbations , optimal feedback , control theory can complement our formulation of invariance principles [63] . This combination of the planning of geometrically invariant trajectories with the anticipation of both geometrical singularities and expected perturbations could allow control of the optimal selection of the relevant parameters . We emphasize that our model relates naturally to the neural encoding of movement because it suggests the possibility that different neural populations represent movement kinematics in terms of the different geometries or combined geometrical representations: Evidence has accumulated for the use of different “reference frames” in movement planning ( [67]–[71] ) . For instance , the parietal cortex codes movement in head- or gaze-centered coordinate frames [72] , [73] , the putamen in a body reference frame [74] and the hippocampus [75] , [76] in an environmental reference frame , etc . Furthermore , there is ample evidence that different or even “mixed” coordinate frames are used within the posterior parietal cortex which may be well addressed by the concept of the mixture of geometries suggested here . Target and hand locations during arm movements are represented in terms of eye-centered coordinates , while the motor error between target and hand positions are represented with respect to a hand-based coordinate frame ( for review see [77] ) . In relation to the theory presented here , it may be suggested that the target and the initial hand position are coded in terms of an absolute eye-centered or visually based Euclidian coordinate frames while an evolving coordinate frame , using a motor error between the instantaneous current and the immediate next hand positions , is better characterized as an affine moving frame . The notion of moving frames ( as in section Mathematical preliminaries ) , particularly the affine geometrical representation , may throw new light on the currently available neurophysiological observations and on the roles of different cell populations in representing movement . Schwartz and colleagues [78] , [79] have reported observations consistent with the notion that arm trajectories are well encoded by motor cortical activity in monkeys . A key finding was that the endpoint velocities ( including the speed and movement direction ) are well represented by single cells and by neuronal populations . This is an instantaneous , relative representation and the magnitude of the population vector was shown to obey the 2/3 power law , while the instantaneous movement direction matched the direction of the population vector [78] , [79] . In a recent study Polyakov et al . [52] analyzed the kinematic properties of monkey scribbling movements and the related neural activities of motor cortical units . The scribbling movements were found to be composed of parabolic segments . Using the partial cross-correlation method developed by Stark et al . [80] , Polyakov et al . [52] showed that equi-affine velocity was represented more strongly than the Euclidian speed in the activity of several recorded units and the segmentation of the neural activities predicted parabolic segments . Therefore parabolic segments constitute geometrically defined motion primitives which subserve the construction of scribbling movements . This study has also provided the first direct evidence that equi-affine geometry may be used in the neural coding of arm movements . Consistent with the theory presented here we speculate that there must be many dynamically interconnected neuronal populations , either within one area or more probably within different areas , which use different geometrical representations . These assemblies would be selective for parameters intrinsic to a particular geometry . Some supporting evidence has been obtained in a recent fMRI study [81] . A large number of brain areas responded more strongly to a dot moving along elliptical trajectories with velocities consistent with the 2/3 power law; activity was seen particularly in motor areas ( M1 , PMd , cerebellum and the basal ganglia ) as well as in frontal , cingulate and parietal regions . Brain regions responding more strongly to a dot moving in elliptical trajectories with constant Euclidian were found in the occipital visual areas , the fusiform gyrus and the right parahippocampal gyrus [81] . Neural assemblies within these areas may therefore generate different possible combinations of geometries which may influence movement timing . Analyzing how children draw simple ellipses Viviani and Schneider [43] have shown that both the 2/3 law and the isochrony principle are qualitatively present at 5 years of age and evolve further until age 12 . Variability , and geometrical and kinematic distortions of the drawn trajectories diminish greatly between ages 5 and 7 and continue to diminish thereafter . The exponent of the radius of curvature in the formula , increased from about 0 . 25 at age 5 to 0 . 33 at age 12 . The exponent for the perimeter decreased from 0 . 4 at age 5 to 0 . 2 at age 12 . These findings suggest that Euclidian geometry develops first , followed by equi-affine or affine geometries . Piaget and Inhelder [82] suggested that the chronological sequence of development of geometric intuition in children is: 1 ) topology for the most elementary stages of perception , for which only continuity within the spatial field is important; 2 ) projective geometry , subserving the coordination between prehension and vision through operations that depend on and integrate different points of view; 3 ) Euclidian geometry utilizing proportions and distances , for perception and storage in memory of places and distances , e . g . , in navigation . Piaget and Inhelder suggested that progress is made through the use of concrete operations associated with these different geometries . In their view , affine geometry would be used during the intermediate phases between those associated with projective and Euclidian geometries , e . g . , at the beginning of coordination between gaze direction and the direction of body motion during active exploration . Our data on voluntary movements suggest a different order of development of the different geometries . Implicit motions , unlike explicit or iconic descriptions , seem to be acquired initially using more Euclidian reference frames . This is suggested by the exponent initially being closer to zero in young subjects . Thereafter these exponents move closer to 1/3 [43] , [83] , [84] possibly reflecting the use of equi-affine or affine representations . Every action is a specific solution to a problem . What is a priori undetermined by this solution before it is selected is partly encoded by a particular set of symmetries of space and time , a set permitting possible actions at a given level of computation . Any given decision confines the original symmetry group to a specific subgroup , and an action is ultimately chosen when the symmetry is further reduced to the identity . Similarly to perception , geometrical invariance gives motor actions a structure . The most familiar instance of a particular invariance is the global isochrony principle that we interpret as being a manifestation of the use of full affine geometry . Another instance is the 2/3 power law for parabolic segments . However , to be compatible with the strong Euclidian constraint of the physical world and with the restrictions on the computational capacity of the system , full affine invariance is only achieved through the mixing of invariant canonical durations specified by several geometries , such as the equi-affine and Euclidian geometries . In full affine geometry , time is a pure number ( e . g . , going around any ellipse takes ) . In equi-affine geometry time corresponds to area ( the period of an ellipse is proportional to the cubic root of its area ) , while in Euclidian geometry time corresponds to length . We propose that movement time is continuously selected by the brain based on the combination of these geometrical measures along curves . Still , each individual trajectory is different from all others , since it is associated with a different combination of geometries . Sensation , intention , and cognition can generate particular combinations . The principle of invariance is also compatible with different optimization principles such as the minimum-jerk [50] , [61] or the minimum variance principles [62] , and with optimal feedback control [63] , [85] . It can even offer a framework within which such principles can be formulated . Invariance , information , feedback and optimality must work together in the selection and adaptation of any movement through evolution and development , but we suggest that by constructing the appropriate spaces at each instant of time along the trajectory , geometric invariance is the main component for determining movement timing .
In experimental test no . 1 three young adult men were instructed to draw 10 types of ellipses at 3 different speeds , slow , natural and fast . The ellipses , prescribed in advance , had 3 different eccentricities , , and 3 different sizes , small , medium and large , plus one ellipse which was “as large as possible” called huge . Within each session , each ellipse was drawn 10 times and statistical analysis was performed based on 8 repetitions , ignoring the first and last drawings . ( For further details , see section B . 1 in Text S1 . ) In experimental tests no's 2 and 3 , we analyzed a series of drawings and locomotion trajectories of cloverleaves , limaçons and lemniscates , taken from the studies of Viviani and Flash [4] for drawing movements and from Hicheur et al . [17] for locomotion . For drawing , the trajectories were those of the stylus position along the tablet . For locomotion the trajectories measured were those of the orthogonal projection on the ground of a point corresponding to the mid-point between the subject's shoulders . To verify the stability of the geometrical models , the trajectory of a point marked as the R-point was also considered . The R-point is the intersection on the floor of the line connecting the M-point with the mid-point between the sensors positioned on the subject's chest and back ( see Figure S3 . For the results of the R-point see section D . 4 in Text S1 , and related figures . In particular , Figure S9 shows the results we obtained for the locomotion data using the R-Point . Figure S10 shows for both drawing and locomotion ( M-Point and R-Point ) , the mean values of the functions for the different figural forms and for different subjects . Figure S11 shows the mean values of the functions , separately for the small and big loops of the limaçons and lemniscates , for drawing and locomotion ( both for the M-Point and R-Point ) . In both analyses we started with a collection of point coordinates , registered at time intervals of for drawing and of for locomotion ( 200 Hz and 60 Hz respectively ) . This gives N points . From the total sample set , a smooth geometric trajectory was constructed , without considering the actual timing . This was achieved by separately approximating the position data for the and coordinates using two Fourier series and with 8 harmonics ( each with 17 independent real coefficients ) , being the parametrization used for the Fourier series . From the smooth path we derived the various curvatures ( Euclidian , equi-affine , affine ) and deduced the 3 monotonic velocities ( see equation set 7 ) . In addition we computed the velocity predicted by the constrained minimum-jerk model [44] . For both types of calculation we selected the corresponding time parametrization which is independent of the actual experimental one . We now needed to find the correspondence between the experimental time series of position coordinates and the position on the smooth path obtained from the Fourier approximation . Hence , we calculated by projecting each point of the experimental trajectory on the Fourier approximated path . We then used the new parametrization to derive velocities for the smoothed experimental data ( for more details see section D . 2 in Text S1 ) . These calculations were conducted on the entire N samples obtained from the drawing and locomotion data . For locomotion we called this data set the complete sampled data set ( CSDS ) . For locomotion , we also extracted the data corresponding to positions where the point attained a local minimum altitude above the ground , giving points . We called this a stepwise sampled data set ( SSDS ) . To compare the different velocity profiles we needed to compare velocities occurring at the same points along the same curve . To do this , we found a set of points located at an equal Euclidian distance from each other . For all models , the velocities at these points were calculated using a standard cubic spline interpolation . Note that the number of independent raw data points used for calculating each value of experimental velocity profile was 5 or less , so the number of “statistically independent degrees of freedom” used below was estimated as for drawing and for CSDS for locomotion . For SSDS for locomotion all were used . The velocity derived according to the constrained minimum-jerk model depends only on one parameter corresponding to the total time . The other theoretical velocities , termed affine , equi-affine and Euclidian “uniform velocities” were computed based only on the path coordinates . To choose the mixture of these uniform velocities which results in the predicted combined velocity , we looked for segments of the experimental velocity during which we could set at least one of the weight function β's as a constant . We then used a cubic spline interpolation for computing the remaining functions' between these segments . ( For locomotion we used the experimental step velocity , based on the SSDS samples ) . Seven different algorithms were used for this calculation . The geometrical combination chosen was that giving the best theoretical velocity profile compared with the experimental velocity and which involved the lowest number of parameters . In the first algorithm we used a linear regression in logarithms of velocity and found segments between points where we could determine and , such that the experimental velocities could be well approximated by: ( 15 ) representing straight lines ( ) with a length of at least 30 points . Here and mark the absolute values of the Euclidian and equi-affine curvatures , respectively . This equation is based on equations ( 7 ) , ( 8 ) and the fact that . For the second algorithm we a priori imposed . The new equation we obtained from equation ( 15 ) is: ( 16 ) As in the first algorithm we found segments during which the equation represents a straight line ( ) . We then used spline interpolation to set the values of the weight function between those segments . In the third and fourth algorithms we considered the combination of affine and Euclidian geometries and the equi-affine and Euclidian velocities respectively . The equations used were , respectively , ( 17 ) and ( 18 ) As in the second case , we constructed the theoretical velocity from the segments of straight lines . The velocities constructed in cases two , three and four were marked as , and , respectively . we explicitly used these velocities for the last three algorithms . By dropping the assumption that equals zero we obtained . Hence , for we looked for segments of straight lines in the new equation ( 19 ) All these algorithms were based on the following arguments . First , we expected to find segments during which a constant combination of geometries appears the primary source for movement segmentation . Second , we had no reason to believe that constant combinations of all three geometries would appear at the same time , so we looked for two-by-two constant combinations . However , to reduce the number of parameters , based on the data we limited the algorithm to the same pair of geometries all along the trajectory . This procedure required verification that these segments ( primary and secondary ) were statistically non-trivial . We therefore used a Fisher's test ( see below ) , as explained in section Experimental tests . Our modeling approach also required verification that the success of the model was not only a consequence of our using a large number of fitted parameters . For this purpose we used the Akaike criterion ( AIC ) as explained in section Experimental tests . The F-test: The data used were those of the logarithms of the velocities . For each curve and for each of the seven computational scenarios , let denote the union of the special intervals of total length and the complementary part of the curve , of total length . Recall that on the model logarithmic velocity was directly extracted using the values of and a linear combination of two of the calculated . The number is the residual sum of squares , , on : ( 20 ) The quantity is the total sum of squares , , for the entire curve: ( 21 ) where is the mean value of the experimental velocity . As usual we can write where , and we hypothesize that the random variables within are independent of the variables within . With this hypothesis the scaled ratio follows a Fisher law ( see [86] ) . Note that the number of degrees of freedom ( df ) for is because . We fixed two independent parameters for each connected special segment The df is as usual for , so we obtain and . The second formula comes from the following decompositionWe then repeated the above computation by replacing the mean value of the experimental velocity by its approximation using a trigonometric approximation up to the fourth order . When using the trigonometric polynomials of degree 4 , we still have , changes from to , because a trigonometric polynomial of degree 4 depends on 9 real numbers , the constant being the mean of the function . For drawing , 61 of 78 trials ( 78% ) showed a P-value of significance equal to 0 . 005 in the F-test . For locomotion 65 of 91 trials ( 71% ) satisfied the test . All the results are shown in Table 4 . This gives the probability that the variance with respect to the mean or , respectively , with respect to the trigonometric approximation of degree four , is sufficient to explain the presence of the detected segments . We verify that this probability is very small according to the standard linear F-test . The results confirm the non-triviality of the existence of segments . The Akaike test [45] , [87]: if is the number of data samples , is the number of independent data samples , and is the number of parameters adapted from the data and used by the tested model plus one , we used the following expression: ( 22 ) | No existing theory successfully accounts for several amazing properties of biological movements: dependence of movement speed on path curvature , isochrony ( movement duration is nearly independent of its size ) and the construction of more complex movements from simpler building blocks . Here we present a new theory of movement generation , based on movement invariance with respect to geometrical transformations . Several types of transformations are considered . Euclidian transformations preserve lengths and angles; affine transformations , which are less restricted , preserve parallelisms between lines , while equi-affine transformations preserve both parallelism and area . Each geometry is associated with a different measure of distance along curves . Movement timing is continuously prescribed by the brain by combining different “geometrical times” each assumed to be proportional to the measure of distance of the corresponding geometry . Movements are constructed by using a series of instantaneous ( Cartan ) coordinate frames . The predictions of the theory compared well with experimental observations of human drawing and walking . Equi-affine geometry was found to play a dominant role in both tasks and is complemented by affine geometry during drawing and by Euclidian geometry during locomotion . The proposed theory has far reaching implications with respect to brain representations of motion for both action production and perception . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"neuroscience/motor",
"systems",
"mathematics",
"neuroscience/theoretical",
"neuroscience"
] | 2009 | Movement Timing and Invariance Arise from Several Geometries |
Stress , pervasive in society , contributes to over half of all work place accidents a year and over time can contribute to a variety of psychiatric disorders including depression , schizophrenia , and post-traumatic stress disorder . Stress impairs higher cognitive processes , dependent on the prefrontal cortex ( PFC ) and that involve maintenance and integration of information over extended periods , including working memory and attention . Substantial evidence has demonstrated a relationship between patterns of PFC neuron spiking activity ( action-potential discharge ) and components of delayed-response tasks used to probe PFC-dependent cognitive function in rats and monkeys . During delay periods of these tasks , persistent spiking activity is posited to be essential for the maintenance of information for working memory and attention . However , the degree to which stress-induced impairment in PFC-dependent cognition involves changes in task-related spiking rates or the ability for PFC neurons to retain information over time remains unknown . In the current study , spiking activity was recorded from the medial PFC of rats performing a delayed-response task of working memory during acute noise stress ( 93 db ) . Spike history-predicted discharge ( SHPD ) for PFC neurons was quantified as a measure of the degree to which ongoing neuronal discharge can be predicted by past spiking activity and reflects the degree to which past information is retained by these neurons over time . We found that PFC neuron discharge is predicted by their past spiking patterns for nearly one second . Acute stress impaired SHPD , selectively during delay intervals of the task , and simultaneously impaired task performance . Despite the reduction in delay-related SHPD , stress increased delay-related spiking rates . These findings suggest that neural codes utilizing SHPD within PFC networks likely reflects an additional important neurophysiological mechanism for maintenance of past information over time . Stress-related impairment of this mechanism is posited to contribute to the cognition-impairing actions of stress .
The prefrontal cortex ( PFC ) plays a central role in a diverse set of cognitive and behavioral processes , including sustained attention , working memory , and behavioral inhibition . In rat , the prelimbic region of the PFC ( plPFC ) is a crucial subregion for these cognitive processes [1]–[4] . Delayed-response tasks of working memory have been extensively used to study the neurobiological basis of PFC-dependent function , in which information is retained during short delay intervals and used to guide subsequent behavior [5]–[11] . Seminal electrophysiological studies identified a subset of PFC neurons that display persistent spiking activity during delay periods of these tasks [7] , [12] . Spiking rates during delay periods are correlated with both specific task-related cues and the number of cues required to be maintained during the delay period . Based on these observations , delay-related spiking activity is posited to reflect the maintenance of attentional processes , abstract rules , or past stimuli and events [13]–[15] . Additional evidence indicates that firing rates of PFC neurons during the response and reward phase of these tasks may reflect decision-related or reward/response outcome evaluation [12] , [13] , [16] , [17] . Currently there are competing hypotheses in the literature regarding the potential effects of stress on PFC spiking activity . One view proposes that stress related increases in norepinephrine ( NE ) α1- and dopamine ( DA ) D1-receptor signaling within the PFC will act to inhibit persistent spiking rates during delay intervals of these tasks [18] . In contrast , it is also posited that stress-related increases in glucocorticoid-receptor signaling will enhance spiking rates by facilitating or increasing glutamatergic neurotransmission [19]–[22] . NE , DA , and glucocorticoids activate multiple receptor subtypes , each producing complex concentration- and receptor-dependent modulatory actions on spiking activity of target neurons [22]–[30] . Moreover , the combined actions of these neuromodulators on target neuron spiking rates during stress are difficult to predict . Indirect evidence also predicts that during stress , high levels of NE and DA may act to disconnect PFC neurons from excitatory recurrent feedback and suppress recursive , delay-related discharge of PFC neurons [18] . These actions are posited to involve the degradation of intrinsic neuronal mechanisms and excitatory recurrent neural connectivity that likely support the maintenance of information over time within PFC networks [2] , [31]–[37] . Although there exists a large body of evidence demonstrating that stress impairs higher cognitive processes dependent on the PFC [38]–[40] , surprisingly , to date the actions of stress on PFC neuronal discharge in animals engaged in tasks of working memory remain unknown . To address this gap in our understanding , we examined the relationship between acute noise stress-related impairment of performance during a PFC-dependent T-maze based delayed-response task of spatial working memory and stress-related changes in plPFC neuronal spiking rates during the delay-period and other components of this task . Additionally , we directly determined the degree to which PFC neurons retain representations of past events over time by quantifying spike history-predicted discharge ( SHPD ) using a conditional intensity-generalized linear model statistical framework ( CI-GLM ) [41]–[45] . With the CI-GLM framework , we addressed several questions related to the actions of stress on PFC neuron function . First , to what degree does past spiking activity of PFC neurons predict or modulate ongoing activity of these neurons ? Second , does stress have an overall impact on the predictability of PFC neuron discharge given a cell's intrinsic spiking history ? Third , do specific task components ( i . e . delay-period vs . behavioral response ) interact with or modulate SHPD during baseline and acute stress ? Combined , these studies represent a first characterization of the predictability of neural discharge from intrinsic spiking history within PFC networks of animals engaged in a cognitive task under normal and acute stress conditions .
Five animals were tested in a T-maze based delayed-response task of spatial working memory ( Fig . 1 and Fig . S2 ) . This task has been previously shown to be dependent on a functionally intact PFC and require PFC-dependent cognitive processes [4] , [24] , [30] , [38] . In this task , animals were required to enter the T-maze arm opposite from the one last visited , following a delay period to obtain food rewards ( chocolate chips 1 . 6 gm ) delivered by the experimenter's hand . Delay length was specific to each animal ( range 10–40 sec ) and chosen to maintain baseline performance near 90% correct . Animals were tested for two sessions a day separated by two hours during which the animal remained tethered to the recording equipment . Each session consisted of 41 trials; the 1st trial of each session was always rewarded and not analyzed . During the first session ( baseline ) , low-level masking white noise ( 60 db ) was presented continuously and animals performed well ( average 93% accuracy ) with 3–4 errors occurring sporadically throughout the session ( Fig . 1B ) . Intense white noise ( 93 db ) , presented throughout the second testing session , significantly impaired performance in this task , with animals performing at an average of 64 . 8% ( −28%; p<0 . 0005 pairwise t-test ) . Acute white noise is a well-characterized stressor that elicits the physiological responses of stress and impairs PFC-dependent cognitive function in both humans and animals [18] , [40] , [46]–[51] . In these same animals , bilateral implants of electrode arrays permitted simultaneous recordings of extracellular discharge activity from layer V of the plPFC yielding 491 spike trains from single neurons ( Fig . 1C–D ) [46] , [52] . A subset of 339 of these neurons were classified as “wide spike” ( WS–type ) based upon action potential features , thus putative glutamatergic pyramidal neurons [53] , and exclusively used for these analyses . Analyses of spiking activity were limited to trials containing correct responses , given so few error trials occurred during baseline testing sessions and reliable estimates of neuronal spiking activity were difficult to obtain . Discharge rates of WS-type plPFC neurons were characterized for intervals of the T-maze task ( e . g . delay , run , and reward ) using a peri-event time histogram ( PETH ) approach . Under baseline conditions , the discharge rate of plPFC neurons fluctuated throughout the time-course of each trial . The pattern of spiking activity was neuron-specific , with neurons exhibiting selective increases in discharge rate during single or adjacent behavioral intervals . A significantly large number of plPFC neurons exhibited delay-related spiking activity during baseline conditions ( 48 . 7%; χ2 = 21 . 4 , p<0 . 001 ) . During delay periods , the discharge rates of these neurons averaged 0 . 55 Hz+/−0 . 047 SEM . The illustrative cases shown in Fig . 2A demonstrate task-related fluctuations in discharge rate corresponding to delay periods , components of the behavioral response , and reward intervals across the recorded population of plPFC neurons during baseline sessions . As illustrated by the examples shown in Fig . 2B–D , the effects of stress on plPFC neuron spiking rates were dependent on the behavioral intervals of the T-maze task . For these cases , stress increased delay-related activity 180% of baseline ( 0 . 018 to 0 . 033 Hz; Fig . 2B ) , whereas response related activity was suppressed during the run phase ( 42% of baseline , 0 . 092 to 0 . 039 Hz; Fig . 2C ) with little effect during the choice phase ( 108% of baseline , 0 . 748 to 0 . 811 Hz; Fig . 2D ) . These opposing actions of stress on delay-related activity versus response-related activity were frequently observed across simultaneously recorded neurons within an animal . Within each component of the task , stress produced cell-specific modulatory actions on spiking rate of plPFC neurons , similar to prior reports on catecholamine neuromodulatory effects [25] , [29] , [52] . Nonetheless , across all recorded WS-type plPFC neurons , stress significantly affected task-related spiking rates of correct trials ( rmANOVA ( TaskComponent ) F ( 5 , 3810 ) = 579 . 05 , p<0 . 0001 ) . During delay-periods , stress significantly increased the average discharge rate of these neurons ( 127% of baseline , Fig . 3 ) . In contrast , stress suppressed the average spiking activity during the run and branch components of the behavioral response ( 78% and 92% of baseline respectively ) . During choice intervals , stress produced a modest increase in the average discharge rate ( 115% of baseline ) but this effect was not statistically significant . Lastly , similar to that seen during the delay period , PFC neuron discharge rates during the reward and pickup components of the task were also facilitated by stress ( 135% and 130% of baseline ) . No differences were observed between the effects of stress on right vs . left trials . As such , these data suggest that intense acute stress generally enhances delay-related activity across PFC neurons . Moreover , the combined actions of acute stress across task intervals support the hypothesis that stress could differentially affect PFC-dependent processes that occur during the delay versus response period of these tasks . PFC neuron spiking activity during the T-maze task was further studied using CI-GLM's to assess whether SHPD significantly contributed to the ongoing activity of PFC neurons and the degree to which acute noise stress altered the predictability of PFC neuron discharge given a cell's intrinsic spiking history . The CI-GLM model ( 1a ) , included covariates representing the background level of spiking activity ( intercept; μ ) , an index of baseline and auditory stress conditions within a testing day , the T-maze behavioral intervals ( delay , run , branch , choice , reward , and pickup ) , and a tenth order autoregressive process during baseline as well as during noise stress conditions with discretized time for increasing spike history durations represented as . Detailed descriptions of this and subsequent equations are presented in the Materials and Methods . From this model , the “spiking gain” of predicted discharge activity during baseline ( α ) and stress ( η ) conditions was calculated for each 250 ms bin back in time . Explicitly , the spiking gain is equivalent to the rate-ratio ( exponentiated covariate weights i . e . α , η ) [54] and represents the fold-change in predicted discharge activity given all other spiking activity occurring during the T-maze task . Individual plPFC neurons exhibited unique patterns of SHPD for each point back in time . For the illustrative cases shown in Fig . 4A , SHPD gains generally decayed with increasing time points in the past . Nonetheless , plPFC neurons frequently demonstrated non-monotonic changes in SHPD gains at specific time points . For example , the highlighted pattern of SHPD gains of a single plPFC neuron illustrates both a decay in SHPD gains over time and a selective increase in SHPD gain at 1 . 75 seconds . Similar to these individual cases the pattern of SHPD gains , averaged across the recorded population of WS-type neurons , decayed exponentially with increasing points further back in time during baseline conditions ( f ( x ) = 1 . 16+0 . 544 ( −x/0 . 25 ) +0 . 51; Fig . 4B ) . Furthermore , although SHPD gains decayed over time , gains remained significantly greater than 1 . 0 , indicating that spike history positively predicted future discharge for at least 2 . 5 seconds . Interestingly , the overall pattern of SHPD gains observed from plPFC neurons differed from other areas of the cortex where SHPD does not significantly modulate ongoing spiking activity after 100 ms [55] . This difference may reflect unique intrinsic neuronal or circuit properties of the PFC . Although acute noise stress did not alter the pattern of SHPD gains over time ( ANOVA ( Stress*Time ) F ( 9 , 3523 ) = 0 . 86 , p = 0 . 560 ) , stress did significantly reduce the magnitude of SHPD gains across all time intervals tested ( ANOVA ( Stress ) F ( 1 , 3523 ) = 8 . 78 , p = 0 . 0031; Fig . 4B inset ) . To confirm that task-related fluctuations in spiking rates were important to include in these models , Model 1a can be compared to the reduced model lacking T-maze behavior intervals covariates ( i . e . ) . A comparison of these two models for each plPFC neuron , demonstrated that the measure of deviance was significantly reduced for the majority of neurons ( 85 . 7% , 187 of 218 neurons; χ2 test , FDR corrected p value <0 . 0113 ) when CI-GLM's included the T-maze behavior intervals ( Model 1a ) . When analyses were replicated with the reduced model , a similar stressor-induced significant suppression of SHPD was found across individual plPFC neurons ( ANOVA ( Stress ) F ( 1 , 4448 ) = 19 . 71 , p<0 . 0001; Fig . S1 ) . Together these results support the general hypothesis that , in addition to task-related fluctuations in spiking rates , spike history plays an important role in shaping plPFC neural discharge . Furthermore , the fact that stress suppressed SHPD in animals performing the PFC-dependent T-maze task likely suggests that SHPD may be an important mechanism supporting PFC-dependent cognitive functions . However , the degree to which these effects are specific to the T-maze task remains to be determined . Reverberations in recurrent cortical circuitry ( Fig . 4C ) or intrinsic neuronal mechanisms likely contribute to SHPD [31] , [35] , [39] , [56] and permit PFC neurons to maintain and preserve information within spiking patterns across large time intervals . However , the fact that stress did not alter the pattern of decay for SHPD gains throughout each trial suggests that stress may not change the underlying mechanism ( s ) generating SHPD . Delay-related spiking activity is posited to serve a pivotal role in the accurate performance of delayed response tasks of working memory [2] , [3] , [18] , [30] , [57] . To determine if stress preferentially affects SHPD during delay intervals , a second CI-CLM was formulated to examine the interaction between spike history and the extended delay interval ( Pickup-Delay; Model 2 ) . This model included the intercept ( μ ) , the main effects of SHPD , baseline and stress conditions within a testing day , and the first-order extended delay interval interaction . The second-order interaction terms of this model , , correspond to delay-specific SHPD during baseline ( α ) or stress ( η ) conditions beyond the main effects of stress accounted for in Models 1a , b . During baseline conditions , the gains of delay-specific , SHPD ( α ) averaged across all WS-type plPFC neurons were all positive and significantly greater than one ( Fig . 5A ) . For the first 1 . 5 seconds of spiking history , gains were essentially flat and averaged 1 . 08+/−0 . 0045 ( SEM ) . During acute noise stress , gains for the most recent spike history intervals ( η ) were suppressed from baseline levels ( ANOVA ( Stress*Time ) F ( 9 , 4671 ) = 2 . 00 , p = 0 . 035 ) . Although stress-related suppression of delay-specific , SHPD gains was observed for intervals up to 1 . 5 seconds , a statistically significant difference from baseline levels was only observed at 0 . 5 and 1 . 25 seconds ( FDR corrected ) . Moreover , delay-specific SHPD gains at the 1 . 25 second time point were not statistically different than 1 . 0 during conditions of noise stress . Given that the comparison between baseline and stress conditions was determined in the same CI-GLM , no deviance tests were performed . We posit that suppression of delay-specific SHPD of PFC neurons during conditions of stress contributes to stress-dependent impairment in delay-related PFC-dependent functions . We next examined the interaction between spike history and the response interval ( Run-Branch-Choice; Model 3 ) :Similar to Model 2 , gains associated with these model interaction terms represent response-specific gain of SHPD beyond the main effects of stress accounted for in Models 1a , b . Response-specific interaction term gains exhibited several important stress-related effects , even though these effects were more complex than the effects of stress on delay-specific interaction gains . First , during baseline recordings , response-related gains for intervals up to 1 . 0 second were not significantly different from one , making no contribution to the prediction of discharge activity ( Fig . 5B ) . Stress significantly increased these response-specific SHPD gains at 0 . 25 and 0 . 75 second intervals ( ANOVA ( Stress*Time ) F ( 9 , 6894 ) = 7 . 3 , p<0 . 0001 ) . Second , under baseline conditions spike history response-related gains at longer intervals increased gradually to become significantly different than one . Stress significantly reduced long spike history interval response-specific gains to values that were equivalent to one ( 1 . 75 , 2 . 0 , and 2 . 5 seconds ) . We posit that the different effects of stress , on delay-specific versus response-specific gain of SHPD , likely reflect differing roles for SHPD in the cognitive or behavioral processes that occur during these behavioral intervals .
The T-maze delayed alternation task embodies a number of important cognitive/behavioral processes and neurophysiological features associated with PFC function of human and non-human primates . This task is highly dependent on the PFC and sensitive to the effects of stress [30] , similar to tasks used in humans to probe PFC function . Additionally , it is posited that the T-maze task requires PFC-dependent processes including working memory , attention , inhibition of proactive interference , and inhibition of distracter interference ( generated by handling the animal between each trial ) , similar to tasks used to probe PFC-dependent function in primates and humans [1] , [33] , [58] , [59] . In animals performing the T-maze task , we observed a significant number of PFC neurons exhibiting delay-related spiking activity , similar to that observed in primates performing delayed-response tasks [33] , [60] , [61] . More automated versions of these tasks , including the Figure-8 maze task , lack distracter interference and possibly other processes requiring significant engagement of the PFC . Such differences between the T-maze task and Figure-8 tasks may explain why automated delayed-response tasks have shown only few delay-related cells in rodent PFC [62] . Intense white noise is a well-characterized audiogenic stressor that impairs working memory , attention , and other PFC-dependent functions in rats , monkeys , and humans [38] , [48] , [51] , [57] , [63] . In the current study , continuous presentation of intense white noise ( 93 db ) impaired performance of the T-maze based delayed-response task of spatial working memory and increased PFC neuron firing rates during the delay period . Similar impairment in T-maze performance is seen when animals are exposed to restraint stress immediately prior to testing [64] , [65] , but not after 4 hours of recovery from restraint stress [22] . Although previous studies have demonstrated that non-stressful white noise can activate PFC neurons [66] , these cells are few in number ( approximately 2% of PFC neurons ) . For these PFC neurons , responses to white noise are phasic , quickly adapting , and are linked to the onset of the stimuli . During presentation of noise stress in the current study , PFC neuron firing rates were increased during the delay period but , importantly , were suppressed during the behavioral response . Together , the above observations provide strong evidence that the effects of noise stress on PFC neuron spiking activity were not induced by a continued sensory response to intense white noise . In the present study , a CI-GLM framework was used to examine the degree to which past spiking activity of plPFC neurons contributes to ongoing neural discharge patterns . Although the CI-GLM approach has been used successfully to distinguish between intrinsic spiking-history related discharge and extrinsic activity in motor cortex [67] , [68] , here we extend its use to examine SHPD in PFC networks . An advantage of this approach over peri-event time histogram analysis or other univariate analyses , including autocorrelegram analysis , is that the CI-GLM approach can disambiguate the relative contributions of spiking history from that of experimental and task-related variables to spiking activity . In the current study , task- and stress-related changes in overall firing rates were captured in the β/B terms of the model separately from the effects of the interaction terms ( α or η ) . Thus , SHPD during baseline could be directly compared to SHPD during stress conditions in a manner that accounted for the effects of stress on the overall discharge rate and task-related fluctuations in spiking activity within each trial . A second consideration for the CI-GLM approach is the use of a Poisson distribution to fit the CI-GLM to neural data over other distributions , including Gaussian or Bernoulli . By doing so , this does not imply that a Poisson process generates plPFC neuronal spiking activity . Instead , the Poisson distribution is the appropriate distribution for what is , in simplified terms , a spiking count-based multivariate regression and provides a computationally tractable solution to fit plPFC neuronal spiking activity . A number of excellent reviews described these statistical modeling methods and the appropriate use of these models to characterize spike trains ( e . g . [45] ) . The present findings demonstrate that under baseline conditions , past spiking activity of a plPFC neuron positively predicts future neuronal discharge . The contribution of past spiking activity to ongoing discharge of plPFC neurons decayed exponentially with time during baseline and stress conditions; spiking-history at time points up to 1–1 . 5 seconds comprised most of the predictive power . With delay periods ranging from of 10–40 seconds and behavioral responses lasting several seconds , we posit that the time scale of SHPD is likely an important PFC neural process for the stable maintenance of information for short intervals during delay periods of working memory tasks while providing a mechanism that allows integration of new spiking activity patterns within PFC networks that permits flexible goal-directed behaviors . This time scale and pattern of decay is an order of magnitude longer to what has been observed in primary motor cortex [67] , [68] . In the motor cortex , SHPD does not contribute to the ongoing activity of those neurons after 100 ms and reflects the refractory and recovery periods of those motor neurons . Thus , these differences likely reflect the excitatory recurrent network connectivity of PFC networks and the time scale on which these cortical networks must maintain information to perform their requisite neurocomputations . Furthermore , the current results suggest that neural codes involving SHPD are likely a complement to firing rate-based codes within the PFC . We found that CI-GLMs which account for modulation in firing rates across different task intervals were significantly better at modeling the spiking activity of the neurons . Moreover , SHPD gains were reduced by approximately 10% when the behavioral intervals are included as CI-GLM covariates , suggesting that both SHPD and components of the task account for the variance in spiking activity . The current study also extends the idea that excitatory recursive activity within the PFC may be critical to sustain spiking activity and may act as the neurobiological basis of working memory and/or other cognitive processes occurring during these tasks [15] , [69]–[71] . Early anatomical and recent computer models have suggested that recurrent neural connections within layers II/III and V of the PFC may support recursive , sustained activity during delay periods of working memory tasks [31]–[35] , [72]–[74] . Together , those studies concluded that sustained discharge generated by reverberating excitatory feedback among anatomically connected networks of neurons during delay periods is the realization of maintaining past stimuli and events , attentional processes , and abstract task rules [13]–[15] , [18] , [75] . However , a large body of research further implicates AMPA and NMDA receptors and intrinsic calcium-dependent mechanisms in the generation and preservation of delay-related sustained discharge [56] . Specifically , the decay of NMDA currents are generally in the range of >80 ms , but some components require seconds to decay [76] , suggesting that NMDA channels and associated maintenance of excitatory currents could explain the maintenance of delay-related sustained discharge . Thus , although the current study confirms that spiking history contributes to PFC neuronal discharge activity , the precise network or cellular mechanisms underlying SHPD remains to be identified . Moreover , the timescale of SHPD for plPFC neurons could result from a combination of intracellular intrinsic calcium-dependent mechanisms [56] , [69] , [76]–[78] , recurrent excitatory connections between neighboring neurons [13] , [32] , [34] , as well as long recurrent paths between other cortical and subcortical brain regions , which remains to be tested . The current study found that acute noise stress increased medial PFC neuron discharge rates during delay periods of rats performing a T-maze task of spatial working memory . In contrast , a number of prior observations predicted that during stress , high levels of extracellular catecholamines were likely to result in a suppression delay related firing rates . During acute stress , NE and DA neurotransmission in the PFC is elevated , activating low-affinity noradrenergic α1- and β-receptors as well as the dopamine D1-receptor [23] , [26] , [27] , [30] . Activation of α1 receptors in the PFC has been shown to suppress delay-related sustained discharge to specific cued target directions during working memory tasks [24] . Similarly , high levels of D1 activation in the PFC suppresses delay-related activity to cued and non-cued target directions [29] . Extensive evidence suggests that the discrepancy between this prior prediction and the current observations is most likely due to stress-related increases in glutamate/glucocorticoid signaling within the PFC during acute stress [19]–[22] . Nonetheless , stress increases signaling within the PFC of a number of neuromodulators that may also contribute to changes in PFC neural activity during stress [79] . Increased delay-related discharge rates could reflect that , during stress , competing patterns of activation are instantiated across PFC neural populations during delay-periods that contribute to a disruption of working memory and sustained attention . Such interpretation of these data is supported by recent findings that a stress-related peptide , corticotrophin releasing factor , and activation of α1- receptors likely facilitates behavioral flexibility and processes involving attentional shifting [79] , [80] . Although delay-related neuronal discharge rates are an important measure of PFC function , neural codes involving SHPD are also likely important for computations driving cognitive/behavioral processes in a number of cortical regions , including the PFC . In the current study , we demonstrate that stress impairs SHPD , principally during delay periods . These observations are consistent with theories that stress likely impairs the ability of PFC neurons/networks to continually update and maintain information necessary for appropriate behavioral responses through the delay period [7] , [13] , [32] , [34] , [36] , [56] , [81] . For example , it has been hypothesized that stress-related increases in NE and DA neurotransmission within the PFC act to modulate intracellular cyclic adenosine monophosphate signaling and hyperpolarization-activated cyclic nucleotide-gated channel function thereby disconnecting PFC neurons from excitatory recurrent feedback [18] . Thus , regulation of recurrent excitatory activation within PFC recurrent circuitry [31] , [39] or regulation of NMDA related intracellular mechanisms [21] , [35] may underlie the current findings that stress impairs SHPD and the ability of plPFC neurons to retain representations of past events over time . Lastly , we found that during the response phase of the task , firing rates of PFC neurons were suppressed during noise stress . Such actions are likely important for the complex repertoire of effects of acute stress across a range of cognitive functions . Suppression of PFC activity during the behavioral response could suggest that during stress , the PFC fails to appropriately inhibit behaviors mediated by other brain regions such as the dorsomedial striatum [82] . Such a loss of behavioral inhibition is supported by observations that habit-based actions , requiring little working memory , are favored under conditions of acute stress [40] . Furthermore , these data also support recent findings that rats prefer choices in decision-making tasks that require the least amount of work for reward following acute stress [83] . During the response phase of the task , SHPD interaction gains were enhanced at short intervals and suppressed at long intervals during acute stress . Such enhancement of short interval gains may reflect PFC activity patterns generated in the absence of ongoing inputs that could result in uncertainty/ambiguity of goal selection [84] . Additionally , it is possible that suppression of SHPD generated from spiking events in the distant past ( i . e . >1 . 5 seconds ) by stress could reflect impairment of the maintenance of information from the delay period into the response interval . Brain regions outside the PFC , including the dorsomedial striatum , could use PFC neural codes involving SHPD to guide selection of appropriate behavioral responses towards a rewarded goal . Nonetheless , the current findings are highly consistent with studies demonstrating that acute stress impairs cognitive functions requiring the PFC , whereas functions not dependent on the PFC such as hippocampal- and amygdala-related processes may be facilitated with stress [18] . In summary , these results provide the first evidence that stress impairs the ability of past spiking activity to predict or modulate ongoing activity within PFC neuronal networks during delay periods of working memory tasks . Regardless of whether spike history-predicted discharge reflects an intracellular [18] , [35] or neural circuit mechanism [31] , [39] , the current observations suggest that stress impairs the ability of PFC neurons/networks to continually update and maintain information through the delay period likely necessary for appropriate behavioral responses . Outside of the delay period , we conjecture that stress-related changes in spike history gain during the response period could represent inappropriate network reactivation related to an inappropriate goal selection [84] , [85] or uncertainty/ambiguity of goal selection . Combined , these studies identify broad effects of stress on PFC neuronal activity that likely represent a key aspect of the neurophysiological bases of stress-related cognitive impairment .
Five male Sprague-Dawley rats ( 300–400 g; Charles River , Wilmington MA ) were individually housed in an enriched environment ( Nylabone® chews ) on a 13/11-hour light-dark cycle ( light 0600-2000 ) . Animals were maintained on a restricted feeding schedule ( 15–20 g of standard chow available immediately after training/testing ) . All procedures were in accordance with NIH guidelines and were approved by the University of Wisconsin Institutional Animal Care and Use Committee . Under halothane anesthesia ( Halocarbon Laboratories , River Edge , New Jersey; 1%–4% in air ) , animals were implanted bilaterally with linear electrode arrays ( n = 8 electrodes/array; 250 µm separation; SB103 , NB Labs , Dennison , TX ) targeting layer V of the prelimbic region of the PFC ( plPFC ) as previously described [52] . Electrode arrays contained 50 µm stainless-steel electrodes orientated in a rostral-caudal direction . Electrodes were attached to skull screws ( MX-0080-16B-C , Small Parts , Inc . ) with dental acrylic ( Plastics One , Roanoke , Virginia ) , the wound was closed with wound clips ( 9 mm Autoclip; BD Diagnostic Systems , Sparks , Maryland ) , and animals were allowed to recover for 7–10 days . On testing days , animals were brought into the T-maze testing room and tethered to the Multichannel electrophysiology Acquisition Processor ( MAP , Plexon , Dallas , Texas ) . During the 2 hour habituation period , putative single “units” of the plPFC were discriminated in real time using online template matching algorithms to preliminarily discriminate action potentials exhibiting a 3∶1 signal to noise ratio . Following discrimination of plPFC units , animals remained tethered to recording hardware and the quality of the discrimination was monitored throughout the remainder of the day . During baseline and noise stress conditions , neural activity was simultaneously amplified , discriminated , time stamped , and recorded from these putative single units of the plPFC as previously described [46] , [52] . Additionally , video recordings were made of animal behavior during testing sessions ( resolution = 0 . 0125 sec ) with time-stamp overlays synchronized to the electrophysiological hardware . During the 2-hour inter-session interval , animals remained tethered and neuronal activity was monitored for drift in the quality of discrimination of action potentials . At the end of the study , animals were deeply anesthetized and cathodal current ( 60 µA ) was passed across microwire pairs within a bundle for 45 sec . Animals were perfused with a 10% formalin + 5% potassium ferrocyanide solution that produced a Prussian blue reaction product at the electrode tip . Brains were removed and immersed in 10% formalin for 24 hr . Frozen 40-µm coronal sections were collected through the plPFC and counterstained with Neutral Red . Representative placements of recording electrode bundles within the plPFC are illustrated in Fig . 1C . | When faced with stressful situations , normal thought processes can be impaired including the ability to focus attention or make decisions requiring deep thought . These effects can result in accidents at the workplace and in combat , jeopardizing the lives of others . To date , the effect of stress on the way neurons communicate and represent cognitive functions is poorly understood . Differing theories have provided opposing predictions regarding the effects of stress-related chemical changes in the brain on neuronal activity of the prefrontal cortex ( PFC ) . In this study , we show that stress increases the discharge rate of PFC neurons during planning and assessment phases of a task requiring the PFC . Additionally , using a point process model of neuronal activity we show that stress , nonetheless , impairs the ability of PFC neurons to retain representations of past events over time . Together these findings suggest that stress-related impairment of cognitive function may involve deficits in the ability of PFC neurons to retain information about past events beyond changes in neuronal firing rates . We believe that this advancement provides new insight into the neural codes of higher cognitive function that may lead to the development of novel treatments for stress-related diseases and conditions involving PFC-dependent cognitive impairment . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"working",
"memory",
"cognitive",
"neuroscience",
"mathematics",
"neurophysiology",
"computational",
"neuroscience",
"single",
"neuron",
"function",
"statistics",
"biology",
"neuropsychology",
"neuroscience",
"learning",
"and",
"memory",
"probability",
"theory"
] | 2012 | Stress-Induced Impairment of a Working Memory Task: Role of Spiking Rate and Spiking History Predicted Discharge |
A blend of compounds ( pentanoic acid , guaiacol , δ-octalactone and geranylacetone ) identified in waterbuck ( Kobus defassa ) body odour referred to as waterbuck repellent compounds ( WRC ) and a synthetic repellent 4-methylguaiacol have previously been shown to repel tsetse flies from the morsitans group . However , these repellents have not been evaluated on palpalis group tsetse flies . In this study , we evaluated the effect of these repellents on catches of Glossina fuscipes fuscipes ( major vector of human sleeping sickness ) in biconical traps and on sticky small targets which are visually attractive to palpalis group flies . The attractive devices were baited with different doses and blends of the repellent compounds . We also assessed the effect of removal of individual constituents in the synthetic blend of WRC on catches of G . f . fuscipes . The study was conducted in western Kenya on four islands of Lake Victoria namely Big Chamaunga , Small Chamaunga , Manga and Rusinga . The tsetse fly catches from the treatments were modeled using a negative binomial regression to determine their effect on catches . In the presence of WRC and 4-methylguaiacol ( released at ≈2 mg/h and ≈1 . 4 mg/h respectively ) , catches of G . f . fuscipes were significantly reduced by 33% ( P<0 . 001 ) and 22% ( P<0 . 001 ) respectively in biconical traps relative to control . On sticky small targets the reduction in fly catches were approximately 30% ( P<0 . 001 ) for both 4-methylguiacol and WRC . In subtractive assays , only removal of geranylacetone from WRC significantly increased catches ( by 1 . 8 times; P <0 . 001 ) compared to the complete blend of WRC . We conclude that WRC and 4-methylguaiacol reduce catches of G . f . fuscipes at stationary visually attractive traps and suggest that they may serve as broad spectrum repellents for Glossina species . We recommend further studies to investigate the effects of these compounds on reduction of G . f . fuscipes attracted to human hosts as this may lead to development of new strategies of reducing the prevalence and incidence of sleeping sickness .
Tsetse flies ( Diptera: Glossinidae ) feed on blood and are biological vectors of African trypanosoma parasites that cause human and animal African trypanosomiasis [1] . They find their vertebrate hosts through olfactory and visual cues [2] . Beyond its visual range , the fly is activated by the odour from the host and orients upwind following the odour plume until it comes near the host where visual cues of colour , shape and size may elicit a landing response [2–5] . It is while taking a blood meal from a host that an infected tsetse fly transmits the parasites that cause African trypanosomiasis [6] . However , not all vertebrates found in tsetse fly habitats are fed on [7 , 8] . The differential preference of vertebrate hosts has been attributed to a combination of specific compounds found in the vertebrate’s body odour which could either attract or repel tsetse flies [2 , 4 , 8–10] . Consequently , research on identification of repellents to break the host-tsetse fly contact as a method of control against African trypanosomiasis ( AT ) has been ongoing since the 1970s [11] . Pioneer work on repellents showed that , humans are poorly attractive to Glossina pallidipes and G . morsitans morsitans which are tsetse species that belong to the morsitans group [2] . Variations in lactic acid concentrations in human odour were identified to be responsible for this repellency [12] . Since then , a number of synthetic and naturally occurring repellent compounds have been identified [9–11 , 13–15] . Among these are; acetophenone and 4-methylguaiacol which have been shown to reduce catches of G . pallidipes by 69% and 80% respectively [11 , 13] , and δ-nonalactone which reduce G . pallidipes catches by 76% when used in attractant odour baited traps [15] . Furthermore some naturally occurring tsetse repellents found in the body odour of waterbuck ( Bovidae: Kobus defassa ) , a non-preferred host were identified [7 , 8] including 15 compounds comprising of straight chain carboxylic acids ( C5-C10 ) , phenols ( guaiacol and carvacrol ) , 2-alkanone homologues ( C8-C12 ) , geranylacetone and δ-octalactone [8 , 10] . A blend of all these compounds was found to significantly reduce catches of G . pallidipes in traps baited with odourant by 84% [10] . The blend of these compounds was reduced to a four component-blend [16] comprising pentanoic acid , guaiacol , δ-octalactone and geranylacetone referred to as waterbuck repellent compounds ( WRC ) which was found to reduce levels of animal African trypanosomiasis transmitted by G . pallidipes by 80% [16] . Most reports on tsetse fly repellents have been associated with important species belonging to the morsitans group . However , important tsetse fly species belonging to the palpalis group also responsible for human and animal African trypanosomiasis transmission in central and western Africa have received less attention . The important tsetse species belonging to the palpalis group include G . fuscipes subspecies , G . palpalis subspecies and G . tachinoides . Palpalis group tsetse species account for over 95% of transmissions of all human African trypanosomiasis ( HAT ) cases [17 , 18] . Among these , G . fuscipes subspecies with G . f . fuscipes having the widest distribution account for about 90% of transmissions in the HAT foci of central and west Africa [17–19] . Though they are opportunistic blood feeders , tsetse from the palpalis group have shown preferences to vertebrate hosts as observed from blood meal analysis [7 , 20] . For example , monitor lizards were consistently found to be the main hosts in Central African Republic , Kenya and Uganda accounting for about 40% of the blood meals [7] . Generally , tsetse flies from the palpalis group are reported to exhibit weak responses to host odours compared to flies from the morsitans group [5] . However , there is evidence from studies that have shown a general conservation of chemosensory gene families across five tsetse species which include G . austeni , G . brevipalpis , G . pallidipes , G . m . morsitans and G . f . fuscipes [21–24] . In this study , we hypothesised that repellents previously shown to be effective against morsitans group tsetse flies could also be repellent to palpalis group tsetse flies . Therefore , we evaluated the responses of G . f . fuscipes to visually attractive stationary traps baited with WRC and 4-methylguaiacol . We also assessed the effect of individual constituents of WRC on trap catches through subtractive assays .
The experiments were carried out on four islands of Lake Victoria in western Kenya from April 2016 to December 2017 . The Islands included: Small Chamaunga ( latitude -0 . 431° , longitude 34 . 227°; surface area of about 0 . 2km2 ) , Big Chamaunga ( latitude -0 . 426° , longitude 34 . 227°; surface area of about 0 . 2km2 ) , Manga ( latitude -0 . 353° , longitude 34 . 253°; surface area of about 1km2 ) and Rusinga ( latitude -0 . 358° , longitude 34 . 218°; surface area of about 43km2 ) [18 , 25 , 26] . Big and Small Chamaunga Islands are not inhabited by humans while Manga and Rusinga Islands are . The vegetation on the islands mainly consists of Aeschynomene eraphyroxylon ( fresh water mangroves ) , Lantana camara ( Tickberry ) and Dombeya spp . ( tropical hydrangea ) [18] . These islands exclusively harbor G . f . fuscipes which mainly feeds on monitor lizards ( Varanus niloticus ) [18 , 26 , 27] . However , no case of sleeping sickness has been reported for the last 30 years in the study area [18] . Tsetse fly catches were made using biconical traps [28] and sticky small targets [29] . These were placed at sites that had either open or dense vegetation previously shown to have apparent fly densities of more than twenty flies per biconical trap per day [25] . The compounds: pentanoic acid , guaiacol , δ-octalactone and geranylacetone were blended in similar proportions ( 3:2:3:1 ) as found in waterbuck odour [8 , 10 , 16] . All individual compounds and blends of repellent compounds from waterbuck and 4-methylguaiacol were dispensed passively in natural environmental conditions from sealed polythene sachets ( Audion Elektro , Derby , UK ) with 0 . 125 mm thick walls , 50 mm × 75 mm in width and height placed next to the biconical trap and underneath the sticky small target [30] . All compounds evaluated were of 98–99% purity and sourced from ChemSamp Co , LLC , Trenton , USA . A series of subtractive assays to achieve blends without one constituent of WRC ( Table 1 ) were prepared in sachets . Three sachets , each containing WRC , blends that resulted from subtractive assays of WRC and individual constituents ( Table 1 ) were subjected to field conditions . The weight of each dispenser was then taken every 24 hr for two days to come up with six replicates ( S2 Table ) of each treatment in order to determine the release rates . The average release rate of 4-methylguaiacol was obtained from nine sachets by measuring the difference in masses between freshly prepared sachets and their masses after 24 hr for three days ( 27 replicates in total; S2 Table ) under field conditions . In order to confirm the release of all compounds from the blends , a sachet containing either WRC or blends that resulted from subtractive assays were placed in a 700 ml glass bottle covered with aluminium foil tightly held to the bottle with two tight fitting rubber bands at room temperature . A pre-cleaned ( through thermal desorption at 250°C for 30 min to remove any contaminants ) 65 μm polydimethylsiloxane ( PDMS ) solid phase micro extraction ( SPME ) fibre ( Supelco , Bellefonte , USA ) was inserted through the aluminium foil into the bottle and the PDMS fibre exposed to the headspace for 5 min to adsorb the volatiles . Thereafter , the volatiles collected on the SPME fibre were subjected to GC/MS analysis . The fibre was manually inserted into the injection port of a 7890B Agilent gas chromatograph ( Agilent Technologies , Wilmington , DE ) coupled to an Agilent mass spectrometer ( MSD 59977A , Agilent Technologies , Wilmington , DE ) operated in split-less mode with the injector at 250°C to desorb the trapped volatiles for 2 min . The separation of compounds were done on an Agilent HP-5 MS capillary column ( 30 m × 0 . 25 mm id × 0 . 1 μm film thickness; Agilent Technologies , Santa Clara , US ) using the following temperature programme: 35°C for 5 min , then raised at 10°C/min to a final temperature of 280°C and held for 10 . 5 min . Helium was used as the carrier gas at a constant flow rate of 1 ml/min . The compounds were detected using the electron ionisation mode ( 70eV; Ion source 230°C; quadrupole 150°C; mass scan range , 30–350 amu ) . Experiments where biconical traps were used ran from 8:00 to 18:00 hr [31] while those that used the sticky small target as the trapping device ran from 08:00 to 12:00 hr during the period when G . f . fuscipes is most active [27 , 29] . The treatments were incorporated into a series of randomised block design experiments comprising groups of near or adjacent days at a site as different blocks [32] . Treatments were randomly allocated to days within these blocks . The number of blocks for each experiment which serve as replicates are shown in S1 Table . Tsetse fly catches were sexed and recorded for each treatment and experiment with freshly prepared repellents used for each experiment . Different release rates for the WRC and 4-methylguaiacol as treatments were achieved by varying the number of dispensers between one , two and four sachets per trap with the unbaited traps serving as controls . The number of sachets containing the compounds that effectively reduced both male and female catches of G . f . fuscipes in odourant baited biconical traps was used to bait sticky small targets to assess if the effect was similar to that observed in biconical traps . The blends that resulted from subtractive assays ( two sachets ) , WRC ( two sachets ) and individual constituent ( one sachet ) of WRC served as treatments were compared to the control being a biconical trap alone without odour bait . All statistical tests were done with R version 3 . 2 . 5 [33] . Analysis of variance ( ANOVA ) and Student Newman Keuls ( SNK ) test were used for multiple comparisons of average release rates from single sachets of the individual constituents of WRC; resultant blends from subtractive assays and WRC . A negative binomial model was used to measure the effect of various treatments on the fly catch while taking into account the block and experimental day . Only the detransformed means ( effects display ) of treatments are reported and were obtained from the negative binomial regression using the “effects” package in R [34] . Statistical significance was considered at α less than 0 . 05 . Permission was given to undertake entomological gathering on Big Chamaunga and Small Chamaunga islands by the owners ( International Centre of Insect Physiology and Ecology ) . Other entomological collection was done on public land . This study was conducted in conformity with the International Centre of Insect Physiology and Ecology ethical rules for animals .
The average release rate from each polyethene sachet of WRC was 0 . 83 mg/h ( 95% CI: 0 . 65–1 . 01 ) while that of 4-methylguaiacol from each polythene sachet was about 1 . 4 mg/h ( 95%CI: 1 . 30–1 . 50 ) . From single sachets of individual constituents , δ-octalactone had the lowest release rate ( 0 . 26 mg/h; 95% CI: 0 . 08–0 . 44 ) , while pentanoic acid had the highest ( 3 . 83 mg/h; 95% CI: 2 . 29–4 . 01 ) ( Table 2 ) . For single sachets of the blends , WRC without pentanoic acid had the lowest release rate , whereas WRC without δ-octalactone had the highest release rate ( Table 2 ) . There were significant differences in the release rates from single sachets of the different blends and those of individual constituents of WRC ( ANOVA , df44 , F = 175 . 81 , P<0 . 001 ) . The overall release rate in mg/h of WRC without pentanoic acid or guaiacol was not significantly different from that of WRC ( SNK: P>0 . 05; Table 2 ) . However , when δ-octalactone or geranylacetone were removed from WRC , release rates differed significantly ( SNK: P<0 . 05; Table 2 ) . Samples of sachets containing WRC and WRC without a specific constituent subjected to GC/MS confirmed that all the individual constituents were dispensed from the polyethene sachet dispensers as volatiles ( Fig 1 ) . The total numbers of tsetse flies trapped in experiments with WRC were 3 , 983 of which 1 , 664 ( 41 . 8% ) were males and 2 , 319 ( 58 . 2% ) were females ( S1 Table ) . The overall detransformed means of flies trapped in the control ( biconical trap only ) were higher than those collected from biconical traps with varying number of WRC dispensers as treatments ( Fig 2A ) . The tsetse fly catches for both male and female G . f . fuscipes were significantly reduced when WRC was dispensed from two sachets at a biconical trap by 23% ( 95% CI: 6–37%; P<0 . 05 ) and 37% ( 95% CI: 25–47%; P<0 . 001 ) respectively and overall by 33% ( 95% CI: 20–44%; P<0 . 001 ) . However , when WRC was dispensed from one and four sachets at biconical traps , only the female catches were significantly reduced ( Fig 2B ) . In experiments with 4-methylguaiacol a total of 2 , 589 tsetse flies were caught in traps comprising of 1 , 302 ( 50 . 3% ) males and 1 , 287 ( 49 . 7% ) females ( S1 Table ) . The detransformed means of flies caught in the control ( biconical trap only ) were again higher than for traps with treatments ( Fig 3A ) . Dispensing 4-methylguaiacol from one and two sachets significantly reduced catches of male G . f . fuscipes by 18% ( 95% CI: 3–30%; P<0 . 05 ) and 16% ( 95% CI: 2–28%; P<0 . 05 ) respectively while those of females were reduced by 25% ( 95% CI: 12–36%; P<0 . 001 ) and 19% ( 95% CI: 5–30%; P<0 . 01 ) respectively . Overall , when 4-methylguaiacol was dispensed from one and two sachets , the reduction in catches were 22% ( 95% CI: 13–31%; P<0 . 001 ) and 18% ( 95% CI: 8–26%; P<0 . 001 ) respectively . However , dispensing 4-methylguaiacol from four sachets only reduced female catches significantly ( 26%; 95%CI: 14–33%; P<0 . 001 ) ( Fig 3B ) . Two sachets of WRC and , a sachet of 4-methylguaiacol shown to be effective in reducing catches of both male and female G . f . fuscipes in previous experiments at biconical traps were dispensed from sticky small targets . This was done in order to test the effectiveness of the test compounds at different trapping device . The total number of tsetse flies caught in the experiment was 1 , 695 comprising of 610 ( 36 . 0% ) males and 1 , 085 ( 64 . 0% ) females ( S1 Table ) . The results showed lower detransformed means ( Fig 4A ) and significant reductions in catch indices of both sexes of G . f . fuscipes compared to the control ( Fig 4B ) . During these experiments , a total of 5 , 489 G . f . fuscipes were caught comprising of 2 , 923 ( 53 . 3% ) males and 2 , 566 ( 46 . 7% ) females ( S1 Table ) . The removal of pentanoic , guaiacol or δ-octalactone from WRC neither lowered the overall detransformed mean nor reduced the catches of G . f . fuscipes compared to biconical traps baited with WRC ( P>0 . 05; Figs 5A–5D and 6A–6D ) . Dispensing WRC without geranylacetone from two sachets significantly increased the catch of male and female G . f . fuscipes by 1 . 76 times ( 95%CI: 1 . 36–2 . 29 times; P<0 . 001 ) and 1 . 71 times ( 95% CI: 1 . 31–2 . 25 times; P<0 . 001 ) respectively compared to WRC .
Assessing the responses of tsetse flies from the palpalis group to synthetic compounds and natural odours that repel flies from the morsitans group is important as it may lead to development of novel African trypanosomiasis control methods . Such control methods could be effective by reducing host-vector contact particularly in areas with low density and infection rates in tsetse populations , such as those of the HAT foci in central and west Africa [5] . In this study , we report responses of G . f . fuscipes , a tsetse species from the palpalis group , to visually attractive stationary traps baited with WRC and 4-methylguaiacol . We observed that both male and female catches of G . f . fuscipes in biconical traps were reduced at specific dispensing rates of WRC ( 2 dispensers approximately 2 . 0 mg/h ) and 4-methylguaiacol ( 1 or 2 dispensers approximately 1 . 4 and 2 . 8 mg/h respectively ) , indicating that they are true repellents as defined by Dethier et al . [35] . The repellency of WRC and 4-methylguaiacol to G . f . fuscipes was confirmed by the reduction of catches on sticky small targets baited with these compounds . These results seem to indicate that the release rates of odorants influence the responses of G . f . fuscipes . This conforms to findings from other studies which showed that it was only at release rates of 100 mg/h of lactic acid that G . pallidipes were repelled compared to a release rate of 10 mg/h [13] . In addition , other blood feeding Diptera such as mosquitoes , also demonstrate responses that are dependent on odorant concentrations [36] . We also observed a differential sex response when WRC was dispensed from a single or four sachets ( approximately 1 . 0 or 4 . 0 mg/h respectively ) with only female catches being reduced . A similar observation with 4-methylguaiacol was made with only female catches reducing when it was dispensed from four sachets ( approximately 5 . 6 mg/h ) . Related differential sex responses have been reported with G . pallidipes , a tsetse fly from the morsitans group , to constituents of human odour where the repellent effect was greater for females than males [2] . This could be an indication that female G . f . fuscipes are more sensitive to repellent odours than males . Consistent to the observed differential sex response to odorants in our study , is the finding by Otter et . al . [37] where female G . f . fuscipes had higher electroantennogram responses than males . Nevertheless , there is need to investigate further the role of this differential sex response in the biology of the fly as it could be exploited for development of control methods that target female G . f . fuscipes . Furthermore , the observed repellency of G . f . fuscipes to WRC and 4-methylguaiacol could be exploited for “push-pull” disease control strategies where the repellents could be used to “push” the flies towards stationary visual attractive ( “pull” ) devices that eventually kill the flies . The consistent response to repellent odours in Glossina species observed in this and other studies provides support to a study that showed that there is a general conservation of chemosensory gene families across five tsetse species that includes G . f . fuscipes and G . pallidipes [10 , 11 , 21] . The reduction in G . f . fuscipes catches of ~33% by WRC we observed is less than ~84% reported for G . pallidipes [10] . This variation could be due to the differences in formulation of WRC , where in our case , hexanoic acid was not added as a constituent . However , the repellency of hexanoic acid was reported not to be significantly different from that of pentanoic acid for G . pallidipes [10] . Additionally , the previous study used traps baited with odour attractants as the controls ( reference ) [10] . In this study , the use of traps without odour attractants as controls could also explain the differences in reduction in fly catches . However , the reduction in G . f . fuscipes catches of ~22% by 4-methylguaiacol is also less than ~70% reported for G . pallidipes at traps without odour attractants [11] . This could be an indication that G . f . fuscipes is less sensitive to 4-methylguaiacol than G . pallidipes . Even though this is consistent with other reported observations that tsetse flies in the palpalis group show , markedly weaker responses to host odours compared to those from the morsitans group [5] , further studies to ascertain why this is the case are needed . Despite being dispensed from sachets of relatively consistent measurements , WRC without δ-octalactone or geranylacetone had significantly higher release rates compared to WRC . Clearly , indicating that in blends , the relative diffusion of different constituents across the walls of the polyethene sachet dispensers and subsequent evaporation from the surface is not only influenced by their respective vapour pressures , but also by the presence of other components in the blend [10] . This is further supported by the observed significant variation in release rates of the individual constituents of WRC . Our results also indicate that when geranylacetone is removed from WRC; the catch of the resultant blend ( pentanoic acid , guaiacol and δ-octalactone ) increased by 1 . 8-fold , showing less potency in repellency . This suggests that geranylacetone may be playing an important role to the overall repellent effect of WRC to G . f . fuscipes . In other dipteran vectors , repellence by geranylacetone has been implicated in the differential attraction of humans to various species of mosquitoes and Culicoides midges [38 , 39] . Geranylacetone has also been reported to enhance the repellency of a mixture of ammonia and lactic acid against mosquitoes [38] . Conversely , pentanoic acid enhances attraction at flow rates of 100ml/min of a mixture of ammonia and lactic acid [40] . Interestingly , fly catches in traps baited with the blends that result after removal of pentanoic acid ( geranylacetone , guaiacol and δ-octalactone ) , guaiacol ( geranylacetone , pentanoic acid and δ-octalactone ) or δ-octalactone ( geranylacetone , guaiacol and pentanoic acid ) from WRC ( pentanoic acid , guaiacol , δ-octalactone and geranylacetone ) did not significantly differ with those from traps with WRC . Additionally , the catches at traps with the individual constituents did not also significantly differ from those of WRC . These results suggest that the individual constituents could substitute WRC as repellents at biconical traps . Pentanoic acid and guaiacol have been shown to reduce tsetse fly catches of G . pallidipes at Epsilon traps but not significant effect on the feeding efficiency of the fly [13] . This indicated that what works at traps may not work to protect hosts against tsetse fly bites . With G . fuscipes being a major vector of HAT [18] and evidence that it can be repelled from traps baited with synthetic and allomonal compounds from waterbuck odour , we recommend further studies that will evaluate these compounds to protect human hosts and dwellings . Most studies that compare trap catches of tsetse flies exposed to several treatments of odorants use Latin square design experiments [10 , 11 , 13 , 15] . This is because it can control for variation of fly catches due to the trap site and day at study design stage [41] . One of the important assumptions of the Latin square design is that all trapping sites should have the same sort of vegetation [41] . However , in this study trapping sites were in both open and dense vegetation . Therefore , we used a randomised block design where the site was blocked and treatments were randomly assigned to days as experimental units in each block . This way the randomisation controlled for unknown confounds while known confounds were addressed during statistical analysis by accounting for the block and day of the experiment . Additionally , our study did not consider the effect of age of the polythene sachet dispensers as it could affect the release rates of the repellents [42] . However , for each experiment , freshly prepared repellents dispensers where used . This could have minimised the effect of age of the sachets on the release rates of repellents . Studies with other Dipteran vectors , such as mosquitoes have used fabric-based dispensers for odorants [43] . Further studies should explore how these dispensers could also perform for tsetse flies . Furthermore , on account that the release rates reported were not concurrently obtained from experiments that compared catches of tsetse flies from traps baited with odorants , it is possible that these could be biased . However , the setting and odorant dispensers used to obtain the release rates were similar to those used in the field experiments of comparing the effect of odorants on trap catches of tsetse flies . Thus , we are confident that these could reflect the actual field release rates of the odorants . In conclusion , the present study has shown that WRC and 4-methylguaiacol released at specific rates reduced catches of both sexes of G . f . fuscipes in unbaited traps , an indication that they are true repellents . It also showed that sex of G . f . fuscipes could play a role in its responses to repellents . Additionally , geranylacetone seems to play an important role in the overall repellency of WRC . Furthermore , individual WRC constituents: pentanoic acid , guaiacol , geranylacetone and δ-octalactone repel G . f . fuscipes just as well as WRC at biconical traps . Therefore , we recommend further studies to evaluate the repellency of 4-methylguaiacol , WRC and its individual constituents against G . f . fuscipes in the presence of hosts as it may lead to the development of novel control methods especially in HAT foci . | Tsetse flies are divided into three taxonomic groups: morsitans , palpalis and fusca . Flies from the morsitans and palpalis groups are the main vectors of trypanosoma parasites that cause human and animal African trypanosomiasis . The chemical 4-methylguaiacol and waterbuck ( Kobus defassa ) , a known non-preferred host of tsetse , body odour blend ( pentanoic acid , guaiacol , δ-octalactone and geranylacetone ) , have been shown to repel morsitans group species G . pallidipes and significantly reduce levels of animal African trypanosomiasis . However , these repellents have not been evaluated against other groups of tsetse , for example , those in the palpalis group . Here , we show that visually attractive stationary devices ( biconical traps and sticky small targets ) when baited with these repellent compounds and some of their blends significantly repel G . f . fuscipes , one of the important vectors of human sleeping sickness belonging to the palpalis group . The results provide the foundation for future studies of these compounds on their repellency of other Glossina species and their use in ‘push-pull’ strategies for manipulation of attraction to human and animal hosts and also in disease reduction strategies especially for riverine tsetse flies which are major vectors of parasites that cause sleeping sickness . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"geomorphology",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"landforms",
"african",
"trypanosomiasis",
"topography",
"tropical",
"diseases",
"parasitic",
"diseases",
"animals",
"parasitic",
"protozoans",
"glossina",
"protozoans",
"odorants",... | 2019 | Responses of Glossina fuscipes fuscipes to visually attractive stationary devices baited with 4-methylguaiacol and certain repellent compounds in waterbuck odour |
The three-dimensional positions of immune cells can be tracked in live tissues precisely as a function of time using two-photon microscopy . However , standard methods of analysis used in the field and experimental artifacts can bias interpretations and obscure important aspects of cell migration such as directional migration and non-Brownian walk statistics . Therefore , methods were developed for minimizing drift artifacts , identifying directional and anisotropic ( asymmetric ) migration , and classifying cell migration statistics . These methods were applied to describe the migration statistics of CD8+ T cells in uninflamed lymph nodes . Contrary to current models , CD8+ T cell statistics are not well described by a straightforward persistent random walk model . Instead , a model in which one population of cells moves via Brownian-like motion and another population follows variable persistent random walks with noise reproduces multiple statistical measures of CD8+ T cell migration in the lymph node in the absence of inflammation .
A primary challenge of immunological imaging experimentation is to understand the nature of cell migration statistics , and the role that these statistics play in immune function . Over the last decade , two-photon microscopy has transformed the understanding of the role of cell migration in the immune response [1–3] . However , although improved statistical approaches are still being developed [4–6] , many existing methods for analyzing migration statistics are susceptible to experimental artifacts that can lead to inaccurate conclusions about leukocyte behavior . Similar questions arise in analyzing the migration of organisms ranging from bacteria [7] to vultures [8] and human hunter-gatherers [9] . Migration tracks can be directional or random and can be characterized by a bewildering array of models . This poses the question of how best to analyze migration tracks in an unbiased fashion , given experimental data that is often gathered in a limited field of view over a short period of time . Many immune functions are thought to be directed by chemotactic signals , and directional migration has been observed in numerous cases , such as neutrophil response to sterile inflammation [10] , migration of positively selected T cells in the thymus [11] , and T cell priming by dendritic cells in lymph nodes [12 , 13] . While the directional bias in these studies is clear , they use measures of directionality that can be susceptible to experimental artifacts . These issues range from technical constraints such as the finite imaging field and global drift to the intrinsic limitations of widely-used quantities such as the meandering index and motility coefficient [2 , 4 , 5 , 14 , 15] . Such artifacts can affect quantitative analyses and can even lead to inaccurate qualitative conclusions in cases where directional motility is subtle . Immune cell migration also has a stochastic component [1 , 3 , 6 , 16–18] . Commonly used random walk analyses [14 , 15] assume that cells obey Brownian statistics . In several cases , however , it has been argued that cells migrate via persistent random walks [5 , 16 , 19–22] or even exhibit Lévy behavior [6 , 23] . Despite this knowledge , many analyses of random motion implicitly assume that the statistics are described by Brownian walks even at short time scales , by assuming that the mean-squared displacement increases linearly in time and extracting a motility coefficient [14 , 15] . The accurate identification of the persistence time for persistent random walks [5 , 16 , 19–22] , or of more exotic forms of migration statistics such as Lévy behavior of migrating microglia [23] or of CD8+ T cells in Toxoplasma gondii-infected mouse brains [6] , requires a description that goes beyond use of the mean-squared displacement as a distinctive identifier of migration statistics . A more complete and accurate description of migration statistics requires methods capable of detecting subtle directional bias that can also handle more general forms of random walks , given experimental data gathered over rather short time periods . Furthermore , in order to investigate the correlation between cell migratory behavior and immune function , it is necessary to develop a description that rigorously characterizes both stochastic and directional migration without any initial assumptions regarding the location of possible targets . Current methodologies are of limited use in achieving these goals . Here , we describe a set of analytical and computational methods that can be used to identify various types of directional , anisotropic ( asymmetric ) , and stochastic migration . These methods can be applied to any type of motile organism or cell . In order to demonstrate the practical implementation of the methods , we apply them to the migration of CD8+ T cells in uninflamed mouse lymph nodes . Unexpectedly , this system is well-described by a model containing two populations of T cells , in which one population obeys Brownian statistics and the other population migrates via heterogeneous , persistent random walks . While this model shares similarities with previous persistent random walk and run/pause models [5 , 16 , 19–22] , we show that it reproduces several key statistical measures beyond the mean-squared displacement alone . Our results show that CD8+ T cells in uninflamed mouse lymph nodes migrate differently from those in the brains of mice chronically infected with Toxoplasma gondii .
A pivotal question for many studies of immune cells is whether they migrate and respond to signals with specific directional motion or biases [2 , 3 , 14 , 15] . More generally , anisotropic motion — motion that is not statistically identical in all directions — can indicate a directional bias , such as chemotaxis or chemokinesis due to a chemical gradient , or to an asymmetric feature of a particular direction , such as confinement . However , since even isotropic trajectories appear directed on short enough time scales , and , conversely , directed tracks typically contain a stochastic component , discerning directionality and anisotropy is not a simple task [15 , 16] . The most commonly used method for identifying directional motion currently consists of plotting cell tracks with their starting points translated to the origin , measuring the MSD , calculating the meandering index , and measuring the mean displacement vector [5 , 10–13 , 15 , 18] . However , as discussed in the following sections , these methods are only sensitive to obvious directionality , suffer from several experimental artifacts , and depend quantitatively on details of the experimental set up . Of these methods , even the quantitative tests are only capable of identifying global drift . Thus , they cannot detect other types of anisotropic motion , such as directed motion towards a single target ( or scattered collection of targets ) or Brownian-walk-like motion with different motility coefficients for different spatial directions . Other methods for identifying motion in a specific direction , such as measuring the component of velocity in that direction or angle of motion with respect to a target , require prior knowledge of the existence of a target or special direction [15] . Because of these issues , we developed and tested several techniques to detect and determine the amount and type of anisotropy . These methods are sensitive to small anisotropies and do not rely on having prior knowledge about the directional motion . The mean-squared displacement , while straightforward to calculate , is highly susceptible to artifacts that can lead to misinterpretations of data . For example , typical imaging experiments can only visualize cells within a relatively small part of the space that cells can actually explore . Thus , as noted previously , cells may exit the field of view before the time series has ended , which can bias the analysis [5 , 6 , 15 , 21 , 30] . The magnitude of this effect can be estimated by calculating the mean time , 〈texit〉 , for a Brownian-random-walking cell to reach the boundary of the imaging field , which has a shortest dimension ( typically , depth ) of length L [27]: 〈 t exit 〉 = L 2 12 D , ( 1 ) where D is the diffusion ( motility ) coefficient . For typical values derived from multiple studies of T cell movement , L ≈ 40 μm and D ≈ 30 μm2/min ( e . g . , refs . [1 , 5 , 6 , 15 , 28–30 , 34] ) , 〈texit〉 is just 4 . 4 minutes , with some cells exiting the field of view even more quickly . This limitation of imaging has especially significant consequences for the mean-squared displacement ( MSD ) . Since fast-moving cells tend to leave the imaging field more quickly than others , data at late times becomes biased toward slow-moving cells [5 , 15 , 21 , 30] . This can distort the shape and magnitude of the MSD as a function of time . Furthermore , this issue plagues alternatives to the standard practice [5 , 14 , 15] of measuring the motility coefficient from the slope of the best-fit line to the MSD versus time curve . Since the standard motility coefficient method is inaccurate due to short-time directional persistence [5] , one may try either fitting only the MSD at late times to a line or fitting the MSD with a function of both the motility coefficient , D , and persistence time , tp [5 , 21] . However , the first option exacerbates the finite imaging field problem , underestimating the diffusion coefficient by as much as 20% under simulated typical conditions . The second option is also inadequate because the fit parameters ( motility coefficient and persistence time ) are sensitive to the duration of the “early time” segment used . Alternatively , if long time segments are used , the fits converge on common parameters [5] , but we have found in simulations that they underestimate the motility coefficient and persistence time by as much as 20% and 40% , respectively , due to cells exiting the imaging volume . When considered together , these limitations have several practical implications . Since the mean time 〈texit〉 , for a cell to leave the imaging volume is typically a few minutes , and many cells exit earlier than 〈texit〉 , the MSD and quantities derived from the MSD are unreliable for rigorously assessing migration data . Counterintuitively , in a finite imaging field , measuring the MSD over longer time intervals can lead to erroneous conclusions rather than deeper insights . It should also be noted that the MSD does not uniquely specify an underlying model for random ( or directional ) motion [15 , 35 , 36] . Therefore , this measurement should only be used for qualitative comparisons between experiments with identical imaging dimensions or as a complementary consistency check for any proposed migration model . Instead of focusing on the MSD , we rely on two main quantities to characterize the displacements: ( 1 ) the probability distribution , PΔt ( r ) , of cell displacements , r , as a function of the time interval Δt , over which the displacement occurs and ( 2 ) the correlations , C ( t1 , t2;τ ) , between the displacements during one time interval τ starting at time t1 , with displacements during a later time interval τ , starting at t2 . Utilizing these quantities mitigates problems due to the finite field of view . Probability distributions , PΔt ( r ) , reveal displacements of various sizes and at all times , instead of focusing on large displacements at late times , which are the most profoundly impacted by the limited field of view . The correlation function typically reveals features such as persistence in the early-time data , while minimizing significant artifacts .
The last decade has seen tremendous advances in the ability to image the behavior of lymphocyte populations [3] , but there are several important limitations to imaging experiments and data analyses of cell tracks that have hindered efforts to interpret cell migration quantitatively . For example , the finite imaging volume and z-depth of experiments can skew the interpretation of migration data , and the effects of the finite image volume negate any advantage gained by imaging cell populations for longer times . Due to the effects of the shallow depth of the imaging volume , there are severe truncation effects due to cells prematurely leaving the field of view . It is important to recognize that these effects cannot be remedied by simply imaging over a longer time interval; instead , improvements require experiments with greater imaging volume and thus increased z-depth . Less obvious effects such as global drift can further obscure the true nature of cell migration; this may be mitigated by adjusting tracks according to the motion of auto-fluorescent particles . These experimental issues necessitate a comprehensive analysis that goes beyond standard measures such as the mean-squared displacement and meandering index . In addition , several tests for anisotropy and directionality are required because different measurements capture different types of anisotropic behavior . To directly analyze migration statistics , one should construct probability distributions of displacements over various time intervals and measure correlations in cell trajectories . Even without experimental artifacts such as the limited field of view , common measures such as the mean-squared displacement are not sufficient to distinguish details of cell migratory behavior without further information . For instance , a major drawback of the MSD analysis is that different models may produce identical MSD curves , yet differ in a variety of key aspects including mean velocity or persistence time . We have presented quantitative methods to minimize and account for these limitations , and quantitatively describe cell migration . Together , these techniques minimize errors due to drift ( Fig . 1 ) , reliably detect anisotropic cell migration ( Fig . 2 ) , and provide a strong connection between migration models and experimental data ( Fig . 3 ) . With these methods , we have found that the migration of CD8+ T cells in lymph nodes in the absence of inflammation is reasonably well-described by a model with two distinct populations of stochastic walkers ( Fig . 3 ) . The first population migrates by a pure Brownian walk . The other population , comprising most of the walkers , migrates by a persistent random walk and is subject to a small amount of Brownian noise . This model can also be interpreted as the aggregate of a paused population and an active population with a small amount of overall noise . While this model has some similarities with existing models , it differs from previous models for CD8+ T cell migration in the lymph node , which describe cells as a single , homogeneous population of persistent random walkers [5 , 19–22] . Our model explicitly incorporates heterogeneity in the CD8+ T cell population . Furthermore , in contrast to previous run/pause models for T cells in the lymph node [21] , runs and pauses are not well-mixed . Instead , in our refined model , cells that are essentially paused except for slow Brownian-like motion , remain in the paused state for many minutes at a time ( the entire duration ) , and similarly , cells migrating by persistent random walks move continuously for at least ten minutes . This model , in contrast to existing models and common practice [1 , 41–43] , does not choose an arbitrary speed cutoff below which cells are assumed to be paused and data is discarded . Most importantly , this model successfully describes multiple cell migratory statistical measures ( Fig . 3 ) , rather than just the MSD . There are a variety of possible explanations for the observed heterogeneity in migratory behavior . For instance , in these experiments , CD8+ T cells were imaged throughout the lymph node , and thus , likely in multiple zones within the lymph node . Thus , it is possible that the two populations in the model represent migration in distinct regions of the lymph node . Alternatively , previous studies using this experimental system have shown that this population of OT-I CD8+ T cells express variable levels of the chemokine receptors CCR5 and CCR7 [24] , which could impact migration . Finally , it is unlikely that the paused population of cells arises from an experimental artifact such as phototoxicity . Indeed , paused cells are alive since they exhibit shape fluctuations and even , after very long time scales , begin to migrate . Interestingly , T cells in uninflamed lymph nodes do not migrate via generalized Lévy walks , as activated T cells do in the brain during chronic toxoplasmosis [6] . While generalized Lévy walks may enable T cells to efficiently find rare target parasites [6] , the long runs in the Lévy walk may be less beneficial for T cells that must frequently interact with dendritic cells . The observed differences in migration statistics may be indicative of the cell extrinsic or intrinsic differences between the two cell populations . For instance , it is likely that the structural features within these tissues , which may act as a scaffold for cell crawling [1 , 21 , 44] , are different in the two tissues . In addition , the T cell population in lymph nodes in the steady state is not as activated as the cells in the brain , which could also affect migratory behavior . Together , these observations suggest that CD8+ T cell populations with distinct functions migrate differently . While the model is successful in characterizing many aspects of cell migration ( Fig . 3 ) , there are deviations from measured cell statistics . These differences reflect the difficulty in systematically identifying a simple model that accurately and comprehensively describes walk statistics . Both fluctuations within individual cells and variations within cell populations complicate the overall behavior and ensuing analysis . This problem may be mitigated by accumulating better statistics; for example , if individual cells could be followed for hours throughout the entire lymph node , instead of minutes in a small volume , the whole-population analysis could instead be carried out for individual cell tracks . Additionally , further improvements to analytical and computational methods will lead to more accurate cell migration modeling . However , while various measures and techniques to understand and describe migratory behavior have been developed [4–6 , 15 , 36 , 45–49] , less has been done to systematically build robust migration models . The methodology described in this paper , combined with generalizations of new techniques for analyzing heterogeneous migration statistics [36 , 46 , 48 , 49] , could achieve this goal . In general , use of these statistical approaches require relatively large amounts of data ( more than 100 cell tracks ) and numerical simulations of random walk models . Despite these difficulties , these methods can provide powerful insights into cell migratory behavior , and they will be useful for characterizing migration in future studies . In turn , developing these more accurate models will help connect cell migration to immune function and lead to a deeper understanding of immune response .
All procedures involving mice were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania ( Animal Welfare Assurance Reference Number #A3079–01 ) and were in accordance with the guidelines set forth in the Guide for the Care and Use of Laboratory Animals of the National Institute of Health . CD8+ T cells were isolated as previously described [24] . Briefly , cells were isolated from the spleen and peripheral lymph nodes of DPE-GFP OT-I transgenic mice ( OT-IGFP ) . Single cell suspensions were obtained by mechanical homogenization . Red blood cells were removed by hypotonic lysis . T cells were purified using the mouse T cell enrichment columns ( R&D systems , Minneapolis , MN ) . 2–5 × 106 purified OT-IGFP cells were injected into recipient mice intravenously . Mice were euthanized by CO2 asphyxiation 16–24 hours following T cell transfer . The mesenteric lymph nodes were removed immediately , with minimal mechanical disruption . The lymph nodes were immobilized in 1% agarose in a heated chamber where specimens were constantly perfused with warmed ( 37°C ) , oxygenated media ( phenol-red free RPMI 1640 supplemented with 10% FBS , Gibco ) . The temperature in the imaging chamber was maintained at 37°C using heating elements and monitored using a temperature control probe . Imaging was performed using a Leica SP5 multi-photon microscope system ( Leica Microsystems , Mannheim , Germany ) equipped with a resonant scanner , picosecond laser ( Coherent Inc . , Santa Clara , CA ) , and external detectors that allow simultaneous detection of emissions of different wavelengths . Enhanced GFP was excited using laser light of 920 nm . Images were obtained using a 20X water-dipping lens . Four-dimensional imaging data was collected by obtaining images from the x- , y- , and z-planes , with a z-thickness of 68 μm and step size of 4 μm to allow for the capture of a complete z-series every 20 seconds for period of 15 minutes . 8 separate image series were taken . Individual cells were tracked using Volocity software ( PerkinElmer , Waltham , MA ) , giving the x- , y- , and z-coordinates of each cell at every time point . We implement standard Brownian dynamics algorithms for numerical walker models [50] . For each walker , we draw a run time , t , and speed , | v ⃗ | , from distributions using pseudorandom number generators . A direction for the velocity vector v ⃗ is also uniformly randomly chosen . At each time step , we add v ⃗ δ t to the walker position . We use a time step of δt = 0 . 001 s . When the run time , t , has passed , we randomly draw new run times , speeds , and directions . Diffusive noise is added by adding ( 6 D δ t ) r ^ to the walker position at each time step , where r ^ is a unit vector that points in a random direction . We simulate 5 , 000 walkers in a 600 μm × 600 μm × 170 μm volume , but only collect data on walkers if they are within a particular 500 μm × 500 μm × 68 μm “imaging” volume . Prior to data collection , we simulate walkers for a short period of time ( typically a minute ) in order to avoid artifacts at early times . The 2D moment of inertia tensor , I , is given by [31 , 32]: I = ( I x x I x y I y x I y y ) , I x x = ∑ i N y i 2 , I x y = I y x = − ∑ i N x i y i , I y y = ∑ i N x i 2 . ( 4 ) Ixx , Ixy , Iyx , and Iyy are summations of products of displacements , xi and yi . For the average moment of inertia tensor , I ¯ , the summation runs over all cellular motions , so that i indexes individual steps and N is the total number of steps . For individual track tensors , In , the summation is over all displacements for the individual cell , so while i still indexes individual displacements , N is now the number of frames for the track . An inertia tensor , In , can be calculated for each track , but only tracks of the same length ( e . g . , complete tracks ) should be averaged together . In order to find the eigenvalues of I , follow standard linear algebra methods [31 , 32] . The eigenvalues are found by the solving det ( I − λ 1 ) = 0 so that λ ± = I x x + I y y ± I x x 2 − 2 I x x I y y + I y y 2 + 4 I x y 2 , where λ1 = λ+ and λ2 = λ− . Finally , note that for the asphericity calculation , one typically takes the moment of inertia tensor about the center of the track [33]; to do this replace xi and yi in Eq . 4 with x i − x ¯ and y i − y ¯ , respectively , where x ¯ = 1 N ∑ i N x i and y ¯ = 1 N ∑ i N y i . | Migration is fundamental to immune cell function , and accurate quantitative methods are crucial for analyzing and interpreting migration statistics . However , existing methods of analysis cannot uniquely describe cell behavior and suffer from various limitations . This complicates efforts to address questions such as to what extent chemotactic signals direct cellular behaviors and how random migration of many cells leads to coordinated immune response . We therefore develop methods that provide a complete description of migration with a minimum of assumptions and describe specific quantities for characterizing directional motion . Using numerical simulations and experimental data , we evaluate these measures and discuss methods to minimize the effects of experimental artifacts . These methodologies may be applied to various migrating cells or organisms . We apply our approach to an important model system , T cells migrating in lymph node . Surprisingly , we find that the canonical Brownian-walker-like model does not accurately describe migration . Instead , we find that T cells move heterogeneously and are described by a two-population model of persistent and diffusive random walkers . This model is completely different from the generalized Lévy walk model that describes activated T cells in brains infected with Toxoplasma gondii , indicating that T cells exhibit distinct migration statistics in different tissues . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2015 | Heterogeneous CD8+ T Cell Migration in the Lymph Node in the Absence of Inflammation Revealed by Quantitative Migration Analysis |
The Schistosoma mansoni Venom-Allergen-Like proteins ( SmVALs ) are members of the SCP/TAPS ( Sperm-coating protein/Tpx-1/Ag5/PR-1/Sc7 ) protein superfamily , which may be important in the host-pathogen interaction . Some of these molecules were suggested by us and others as potential immunomodulators and vaccine candidates , due to their functional classification , expression profile and predicted localization . From a vaccine perspective , one of the concerns is the potential allergic effect of these molecules . Herein , we characterized the putative secreted proteins SmVAL4 and SmVAL26 and explored the mouse model of airway inflammation to investigate their potential allergenic properties . The respective recombinant proteins were obtained in the Pichia pastoris system and the purified proteins used to produce specific antibodies . SmVAL4 protein was revealed to be present only in the cercarial stage , increasing from 0–6 h in the secretions of newly transformed schistosomulum . SmVAL26 was identified only in the egg stage , mainly in the hatched eggs' fluid and also in the secretions of cultured eggs . Concerning the investigation of the allergic properties of these proteins in the mouse model of airway inflammation , SmVAL4 induced a significant increase in total cells in the bronchoalveolar lavage fluid , mostly due to an increase in eosinophils and macrophages , which correlated with increases in IgG1 , IgE and IL-5 , characterizing a typical allergic airway inflammation response . High titers of anaphylactic IgG1 were revealed by the Passive Cutaneous Anaphylactic ( PCA ) hypersensitivity assay . Additionally , in a more conventional protocol of immunization for vaccine trials , rSmVAL4 still induced high levels of IgG1 and IgE . Our results suggest that members of the SmVAL family do present allergic properties; however , this varies significantly and therefore should be considered in the design of a schistosomiasis vaccine . Additionally , the murine model of airway inflammation proved to be useful in the investigation of allergic properties of potential vaccine candidates .
Schistosomiasis is an important parasitic disease , caused by trematode worms of the genus Schistosoma , affecting more than 200 million people worldwide , with a further 650 million individuals living at risk of infection , remaining a major public health problem in many developing countries [1] . Transmission occurs through human contact with water containing the cercariae , the infective larval stage . These penetrate the skin , maturing into schistosomula , which reach the lungs via the systemic circulation . In the lungs , the young parasites undergo morphological transformations , gathering in the portal system , where they mature into adult worms . After pairing , the onset of egg deposition in the intestinal lumen , leads to a range of morbidities , such as granulomatous inflammation and periportal fibrosis [2] . A fraction of the eggs is eliminated with excreta , reaching the fresh water supply , where the miracidia hatch , infecting Biomphalaria snails . From these intermediate hosts the cercariae are released into the water to infect the definitive human host , closing the cycle [3] . Within the publication of the transcriptome data for Schistosoma mansoni , a series of novel genes/proteins were selected as potential vaccine candidates based on their functional classification by Gene Ontology functions , which would indicate their surface exposure to allow interaction with the host immune system [4] . Among them , four members of a family of wasp venom allergen orthologs were identified , raising the question of what benefits would there be to the parasite in amplifying allergic or other inflammatory responses in the host interface . Recently , this gene family was formally named as Schistosoma mansoni Venom Allergen-Like proteins ( SmVALs ) and its individual members analyzed concerning the phylogenetic relationships , genomic organization and mRNA expression profile across the life cycle [5] . This work revealed that it is a large family of genes composed by 28 members , with at least 24 members transcriptionally active , which can be divided into two groups; those with a signal peptide that may be released and interact with their immediate environment ( group 1 ) , and those without a secretion signal that should play an intracellular role ( group 2 ) [5] . Since it was the first article to deal specifically with this schistosome gene family , herein we will follow their proposed numbering and nomenclature . Following the transcriptome work , a series of proteomic studies describing different aspects of schistosome life cycle and biology were reported [6] , [7] , [8] , [9] , [10] , [11] , [12] , [13] , [14] , mostly using the sequence database of the S . mansoni and S . japonicum transcriptomes [4] , [15] , as well as the recently published genome databases [16] , [17] . Noteworthy , were the studies on the released proteins ( RP ) into the skin during the transition from cercariae to schistosomula [7] , [10] , since these proteins could be the first ones to be accessible to the immune system . In one of these studies , three different members of the previously described wasp venom allergen orthologs family ( SmVAL4 , 10 and 18 ) were identified as potential immunomodulators [7] and , more recently , SmVAL10 and 18 were characterized as glycosylated secreted proteins after cercarial transformation [9] . Moreover , in a report using a more accurate model to mimic cercariae penetrating human skin , SmVAL4 was detected in the forming tunnels as a secreted protein , 2 hours post cercariae invasion [8] . In a study integrating the transcriptome and proteomic data from S . japonicum , several orthologs of this protein family were identified [11]; worthy of notice , was an SjVAL ortholog detected in the tegument of schistosomula , sharing 73% of identity with SmVAL26 . Also of interest , SmVAL6 , a group 2 family member , was identified as a tegument-exclusive protein in a sub-proteome analysis of S . mansoni [13] ( Figure S1 ) . In schistosomiasis , morbidity and mortality have been associated with egg deposition , therefore identifying the components of Egg Secreted Proteins ( ESP ) is important to understand how these antigens can regulate the surrounding cytokine environment . Using the proteomics approach , four different members of the SmVAL family ( SmVALs 2 , 3 , 5 and 9 ) were identified as ESP [6] . However , in marked contrast , more recently , no one of these proteins were identified in egg secretions by another proteomic study [12] , emphasizing how methodological differences can result in diverse conclusions . An additional proteomic study reinforced the wide-spread distribution of this family along the life cycle , by the identification of several different SmVALs ( 2 , 3/23 , 9 , 15 , 26/28 , and 27 ) released during in vitro miracidium-to-sporocyst transformation [14]; most of this data are summarized in Figure S1 . A natural question that emerged from all these studies is the biological function of these genes in the host-parasite interface . Some of these molecules were suggested by us and by other groups as potential vaccine candidates or immunomodulators , due to their functional classification , expression profile and predicted localization [4] , [5] , [7] , [8] , [9] . Additionally , SmVALs members present sequence similarity to the hookworm lead vaccine candidate NaASP-2 [18] , [19] , [20] . From a vaccine perspective , a major concern is the potential allergic effects of these molecules . Herein , we tried to investigate the immunomodulatory properties of some SmVALs by exploring the murine model of airway inflammation [21] . The investigation of localized inflammation in tissues is often difficult because it is hard to isolate the immune response against a particular stimulus . Therefore , the utilization of an inflammatory model in a confined location can be useful to monitor changes in cell population and to identify regulatory mechanisms . We selected SmVALs from group 1 , which would be putatively expressed in intra-host stages preferably exposed to interaction with the immune system . Therefore , we selected SmVAL4 , which would be released in the transition between cercariae and schistosomula [7] , [8] and SmVAL26 , which would probably be in the tegument of schistosomula due to its S . japonicum ortholog identification in the tegument [11] . The respective recombinant proteins were obtained in an eukaryotic expression system and the purified proteins used to produce specific antibodies . The protein expression profile was characterized across the life cycle stages . The allergic properties of SmVAL4 and SmVAL26 proteins were investigated in the murine model of airway inflammation . Our results show that the allergic properties of these molecules vary significantly and this should be considered in the putative design of a schistosomiasis vaccine .
Schistosoma mansoni adult worms ( BH strain ) were obtained by perfusion of hamsters , 6 weeks after infection with 200 cercariae; eggs were extracted from infected hamster liver by maceration and partial digestion with collagenase followed by washes and passage through sieves and percoll gradients as previously described [22]; miracidia were obtained by exposing purified eggs to a bright light; cercariae were harvested from infected B . glabrata snails exposed to light . Following in vitro transformation of cercariae , schistosomula were cultured for 0–6 hours or 7 days prior to recovery [23] . The procedures involving animals were carried out in accordance with the Brazilian legislation ( 11790/2008 ) . All animals were handled in strict accordance with good animal practice and protocols were previously approved by the Ethical Committee for Animal Research of Butantan Institute , under the license number 604/09 . The signal peptide prediction was performed using the SignalP 3 . 0 server ( http://www . cbs . dtu . dk/services/SignalP/ ) , N-glycosylation sites were analyzed using the NetNGlyc 1 . 0 ( www . cbs . dtu . dk/services/NetNGlyc/ ) , O-glycosylation sites were analyzed using the OGPET ( http://ogpet . utep . edu/OGPET/ ) , and transmembrane helices were analyzed by TMHMM version 2 . 0 ( http://www . cbs . dtu . dk/services/TMHMM-2 . 0/ ) . Molecular weight ( MW ) and isoelectric point ( pI ) were calculated with the Compute pI/Mw tool ( http://www . expasy . ch/tools/pi_tool . html ) . The plasmids containing SmVAL4 ( optimized sequence ) and SmVAL26 ( optimized sequence ) were linearized with SacI and the P . pastoris strain GS115 ( Invitrogen ) was transformed by electroporation following the instructions of the manufacturer . Twenty colonies were first isolated and purified in YPDS plates containing 100 µg/mL Zeocin , to select putative multi-copy recombinants in YPDS plates containing 500 , 1000 , and 2000 µg/mL Zeocin . To verify production of the relevant proteins , initial studies were done in small-scale expression conditions , followed by Western blot with anti-His-tag antibody ( GE ) . Fermentation conditions were carried out as per manufacturer's recommendations . Briefly , selected P . pastoris cells were grown in 15 mL of BMGY 28–30°C in a shaking incubator ( 250–300 rpm ) until cultures reached an OD600 = 2 . 0 ( approximately 16–18 h ) . The cells were harvested by centrifuging at 3000× g for 5 min at room temperature , the supernatant was decanted and cells resuspended in 20 mL of BMMY medium to induce expression . Methanol was added to a final concentration of 0 . 5% methanol every 24 h to maintain induction; expression was monitored at 48 and 96 h time points . The supernatants and cell pellets for 15 colonies of each SmVAL was analyzed for protein expression by Western blot . Those colonies that presented the highest expression level were selected for scale-up fermentation . For protein expression and purification , selected clones for SmVAL4 and SmVAL26 were grown ( 28°C , 250 rpm ) in 25 mL of BMGY in a 250 mL baffled flask until OD600 = 2 . 0 ( approximately 18 h ) , then inoculated in 300 mL of BMGY in a 2 . 0 L baffled flask and grown in the same conditions until culture reached OD600 = 2−6 . The cells were harvested ( 3000× g , 5 min at room temperature ) and resuspended in 600 mL of BMMY to start induction . Methanol was added to a final concentration of 0 . 5% every 24 h to maintain the induction . Cells were harvested after 48 h ( SmVAL26 ) and 96 h ( SmVAL4 ) by centrifugation . The culture medium containing the secreted proteins ( rSmVAL4 and rSmVAL26 ) were filtered through a 0 . 22 µm membrane , and diluted with 3 volumes of equilibration buffer ( 50 mM sodium phosphate pH 5 . 8 ( for rSmVAL4 ) and pH 7 . 2 ( for rSmVAL26 ) , 150 mM NaCl , 20 mM imidazole ) . The recombinant proteins were then purified by metal affinity chromatography using the Akta Prime system ( GE Healthcare ) under native conditions . Briefly , the sample was loaded onto a Ni2+-NTA column ( 5 mL bed volume ) pre-equilibrated with the same buffer . The column was washed with 30 bed volumes of the equilibration buffer and then eluted with 20–500 mM imidazole linear gradient . Fractions encompassing the main peak and the purity of the preparation were assessed by SDS-PAGE . Before its use the proteins were dialyzed in Phosphate Buffer Saline pH 7 . 4 ( PBS ) . CD measurements were carried out on a Jasco J-810 Spectropolarimeter at 20°C equipped with a Peltier unit for temperature control . Far-UV CD spectrum was acquired using a 1 mm path length cell at 0 . 5 nm intervals over the wavelength range from 190 to 260 nm . Five scans were averaged for each sample and subtracted from the blank average spectra . The protein concentration was kept at 10 µM in 10 mM sodium phosphate buffer pH 7 . 4 . Total protein extracts from whole parasite stages ( eggs , miracidia , cercariae , in vitro 7-day-old schistosomula and adult worms ) were prepared in 40 mM Tris , pH 7 . 4 , 2% SDS , plus protease inhibitor cocktail ( Sigma ) by sonication ( 4 cycles of 2 min , with pulses of 0 . 75 s , 40% amplitude ) . The samples were centrifuged at 20 , 000× g for 30 min at 4°C and the supernatant was recovered and used for the assays . The tegument extract was obtained by a freeze/thaw/vortex procedure , as previously described [24] . Their protein concentrations were determined with a DC Protein Assay Kit ( Bio-Rad , CA , USA ) . Purified rSmVALs ( 50 ng ) , total parasite protein extracts ( 20 µg ) , and total tegument extract ( 20 µg ) were subjected to SDS-PAGE . The gel was electroblotted onto a PVDF membrane , which was blocked with 0 . 02 M Tris ( pH 7 . 5 ) and 0 . 3% Tween 20 containing 5% dry milk for 16 h at 4°C . Subsequently , the membrane was incubated in 1∶4000 or 1∶3000 dilution of anti-rSmVAL4 and anti-rSmVAL26 primary antibody , respectively , in blocking buffer plus 150 mM NaCl for 3 h at room temperature . After three washes using 150 mL of 10 mM Tris ( pH 7 . 5 ) , the membrane was incubated in a 1∶2000 dilution with secondary goat anti-mouse IgG conjugated to horseradish peroxidase ( Sigma ) for 1 h , followed by another three washes using the same buffer . Antibody reactivity was developed with ECL reagent ( GE Healthcare ) according to the manufacturer's instructions and imaged using Hyperfilm ( GE Healthcare ) . N-deglycosylation of native and recombinant proteins was carried out as previously described [25] . Briefly , 20 µg of parasite extracts ( in 40 mM Tris , pH 7 . 4 , 0 . 7% SDS , 1% 2-mercaptoethanol ) or 10 µg of recombinant proteins ( in PBS , pH 7 . 2 , 0 . 7% SDS , 1% 2-mercaptoethanol ) were denatured by boiling for 10 min . NP-40 ( Sigma ) was added ( 1% final concentration ) , 2 µL of recombinant N-glycosidase F ( 500 U/µL ) ( New England Biolabs ) , 20 mM sodium phosphate ( pH 7 . 5 ) for a final volume of 20 µL , and incubated overnight at 37°C . For subsequent O-deglycosylation of native proteins the following enzymes were used α ( 2→3 , 6 , 8 , 9 ) Neuraminidase , O-glycosidase , β ( 1→4 ) -Galactosidase and β-N-Acetylglucosaminidase , and the reactions were carried out as per the manufacturer's recommendations ( Sigma ) . Samples of purified rSmVALs and protein parasite extracts digested and non-digested were submitted to SDS-PAGE . The gels containing the recombinant proteins before and after treatment with glycosidases were stained with Schiff's reagent ( Sigma ) for detection of glycoproteins as per the manufacturer's recommendations; as control for the specificity of the reaction , we used Bovine Serum Albumin ( Bio-Rad ) at the same concentration of the recombinant proteins . To analyze the glycosylation pattern of native SmVAL proteins , the protein parasite extracts treated with glycosidases were electroblotted onto a PVDF membrane and Western blot developed as described above . Shedding of cercariae from snails was stimulated by exposure to bright light . Mechanical transformation by vortexing was used to stimulate the release of gland cell contents of 2000 parasites during a time course of 0–6 h culture period in 2 mL of RPMI 1640 medium ( Invitrogen ) containing 300 units/mL penicillin and 300 g/mL streptomycin at 37°C in 5% CO2 . Parasites were collected by centrifugation at 200× g at 4°C for 5 min and processed as previously described; the supernatant ( medium containing released proteins ) was stored at −20°C with the addition of 2 µL of 10× general use protease inhibitor mixture ( Sigma ) . This soluble preparation , termed the 0–6 h released proteins ( RP ) , was precipitated with trichloroacetic acid ( TCA ) and used for Western blot analysis . For egg released protein collection , eggs were isolated as previously described [22] , followed by a separation of mature and immature eggs as previously described [26] . ESP was collected by incubating 3 . 0 million mature eggs in 10 mL of serum-free RPMI ( Invitrogen ) for 72 h at 37°C in 5% CO2 . Post-culture viability of eggs was >85% , as assessed by observation of muscular and flame cell activity in unhatched miracidia . Culture medium containing ESP was precipitated with TCA and resuspended in 40 mM Tris , pH 7 . 4 , 2% SDS , plus protease inhibitor cocktail ( Sigma ) for Western blot analyses . The Hf was obtained by hatching mature eggs through exposure to bright light for 1 h in pound water . Miracidium and egg shell were pelleted by centrifugation at 200× g at 4°C for 10 min , and the protein water-soluble content was carefully collected to avoid turbulence , the sample ( Hf ) was filtered through a 0 . 22 µm polyethersulfone filter ( Millipore ) and concentrated through precipitation with TCA . For elimination of the protein moiety , 12 µg of rSmVAL4 was incubated with 0 . 05% of pronase ( Calbiochem , San Diego , CA ) for 2 h at 37°C . A separate sample of rSmVAL4 was incubated for 2 h at 37°C without pronase to provide an undigested control . Successful pronase digestion was confirmed by gel electrophoresis followed by Coomassie staining , which revealed the absence of any remaining protein ( data not shown ) . Female BALB/c mice ( 6–8 wk old ) were weight matched and used throughout this study . To test whether rSmVALs could induce an inflammatory reaction in the lungs , we adapted the well-established murine model of asthma based on Ovalbumin ( OVA ) /alum sensitizations and challenges [21] . Briefly , mice were sensitized on days 0 , 7 and 14 by subcutaneous ( s . c . ) injection ( 0 . 4 mL total volume ) in the nape of the neck with 10 µg rSmVALs adsorbed to 2 . 0 mg of aluminum hydroxide ( Alhydrogel-Brenntag Biosector ) . On days 21 and 28 , mice were challenged i . n . with 10 µg of rSmVALs in 50 µL of PBS . Two additional groups were used: Control group received only the intranasal challenge with rSmVAL and naïve mice received only PBS . It is important to mention that mice were anesthetized intramuscularly with 100 µL of a solution containing ketamine ( Ketamina Agener , Uniao Quimica Farmaceutica Nacional ) and xylazine ( Bayer ) before any immunization or challenge , to ensure a complete instillation to the lungs , as previously described [27] . Twenty four hours after the last challenge , mice were deeply anesthetized by an i . p . injection of urethane ( Sigma-Aldrich ) at 15 mg/10 g body weight , the abdominal cavity was opened , and blood samples from the inferior cava vein were collected for serum antibody determinations . The trachea was cannulated and lungs were lavaged twice with 0 . 5 and 1 . 0 mL of cold PBS . After total cell counting , cytospin preparations of bronchoalveolar lavage ( BAL ) cells were stained with Instant-Prov ( Newprov ) and differential cell counts were performed on 200 cells on the basis of morphology and staining characteristics . Supernatants from BAL were collected and frozen at −80°C for cytokines measurements . Anti-SmVALs antibodies were assayed by sandwich ELISA , as previously described [21] . Briefly , serum samples were titrated for optimal dilutions for testing different isotypes . For SmVALs-specific IgE determinations , plates were coated with goat anti-mouse unlabelled IgE ( 1∶250; BD Bioscience ) following the manufacturer's recommendations; serum samples were incubated ( 1/10 dilution ) for 2 h at room temperature and subsequently biotin-labeled rSmVALs ( 5 µg/mL ) were added to the wells . The biotinylated rSmVALs were prepared by reacting 1 mL rSmVALs ( 1 mg/mL ) in PBS with 100 µL of N-hydroxysuccinimidobiotin in dimethyl sulfoxide ( DMSO ) ( 4 mg/mL ) for 4 h at room temperature , followed by overnight dialysis against PBS at 4°C . The bound rSmVALs-biotin was coupled to streptavidin-peroxidase 1∶250 , for 15 min incubation at room temperature and revealed as per the manufacturer's recommendations . SmVAL-specific IgE levels of samples were expressed by OD . For SmVALs-specific IgG1 and IgG2a antibodies , serum samples were plated on 96-well plates previously coated with rSmVALs ( 0 . 5 µg/well ) . The bound antibodies were revealed with goat anti-mouse IgG1 or IgG2a followed by peroxidase-labeled rabbit anti-goat antibodies ( all from Southern Biotech ) . The concentration of each specific isotype was estimated by comparison with IgG1 and IgG2a standards run in parallel and expressed as the mean ± SEM of the antibody concentration of 4 mice per group . The cytokine concentration in the BAL fluid was quantified by ELISA kits specific for IL-5 and IL-10 ( BD Biosciences PharMingen ) and for IFN-γ ( Peprotech INC . ) . The values are expressed as picograms per milliliter deduced from standards , run in parallel with the recombinant cytokines . The limit of detection values were 10 pg/mL for IL-5 and IL-10 and 16 pg/mL for IFN-γ . The anaphylactic activity of reactogenic antibodies was evaluated by passive cutaneous anaphylactic reaction in mice as described by Ovary et al . [28] . Previously shaved mice were injected intradermally with 50 µL of three serial dilutions of serum in each side of the dorsal skin . After 2 h , they were challenged intravenously with 250 µg of rSmVAL4 , rSmVAL4-Pro ( Pronase-digested ) or rSmVAL26 , all plus 0 . 25% of Evans blue solution . All determinations were made in triplicate and the PCA titers were expressed as the reciprocal of the highest dilution that gave a lesion of >5 mm in diameter . The detection threshold of the technique was established at 1/5 dilutions . Polyclonal mouse sera were produced against preparations of rSmVAL4 and rSmVAL26 . BALB/c mice were immunized three times , subcutaneously in the nape of the neck , at 14-day intervals with 25 µg of rSmVAL4 or rSmVAL26 formulated with TiterMax adjuvant ( CytRx Corporation; first dose ) or PBS 1x ( in subsequent doses ) . Fifteen days after the last inoculation , rodents were exsanguinated . The sera were used at a dilution of 1∶4000 ( anti-rSmVAL4 ) and 1∶3000 ( anti-rSmVAL26 ) in Western blots . Student's t-test was used to compare experimental and control groups on antibody and cytokine levels . For cellular migration assays and analysis involving more than two groups , statistical comparisons were performed with one-way ANOVA followed by a Bonferroni pairwise comparison . A ρ value<0 . 05 was considered statistically significant .
In order to analyze SmVALs that could interact with cells in the definitive host , we chose 2 members of group 1 . SmVAL4 would be released in the transition between cercaria and schistosomula . A phylogenetic analysis of SmVALs and SjVALs ( Figure S2 ) , revealed that SmVAL26 branched together with a SjVAL ortholog , with 73% of identity detected in the tegument of hepatic schistosomula ( Gene Bank accession AAW27353 . 1 ) ( Figure S3 ) . SmVALs contain the sperm-coating protein ( SCP ) signature sequence ( outlined by a dark grey box in Figure 1A ) and are recognized as part of the Pfam SCP family ( PF00188 ) with an E-value ranging from 6 . 66×10−24 to 8 . 20×10−31 . The SmVALs under investigation , once belonging to group 1 , also contain a putative N-terminal signal peptide ( outlined by a grey box in Figure 1A ) . Putative N and O-glycosylation sites were identified and investigated ( identified by * and # , respectively in Figure 1A ) . The predicted molecular mass and isoelectric points of these proteins are presented in Table 1 . The recombinant proteins rSmVAL4 and rSmVAL26 were expressed using codon optimization in P . pastoris GS115 strain and secreted into the culture supernatant ( products around 30 kDa and 20 kDa , respectively ) ( Figure 1B and 1C ) . rSmVAL4 and rSmVAL26 were purified by affinity chromatography on nickel-charged columns eliminating the main contaminant from both samples ( around 66 kDa ) in the flow through ( Figure 1B and 1C ) . The eluted fractions of rSmVAL4 showed two main bands at ∼30 and 34 kDa ( Figure 1B lanes 5–8 ) , while rSmVAL26 eluted fractions presents only one product ( ∼20 kDa ) ( Figure 1C , lanes 4–7 ) . Eluted fractions were pooled and submitted to extensive dialysis in PBS pH 7 . 2; protein yield after dialysis were estimated to be around 10 . 0 mg of rSmVAL4/L and 6 . 0 mg of rSmVAL26/L of culture . These samples were used in the sensitization and challenge assays and to generate polyclonal antibodies in mice . Both Pichia-secreted proteins migrate with a higher molecular mass than that predicted ( ∼30 . 0 kDa for rSmVAL4 and ∼20 kDa for rSmVAL26 ) ( Table 1 ) , which could reflect a likely product of post-translational glycosylation . To test this hypothesis , we digested the purified proteins with a recombinant N-glycosidase F , and also stained the gels with Schiff's reagent to reveal the presence of glycans . The SDS-PAGE showed that , after digestion , rSmVAL4 migrated to a lower MW ( ∼25 kDa ) , demonstrating the removal of N-glycans ( Figure 1D ) , while rSmVAL26 does not show any mobility shift ( Figure 1F ) , suggesting that the protein was not N-glycosylated . Additionally , for both proteins , the Schiff's reagent staining procedure revealed the presence of glycans after treatment with PNGase F , suggesting the presence of a PNGase F insensitive N-linked glycan site or an O-linked glycosylation site ( Figure 1E and G ) . Circular dichroism spectra revealed that rSmVAL4 and rSmVAL26 display an ordered secondary structure , which resembles the NaASP-2 protein ( structure resolved – three layer α-β-α-sandwich ) ( Figure 1H ) , although the proportions of secondary structure elements ( α-helix and β-sheet ) were not calculated . Samples prepared from cercariae , schistosomula , adult worms , eggs and miracidia stages of S . mansoni , and tegument isolated by the freeze/thaw method , were all separated by SDS-PAGE . Immunoblotting was performed using mouse anti-rSmVAL4 and rSmVAL26 antisera . The protein expression profile of SmVALs revealed a very specific stage associated expression . Briefly , the expression of SmVAL4 seems to be restricted to the cercariae stage . In the case of SmVAL26 , although we expected it to be in the tegument due to its similarity with the S . japonicum ortholog , it is actually detected in the egg , but not in the miracidium stage ( Figure 2A and B ) . It is important to note that in the assessed experimental conditions , no sign of cross reactivity was observed with SmVALs in other stages or in the tegument fraction . Noteworthy , the SmVAL4 and SmVAL26 native proteins detected in schistosome extracts migrate with a higher molecular mass than that predicted ( Table 1 ) , which again could reflect a likely product of post-translational glycosylation . To test this hypothesis , we digested total cercariae and egg extracts with the recombinant N-glycosidase F . The immunoblot showed that , after digestion , the native proteins displayed a shift in the migration , demonstrating that native SmVAL4 is N-glycosylated , whereas SmVAL26 is not ( Figure 2C and D ) . In order to investigate possible O-glycosylations , the extracts were treated with neuraminidase , O-glycosidase , beta ( 1–4 ) galactosidase , or N-acetylglucosaminidase , all of which had no effect on the proteins' mobility on SDS-PAGE ( data not shown ) . To investigate the presence of SmVAL4 in the secretions of newly transformed schistosomula , we collected the proteins released by 2000 parasites in a time course manner . We were able to detect the secretion of SmVAL4 as early as 30 min after transformation and its increasing secretion in the medium of cultured schistosomula from 0–6 h ( Figure 3A ) . However , after this period , there are still significant amounts of protein within the parasite or associated to its surface ( Figure 3A ) . We also evaluated the presence of SmVAL26 in the egg secretions and in the hatched fluid , using the respective antibody . SmVAL26 was detected in both extracts ( Figure 3B and C ) . In order to investigate a putative immunomodulatory effect of rSmVALs , we explored the well-established murine model of airway inflammation induced by OVA/Alum sensitization and OVA intranasal challenge , replacing OVA by the rSmVAL4 and rSmVAL26 proteins . Our data revealed that mice sensitized and challenged with rSmVAL4 present an increased number of total cells in the bronchoalveolar lavage ( BAL ) , when compared with the control group . This effect comprises mainly an increase in eosinophils ( 55% ) and macrophages ( 32% ) , which resembles an allergic airway inflammatory response . Mice that received rSmVAL26 show a discrete increase in total cell counting , mostly macrophages ( Figure 4A and B ) . When mice received up to 6 doses of rSmVAL26 , this profile does not change significantly . The inflammatory response induced by rSmVALs was also evaluated in the lungs of mice by histopathology . A dense mixed-cellular infiltrate surrounding the airway ( peribronchovascular inflammation ) was evident only in the group treated with rSmVAL4 ( Figure 4D ) . Since the allergic effect could be due either to the proteic or the carbohydrate moieties of rSmVAL4 produced in Pichia pastoris , we eliminated the protein moiety by pronase treatment and performed the sensitization step with the pronase-treated rSmVAL4 ( rSmVAL4-Pro ) . As shown in Figure 4A and 4E , pronase treatment of rSmVAL4 totally abolished airway inflammation . The serum levels of SmVALs-specific IgG1 , IgG2a and IgE were measured in sensitized and challenged mice and in those only challenged with different rSmVAL proteins . The production of SmVAL-specific IgG1 , a Th2-affiliated antibody isotype , was significantly higher in all groups analyzed as compared to the control challenge group ( Figure 5A ) . Additionally , rSmVAL4 showed higher levels of IgG1 , and no antibody production was observed in response to immunization with pronase-treated rSmVAL4 . Concerning the production of SmVALs-specific IgG2a antibodies , very low levels and no significant differences were detected between experimental and control groups ( data not shown ) . The serum levels of SmVALs-specific IgE , a Th2 allergic associated isotype , were also measured , revealing significantly higher levels only in the rSmVAL4 group as compared to all other groups ( Figure 5B ) . We evaluated the secretion of IL-5 , IL-10 and IFN-γ in the BAL fluid . The levels of IFN-γ were below detectable levels ( 50 pg/mL ) in all analyzed groups , while the levels of IL-10 did not differ significantly from the sensitized and challenged groups to those only challenged ( data not shown ) . The levels of IL-5 were significantly higher in the rSmVAL4 immunized mice as compared to the control groups ( naïve or only challenged ) . SmVAL4-Pro group revealed intermediate levels of this cytokine , not differing statistically from control or rSmVAL4 ( Figure 6A ) . It is important to mention that this IL-5 secretion observed in SmVAL4-Pro was not sufficient to induce/mediate the eosinophil infiltration or other parameters of the allergic response . Taking into account the eosinophil migration , the presence of IL-5 in the BAL and the high levels of systemic IgE , we performed Passive Cutaneous Anaphylactic ( PCA ) assays to evaluate the production of anaphylactic IgG1 in heat-inactivated serum of mice sensitized and challenged with rSmVALs . Our results demonstrated that IgG1 antibodies produced by the rSmVAL4 group exhibited strong PCA activity with a very high titer of 1∶1250 , whereas in the other groups , including the pronase-treated rSmVAL4 or the control groups , low levels of anaphylactic IgG1 antibodies were observed ( Figure 6B ) . The induction of high levels of systemic IgE by rSmVAL4 in the model of airway inflammation led us to evaluate the use of a less Th2-prone adjuvant in a conventional protocol of immunization . Therefore , mice were immunized with rSmVALs formulated in TiterMax Gold , which is described to produce considerable levels of IgG2a in addition to IgG1 . The rSmVAL4 group showed higher levels of IgG1 antibody in relation to rSmVAL26 . Concerning the production of SmVALs-specific IgG2a antibodies , significant levels were detected in both immunized groups as compared to the control , with no differences between them . The levels of specific IgG1 and IgG2a and the IgG1/IgG2a ratio indicate that immunization with TiterMax Gold induced a predominant Th2 immune response to rSmVAL4 and a more balanced ( Th1/Th2 ) response to rSmVAL26 ( Figure 7A ) . The serum levels of SmVALs-specific IgE , were also measured , revealing significantly higher levels only in the rSmVAL4 group as compared to all other groups ( Figure 7B ) . Animals sensitized with rSmVAL4 and rSmVAL26 formulated in Titermax were also challenged intranasally and airway inflammation evaluated . It was observed that only the group receiving rSmVAL4 developed airway eosinophilic inflammation ( data not shown ) .
Recently , SmVALs have emerged from transcriptoma , microarray and proteomic studies as potential targets for immune intervention . In the present manuscript , we extend the previous molecular characterization performed by Chalmers et al . ( 2008 ) , focusing on the protein products of SmVAL4 and SmVAL26 , which may play different roles in the parasite-host interface . Our results describe the expression of the codon-optimized versions of SmVAL4 and SmVAL26 in Pichia pastoris . The secretion of these proteins in the yeast expression system allowed purification of the proteins in soluble form . Maybe the most relevant feature when producing a recombinant protein for functional assays is its correct folding . Our CD data indicated that the soluble secreted forms of rSmVAL4 and rSmVAL26 contained an ordered secondary structure , with similar proportions of structural elements ( α-helix and β-sheet ) as determined for NaASP-2 , which presents a three-layer ( α-β-α ) sandwich flanked by an N-terminal loop and a short cystein-rich C terminus [18] . The presence of glycans in such a class of secreted molecules is somewhat expected . Interestingly , both rSmVAL4 and rSmVAL26 were revealed to be glycosylated , which could help the proper folding and stabilization of the protein . It is important to note though , that the profile of glycosylation obtained in the Pichia system will not be equivalent to the native pattern in schistosomes . Concerning the native SmVAL4 protein , we confirmed the in silico predictions of its N-glycosylation by mobility shift . On the other hand , the native SmVAL26 was not N-glycosylated , which is also in conformity with the N-glycan in silico predictions . We did not detect the presence of O-glycans in the native SmVALs analyzed . However , we cannot completely exclude the presence of this class of carbohydrate , since the deglycosidases used were not specific for schistosomes , which could have modifications of the core structure , impairing or blocking enzymatic digestion . The previously determined mRNA expression profile across the life cycle suggested that SmVAL4 could be involved in the invasion of the definitive host [5]; this stage associated expression was confirmed at the protein level by our Western blot analysis . SmVAL4 protein was revealed to be present only in the cercarial stage , increasing in the secretions of newly transformed schistosomulum in a time course manner . This data could reflect the fact that not all parasites were at the same metabolic stage during transformation or that the protein continues to be released even after 6 hours post-transformation . The identification of SmVAL4 in cercariae secretions by proteomics suggested that it could be localized in the acetabular glands . However , the detection of a significant amount of protein in 6 h schistosomula implies another localization , which is being investigated by imunohistochemistry and whole mount in situ hybridization . Based on its similarity to a S . japonicum ortholog , we predicted that SmVAL26 could be in the tegument of young adults or hepatic schistosomula ( 14 day-old ) . However , contrary to our expectations , SmVAL26 was identified in the egg stage and in the hatched eggs' fluid . This is in accord with the proteomic study of Mathieson and Wilson ( 2010 ) , which compared the contents of egg , miracidium , egg secretions and hatch fluid . Interestingly , we also detected SmVAL26 in the secretions of cultured eggs . Based on our data , we hypothesized that SmVAL26 could be localized in the egg envelope , or as proposed by Mathieson and Wilson ( 2010 ) , in the protective fluid located between the miracidium and the envelope . The biological function of SmVALs and orthologs belonging to the SCP/TAPS domain super-family remains unclear . The NaASP-2 protein from N . americanus , secreted by the infective larvae , was described to induce neutrophil recruitment in vivo in the air pouch model of acute inflammation [29] . Two additional hookworm SCP/TAPS proteins were reported to have immunomodulatory activities . Ancylostoma caninum hookworm platelet inhibitor ( Ac-HPI ) exhibits an inhibitory effect on platelet aggregation , acting via glycoprotein Ia/IIa [30] , and A . caninum neutrophil inhibitory factor ( Ac-NIF ) showed ability to inhibit CD11b/18-dependant leukocyte function [31] , [32] . Concerning the SCP orthologs from filarial nematodes , the rOv-ASP-1 from Onchocerca volvulus , showed striking features ranging from angiogenic activity in mouse corneas , to being proposed as an adjuvant for bystander proteins due to its ability to bind to APCs and trigger Th1 proinflammatory cytokines [33] . It is clear that no common biological function or activity has been associated with this protein family . And it is possible , as proposed by Hewitson et al . [34] , that the SCP domain operates as an adaptable protein framework facilitating the evolution of various specialized functions . In principle , schistosome proteins implicated in the tissue invasion process , are particularly good candidate antigens for the development of vaccines and drugs . One major concern on the use of these and other SmVALs as vaccine candidates is the potential allergic effects of these molecules . Herein , we have explored the murine model of airway inflammation , substituting OVA for the SmVALs , to investigate the putative allergic responses induced by these proteins . Our results demonstrate that following sensitization and challenge with the different proteins , they present varying properties in regards to the recruitment of inflammatory cells to the BAL fluid . SmVAL4 was shown to induce a marked increase in total cells in the BAL fluid , mostly due to an increase in eosinophil and macrophages , which correlated with increases in IgG1 , IgE and IL-5 , characterizing a typical allergic response , while SmVAL26 showed no alterations in the allergic parameters in the lungs . Furthermore , the use of a Pronase-treated SmVAL4 , strongly supports the conclusion that the allergic properties are due to the protein itself and not to the carbohydrate moiety . We also demonstrated that anti-rSmVAL4 sera presented high titers of anaphylactic IgG1 antibody . Human Schistosoma infections have been associated with the inhibition of allergies . In this context , the murine model of ovalbumin induced airway inflammation has been used to demonstrate that Schistosome infection , egg extracts or some purified antigens modulate negatively the allergic response induced by OVA treatment by a mechanism mediated by T regulatory cells [35] , [36] . Intrinsic properties of the antigens must be important for this modulation , since not all molecules modulate equally the allergic response . It is important to note that our model is quite different , since we have investigated the allergenic properties of the antigens per se . On the other hand , in the course of a S . mansoni infection , two phases of immune responses have been recognized . In the first 3–5 weeks , during which the host is exposed to immature parasites , the immune response has been shown to be Th1-predominant . As the parasites mature , mate and begin to produce eggs at weeks 5–6 , the immune response alters markedly to a strong Th2 profile [37] . However , there are reasons to believe that responses to schistosome worms during established infections are more complex . For example , immune responses classically mediated by the Th2 cytokine IL-5 ( eosinophilia and eosinopoiesis ) were reported in the first weeks of infection [38] . Moreover , the induction of CD4+ T cells , specific IgE and basophils that produce IL-4 in response to worm antigens ( e . g . SmCB1 – Cathepsin B ) have also been described in this phase [39] . Finally , there is increasing evidence that excretory/secretory molecules from schistosome larvae can stimulate a mixed T helper response in the skin , with evidence of both Th1 and Th2 skewed responses at the site of infection [40] . Therefore , it is possible that SmVAL4 could be an antigen involved in the initial stage of the infection inducing a Th2-predominant response , whereas SmVAL26 may be more important in the late phase of infection . In a vaccine context , the immunization of mice with rSmVAL4 and TiterMax Gold ( a more balanced adjuvant ) produced high levels of IgE in a conventional immunization protocol . It is interesting to note , that when challenged i . n . after this protocol , only rSmVAL4 induced airway inflammation ( data not shown ) . This in itself poses risks , since a vaccination regime that promotes IgE production may well elicit undesirable side-effects such as exacerbation of allergy . Furthermore , in a recent report of a second Phase I trial , adult volunteers did experience allergic responses following immunization with the Necator ortholog , NaASP-2 [41]; this data advocates for the presence of IgE epitopes in this class of molecules . Therefore , it would be interesting to evaluate the levels of specific IgE antibodies for SmVAL4 in the sera of individuals resident in a schistosomiasis endemic area . The IgE epitopes for the SmVAL family remain to be determined , but it is tempting to conclude that the SmVAL4 protein contains some IgE epitopes absent on SmVAL26 . The functional evaluation of other cercariae secreted SmVALs closely related to SmVAL4 ( e . g . SmVAL10 and SmVAL18 ) ( Figure S4 ) in this model of airway inflammation could help the identification and mapping of such IgE epitopes . After the sequencing and assembly of the Schistosoma genome , efforts in schistosome research should shift from the simple identification of genes to the characterization of their functions and interactions with host cells . From this perspective , data presented here should be taken as a first insight on possible functions for members of the SmVAL family . They could have an immunomodulatory role that may be important during parasite penetration . One could hypothesize that SmVAL4 would be involved in the recruitment of mast cells and basophils , inducing secretion of histamine , which could facilitate parasite invasion through vessel dilatation . Additional studies , such as in situ hybridization and the evaluation of native proteins should further elucidate the localization and the role of these proteins . Finally , we believe that , although the airway inflammation model explored here presents some divergences from physiological conditions , it reveals and differentiates the allergic properties of molecules , proving to be useful for studying molecules with allergic potential . In selecting SmVAL molecules to be further investigated as vaccine candidates , we can eliminate those that display allergic potential in this model , such as SmVAL4 . | The Schistosoma mansoni Venom Allergen Like proteins ( SmVALs ) have been identified in the Transcriptome and Post-Genomic studies as targets for immune interventions . Two secreted members of the family were obtained as recombinant proteins in the native conformation . Antibodies produced against them showed that SmVAL4 was present mostly in cercarial secretions and SmVAL26 in egg secretions and that only the native SmVAL4 contained carbohydrate moieties . Due to concerns with potential allergic characteristics of this class of molecules , we have explored the mouse model of airway inflammation in order to investigate these properties in a more confined system . Sensitization and challenge with rSmVAL4 , but not rSmVAL26 , induced extensive migration of cells to the lungs , mostly eosinophils and macrophages; moreover , immunological parameters were also characteristic of an allergic inflammatory response . Our results showed that the allergic potential of this class of proteins can be variable and that the vaccine candidates should be characterized; the mouse model of airway inflammation can be useful to evaluate these properties . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"vaccination",
"clinical",
"immunology",
"immunity",
"immunology",
"biology"
] | 2012 | Schistosoma mansoni Venom Allergen Like Proteins Present Differential Allergic Responses in a Murine Model of Airway Inflammation |
Diverse Polycomb repressive complexes 1 ( PRC1 ) play essential roles in gene regulation , differentiation and development . Six major groups of PRC1 complexes that differ in their subunit composition have been identified in mammals . How the different PRC1 complexes are recruited to specific genomic sites is poorly understood . The Polycomb Ring finger protein PCGF6 , the transcription factors MGA and E2F6 , and the histone-binding protein L3MBTL2 are specific components of the non-canonical PRC1 . 6 complex . In this study , we have investigated their role in genomic targeting of PRC1 . 6 . ChIP-seq analysis revealed colocalization of MGA , L3MBTL2 , E2F6 and PCGF6 genome-wide . Ablation of MGA in a human cell line by CRISPR/Cas resulted in complete loss of PRC1 . 6 binding . Rescue experiments revealed that MGA recruits PRC1 . 6 to specific loci both by DNA binding-dependent and by DNA binding-independent mechanisms . Depletion of L3MBTL2 and E2F6 but not of PCGF6 resulted in differential , locus-specific loss of PRC1 . 6 binding illustrating that different subunits mediate PRC1 . 6 loading to distinct sets of promoters . Mga , L3mbtl2 and Pcgf6 colocalize also in mouse embryonic stem cells , where PRC1 . 6 has been linked to repression of germ cell-related genes . Our findings unveil strikingly different genomic recruitment mechanisms of the non-canonical PRC1 . 6 complex , which specify its cell type- and context-specific regulatory functions .
Polycomb group ( PcG ) protein complexes play crucial roles in many physiological processes , including stem cell maintenance , differentiation , cell cycle control and cancer [1–4] . PcG complexes repress transcription through various mechanisms including changes in histone modification , polynucleosome compaction and direct interaction with the transcription machinery [1 , 5] . Two major complexes exist in mammals , the Polycomb repressive complexes 1 and 2 ( PRC1 and PRC2 ) , which differ in their enzymatic activity . PRC1 contains the E3 ligase RING1/2 , which catalyzes ubiquitination of histone H2A at lysine 119 ( H2AK119ub1 ) , while PRC2 contains the methyltransferase EZH2 ( Enhancer of Zeste Homolog 2 ) that catalyzes tri-methylation of histone H3 ( H3K27me3 ) . It has long been considered that H3K27me3 is required for PRC1 binding to chromatin . However , this view was challenged when it was found that a number of PRC1 complexes exist , which lack H3K27me3-binding CBX ( Chromo Box ) subunits [6–9] . Six major PRC1 complexes have been described and each contains a defining PCGF ( Polycomb Group Ring Finger ) subunit ( PCGF1-6 ) , the RING1/2 E3 ubiquitin ligase , RYBP/YAF2 ( RING1 and YY1 binding protein/YY1 Associated Factor 2 ) or a CBX protein , and a unique set of associated proteins [7 , 8] . The canonical PCR1 complexes are PRC1 . 2 , which contains PCGF2 ( MEL-18 ) , and PRC1 . 4 , which contains PCGF4 ( BMI1 ) . They are recruited to chromatin by the H3K27me3 mark deposited by PRC2 . By contrast , the non-canonical ( ncPCR1s ) PRC1 . 1 , PRC1 . 3 , PRC1 . 5 and PRC1 . 6 are targeted to chromatin by H3K27me3-independent mechanisms . Significantly , ncPRC1s are responsible for H2A ubiquitination ( H2AK119ub1 ) , which leads to recruitment of PRC2 and downstream H3K27me3 deposition [10] . Our knowledge of the targeting of ncPRC1 complexes to their genomic sites is limited . The PCGF1-containing PRC1 . 1 variant is recruited to non-methylated CpG islands via the histone methyltransferases Kdm2b ( Lysine ( K ) -specific demethylase 2b ) , which binds to non-methylated CpG islands [6 , 9] . Recruitment of the PCGF3/5-containing ncPRC1s to the inactive X-chromosome is mediated by the Xist-RNA [11] . The subunit composition of the different ncPCR1s is specific and potentially revealing . While PRC1 . 6 ( also known as E2F6-PRC1 and PCGF6-PRC1 ) is similar if not identical to L3MBTL2 ( Lethal ( 3 ) Malignant Brain Tumor-Like 2 ) -containing complexes [12 , 13] and the E2F6 repression complex [14] , it is specifically associated with several proteins that are not found in other ncPRC1s ( Fig 1A ) , [7 , 15] . MGA ( MAX Gene-Associated protein , also abbreviated as MGAP by UniProt ) contains two DNA-binding domains , a T-box domain and a bHLH ( basic helix-loop-helix ) domain . MGA interacts with MAX ( Myc-associated Factor X ) , and E2F6 interacts with DP-1 or DP-2 ( transcription factor DP-1 or DP-2 ) . Heterodimeric MGA/MAX binds E-boxes , and heterodimeric E2F6/DP-1/2 binds to E2F recognition sequences in vitro [16–18] . L3MBTL2 contains four MBT domains that bind to mono- and di-methylated histone H3 and H4 tails in vitro [19–21] . Full-length L3MBTL2 can also interact with histones independent of their lysine methylation state [12 , 20] . The association of PRC1 . 6 with the sequence-specific DNA binding proteins MGA/MAX and E2F6/DP1 , and with the histone-interacting protein L3MBTL2 suggests that these proteins could play a role in locus-specific recruitment of PRC1 . 6 . Crucially , this notion has not been addressed experimentally . In mouse embryonic stem cell ( ESC ) an essential role for the corresponding PRC1 . 6 subunits in specification and proliferation has been demonstrated . Mga and Pcgf6 were identified as essential self-renewal genes in ESCs by a genome-wide RNAi screen [22] . A more recent knockout study revealed that Mga is essential for survival of mouse pluripotent cells during peri-implantation development and for growth of ESC cultures [23] . L3mbtl2-deficient ESCs retain characteristics of pluripotent cells but are severely impaired in proliferation [13] . Finally , the defining subunit of PRC1 . 6 , Pcgf6 , is expressed at high levels in mouse ES cells , where it is required for ESC identity [24 , 25] . The mechanism by which this occurs remains controversial . Two reports suggested a repressive function of Pcgf6 on mesodermal-specific [24] and on endodermal lineage genes [26] , while Yang et al . suggested an PRC1 . 6-independent direct activator function of Pcgf6 on core ESC regulators such as Oct4 , Sox2 and Nanog [25] . Here we describe the targeting mechanism of PRC1 . 6 , an exemplar of the non-canonical PRC1 class , by detailing the role of MGA , L3MBTL2 , E2F6 and PCGF6 in genomic binding site selection . We show that MGA , L3MBTL2 , E2F6 and PCGF6 colocalize genome-wide in the context of PRC1 . 6 . Taking advantage of CRISPR/Cas-mediated genetic ablation in HEK293 cells , we demonstrate that MGA is absolutely essential for binding of PRC1 . 6 . By expression of MGA mutants in MGAko cells , we found that the bona fide T-box and bHLH DNA-binding domains of MGA mediate binding to a subset of loci but are dispensable for others . We further demonstrate that L2MBTL2 and E2F6 determine differential binding of PRC1 . 6 to distinct promoters . Finally , we demonstrate that Mga , L3mbtl2 and Pcgf6 colocalize also in mouse ESCs . In particular , we found enrichment at promoters of meiosis-and germ-line-specific genes that were shown to be de-repressed on Max- , L3mbtl2- or Pcgf6-depletion . Together , our findings unveil strikingly different genomic recruitment mechanisms for a non-canonical Polycomb repressive complex , which specify its cell type- and context-specific regulatory functions .
To identify the genomic binding sites of PRC1 . 6 and to gain mechanistic insights into its targeting , we focused on the roles of MGA , L3MBTL2 , E2F6 and PCGF6 as these factors are specific to PRC1 . 6 ( Fig 1A ) and were not found in other ncPRC1s . We established HEK293 cell clones in which each of these four proteins was depleted individually using the CRISPR/Cas9-sgRNA system ( S1 Fig ) as controls at key steps in the analysis . By Western blotting we confirmed successful depletion of MGA , L3MBTL2 , E2F6 and PCGF6 in several clones ( Fig 1B ) . Next , we determined MGA , L3MBTL2 , E2F6 and PCGF6 occupancy by ChIP-seq using chromatin of the corresponding knockout cell lines as a reference for peak selection . Thereby , we were able to remove a number of perfectly shaped false positive ChIP-seq signals ( S2A Fig ) from the classified lists of binding sites . We obtained different peak strengths and different numbers of peaks for the different factors ( L3MBTL2 > E2F6 > MGA > PCGF6 ) , possibly due to the different performance of the antibodies resulting in different ChIP efficiencies . Stringent filtering of uniquely mapped reads ( ≥30 tags and ≥3-fold enrichment over the corresponding knockout control ) yielded lists of high-confidence binding sites for each factor . Comparison of the MGA , L3MBTL2 , E2F6 and PCGF6 data sets revealed a very high degree of overlap ( Fig 1C and 1D ) reflecting colocalization ( Fig 1E ) . Consistent with the role of PRC1 in regulating gene expression , the large majority of these sites were located close to the 5´-end of annotated transcripts ( Fig 1F ) . We also confirmed colocalization of MGA , L3MBTL2 , E2F6 and PCGF6 to a set of selected target promoters by conventional ChIP-qPCR analysis ( Fig 1G ) . The overlap of the MGA , L3MBTL2 , E2F6 and PCGF6 ChIP-seq peaks shown in Fig 1C also suggests the existence of some genomic sites bound by only one of the four factors . However , the majority of the potential factor-specific sites was removed when we compared filtered peaks with unfiltered MACS peaks ( S2B–S2E Fig ) . Moreover , visual genome browser inspection of the remaining potential subunit-specific peaks indicated the shared presence of MGA , E2F6 , L3MBTL2 and PCGF6 at all examined sites ( S2B–S2E Fig ) . Hence , our ChIP-seq results indicate that all four factors bind to the same genomic loci in vivo . This conclusion is strongly supported by the complete absence of genomic L3MBTL2 , E2F6 and PCGF6 binding events in MGA-depleted cells ( see below ) . A de novo sequence motif analysis of the top 600 ranked MGA , L3MBTL2 , E2F6 and PCGF6 binding sites revealed centrally enriched motifs that match in vitro recognition sequences for MGA/MAX ( the E-Box , CACGTG ) [17] and for E2F6/DP1 ( GCGGGAA ) [18] ( Fig 1H ) . The abundant occurrence of the E-box and the E2F6 binding motif indicated that both , MGA and E2F6 , could be important for recruitment of PRC1 . 6 to its specific sites in chromatin . MGA and E2F6 are sequence-specific DNA binding factors; and L3MBTL2 is a histone-interacting protein . Having found that they colocalize genome-wide , we set out to investigate their interdependence in genomic targeting of PRC1 . 6 . At first we focused on the role of MGA and examined whether binding of other PRC1 . 6 subunits was affected in MGA-depleted cells . ChIP-seq analysis revealed that MGAko cells lack genome-wide binding of both L3MBTL2 and E2F6 ( Fig 2A and 2B ) indicating that MGA is crucial for genomic targeting of L3MBTL2 and E2F6 and potentially for the entire PRC1 . 6 complex . This finding was particularly unexpected since the E2F6/DP2 heterodimer binds E-box motifs readily in vitro [18] . Western blot analysis revealed that MGA-depleted cells contained markedly less E2F6 as well as PCGF6 ( Fig 2C ) . The reduced protein levels of E2F6 and PCGF6 in MGAko cells were likely due to impaired protein stability , as the transcript levels of E2F6 and PCGF6 were not reduced in MGAko cells ( Fig 2D ) . The protein level of L3MBTL2 in MGAko cells was similar as in wild type cells . However , the fraction of SUMO-modified L3MBTL2 [20] was strongly reduced ( Fig 2C ) , which may indicate that SUMOylation of L3MBTL2 in wild type cells takes place at the level of chromatin . Finally , the level of RING2 protein was unchanged in MGA-deficient cells . To exclude that the lack of any E2F6 and L3MBTL2 binding in MGAko cells was the result of inefficient ChIPs , we also probed a panel of selected target promoters by ChIP-qPCR . These experiments validated the lack of genomic L3MBTL2 and E2F6 binding in two different MGAko clones ( Fig 2E ) . We also analyzed for the presence of other PRC1 . 6 components including PCGF6 , MAX , RING2 , RYBP , HP1γ and the H2AK119ub1 mark . All factors as well as the H2AK119ub1 mark were present at the MGA target sites in wild type cells but were absent or , in the case of the H2AK119ub1 mark , markedly reduced in both MGA-depleted cell clones ( Fig 2E ) . The global H2AK119ub1 levels were similar in wild type , MGAko , L3MBTL2ko , E2F6ko , and PCGF6ko cells ( S3 Fig ) showing that the observed reduction of the H2AK119ub1 mark at the PRC1 . 6 target regions is due to changes in local RING2 deposition . Collectively , these results demonstrate that MGA is absolutely crucial for genomic loading of the entire PRC1 . 6 complex . Importantly , these results also indicate that E2F6 , L3MBTL2 and PCGF6 bind to their genomic sites exclusively in the context of the PRC1 . 6 complex and are not recruited to chromatin independently of PRC1 . 6 . Since previous studies reported that H2AK119ub1 plays a critical role for recruitment of PRC2 followed by downstream deposition of H3K27me3 [10] , we also tested for the presence of the catalytic PRC2 component EZH2 and for H3K27me3 ( Fig 2F ) . Neither EZH2 nor H3K27me3 were enriched at the selected PRC1 . 6 loci suggesting that PRC1 . 6 binding is not generally interconnected with PRC2 binding . Importantly , we found considerable enrichment of EZH2 and H3K27me3 at known PRC2-dependent canonical PRC1 target sites . These canonical PRC1 binding sites were not bound by MGA , and the levels of EZH2 and H3K27me3 at these sites remained unchanged in MGAko cells ( Fig 2F ) . The absence of MGA at canonical PRC1 binding regions is consistent with genome-wide data that revealed only a low level of overlap between PCGF6 and other PCGFs in HEK293 cells [7] . Given that MGA is essential for targeting of PRC1 . 6 , it would be expected that re-expression of MGA would restore not only genomic binding of MGA but also of the other PRC1 . 6 components . To test this prediction , we expressed full-length MGA in MGAko cells , and subsequently analyzed a panel of PRC1 . 6 target promoters for binding of exogenous MGA and of endogenous L3MBTL2 , E2F6 , PCGF6 , MAX and RING2 . Indeed , re-expression of MGA in MGAko cells not only restored specific binding of MGA but also of the other PRC1 . 6 subunits ( Fig 3A ) . We did not observe an increase of H2AK119ub1 levels at these promoters . Potentially , the short time span of transient MGA expression was not sufficient for the H2AK119ub1 mark to be deposited efficiently . MGA contains two different DNA binding domains , a T-box domain close to the N-terminus and a bHLH domain in its C-terminal part ( Fig 3B ) . To test whether these DNA binding domains account for genomic loading of PRC1 . 6 , we generated two different types of DNA binding-deficient MGA mutants by deleting the entire T-box domain ( MGA-ΔT , aa 79 –aa 264 deleted ) , and by replacing several critical amino acids in the bHLH domain [27 , 28] by alanine residues ( MGA-bHLHmut , MGA-H2477A_ER2481/2482AA_R2485A ) . Compared with wild type MGA , binding of the MGA-bHLH mutant to several target promoters ( AEBP2 , ZFR , CDIP , CCND2 and TFAP4 ) was strongly reduced ( Fig 3D ) . We also observed reduced binding of the MGA-T-box deletion mutant to the SPOP promoter . However , both MGA mutants still bound to the RFC1 , PHF20 , RPA2 , RNF130 and CDC7 promoters as efficiently as wild type MGA ( Fig 3D ) . We also tested binding of an MGA double mutant in which both DNA binding domains were mutated simultaneously ( MGA-ΔT-bHLHmut ) . Remarkably , the MGA-ΔT-bHLHmut double mutant still bound to these promoters as efficiently as wild type MGA . Importantly , the DNA binding-deficient MGA mutants also rescued binding of endogenous L3MBTL2 to the RFC1 , PHF20 , RPA2 , RNF130 and CDC7 promoters but not to the MGA-bHLH-dependent AEBP2 , ZFR , CDIP , CCND2 and TFAP4 promoters and to the MGA-T-Box-dependent SPOP promoter ( Fig 3D , right panel ) . These results suggest that MGA can recruit PRC1 . 6 to specific target sites by DNA-binding-dependent and by DNA-binding-independent mechanisms . As MGA is able to bind a subset of PRC1 . 6 loci independent of its DNA-binding activity , we investigated the potential contribution of E2F6 , L3MBTL2 or PCGF6 to the recruitment of PRC1 . 6 to its target sites . To address this issue , we profiled binding of MGA , L3MBTL2 and E2F6 in cells lacking L3MBTL2 , E2F6 or PCGF6 ( L3MBTL2ko , E2F6ko or PCGF6ko cells ) . Importantly , the level of MGA and of other PRC1 . 6 subunits in L3MBTL2ko- , E2F6ko- , and PCGF6ko cells was unaffected ( Fig 4A and S4 Fig ) . Analysis of the ChIP-seq data sets revealed that the overall genomic positions of the PRC1 . 6 binding sites in E2F6ko- , L3MBTL2ko- and PCGF6ko cells are similar to those in wild type cells ( Fig 4B and 4C ) . However , the signal strengths of the MGA and L3MBTL2 peaks in E2F6ko cells and the signal strength of the MGA and E2F6 peaks in L3MBTL2ko cells were significantly reduced , but only slightly affected in PCGF6ko cells ( Fig 4C and 4D ) . Notably the extent of reduction of MGA binding in E2F6ko cells correlated well with the extent of reduction of L3MBTL2 binding in E2F6ko cells ( Fig 4E , left panel ) . Equally , the extent of reduction of MGA binding in L3MBTL2ko cells correlated well with the extent of reduction of E2F6 binding in L3MBTL2ko cells ( Fig 4E , right panel ) . These results demonstrate that the genomic localization of PCR1 . 6 requires the simultaneous association of MGA , L3MBTL2 , E2F6 and PCGF6 in a single complex . Binding of MGA to the majority of its genomic sites was greatly reduced in L3MBTL2ko as well as in E2F6ko cells , indicating that both , L3MBTL2 and E2F6 contributed to genomic binding of PRC1 . 6 . Importantly , however , the extent of reduction of MGA and E2F6 binding in L3MBTL2ko cells , and the extent of reduction of MGA and L3MBTL2 binding in E2F6ko cells did not correlate ( Fig 5A ) . Rather , the shape of these plots revealed three distinct types of PRC1 . 6 binding site ( i ) loci where binding of MGA was reduced in both , L3MBTL2ko and E2F6ko cells ( ii ) loci where binding of MGA was reduced in L3MBTL2ko cells but not in E2F6ko cells; ( iii ) loci where binding of MGA was reduced in E2F6ko cells but not in L3MBTL2ko cells . Thus , we were able to identify L3MBTL2-dependent and E2F6-dependent PRC1 . 6 binding sites ( Fig 5B and S5A Fig ) . We also probed a panel of PRC1 . 6 target sites in two different L3MBTL2ko and E2F6ko cell clones by conventional ChIP-qPCR . We tested for the presence of MGA , L3MBTL2 , E2F6 and PCGF6 , MAX , RING2 and H2AK119ub1 ( Fig 5C and S5B Fig ) . This analysis confirmed L3MBTL2- and E2F6-dependent binding of PRC1 . 6 to the RFC1 , PHF20 and SPOP promoters; L3MBTL2-dependent but E2F6-independent binding to the AEBP2 and ZFR promoters; and E2F6-dependent but L3MBTL2-independent binding to the ALDOA , RNF130 and CDC7 promoters ( S5B Fig ) . In all cases the levels of H2AK119ub1 correlated with PRC1 . 6 binding . Finally , we performed rescue experiments in which we found that expression of L3MBTL2 in L3MBTL2ko cells not only restored binding of ectopically expressed L3MBTL2 but also of endogenous MGA , E2F6 and PCGF6 to the L3MBTL2-dependent RFC1 , PHF20 , SPOP , AEBP2 and ZFR promoters ( Fig 5D ) . Of note , the enrichment levels observed in rescued L3MBTL2ko cells were approximately 1 . 5- to 3-fold lower than in wild type cells ( compare Fig 5D with in Fig 1G ) . This is not surprising given that not all cells in the population express L3MBTL2 after transient transfection . Importantly , L3MBTL2 also re-occupied the E2F6-dependent ALDOA , RNF130 and CDC7 promoters; however , binding of MGA , E2F6 and PCGF6 to these promoters was not enhanced . This result strongly supports our model in which L3MBTL2 is only essential for recruitment of PRC1 . 6 to a subset of loci despite its presence at all PRC1 . 6 binding sites . To gain further insight into the E2F6-dependent recruitment of PRC1 . 6 , we examined whether the DNA-binding activity of E2F6 is necessary for PRC1 . 6 binding . Wild type E2F6 expressed in E2F6ko cells re-occupied all tested PRC1 . 6 target loci , and also resulted in slightly increased binding of endogenous MGA and L3MBTL2 to E2F6-dependent promoters but not to the L3MBTL2-dependent promoters ( Fig 5E ) . The DNA binding-deficient E2F6 mutant ( E2F6-L68E , V69F ) did not bind to the E2F6-dependent promoters , and did not re-occupy the L3MBTL2-dependent promoters . This observation indicates that the DNA binding domain of E2F6 is not only necessary for DNA recognition but also for association with PRC1 . 6 . By ChIP-qPCR analysis of selected promoters we also validated that PCGF6 has a limited role in the recruitment of MGA , L3MBTL2 and E2F6 ( S6 Fig ) . However , it is important to note that binding of RING2 in PCGF6ko cells was nearly reduced to levels at a negative control region . Consistently , H2AK119ub1 levels were also reduced at these promoters ( S6 Fig ) . Thus , PCGF6 , albeit not essential for binding site selection by PRC1 . 6 , it is required to recruit RING2 to these loci . This observation is in line with a recent study that revealed recruitment of RING2 by a PCGF6-TET repressor fusion protein tethered to a Tet operator array in vivo [29] . We surveyed the DNA sequences of L3MBTL2- and the E2F6-dependent PRC1 . 6 loci and found specific enrichment of the E2F binding motif ( GCGGGA ) in the E2F6-dependent PRC1 . 6 binding sites , and specific enrichment of the E-box ( CACGTG ) and T-box ( AGGC/TGC/TGAGG ) binding motifs in the L3MBTL2-dependent PRC1 . 6 binding sites ( Fig 5F ) . The strong association of E-box and T-box motifs with L3MBTL2-dependent PRC1 . 6 binding sites point to an important role of L3MBTL2 in facilitating or stabilizing an interaction of MGA/MAX with DNA . Finally , we examined whether there are specific functional features shared amongst E2F6-dependent and L3MBTL2-dependent PRC1 . 6-bound genes . E2F6-dependent PRC1 . 6 target genes but not L3MBTL2-dependent PRC1 . 6 target genes were highly enriched in Gene Ontology ( GO ) terms related to cell cycle control ( Fig 5G ) . This finding is in line with the role of E2F6 as an RB-independent transcriptional repressor during cell cycle progression [30] . GO terms associated with L3MBTL2-dependent PRC1 . 6 target genes included quite different biological processes such as “positive regulation of neurotransmitter secretion” , meiotic “synaptonemal complex assembly” and “ribosome assembly” . Altogether these results suggest that E2F6 and L3MBTL2 recruit PRC1 . 6 to distinct gene sets that regulate different biological processes . We went on to investigate the role of PRC1 . 6 in cell growth and gene expression in HEK293 cells by comparing the proliferation potential of wild type , MGAko , L3MBTL2ko , E2F6ko and PCGF6ko cells . The growth rates of wild type cells and PCGF6ko cells were similar; however , we observed reduced proliferation of MGAko , L3MBTL2ko and E2F6ko cells . ( Fig 6A ) . Next , we examined the transcriptional impact of PRC1 . 6 . RNA-seq of three independent wild type cell cultures and three independent MGAko clones identified 587 genes with ≥2-fold altered expression levels in MGAko cells . Expression of 485 genes was reduced in the MGAko cells , while expression of 102 genes was increased ( Fig 6B ) . Comparison of the set of de-regulated genes with the gene set bound by MGA revealed that MGA was not bound to the majority of the down-regulated genes ( 434/485 , 89% ) suggesting an indirect role of MGA in the regulation of these genes . In contrast , MGA was bound to the majority of the up-regulated genes ( 71/102 , 70% ) suggesting that PRC1 . 6 acts as a direct repressor on these genes . Representative ChIP-seq and RNA-seq genome browser screenshots of de-repressed genes bound by PRC1 . 6 are shown in Fig 6C . Interestingly , the top up-regulated genes in MGAko cells included several critical regulators and effectors of meiosis such as CNTD1 , SMC1B , SYCE2 , YBX2 , MEIOC ( C17orf104 ) , RAD9B , TAF7L , STAG3 , CPEB1 and ALDH1A2 , as well as several testis-enriched genes such as PRSS50 , TRIM71 , C19orf57 , ZCWPW1 , ZNF239 , RIBC2 , NEUROG2 ( http://www . proteinatlas . org/ ) . To test whether L3MBTL2 , E2F6 and PCGF6 contribute to PRC1 . 6 target gene repression , we examined the expression of a panel of fourteen genes in L3MBTL2ko , E2F6ko and PCGF6ko cells by locus-specific RT-PCR assays . We found increased transcript levels of these genes also in L3MBTL2ko , E2F6ko or PCGF6ko cells , yet to different degrees ( Fig 6D ) . For example , compared with wild type cells , transcript levels of CNTD1 were also increased in L3MBTL2ko and in E2F6ko but not in PCGF6ko cells . Transcript levels of SMC1B , however , were increased in E2F6ko cells but not in L3MBTL2ko and PCGF6ko cells . Conversely , transcript levels of STAG3 were increased in L3MBTL2ko and PCGF6ko cells but not in E2F6ko cells ( Fig 6D ) . Thus , L3MBTL2 , E2F6 and PCGF6 contributed to repression of these genes differentially in a gene-specific manner . Interestingly , specific de-repression of these genes in L3MBTL2ko or E2F6ko cells , respectively , correlated well with the contribution of L3MBTL2 and E2F6 to PRC1 . 6 binding . Binding of PRC1 . 6 to the CNTD1 promoter was diminished in L3MBTL2ko as well as in E2F6ko cells . Binding of PRC1 . 6 to the SMC1B promoter was lost in E2F6ko cells but remained in L3MBTL2ko cells , and binding of PRC1 . 6 to the STAG3 promoter was lost in L3MBTL2ko cells but remained in E2F6ko cells ( S7 Fig ) . Several recent studies revealed critical roles for Mga , L3mbtl2 and Pcgf6 in the regulation of mouse ES cell pluripotency , proliferation and differentiation [13 , 23 , 24 , 26 , 29] . Therefore , we investigated the genomic localization of PRC1 . 6 components also in mouse ES cells . We focused on Mga , L3mbtl2 and Pcgf6 as there is no available antibody , which efficiently recognizes murine E2f6 . Also of note is that we failed to generate Mga-deficient ESC clones , which is in line with a previous study suggesting that Mga plays an essential role in ESCs [23] . Therefore , we used an IgG control ChIP-seq dataset as a reference for peak selection . Similar to the ChIP-seq results with chromatin of HEK293 cells , we obtained different numbers of filtered ( ≥30 tags and ≥3-fold enrichment over IgG ) peaks for Mga ( 14 . 183 ) , L3mbtl2 ( 17 . 007 ) and Pcgf6 ( 4817 ) ( Fig 7A ) . The vast majority of the Pcgf6 peaks ( 90% ) overlapped with Mga and L3mbtl2 peaks; and the majority of the Mga peaks ( 82% ) overlapped with the L3mbtl2 peaks . Genome browser track and heatmap views of binding densities also revealed clear colocalization of Mga , L3mbtl2 and Pcgf6 ( Fig 7B and 7C ) . Moreover , visual inspection of genome browser tracks did not confirm any Mga- , L3mbtl2- or Pcgf6-specific binding site ( S8A Fig ) . Collectively , our ChIP-seq data sets reveal that Mga , L3mbtl2 and Pcgf6 colocalize in mouse ESCs suggesting strongly that the function of PRC1 . 6 is conserved in murine and human cells . This conclusion was further supported by a de novo sequence motif analysis of the top 600 ranked Mga/L3mbtl2/Pcgf6 peak regions , which revealed the presence of centrally enriched E-box as well as T-box and E2F6/DP1 recognition sequences as prevalent motifs ( Fig 7D ) . Finally , as in HEK293 cells the majority of the Mga/L3mbtl2/Pcgf6 binding sites were located close to transcriptional start sites ( Fig 7E ) . However we observed that the genomic distribution of the Mga/L3mbtl2/Pcgf6 peaks in mouse ESCs differ to some extent from the distribution in HEK293 cells . In many instances , we observed multiple Mga/L3mbtl2/Pcgf6 peaks within a gene locus including the promoter , exons , and the 3´-end ( S8B Fig ) . Potentially , the peaks within gene bodies were not direct PRC1 . 6 binding sites but reflect local intragenic loops within genes that fold exons close to cognate promoters . The capture of such structural features by ChIP-seq has been reported previously [31] . It is also possible that these intragenic peaks reflect discrete compacted chromatin structures similar to those generated by canonical cPRC1 [32] . To examine the impact of the Mga/L3mbtl2/Pcgf6 binding sites on gene expression in mouse ESCs , we compared our ChIP-seq data sets with genes that were deregulated in L3mbtl2-depleted [13] or in Pcgf6-depleted ESCs [26] . We found that two-third of the genes ( 882 out of 1354 ) that were up-regulated in Pcgf6-depleted ES cells , and 71% of the genes that were up-regulated in L3mbtl2-depleted cells ( 421 out of 587 ) by ≥2-fold were bound by Mga , L3mbtl2 and Pcgf6 ( Fig 7F ) . Interestingly , genes aberrantly expressed in Pcgf6ko and in L3mbtl2ko ESCs largely do not overlap ( Fig 7F ) . Nevertheless , a GO analysis revealed that Pcgf6-dependent , as well as L3mbtl2-dependent repressed PRC1 . 6 target genes were strongly associated with germ-line development ( meiosis and spermatogenesis ) . Also the small group of 78 direct PRC1 . 6 target genes that were up-regulated in Pcgf6ko as well as in L3mbtl2ko ES cells ( Fig 7F ) encode several meiotic genes including Syce3 , Stk31 , Slc22a , Mei1 and Tdrkh . Remarkably , promoters of the meiotic genes were within the top ranked 200 Mga , L3mbtl2 and Pcgf6 peaks . Genes specifically de-repressed in Pcgf6-depleted ES cells but not in L3mbtl2-depleted ES cells were related to the wnt signaling pathway and to neuron differentiation . Conversely , specific L3mbtl2-repressed genes were associated with angiogenesis and positive regulation of cell migration ( Fig 7F ) . This finding suggests that Pcgf6 and L3mbtl2 repress common as well as different sets of genes despite the presence of both factors at all target genes within the PRC1 . 6 complex . Mouse ESCs also express other polycomb complexes including canonical cPRC1 , the non-canonical PRC1 . 1 ( PRC1-Kdm2b , PRC1-Fbxl10 ) complex and PRC2 . To determine whether these complexes also bind to PRC1 . 6 target genes we compared our Mga/L3mbtl2/Pcgf6 data sets with published ChIP-seq data sets of Ring1b , Rybp , Cbx6 , Pcgf2 ( Mel18 ) , Cbx7 , Suz12 , H3K27me3 and Kdm2b ( Fbxl10 ) . Ring1b is a constituent of all PRC1 complexes . Cbx6 is associated with cPRC1 as well as with PRC1 . 6 [33] . Rybp is found in canonical as well as in non-canonical complexes; however the presence of Rybp or Cbx7 in PRC1 complexes is mutually exclusive [8 , 34] . Pcgf2 and Cbx7 are subunits of the cPRC1 complex and require the H3K27me3 mark to localize to chromatin [34 , 35] . Suz12 is a subunit of the PRC2 complex , which deposits the H3K27me3 mark . Kdm2b together with Pcgf1 forms the non-canonical variant PRC1 . 1 [6 , 9] . A heatmap view of binding densities shows tight colocalization of Rybp and Mga-L3mbtl2-Pcgf6 ( Fig 8A ) . This is consistent with the presence of Rybp in the PRC1 . 6 complex [13 , 26] . Also the Cbx6 ChIP-seq data set displayed colocalization with Mga-L3mbtl2-Pcgf6 binding sites despite its weak density . This finding is consistent with a recent report that revealed an interaction of Cbx6 with Pcgf6 and L3mbtl2 [33] . A considerable number of PRC1 . 6 target regions was also occupied by canonical PRC1 ( Pcgf2 and Cbx7 ) , PRC1 . 1 ( Kdm2b ) and PRC2 ( Suz12 ) , and was decorated with H3K27me3 ( Fig 8A ) . Unlike Rybp , however , Pcgf2 , Cbx7 , Kdm2b , Suz12 and H3K27me3 displayed a broader distribution at the Pcgf6 target regions . We assessed that approximately 30% , 50% and 90% of the high confidence Pcgf6 target genes were co-occupied by Cbx7 ( cPRC1 ) , Suz12/H3K27me3 ( PRC2 ) and Kdm2b ( PRC1 . 1 ) , respectively ( Fig 8B ) . Interestingly , meiosis-related genes that were de-repressed in Pcgf6- and L3mbtl2-depleted cells were largely bound exclusively by PRC1 . 6 , whereas “typical” cPRC1 target genes such as Nkx-2 or Hoxa7 were decorated with both PRC1 . 6 and cPRC1 as well as with PRC2 ( Fig 8C and 8D ) . Together , these results suggest that PRC1 . 6 and cPRC1 have both unique and common target genes in mouse ESCs .
In this study , we provide insights into the genomic targeting mechanism of the non-canonical PRC1 complex PRC1 . 6 . We find that MGA , L3MBTL2 , E2F6 and PCGF6 colocalize genome-wide in the context of PRC1 . 6 . MGA is absolutely crucial for binding of the complete PRC1 . 6 since genome-wide binding of E2F6 , L3MBTL2 and PCGF6 is lost in MGA-depleted cells ( Fig 2 ) . Mechanistically , we provide strong evidence that MGA executes recruitment of PRC1 . 6 to its target sites through two distinct functions ( Fig 9 ) . On the one hand , MGA acts as a sequence-specific DNA-binding factor mediating recruitment of PRC1 . 6 to E-box and T-box containing promoters . On the other hand , MGA has a scaffolding function , which is independent of its DNA binding capacity ( Fig 3B and 3C ) . The scaffolding function of MGA may protect E2F6 and PCGF6 against degradation ( Fig 2C ) . The other components of PRC1 . 6 have distinct functional roles . L3MBTL2 is also involved in genomic targeting of PRC1 . 6 since in L3MBTL2ko cells , MGA , E2F6 and PCGF6 fail to bind to a large fraction of promoters ( Figs 4 and 5 ) . These L3MBTL2-dependent PRC1 . 6 binding sites are enriched for the bHLH E-box motif but not for the E2F6-binding motif ( Fig 5 ) . The MBT domains of L3MBTL2 are known to bind preferentially mono- , and di-methylated histone H3 and H4 marks [19–21]; and full-length L3MBTL2 interacts with histone tails independent of their lysine methylation state [12 , 20] . Thus , we propose that L3MBTL2 promotes binding site selection of PRC1 . 6 by facilitating and stabilising the interaction of MGA/MAX with E- or T-box-containing promoters . PCGF6 has a minor role in genomic loading of PRC1 . 6 ( Fig 4 ) , but significantly it interacts with RING1B [36] , recruits it to genomic PRC1 . 6 sites and facilitates downstream H2AK119 ubiquitination and transcriptional repression [29] . Consistently , PRC1 . 6-bound promoters are enriched of H2AK119ub1 ( Fig 2E and S6 Fig ) . Importantly , these targets are not bound by PRC2 and are not enriched in H3K27me3 ( Fig 2F ) . The absence of PRC2 and H3K27me3 at PRC1 . 6 target promoters is consistent with a previous report that revealed lack of H3K27me3 at L3MBTL2-E2F6 binding sites in K562 cells [12] . Yet , H3K27me3 enrichment was found at a subgroup of Pcgf6 binding sites in mouse ESCs ( Fig 8 and [29] ) suggesting a potential cell type-specific role of Pcgf6-dependent recruitment of PRC2 and downstream H3K27me3 deposition . A large fraction of PRC1 . 6 binds to promoters that regulate mitotic cell cycle genes . Since binding is largely unaffected in L3MBTL2-depleted cells targeting of PRC1 . 6 to this class of genes is more likely mediated by E2F6 ( Fig 5F and 5G ) . This finding is consistent with a previous report showing that E2F6 functions as an RB-independent transcriptional repressor by controlling E2F1-3-dependent transcription during cell cycle progression , particularly by counteracting the activating E2Fs during S phase [30] . These cell cycle-regulated genes are not upregulated in MGA-depleted cells . Likely , E2F4 , another repressive E2F family member , compensates for the loss of E2F6-mediated PRC1 . 6 binding . Indeed , it has been shown that only simultaneous inhibition of both , E2F6 and E2F4 activity , results in depression of these PRC1 . 6 target genes [30] . Mga , L3mbtl2 and Pcgf6 colocalize also in mouse ESCs ( Fig 7 ) strongly suggesting that the core components of PRC1 . 6 are evolutionarily and functionally conserved . In addition , knockdown of Mga and Max in mouse ESCs leads to the loss of Pcgf6 binding at several promoters of the genes that are up-regulated in Pcgf6ko cells [29] indicating that also the recruitment mechanisms in ESCs are similar to those observed in human cells . Notably , Pcgf6 is the most highly expressed Pcgf paralog in undifferentiated ESCs [24] , and Pcgf6 is the predominant Ring1b-interactor in ESCs [37] . These observations indicate that PRC1 . 6 is a major PRC1 complex in ESCs . PRC1 . 6 components play essential roles in ESCs including regulation of ESC pluripotency , proliferation and differentiation . Most significantly , Mga depletion leads to the death of proliferating pluripotent ICM cells in vivo and in vitro , and the death of ESCs in vitro [23] . Also Pcgf6ko and L3mbtl2ko as well as Maxko ESCs have defects in proliferation and differentiation [13 , 26 , 38] but less severe as Mgako ESCs . The most severe phenotype of Mgako ESCs is in line with the crucial importance of Mga for genomic PRC1 . 6 binding . Consistent with published reports , we have found that in ESCs , PRC1 . 6 is involved in the repression of meiotic genes . Ablation of Max , the dimerization partner of Mga , activates meiotic genes in ESCs and induces cytological changes , which are reminiscent of germ cells at the leptotene and zygotene stages of meiosis [39 , 40] . Meiotic and germ-line-specific genes are also activated in Pcgf6ko and L3mbtl2ko cells [13 , 26] . Mga , L3mbtl2 and Pcgf6 bind to the promoters of these meiosis-specific genes ( Fig 7 ) strongly suggesting that PRC1 . 6 directly represses these genes in ESCs thereby safeguarding/preventing meiosis . Interestingly , several meiotic genes are also de-repressed in MGA-deficient 293 cells ( Fig 6 ) . De-repression of a limited number of meiotic and germ cell-specific genes is also observed in E2F6-deficient MEFs indicating that the repressive function operates in somatic cells [41–43] . Knockdown of Pcgf6 results also in strongly increased expression levels of several mesodermal genes including T ( Brachyury ) , the Runx transcription factor Mlf1 and the vascular endothelial growth factor receptor 2 ( Vegfr-2 , Flk ) encoded by the Kdr gene [24] . Our ChIP-seq data revealed binding of Mga , L3mbtl2 and Pcgf6 to these genes suggesting that PRC1 . 6 also directly represses mesodermal lineage genes in mouse ESCs . Apart from meiosis-specific and germ-line-specific genes , quite different gene sets are de-repressed on Pcgf6- and L3mbtl2-depletion in ESCs . Based on this observation it was suggested that Pcgf6 acts independently of L3mbtl2 [24] . However we provide strong evidence that all Pcgf6 binding sites are also bound by L3mbtl2 . We speculate that L3mbtl2 facilitates binding of PRC1 . 6 to specific loci , while Pcgf6 acts through recruitment of Ring1b and downstream H2AK119ub1 [29] . Since L3mbtl2 associates with the methyltransferases G9A and GLP [13 , 14] it may also facilitate H3K9 dimethylation . Indeed , G9A and GLP are required for repression of germ cell-specific genes [44] . It is also possible that L3mbtl2 , known to compact nucleosomal arrays in vitro [12] , represses transcription directly by chromatin compaction making promoters inaccessible for the transcription machinery .
Rabbit polyclonal antibodies against MGA for use in ChIP experiments and immunoblotting were generated by immunizing with a bacterially expressed GST fusion protein carrying the 300 C-terminal amino acids of human MGA . Immunization was carried out by Eurogentec ( Seraing , Belgium ) using the 28-day Speedy immunization protocol . Antisera were affinity-purified according to a protocol described in [45] using the matrix-coupled GST-MGA fusion protein . The commercially available antibodies used in this study are shown in Table 1 . HEK293 cells were transfected using FugeneHD ( Promega , Madison , WI ) with plasmids expressing mammalian-codon optimized Cas9 and sgRNAs targeting the coding region of human MGA , E2F6 , L3MBTL2 or PCGF6 ( S1 Fig ) . The parental vector pSpCas9-2A-Puro ( pX459 ) was a gift from Feng Zhang ( Addgene plasmid # 48139 ) [46] . The sequences of the oligonucleotides used for targeting MGA , L3MBTL2 , E2F6 and PCGF6 were as follows . MGA-gRNA6: CATCTGGAAAGGTACTCCCA , MGA-gRNA7: GTCATACTTGAATTGTATAC; L3MBTL2-gRNA6: GGATGTGATGAAAGGGATGA , L3MBTL2-gRNA7: GCCTCTGTCATCCAGACAGC; E2F6-gRNA1: GGGTATTCTTGACTTAAACA , E2F6-gRNA2: GTTTAAGTCAAGAATACCCC , E2F6-gRNA3: GTCGATTCCATCTAAGACAT; PCGF6-gRNA3: GGTATGAAGACATTCTGTGA , PCGF6-gRNA4: TGTACTACTATATTGCATTT . The empty pX459 vector was transfected as a negative control . Puromycin selection ( 3 μg/ml ) was carried out 48 hours after transfection for 3 to 6 days . Individual colonies were isolated and the targeted loci were genotyped by PCR ( see S1 Fig ) and sequenced . Cell clones with indels in the targeted locus were further analyzed by Western blotting . The expression vector for 3xFLAG-L3MBTL2 has been described in [20]; and the expression vectors for HA-tagged wild type E2F6 and the DNA-binding-deficient E2F6 mutant in [18] . The HA-tag was removed by BamHI/HindIII digestion and blunt-end re-ligation . For expression of 3xFLAG-tagged MGA under control of the CMV promoter , several MGA cDNA fragments were amplified from poly ( A ) - and random-primed HEK293 cell cDNA libraries , and placed stepwise into pN3-3xFLAG using conventional restriction cloning procedures . The sequence of the cloned MGA cDNA is identical to the NCBI reference sequence XM_005254246 . 2 and encodes the 3115 amino acid full-length MGA isoform XP_005254303 . 1 . Mutations of the MGA T-box and bHLH domains were introduced into the wild type MGA construct by replacing appropriate wild type fragments with corresponding mutant gBlock DNA fragments ( IDT , Leuven , Belgium ) using internal restriction sites of the MGA cDNA . For expression of MGA , L3MBTL2 or E2F6 , the respective knockout clones were transiently transfected with the corresponding expression plasmid using the FugeneHD transfection reagent ( Promega ) . Five million cells on a 15-cm dish were transfected with 20 μg of plasmid DNA , harvested 48 hours after transfection and cross-linked chromatin was prepared . Expression of the proteins was monitored by Western blotting . HEK293 cells were cultured in DMEM/F-12 + GlutaMax medium ( Gibco , Thermo Fisher , Waltham , MA ) supplemented with 10% fetal bovine serum ( Sigma Aldrich , St . Louis , MO ) and 1% Penicillin/Streptomycin ( Sigma Aldrich ) . Mouse J1 ES cells were cultivated feeder-cell free on gelatin-coated plates in DMEM + GlutaMax ( Gibco , Thermo Fisher ) , supplemented with 15% fetal bovine serum ( Biochrom , Berlin , Germany ) , 1% non-essential amino acids ( Gibco , Thermo Fisher ) , 1% Penicillin/Streptomycin ( Sigma Aldrich ) , 50 mM ß-Mercaptoethanol and 1000 U/mL ESGRO leukemia inhibitory factor ( Merck Millipore , Billerica , MA ) . For determination of growth rates of wild type and corresponding knockout HEK293 cell lines , 3x105 cells were plated on a 6-well dish and counted in two or three days intervals as indicated in Fig 6A . Cumulative cell numbers were calculated by multiplying the initial cell number with the fold-increase in cell numbers in each interval . ChIP experiments were performed with the One Day ChIP kit ( Diagenode , Seraing , Belgium ) . ChIP-qPCRs with gene-specific primers ( Table 2 ) were performed using the ImmoMix PCR reagent ( Bioline , Luckenwalde , Germany ) in the presence of 0 . 1 x SYBRGreen ( Molecular Probes , Thermo Fisher , Waltham , MA ) . Enrichment was calculated relative to input . Three to four individual ChIPs were pooled and purified on QIAquick columns ( Qiagen , Hilden , Germany ) . Five nanograms of precipitated DNA were used for indexed sequencing library preparation using the Microplex library preparation kit v2 ( Diagenode ) . Libraries were purified on AMPure magnetic beads ( Beckman Coulter , Brea , CA ) and quantified on a Bioanalyzer ( Agilent Technologies , Santa Clara , CA ) . Pooled libraries were sequenced on an Illumina HiSeq1500 platform ( Illumina Inc . , San Diego , CA ) , rapid-run mode , single-read 50 bp ( HiSeq SR Rapid Cluster Kit v2 , HiSeq Rapid SBS Kit v2–50 cycles ) according to manufacturer´s instructions . Raw ChIP-seq reads were aligned using Subread [47] version 1 . 4 . 3-p1 . Reads matching multiple locations were discarded during alignment . Peaks were called with MACS [48] version1 . 4 . 0rc2 against the respective knockout control or against IgG for mouse ES cell data . Filtered peaks were required to have at least 30 tags and a sequencing depth-corrected ratio over control of 3x . Published mESC datasets ( Fig 8 ) were retrieved from GEO and processed as above using Subread and MACS , but were not filtered . Unions and overlaps were calculated on an ‘at least 1bp overlap’ basis . For motif search and heatmaps , peaks were centred at their summits and fixed sized regions extracted . Summits were defined as the point of highest read overlap after extending the reads to 200 bp . Heatmaps show number of reads extended to 200 bp , normalized for sequencing depth . The signal distribution was truncated at the 99th percentile in each sample in order to increase contrast . Regions for heatmaps were ordered by the sum of signal in the first sample depicted . ChIP-seq signal plots shown in Figs 1F and 7E are also based on reads extended to 200 bp . Genes were associated with a peak if the peak was located within -2 . 5 kb of TSS to TES . De novo motif search including Tomtom and CentriMo was performed online with MEME-ChIP versions 4 . 11 . 3 and 4 . 11 . 4 ( http://meme-suite . org/meme_4 . 11 . 4/tools/meme-chip ) [49] within the MEME Suite ( http://meme-suite . org ) [50] using 300 bp sequences surrounding peak summits ( +/- 150 bp ) . Gene Ontology ( GO ) analyses were performed using Enrichr ( http://amp . pharm . mssm . edu/Enrichr/ ) [51 , 52] and the DAVID 6 . 8 web-based tool ( https://david . ncifcrf . gov ) [53 , 54] . For RNA-seq , total RNA was extracted from HEK293 cells stably transfected with the empty pX459 vector and three different MGAko clones by using the RNeasy Mini system ( Qiagen ) including an on-column DNaseI digestion . RNA integrity was assessed on an Experion ( Bio-Rad Laboratories , Hercules , CA ) . Sequencing libraries were generated using the TruSeq stranded mRNA Library Preparation Kit ( Illumina Inc . ) . Libraries were quantified on a Bioanalyzer ( Agilent Technologies ) and subsequently sequenced on an Illumina HiSeq1500 platform ( Illumina Inc . ) , rapid-run mode , single-read 50 bp ( HiSeq SR Rapid Cluster Kit v2 , HiSeq Rapid SBS Kit v2–50 cycles ) according to manufacturer´s instructions . Quantitative RT-qPCR was performed essentially as described in [20] . cDNA was synthesized with the Tetro reverse transcriptase ( Bioline ) using one to two microgram of total RNA . Quantitative PCR was performed in triplicates by using the ImmoMix PCR reagent ( Bioline ) with gene-specific primers ( Table 3 ) . Values were normalized to GAPDH and/or B2M mRNA content . The source for genome sequences and annotation was Ensembl revision 83 [55] . Our ChIP-seq and RNA-seq were deposited at ArrayExpress under accession numbers E-MTAB-6006 ( ChIP-seq , HEK293 ) , E-MTAB-6007 ( ChIP-seq , mouse ES ) and E-MTAB-6005 ( RNA-seq , HEK293 ) . For assessing the overlap of PRC1 . 6 with other polycomb complexes in mESCs , the following ChIP-seq data sets were used: Ring1b ( GSM1041372 ) , Rybp ( GSM1041375 ) , Cbx6-HA ( GSM2610616 ) , Pcgf2 ( GSM1657387 ) ; Cbx7 ( GSM2610619 ) , Suz12 ( GSM1041374 ) , Kdm2b ( GSM1003594 ) and H3K27me3 ( GSM1341951 ) . | Polycomb group proteins assemble in two major repressive multi-subunit complexes ( PRC1 and PRC2 ) , which play important roles in many physiological processes , including stem cell maintenance , differentiation , cell cycle control and cancer . In mammals , six different groups of PRC1 complexes exist ( PRC1 . 1 to PRC1 . 6 ) , which differ in their subunit composition . The mechanisms that target the different PRC1 complexes to specific genomic sites appear diverse and are poorly understood . In this study , we have investigated the genomic targeting mechanisms of the non-canonical PRC1 . 6 complex . In PRC1 . 6 , the defining subunit PCGF6 is specifically associated with several proteins including the transcription factors MGA and E2F6 , and the histone-binding protein L3MBTL2 . We found that MGA is absolutely essential for targeting PRC1 . 6 . MGA executes recruitment of PRC1 . 6 to its target sites through two distinct functions . On the one hand it acts as a sequence-specific DNA-binding factor; on the other hand it has a scaffolding function , which is independent of its DNA binding capacity . E2F6 and L3MBTL2 are also important in genomic targeting of PRC1 . 6 as they promote binding of PRC1 . 6 to different sets of genes associated with distinct functions . Our finding that different components specify loading of PRC1 . 6 to distinct sets of genes could establish a paradigm for other chromatin-associated complexes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cell",
"binding",
"cell",
"physiology",
"chemical",
"characterization",
"dna-binding",
"proteins",
"cloning",
"sequence",
"motif",
"analysis",
"mammalian",
"genomics",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",... | 2018 | MGA, L3MBTL2 and E2F6 determine genomic binding of the non-canonical Polycomb repressive complex PRC1.6 |
The most frequent form of pairwise synthetic lethality ( SL ) in metabolic networks is known as plasticity synthetic lethality . It occurs when the simultaneous inhibition of paired functional and silent metabolic reactions or genes is lethal , while the default of the functional partner is backed up by the activation of the silent one . Using computational techniques on bacterial genome-scale metabolic reconstructions , we found that the failure of the functional partner triggers a critical reorganization of fluxes to ensure viability in the mutant which not only affects the SL pair but a significant fraction of other interconnected reactions , forming what we call a SL cluster . Interestingly , SL clusters show a strong entanglement both in terms of reactions and genes . This strong overlap mitigates the acquired vulnerabilities and increased structural and functional costs that pay for the robustness provided by essential plasticity . Finally , the participation of coessential reactions and genes in different SL clusters is very heterogeneous and those at the intersection of many SL clusters could serve as supertargets for more efficient drug action in the treatment of complex diseases and to elucidate improved strategies directed to reduce undesired resistance to chemicals in pathogens .
In metabolic networks , phenotypic responses to mutations that block the activity of nonessential biochemical reactions imply a fast rearrangement of fluxes . This metabolic plasticity , understood as a reorganisation or repair of damage in response to a disruption , is a signal of the robustness of metabolism against perturbations [1] . Further indications of metabolism robustness is provided by experimental results showing that more than 90% of the genes in Escherichia coli K-12 are probably not essential , with metabolic genes presenting no exception [2] . However , among the viable mutations , some are critically fragile . If a mutated enzyme-coding gene or a disrupted reaction forms a synthetic lethal ( SL ) pair with a partner , meaning that their simultaneous deletion becomes lethal for the organism even though the individual removals are not [3–7] , metabolic plasticity becomes essential to ensure viability . These synthetic lethalities provide the mutant with new vulnerabilities , exploitable for antimicrobial drug target identification [8] or , in the case of eukarya , for cancer therapy [9] . Due to the complex interconnectivity of metabolic networks [10–12] , the critical reorganisation of fluxes in most SL mutants may affect a significant fraction of reactions other than the SL pair . This would be specially relevant for SL interactions formed by a functional or active ( non-zero flux ) and a silent or inactive ( zero flux ) reaction in the native state . These interactions are called plasticity synthetic lethal ( PSL ) pairs , the dominant SL category in Escherichia coli [7] . Mutants of PSL interactions , in which the inactive reaction in the pair activates as a backup when the active reaction fails , could be impaired by a metabolic burden which should be compensated by specific metabolic/genetic mechanisms . A parallel situation has been observed , for instance , in antibiotic resistant mutants . A nonspecific metabolic stress leading to a potential fitness cost [13] has been proposed to be compensated for by adjusting metabolism without the need for acquiring compensatory mutations [14] . In the following , we use computational techniques on genome-scale metabolic reconstructions [15] of three bacteria in the Escherichia genus to define and characterize the metabolic reorganisation related with essential plasticity in PSL mutants . We quantify the increased structural and functional metabolic burden of essential plasticity , and we explore the mechanisms that buffer against the huge costs expected for sustaining alternative backup mechanisms for all PSL pairs in these organisms .
We name the WT functional reaction in a PSL pair a metabolic switch , and the WT silent coessential partner its backup . When the metabolic switch fails , phenotypic reorganization takes place to allow viability . More specifically , the inactive reaction in the PSL pair turns on acting as a functional backup which buffers against the mutation . Notice that one switch can have more than one backup , with proportions close to 4 backups for every switch in E . coli and S . sonnei to 1 in S . enterica ( Table 1 ) . For the three bacteria , all the backups associated with the same switch activate together , forming what we call a backup system . At the same time , other fluxes reorganize such that reactions that were inactive in the WT become active in the mutant and vice versa . For every switch of E . coli , S . sonnei and S . enterica , we identified the differentially activated reactions in the mutant which changed from active to inactive or from inactive to active . The whole set is named the SL cluster , ( Fig 1a ) . By definition , the SL cluster contains the switch and its backup system and the switch is unique to a SL cluster and cannot enter as a backup in any other . However , backups can serve more than one switch , and both switches and backups can enter as differentially activated reactions in other SL clusters as well . In the studied organisms , the SL clusters have a small but significant size , comprising groups of about 5% of the total number of reaction in the genome-scale reconstructions . The size distributions for all SL clusters in the three bacteria are shown in ( Fig 1b–1d . As a control , we use the differentially activated reaction sets resulting in mutants obtained when reactions which are active in WT but not essential or coessential are knocked out . To discern whether the distributions of observed SL cluster sizes were compatible with the distributions given by the control we calculated the p-value given by the Kolmogorov-Smirnov ( K-S ) test . The results in Table 1 show that we can reject that the distributions are the same for E . coli and S . sonnei , since the K-S p-value is well below 10% , while it is not the case for S . enterica . The comparison of the curves shows that the size distribution of SL clusters has a long tail in E . coli and S . sonnei which is absent in the control . Therefore , essential plasticity is more complex and involves more metabolic reactions than reorganizations induced by nonessential mutations . This is also valid for results in rich medium , see inset ( Fig 1b ) . Reactions in SL clusters form cohesive structures . First , the switch belongs to a connected component of differentially activated reactions which includes its backups The only exception corresponds to the SL cluster associated to the switch reaction phosphoenolpyruvate carboxylase in S . enterica; it has Isocitrate lyase as its only backup , which is not included in the internal connected component but forms an isolated component inside the corresponding SL cluster . Second , this connected component is the largest in the SL cluster and on average includes more than 92% of its reactions , ( Fig 1e–1g ) . The rest form residual disconnected components scattered through the metabolic network . Only 2 clusters in E . coli and S . sonnei and 4 in S . enterica out of approximately 60 in each organism ( Table 1 ) deviate from this behavior . The explanation for the divergence of the sizes of the SL cluster and its connected component is most frequently a change of strategy in mutants , like the switch from aerobic to anaerobic metabolism associated with the extracellullar transport of oxygen in E . coli , and to the reaction protoporphyrinogen oxidase in S . sonnei . Special mention deserves the SL cluster in the three organisms associated with the extracellular transport of glucose via diffusion . Both the switch and the backup are connected to exactly the same reactants and products . Therefore , they are equivalent alternatives for FBA , meaning that we cannot distinguish between switch and backup because both reactions in the coessential pair can indistinctly play both roles . As a consequence , the corresponding internal connected component is only formed by the switch and the backup while the SL cluster contains other reactions . We found no other topologically invariant coessential PSL pairs in the organisms . The fact that coessential partners of switch reactions in PSL pairs remain silent in the WT and only change their activation state when needed to ensure the viability of the organism points to possible costs associated to the activation of these backup systems . To check this hypothesis , we measured both structural and functional costs associated to the viability of PSL mutants . Notice that the biomass yield is not a good indicator of possible functional costs . In more than 80% of the mutants the FBA solution is suboptimal but very close to optimal , while growth is not changed with respect to the WT in the rest . We quantified the flux and energetic requirements of PSL mutants as compared to the WT . The flux disparity measure is given by the ratio of the total flux running through reactions in SL clusters in mutants relative to that of the same reactions in the WT , [Fmutant/FWT]cluster . The results in ( Fig 2a–2c ) show that this quantity is specific to the mutant . However , in E . coli and S . sonnei more than 70% of the clusters have a mutant-to-WT flux ratio larger than one with average values 542 . 7 and 6 . 6 , respectively . Conversely , the flux through more than half of the SL clusters of S . enterica is lower in PSL mutants than in the WT , but flux decreases are relatively minor in most PSL mutants while flux increases are dramatic , such that the average of the ratio over mutants is more than 90000 ( Table 1 ) . The observed flux increase is not only related to the larger number of active reactions in SL clusters of mutants as compared to the WT , ( Fig 2a–2c ) and S1 Fig , but also to a higher average flux per active reaction , see S2 Fig . More specifically , 53 mutants in E . coli have more active reactions in their SL cluster than in the WT , and the average flux per active reaction is higher in 69% of the clusters . In S . sonnei , the number of mutants with more active reactions in the SL cluster than in the WT is still very high , 49 of 63 , but the situation is more balanced regarding the average flux per active reaction . In S . enterica , still 40 of the 61 mutants have the same or more active reactions in SL clusters but the average flux per active reaction is generally lower . The second magnitude calibrating the functional cost of viability in PSL mutants is given by the ratio of ATP production in the mutants as compared to the WT , Emutant/EWT . ATP production is defined as the flux through the ATP maintenance reaction –which is a balanced ATP hydrolysis reaction used to simulate energy demands not associated with growth– versus the intake flux of oxygen , both obtained by optimizing biomass yield [7] . The flux through the ATP maintenance reaction is preserved and deviations of the ATP production in some mutants as compared to WT are basically due to changes in oxygen consumption , see S3 Fig . The number of mutants showing this altered phenotype is 35% in E . coli and 41% in S . sonnei . Oxygen needs for the production of the same ATP amount in the affected E . coli mutants is always increased , while it is always decreased in altered S . sonnei mutants . Curiously , the variations in oxygen consumption in mutants of both bacteria is in all cases a fixed amount . The increased number of active reactions in most mutants gives a rough indication of the cost of essential plasticity at the structural level . A different estimation can be obtained by evaluating the degree centrality [23] of reactions within the internal connected component of a SL cluster , which informs about their role in mediating interactions among other reactions . The investigation of this quantity for the three bacteria , shown in ( Fig 2d–2f ) ( the results are qualitatively similar for other centrality measures ) , reveals that PSL coessential reactions , and especially backups , have higher centrality as compared to noncoessential reactions . The centrality distribution of switches features a tendency towards lower centrality values , but they always possess a marked peak at very large values . This makes them , on average , more central than noncoessential reactions . Finally , backup reactions have a distribution which is more concentrated at intermediate-to-large centrality values . In fact , the probability that backups have higher centrality than switches and than noncoessential reactions in E . coli and S . sonnei is approximately 0 . 64 and 0 . 55 , respectively , while the value for S . enterica is notably smaller , 0 . 37 . SL clusters are not disjoint but strongly coalescent in their component reactions and associated genes . The first observation supporting this is given by the repeated usage of the same backup system for different switches in an organism . This redundancy is intensive in E . coli and S . sonnei , in which ≈ 20% of the switches share what we call the “fatty acid biosynthesis backup system” formed by eleven reactions ( the “fatty acid biosynthesis backup system” includes the four 3-oxoacyl-[acyl-carrier-protein] synthase , the three 3-oxoacyl-[acyl-carrier-protein] reductase reactions , and the four 3-hydroxyacyl-[acyl-carrier-protein] dehydratase reactions in the bacteria ) . This specific redundancy explains the prominent function of Cell Envelope Biosysthesis as a backup for Membrane Lipid Metabolism , with a concentration of PSL pairs at their interface [7] . The redundancy not only affects backup reactions . We ranked all reactions participating in SL clusters by their occurrence in different clusters . Results are shown in ( Fig 3a–3c ) . We compared the rankings of switches , backups and noncoessential reactions with the control given by the sets of differentially activated reactions resulting when knocking down reactions which are active in WT , but which are not essential or coessential . As expected , the profiles of participation in SL clusters of switches and backups are similar . In fact , the p-values of the K-S tests ( Table 1 ) do not allow to exclude that switch and backup reactions have the same distribution for all organisms , while these are different from the curves of the control test , which decreases much more quickly in E . coli and S . sonnei indicating a stronger overlap of SL clusters as compared with rearrangements caused by random deletion of reactions . In line with what has been observed so far , the picture changes in the case of S . enterica , with no clear distinction between switches , backup , and control reactions . As a common trend in the three bacteria , we observe a heterogeneity of participation values such that a small number of reactions participates in a very large fraction of clusters . This suggests that metabolic reorganizations caused by essential plasticity happen mainly by leveraging on some key reactions . For instance , the reductase reactions producing Isopentenyl diphosphate and Dimethylallyl diphosphate , used by organisms in the biosynthesis of terpenes and terpenoids , form a SL cluster in E . coli and coappear in 70% of all SL clusters in this organism . The two reactions of this SL cluster are totally exchangeable in terms of assuming the switch or backup role . All the statistics measured to characterize the functional and structural cost of essential plasticity remain invariant when the roles are swapped , and their FVA flux ranges in optimal states are identical . Another intriguing result is related with oxygen consumption in mutants of E . coli and S . sonnei . Both organisms display a strong overlap of SL clusters of the mutants which display an altered consumption of oxygen in comparison with WT . Strikingly , all of them share a module of three reactions for the exchange , transport and oxidation of iron , which never participate in SL clusters of mutants which do not show alteration of oxygen consumption . This is in agreement with observations that report oxygen as a signal for regulating iron acquisition in Shigella [24] . The strong entanglement of SL clusters can also be explored by constructing the participation backbone . This is a graph representations in which two clusters i and j are connected by a directed link from i to j if the proportion of reactions in cluster i which are also in cluster j is larger than a certain significance level . The unfiltered participation graph in which two clusters are linked whenever they share at least one reaction is almost fully connected and not very informative . In order to give a meaningful backbone , the significance level is chosen to optimize the trade-off between preserving the maximum number of clusters while reducing the number of links to just those which are significant [25] , see S4 Fig . For the three bacteria , we have used a significance level of 75% , meaning that two clusters are connected if more than 75% of the reactions in one are also in the other , ( Fig 3d–3f ) . We have performed other complementary tests , see the network representation of pathways entangled through PSL pairs S5 Fig and reaction entanglement matrices showing co-occurrences of reactions in SL clusters , S6 , S7 and S8 Figs . The sparsity of the backbones allows us to study quantitatively the hierarchical ordering of SL clusters using the k-core decomposition [26] , which identifies groups of clusters having k-c or more connections among them ( see Materials and methods ) . Each SL cluster is annotated with the k-c index given by the maximum k-core it belongs to , computed on the basis of outgoing connections . Notice that SL clusters in the highest k-cores are very densely connected among them . In E . coli and S . sonnei , the k-core partition denotes a hierarchical core-pheriphery structure . SL clusters are clearly separated into two groups with low and high k-c value , the latter typically formed by the largest SL clusters . The only exception is the size-two SL cluster in E . coli Isopentenyl- Dimethylallyl diphosphate mentioned above , which acts as an important interface between the core and many peripheral clusters , see ( Fig 3d ) and S9 Fig . The core contains approximately 1/5 of the SL clusters and forms an almost fully connected set of reactions participating in many other SL clusters . In contrast , the k-core layout of S . enterica is almost flat with no relevant core-periphery structure . At the level of genes , the entanglement of SL clusters is even stronger . Here , we consider genetic units , which can be single genes or gene complexes ( sets of functionally related genes that regulate together a metabolic reaction in a SL cluster via an AND logical relation ) . To clarify the question whether there is a common set of regulatory genes for SL clusters we ranked metabolic genes according to the number of SL clusters in which they participate , results in ( Fig 4a–4c ) . Hub genes participate in more than 70% of SL clusters and the top 10% enter in about 50% of the sets . One example is the gene regulating the function of the reductase reactions Isopentenyl and Dimethylallyl diphosphate in E . coli . This gene appears to be involved in cell lysis and in the stress response of bacteria in reaction to amino-acid starvation , fatty acid limitation , iron limitation , heat shock and other stress conditions . Its action causes the cell to divert resources away from growth and division toward amino acid synthesis in order to promote survival until nutrient conditions improve . However , more than 50% of the genes is specific to up to three SL clusters . Interestingly , the gene participation curve decays faster for S . enterica than for the other two bacteria , highlighting again a more limited organization . In ( Fig 4d–4f ) , we plot both the number of unique genes entering in a cluster ( Unique ) and their total number by counting repetitions ( Occurrences ) . The number of occurrences grows linearly with the number of reactions in each set , with an approximate slope of 1 . 4 in the three bacteria , which indicates that complexes are frequently associated to the regulation of SL clusters . Interestingly , the number of unique genes follows the same linear growth up to a ‘critical’ value from which it saturates to a constant around 40 for SL clusters with more than ∼60 reactions ( around half the maximum size of SL clusters ) . The saturation effect implies that large SL clusters are regulated by a reduced number of different genes and that the basis of regulation of genes and the role of complexes grows with the number of reactions in the set . These features are common to the three organisms , although SL clusters in S . enterica are smaller than in the other two bacteria , so that the saturation effect is less evident and the redundancy of genes is more limited . Pairs of genetic units show also a clear tendency to co-occurrence in SL clusters , see the gene entangle matrices for the three bacteria in S10 , S11 and S12 Figs . To check the dependency of SL clusters on alternative optimal FBA solutions in glucose minimal medium , we have analyzed the congruency of the detected SL clusters in E . coli with the range of possible reaction fluxes , as determined by FVA [27] fixing the optimal growth rate ( see Materials and methods ) . We found that 68% of the clusters , 41 out of 60 , present optimal flux ranges which are fully consistent with the solution reported in the previous sections , meaning that the minimum flux attainable by switch reactions ensures that they are active while , simultaneously , the flux of the corresponding backup reactions is basically confined to be zero in optimal states . In 5 more cases the maximum allowed flux through the switch reaction is at least a hundred times larger than that of the corresponding backups . Only 4 clusters , less than 7% , show backup systems with all reactions practically bounded to zero except for one of them , which presents a potential maximum flux comparable to that of the switch reaction . However , the biomass yield is slightly decreased in the corresponding mutant as compared to the WT , as a consequence of the activation in the backup system of reactions which are bounded to zero flux in the optimal states . Finally , the remaining 10 SL clusters correspond to alternative optimal solutions in which the switch and the backup system can exchange role in terms of optimal biomass production , although flux and energetic requirements of the mutants can be different . Results were also obtained for E . coli in rich medium ( see Materials and methods ) . The total number of different reactions in SL clusters is a factor 1 . 8 larger than in minimal medium and the average number of backups per switch increases approximately in the same proportion . The number of switches is slightly increased ( Table 1 ) . Similar to glucose minimal medium , the size distribution of SL clusters has a longer tail as compared with the control , inset ( Fig 1b ) , and SL clusters of differentially activated reactions are formed basically by a connected component , inset ( Fig 1e ) . Of the 287 reactions in SL clusters in glucose minimal medium , 85% ( 244 ) also belong to SL clusters in rich medium . Of them , only 8 . 6% ( 21 ) change role . In particular , 15 which were coessential are rescued and become nonessential in rich medium . More than 70% of switches and backups in glucose minimal medium are conserved , 43 and 42 respectively . Of them , 40 switches preserve exactly the same backup system , 2 switches acquire an extra backup reaction , and 1 , corresponding to the switch reaction isopentenyl pyrophosphate isomerization , changes backup . In this case , the isomers Isopentenyl diphosphate , less reactive , and Dimethylallyl diphosphate , more reactive , can only be produced by the corresponding reductase reaction and by the isopentenyl pyrophosphate isomerization reaction . In glucose minimal medium , the more reactive form is obtained via the less reactive isomerase and not directly by the action of the corresponding reductase , so that the reductase producing Isopentenyl diphosphate and the isomerization reaction act as switches with the reductase reaction producing directly Dimethylallyl diphosphate as a backup . The reverse is observed in rich medium , where the less reactive form is obtained from the more reactive one so that the reductase producing Dimethylallyl diphosphate and the isomerization reaction act as switches with the reductase reaction producing directly Isopentenyl diphosphate as a backup . Interestingly , in two more SL clusters the switch and the backup swap roles so that alternative mechanisms are used to produce specific metabolites . One of them is related to the production of deoxyuridylic acid , an intermediate in the metabolism of deoxyribonucleotides . The hidrolase reaction , without the direct intervention of ATP , is active in glucose minimal medium while the active reaction in rich medium is the thymidine kinase catalysed reaction , which involves the direct consumption of ATP . We see that the cost and energetic requirements of rearrangements from WT to mutant do vary between minimal and rich medium conditions ( Fig 2 ) . The flux ratio per module Fmutant/FWT tends in fact to generally decrease , at odds with the minimal medium case . This is however reasonable , since in the rich medium there are several metabolic routes that connect nutrients to biomass . Introducing mutations may disrupt many of these routes ( literally switching off the metabolism of some of the redundant nutrients ) without impairing survival , decreasing the total flux running through the modules as an effect . In the minimal medium case , instead , it is impossible to disrupt these routes without incurring in lethality . Besides this discrepancy , switches and backup participations in different SL clusters decrease faster than non coessential and control reactions , suggesting that SL cluster entanglement decreases dramatically in the presence of multiple nutrients , as also suggested by the reduced average number of backups per switch .
In summary , we propose a change of paradigm in the approach to understand the phenomenon of synthetic lethality . The complexity of molecular interactions at the cell level urge us to go from the mere screening of SL reaction or gene pairs , or even of triplets or higher order motifs , to the study of SL clusters and their entanglement . Approaching directly SL pairs of reactions or genes without their multifunctional integration in clusters is like drawing paths between pairs of geographical places without the scaffold of a map telling how the different paths relate to each other . The complete portray at the systems level is far more complex than a collection of separate PSL pairs . Beyond theoretical implications for the understanding of plasticity in metabolic networks , our results could help to identify drug action and to design improved strategies that reduce undesired resistance in synthetic lethal interactions to chemicals in pathogens . We believe that SL clusters will be also found in human cells , with important implications for biomedicine and biotechnology . Our work reveals that backups that belong to the same SL cluster offer alternative but equivalent targets , a clear advantage in cases in which the experimental targeting of some specific reaction is technically more difficult . At the same time , not all computationally detected SL pairs have the same quality as potential therapeutic targets in complex diseases such as cancer or to fight infections of pathogens . We expect that more redundant coessential reactions with a higher participation in different SL clusters can become efficient and reliable supertargets .
In Table 1 , and for each of the three bacteria , we report the metabolic reconstruction ( Organism , Model , Reference ) , the number of reactions included in the genome-scale reconstruction ( NR all ) , the number of reactions possibly active according to FVA ( NR FVA ) , the number of active reactions in the FBA solution in glucose minimal medium ( NR active ) , the number of metabolites in the reconstruction ( NM actual ) and those resulting when considering compartments ( NM synth ) , the number of single essential reactions , the number of Plasticity SL pairs , and of Redundancy SL pairs , the number of SL clusters and switches ( Rswitches ) , and the number of backups ( Rbackups ) , noncoessential reactions ( Rnoncoess ) , total different reactions ( Total R ) , and Gene units associated to the SL clusters . We also report the ratio of the total flux running through reactions in SL clusters of mutants and WT ( [Fmutant/FWT]cluster ) , the ratio of active reactions in SL clusters of mutants and in the WT ( ANG , mutant/AWT ) , and the ratio of ATP production in PSL mutants and in the WT ( ENG , mutant/ENG , WT ) . In relation to centrality measures , we report the fraction of times that backups have higher centrality than switches inside the internal connected component of SL clusters CBgtS , the fraction of times that backups have higher centrality than non coessential reactions inside the internal connected component of SL clusters CBgtNC , the average centrality of backup reactions in the internal connected component of SL clusters 〈CB〉 , the average centrality of switch reactions in the internal connected component of SL clusters 〈CS〉 , the average centrality of noncoessential reactions in the internal connected component of SL clusters 〈CNCE〉 . For E . coli , values in parentheses correspond to rich medium , see next subsection . We have also computed the p-values of the Kolmogorov-Smirnov ( K-S ) test for the different distributions that we obtain in our analysis . More specifically , we have calculated the distances between the distributions of observed SL cluster sizes and the distributions given by the control shown in ( Fig 1b–1d ) ( p-values in row SL cluster sizes ) , and between the curves associated to the rankings of occurrences of reactions in different SL clusters as shown in ( Fig 3a–3c ) ( p-values in rows Ranking in clusters , where S stands for Switch , B for Backups , and NC for Noncoessential . The rest of potential comparison , for instance between S or B with the control , have p-values which are virtually zero and are not included in the table ) . Flux Balance Analysis ( FBA ) [15] is a technique which allows to compute metabolic fluxes without the need of kinetic parameters , just by using constrained-optimization . The vector of the time variation of the concentrations of metabolites c ˙ is related with the stoichiometric matrix S of the whole network ( it contains the stoichiometric coefficients of each metabolite in each reaction of the network ) and the vector of fluxes ν , c ˙ = S · ν . Steady-state is assumed , thus S ⋅ ν = 0 . In general , metabolic networks contain more reactions than metabolites , and hence the system of equations for the fluxes is underdetermined . Hence , a biological objective function must be defined in order to select a biologically meaningful solution . In this work , we use FBA to find the solution that optimizes the growth of the organism , which is equivalent to maximize biomass formation . Reversibility of reactions is also added in order to constrain the solutions . Since we have a linear system of equations with linear constraints , Linear Programming is used in order to compute a flux solution in a small amount of time ( of the order of 1 s ) , which implies a computationally cheap method . Flux variability analysis [27] is computed by fixing the biomass yield to its optimal value and by extracting the maximum and minimum flux value associated to each reaction in this fixed optimal condition via linear programming . According to the specifications in each metabolic reconstruction , growth in glucose minimal medium was simulated by fixing the lower bound of the glucose exchange reaction to −10mmol/ ( gDW ⋅ h ) for E . coli and S . sonnei , and to −5mmol/ ( gDW ⋅ h ) for S . enterica . For the rich medium in E . coli , we used a Luria-Bertani Broth [34] , which contains as additional compounds purines and pirimidines apart from amino acids . We also added vitamins , namely biotin , pyridoxine , and thiamin , and also the nucleotide nicotinamide monocleotide [31] . Other compounds , like PABA or chorismate , cannot be uptaken by the E . coli model that we are using . The exchange constraints bounds of these compounds are set to −10mmol/ ( gDW ⋅ h ) ( ν e x c h a n g e c o m p o u n d ≥ - 10 ) . A detailed list of the added compounds is given in S4 Table . From the set of reactions in the genome-scale reconstructions , we excluded essential reactions detected computationally and also spontaneous reactions ( e . g transport reactions ) . We focused exclusively on reactions catalyzed by enzymes with an associated gene . In this way , we identified a set of candidate reactions in each organism that can be removed individually , but whose pair deletion may be lethal for the organism . We checked every possible pair by applying FBA to the double mutant . As in [7] , we classified the detected SL pairs into plastic and redundant , depending on whether only one or both reactions are active in the FBA solution in the given medium . We modelled metabolism as a bipartite directed network [35] , where directed links connect metabolites with reactions in which they participate as reactants or products . The degree centrality of a reaction is simply given by its degree k , measuring the number of other reactions connected to it by shared metabolites . We define the normalized degree centrality of a reaction in the internal connected component of a SL cluster as kr/ ( R − 1 ) , where kr stands for the number of reactions connected to reaction r inside the connected component , and R is its total number of reactions . In complex networks , the k-core decomposition of a graph allows to disentangle the hierarchical structure of networks by progressively focusing on their central cores . It is a threshold-based hierarchical decomposition of a graph . A k-core of level c is defined as the maximal subgraph in which all the nodes have at least c internal connections [26] . It can be obtained by recursively deleting all nodes with less than c connections until all nodes in the remaining graph have at least c neighbors . Notice that the obtained cores for different values of c form a nested hierarchy of subgraphs . As a consequence , each node in a network can be labeled with an index k − c called its coreness , meaning that the node belongs to the k − c core but has been removed by the process in the k − c + 1 core . In the case of directed networks , in which nodes have both incoming and outgoing neighbours , the k-core decomposition can be based on the incoming connections , the outgoing connections , or a combination of the two . In this work , we based our definition on outgoing connections , such that the coreness of a node , which represents a SL cluster , is given by its maximum k-core out , where a k-core out with value k − c is defined as the maximal subgraph of SL clusters such that all the SL clusters in it have at least k − c outgoing connections inside the subgraph . | Synthetic lethality ( SL ) , in which the combined knockout of two nonessential genes or reactions is lethal , has direct applications in recognising targets for therapeutic treatment of complex diseases and for fighting against undesired resistance . Typically , SL interactions are reported in pairs . We propose a change of paradigm based on the fact that SL interactions in metabolism are not independent of each other but form complex backup systems involving the rearrangement of a significant SL cluster of metabolic fluxes to ensure viability . This robustness comes at the expenses of acquired vulnerabilities and increased costs , mitigated by the entanglement of SL clusters in terms of shared reactions and genes , which could serve as supertargets for a new generation of therapeutic treatments . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"bacteriology",
"carbohydrate",
"metabolism",
"medicine",
"and",
"health",
"sciences",
"gut",
"bacteria",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"metabolic",
"networks",
"microbiology",
"glucose",
"metabolism",
"escherichia",
"coli",
"bacterial",
"dise... | 2018 | Metabolic plasticity in synthetic lethal mutants: Viability at higher cost |
MicroRNA-22 ( miR-22 ) is emerging as a critical regulator in organ development and various cancers . However , its role in normal hematopoiesis and leukaemogenesis remains unclear . Here , we detected its increased expression during monocyte/macrophage differentiation of HL-60 , THP1 cells and CD34+ hematopoietic stem/progenitor cells , and confirmed that PU . 1 , a key transcriptional factor for monocyte/macrophage differentiation , is responsible for transcriptional activation of miR-22 during the differentiation . By gain- and loss-of-function experiments , we demonstrated that miR-22 promoted monocyte/macrophage differentiation , and MECOM ( EVI1 ) mRNA is a direct target of miR-22 and MECOM ( EVI1 ) functions as a negative regulator in the differentiation . The miR-22-mediated MECOM degradation increased c-Jun but decreased GATA2 expression , which results in increased interaction between c-Jun and PU . 1 via increasing c-Jun levels and relief of MECOM- and GATA2-mediated interference in the interaction , and thus promoting monocyte/macrophage differentiation . We also observed significantly down-regulation of PU . 1 and miR-22 as well as significantly up-regulation of MECOM in acute myeloid leukemia ( AML ) patients . Reintroduction of miR-22 relieved the differentiation blockage and inhibited the growth of bone marrow blasts of AML patients . Our results revealed new function and mechanism of miR-22 in normal hematopoiesis and AML development and demonstrated its potential value in AML diagnosis and therapy .
Hematopoiesis is a highly ordered multistep process that is orchestrated by various regulators including transcriptional factors [1] , cytokines [2] , and noncoding RNAs [3] . Deregulation of important regulators in hematopoiesis could induce hematopoietic cancers including acute myeloid leukemia ( AML ) [4] . In recent years , microRNAs ( miRNAs ) are emerging as novel regulators in myelopoiesis and AML . The aberrant expression of the miR-17-92 cluster which is epigenetically regulated by PU . 1 contributes to leukaemogenesis [5] . MiR-223 acts as a fine-tuner of granulocyte production and the inflammatory response in mice [6 , 7] , while miR-142-3p and miR-29a promote myeloid differentiation [8] . Abnormally expressed miRNAs are associated with AML , serving as powerful prognostic indicators [9–13] . Several miRNAs , such as the miR-29 family [14] and the miR-181 family [15] have proven promising in AML therapy . To reveal more miRNAs participating in myelopoiesis and AML development , we performed miRNA chip analysis and identified several miRNAs with expression change during monocyte/macrophage differentiation and in AML patients . Among them , miR-22 exhibited evident up-regulation during the differentiation and abnormal down-regulation in AML patients . MiR-22 is reported to play an important role in several physiology processes and cancers . In mice , miR-22 targets Irf8 mRNA and controls the differentiation of dendritic cells [16] . Moreover , miR-22 is a critical regulator in cardiomyocyte hypertrophy and cardiac remodeling [17 , 18] . Interestingly , miR-22 can act as either a tumor suppressor or an oncogene in different cancers . miR-22 inhibits cell growth and induces cell-cycle arrest , apoptosis and senescence in breast cancer , colon cancer and lung cancer [19–22] . miR-22 was also reported to promote chronic lymphocytic leukemia B cell proliferation via activation of the PI3K/AKT pathway [23] . MECOM ( MDS1 and EVI1 complex locus ) , also termed EVI1 ( Ecotropic viral integration site 1 ) , was first identified as a murine common locus of retroviral integration in myeloid leukemia [24] . Several studies have demonstrated MECOM as a regulator in the maintenance [25] and differentiation [26] of mouse hematopoietic stem cells . However , the function of MECOM in human hematopoiesis is poorly understood . The inappropriate high expression of MECOM is an adverse prognostic marker in AML [27] . MECOM can act as a transcriptional factor [25] , epigenetic regulator [28] , or repressor of key transcriptional factors in hematopoiesis such as PU . 1 and GATA1 via protein—protein interaction [26 , 29] . MECOM mRNA was previously identified as a miR-22 target in metastatic breast cancer cells [30] . Here , we showed that miR-22 is transcriptionally activated by PU . 1 during monocyte/macrophage differentiation , and that miR-22 promotes the differentiation by targeting MECOM mRNA and further increasing interaction between c-Jun and PU . 1 . We also showed miR-22 to be a repressor miRNA in AML development and examined whether it could be a therapeutic target for AML therapy .
We performed quantitative real-time PCR ( qRT-PCR ) to detect miR-22 expression in peripheral blood ( PB ) mononuclear cells ( MNCs ) derived from 79 primarily diagnosed AML patients ( S1 Table ) and 114 healthy donors , as well as in bone marrow ( BM ) MNCs and in BM CD34+ hematopoietic stem cells and progenitors ( HSPCs ) derived from limitary healthy donors and AML patients . Significantly decreased miR-22 levels were observed in the AML patients as compared with the healthy donors for each kind of the materials ( Fig 1A ) . Receiver-operating characteristic curve analysis of miR-22 suggested that the miR-22 level in each kind of the materials could be as a reference marker with high sensitivity and specificity for AML diagnosis ( S1 Fig ) . We performed qRT-PCR and Northern blot analyses to detect changes in miR-22 expression during monocyte/macrophage differentiation . The results revealed that miR-22 was gradually elevated during phorbol myristate acetate ( PMA ) -induced monocyte/macrophage differentiation of HL60 and THP1 cells , as well as during monocyte/macrophage induction of CD34+ HSPCs derived from human umbilical cord blood ( UCB ) ( Fig 1B ) . As miR-22 increased substantially during monocyte/macrophage differentiation , we examined if it was regulated by key transcriptional factors . We first performed rapid amplification of cDNA 5’ end ( 5’ RACE ) and identified the transcriptional start site ( TSS ) of miR-22 at 60 base-pair downstream of the predicted 5’ end of C17orf91 where miR-22 located ( Fig 1B ) . Two potential PU . 1 binding sites PB1 and PB2 ( Fig 1C ) were identified within the region of -2 to +0 . 2 kb from the TSS using the Transcription Element Search System . Moreover , PU . 1 inhibition by siPU1 transfection resulted in obviously reduced miR-22 expression in HL60 and THP1 cells ( Fig 1D ) . Furthermore , PU . 1 knockdown by Lenti-shPU . 1 infection impaired miR-22 primary transcript ( Pri-miR-22 ) expression , while ectopic expression of PU . 1 by Lenti-PU . 1 infection induced Pri-miR-22 expression ( Fig 1E ) . To examine whether PU . 1 physically interacts with miR-22 promoter in vivo , we performed chromatin immunoprecipitation ( ChIP ) assays in HL60 and THP1 cells . The DNA fragments immunoprecipitated by the PU . 1 antibody were amplified with two pairs of PCR primers surrounding PB1 and PB2 . Only the fragments containing PB2 were detected ( Fig 1F ) . Moreover , ChIP-qPCR showed that the interaction between PU . 1 and miR-22 promoter became stronger with PMA induction . The specificity of PU . 1 occupancy in PB2 was demonstrated by the absence of immunoprecipitated chromatin fragments corresponding to an unrelated genomic region ( UR ) and the presence of immunoprecipitated chromatin fragments corresponding to a known PU . 1-binding site within the CSF1R promoter ( pro-CSF1R ) [31] ( Fig 1G ) . We then cloned the genomic fragment surrounding PB2 ( PB2_WT ) or a mutant version ( PB2_Mut ) , into a promoterless luciferase reporter plasmid pGL3-basic . The recombinant plasmid expressing PU . 1 pcDNA3 . 1-PU . 1 ( or pcDNA3 . 1 control ) , together with the luciferase reporter plasmid pGL3-PB2_WT ( or pGL3-PB2_Mut ) , were co-transfected into 293T cells . Notably , the PB2_WT fragment yielded robust PU . 1-dependent transcriptional activity , while the mutation eliminated the transcriptional activity ( Fig 1H ) . These findings provide compelling evidence that miR-22 is transcriptionally regulated by PU . 1 . To examine the effect of miR-22 on monocyte/macrophage differentiation , we firstly transfected THP1 and HL60 cells with miR-22 mimics or anti-miR-22 , and then induced monocyte/macrophage differentiation . The overexpression or inhibition of miR-22 was confirmed by qRT-PCR ( S2 Fig ) . Remarkably , the flow cytometry data revealed a higher percentage of CD14-positive cells in the miR-22 mimics-transfected cells , while a lower percentage was observed in the anti-miR-22-transfected cells ( Fig 2A ) . qRT-PCR analysis also demonstrated that the ectopic expression of miR-22 increased while the inhibition of miR-22 expression impaired CD14 and CSF1R mRNA levels in the cells ( Fig 2B ) . In addition , May-Grünwald Giemsa staining demonstrated that the overpresence of miR-22 promoted monocyte/macrophage cell development in the PMA-induced cells ( Fig 2C ) . Next , we infected HSPCs with a recombinant lentivirus harbouring a miR-22 precursor ( Lenti-miR-22 ) or expressing hairpins that exhibited anti-miR-22 activity ( Lenti-ZIP-miR-22 ) and a relative control ( Lenti-Con or Lenti-ZIP-Con ) , and induced monocyte/macrophage differentiation . As shown in Fig 3A , miR-22 overexpression increased while miR-22 knockdown decreased CD14 mRNA levels . Flow cytometry analysis revealed that miR-22 overexpression increased ( Fig 3B , left ) while knockdown of miR-22 decreased the percentage of CD14-positive cells ( Fig 3B , right ) . Furthermore , Lenti-miR-22 infection significantly promoted the colony-forming activity of CFU-M and CFU-GM , both in clone size and number ( Fig 3C ) . May-Grünwald Giemsa staining confirmed that Lenti-miR-22 infection led to higher percentage of more mature monocyte/macrophage cells ( Fig 3D ) . Altogether , these data indicated that miR-22 is a positive regulator in monocyte/macrophage differentiation . To identify miR-22 targets contributing to the phenotypes observed , we looked over the potential targets predicated by TargetScan and noticed MECOM , a transcription factor and oncoprotein , which was documented to be involved in hematopoietic stem cell function [32] and blood generation [33] ( Fig 4A ) . To determine whether MECOM is directly regulated by miR-22 , the 3’UTR of MECOM was inserted downstream of the luciferase ORF of pMIR-REPORT . Significant repression of luciferase activity caused by miR-22 transfection was observed and the repression effect was abrogated by mutagenesis of the core miR-22 binding site in the 3’UTR of MECOM ( Fig 4B ) . Furthermore , enforced miR-22 expression reduced the MECOM protein level , while inhibition of miR-22 expression elevated the MECOM protein level in THP1 and HL60 cells ( Fig 4C ) . These data demonstrated that MECOM is a target of miR-22 in the AML cell lines . We then examined the function of MECOM in monocyte/macrophage differentiation of THP1 cells . MECOM knockdown by Lenti-shMECOM infection in THP1 cells significantly increased the percentage of CD14-positive cells ( Fig 4D ) , and also mRNA levels of the differentiation markers CD11b , CD14 and CSF1R ( Fig 4E ) . In addition , May-Grünwald Giemsa staining showed that MECOM knockdown promoted monocyte/macrophage development ( Fig 4F ) . To confirm whether the miR-22 regulation of monocyte/macrophage differentiation occurred via its regulation on MECOM , we performed rescue assays . As shown in Fig 4G , the increase in MECOM expression ( left ) was accompanied by decreased CD14-positive cells ( middle and right ) after anti-miR-22 transfection ( b vs . a ) . As expected , retransfection with siMECOM reduced the increase of MECOM expression resulting from anti-miR-22 treatment ( left , c vs . b ) , which was accompanied with restoration of the percentage of CD14-positive cells ( middle and right , c vs . b ) . Collectively , these results demonstrated that the enhancement of monocyte/macrophage differentiation induced by miR-22 occurred at least partially via its negative regulation on MECOM . MECOM was reported to impair myeloid differentiation via blocking the association of PU . 1 with c-Jun [26] , a critical coactivator of PU . 1 transactivation . A recent report demonstrated antagonism between PU . 1 and GATA2 in the transcriptional regulation of some genes [34] . GATA2 was also reported to inhibit the binding of PU . 1 to c-Jun [35] . Interestingly , MECOM was reported to directly target the GATA2 promoter to promote its transcription [25] . In addition , MECOM can interfere with interaction between JNK ( c-Jun N-terminal kinases ) and c-Jun , thus reducing the level of phosphorylated c-Jun ( p-c-Jun ) [36] , which is capable of inducing c-Jun expression via the formation of the heterodimer AP-1 with c-Fos [37] . Based on this evidence , we examined whether miR-22 affects the interaction between c-Jun and PU . 1 through regulating MECOM in HL60 and THP1 cells . As shown in Fig 5A , the levels of MECOM and GATA2 decreased , whereas the levels of p-c-Jun , c-Jun and PU . 1 increased following the PMA-induced monocyte/macrophage differentiation of THP1 and HL60 cells . We assessed the effects of miR-22 on the expression of these factors . In THP1 cells , ectopic expression of miR-22 reduced MECOM and GATA2 levels and increased p-c-Jun and c-Jun levels; but it barely affected PU . 1 expression . In contrast , anti-miR-22 transfection led to an increase of MECOM and GATA2 levels , and a concomitant reduction of p-c-Jun and c-Jun levels , but little change was observed in PU . 1 expression ( Fig 5B ) . Similar results were obtained from the transfected HL-60 cells ( S3 Fig ) . Furthermore , MECOM knockdown by Lenti-shMECOM infection showed similar results as miR-22 overexpression ( Fig 5C ) in THP1 cells . We next performed co-immunoprecipitation analysis on the infected THP1 cells . As shown in Fig 5D , miR-22 overexpression increased c-Jun levels but barely affected PU . 1 expression ( lane 1 and 2 ) . Normal mouse IgG was not able to immunoprecipitate PU . 1 and c-Jun ( lane 3 and 4 ) . However , anti-PU . 1 antibody immunoprecipitated more endogenous c-Jun in the cells infected with Lenti-miR-22 than those infected with Lenti-Con ( IP-PU . 1 , up , lane 5 and 6 ) , while the immunoprecipitated endogenous PU . 1 level was almost the same in the two groups ( IP-PU . 1 , down , lane 5 and 6 ) . Moreover , the anti-c-Jun antibody immunoprecipitated more endogenous c-Jun and PU . 1 in the cells infected with Lenti-miR-22 than in the Lenti-Con-infected cells ( IP-c-Jun , lane 5 and 6 ) . These results demonstrate that miR-22 increases interaction between PU . 1 and c-Jun . To further determine if the mechanism by which miR-22 regulates monocyte/macrophage differentiation revealed in THP1 and HL-60 cells also exists in normal hematopoiesis , we analyzed the expression of miR-22 and its target protein MECOM . A gradual increase in miR-22 levels whereas a decrease in MECOM mRNA and protein levels were detected during the monocyte/macrophage induction culture of CD34+ HSPCs derived from human HCB ( Fig 6A ) . Western blotting revealed a decrease in MECOM and GATA2 , and an increase in c-Jun levels in the induction culture of the Lenti-miR-22-infected HSPCs ( Fig 6B , left ) . Conversely , Lenti-ZIP-miR-22 infection caused increased MECOM and GATA2 levels and decreased c-Jun levels ( Fig 6B , right ) . We also examined the effect of MECOM on monocyte/macrophage differentiation of HSPCs . Flow cytometry demonstrated that knockdown of MECOM by Lenti-shMECOM in the HSPCs increased percentages of CD14-positive cells ( Fig 6C ) . Western blot analysis revealed a reduced GATA2 levels but increased c-Jun levels ( Fig 6D ) in the induction culture of the Lenti-shMECOM-infected HSPCs . These results confirmed that the mechanism by which miR-22 promotes monocyte/macrophage differentiation of HSPCs is identical to that in the cell lines . We performed Taqman real-time PCR to detect MECOM mRNA expression [40] in PBMNCs derived from 40 AML patients and 43 healthy donors . Significantly higher MECOM mRNA levels were detected in the AML patients compared to the healthy donors ( Fig 6E , left ) , while miR-22 levels were much lower in the same AML samples compared to the healthy donors ( Fig 6E , middle ) . Moreover , miR-22 expression was conversely associated with MECOM mRNA expression in the tested projects ( Fig 6E , right ) . As PU . 1 has been proved to regulate miR-22 expression , we questioned whether the down-regulation of miR-22 was related to PU . 1 expression in AML . We examined their expression in PBMNCs derived from 42 AML patients and 39 healthy donors and found decreased PU . 1 levels in AML patients ( Fig 6F , left ) . Moreover , the PU . 1 levels were positively associated with miR-22 levels in the tested projects ( Fig 6F , right ) . Collectively , these results at least partially confirm PU . 1-miR-22-MECOM regulation in AML development . Since a remarkable decrease of miR-22 was observed in AML patients , and since myeloid differentiation blockage is one of the key characterizations in AML , we examined whether reintroduction of miR-22 could relieve the differentiation blockage . The BM CD34+ HSPC samples derived from seven AML patients were infected with Lenti-miR-22 or Lenti-Con and subjected to monocyte/macrophage induction . Flow cytometry demonstrated that Lenti-miR-22 infection significantly improved the differentiation of HSPCs from all seven patients ( Fig 7A and S4A Fig ) . May-Grünwald Giemsa staining also showed that Lenti-miR-22 infection improved monocyte/macrophage development of the AML HSPCs ( Fig 7B and S4B Fig ) . Analysis with qRT-PCR confirmed miR-22 overpresence in the Lenti-miR-22-infected cells ( Fig 7C and S4C Fig ) . Western blot analysis displayed significantly decreased MECOM and GATA2 levels and increased c-Jun levels in the induction cultures of Lenti-miR-22-infected-AML HSPCs as compared with the control infection samples ( Fig 7D ) . These results demonstrated that the reintroduction of miR-22 could partially relieve differentiation blockage in AML BM blasts . We also examined the effects of miR-22 on the growth of HL60 and THP1 cells , and found that miR-22 significantly inhibited cell growth ( S5 Fig ) . Following this observation , we further demonstrated that lentivirus-mediated miR-22 reintroduction inhibited cell growth during the monocyte/macrophage induction culture of AML BM CD34+ HSPCs ( Fig 7E ) .
PU . 1 has been shown to play a decisive role in lympho-myeloid development and its stage-specific expression is critical to prevent leukemic transformation [39 , 40] . Other studies have revealed that monocyte/macrophage development from hematopoietic stem cells requires PU . 1-coordinated miRNA expression [41 , 42] . It was also reported that miR-22 was transcriptionally regulated by P53 and c-Myc [43–45] . However , the TSS of miR-22 has not been confirmed . In this paper , we identified the TSS and showed that PU . 1 activates miR-22 transcription by directly binding to the miR-22 promoter . MiR-22 has been reported to play an important role in several physiologic processes and cancers , and several target genes of miR-22 have been identified in different cell types [16–23 , 30 , 46 , 50 , 51] . Here , we demonstrated that miR-22 is a positive regulator and MECOM a negative regulator in monocyte/macrophage development . We also showed that miR-22 promotes the differentiation via targeting and downregulating MECOM mRNA , at least partially . MECOM was reported to impair the function of PU . 1 by competing with c-Jun , a critical coactivator of PU . 1 [47] Similarly , GATA2 , which can be transcriptionally activated by MECOM [27] , is able to interfere with the interaction between c-Jun and PU . 1 [37] . In addition , other studies have revealed that MECOM blocks JNK-dependent phosphorylation of c-Jun [38] , thus reducing p-c-Jun levels , which can form heterogeneous or homogeneous AP-1 to activate c-Jun transcription [39] . In this study , we found that a decrease in miR-22-mediated MECOM resulted in increased c-Jun-PU . 1 protein complexes via increasing c-Jun levels and by relieving MECOM- and GATA2-mediated interference in the interaction between c-Jun and PU . 1 , which promotes monocyte/macrophage differentiation . In the present study , we also detected abnormally decreased expression of miR-22 in de novo AML patients , suggesting that it acts as a tumor suppressor in AML development . Additionally , we detected a negative association between MECOM mRNA and miR-22 expression and a positive association between miR-22 and PU . 1 expression in AML patients , which suggests that PU . 1-miR-22-MECOM regulation is involved in AML development . According to the above results , we summarized molecular models underlying miR-22’s involvement in monocyte/macrophage differentiation regulation ( Fig 8A ) and AML development ( Fig 8B ) . Until now , there have been two published leukemia/miR-22-related reports with opposite conclusions . Song et al . reported that miR-22 is an oncogenic miRNA and is abnormally upregulated in myelodysplastic syndrome ( MDS ) and MDS—derived leukemia [50]; they also showed that miR-22 transgenic mice developed MDS and hematological malignancies [50] . Jiang et al . reported that miR-22 plays an anti-tumor role and is abnormally downregulated in de novo AML [51] , which is consistent with our results . Mechanistically , Song et al . reported that miR-22 regulated methylation status via targeting TET2 mRNA [50] , while Jiang et al . reported that TET1 could repress miR-22 transcription , and that miR-22 targets multiple oncogenes , including CRTC1 , FLT3 and MYCBP , and thus repressing the CREB and MYC pathways [51] . Our present paper demonstrates that miR-22 is transcriptionally activated by PU . 1 , and can enhance PU . 1–c-Jun interaction by targeting MECOM and thus affecting GATA2 and c-Jun levels . These findings illustrate how miR-22 and the transcription factors MECOM , GATA2 , c-Jun , and PU . 1 are orchestrated in normal monocyte/macrophage differentiation regulation and AML development . Using oncogenes-transformed mouse models , Jiang et al . demonstrated miR-22’s therapeutic potential in AML . Using BM CD34+ cells obtained from AML patients , our present paper shows that the reintroduction of miR-22 could relieve the differentiation blockage and inhibit the growth of AML BM blasts , which also suggests its potential in AML therapy . High MECOM expression defines a subgroup of AML with a poor prognosis [38 , 48 , 49] . We found that the reintroduction of miR-22 significantly improved monocyte/macrophage differentiation in the patients with either high or low MECOM expression . Interestingly , it seems that Lenti-miR-22 infection improved differentiation better in MECOMhigh patients than in MECOMlow patients ( 18 . 63 ± 5 . 41% vs . 9 . 13 ± 2 . 39% , p = 0 . 002 , see S6 Fig ) ; however this finding needs further demonstration . In conclusion , our data revealed new function and mechanism of miR-22 in human monocyte/macrophage differentiation and AML development , and demonstrated its potential value in AML diagnosis and therapy .
Human UCB was obtained from normal , full-term deliveries from Beijing Hospital . The PB and BM samples of AML patients and normal volunteers were obtained from the 303 hospital and the 307 Hospital according to the protocols approved by the Ethics Committees of the Institutional review Board of Institute of Basic Medical Sciences , Chinese Academy of Medical Sciences . The informed consent was obtained from all of the examined subjects . The human promyelocytic cell line HL60 was maintained in IMDM ( Gibco-BRL , Paisley , UK ) containing 2 mM glutamine , 25 mM HEPES , 1 . 5g/L sodium bicarbonate , 50 U/mL penicillin and 50 μg/mL streptomycin ( Sigma , St . Louis , MO , USA ) , supplemented with 10% FCS ( PAA , Pashing , Austria ) , at 37°C in 5% CO2 . Acute monocytic leukemia cell line THP-1 was maintained in RPMI-1640 medium ( Gibco-BRL ) containing 2 mM glutamine , 25 mM HEPES , 1 . 5 g/L sodium bicarbonate , 50 U/mL penicillin , and 50 μg/mL streptomycin ( Sigma ) , supplemented with 10% fetal bovine serum ( FBS ) , at 37°C in 5% CO2 . Lentivirus packaging cell line 293TN was cultured in DMEM medium ( Gibco-BRL ) , supplemented with 10% FBS . For monocytic/macrophagic induction , PMA ( Sigma ) was added to a final concentration of 16 nM . Total RNA was isolated from the cell harvest using Trizol ( Invitrogen , CA , USA ) according to the manufacturer’s instructions . One μg of total RNA was used to generate cDNA by M-MLV reverse transcriptase ( Invitrogen ) . Stem-poop RT primers were used for the reverse transcription of miRNAs , and Oligo ( dT ) 18 was used for the reverse transcription of mRNAs . qRT-PCR was carried out in the Bio-Rad IQ5 real-time PCR system ( Bio-rad , CA , USA ) or in the ABI PRISM 7900HT Sequence Detection System ( Applied Biosystem , CA , USA ) according to the manufacturer’s instructions . Each qRT-PCR assay was performed in triplicate . The data were normalized using the endogenous GAPDH mRNA or U6 snRNA . The primers for reverse transcription of miRNAs and qRT-PCR as well as the Taqman probes are described in S2 Table . Human CD34+ cells that contain HSPCs were collected using a human CD34 MicroBead Kit ( Miltenyi Biotec , Cologne , Germany ) from MNCs isolated from UCB , PB or BM by percoll density gradient ( d = 1 . 077 ) ( Amersham Biotech , Little Chalfont , UK ) . The CD34+ cells were cultured in IMDM ( Gibco-BRL , Paisley , UK ) with 30% FBS , 1% bovine serum albumin , 2 mM L-glutamine , 0 . 05 mM 2-mercaptoethanol , 50 U/ml penicillin , 50 μg/ml streptomycin , 50 ng/ml stem cell factor and 20 ng/ml IL-3 . To induce monocyte/macrophage differentiation , a cytokine cocktail of 50 ng/ml M-SCF , 1 ng/ml IL-6 and 100 ng/ml Flt-3 L was used . All of these cytokines were purchased from Peprotech ( Rocky Hill , NJ , USA ) . The total RNA isolated from THP1 cells treated with PMA for 72 hours was used and RACE was performed using a 5’-Full RACE kit with TAP ( Takara , Dalian , China ) . Primer sequences are listed in S2 Table . Twenty μg of denatured total RNA was loaded onto a 15% polyacrylamide TBE gel and separated in a 1 X TBE running buffer , followed by transfer onto a N+ membrane ( Amersham , London , UK ) at 200 mA for two hours in an electro-transferring system and crosslinking under ultraviolet radiation for 150 seconds . The miRNA-specific oligo was 5’ end labelled with γ-32P-ATP through T4 polynucleotide kinase ( Takara ) , according to the manufacturer’s protocol . The oligo probes were designed based on individual miRNA sequence information deposited in miRBase ( http://microrna . sanger . ac . uk ) . An antisense oligo of U6 snRNA was used to detect U6 snRNA from each sample as a loading control . After prehybridisation using hybridizing buffer ( BioDev , BJ , China ) , blots were hybridized with 32P-labelled DNA probes ( 2 μmol/ml ) overnight at 37°C . After washing , the hybridized membranes were exposed to Kodak X-omat BT film . miR-22 mimics , anti-miR-22 ( miR-22 inhibitor ) , small interference RNAs ( siPU . 1 ) and negative controls ( scramble control , inhibitor control and siRNA control ) were purchased from Dharmacon ( IL , USA ) . Small interference RNAs ( siMECOM ) were purchased from Origene ( MD , USA ) . These oligonucleotides were transfected into HL-60 and THP1 cells using a DharmaFECT1 reagent ( Dharmacon ) at a final concentration of 100 nM . Total proteins were extracted from cells or tissues using a RIPA buffer ( 50 mM Tris-HCl , pH 7 . 4 , 150 mM NaCl , 1 mM EDTA , 1% Triton X-100 , 1% sodium deoxycholate , 0 . 1% SDS ) supplemented with 1 mM PMSF , 5 μg/ml aprotinin and 5 μg/ml leupeptin . The protein concentration was determined with a BCA Protein Assay Kit ( Vigorous , China ) . The total protein ( 15–30 μg ) was loaded onto a 10% SDS-PAGE gel , probed with mouse or rabbit mAb against MECOM ( Epitomics , MA , USA ) , GATA2 ( Proteintech , IL , USA ) , p-c-Jun ( Bioworld Technology , St . Louis , USA ) , c-Jun ( Bioworld Technology ) , PU . 1 ( Cell Signaling Technology ) and GAPDH ( Proteintech ) followed by horseradish peroxidase-conjugated sheep anti-mouse or rabbit Ig ( ZSGB-BIO ) . GAPDH was detected as a loading control . The harvested cells were washed twice with PBS , resuspended in 100 μl cold PBS and stained with PE- or APC-conjugated anti-CD14 ( eBioscience , San Diego , CA , USA ) for 30 minutes at 4°C in the dark . Finally , Stained cells were washed using cold PBS and analyzed on an Accuri C6 flow cytometer ( Becton Dickinson Biosciences , San Jose , CA , USA ) . Usually , in an identical experiment group , the fluorescence gate was set based on untreated cells which endured neither DNA delivery treatment nor induction , and at the place where the percentage of differentiated cells in untreated cells was less than 1% and fixed . The unstained cells , which endured the same treatment except for antibody staining , were used for controls to exclude the effect of the background fluorescence of the cells caused by treatment . Cells were re-seeded into 96-well plates at a density of 10000 cells per well after transfection or infection for 12 hours . Cell viability was measured every 24 hours for HL60 and THP1 cells or every three days for AML BM CD34+ cells by adding 10% CCK-8 ( DOJINDO , Japan ) and then incubating at 37°C for three hours . The optical density was read at 450 nm with a microplate spectrophotometer . Each experiment was carried out three times . The 3′UTR of MECOM containing the predicted binding site of miR-22 was amplified from human cDNA by PCR using the primers described in S2 Table , and then inserted into the pMIR-REPORT luciferase reporter vector ( MECOM_WT ) . MECOM_Mut contained the sequences with mutations in the binding site . Fragments containing wild-type PB2 ( TTCCTC ) were amplified from genomic DNA by PCR using the primers described in S2 Table and then cloned into pGL3-basic to get pGL3-PB2_WT , while those containing mutant PB2 ( CATGAG ) were also cloned into pGL3-basic to get pGL3-PB2_Mut . The recombinant construct ( pGL3-PB2_WT or pGL3-PB2_Mut ) , pRL-TK and miR-22 mimic or pcDNA3 . 1-PU . 1 and related controls were co-transfected into HEK293T cells using Lipofectamine 2000 ( Invitrogen , CA , USA ) . The plasmid pRL-TK containing Renilla luciferase was used as an internal control . The cells were harvested after transfection for 48 hours and the luciferase activity was measured using a Dual Luciferase Assay System ( Promega , WI , USA ) according to the manufacturer’s instructions . Data were obtained by normalization of firefly luciferase activity to renilla luciferase activity . All transfection assays were performed three times . CD34+ HSPCs infected with Lenti-miR-22 or Lenti-Con were cultured in a 6-well plate with human methylcellulose medium without EPO ( R&D , SystemsGmbH , MN , USA ) according to the manufacturer’s instructions . After 12 days of incubation at 37°C in a 5% CO2 incubator , colony-forming unit granulocyte/macrophages ( CFU-GM ) and colony-forming unit-macrophages ( CFU-M ) were analyzed and quantified using Eclipse TS100 phase-contrast microscopy ( Nikon , Tokyo , Japan ) . The HL60 , THP1 or CD34+ HSPCs induced monocyte/macrophage differentiation were harvested at the indicated time and stained with May-Grünwald for 5 min and Giemsa for 30 minutes . Then , the cell smears were washed with water , air-dried , and observed under Olympus BX51 optical microscopy ( Olympus , Tokyo , Japan ) . Cells ( 2 x 107 ) were treated with 1% formaldehyde in a medium for 10 minutes at 37°C , followed by addition of glycine ( final concentration , 0 . 125 M ) . After washing with PBS , cells were lysed on ice and sonicated to obtain 500–1000-bp sheared chromatin fragments . Subsequent ChIP steps were performed according to the protocols from Upstate Biotechnology ( Charlottesville , VA , USA ) . Each reaction included 2 μg anti-PU . 1 ( Cell Signaling Technology , Beverly , MA , USA ) ; anti-IgG ( Santa Cruz Biotechnology , Santa Cruz , CA , USA ) served as the unspecific control . The presence of target DNA sequences was detected by PCR and qRT-PCR . PCR products were resolved by 2% agarose gel electrophoresis . qRT-PCR analysis of fragments containing validated PU . 1-binding site , the positive control ( pro-CSF1R ) and the negative control ( UR ) were carried out three times with the primers listed in S2 Table . The relative occupancy of the immunoprecipitated factor at a locus is examined via the comparative threshold method [52] . For every promoter studied , a ΔCt value was calculated for each sample by subtracting the Ct value for the input ( to account for differences in amplification efficiencies and DNA quantities before immunoprecipitation ) from the Ct value obtained for the immunoprecipitated sample . A ΔΔCt value was then calculated by subtracting the ΔCt value for the sample immunoprecipitated with PU . 1 antiserum from the ΔCt value for the corresponding control sample immunoprecipitated with normal rabbit serum . Fold differences ( PU . 1 ChIP relative to control ChIP ) were then determined by raising 2 to the ΔΔCt power . The equation used in these calculations is summarized as fold difference ( PU . 1 ChIP relative to control ChIP ) = 2[Ct ( control ) —Ct ( PU . 1 ) ] , where Ct = Ct ( immunoprecipitated sample ) –Ct ( input ) . THP1 cells were infected with a lentivirus overexpressing miR-22 or GFP control , then treated with PMA for 48 hours for co-immunoprecipitation . The Dynabeads Protein G ( Invitrogen ) was incubated with anti-PU . 1 antibody ( Santa Cruz Biotechnology ) or anti-c-Jun antibody ( Santa Cruz Biotechnology ) or IgG ( Santa Cruz Biotechnology ) in antibody binding and washing buffer at room temperature with a 20-minutes rotation . The Dynabeads-antibody complexes were washed one time using antibody binding and washing buffer then incubated with the whole cell lysates at 4°C overnight . For Western blot analysis , the Dynabeads-antibody-antigen complexes were washed four times with washing buffer , and the proteins were separated by SDS-PAGE . For construction of the recombinant lentiviruses that expresses specific shRNAs against MECOM or PU . 1 , the targeted sequences ( see S2 Table ) were synthesized and inserted into the pLentiLox 3 . 7-RNAi plasmid ( Invitrogen ) following the manufacturer’s protocols . For construction of the recombinant lentivirus that expresses miR-22 , a 300-bp DNA fragment containing the miR-22 precursor was amplified and inserted into pMiRNA1 vector . The miRZip lentivector construct expressing miRZip shRNAs targeting miR-22 ( Lenti-ZIP-miR-22 ) was purchased from SBI ( Mountain View , CA , USA ) . The virus packaging was performed using a packaging kit from SBI ( Mountain View ) according to the manufacturer’s instructions . The lentivirus particles ( Lenti-miR-22 , Lenti-Con , Lenti-ZIP-miR-22 , Lenti-ZIP-Con , shMECOM , shPU . 1 , shCon ) were harvested and concentrated using PEG-it Virus Precipitation Solution ( SBI ) . The lentiviral particles were added into the THP1 cells or CD34+ cells in the presence of Polybrene ( 5 μg/mL; Sigma , St . Louis , MO , USA ) . The cells were washed with PBS 24 hours after infection and exposed to lineage-specific differentiation cultures or plated for colony-forming assay . A Student’s t-test ( two-tailed ) was performed to analyze the data . The correlation between miR-22 and MECOM mRNA as well as between miR-22 and PU . 1 mRNA was examined by Pearson correlation analysis . P-values <0 . 05 were considered to be significant . | We found that miR-22 is transcriptionally activated by PU . 1 during monocyte/macrophage differentiation and miR-22 promotes the differentiation via targeting MECOM ( EVI1 ) mRNA and further increasing interaction between c-Jun and PU . 1 . We also show that miR-22 is a tumor repressor and that PU . 1-miR-22-MECOM regulation is involved in AML development; moreover , we demonstrate that reintroduction of miR-22 relieves the differentiation blockage and inhibits the growth of AML bone marrow blasts . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"acute",
"myeloid",
"leukemia",
"flow",
"cytometry",
"transfection",
"leukemias",
"chromosome",
"staining",
"medicine",
"and",
"health",
"sciences",
"luciferase",
"myeloid",
"leukemia",
"gene",
"regulation",
"enzymes",
"messenger",
"rna",
"cancers",
"and",
"neoplasms",
... | 2016 | The PU.1-Modulated MicroRNA-22 Is a Regulator of Monocyte/Macrophage Differentiation and Acute Myeloid Leukemia |
Most known thioredoxin-type proteins ( Trx ) participate in redox pathways , using two highly conserved cysteine residues to catalyze thiol-disulfide exchange reactions . Here we demonstrate that the so far unexplored Trx2 from African trypanosomes ( Trypanosoma brucei ) lacks protein disulfide reductase activity but functions as an effective temperature-activated and redox-regulated chaperone . Immunofluorescence microscopy and fractionated cell lysis revealed that Trx2 is located in the mitochondrion of the parasite . RNA-interference and gene knock-out approaches showed that depletion of Trx2 impairs growth of both mammalian bloodstream and insect stage procyclic parasites . Procyclic cells lacking Trx2 stop proliferation under standard culture conditions at 27°C and are unable to survive prolonged exposure to 37°C , indicating that Trx2 plays a vital role that becomes augmented under heat stress . Moreover , we found that Trx2 contributes to the in vivo infectivity of T . brucei . Remarkably , a Trx2 version , in which all five cysteines were replaced by serine residues , complements for the wildtype protein in conditional knock-out cells and confers parasite infectivity in the mouse model . Characterization of the recombinant protein revealed that Trx2 can coordinate an iron sulfur cluster and is highly sensitive towards spontaneous oxidation . Moreover , we discovered that both wildtype and mutant Trx2 protect other proteins against thermal aggregation and preserve their ability to refold upon return to non-stress conditions . Activation of the chaperone function of Trx2 appears to be triggered by temperature-mediated structural changes and inhibited by oxidative disulfide bond formation . Our studies indicate that Trx2 acts as a novel chaperone in the unique single mitochondrion of T . brucei and reveal a new perspective regarding the physiological function of thioredoxin-type proteins in trypanosomes .
Trypanosomatids , the causative agents of human sleeping sickness and Nagana cattle disease in Africa ( Trypanosoma brucei species ) , South-American Chagas’ disease ( T . cruzi ) and the different forms of leishmaniasis ( Leishmania species ) , possess an unusual thiol redox metabolism . While most organisms rely on the glutathione/glutathione reductase and thioredoxin ( Trx ) /thioredoxin reductase systems to maintain cellular redox homeostasis , the main low molecular mass thiol in trypanosomatids is trypanothione , bis ( glutathionyl ) spermidine; T ( SH ) 2 [1–5] . The dithiol is kept in the reduced form by the NADPH-dependent trypanothione reductase which catalyzes the only reaction directly connecting the thiol and dinucleotide redox metabolism in these parasitic protozoa . The T ( SH ) 2 system provides reducing equivalents for vital processes such as the synthesis of DNA precursors and the detoxification of hydroperoxides by different types of tryparedoxin peroxidases [6–9] . An important mediator for T ( SH ) 2-dependent reactions is tryparedoxin ( Tpx ) , an essential and parasite-specific distant member of the thioredoxin ( Trx ) -type protein family [10] . In addition to Tpx , African trypanosomes possess conventional small oxidoreductases such as glutaredoxins [11–13] and Trxs [14] . Trx-type proteins are ubiquitous in nature . They are characterized by a common structural fold and a catalytic CXXC motif located on the surface of the protein [15 , 16] . Trx was originally isolated from Escherichia coli as an electron donor for ribonucleotide reductase [17] . However , members of the Trx protein family have since been shown to exhibit multiple physiological functions . In addition to their role as general thiol-disulfide oxidoreductases , they can function in transcription factor regulation , protein binding , protein folding facilitation and chaperone-type activities [16] . So far , only one Trx has been characterized in T . brucei . This Trx1 ( TriTrypDB gene ID Tb927 . 9 . 3370 ) has the canonical WCGPC motif of classical Trxs and catalyzes the electron transfer from T ( SH ) 2 onto the parasite ribonucleotide reductase as well as tryparedoxin peroxidases but with much lower efficiency compared to Tpx [18] . Since thioredoxin reductases are missing in trypanosomatids , Trx1 is most likely kept reduced by thiol-disulfide exchange with T ( SH ) 2 . The protein does not seem to be essential at least under standard culture conditions and may be functionally replaced by Tpx [14] . African trypanosomes are obligate free-living parasites that multiply as bloodstream ( BS ) form in the blood and body fluids of their mammalian hosts and as procyclic ( PC ) insect form in the tsetse fly vector . BS T . brucei harbor a mitochondrion that lacks a cytochrome-containing electron transport chain and rely exclusively on glycolysis for energy supply . In contrast , in the PC stage , the single mitochondrion is fully elaborated and the parasites gain ATP via oxidative phosphorylation . All enzymes involved in the biosynthesis of T ( SH ) 2 as well as trypanothione reductase and Tpx are located in the cytosol . The nature of the thiol metabolism in the mitochondrion of trypanosomatids is still largely unknown [19] . The primary aim of this work was to identify the missing oxidoreductase in the redox metabolism of the mitochondrion in African trypanosomes . In addition to the gene for Trx1 , the genome of T . brucei encodes another putative Trx ( TriTrypDB gene ID Tb927 . 3 . 4240 ) . This previously uncharacterized Trx2 is larger than classical Trxs with a predicted molecular mass of 21 . 8 kDa and its amino acid sequence contains five cysteine residues , two of which form a CKPC motif . The protein is only distantly related to T . brucei Trx1 and other classical Trxs and does not appear to have any counterpart outside trypanosomatids . Different prediction algorithms ( see below ) suggest a mitochondrial localisation and a genome-wide RNA interference approach in T . brucei strongly suggests that the protein is essential [20] . Here we investigated the molecular properties of Trx2 both in vitro and in vivo . We show that Trx2 is located in the mitochondrion of BS and PC T . brucei . Depletion of Trx2 results in a proliferation defect in both parasite stages and PC cells virtually lacking a Trx2 species are unable to sustain long-term exposure to 37°C . However , a mutant in which all five cysteines were replaced by serine residues ( 5S-Trx2 ) can replace the authentic protein both in vitro and in vivo in the mouse model . Remarkably , recombinant Trx2 and 5S-Trx2 lack insulin reductase activity but , instead , slow down the formation of insulin aggregates . Reduced Trx2 and the 5S-Trx2 mutant prevent the aggregation of thermally unfolding proteins and preserve the folding competent state of client proteins . Activation of the chaperone function appears to be induced by conformational changes at elevated temperatures . This structural reorganization is impaired by disulfide bond formation , keeping Trx2 in a chaperone-inactive state . Taken together , we conclude that Trx2 plays a vital , redox-regulated , role in the mitochondrion of T . brucei which becomes even more important when the parasites face a long-term heat stress .
The protein sequence of Trx2 as deduced from the genome of the T . brucei 427 strain comprises 200 amino acid residues . It is identical to the T . b . gambiense ortholog and differs in just two and one position ( s ) from the T . b . brucei 927 and T . evansi sequences , respectively ( S1A Fig ) . With only 32% overall identity , the T . brucei and T . cruzi Trx2 sequences are barely related , in striking contrast to the 74% identical Trx1 sequences from these parasites ( S1B Fig ) . The situation in Leishmania is not clear . For L . major and L . braziliensis , proteins with more than 450 residues are annotated as orthologs/paralogs , whereas in L . infantum and L . donovani proteins with only 112 residues , corresponding to the C-terminal part of T . brucei Trx2 , are described as putative Trxs . None of the proteins has been studied so far and future work is required , especially because the overall genomic locus is highly conserved between the various Leishmania species . T . brucei Trx2 has five cysteine residues of which Cys63 and Cys66 form a CXXC motif , one of the hallmarks of Trx-type proteins [16] . The long putative Leishmania proteins—but not the putative T . cruzi ortholog—display the motif as well . The other cysteine residues are conserved only in the proteins from African trypanosomes . T . brucei Trx2 is considerably longer than classical Trxs and shares only 18% of all residues with T . brucei Trx1 which is a classical Trx and more closely related to human Trx1 [21] ( S1B Fig ) . Structural modeling of T . brucei Trx2 ( https://swissmodel . expasy . org/interactive ) revealed that the protein may adopt a Trx-fold with a long central insertion ( see the alignment in S1A Fig ) and with the CXXC motif correctly placed at the N-terminus of an α-helix . A Blast-P search with T . brucei Trx2 against all non-redundant GenBank CDS translations retrieved exclusively Trx-like proteins but no glutaredoxins or tryparedoxins . The top hits belonged to Kinetoplastid Trx2-homologues ( S1C Fig ) . These sequences together with those from canonical Trx proteins were subjected to multiple sequence alignment and phylogenetic tree analysis ( S1D Fig ) . The phylogenetic analyses strongly suggest that the Trx2-like sequences from Kinetoplastida form an out-group with respect to canonical ( group I ) and non-canonical ( group II ) Trx-like proteins . Table 1 summarizes the cells used in this work to evaluate the physiological role of Trx2 in BS and PC T . brucei . The molecular biology procedures are described in the Materials and methods section . Several attempts to generate homozygous trx2 knockout ( KO ) cell lines were unsuccessful . PCR analysis of BS and PC resistant to both selection markers revealed that all clones had retained a trx2 copy ( S2A and S2B Fig ) which strongly suggested that Trx2 is essential in both developmental stages . Therefore , we generated cKO cell lines that expressed a Tet-inducible ectopic copy of WT-Trx2 ( S2C Fig ) . To unveil if the parasites require a redox-active form of Trx2 , we produced cKO cell lines that expressed mutants in which Cys63 and Cys66 of the CXXC motif ( C63/66S-Trx2 ) or all five cysteines ( 5S-Trx2 ) were replaced by serine residues ( S2D Fig ) . The prediction algorithms MitoProt II , PredSL , PROlocalizer , TargetP1 . 1 and PSORT 2 suggest a mitochondrial localisation for T . brucei Trx2 . To experimentally verify the subcellular localisation , BS 2T1 cells expressing WT-Trx2 and PC WT-Trx2 cKO ( Table 1 ) were cultured in the presence of Tet and subjected to immunofluorescence microscopy ( Fig 1A and 1B ) . The myc signals displayed a punctate pattern overlaying with the Mitotracker signal . The mitochondrial localisation of Trx2 was independently confirmed by fractionated cell lysis ( Fig 1E ) and is in accordance with recent proteome analyses [22 , 23] . To detect the authentic protein in the parasites , antibodies against recombinant Trx2 were raised . Since the prediction algorithms described above did not provide a clear cleavage site for the mitochondrial pre-sequence , we decided to generate a recombinant Trx2 species with the 32 N-terminal residues cleaved off and an Arg at the -2 position , a characteristic of the cleavage sites of mitochondrial processing peptidase ( MPP [24 , 25] ) ( S1A Fig ) . The molecular mass of recombinant Trx2 derived from Western blot analysis ( 18 . 55 ± 0 . 4 kDa ) was in agreement with its calculated mass of 18 , 467 Da ( -32 residues , plus an additional N-terminal GAMG stretch due to the TEV-cleavage ) . In cell lysates of T . brucei , the antibodies consistently detected a doublet of bands that migrated slightly above and below the recombinant protein ( Fig 1C ) . Both bands were diminished/absent when the cells were subjected to Trx2 mRNA depletion and re-appeared when the cells had lost the RNAi regulation ( see below ) . The apparent average molecular masses of the bands were 20 . 1 and 17 . 5 kDa ( Fig 1D ) . As full-length Trx2 has a theoretical mass of 21 . 82 kDa and the recombinant full-length protein runs at a much higher mass than the truncated Trx2 ( see S6B Fig ) , we concluded that both species detected in the cells are processed forms . To evaluate the intra-mitochondrial localisation of Trx2 in more detail , PC WT-Trx2 cKO cells were treated with increasing concentrations of digitonin , which results in the gradual permeabilization of cellular membranes [26] . The pellet and supernatant fractions were subjected to Western blot analysis . Antibodies against Trx2 or the myc-tag revealed a prominent band and a very weak band with slightly higher mass ( Fig 1E ) . Up to the highest digitonin:protein ratio ( 0 . 55:1 mg/mg ) applied , both Trx2 bands remained largely in the pellet fraction . The same behavior was observed for the mitochondrial 2-Cys-peroxiredoxin ( mPrx ) , used as marker protein for the mitochondrial matrix . At digitonin:protein ratios of ≥ 0 . 35:1 , Trx2 and mPrx became only partially solubilized whereas the intermembrane space proteins Grx2 and Cyt c as well the cytosolic Tpx were virtually completely in the supernatant fractions . Thus , we conclude that the two Trx2 bands detected in the Western blots represent the partially and fully processed protein in the mitochondrial matrix . To estimate the cellular concentration of Trx2 , total lysates of PC T . brucei and different amounts of the recombinant protein were studied by comparative Western blot analyses . Based on the data from seven independent analyses and using 96 fl as volume of PC cells [27] , we calculated a cellular concentration of Trx2 of 131 ± 49 nM . The mitochondrion of PC T . brucei has been reported to occupy about one fourth of the total cell volume [28] . Thus , the local concentration of Trx2 in the mitochondrion would be in the order of 0 . 5 μM . A genome-wide RNAi library screen suggested that Trx2 is essential for T . brucei [20] . BS Trx2 RNAi cells ( Table 1 ) as well as WT parasites were cultured in the presence or absence of Tet . Whereas proliferation of WT parasites was unaffected by Tet , the induced Trx2 RNAi cell lines displayed a minor but significant growth retardation between day 4 and 7 , when compared to the respective non-induced cells ( Fig 2A ) . For example , between day 5 and 6 , WT parasites , non-induced and induced RNAi cells multiplied by a factor of 25 , 20 and 12 , respectively . Western blot analyses of cells harvested after five days of Tet exposure revealed a partial depletion of the protein . Apparently , the remaining comparably low level of Trx2 still allowed the parasites to proliferate . Cultivation of BS WT-Trx2 cKO cells ( Table 1 ) in Tet-free medium for seven days did not affect proliferation ( Fig 2B ) . Western blot analyses confirmed that the cells , albeit at significantly lower levels , still expressed the ectopic copy of Trx2 . As tightly regulated RNAi or WT-Trx2 cKO cell lines could not be obtained , we asked if the parasites require a redox-competent form of the protein . In the presence of Tet , C63/66S-Trx2 cKO cell lines ( Table 1 ) were viable and proliferative ( S3A Fig ) . In the absence of Tet , the cells showed a minor and only transient growth defect . Finally we generated BS cells that harbored solely a cysteine-free mutant of the Trx2 . Tet-removal from cultures of these 5S-Trx2 cKO cells ( Table 1 ) also resulted in only a very minor proliferation defect . Again low levels of the protein were still detectable after five days in Tet-free medium ( Fig 2C ) . In order to assess the biological relevance of Trx2 for parasite proliferation and survival in a mammalian host , mice ( five per group ) were infected with BS T . brucei cell lines that allowed the Tet-inducible downregulation of the endogenous protein ( Trx2 RNAi ) or expression of a myc-tagged ectopic copy of WT-Trx2 ( WT-Trx2 cKO ) or 5S-Trx2 ( 5S-Trx2 cKO ) in a Trx2 KO genetic background ( see Table 1 for the different cell lines ) . Animals infected with WT T . brucei served as controls . Parasitemia and animal survival were monitored ( S4 Fig and Fig 2D ) . In agreement with our earlier studies [27 , 29] , feeding the animals with water containing oxytetracycline ( 1 mg/ml , +Tet ) did not have any effect on the outcome of the infection for mice challenged with WT parasites ( S4 Fig and Fig 2D; p values >> 0 . 05 for parasitemia and survival ) . As observed in previous in vivo studies [13 , 29 , 30] , the parasitemia dropped to almost undetectable levels about one week post infection ( S4 Fig , day 6 ) , a phenomenon associated with the acute host immune response aimed to control infection [31] . The survival of +Tet mice infected with Trx2 RNAi parasites was significantly extended when compared to the animals where Trx2 expression was not silenced ( Fig 2D; p = 0 . 044 ) . In fact , animals from the non-induced RNAi group died between day 11 and 14 post infection , similar to those infected with WT parasites . In comparison , the fatal outcomes in mice of the RNAi +Tet group occurred between day 11 and 49 post infection . The median survival time was estimated to be 13 and 17 days for the non-induced and induced RNAi-group , respectively . Despite the prolonged survival , all mice from the +Tet RNAi group died . This can be ascribed to the stochastic appearance in vivo ( and also in vitro , Fig 3 ) of mutant parasites that are refractory to the RNAi silencing . The phenomenon is often observed for essential proteins in T . brucei [32–34] . Although statistically not significant , likely due to the low animal number from day 9 onwards , the median parasitemia was consistently lower in mice from the +Tet group compared to the non-induced RNAi group ( S4 Fig , day 9 , 11 , and 13 ) . As shown in Fig 2A , RNAi partially depleted Trx2 in BS cells which resulted in a minor but significant in vitro proliferation defect . Assuming a similar RNAi efficiency under in vivo conditions , low levels of Trx2 are probably sufficient for the parasite to establish host infection . Nonetheless , the sustained lower parasite load in +Tet mice ( S4 Fig ) suggests an impaired capacity of the pathogen to proliferate in the host when Trx2 is downregulated , and correlates well with the longer animal survival time exhibited by this group ( Fig 2D , left ) . As shown in Fig 2B and 2C , after 4 or 5 days of cultivation in the absence of Tet , the WT-Trx2 and 5S-Trx2 cKO cell lines still expressed Trx2 . Thus , as in the RNAi approach , this system did not allow to generate cells that lacked the protein completely . Yet , the cKO cell lines offered the possibility to address a potential redox role of Trx2 as a virulence or signaling factor in vivo . Both cKO cell lines were able to establish mouse infection , which displayed a similar progress independently whether the animals were fed with oxytetracycline to sustain the expression of the ectopic Trx2 species or not ( Fig 2D , right ) . Notably , under +Tet and -Tet conditions , the cKO cell lines displayed a significantly ( p values < 0 . 05 ) lower virulence ( median survival time from 17 to 19 days ) than WT parasites +Tet ( median survival time of 11 days ) or the non-/induced RNAi cell line ( 14 days ) . Maybe the genetic manipulations required to generate the cKO cells affect their in vivo virulence or the C-terminal tag slightly interferes with the physiological role ( s ) of the protein during in vivo infection . Animals infected with WT-Trx2 cKO cells displayed an identical median survival time of 19 days for the + and -Tet groups ( p value = 0 . 759 ) and no differences in parasitemia were observed ( S4 Fig ) . In contrast , mice infected with the 5S-Trx2 cKO cells and fed with oxytetracycline showed a mean survival time ( 17 days ) that was significantly shorter ( p value = 0 . 044 ) than that of the -Tet group ( 19 days ) . This could not be ascribed to differences in parasite load because statistical analysis did not yield significant results ( p value >> 0 . 05 ) . In summary , down-regulation of Trx2 impairs development of T . brucei in a mammalian host and thus shows that Trx2 is required for full infectivity under in vivo conditions . Strikingly , cKO parasites that exclusively expressed 5S-Trx2 were at least as infectious as those expressing an ectopic copy of WT-Trx2 . Thus , the in vivo Trx2 functions also do not involve thiol-redox reactions . Depletion of Trx2 in PC cell lines caused a significant growth defect ( Fig 3A ) . In Western blots , the protein was no longer detectable 48 h after RNAi induction . Upon prolonged cultivation , the PC Trx2 RNAi cell lines ( Table 1 ) resumed proliferation despite the presence of Tet which was accompanied by re-appearance of the protein . This phenomenon is often observed for essential T . brucei proteins and implies that the parasites managed to escape from the inducible RNAi system [33 , 35] . Cultivation of three distinct PC WT-Trx2 cKO cell lines ( Table 1 ) for four days in Tet-free medium did not affect proliferation ( Fig 3B ) . Also C63/66S-Trx2 cKO cell lines showed only a minor growth defect when transferred into medium without Tet ( S3B Fig ) and Western blot analyses of cells that were kept for five days in Tet-free medium still revealed the presence of the protein ( S3C Fig ) . Immunofluorescence microscopy showed that the ectopically expressed C63/66S-Trx2 was correctly targeted to the mitochondrion ( S3D Fig ) . Finally , we generated PC cell lines that harboured a cysteine-free mutant as sole Trx2 species ( Table 1 ) . Immunofluorescence microscopy confirmed the specific targeting of the protein into the mitochondrion ( Fig 3C ) . In contrast to the respective BS mutants , PC 5S-Trx2 cKO cells lines stopped proliferation about four days after Tet removal and at day 7 , the protein was almost undetectable ( Fig 3D ) . Upon continuous cultivation in Tet-free medium , the cells resumed normal growth . Reappearance of 5S-Trx2 at a level comparable to that in the constantly induced control cells at day 12 indicates that regulation of the Tet-inducible system was lost . In summary , our data show that Trx2 is required for the proliferation of both BS and PC T . brucei , and demonstrate that a Trx2 species that is devoid of any cysteine residue is sufficient to maintain cell viability and proliferation in vitro . In a transcriptome analysis of PC T . brucei that were treated for 1 or 3 h with 4 mM DTT in order to induce an ER stress response , the Trx2 mRNA level was reported to be 2-fold increased [36] . Due to the known communication between mitochondria and the ER in eukaryotic cells [37] , we evaluated if the higher transcript level correlates with an increased protein concentration . For this purpose , we treated WT parasites with 4 mM DTT and subjected the total lysates of stressed and non-stressed cells to Western blot analysis . No significant difference in the Trx2 protein level was observed ( S5A Fig ) . Subsequently , WT and induced 5S-Trx2 cKO cells ( Table 1 ) were treated with DTT and cell viability was followed . Under our experimental settings , PC T . brucei were highly sensitive towards DTT in accordance with a previous study [38] . In the presence of 150 μM DTT , the parasites started to die within 2 h , and after 8 h , less than 50% of cells were still viable ( S5B Fig ) . Under all conditions studied , WT and 5S-Trx2 cKO cells displayed identical sensitivity . Whereas in most organisms peroxiredoxins ( Prxs ) use Trxs as an electron source , the known trypanosomatid 2-Cys-peroxiredoxins are Tpx-dependent [19 , 39 , 40] . The parasite Prxs occur in the cytosol and mitochondrion , but Tpx appears to be restricted to the cytosol [39] . To get an insight if the mitochondrial Trx2 could play a role in hydroperoxide detoxification , we treated WT parasites and 5S-Trx2 cells with H2O2 . Both cell types showed a very similar behavior when exposed to short- or long-term peroxide stress ( S5D Fig ) . Paraquat is a redox-cycling compound that generates superoxide anions which are converted into hydrogen peroxide , and the mitochondrion is a main compartment of its mode of action [41 , 42] . At all time-points studied and both concentrations applied , paraquat affected the proliferation of WT parasites , WT-Trx2 cKO and 5S-Trx2 cKO cells ( Table 1 ) to the same degree ( S5D Fig ) . In addition , cKO cells that were grown for five days in the absence of Tet and thus had strongly reduced levels of the protein ( see Fig 3B and 3D ) , did not reveal an increased sensitivity . Taken together , Trx2 does not seem to play a role in the response of T . brucei to exogenous reductive or oxidative stresses . As PC cells displayed a clearer proliferation defect upon down-regulation of Trx2 and the Tet-inducible expression was more tightly regulated in PC 5S-Trx2 cKO clones than in BS cells , we decided to use the insect form with its fully elaborated mitochondrion to study a putative role of Trx2 in the heat response of the parasites . In the first approach , 5S-Trx2 cKO cells that were grown for five days in the absence of Tet as well as 5S-Trx2 cKO constantly cultured in the presence of Tet and WT parasites were exposed to a 41°C heat shock for 1 h , re-transferred to 27°C and cell viability/proliferation was followed . WT parasites and the 5S-Trx2 cKO cells +Tet readily recovered from the stress . One of the 5S-Trx2 cell lines -Tet depicted in Fig 4A , had a strong proliferation defect and the protein was practically undetectable . The other clone had already resumed growth in accordance with the reappearance of the protein . The 1 h 41°C treatment had no significant additional effect on the viability of both cell lines , strongly suggesting that Trx2 does not play a significant role in the recovery of procyclic cells from a short-term heat stress ( Fig 4A ) . Next , we monitored the long-term proliferation of WT-Trx2 cKO and 5S-Trx2 cKO cell lines and WT parasites at 37°C ( Fig 4B ) . Under standard conditions of 27°C , WT parasites multiplied by a factor of 10–13 every 24 h , and the induced WT-Trx2 cKO and 5S-Trx2 cKO cell lines by a factor of 8–9 . At 37°C , WT parasites ± Tet and the induced cKO cells multiplied about 4-fold within the first 24 h , and 2-3-fold within the next 24 h . After 48 h exposure to 37°C , proliferation of the cells stopped until the experiment was terminated after 120 h . The response of WT parasites ( ± Tet ) and the induced cKO cell lines to 37°C heat stress was practically identical . In contrast , 5S-Trx2 cKO cells that had been pre-cultured for 6 to 9 days in the absence of Tet ( and thus lacked virtually any Trx2 ( see Fig 3D ) were significantly more sensitive to heat stress when compared to the induced cells ( Fig 4B ) . Within 24 and 48 h , the cell density dropped to 43% and 20% of the starting cell number , and less than 2% of the cells were still living after 120 h . In contrast to the phenotype observed at 27°C , 5S-Trx2 cKO cells grown in the absence of Tet at 37°C did not resume growth , indicating that they were unable to overcome the Tet-regulation . The data showed that PC cells require Trx2 to withstand a prolonged 37°C heat stress but a cysteine-free form of Trx2 is able to replace the authentic protein . T . brucei Trx2 , C63/66S-Trx2 , and 5S-Trx2 , lacking the 32 N-terminal residues as well as full length Trx2 were expressed as TEV-cleavable fusion proteins in E . coli . The tag-free proteins were obtained by metal affinity chromatography as outlined in Materials and methods and their purity was verified by SDS-PAGE ( S6 Fig ) . Because the recombinant full-length protein was unstable and constantly precipitated it was not further characterized . Freshly prepared recombinant Trx2 yielded 0 . 5–1 . 5 thiols/protein in the DTNB assay , independent of the presence or absence of 8 M urea . After treatment with TCEP or DTT , five thiol groups were detected in accordance with all cysteine residues being accessible and present in reduced form . Upon storage in the absence of a reducing agent , free thiols virtually completely disappeared . This sensitivity to spontaneous oxidation could be visualized upon SDS-PAGE . Under reducing conditions , Trx2 migrated at about 19 kDa corresponding to its theoretical mass . Under non-reducing conditions , the protein showed several bands with higher masses and , remarkably , the monomer was shifted to an apparently lower mass indicating that Trx2 formed inter- and intramolecular disulfides ( S6C Fig ) . The C63/66S-Trx2 mutant showed a similar behavior suggesting that the cysteines that cause the shift of the oxidized monomer are distinct from those of the CXXC motif . As expected , the 5S-mutant migrated at the theoretical mass of the monomer , independent of the presence or absence of DTT . Recombinant Trx2 displayed a brownish color . The UV-visible spectrum revealed , in addition to the maximum at 280 nm , absorption peaks at 320 and 420 nm and resembled that of other Trx family proteins found to coordinate iron sulfur clusters [43–45] . The C63/66S-Trx2 variant had very low absorption at 320 and 420 nm which did not allow us to decide if the mutant is still able to coordinate a cluster . Clearly , 5S-Trx2 showed an absorption maximum at 278 nm but no absorption at higher wavelengths ( Fig 5A ) . To get a deeper insight into the putative iron sulfur cluster complex , freshly prepared Trx2 was subjected to gel filtration and the absorption recorded at 280 , 320 , and 420 nm . The 280 nm profile revealed peaks at about 26 , 45 , and 70 kDa ( Fig 5B , solid black line ) which probably represent monomeric , dimeric and oligomeric forms of Trx2 . The 26 kDa peak had absorption at 320 nm ( solid red line ) which might be due to some iron binding [46] . The absence of any absorption at 420 nm ( solid blue line ) indicates that the monomer does not bind an iron sulfur cluster . The 45 and 70 kDa peaks showed absorption at all three wavelengths suggesting that dimeric and oligomer forms of Trx2 can coordinate an iron sulfur cluster . When Trx2 was pre-reduced and the column run in the presence of DTT , the 280 nm profile displayed two broad peaks with maxima around 40 and 52 kDa ( black dashed line ) . Interestingly , the 320 and 420 nm maxima ( red and blue dashed lines ) did not overlay with the 280 nm maximum but remained virtually at the position observed under non-reducing conditions . The Trx2 species with bound iron sulfur cluster appears to be unaffected by DTT . Next , we compared the oligomeric state of Trx2 and 5S-Trx2 . Under non-reducing conditions , the elution pattern of Trx2 displayed peaks at 25 , 32 , 50 , and 70 kDa which are supposed to reflect the monomer with intramolecular disulfide , fully reduced monomer , various dimeric and oligomer forms , respectively ( Fig 5C , black solid line ) . The 45 kDa peak seen in Fig 5B was resolved into two peaks , maybe because an only minor fraction of Trx2 was still in complex with an iron sulfur cluster or because less protein was applied onto the column . When pretreated with DTT , Trx2 eluted in one prominent peak with an apparent mass of 32 kDa ( Fig 5C , black dashed lines ) . 5S-Trx2 eluted in a single peak independent of the presence or absence of DTT ( Fig 5C , red lines ) . Both 5S-Trx2 and reduced Trx2 displayed an apparent mass of about 32 kDa in comparison to their theoretical masses of 18 . 5 and 18 . 4 kDa , respectively . To verify that these species are indeed monomeric forms of the protein , 5S-Trx2 was subjected to size-exclusion chromatography with multi-angle light scattering ( SEC-MALS ) . The analysis yielded a molecular mass of 20 , 730 Da ( Fig 5D ) and confirmed that the 32 kDa peak contains monomeric 5S-Trx2 . The relative intensity of the Rayleigh ratio indicated the presence of a high molecular mass species . However , the differential refractive index ( dRI ) showed that the concentration of this species is low in comparison to that of the monomeric form . Taken together , our data revealed that recombinant T . brucei Trx2 can assume a variety of forms such as the monomeric protein with and without intramolecular disulfides , dimers between reduced and oxidized subunits as well as covalent dimers and finally dimers and oligomers with bound iron sulfur cluster . Most Trxs are known for their high protein disulfide reductase activity [15 , 16] . Insulin reduction is the most convenient assay to measure the activity in vitro . Cleavage of the inter-chain disulfide bridges results in the release of the insoluble free B-chain which can be monitored by an increase in turbidity at 650 nm [47] . We used T . brucei Tpx as positive control [26] . Indeed , presence of 0 . 21 μM Tpx strongly accelerated the reduction of insulin by DTT and hence the aggregation process ( Fig 6 ) . In contrast , the presence of up to 2 μM Trx2 had no effect on the B-chain aggregation , and higher Trx2 concentrations slowed down the process ( Fig 6A ) . These results strongly suggested that T . brucei Trx2 is not acting as protein disulfide reductase but instead might function as a molecular chaperone that protects the B-chain against irreversible aggregation . To get a deeper insight in the underlying mechanism , Tpx and Trx2 were treated with DTT and all cysteine residues were alkylated by NEM . As expected , NEM-treated Tpx completely lost its insulin reductase activity ( Fig 6B ) . The alkylated Trx2 , on the other hand , maintained its anti-aggregation activity , and seemed to be even more effective than the non-modified protein ( Fig 6B ) . To corroborate that the putative chaperone activity of Trx2 is really independent of the cysteine thiols , we studied insulin reduction in the presence of 5S-Trx2 . Indeed , this mutant was even more effective than NEM-treated WT Trx2 in slowing down the Tpx-catalyzed reaction and almost completely prevented the spontaneous turbidity development when present at 30 μM ( Fig 6C ) . Based on the observation that Trx2 , instead of stimulating the precipitation of the insoluble insulin B-chain , slowed down its aggregation ( Fig 6 ) , we wondered whether the protein might function as a molecular chaperone . To test this idea , we investigated the influence of Trx2 on the aggregation of thermally unfolding luciferase ( Fig 7 ) . Fully reduced ( Trx2red ) and oxidized ( Trx2ox ) protein was obtained as described in Materials and methods . DTNB assays revealed five moles of thiols per mole of Trx2red , and thus confirmed the complete reduction of the protein , and less than 0 . 8 thiols per Trx2ox molecule . Firefly luciferase rapidly unfolds and aggregates at temperatures above 42°C , a process that can be monitored by light scattering ( Fig 7 ) . In the presence of Trx2red , the aggregation was significantly decreased and almost completely suppressed at a 20:1 ratio of Trx2red to luciferase ( Fig 7A and 7B ) . In contrast , Trx2ox only marginally affected luciferase aggregation even when used at a 20-fold excess . The cysteine-free 5S-Trx2 , showed significant chaperone activity . The activity was slightly lower compared to Trx2red but was unaffected by a treatment with DTT or H2O2 ( Fig 7A ) . To study how rapidly Trx2 loses its chaperone activity when exposed to non-reducing conditions , we incubated Trx2red in the assay buffer at 44°C either in the presence or absence of DTT prior to measuring its chaperone activity ( Fig 7C ) . While Trx2red did not show any change in chaperone activity when kept under reducing conditions , absence of DTT in the assay buffer led to a gradual loss in chaperone activity with time . From these results , we concluded that Trx2 functions as a molecular chaperone , whose activity appears to be controlled by the oxidation status of its cysteine residues . Next , we studied if Trx2 is also able to prevent the aggregation of chemically denatured proteins at 30°C . For these experiments , citrate synthase was treated with 6 M guanidine hydrochloride overnight . Light scattering was monitored upon dilution of denatured citrate synthase into HEPES buffer at 30°C . When the assays were conducted in the presence of Trx2red or 5S-Trx2 that had been kept at 25°C , we did not observe any effect on the rate or degree of protein aggregation ( S7 Fig ) . In contrast , when the two Trx2 variants were pre-heated for 5 min at 40°C before addition to the assay buffer , both Trx2red and 5S-Trx2 slowed down the aggregation . These results suggested that elevated temperatures might activate the chaperone function of Trx2 . Analysis of WT-Trx2 by size exclusion chromatography revealed that the protein adopts a number of distinct forms ( Fig 5 ) . To investigate whether the redox and oligomerization states of the protein correlate with its chaperone function , we subjected Trx2red , Trx2ox and 5S-Trx2 to gel filtration and analyzed the individual peak fractions by non-reducing SDS-PAGE and chaperone assays . As shown in Fig 8A and 8B , both Trx2red and Trx2ox eluted in several distinct peaks . SDS-PAGE of the four major peak fractions of Trxred revealed the monomeric form , indicating that the protein is able to form non-covalent oligomers ( Fig 8A ) . In contrast , SDS-PAGE of the peak fractions of Trxox showed an increasing number of covalently linked dimers , tetramers , and higher oligomeric species in the earlier fractions compared to primarily oxidized Trx2 monomers eluting in the later fractions ( Fig 8B ) . These results confirmed the data shown in S6C Fig , and suggested that Trx2 forms intra- and intermolecular disulfide bonds upon oxidation . We then tested the major fractions of Trxred for chaperone activity , and found them all to be similarly effective in preventing the thermal aggregation of luciferase at 44°C ( Fig 8A ) . This result is not unexpected given that the different oligomeric states that we isolated from the size exclusion column will likely all adopt the same oligomeric state once diluted into the chaperone assay . None of the Trx2ox-containing fractions revealed any significant chaperone activity in vitro ( Fig 8B ) . The 5S-Trx2 mutant eluted in a single prominent peak ( Fig 8C ) , which appeared to correspond to peak 1 of Trx2red and most likely represents the monomeric form of the protein ( Fig 5 ) . As before , the 5S-mutant showed chaperone activity similar to Trx2red analyzed in the presence of DTT ( Fig 8A and 8C ) . Together , these results suggest that Trx2 is chaperone-active in its monomeric conformation , and demonstrate that the formation of intra- and intermolecular disulfides abolishes the chaperone function . Our studies showed that Trx2red and 5S-Trx2 prevent the thermal aggregation of luciferase in vitro whereas Trx2ox had no significant effect . To further examine whether the chaperone-active conformations of Trx2 affect the rate of thermal inactivation and/or the refolding of luciferase upon return to non-stress temperatures , we incubated luciferase at 42°C in the absence or presence of Trx2red , Trx2ox or 5S-Trx2 . At this temperature , luciferase activity decreased within 20 min to less than 1% of its initial activity ( S8 Fig ) . After resetting the temperature to 25°C and adding ATP , we did not observe any significant luciferase reactivation in the absence or presence of Trx2 . Since many ATP-independent chaperones are unable to refold their client proteins but can maintain them in a refolding competent state and hand them over to ATP-dependent foldases , we supplemented the reaction mixture after its shift to 25°C with the DnaK/DnaJ/GrpE ( KJE ) foldase system . This chaperone system is highly homologous to the mitochondrial Hsp70 system [48] . It is known to promote reactivation of unfolding protein intermediates that were kept refolding-competent by chaperone holdases , such as Hsp33 or the small Hsps , but is unable to refold fully aggregated proteins [49] . Upon addition of the KJE system , we observed substantial refolding of luciferase that was heat-inactivated in the presence of Trx2red or 5S-Trx2 but not of luciferase that was heat-inactivated in the absence of Trx2red or in the presence of Trx2ox ( Fig 9A ) . Since Trx2red was highly sensitive to spontaneous oxidation ( Figs 5 and 7C and S6 Fig ) , we also performed the luciferase heat-inactivation and refolding with different amounts of Trx2red in the absence or presence of 0 . 2 mM DTT ( Fig 9B ) . Reactivation of luciferase , inactivated in the absence of any chaperone or in the presence of the KJE-system , was not affected by DTT . In contrast , the ability of Trx2 to maintain luciferase in a folding competent state was significantly enhanced when the assay was performed in the presence of DTT . These results strongly suggested that chaperone-active Trx2red and 5S-Trx2 are able to maintain thermally unfolding luciferase in a refolding-competent conformation , and can transfer the client protein to the KJE-system for refolding once non-denaturing conditions are restored . As described above , Trx2red and 5S-Trx2 prevent aggregation of unfolding proteins when incubated at elevated temperatures ( Fig 7 and S7 Fig ) . To test whether elevated temperatures result in structural changes that might be required for the activation of the chaperone function , we compared secondary structure elements in Trx2ox , Trx2red and 5S-Trx2 by measuring far-UV circular dichroism ( CD ) spectra and surface hydrophobicity via 4 , 4’-bis-anilino-1 , 1’-binaphthyl-5 , 5’-dissulfonic acid ( bis-ANS ) binding at both 25°C and 42°C . We did not detect any significant differences in the secondary structure arrangement of the proteins at either temperature , and no major difference in the temperature transition ( S9 Fig ) . However , the bis-ANS fluorescence signal of both Trx2red and 5S-Trx2 samples showed a significant increase when the temperature was raised from 25°C to 42°C ( Fig 10 ) indicating that the proteins expose hydrophobic patches as the temperature increases . This finding was entirely consistent with the behavior of other temperature-activated chaperones [50–52] . In contrast , the chaperone-inactive Trx2ox showed a much lower bis-ANS fluorescence compared to Trx2red and 5S-Trx2 at 25°C and an even further reduction of the signal upon incubation at 42°C . These results suggest that oxidative disulfide bond formation causes pronounced structural rearrangements in Trx2 that reduce surface hydrophobicity and may explain why Trx2ox is unable to interact with unfolding client proteins .
Here we show that T . brucei Trx2 , a Trx-type protein that lacks any counterpart outside the order Kinetoplastida , is important for proliferation , cell survival upon prolonged heat stress and infectivity of the parasites . Trx2 is a mitochondrial protein . Indeed , ablation of archaic translocase of the mitochondrial outer membrane ( ATOM 40 ) , which is essential for the import of mitochondrial precursor proteins into the mitochondrion of T . brucei , strongly reduces the abundance of Trx2 in the mitochondrion [23] . Cell lysates often revealed two Trx2 bands which most likely are both processed forms with the upper one representing an intermediate . T . brucei possesses a canonical mitochondrial processing peptidase ( MPP ) as well as a mitochondrial intermediate peptidase ( corresponding to Oct1 in yeast ) [25 , 53] which appears to act synergistically with MPP [53] . Alternatively , the pre-sequence may be removed by two consecutive cleavage steps both performed by MPP as shown for yeast and human frataxin [54] . The 42 N-terminal residues of Trx2 comprise ten arginine residues of which several may serve as cleavage sites of MPP [24] . Interestingly , silencing of Mic20 , a Trx-like subunit of the mitochondrial contact site and cristae organization system ( MICOS ) complex in the mitochondrial IMS , induces the upregulation of Trx2 in PC T . brucei together with several chaperone-like proteins [55] . A role of Trx2 in cristae formation is unlikely as the mitochondrion of BS cells is devoid of cristae . However , Mic20 appears to be involved also in the import of proteins into the IMS and matrix of the mitochondrion [55] . A putative role of Trx2 in folding/stabilization of proteins that are imported into the mitochondrion would be in accordance with its chaperone-like activities in vitro . Remarkably , a cysteine-free mutant of Trx2 was able to fully substitute for the authentic protein . The finding that cells in which Trx2 was down-regulated displayed the same sensitivity towards paraquat as WT parasites and WT-Trx2 or 5S-Trx2 cKO cells , indicates that Trx2 does not play a role in the oxidative stress response of the parasite . Depletion of Trx2 by RNAi attenuated the infectivity and viability of T . brucei in the mouse model . Importantly , 5S-Trx2 cKO parasites were at least as infectious as WT-Trx2 cKO cells . This clearly showed that the physiological role of Trx2 is thiol-independent . The absorption spectra of dimeric and oligomeric forms of recombinant Trx2 suggested that the protein can coordinate iron sulfur clusters . Only a small subset of Trx-fold proteins , most of them representing mono- or dithiol glutaredoxins [13 , 26 , 43 , 56 , 57] , have been shown to bind iron sulfur clusters . The first natural Trx found to bind an iron sulfur cluster is IsTRP , a protein from the tapeworm Echinococcus granulosus [44] . However , the primary structures of IsTRP and T . brucei Trx2 do not display any pronounced similarity . The fact that the 5S-mutant could replace the authentic Trx2 in vitro and in vivo indicates that iron sulfur cluster binding is not an essential physiological role of T . brucei Trx2 . Recombinant T . brucei Trx2 lacked protein disulfide reductase activity but , instead , slowed down precipitation of the insoluble B-chain in the insulin reduction assay . AtTDX , a Trx-like protein from Arabidopsis thaliana , has been reported to have insulin reductase activity at low concentrations which is lost when higher concentrations are applied [58] . This is not the case for Trx2 as the parasite protein did not display reductase activity at any of the concentrations tested . Trx2 species that lacked free cysteine residues were even more efficient in preventing precipitation suggesting that the protein can function as a thiol-independent molecular chaperone . Indeed , Trx2red and the 5S-Trx2 mutant , but not Trx2ox , slowed down the aggregation of heat-denatured luciferase and were able to maintain a thermally unfolding protein in a conformation that allows transfer to the DnaK/DnaJ/GrpE system for refolding . When pre-heated at 40°C , Trx2red and 5S-Trx2 prevented also the aggregation of chemically denatured citrate synthase indicating that elevated temperatures trigger the chaperone function . Trx2red and Trx2ox eluted from gel filtration columns as several distinct oligomeric forms . None of the Trxox species was chaperone-active . In contrast , all Trx2red species were composed of reduced monomers and displayed chaperone activity . In AtTDX , the holdase activity is associated with the formation of oligomeric forms of the protein [58] . However , oligomerization does not appear to be a general mechanism for Trx-type proteins to act as chaperones . The expression of recombinant proteins as fusion proteins with Trx , to increase their solubility and favor folding , supports our conclusion that Trx2red is chaperone-active in monomeric form . Interestingly , 5S-Trx2 and the monomeric form of Trx2red displayed an apparent mass that was nearly twice that of the calculated protein mass . In addition , the surface hydrophobicity of both species proved to be significantly higher than that of Trx2ox and further increased when the temperature was raised from 25°C to 42°C . This indicates that Trx2red and 5S-Trx2 adopt a conformation that supports binding of client proteins which becomes even more pronounced at higher temperature and may explain why Trx2ox , for which the surface hydrophobicity was low and further dropped at elevated temperatures , lacks chaperone activity . Thiol-independent chaperone functions have been shown for E . coli Trx [59] and various other Trx-type proteins [60 , 61] whereby most of these proteins display both reductase and chaperone activity . The properties of T . brucei Trx2 revealed in this work are remarkably reminiscent of those reported for the mitochondrial 2-Cys-peroxiredoxin ( mPrx , mTXNPx ) from Leishmania infantum [51 , 62 , 63] . Both proteins are located in the single mitochondrion of the respective parasites , confer heat tolerance to the cells and are critical for parasite infectivity . Recombinant Trx2 and mPrx have a high propensity to air-oxidize and act as thiol-independent chaperones for putative client proteins under reducing , but not under non-reducing conditions . Both proteins cooperate with the bacterial DnaK/DnaJ/GrpE ( KJE ) system , which is highly homologous to the mitochondrial Hsp70 system [51] , and their chaperone-active reduced forms maintain client proteins in a refolding-competent conformation . An in silico survey of the Hsp70/J-protein machinery of African trypanosomes revealed 12 putative Hsp70 proteins and 67 putative J-proteins [64] . Many of these proteins were predicted or experimentally shown to be localized in the mitochondrion . It will be interesting to see which of these Hsp70/J-protein couples may interact with Trx2 . In a first attempt to identify proteins that may interact with Trx2 , we subjected WT-Trx2 , C63/66S-Trx2 and 5S-Trx2 cKO cells to co-immunoprecipitation and label-free quantitative mass spectrometry ( S10 Fig and S1 Text ) . This approach did not yield a specific interaction partner . Many stress-activated chaperones serve as dual-function proteins with distinct and mutually exclusive activities under non-stress conditions [65] . For the parasite mPrx and Trx2 , an essential redox activity can be ruled out as both proteins can be replaced by cysteine-free mutants . The mPrx is required for long term stability of the insect stage of L . infantum at 37°C but is dispensable when the parasites are cultured at 25°C [62] . In contrast , the presence of Trx2 or 5S-Trx2 was required for proliferation of PC T . brucei under standard culture conditions at 27°C and thus in the absence of a known stress . T . brucei mPrx shares 70% of all residues with the Leishmania protein suggesting largely conserved function ( s ) . RNAi against mPrx in BS T . brucei does not cause any proliferation defect [8] . Our finding that Trx2 is required in both BS and PC T . brucei indicates that the mPrx is unable to functionally substitute for Trx2 . Trx2 and mPrx may act as chaperones on distinct client proteins and/or Trx2 plays an additional constitutive physiological role . Such a dual function has been shown for the mitochondrial Hsp70 . A small fraction of the protein is involved in pre-protein import whereas the majority of mitochondrial Hsp70 is dedicated to protein folding in the mitochondrial matrix [66 , 67] . As shown recently , heat-stress does not induce a general mitochondrial protein aggregation but , remarkably , decreases the import efficiency for cytosolic proteins [68] . If the proliferation defect of Trx2-depleted cells observed at 27°C is indeed due to an impaired mitochondrial protein import one may expect that this process is even more affected when the cells were exposed to 37°C . Future work should focus on the specific role of the novel molecular chaperone in the mitochondrion of African trypanosomes .
Puromycin dihydrochloride , hygromycin B and blasticidin were purchased from Roth , Karlsruhe , Germany . DAPI , tetracycline ( Tet ) , insulin , paraquat and 2-mercaptoethanol were from Sigma-Aldrich . Fetal calf serum ( FCS ) was from Biochrome . H2O2 was from Merck . All restriction enzymes were purchased from ThermoFisher . Primers were synthesized by Eurofins MWG Operons , Ebersberg , Germany . Plasmids were sequenced by GATC Biotech AG , Konstanz , Germany . The pET vectors were a gift of Gunter Stier , BZH , Heidelberg University . The pHD vectors as well as rabbit antibodies against T . brucei aldolase were kindly provided by Dr . Christine Clayton , ZMBH , Heidelberg University . The pRPa vector was obtained from Dr . Keith Matthews , Edinburgh . Antibodies against T . brucei cytochrome c ( Cytc ) were provided by Dr . André Schneider , Bern . Recombinant T . brucei Tpx [69] , rabbit antibodies against T . brucei lipoamide dehydrogenase [27] and guinea pig antibodies against mitochondrial peroxiredoxin ( mPrx; TriTrypDB: Tb927 . 8 . 1990 ) [12] were obtained previously . Guinea pig antibodies against T . brucei Trx2 ( TriTrypDB: Tb427 . 03 . 4240 ) were generated by Eurogentec , Seraing , Belgium . Mouse anti-c-myc and HRP-conjugated goat antibodies against mouse IgGs were from Santa Cruz Biotechnology . HRP-conjugated goat antibodies against rabbit IgGs were purchased from ThermoFisher . HRP-conjugated donkey antibodies against guinea pig IgGs were from Merck . Culture-adapted BS and PC Trypanosoma brucei of cell line 449 were used . The cells are descendants of strain Lister 427 that were stably transfected with pHD449 encoding the tetracycline repressor [70] . 2T1 cells derived from BS 449 cells , which contain a 3’-HYG fragment and VSG expression site promoter-driven PAC ORF [71] , were used to generate BS cell lines overexpressing Trx2-myc6 . Unless otherwise stated , BS cells were grown in HMI-9 medium at 37°C in a humidified atmosphere with 5% CO2 , and PC cells were cultivated at 27°C in MEM-Pros medium , both supplemented with 50 U/ml penicillin , 50 mg/ml streptomycin and 10% FCS as described previously [35] . HMI-9 and MEM-Pros media contained 0 . 2 μg/ml and 0 . 5 μg/ml phleomycin , respectively . BS and PC 449 cells were used to generate Trx2 RNAi and cKO cell lines . For phenotypic analyses , cells were grown in the presence or absence of 1 μg/ml Tet without selecting antibiotics . Genomic DNA was isolated from BS T . brucei parasites using the Qiagen DNeasy Blood and Tissue kit . Unless otherwise stated , E . coli NovaBlue competent cells ( Merck ) were used for plasmid amplification . All primer sequences are listed in S1 Table . For Tet-inducible overexpression of Trx2-myc6 in BS 2T1 cells , the coding region of trx2 was amplified by PCR using Trx2-myc6-HindIII-F and Trx2-myc6-XbaI-R as primers and genomic DNA as template and cloned into the pRPa vector [71] yielding the pRPa-Trx2-myc6 plasmid . For Tet-inducible expression of Trx2 in PC cells , the coding region of trx2 was amplified using Trx2-myc2-HindIII-F and Trx2-myc2-BamHI-R primers and cloned into the pHD1700 vector yielding the pHD1700-Trx2-myc2 plasmid . For Tet-inducible depletion of Trx2 by RNAi , a stem loop construct targeting the coding region of trx2 was assembled . The fragment was selected using the RNAit primer design algorithm to minimize off-target effects [72] . With the primer pairs Trx2i-HpaI/Trx2i-EcoRI-1 or Trx2i-HindIII/Trx2i-EcoRI-2 and genomic DNA as template , a 345 and 300 bp DNA fragment , respectively , was amplified by PCR . Both fragments were digested with EcoRI , ligated and cloned into the pGEM-T easy vector ( Promega ) . Following HindIII/HpaI digestion , the cassette was cloned into the pHD678 vector ( hygromycin resistance ) [70] . For both cloning steps to generate the pHD678-Trx2i plasmid , SURE E . coli competent cells ( Stratagene , Aglient Technologies ) were used . To replace both trx2 alleles , the 5’untranslated region ( UTR ) was amplified by PCR using the primers Trx2-5’UTR-XhoI and Trx2-5’UTR-HindIII yielding a fragment of 237 bp . The 3’UTR of trx2 was amplified using Trx2-3’UTR-PstI and Trx2-3’UTR-NotI generating a fragment of 287 bp . Both fragments were purified and cloned into the pHD1747 ( puromycin resistance ) and pHD1748 ( blasticidin resistance ) vectors to generate the pHD1747-Trx2-KO and pHD1748-Trx2-KO plasmids . For mutageneses , the QuickChange II site-directed mutagenesis Kit ( Agilent Technologies ) was used . With pHD1700-Trx2-myc2 as template and PfuUltra HF DNA polymerase , the cysteines in Trx2 were replaced by serine residues . The first construct generated was pHD1700-C63/66S-Trx2-myc2 which then served as template for the sequential replacement of the other cysteine residues yielding finally pHD1700-5S-Trx2-myc2 . The amplicons obtained were digested with DpnI and used to transform competent E . coli cells . The mutations were verified by plasmid sequencing . For all transfections , approximately 4 x 107 cells were harvested and 10 μg of digested and ethanol precipitated plasmid DNA was used . Transfections were carried out in the Amaxa Nucleofector electroporator with program X-001 using either the Amaxa transfection solution ( Lonza ) or a buffer developed for the transfection of BS T . brucei [73] . For the generation of BS 2T1 cells overexpressing Trx2-myc6 , the pRPa-Trx2-myc6 was linearized with AscI . Stably transfected cell lines were selected with 2 . 5 μg/ml hygromycin . To produce Tet-inducible Trx2 RNAi cell lines , WT T . brucei were transfected with the NotI-linearised pHD678-Trx2i plasmid . Stably transfected BS and PC clones were selected with 10 μg/ml and 150 μg/ml hygromycin , respectively . To generate Trx2 cKO cells , BS and PC WT T . brucei were firstly transfected with NotI/XhoI-digested pHD1747-Trx2-KO and selected with 0 . 2 μg/ml and 2 μg/ml puromycin , respectively . The single KO cells obtained were subsequently transfected with the NotI linearized pHD1700-Trx2-myc2 , pHD1700-C63/66S-Trx2-myc2 or pHD1700-5S-Trx2-myc2 plasmid DNA and cultured in the presence of puromycin plus 10 μg/ml or 150 μg/ml hygromycin to select for BS and PC cell lines , as well as 1 μg/ml Tet . To replace the second trx2 allele , these clones were transfected with the NotI/XhoI-digested pHD1748-Trx2-KO construct , selecting BS and PC clones with puromycin , hygromycin , Tet , as well as 5 μg/ml and 10 μg/ml blasticidin , respectively . Replacement of both trx2 alleles by the resistance genes was confirmed by PCR analyses . BS and PC RNAi and cKO cells were continuously cultured in the presence of 10 and 50 μg/ml hygromycin , respectively . In all experiments , the starting density was 5 x 105 cells/ml . The cKO cells were maintained in the presence of 1 μg/ml Tet unless otherwise stated . To induce a putative endoplasmic reticulum stress , cells were incubated with different concentrations of DTT . To induce oxidative stress , cells were treated with different concentrations of H2O2 or paraquat . Either short term cell viability or proliferation was followed . For heat shock induction , MEM-Pros medium was pre-heated to 41°C for at least 3 h before adding the cells . For long term heat stress , 5S-Trx2 cKO cells were firstly maintained at 27°C in the presence or absence of Tet . When cKO cells grown in the absence of Tet stopped proliferating ( usually observed after 6 to 9 days -Tet ) , they were transferred to 37°C . Every 24 h , the cells were counted and diluted to the start density . The differential membrane permeabilization was done essentially as described previously [26] , except that the 10 mM Tris-HCl , 150 mM NaCl , 1 mM EDTA , pH 8 . 0 buffer was supplemented with 0 . 1 mM PMSF , 150 nM pepstatin and 4 nM cystatin and the lysis was done for 4 min on ice . On two gels , the equivalents of 1 x 107 cells were applied per lane and the blots developed with antibodies against Trx2 followed by Cytc antibodies . For the next couple of gels , 2 x 107 cells and antibodies against Grx2 and C-Myc were used . For the third two gels , 2 x 106 cells and antibodies against mPrx and Tpx were applied . For the subsequent Western blot analyses , the antibodies against Trx2 , c-Myc , mPrx , Grx2 , Cytc and Tpx were diluted 1:1 , 000 , 1:400 , 1:2 , 000 , 1:1 , 000 , 1:200 and 1:2 , 000 , respectively . In each case , the blots were developed with the second antibody after reactivation without stripping . For the secondary antibodies and further details see next section . Cells were harvested , washed , resuspended in PBS containing 20 mM DTT , incubated for 30 min at 30°C , and mixed with 4 x SDS sample buffer containing 8 M urea or directly mixed with reducing sample buffer and incubated for 30 min at 30°C . Total lysates from 1–3 x 107 cells were separated on 12 or 14% SDS gels . After electrophoresis , proteins were transferred onto a 0 . 2 μm PVDF membrane ( GE Healthcare ) and probed with the T . brucei Trx2 antibodies ( 1:1 , 000 ) overnight followed by HRP-conjugated donkey antibodies against guinea pig IgGs ( 1:40 , 000 ) . For a loading control , membranes were treated with antibodies against T . brucei aldolase ( 1:20 , 000 ) or lipoamide dehydrogenase ( LipDH ) ( 1:20 , 000 ) followed by HRP-conjugated goat antibodies against rabbit IgGs ( 1:20 , 000 ) . For detection of the ectopically expressed myc-tagged Trx2 species , mouse anti-c-myc antibody ( 1:400 ) followed by HRP-conjugated goat antibodies against mouse IgG ( 1:20 , 000 ) were used . Bands were visualized by chemiluminesence using the SuperSignal West Pico or Femto substrate ( ThermoFisher ) or the Western BLoT Ultra Substrate ( Takara ) and a digital imager ( GE Healthcare ) . To estimate the cellular level of Trx2 , between 0 . 5 and 5 ng recombinant Trx2 and the lysate of 1 to 3 x 107 PC T . brucei were separated on 14% SDS gels , blotted and probed with the T . brucei Trx2 antibodies as described for Western blot analysis . ImageJ was used to quantify the raw integrated density ( the sum of the values of the pixels ) of each band . The density of the band corresponding to the highest amount of recombinant Trx2 was set as 100% and the % signal for other protein bands was calculated as a proportion of this . The signals from the two forms of Trx2 detected in the cell lysates were combined to give the total cellular Trx2 . The values from seven independent analyses were averaged . To determine the size of the two Trx2 species detected in the cell lysates , the relative mobility of both bands and of recombinant Trx2 was measured and the molecular mass calculated based on standard curves derived from PageRuler Plus ( ThermoFisher ) and Precision Plus Protein DualColor ( BioRad ) protein standards . Immunofluorescence microscopy was conducted as described previously [7] . Approximately 2 x 106 BS and PC cells inducibly expressing myc-tagged versions of Trx2 grown in the presence of Tet were harvested , washed with PBS , and stained with MitoTracker Red CMXRos ( Life Technologies ) . Afterwards , the cells were fixed , permeabilized and treated with anti-c-myc antibodies ( 1:200 in 0 . 5% gelatin in PBS ) for 1 h at room temperature followed by goat anti-mouse antibodies coupled to Alexa Fluor 488 ( 1:1 , 000 in 0 . 5% gelatin in PBS , Molecular Probes ) . The nucleus and kinetoplast were visualized by DAPI staining . Cells were examined under a Carl Zeiss Axiovert 200 M microscope equipped with an AxioCam MRm digital camera using the AxioVision program ( Zeiss , Jena ) . The animal experimentation protocol used in this work was approved by the Animal Use and Ethic Committee ( CEUA ) of the Institut Pasteur de Montevideo ( Protocol 001–18 ) . It is in accordance with the Federation of European Laboratory Animal Experimentation ( FELASA ) guidelines and the National Law for Laboratory Animal Experimentation ( Law nr . 18 . 611 ) . The infection experiments were carried out using Balb/cJ female mice ( 7–9 weeks old ) hosted at the Transgenic and Experimental Animal Unit ( Institut Pasteur de Montevideo ) as described previously [27] . Half of the animals was fed with water containing 1 mg/ml oxytetracycline 96 h prior to infection and during the course of the experiment , replenishing the water with fresh drug every 48 h . Mice ( five per group ) fed with plain water or oxytetracycline ( +Tet groups ) were infected intraperitoneally with 104 exponentially growing WT BS T . brucei , Trx2 RNAi , WT-Trx2 cKO or 5S-Trx2 cKO cell lines . The health status and survival of the animals were monitored daily . The parasitemia levels were assessed from day 3 post-infection onwards . Briefly , blood taken from the submandibular sinus ( 1–100 μl ) was collected in a tube containing 5 μl of anticoagulant ( equilibrated solution of sodium and potassium EDTA salts at 0 . 342 mol/l , pH 7 . 2; anticoagulant W , Wiener lab ) . For some samples , 10 μl PBS-1% ( w/v ) glucose was added to the blood to extent parasite viability during sample processing . After thorough homogenization , an aliquot was diluted 1:20 in red cell lysis buffer ( 0 . 8% ( w/v ) NH4Cl , 0 . 084% ( w/v ) NaHCO3 and 0 . 038% ( w/v ) Na2-EDTA , pH 7 . 4 ) , incubated for 2 min at room temperature , and further diluted with PBS-1% ( w/v ) glucose when the parasite density was above 106 cells/ml . Parasites were counted under an inverted microscope using a Neubauer chamber , which allows to detect a minimum parasite density corresponding to 2 . 5 × 104 cells/ml . Mice showing an impaired health status and/or a parasite load of ≥ 108 cells/ml blood were euthanized . The Kruskal-Wallis test ( followed by Dunn's multiple comparison ) and/or the Mann Whitney test ( non-parametric , two-tailed ) were applied to assess statistical significance of parasitemia . Survival plots were analyzed using the log rank test . The statistical analysis was performed with GraphPad Prism version 6 . 01 for Windows ( GraphPad Software , La Jolla , California , USA ) . P values < 0 . 05 were considered statistically significant . The trx2 coding region without putative mitochondrial pre-sequence was amplified by PCR from genomic DNA using the primer pair Trx2 Short NcoI-F/Trx2 Acc65I-R . The amplicon was purified and cloned into the pET-MBP-vector . To generate Trx2 species in which either Cys63 and Cys66 or all five cysteines were replaced by serine residues , the pET-MBP-Trx2-short plasmid was subjected to site-directed mutagenesis as outlined above for pHD1700-Trx2 . The full length coding region ( Trx2 fl ) was amplified using the primer couple Long NcoI-F/Trx2 Acc65I-R and cloned into the pET-NusA-Trx2 vector . Competent BL21 ( DE3 ) E . coli cells were transformed with the respective pET-Trx2 plasmid . Three liters of bacterial cell culture were grown in LB medium and overexpression of the different T . brucei Trx2 species induced with 0 . 1 mM IPTG . After overnight cultivation at 18°C , the cells were harvested and suspended in 50 ml buffer A ( 50 mM sodium phosphate , 300 mM NaCl , pH 8 . 0 ) containing 50 μM PMSF , 150 nM pepstatin , 4 nM cystatin , 5 mg lysozyme and 0 . 5 mg DNase . The cells were disintegrated by sonication and the cell debris removed by centrifugation . The recombinant proteins were purified by three consecutive chromatographies on Ni-NTA-Superflow column ( Qiagen ) using buffer A with different imidazole concentrations . The supernatant was loaded on a 12 ml Ni-NTA-column equilibrated with buffer A and connected to an ÄKTA Pure system ( GE Healthcare ) . The column was washed with 25 mM imidazole . The fusion protein was eluted with 150 mM imidazole , re-buffered to buffer A , and concentrated using an Amicon Ultra 30 kDa cut-off concentrator ( Millipore ) . In a total volume of 5 ml , the fusion protein was treated for 1 h at room temperature followed by 16 h at 4°C with 2 mg of His-tagged TEV-protease [74] . The digest was applied onto a 5 ml Ni-NTA column equilibrated in buffer A . The tag-free Trx2 was eluted with 25 mM imidazole , washed with buffer A for buffer exchange and concentrated in an Amicon Ultra 10 kDa cut-off concentrator . To remove any remaining fusion protein or impurity , the last purification step was repeated , Trx2 was eluted with 10 to 25 mM imidazole and treated as before . The concentration of the tag-free Trx2 species was determined by Bradford assay and the pure proteins stored at 4°C in the presence of 0 . 02% sodium azide . The concentration of free SH groups was determined by reaction with 5 , 5’-dithiobis- ( 2-nitrobenzoic acid ( DTNB , Ellman’s reagent , ε412 = 13 . 6 mM-1cm-1 ) . Recombinant Trx2 and 5S-Trx2 ( 200 μM ) were treated with 5 mM DTT in 40 mM HEPES-KOH , pH 7 . 5 for 30 min at 30°C and washed with buffer on an Amicon filter with a 10 kDa cut-off until the flow-through was free of thiols ( measured by adding DTNB ) . To generate oxidized Trx2 ( Trx2ox ) , the reduced Trx2 ( Trx2red ) was again diluted to 200 μM in buffer and incubated with 2 mM H2O2 for 30 min at 30°C . H2O2 was removed and the protein concentrated as described above . The 5S-Trx2 variant was also treated with DTT and subsequently H2O2 to exclude any non-thiol-based oxidative modifications . The absorption spectra of recombinant Trx2 , C63/66S-Trx2 and 5S-Trx2 were recorded on a Jasco 650 spectrophotometer . Trx2red , Trx2ox and 5S-Trx2 were diluted to 0 . 2 mg/ml in 20 mM KH2PO4 , pH 7 . 5 ( filtered and degassed ) and far-UV CD spectra recorded between 260–190 nm in a quartz cell with 1 mm path-length at 25°C and 42°C using a Jasco-J810 spectropolarimeter . To monitor the thermostability of Trx2 , the CD signal at 222 nm was followed from 20°C to 80°C . The temperature was controlled by a Jasco Peltier device and increased at a rate of 1°C/min . All spectra were buffer corrected . Changes in the surface hydrophobicity were monitored by the binding of 4 , 4’-dianilino-1 , 1’-binaphthyl-5 , 5’-disulfonic acid ( bis-ANS , Molecular Probes ) resulting in a fluorescent signal upon excitation at 370 nm . Trx2 was diluted to a final concentration of 3 μM in 10 mM KH2PO4 , pH 7 . 0 containing 15 μM bis-ANS . The emission spectrum was recorded from 400–600 nm at 25°C and 42°C using a Hitachi F4500 fluorescence spectrophotometer . All spectra were buffer corrected . Recombinant Trx2 and 5S-Trx2 ( 450 μM to 900 μM ) in 50 mM sodium phosphate , 300 mM NaCl , pH 8 . 0 were preincubated for 30 min at 25°C in the presence or absence of 25 mM DTT . After centrifugation for 45 min at 13 , 000 rpm and 4°C , 25 to 50 μl of the clear protein solution was loaded onto a Superdex 75 10/300 GL column equilibrated in 50 mM sodium phosphate , 150 mM sodium chloride , pH 7 . 0 ± 1 mM DTT and connected to an ÄKTA Purifier system ( GE Healthcare ) . Gel filtration was performed at room temperature at a velocity of 0 . 3 ml/min and detection at 280 , 320 , and 420 nm . Ribonuclease A ( 13 . 7 kDa ) , chymotrypsinogen A ( 25 kDa +/- 25% ) , ovalbumin ( 44 kDa ) , conalbumin ( 75 kDa ) , and alcohol dehydrogenase ( 150 kDa ) served as molecular mass standards . To measure the chaperone activity of the different oligomeric species , a 200 μM solution of Trx2red , Trx2ox or 5S-Trx2 was loaded onto a Superdex 200 10/300 GL column ( GE Healthcare ) equilibrated with 40 mM HEPES , 140 mM NaCl , pH 7 . 5 ± 5 mM DTT . The gel filtration was run at a flow-rate of 0 . 5 ml/min at 4°C using an Äkta-FPLC system . Absorption at 280 nm was detected and the individual elution fractions were studied by SDS-PAGE and in the luciferase aggregation assay . SEC-MALS was performed to determine the absolute molecular mass of 5S-Trx2 . Typically 100 μl of 80 μM recombinant 5S-Trx2 was injected into a Superdex 200 10/300 column connected to an ÄKTA Purifier . This system was coupled to a light scattering detector ( DAWN8+ HELEOS , Wyatt Technology ) and a refractometer ( Optilab tREX , Wyatt Technology ) to measure the absolute refractive index of the solution . Runs were carried out at 4°C and a flow rate of 0 . 4 ml/min . Data were analyzed using the manufacturer supplied software ( ASTRA 6 . 1 , Wyatt Technology ) . The Rayleigh ratio and differential refractive index were plotted against the elution volume . The assay was conducted essentially as described previously [26] . In a total volume of 200 μl of 100 mM potassium phosphate , 2 mM EDTA , pH 7 . 0 , various concentrations of recombinant Trx2 , 5S-Trx2 and/or Tpx as positive control were incubated with 2–3 mM DTT for 20 min at room temperature . The reaction was started by adding 600 μl of a 1 mg/ml insulin solution in buffer , resulting in a final concentration of 130 μM insulin . The increase in turbidity was monitored at 650 nm and 37°C . To generate Trx2 or Tpx species in which the cysteines were blocked , the pre-reduced proteins were incubated with 15 mM N-ethylmaleimide ( NEM ) for 30 min at room temperature . The effect of Trx2red , Trx2ox and 5S-Trx2 on the aggregation of unfolding proteins was studied using the ( 1 ) luciferase and ( 2 ) citrate synthase aggregation assays . ( 1 ) To measure thermal unfolding , a fresh solution of 12 μM luciferase ( Promega ) was prepared in 40 mM MOPS , 50 mM KCl , pH 7 . 5 ( assay buffer ) and diluted to a final concentration of 0 . 1 μM in assay buffer pre-heated to 44°C to initiate protein aggregation under continuous stirring . ( 2 ) 12 μM citrate synthase ( Sigma-Aldrich ) was denatured by overnight incubation at room temperature in 40 mM HEPES , pH 7 . 5 containing 6 M guanidine hydrochloride . To follow protein aggregation , the denatured citrate synthase was diluted to a final concentration of 0 . 075 μM in 40 mM HEPES , pH 7 . 5 at 30°C under continuous stirring . Light scattering was monitored ( λex/em = 360 nm ) in a Hitachi F4500 fluorescence spectrophotometer equipped with a temperature-controlled cuvette holder and stirrer . The maximum in light scattering signal was reached after ( 1 ) 15 min and ( 2 ) 4 min of incubation and was set to 100% . To study the effect of Trx2 on the protein aggregation , different molar ratios of Trx2 were added into the assay buffer . A freshly prepared solution of 12 μM luciferase in 40 mM MOPS , 50 mM KCl , pH 7 . 5 ( assay buffer ) was diluted to 0 . 1 μM in assay buffer and incubated for 20 min at 42°C either alone or in the presence of 2 μM Trx2red , Trx2ox or 5S-Trx2 . As a positive control , luciferase was incubated in the presence of 2 μM DnaK , 0 . 4 μM DnaJ and 2 μM GrpE . For refolding , the reaction was cooled down to 25°C for 10 min and supplemented with 2 mM MgATP and 0 . 1 mg/ml BSA . Since none of the Trx2 variants mediated reactivation of luciferase on their own , 2 μM DnaK , 0 . 4 μM DnaJ , 2 μM GrpE were added to the refolding reaction . At defined time points , 5 μl aliquots were taken to measure luciferase activity in a 96-well plate . Luminescence was measured upon adding 95 μl of 100 mM KH2PO4 , 25 mM glycyl glycine , 200 μM EDTA , pH 7 . 5 , containing 2 mM MgATP , 0 . 5 mg/ml BSA and 70 μM luciferin at 25°C using the FLUOstar Omega microplate reader ( BMG Labteck ) . | African trypanosomes are the causative agents of human sleeping sickness and Nagana cattle disease . These strictly extracellular pathogens multiply in the blood and body fluids of their mammalian hosts and the tsetse fly vector , where efficient redox regulation is essential for parasite survival . While most organisms use the glutathione/glutathione reductase and thioredoxin/thioredoxin reductase couples to maintain cellular redox balance , trypanosomes rely on a unique trypanothione-based thiol metabolism to survive exogenous and endogenous oxidative stresses . Despite the lack of thioredoxin reductases , the Trypanosoma brucei genome encodes thioredoxins , raising questions for their biological function . Our work is the first report on T . brucei thioredoxin-2 ( Trx2 ) . We show that Trx2 is located in the mitochondrion and its absence affects parasite proliferation and infectivity . Recombinant Trx2 lacks protein disulfide reductase activity but protects proteins against aggregation and maintains them folding-competent . Remarkably , a mutant that is devoid of any cysteine residues is able to fully substitute for the authentic protein under in vitro and in vivo conditions . Our data reveal that Trx2 does not function as a classical thioredoxin but acts as a chaperone that plays a crucial role in the mitochondrion of T . brucei . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"luciferase",
"rna",
"interference",
"chemical",
"compounds",
"enzymes",
"enzymology",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"organic",
"compounds",
"trypanosoma",
"brucei",
"protozoans",
"amino",
"acids",
"energy-pro... | 2019 | An essential thioredoxin-type protein of Trypanosoma brucei acts as redox-regulated mitochondrial chaperone |
During mammalian development , neuromuscular junctions and some other postsynaptic cells transition from multiple- to single-innervation as synaptic sites are exchanged between different axons . It is unclear whether one axon invades synaptic sites to drive off other inputs or alternatively axons expand their territory in response to sites vacated by other axons . Here we show that soon-to-be-eliminated axons rapidly reverse fate and grow to occupy vacant sites at a neuromuscular junction after laser removal of a stronger input . This reversal supports the idea that axons take over sites that were previously vacated . Indeed , during normal development we observed withdrawal followed by takeover . The stimulus for axon growth is not postsynaptic cell inactivity because axons grow into unoccupied sites even when target cells are functionally innervated . These results demonstrate competition at the synaptic level and enable us to provide a conceptual framework for understanding this form of synaptic plasticity .
Physiological evidence that axons completely lose connections with some postsynaptic cells as part of naturally occurring development was first observed at the neuromuscular junction in mammals more than 40 years ago [1] . Since then analogous axonal loss has been seen in many parts of the central and peripheral nervous systems [2] , [3] . While the underlying mechanism is still unclear anywhere , evidence suggests that in the neuromuscular system local events at or near the synapse regulate the process . Evidence for local regulation includes the following: ( 1 ) the axonal inputs that are eliminated from neuromuscular junctions do so by gradually vacating their synaptic contact sites [4] rather than suddenly undergoing degeneration , as occurs when axons are damaged [5]; ( 2 ) the axon that ultimately is maintained increases its synaptic contact area by gradually occupying many of the synaptic sites that were previously occupied by other motor axons [4]; ( 3 ) the loss and acquisition of synaptic sites is paralleled by a local reduction and strengthening in synaptic efficacy [6]; ( 4 ) the loss of axonal branches from one axon that projects to many muscle fibers occurs asynchronously , suggesting that the timing of elimination is not set by a signal from the cell soma but regulated independently at each neuromuscular junction site [7]; ( 5 ) local differences between the synaptic activity of axons converging at the same neuromuscular junction have the ability to cause synapses to be eliminated [8] , [9]; ( 6 ) local changes in target cell signaling can affect synapse maintenance [10]; and ( 7 ) once an axon has vacated all of its synaptic territory at a neuromuscular junction , it locally sheds cytoplasm that is internalized by glia associated with the neuromuscular junction entry zone [4] , [11] . Collectively , these data argue that the ultimate identity of the one permanent presynaptic input to a muscle fiber is determined by events occurring at the level of individual neuromuscular junctions . Other data suggest that neuronal properties ( as opposed to synaptic properties ) such as an axon's biochemical identity or its firing pattern play a role in determining the outcome of synapse elimination , but even these may operate through local synaptic mechanisms [12] . Several different local mechanisms have been proposed to explain what drives this process forward . One idea is that individual axon branch removal occurs randomly from a motor unit and is related to an intrinsic requirement that neurons scale back their initially exuberant arbors [13] . A second idea is that the fate of axons is predetermined by positional or perhaps other molecular cues that specify which axon is the best match for each muscle fiber [14]–[16] . A third possibility is that axons converging at a neuromuscular junction compete with each other causing all but the ultimate victor to be removed . It is also possible that some combination of these forces is at play . The idea that synapse elimination is primarily the result of a competitive interaction between the innervating axons was originally proposed because in many muscles the loss of inputs results in exactly one axon remaining at each junction [17] . A competitive mechanism is also suggested by the fact that increases in the size and strength of one input are related to the shrinkage and weakening of other axons [4] , [6] . But there is no direct evidence supporting such a mechanism at the synaptic level , and while a number of studies have suggested inter-axonal competition as the likely mechanism , to our knowledge none have shown a direct reciprocal causal relationship between the fates of the surviving and eliminated axons during developmental synapse elimination [8] , [18]–[21] . Moreover , in some circumstances multiple axons can remain at the same neuromuscular junction [22]–[24] indicating that in some circumstances either competition can be overridden by other factors or that the whole process is not competitive in the first place . Understanding what drives the process forward is important because this mechanism seems to be one of the strategies at play more generally in the developing mammalian nervous system to help shape it to the particular environment in which it finds itself . Thus we felt that it would be worthwhile to directly test whether or not synapse elimination is driven by interaxonal competition . We reasoned that if competition between two axons vying for the same postsynaptic site was causing the elimination of one of them , then that axon should not ever be removed if its putative competitor was no longer present . We therefore ablated the axon that had the greatest likelihood of being maintained at a neuromuscular junction to see if the weaker input would have a reversal of fate and now be maintained . If this outcome did occur , we were interested to know when the decision for an axon to be eliminated finally becomes irreversible . For example , it was possible that axons compete and set into motion a program of elimination that is irreversible even many days before the axonal loss finally occurs . The cascade that leads to neuronal cell death has such points of no return [25] , which imply that the downstream events are irreversible . Might the same be true for the program leading to synapse elimination ? If conversely the synapse elimination program were readily reversible , even at late stages , it would argue that axons remain viable in an ongoing effort to maintain access to the target muscle fiber . In this latter case , the synaptic reorganization events might be played out with little lag between the competitive actions and their consequences on axonal growth or retraction . For example , if one input “pushed” another off a synaptic site , axon withdrawal would be temporally coordinated with near simultaneous axonal takeover , allowing for a highly dynamic process where axonal territory might wax and wane on timescales of minutes or hours . Indeed time-lapse imaging shows that an axon's synaptic territory can be increased and decreased in a dynamic manner [4] . Despite the relatively high temporal and spatial resolution images in many previous studies , however , the motive force for growth and retraction and the details of these behaviors for interacting axons remain obscure . The experiments reported here allowed us to examine how developing axons respond to vacated synaptic sites . We developed a laser-based technique with which we could remove one of two closely spaced axons that innervated the same neuromuscular junction . This technique showed that axons readily grew to occupy vacant sites even when they appeared to be in the process of withdrawing at the time the sites were vacated . In addition we observed that axons were stimulated to grow even in situations when the muscle fiber was still active . This combination of synaptic vacancy and the axonal takeover it induces allows us to explain a range of complex phenomena associated with synapse elimination .
In order to selectively remove one axonal branch without damaging any neighboring axons in vivo , we used a diode-pumped mode-locked Ti:Sapphire laser oscillator to cause localized phototoxicity in fluorescent protein containing motor axons in living mice . By taking advantage of non-linear aspects of multi-photon excitation , we could damage one axon and leave immediately adjacent axons unscathed . The focused laser spot was positioned over an axon branch using a modified scanning microscope system ( see Materials and Methods and Figure S1 ) , and the axon's fluorescence was bleached at one location . One hundred and seventy-three axons were irradiated ( 71 in adult neuromuscular junctions and 102 in 1-wk-old neonates ) . Damage to axons typically evolved over 30–45 min and the whole process of axon removal required many hours . Even though we observed bleaching of the axon segment at the time of irradiation , evidence for structural damage only became apparent within 10–20 min ( see Figures 1–3 ) . Signs of axon damage included dramatic swelling of the axon distal to the site of laser focus and a progressive widening of the region of non-fluorescence both distal and proximal to the laser irradiation site . Presumably this loss of fluorescence is secondary to leakage of proteins from the cytoplasm at the damage site . This phase which typically lasted up to several hours was followed by the complete disappearance of the distal axon save for occasionally a few small disconnected fluorescent fragments that ultimately all disappeared by 10 h . In the proximal direction the damage initiated a die-back that was reminiscent both in time course and scale of “acute axonal degeneration” of damaged central axons [26] . Typically , the die-back stopped at the proximal branch point ( Figure 2 ) , although sometimes it extended anterogradely from the branch point to cause the disappearance of other terminal branches . If the fluorescence at the laser spot recovered after several minutes , that was an indication that the fluorescence in the axonal branch had been bleached but the axon was not seriously damaged because no subsequent changes were noted over the next half hour to hour , or the following day ( see Materials and Methods for details ) . We attempted to remove one of the axons converging at multiply innervated neuromuscular junction in early postnatal life . In anesthetized mice that were 7–8 d old we located neuromuscular junctions in the superficial ( ventral ) surface of the sternomastoid muscle that were innervated by two axons . In the sternomastoid muscle about half the neuromuscular junctions are multiply innervated at 1 wk of age , whereas at 2 wk the number of multiply innervated neuromuscular junctions is very small ( <0 . 1% ) [8] . At 1 wk nearly all of the multiply innervated junctions are contacted by only two axons [4] , [7] , indicating that each of these junctions will lose one input over the next several days . Using multi-photon laser irradiation we successfully removed one axon from each of 15 multiply innervated neuromuscular junctions . At an additional 87 neonatal muscle fibers , axons were damaged but the experiment failed for other reasons including connective tissue buildup and muscle fiber rotation that obscured details when we returned to the muscle the next day; inadvertent muscle , nerve , or blood vessel damage; or occasional animal mortality post-surgery . After confirming that the axon was damaged during the imaging session ( up to 3 h ) we sutured the neck wound and allowed the animals to recover . In most cases ( 10/15 ) we intentionally irradiated the axon with the larger caliber . In 4/15 junctions the two axons had nearly the same caliber , but based on their appearance at the site of entry into the junction , we could identify the axon that we thought had less territory; we then irradiated the larger input . In the remaining case we intentionally laser irradiated the axon with the smaller caliber . In all of these junctions the selectivity of the laser damage was apparent because whereas the damaged axon disappeared , in no case did the non-targeted axon show any swelling , fragmentation , bleaching , or loss ( Figures 1–3 ) . Because in this experiment the two axons were labeled with the same fluorescent protein and in most cases their territories coalesced at light microscopic resolutions , it was not possible to know precisely how much territory the remaining axon occupied except retrospectively . Once the laser irradiated axon had disappeared , it was easy to see the extent of the territory occupied by the remaining axon ( Figure 3B ) . In the 14 cases where we attempted to remove the stronger input , the remaining axon occupied half or less of the junctional acetylcholine receptor ( AChR ) sites ( mean 12% ) , and in the one case where we irradiated the thinner axon , the remaining axon occupied 80% of junction ( Table 1 ) . The territories occupied by these axons were consistent with previous work showing that terminal axon caliber correlated with synaptic territory [7] . Because the axon occupying the majority of the territory at postnatal day 7 or 8 ( P7 or P8 ) was more than twice as likely to remain at a junction than the axon occupying the minority of the territory [4] , we would anticipate that in most of the 14 cases where we targeted the stronger axon , the remaining undamaged axon would have been eliminated had we not perturbed the system . Nonetheless , when we re-anesthetized the mice a day later and returned to the same muscle fibers , in none of the 14 cases had the remaining axon withdrawn . Nor in any case did we see any evidence of regrowth of the damaged axon . We were certain that the axon remaining at the junction was the axon that was not irradiated on the previous day because its site of entry into the junction was in each case the same as the site where the thin axon was situated on the day of laser irradiation ( see Figure 3 ) . In each case , however , the axon that remained changed in a striking way . Each of these thin axons now extended branches throughout the postsynaptic area to fully occupy the territory formerly overlain by the laser irradiated axon . In all but one case the growth response appeared complete within the first 24 h after laser exposure . The axonal expansion of territory was reminiscent of the “takeover” seen during normal synapse elimination at the neuromuscular junction: the advancing axon specifically enlarged its coverage of the adjacent postsynaptic AChR sites without extending sprouts to new sites [4] . Because ordinarily smaller axons were twice as likely to leave a multiply innervated junction than larger ones [4] , the probability that none of them ( 14 cases ) would have withdrawn is very low ( probability <0 . 0000003 ) . Therefore , the fate of the weaker axon was changed by removing the stronger input , a result that argues that competition is the cause of the elimination of the weaker input . We were interested to know what occurred in the immediate aftermath of removing the other input . In particular , did the remaining axon continue to be eliminated for some time , suggesting , for example , that competitive effects have some momentum and a certain amount of time is required before an axon can change its fate ? We found , however , that at no time after axon removal did the remaining axon show any evidence of continued elimination . In three cases we reimaged junctions less than 24 h after the initial view ( one at 6 h , one at 12 h , and one at 17 h ) . At the 6 h and 12 h views , the remaining axons had not lost any territory ( see Figure 2 ) . The junction viewed at 17 h had already shown signs of expansion . Unfortunately , it was not possible for us to anesthetize the same animal for imaging more than once per day and have it survive , so the exact time axons began to grow following laser damage of the other synaptic occupant remains unclear . We do know , however , that at some point after a period of quiescence that lasted up to 12 h , the territory occupied by the remaining axon changed rapidly . In 13/13 junctions imaged at 24 h after laser induced removal of the stronger input , the remaining axon had expanded dramatically . In all but one case the axon occupied the entire postsynaptic site , and in the one other case , it occupied 75% of it ( Table 1A , Figures 1 and 3 ) . The increase in territory by the remaining axon was often matched by a thickening of the caliber of its preterminal branch ( Table 1A , Figures 1 and 3 ) . In the one junction we studied in which the weaker input was intentionally eliminated , we also noted complete takeover of its territory by the stronger axon at 24 h ( Table 1A ) . Thus by 1 d after the laser removal of one axon at dually innervated neuromuscular junctions , the remaining input had grown and now appeared identical to the axons that survive naturally occurring synapse elimination and singly innervate neuromuscular junctions . In each case the undamaged axon now occupied all or nearly all the postsynaptic territory and possessed a thick preterminal axon . Importantly , in 6 of the 15 cases , the axon that remained had occupied less than 5% of the junction at the time of axon removal . These axons were as effective in taking over the remaining territory as axons that had a larger footprint at the time of axon removal ( Figure 3B and Table 1A ) . We thus conclude that once a competing axon is removed the remaining axon , within hours , and irrespective of the contact area of its terminal arbor , changes its fate to take the position and characteristics of the dominant axon . Once an axon has lost all territory at a neuromuscular junction it undergoes a stereotyped process of withdrawal in which the bulb tipped axon branch sheds some of its cytoplasm and appears to retract away from the junction [11] , [18] , [27] . These “retraction bulbs” are seen frequently in developing muscles at the time of synapse elimination but are not seen at all in adult muscles . Might these structures be irreversibly committed to retraction ? In anesthetized mice we successfully damaged 18 strong axonal inputs of singly innervated junctions where a second axon had recently retracted but was still visible nearby . Previous time lapse studies indicate that retracted axons which were within ∼200 µm of a junction had disconnected at some point over the previous 48 h [4] , [11] . After damaging the axonal input that innervated the junction , we allowed the animals to recover and waited to see if nearby retracted inputs ever attempted to return to the junction over the following days . To our surprise , in 55% of the cases ( n = 10/18 ) , the axon stopped retracting , grew back to the junction , and occupied the entire junctional area ( Table 1B ) . The laser-irradiated axon typically died back to the proximal branch point . Although in most cases given the position of the two axons it was unambiguous that the damaged axon did not reinnervate the junction , a potential ambiguity could occur if the damaged axon rapidly reinnervated the junction while at the same time the retracting axon completely disappeared . In order to directly identify the re-growing axon , we used a doubly transgenic mouse in which individual neurons expressed different concentrations of cyan fluorescent protein ( CFP ) and yellow fluorescent protein ( YFP ) in each axon ( see Materials and Methods ) . In this case , we found that the damaged axon ( identified by its color ) did not return over the next 48 h , whereas the retracted axon ( unambiguously identified by its color and its site of exit from the nerve fascicle ) reinnervated the junction within 24 h and increased in caliber over the next day ( Figure 4A ) . In order to determine why some retracting axons succeeded in regrowth whereas others did not , we compared the fates of retracting axons at various distances from their previous neuromuscular junction . In three of the cases , we infer that the retracting axons had only recently been eliminated because at the time of laser irradiation of the dominant axon they still had a fine filamentous process connecting them to the junctional site . All of these retracting axons reoccupied the junction 24 h later . But such a tendril was not required for reinnervation: 7/16 of the remaining retracting axons also reoccupied the junctions despite not being connected . Generally , we found that retracting axons close to the junction were significantly more likely to grow back than ones farther away ( Figure 4B ) . These results indicate that once an axon has disconnected from a junction , there still may be a window of 1–2 d before it has transitioned to a mode where retraction is irreversible . In summary , the ability of retracted axons to return to a junction suggests that growth stimulating signals and/or signals that disinhibit the retraction process are activated following laser axotomy . The delay in initiation of growth appears to be nearly the same whether or not the undamaged input was in direct contact with the laser damaged axons at the junction , suggesting that the onset of growth response is not a result of the loss of contact inhibition . It is , however , clear that the time required to completely reinnervate a neuromuscular junction is a bit slower for axons that have completely disconnected and have a longer distance to grow ( see Table 1B ) . The results described above show that axons that are in the midst of withdrawing from a synapse can be stimulated to grow and occupy the recently vacated sites of a damaged axonal competitor . We were interested in learning if this method of axonal takeover also underlies the way synaptic competition occurs in normal development . In particular there are two alternative ways axons might enlarge their territory . One way is that an axon expands its territory by “pushing off” a competing axon . If this were the case , each AChR region is sequentially occupied first by one axon and then by another with no lag in takeover and no interval when the AChRs are unoccupied . An alternative way an axon might enlarge its territory is if the stimulus for an axon to take over a site is generated in response to that site's vacancy following withdrawal of another axon . This latter mechanism is what likely stimulates the growth of axon terminals following laser axotomy of a competing axon . Previous studies did not have the necessary spatial resolution to resolve this question at sites of synaptic takeover; however , in one situation where the two axons' territories were widely separated , sites that were vacated by one axon were not taken over by the other . That result rules out the idea that in all cases synapse elimination requires one axon to displace another from a synaptic site [4] . In this work we wanted to extend our analysis to sites of takeover . We reasoned that if synaptic vacancy were the proximate cause for synaptic takeover in naturally occurring synapse elimination , then it should be possible to observe instances of transiently vacated AChR sites at the boundaries of the territories occupied by different inputs . Whereas if one axon only lost its territory when another invaded its site , then no lag should ever be seen . In two neuromuscular junctions we did observe synaptic territories that were unoccupied at the first imaging session and then occupied a day later . The presence of unoccupied receptors suggests but does not prove that one axon had withdrawn and the other axon was responding by taking over those sites . More direct evidence would require seeing one axon withdraw from a site that would then appear vacant before another axon would later reoccupy it . Our previous experience suggested that finding examples , were they to exist , would be difficult because in most cases competing axons overlap in some regions with one sitting on top of the other , meaning that if withdrawal precedes takeover , the axon branches are likely lifting but not retracting before takeover occurs . The distances involved are below the resolution limit of the microscope , making it difficult to know if AChRs are transiently unoccupied or not [12] . In addition , we know that takeover following laser axotomy occurs rapidly , but previous time lapse images of retraction suggest that removal of a detached branch occurs more slowly , meaning that the takeover may be happening before the other branch disappears [4] . We thus devised a photo-bleach method to unambiguously see the extent of overlap at neuromuscular junctions as a way to screen for multiple innervated junctions where the two axons abutted each other but had minimal overlap ( see Materials and Methods ) . Such junctions would be most likely to reveal the transient presence of a vacant AChR site . In one case out of hundreds of attempts we did see the withdrawal of one axon leading to AChR vacancy and then the takeover of that site by a second axon ( Figure 5 ) over the course of 3 d . In the first view the two inputs appear to overlap in some areas but not at the site of axon entry . After 1 d , the smaller input has withdrawn completely from the junction . A vacated AChR region is seen at the site of axon entry . By the following day , this region has become occupied by the remaining input . It is clear in this example that withdrawal preceded takeover and that the takeover occurred after a delay of many hours . The timing of takeover is similar to what we found following removal of an input by laser irradiation . Because we did not observe takeover in the first hours after removal of an axonal input by laser irradiation , we think it is likely that takeover is stimulated by the withdrawal ( i . e . , vacated sites ) and that withdrawal before takeover is possibly the general mechanism by which synaptic contacts are rearranged at junctions . However , based on some of the data in this article ( see section below ) and previous studies [4] , it seems likely that a vacant site is a necessary prerequisite for axonal growth but not a sufficient stimulus . For example , when a vacant site is nearby other innervated sites , it is very likely to be reoccupied . But when the site is off by itself at the edge of a junction or at the nerve entry site , then the vacant site is sometimes not reinnervated and the receptors eventually disappear . The results above suggest that a signal from a vacant site may stimulate the growth response during naturally occurring synapse elimination . In many situations axonal growth is thought to be stimulated by signals that emanate from denervated , and thus inactive , target cells [28] . However , in the cases of normal takeover in development and in the case where we targeted the laser to a weak input , the loss of an axon was unlikely to give rise to target inactivity because we had denervated a minuscule portion of a large synapse . We were thus interested to know if small partial denervations of target cells are generally sufficient to activate an axonal regeneration response . In particular if only a small synaptic bouton is removed and the muscle fiber is not functionally denervated , will sprouting be stimulated ? As described above we found one case in development where a small site became unoccupied and then later reoccupied by the remaining axon ( see Figure 5 ) , but we wanted to see if this was a general trend that occurred if vacant sites were present at any age . We were also interested to know whether the proximate cause for the sprouting within a neuromuscular junction could be explained by release of contact inhibition . We thus did focal laser axotomies in adult neuromuscular junctions to denervate small isolated synaptic boutons while retaining innervation to the rest of the junction . Surprisingly in adult animals more than half ( 55% , n = 12/22 ) of these small laser-targeted axonal surgeries which denervated between 5% and 30% of the AChR sites still induced reinnervation ( Figure 6 ) . The reinnervation started after a delay of at least 1 d following laser exposure and was typically complete by 2 d but sometimes longer ( see Figure 6B ) . In all cases reinnervation occurred by sprouting from adjacent branches in the terminal . The sprouts appeared to be directed specifically to unoccupied AChR sites and not elsewhere , yet the original branching pattern ( pre-irradiation ) was not necessarily preserved , suggesting that regrowth was not necessarily guided by preexisting glial sheaths but by highly localized signaling originating at the vacated sites .
This study was undertaken to better understand the sequence of events that occur during development underlying the transition from multiple to single innervation in skeletal muscle . This phenomenon , which has analogs in other parts of the developing nervous system , occurs by one axon's takeover of most of the postsynaptic sites that were earlier occupied by other axons . However , a number of questions about the underlying mechanism remain unanswered . First , what drives the exchange of territory such that when one axon loses sites another typically gains those sites [4] ? Second , what determines the identity of the eventual surviving input given that an axon that loses territory at one time point sometimes gains it back at a later time [4] ? And third , why do the contacts of an axon within a neuromuscular junction tend over time to cluster to occupy a contiguous segregated territory [29] ? In this work we focused on answering the first question . In so doing we think we have also uncovered explanations for the other questions and believe we now have a framework to interpret many aspects of this form of synaptic plasticity . We show that axons rapidly respond to vacant synaptic sites by growth . In multiply innervated neuromuscular junctions an axon whose elimination appears imminent will , within 1 d , occupy all the sites of an axon that was experimentally removed . Moreover , axons that have recently withdrawn completely from a neuromuscular junction will reverse their fate and reoccupy it if the innervating axon is caused to disappear . These results strongly support the idea that the process leading to single innervation is competitive: an axon destined for elimination always survives if the other innervating axon is removed . This growth response of one terminal axon branch to the damage of another terminal branch is in some ways reminiscent of the reinnervation response following partial denervation of a muscle where an axon that is undamaged sprouts to occupy neuromuscular junctions on denervated muscle fibers [30] . However , these two phenomena seem to be dissimilar in several important respects and may have different underlying mechanisms . First , a number of studies support the idea that sprouting following partial denervation is stimulated by muscle fiber inactivity [30] , [31]; however , several of our results show axons growing into vacated synaptic sites even when the muscle fiber is functionally innervated . It is also clear that in naturally occurring synapse elimination , an axon continues to take over vacated sites even when it already occupies the vast majority of the terminal area so that its growth is not being stimulated by inactivity of the muscle [4] . A second difference between partial denervation of muscle and the growth response described here is that the following partial denervation axons grow through vacated Schwann cell tubes [32] or extend along new Schwann cell processes [33] . Neither of these paths is available within neuromuscular junctions . Third , the growth response following laser axotomy in neonatal animals is fast compared to the response of axons in adults to partial denervation . Another difference is that many of the axons undergoing branch loss in development were atrophic and had to transition from a withdrawing state to a growing state , whereas the axons responding by growth following partial denervation in adults are in a healthy quiescent state before being induced to grow . These differences suggest that the local growth response to synaptic vacancy within a neuromuscular junction is different from the growth response of axons to the loss of all innervation to a subset of muscle fibers ( i . e . , partial denervation ) . Because muscle fiber inactivity is unlikely to be the stimulus that induces axonal growth into vacant synaptic sites in our studies , what then is the signal ? One idea is that Schwann cell processes that no longer are associated with an axon become activated . Previous studies have shown that Schwann cell activation following nerve damage is a potent stimulus for axon growth [34] , [35] . Thus , it is possible that focal loss of nerve-glial contact leads to the release of a glial-derived signal that causes axons to grow . Interestingly glial cell-derived neurotrophic factor ( GDNF ) , a glial based growth factor , is one of the strongest known stimuli for mammalian motor nerve growth [36] , [37] . Growth based on loss of nerve-glial contact is an attractive idea because it does not require muscle inactivity as a stimulus . Such a mechanism could be the same as the one that promotes axon regrowth along nerveless Schwann cell tubes to distant muscle targets following nerve damage far from muscle end organs ( such as the sciatic nerve ) [38] . Another possibility is that Schwann cell activation is downstream of a signal originating in the postsynaptic cell or that the vacant postsynaptic site signals axon growth directly . Ongoing experiments are aimed at deciding between these alternatives . The results presented here suggest the primacy of the withdrawal ( or loss of maintenance of synaptic contacts ) as the initiating event leading to synaptic takeover and ultimately single innervation of neuromuscular junctions . Interestingly , signals from presynaptic , postsynaptic , and glial cells all seem to be able to regulate synaptic maintenance . For example , at the mammalian neuromuscular junction , synaptic loss can be initiated by postsynaptic protein synthesis inhibition [10] , focal blockade of neurotransmission at a synaptic site [8] , exuberant branching of motor axons in development [13] , terminal sprouting [35] , and glial loss [39] , [40] . A molecular understanding of developmental synaptic reorganization may therefore require understanding the relative roles of these diverse signals and cells in causing synapse loss . The important point from our perspective is that by focusing on synapse loss ( as opposed to synaptic addition ) it may be possible to get to root causes for the rearrangements . Based on these results we developed a surprisingly robust graphical model simulating synaptic competition ( Figure 7 ) . In this model we posit that maintenance of synaptic contacts is imperfect and that at regular intervals an axon withdraws from an individual synaptic site at random . However , in each case the consequence of the vacated site is the same: as shown in this work , when an axon withdraws from a synaptic site , signals stimulate nearby axons to grow and attempt to occupy that site . In this model the reoccupation favors axons that have the largest number of nearby synaptic contacts ( Figure 7A and see Materials and Methods for details ) . This proposed mechanism is inspired by models used in evolutionary biology for understanding the survival and extinction of different populations within the same niche [41] . In this case the different populations that are in competition are the several axons converging on the same neuromuscular junction , with each synaptic contact being equivalent to an individual member of one or another population . As shown this process will lead eventually to single innervation ( Figure 7B ) . This simple model also provides insight into the cause of several other features of synaptic competition such as synaptic segregation [29] , flip-flop [4] , and the slowing pace of input loss in the second compared to the first postnatal week ( Figure 7B and C ) . Thus , loss-initiated synaptic takeover suggests a useful conceptual framework for understanding competitive synaptic rearrangements at the neuromuscular junction . An important question is whether the model proposed here also has relevance to synaptic rearrangements occurring in other parts of the nervous system . It is well known that axons are lost from neurons at the same developmental stage that axons are removed from muscle fibers . In several cases it is also clear that a remaining input to a neuron ( such as the climbing fiber on a Purkinje cell or a preganglionic axon on a submandibular ganglion cell ) elaborates new synaptic connections at the time other axons are being eliminated [42] . The complementary nature of the loss and gain of connections by withdrawing and remaining inputs , respectively , raises the possibility that a remaining input is induced to grow and innervate synaptic sites that have become vacant because of axon loss . In the climbing fiber system , the final area occupied by the surviving axon is far greater than the area occupied by the multiple axons initially innervating the Purkinje cell soma . This large increase in area is a consequence of the Purkinje cell elaborating its dendritic tree at the same time climbing fiber axons are being lost from the soma [43] . In this case new ( and hence vacant ) postsynaptic sites on the expanding dendritic arbor may stimulate the climbing fiber to “climb” from the soma and grow out along the dendrites . In autonomic ganglia it is also clear that axons elaborate synapses on the soma and then grow out to innervate sites on the dendrites [44] . Thus in both of these cases rearrangement of synaptic connections has two parts: ( 1 ) loss of some axonal inputs and ( 2 ) concomitant elaboration of additional connections by the axons that survive . Hence the mechanism by which motor axons grow locally in response to vacated sites at the neuromuscular junction may inform on and be analogous to the mechanisms underlying the establishment of both the correct number of innervating axons and their total synaptic drive on neurons ( see also discussion in Tapia et al . [45] elaborating the idea set out above that the mechanism for competitive synaptic rearrangements is consistent with a net increase in synaptic numbers and an important general role for elimination during neural development ) . Moreover , if our evidence of continuing plasticity in adult muscles is any guide ( see Figure 6 ) , a kind of local structural plasticity might continue on multiply innervated neurons even in the adult brain . As such “evolution” of neural network connectivity from a dynamic competitive state to one marked by long-term functional stability could be the basis of indelible alterations in brain function such that occur with learning .
Mice that expressed either cytoplasmic GFP ( line GFP-I+/+ ) , YFP ( line YFP-16+/+ ) , or both CFP and YFP ( lines CFP-5+/+ or CFP-23+/+ crossbred with YFP-16+/+ ) were used for all experiments ( protocols approved by Washington University Animal Studies Committee and Harvard University's Institutional Animal Care and Use Committee ) [46] . The GFP line was used for studies in adult mice . The YFP and CFP lines were used for neonatal mice . The GFP line could not be used for imaging in neonatal mice because the onset of GFP expression in motor neurons was delayed until after the first postnatal week . To be able to visually distinguish one input from another we developed a method of labeling axons in multiple colors using very bright thy1-driven XFP transgenic mouse lines . By crossing two lines in which all motor neurons constitutively express a fluorescent protein , we found that in young animals intrinsic variation in expression level of each fluorescent protein between neurons was sufficient for us to distinguish axons by color . An advantage of this approach was that every multiply innervated junction on the muscle surface was a potential candidate for imaging . A challenge was that the spectral separation of the axons was incomplete when imaged using conventional confocal microscopy . Because every neuron expressed the same two fluorescent markers ( varying only in linear combination ) , all the axons appeared in both fluorescence channels . Thus fluorescence of one input obscured the fluorescence of another where there was overlap between the inputs . To achieve complete color separation , we found that we could temporarily and selectively photobleach one or both fluorescent proteins in an input , changing its color and eliminating its fluorescence completely for a time ( ∼minutes ) . This photobleaching was accomplished using the standard confocal lasers that were using many orders of magnitude less power than the IR irradiation for laser axotomy . A region of interest ( ROI ) in the scanning confocal was located where we could illuminate the terminal of a single axonal input . We used the color differences between axons at their entry point to identify where to focus the laser at the terminal for photobleaching one axon without bleaching the other . The ROI was scanned by zooming the laser scanning confocal to cover that area but not the other axon . We followed the effectiveness of the bleaching by visualization of the decrease in fluorescence emission . CFP or YFP were bleached using 440 nm or 514 nm light , respectively , at 100% power for 2–5 min . The animal ventilator was left on while bleaching . An image stack of the neuromuscular junction was then acquired immediately after bleaching to capture the change in that axon's color before fluorescence recovery . We found that because the fluorescence recovers quickly , presumably by dilution with unbleached fluorescent protein in the more proximal axon , the same region could be photobleached repeatedly and without any apparent phototoxic effects ( see [47] for control experiments ) . We assume that because the fluorescent proteins were freely diffusible in the cytoplasm and not tethered to any important organelles or membranes , we noted no immediate ( minutes to hours ) or delayed ( days ) toxic effects on the irradiated neuron ( see Figure S2 ) . Once bleached , the areas where the two axons overlapped in the junction were disambiguated by taking images of the junction with each laser line . Adult and neonatal mice were anesthetized , intubated , and respirated as previously described [4] , [47] , [48] . The ventral neck skin was incised and retracted laterally to expose the right sternomastoid muscle , which was gently lifted on a flat steel platform . Care was taken not to stretch the muscle in order to prevent damage to the muscle , blood supply , or innervating nerve bundle . The wound was filled with sterile saline and a glass coverslip was placed over the muscle , making a meniscus with the saline to help keep the muscle from drying . The coverslip did not touch the muscle . Axons innervating the central band of neuromuscular junctions were visualized using standard epifluorescence optics ( YFP filter cube ) at high magnification ( 60× 0 . 9NA water immersion objective ) . Once a particular nerve terminal was selected , images were taken using a confocal microscope ( Bio-Rad MRC1024MP or Olympus FV-1000 ) . GFP-filled axons in adult animals were illuminated using 488 nm light . In neonates , YFP-filled axons were illuminated using either 488 nm light or 514 nm light , and axons that were filled with both CFP and YFP were illuminated using 458 nm and 514 nm light simultaneously . The ventilator was turned off temporarily ( 30–60 s ) to acquire a stack of images . Axons and acetylcholine receptors were imaged sequentially . To avoid damage to the muscle and synapse , acetylcholine receptors were labeled only after laser ablation of axons . Receptors were lightly labeled with alexa-647 conjugated α-bungarotoxin ( 5 µg/ml in PBS for 40 s ) , and the muscle was then rinsed well with PBS . This dosage did not paralyze the muscle , as greater than 70% of the receptors remained unlabeled [4] . Movement artifacts were removed from stacks using special alignment software ( Autoquant; Media Cybernetics , Inc ) . A 2-D image of each junction was then obtained by a maximum intensity projection . The same junction was imaged multiple times , before and after laser ablation , and in most cases at 1-d intervals thereafter . The mice were resuscitated after each imaging session as previously described [4] , [47] . In the time lapse views the color balance of the image from different time points was sometimes corrected using the background autofluorescence as a reference . Laser ablation was performed using a Spectra Physics Tsunami Ti:S laser oscillator and pumped by a Millenia 5W solid state laser using scanning mirrors from a laser scanning microscope . The pulsed laser output was tuned to 815 nm . The laser power was approximately 120 mW at the back aperture of the objective . An IR-corrected water dipping cone objective ( LUMPlan W-IR2 60× 0 . 9NA Olympus ) was used to focus the laser onto an axon branch . The scanner's mirrors were parked ( i . e . , maximal zoom ) to position the laser onto a ∼0 . 5 µm spot ( diffraction limited ) . In this way all of the laser's power was focused to the waist of an hourglass-shaped beam giving a football-shaped spot ( long axis ∼2 . 4 µm ) . The damage required 30 s to 1 min of mode-locked operation of the laser ( approximately 1nJ energy per 80 fs pulse ) . This laser axotomy approach differs from a previous technique that used a regenerative optical amplifier to provide pulse powers of 10–40 nJ and oil immersion , high NA short working distance objectives to break down the optical transparency and cause rapid heating and vaporization at the focal spot [49] . That previous approach had the advantage that the damage could be accomplished in less than a second with relatively few pulses of laser irradiation and showed immediate effects . Our approach required many seconds with far more pulses and required many minutes to discern the damage ( although photobleaching was evident quickly ) . However , because we were damaging axons by virtue of their ability to absorb the multiphoton excitation , we may have achieved a degree of selectivity that would not be possible otherwise . We believe this damage depends on absorbance of the two photon excitation because there was no damage at the same power level when the laser was not mode-locked . Fluorescence excitation thus appears critical for damage at the laser intensities we used . At the powers used , no scarring or collateral damage to muscles fibers or nearby axons was observed , nor was there a visible plasma bubble . To visualize the excitation spot and its position in the optical field simultaneously on a Bio-Rad MRC1024MP microscope system , we replaced a mirror above the objective in the excitation and return light path with a dichroic mirror ( Chroma 700DCSPXR ) , which reflected the exciting infrared light and transmitted the fluorescent emission to the eyepieces or a camera ( Figure S1 ) . A YFP filter cube ( Chroma ) was used in the epifluorescence light path , above the specially installed dichroic mirror . The fluorescent axons and site of laser excitation were visualized using a low-light SIT video camera ( Dage-MTI Series 68 ) . Filters were installed to block reflection of the IR laser to the camera and eyepieces . In adults , damage was easier to generate in GFP-expressing than YFP-expressing axons . We think the lower susceptibility of YFP axons to damage may be due to lower absorption of multiphoton light by YFP at 815 nm . At longer wavelengths where YFP might be a better absorber , the laser pulse energy available in our system was substantially less than 1 nJ . In pups interestingly , axons were more easily damaged than in adults . YFP expressing axons in pups were just as susceptible to damage as GFP labeled axons in adults , and the threshold intensity for damage was roughly the same for both fluors . Our model of synaptic rearrangements at a neuromuscular junction is based on evolutionary graph theory [41] . The vertices of the graph represent the individual synaptic sites of the junction . In our simulation , each site is randomly assigned to one of six axons . The axons are assumed to have equal fitness; therefore , the probability of axon withdrawal is the same at all synaptic sites . The axon growth rate is assumed to be constant for all axons . A site is selected at random for axon withdrawal . We then randomly select a neighboring synaptic site . The axon innervating this neighboring site grows to take over the vacated site . Thus an axon that occupies multiple neighboring sites has a higher probability of taking over the vacated site than an axon that innervates only one neighboring site . This process of withdrawal from a randomly selected site and takeover by a randomly selected neighboring site is repeated until all the synaptic sites of the junction are innervated by the same axon . To perform our simulations we used a macro written for Matlab ( The Mathworks , Inc . ; Natick , MA ) based on a Potts model with non-periodic boundary conditions . | Early in development , neurons make multiple synaptic connections with their target cells . Over time , many of these connections disappear , leaving behind a fraction of the original connections . Because this pruning occurs when mammals first leave the uterus , it's thought that this type of remodeling may serve to sculpt the nervous system to match a particular environment . However , what causes synapse elimination is not well understood . In this study , we use in vivo imaging to study the connections between motor neuron axons and their target muscle cells , at the neuromuscular junction ( NMJ ) , during a developmental stage when each NMJ has multiple connections . We find that synapse loss is driven by competition between nerve cells vying to remain in contact with the same target cell . We show that an axon that would have been eliminated can always be spared by removing ( with laser microsurgery ) another axon converging on the same synaptic site . The remaining axon not only survives but rapidly grows to occupy the synaptic sites vacated by the removed axon . These results provide a framework for understanding synaptic rearrangements in the developing nervous system . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"motor",
"systems",
"developmental",
"neuroscience",
"synaptic",
"plasticity",
"biology",
"neuroscience"
] | 2012 | Reversing the Outcome of Synapse Elimination at Developing Neuromuscular Junctions In Vivo: Evidence for Synaptic Competition and Its Mechanism |
West Nile Virus ( WNV ) , a member of the genus Flavivirus , is one of the most widely distributed arboviruses in the world . Despite some evidence for circulation of WNV in countries summarized by the World Health Organization as the Eastern Mediterrian Regional Office ( EMRO ) , comprehensive knowledge about its epidemiology remains largely unknown . This study aims to provide a concise review of the published literature on WNV infections in the Eastern Mediterranean Regional Office of WHO ( EMRO ) . A systematic review of WNV prevalence studies on humans , animals and vectors in the EMRO region was performed by searching: Web of Science , Science Direct , Scopus , PubMed , Embase and Google Scholar . Finally , 77 citations were included , comprising 35 seroprevalence studies on general population ( 24460 individuals ) , 15 prevalence studies among patients ( 3439 individuals ) , 22 seroprevalence studies among animals ( 10309 animals ) , and 9 studies on vectors ( 184242 vector species ) . Of the 22 countries in this region , five had no data on WNV infection among different populations . These countries include Kuwait , Bahrain , Oman , Syria and Somalia . On the other hand , among countries with available data , WNV-specific antibodies were detected in the general population of all investigated countries including Djibouti ( 0 . 3–60% ) , Egypt ( 1–61% ) , Iran ( 0–30% ) , Iraq ( 11 . 6–15 . 1% ) , Jordan ( 8% ) , Lebanon ( 0 . 5–1% ) , Libya ( 2 . 3% ) , Morocco ( 0–18 . 8% ) , Pakistan ( 0 . 6–65 . 0% ) , Sudan ( 2 . 2–47% ) , and Tunisia ( 4 . 3–31 . 1% ) . WNV RNA were also detected in patient populations of Iran ( 1 . 2% ) , Pakistan ( 33 . 3% ) , and Tunisia ( 5 . 3% –15 . 9% ) . WNV-specific antibodies were also detected in a wide range of animal species . The highest seropositivity rate was observed among equids ( 100% in Morocco ) and dogs ( 96% in Morocco ) . The highest seroprevalence among birds was seen in Tunisia ( 23% ) . In addition , WNV infection was detected in mosquitoes ( Culex , and Aedes ) and ticks ( Argas reflexus hermanni ) . The primary vector of WNV ( Culex pipiens s . l . ) was detected in Djibouti , Egypt , Iran and Tunisia , and in mosquitoes of all these countries , WNV was demonstrated . This first systematic regional assessment of WNV prevalence provides evidence to support the circulation of WNV in the EMRO region as nearly all studies showed evidence of WNV infection in human as well as animal/vector populations . These findings highlight the need for continued prevention and control strategies and the collection of epidemiologic data for WNV epidemic status , especially in countries that lack reliable surveillance systems .
West Nile Virus ( WNV ) is one of the most widely distributed arboviruses in the world , and a pathogen of public health significance in both humans and animals [1] . This mosquito-borne virus has been classified in the genus Flavivirus within the family Flaviviridae [2] . In nature , WNV is maintained in a zoonotic transmission cycle between birds and mosquitos , principally the Culex species . Susceptibility to WNV infection has also been indicated for many other vertebrate hosts including mammals , birds , reptiles , and amphibians [3] . Equines and humans are incidental “dead-end” hosts who do not play a role in the transmission cycle of the virus . However , equines and humans may manifest sever disease or death as a consequence of infection [4] . Since the first discovery of the virus in 1937 in the West Nile district of Uganda [5] , it has undergone a substantial geographical migration , and spread around the globe . Infection with WNV was first identified in an EMRO country ( Sudan ) in the 1940s . Since then , infection with the virus has been reported in Egypt ( 1950s ) , Iran ( 1970s ) , and subsequently in several other countries across the region [6] . The prevention and control efforts substantially rely on effective surveillance of the infection in birds , vectors , animals , and humans . Despite several studies on different aspects of WNV epidemiology in the EMRO region , there are still many unknowns about the circulation of the virus and the driving factors of outbreaks [6 , 7] . Understanding the epidemiology of WNV in the EMRO faces a number of challenges including inadequate knowledge of physicians about the nature of the disease , misdiagnosis of other common infectious diseases due to similarity in clinical presentations , poor diagnostic infrastructures and the absence of confirmatory assays for serological tests , and lack of a comprehensive and progressive monitoring and surveillance system in majority of countries . The latter has resulted in a gap in knowledge regading the prevalence of WNV infection in the EMRO region . Therefore , we designed a systematic review to provide a clear and comprehensive presentation of the virus prevalence distribution among human and animal populations as well as infection rate in vectors of the region , based on available data .
Articles were screened and selected according to the PRISMA criteria [8] . The PRISMA checklist completed for this review is presented in S1 File . We made an electronic literature search through Web of Science , Scopus , PubMed , Google Scholar , and Index Medicus for the Eastern Mediterranean region database ( IMEMR ) using different combinations of the following keywords ‘West Nile virus , West Nile Fever , WNV’ and the name of the EMRO countries as: Afghanistan , Bahrain , Djibouti , Egypt , Iran , Iraq , Jordan , Kuwait , Lebanon , Libya , Morocco , Oman , Pakistan , Palestine , Qatar , Saudi Arabia , Somalia , Sudan , Syria , Tunisia , United Arab Emirates , and Yemen ( S2 File ) . All databases were searched for English-language original articles published from database inception to January 30 , 2018 . Choosing multiple sources for article search we aimed to enhance our sensitivity in finding relevant articles . To find citations that were not indexed in our target databases , we reviewed the reference lists of relevant articles . Studies identified through electronic and manual searches were listed in EndNote software ( EndNote X7 , Thomson Reuters ) . After exclusion of duplicate citations , two authors ( MF , FS ) independently reviewed titles and abstracts according to the research question . Relevant studies were obtained in full , and assessed for eligibility and risk of bias as described below . All original articles from peer-reviewed scientific journals with a cross-sectional or survey design that estimated the prevalence of WNV infection in humans , animals , or infection rate in vectors were potentially eligible for inclusion in this review . Relevant studies whose abstract was available but their full-text was not ( even after contacting the authors via e-mail ) , were kept in this review in order to present all available data . Studies from outside of the EMRO region were excluded . Any disagreements between the review team were resolved through discussion . The risk of bias in primary studies was assessed following the Cochrane approach [9] . We also considered individual studies’ sample size ( precision ) as a criterion to assess risk of bias , as proposed by Humphre , et al . [10] . Therefore , we evaluated each WNV prevalence study in three domains: 1 ) sampling method , 2 ) response level ( the proportion of subjects who accept to participate in the study ) , and 3 ) type of assay used for the detection of WNV . Each study was considered to have a low risk of bias if: 1 ) it used probability-base/random sampling methods , 2 ) maintained participants’ response level at ≥80% [11 , 12] , or 3 ) it employed viral neutralization testing ( VNT ) for a prevalence study on the general population or used biological tests including viral genome detection and virus isolation from infected individuals . Studies were classified as having unclear risk of bias for a given domain if they did not provide information for that specific domain . Use of probabilistic sampling methods was only evaluated for studies on the general population , because acute infection studies included individuals attending to healthcare facilities . For studies that were conducted on blood samples collected and stored from blood donors , response rate criteria were not evaluated . Studies on human subjects were considered to have high precision if their sample sizes were ≥ 100 [13] . Moreover , in the studies on WNV vectors , minimum infection rates ( MIR ) , that were calculated for samples of ≥ 1000 specimens , were considered as a reliable representation of the true infection rate in the vector population [14 , 15] . Data was extracted from the selected studies using a researcher-made and piloted data extraction form in excel . For studies on human and animal subjects we extracted data on: first author , year of publication , year of implementation , country , city/governorate , sample size , participants’ age and sex ( for human subjects only ) , animal species ( for studies on animals ) , assay type , and estimated assay-based WNV prevalence . For studies on vector populations , further data was extracted including vector species , number of species ( vectors ) tested , collection methods , number and size of the pools as well as the number of positive pools for each species . WNV minimum infection rate ( MIR ) for each species was calculated by dividing the number of positive pools by the total number of specimens tested for that specific species and multiplied by 1000 . When data was available , assay-specific MIRs were calculated and reported .
Database search resulted in 3298 records . After removal of duplicates , we initially screened the title and abstract of 2667 records , 2488 of which were excluded as they were irrelevant to this review . The remaining 179 papers were reviewed in full , of which 77 eligible reports on the prevalence/MIR of WNV covering 17 countries in the EMRO region were included in this systematic review . We identified two relevant citations by reviewing the reference list of these relevant studies [16 , 17] . Fig 1 shows the literature search process . The full-text of five studies could not be obtained even after contacting the authors [18–22] . These studies were kept in this review to present all available data to the readers . All included studies on WNV entailed 27899 individuals ( 24460 general populations and 3439 patients ) , 10309 animals , and 184242 vector species . A summary of the risk of bias assessment results is shown in Table 1 . In brief , most human studies ( 28 out of 35 ) contained sample sizes of ≥100 participants , yielding a high precision in the reported prevalence measure . Thirty out of thirty-five studies on the general population reported their sampling strategy , fourteen of which utilized some forms of random sampling , and hence , had low risk of bias at this domain . In most studies on the general population ( 24 out of 35 ) , risk of bias assessment was affected by unclear reporting in the ‘response rate’ domain . Six studies were performed on volunteers or on blood specimens stored in national reference laboratories or blood transfusion centers , and hence , were not subjected to risk of bias assessment in the ‘response rate’ domain . Viral neutralization test was performed in 40 . 0% and 13 . 3% of prevalence studies on the general and patient populations , respectively , which entails a low risk of bias for the assays used . A total of 50 human prevalence studies for WNV were identified , 35 of which estimated the seroprevalence in the general population . Furthermore , and 15 of them investigated the presence of WNV antibody or genetic material in patients suspected with WNV infection . Human studies covered 14 of 22 countries of the EMRO region , and were published from 1942 to 2017 . The highest number of human studies were reported from Egypt ( n = 10 ) , Iran ( n = 8 ) , and Pakistan ( n = 9 ) , most of which targeted the general population . ELISAs were the most commonly used diagnostic method for the general and patient populations . Table 2 presents detailed data for these studies . The geographic distribution of human prevalence studies is also illustrated in Fig 2A and 2B . Regarding the general population , WNV antibodies were detected in 11 countries including Djibouti ( n = 2 , 0 . 3–60% ) , Egypt ( n = 7 , 1–61% ) , Iran ( n = 6 , 0–30% ) , Iraq ( n = 1 , 11 . 6–15 . 1% ) , Jordan ( n = 1 , 8% ) , Lebanon ( n = 2 , 0–1% ) , Libya ( n = 1 , 2 . 3% ) , Morocco ( n = 3 , 0–18 . 8% ) , Pakistan ( n = 5 , 0 . 2–65% ) , Sudan ( n = 5 , 2 . 2–47% ) , and Tunisia ( n = 1 , 4 . 3–31 . 1% ) . Since 2010 , seroprevalence of WNV among the general population has been investigated in Djibouti , Egypt , Iran , Iraq , Libya , Morocco , and Sudan among which the lowest and highest median prevalence was found in Iran ( median prevalence = 1 . 4 , range: 0–18%; total SS = 1322; 2010–2012 ) , and Egypt ( median prevalence = 55%; total SS = 160 , 2013–2014 ) , respectively ( Table 2 ) . In addition , the presence of WNV antibody or genetic material in patients was investigated in 15 human prevalence studies . In this regard , seven studies assessed WNV IgM , five of which detected the antibodies in patients’ sera . These studies were from Afghanistan ( n = 1 , 0 . 5% ) , Pakistan ( n = 1 , 6 . 6% ) , Sudan ( n = 2 , 0 . 3–87 . 5% ) , and Yemen ( n = 1 , 14 . 3% ) . Four studies [56 , 59 , 63 , 64] used both serological and molecular assays to detect WNV IgM as well as WNV RNA in patients’ sera ( Table 2 ) . A total of 22 studies investigated the WNV seroprevalence in animals ( Fig 2A and 2B ) . WNV antibodies were detected in 10 countries , including , Djibouti ( n = 1 ) , Iran ( n = 4 ) , Jordan ( n = 1 ) , Morocco ( n = 4 ) , Pakistan ( n = 2 ) , Palestine ( n = 1 ) , Qatar ( n = 2 ) , Saudi Arabia ( n = 1 ) , Tunisia ( n = 5 ) and the United Arab Emirates ( UAE; n = 1 ) . In these studies , serological evidence of WNV infection was detected in a wide range of domestic and wild animals , including Buffalos ( Pakistan , total SS = 33 , prevalence = 15 . 1% ) , Camels ( Morocco , total SS = 2775 , prevalence = 8–23%; Palestine , total SS = 35 , Prevalence = 40%; Tunisia , total SS = 120 , Prevalence = 0–25 . 8% ) , Cows ( Pakistan , total SS = 45 , prevalence = 2 . 2% ) , Goats and sheep ( Pakistan , total SS = 94 , prevalence = 23 . 9%; Palestine , total SS = 95 , prevalence = 14 . 7% ) , Dogs ( Djibouti , total SS = 91 , prevalence = 56 . 5%; Morocco , total SS = 231 , prevalence = 54–96% ) , Ruminants ( Djibouti , total SS = 11 , prevalence = 25 . 3% ) , Equids ( Iran , total SS = 1839 , prevalence = 0 . 8–70 . 3%; Jordan , total SS = 253 , prevalence = 24 . 9%; Morocco , total SS = 1189 , prevalence = 25–100%; Pakistan , total SS = 449 , prevalence = 65%; Palestine , total SS = 585 , prevalence = 75%; Qatar , total SS = 421 , prevalence: 0–27%; Saudi Arabia , total SS = 63 , prevalence = 33 . 5%; Tunisia , total SS = 1473 , prevalence = 28 . 0–45 . 2%; the UAE , total SS = 750 , prevalence = 5 . 4–28 . 6% ) , and different types of wild and domestic birds ( Iran , total SS = 519 , prevalence = 15%; Morocco , total SS = 346 , prevalence = 3 . 5%; Tunisia , total SS = 434 , prevalence = 0 . 7–23% ) . Table 3 provides further details on these studies . Nine studies investigated arthropods in order to analyze the WNV infection rate among vectors . These reports were from Djibouti ( n = 2 ) , Egypt ( n = 2 ) , Iran ( n = 2 ) , Lebanon ( n = 1 ) , Pakistan ( n = 1 ) , and Tunisia ( n = 1 ) . The primary vector of WNV , i . e . , Cx . pipiens s . l . [2] , was detected in Djibouti , Egypt , Iran , and Tunisia , and in all theses countries WNV infection in Cx . pipiens s . l . was identified . WNV infection was also detected in a wide range of other vector species , including Cx . quinquefasciatus ( Djibouti ) , Ae . caspius ( Iran ) , Cx . antennatus ( Egypt ) , Cx . perexiguus ( Egypt ) , and Argas reflexus hermannii ( Egypt ) . Details for studies on WNV infection vectors are provided in Table 4 and Fig 3 .
Seroprevalence of WNV has been investigated in 14 of 22 countries in the EMRO region . Since 1942 , WNV antibodies have been detected in the general population in 11 countries with available data , including: Djibouti , Egypt , Iran , Iraq , Jordan , Lebanon , Libya , Morocco , Pakistan , Sudan , and Tunisia . Our results also suggested that the overall seroprevalence of WNV has been lower in reports from more recent years ( since 2010 ) compared to reports compiled between 1942 and 2009 . Although the presence of WNV infection remains unknown in countries without data in the EMRO region ( n = 14 ) , it can be implied that the virus may probably circulate within these countries as well . Existing evidence suggests cross-country dispersion of a number of viruses such as human immunodeficiency virus ( HIV ) [91] and hepatitis B virus ( HBV ) [92] . These observations can imply the hypothesis in which WNV also have dispersed across countries in the region , affecting localities ( countries ) adjacent to infected areas . The argument is further strengthened if we consider the transmission routes of HIV , HBV , and WNV . The transmission of HIV and HBV depends on effective human-to-human contacts , which acts as a barrier for virus dispersion over large geographic distances . However , similar to other arboviruses like Dengue and Crimean-Congo Hemorrhagic Fever [93 , 94] , the cross-country spread of WNV can be much easier and fast as it can be transmitted through a broad range of vectors and reservoirs . Most of the seroepidemiological studies included in this review used ELISA for the detection of anti-WNV antibodies . Although this assay is simple , sensitive , and commercially available , it suffers from cross-reactivity with antibodies raised against other flaviviruses . So , using the ELISA method for testing individuals with a history of vaccination against , or infection with related flaviviruses can yield false positive results [95] . To achieve a more specific measurement , positive ELISA test results should be confirmed by the plaque reduction neutralization test ( PRNT ) , which is considered as the gold standard method for WNV serological testing . However , PRNT can detect antibodies at levels that neutralize the virus; therefore , it has low sensitivity for seroepidemiological studies in weakly-exposed populations [95] . Approximately , one-fifth of WNV infected individuals demonstrate symptomatic infection [96] . Clinical symptoms are also non-specific to the disease and include fever , malaise , headache , back pain , myalgia , and anorexia . Therefore , WNV infected individuals can be misdiagnosed with other febrile infections . In areas with evidence of WNV circulation , WNV infection should be considered as a differential diagnosis for patients demonstrating non-differential febrile syndroms . Non-specific sympotoms of the WNV infection also highlights the need for laboratory testing of suspected human cases . While WNV IgM is the most common target for confirmation of the infection , viral RNA testing can also be performed . Combining IgM detection and viral RNA testing can enhance the possibility of diagnosis in patints with West Nile fever , as indicated by Tilley et al . [97] . However , among 15 studies on patient populations , only four used a combination of serological and molecular assays for the diagnosis of WNV infection . In this review , we have highlighted serological evidence of WNV infection from 22 independent studies conducted on animal populations in the region . These studies were carried out in 10 countries including , Djibouti , Iran , Jordan , Morocco , Pakistan , Palestine , Qatar , Saudi Arabia , Tunisia , and the UAE . Most studies , have investigated evidence of WNV infection among domestic animals . Since 2010 , the highest prevalence of WNV among domestic animals , has been reported among dogs of Morocco and equids of Morocco , Pakistan , Palestine and Iran . The high rates of animal seropositivity and geographic distribution of animal infection reflect the favorable conditions for the circulation of WNV in these countries . In these areas , stronger preventive measures should be considered to reduce the risk of WNV transmission to humans and horses . High seropositivity among dogs and equids also suggests that these animals can be useful sentinels for WNV surveillance , as discussed by previous studies [98–100] . Resnick , et al . ( 2008 ) reported that WNV seroconversion in dogs happened six weeks prior to the infection in exposed human cases [100] . Only two studies from Pakistan ( on rodents ) [45] and Djibouti ( on wild ruminants ) [66] ) have investigated wild animals’ infection with WNV . The paucity of published studies on the prevalence of WNV infection in wild animals of the EMRO region underlines a gap in current knowledge about the issue . Knowledge about the reservoirs’ infection and virus circulation among wild animals has important implications for forecasting the emergence or re-emergence of WNV epidemics[95] . So , it is recommended future seroprevalence studies include representative samples from wild animals to further illuminate the state of the infection among these hosts . Four studies investigated the infection among birds from Iran , Morocco , and Tunisia , from which only two studies were recently performed ( i . e . , Tunisia , 2015 and 2017 ) . These observations also highlight a gap in current knowledge , this time , on the extend of the infection among birds of the EMRO region . Birds play a critical role in the maintenance and spread of the virus . Prolonged high levels of viremia have been demonstrated in several bird species [101 , 102] . The virus has also been isolated from several migratory birds . Thus , surveillance of WNV infection among birds would be of great importance , especially in areas with favorable ecological conditions for birds and mosquitoes . In this regards , a better understanding of birds migration routes would be helpful in selecting the most probable sites for tracking the virus [102] , and subsequently making judgments on what areas might be focal points for the emergence of WNV outbreaks . Mosquitoes and birds are currently considered to have the key role in the life cycle of the virus [2] . However , there are more than 30 other vertebrates such as lemurs , frogs , hamsters , squirrels , rabbits , and chipmunks that have been reported as possible reservoirs for the virus , since they can provide viremia levels that are sufficient to infect mosquito vectors [103] . The role of these reservoirs in the WNV life cycle and epidemic has been less regarded till now , and is an open area for future research . Despite the critical role of the vector in the life cycle and the epidemic of WNV , only nine studies have investigated vector infection in the region . These studies have been conducted in Djibouti , Egypt , Iran , Pakistan , and Tunisia . The primary vector of WNV , i . e . , Cx . pipiens s . l . [2] was detected in all investigared countries except Pakistan . Although WNV infection has been detected in more than 60 mosquito species , detection of viral infection in a mosquito alone does not indicate that the mosquito is a competent vector for the virus . In addition to Culex species , WNV has also been detected in Aedes and Mansonia mosquitoes . Additional studies are necessary to further clarify the potential role these species in the maintenance and transmission of WNV . Interestingly , WNV infection was observed in ticks Argas reflexus hermannii . Previous studies from other regions of WHO also detected WNV RNA in ticks R . turanicus and mites D . gallinae and O . sylvarum . However , their competency as vectors is less clear [104] . Reducing virus transmission from a vector is one of the main strategies of controlling arboviral diseases . Therefore , more efforts to identify the main vectors and understand virus–vector interaction in burdened countries would benefit disease control strategies [105] . The main limitations of this systematic review relate to the data . First , there is a paucity of prevalence studies in the EMRO region , and the quality of data reported by studies varied . For instance , many available studies on human populations were focused on adults , or did not report age and gender for the study sample . The remaining studies included a broad range of age groups ( including infant , children , and adults ) , most of which did not report age and gender specific prevalence . Prevalence data on healthy infants and young children alone was particularly sparse . Therefore , the state of the epidemic among different age and sex groups remains unknown in this region and requires further study with representative samples . Although current data provides a good basis for an overall judgment about the presence of current/past WNV exposure in most investigated samples , they can hardly be used to infer the actual prevalence and state of the epidemic in most investigated countries . For example , only in four countries with available data on the ‘general population’ ( i . e . , Egypt , Iran , Lebanon , and Pakistan ) , the total number of tested individuals was reasonably representative of the target population ( i . e . , more than 1000 ) . These ‘powerful’ studies , however , were not totally flawless . One of the main limitations of these studies was that some of them had used convenience ( non-random ) sampling methods . In convenience sampling , individuals have unequal and unknown probability of being selected [106] . Hence , the resulting seroprevalence estimates should be generalized to the target population with caution . Few studies available from animal populations in the region also suffered from the abovementioned shortcomings; i . e . , non-random sampling and small sample sizes . For example , the seroprevalence of WNV has been investigated in Morocco , Palestine , and Tunisia , but only the study in Morocco has provided the estimate based on a fairly representative sample of 556 and 836 camels for the years 2003 and 2009 , respectively . The case was even worse for the seroprevalence studies on dogs , cows , sheep , goats , buffalos , and birds as none of the available studies were well-powered enough ( i . e . , had small sample sizes ) . The situation was more satisfactory for the population of horses , where a number of studies with large sample sizes were available from different parts of Iran , Morocco , Pakistan , Palestine , Tunisia , and the UAE . Second , the relative dearth of recent seroprevalence studies , particularly from burdened areas for WNV infection and high-risk population groups is a serious limitation . As the face of WNV disease and its geographic range changes rapidly , WNV prevalence estimated by older studies may not properly reflect the current status of WNV circulation . Less accurate serological tests used by older studies also affect the validity and reliability of the prevalence estimates in these studies . Standardized seroprevalence studies at national levels are critical to best appraise the epidemic status , the impact of interventions and the potentials for future outbreaks . Third , substantial within-country heterogeneity in the prevalence of WNV was noted . This might be due to diversity in the geographical areas , target groups , and the reported sample sizes of studies . Local prevalence estimates , hence , might not be representative of national level prevalence , particularly in large countries with much geographic and ethnic disparities . Finally , our review is limited to reports written in English .
This review provides estimates of the scale of the WNV epidemic at country and regional levels in order to inform efforts for developing and implementing effective future responses . Our results suggested the circulation of WNV in humns , animals , or vectors of most investigated countries in the region . However , there is paucity of data about WNV infection , especially with respect to the burden of the infection in most countries across the region . Hence , further epidemiological studies that take into account the human , reservoir and vector dimension/aspect of the occurrence and distribution of the virus should be conducted particularly in high-prevalent countries . Such research effort will generate robust knowledge and a detailed understanding of the epidemiology of the infection in local populations , and foster in-depth investigations about transmission patterns of the virus . Identification of the geographic distribution of primary reservoirs of the virus and their infection status can also enhance targeted prevention and elimination efforts and aid forecasting attempts . Moreover , surveillance capacities in EMRO countries ought to be established or expanded for better monitoring of WNV infection at national and regional levels . | West Nile Virus ( WNV ) is a mosquito-borne Flavivirus belonging to the Flaviviridae family , which is endemic in a vast geographical area , including the EMRO region . However , the epidemiology of WNV in the EMRO region remains poorly understood . To address this gap , we performed a systematic review on WNV prevalence studies conducted on human populations , animals and vectors across Eastern Mediterranean countries . Our review indicated the infection of most investigated human , animal and vector populations with WNV; however , the paucity of epidemiological data underline the need for integrated surveillance programs as well as continued deployment of prevention and control strategies . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion",
"Conclusions"
] | [
"reverse",
"transcriptase-polymerase",
"chain",
"reaction",
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"geographical",
"locations",
"microbiology",
"vector-borne",
"diseases",
"viruses",... | 2019 | Epidemiology of West Nile Virus in the Eastern Mediterranean region: A systematic review |
Endogenous retroviruses ( ERVs ) , the remnants of retroviral infections in the germ line , occupy ~8% and ~10% of the human and mouse genomes , respectively , and affect their structure , evolution , and function . Yet we still have a limited understanding of how the genomic landscape influences integration and fixation of ERVs . Here we conducted a genome-wide study of the most recently active ERVs in the human and mouse genome . We investigated 826 fixed and 1 , 065 in vitro HERV-Ks in human , and 1 , 624 fixed and 242 polymorphic ETns , as well as 3 , 964 fixed and 1 , 986 polymorphic IAPs , in mouse . We quantitated >40 human and mouse genomic features ( e . g . , non-B DNA structure , recombination rates , and histone modifications ) in ±32 kb of these ERVs’ integration sites and in control regions , and analyzed them using Functional Data Analysis ( FDA ) methodology . In one of the first applications of FDA in genomics , we identified genomic scales and locations at which these features display their influence , and how they work in concert , to provide signals essential for integration and fixation of ERVs . The investigation of ERVs of different evolutionary ages ( young in vitro and polymorphic ERVs , older fixed ERVs ) allowed us to disentangle integration vs . fixation preferences . As a result of these analyses , we built a comprehensive model explaining the uneven distribution of ERVs along the genome . We found that ERVs integrate in late-replicating AT-rich regions with abundant microsatellites , mirror repeats , and repressive histone marks . Regions favoring fixation are depleted of genes and evolutionarily conserved elements , and have low recombination rates , reflecting the effects of purifying selection and ectopic recombination removing ERVs from the genome . In addition to providing these biological insights , our study demonstrates the power of exploiting multiple scales and localization with FDA . These powerful techniques are expected to be applicable to many other genomic investigations .
Endogenous Retroviruses ( ERVs ) are Class I Transposable Elements ( TEs ) considered to be remnants of germ-line retrovirus infections inherited by the next generations [1] . As all Class I TEs , ERVs transpose via an RNA intermediate , i . e . they “retrotranspose” . Because they possess Long Terminal Repeats ( LTRs ) , they are also known as LTR-retrotransposons . Depending on the similarity of their gene content to that of certain retroviruses , ERVs are classified as Gammaretrovirus- , Betaretrovirus- , and Spumaretrovirus-like [1–3] . Full-length ERVs possess three retroviral coding genes ( i . e . gag , pol , and env ) and LTR flanking sequences [4]] . In most cases , the internal genes are deleted by recombination of LTRs , converting ERVs into solo-LTRs [5 , 6] . Most ERVs have accumulated numerous mutations that render them inactive [7] . However , some rare examples of young ERVs that have coding capacity , are expressed and are transpositionally active , have been described in mammals , e . g . in koala [8] , mouse [3] , cat [9] , sheep [10] , and mule deer [7] . Active ERVs are transposition-competent and have integrated recently; hence for them , polymorphic events–in terms of presence/absence–are observed at the population level but the allele frequencies of integrations are low . For instance , CrERVγ is an endogenous gammaretrovirus that was recently detected in mule deer [7] . This ERV has been invading the germ line of mule deer since its speciation from white-tailed deer approximately 1 . 1 million years ago ( MYA ) , and the copies found display polymorphisms in the wild mule deer population . In total , in this species , there are on average 100 full-length copies of the CrERVγ per haploid genome [7] . However , if solo-LTR elements are included , this estimate increases two-to-three-fold [7] . ERVs occupy ~8% of the reference human genome ( they are called HERVs for Human ERVs ) , and have been integrating in it starting more than 35 MYA [6 , 7 , 11 , 12] . However , only the HERV-K family has been active during the past 6 MY–since the divergence of human and chimpanzee [13] . Moreover , among 113 human-specific HERV-K elements only 15 are full-length and none is infectious [13] , though about a dozen were found to be polymorphic in 100 individuals from diverse populations indicating retrotransposition activity in the recent past [13 , 14] . In cell lines , however , two HERV-K named Phoenix [15] and HERV-KCON [16] were reconstructed to be infectious , producing retroviral particles and causing in vitro integrations . Expression of HERV mRNA varies among tissues . Importantly , significant expression levels were detected in testis as well as placenta [17 , 18] . Some analyses are available for HERV-Ks embryonic expression [19 , 20] . Approximately 10% of the reference mouse genome is derived from LTR elements , including ERVs [21] . In mouse oocytes , approximately 13% of transcripts were reported to be derived from MaLRs ( a type of LTR elements ) as detected from ESTs [22] . Also , mice have highly active ERVs causing up to 10–12% of spontaneous germ-line insertional mutations–most of which are due to activity of IAP ( Intracisternal A Particle ) and MusD/ETn ( or ETns in short; Early Transposon family ) elements [23] . IAPs and ETns are both non-infectious betaretroviruses . In the mouse genome , full-length IAPs contain retroviral genes needed for retrotransposition; however there are also partially deleted copies ( ERVs missing genes or other sequences ) . ETns consist of non-coding sequences and are aided by MusD proteins to retrotranspose [3] . Insertional polymorphisms have been detected for both IAPs and ETns in multiple mouse strains; additionally , some insertions arose prior to the divergence of these strains [24 , 25] . In the rodent lineage , out of seventeen species studied , three ( Mus , Spermophilus , and Cavia ) possess 80% of all IAP loci found in these species [26] . These elements are absent from monkeys and apes [26] . Mouse IAPs and ETns are known to transpose in different mouse strains causing mutations in the germ line; both polymorphic ( in terms of presence/absence ) and fixed elements are known for each mouse strain [25] . The exaptation of ERVs–i . e . the recruitment of their sequences to perform a new function as regulatory or coding sequences–has influenced the evolution of genomes in multiple ways . Some enhancers and promoters derived from ERVs assume new roles in gene regulation; e . g . , the alternative promoter of the CYP19 gene–an enzyme important for estrogen biosynthesis–leads to its high expression levels in the primate placenta [3 , 27] . Another interesting example of ERV exaptation associated with the evolution of placenta is syncytin , a gene derived from the env gene of HERV-W [28] . Other ERV genes were exapted to function as proteases , RNA-dependent DNA polymerase with RNAse H , and integrases , as well as structural proteins , in diverse organisms [29] . Importantly , while ERVs have been relevant to genome evolution , they have also been implicated in the development of multiple diseases by disrupting genes , modifying regulatory sequences or altering gene expression . Though causal links have not been definitely established , the diseases that have been associated with ERV retrotransposition or expression include multiple sclerosis , cancer and psoriasis in human [3 , 11 , 29 , 30]; and obesity , diabetes , and cancer in mouse [24] . Notwithstanding the role ERVs play in the architecture , evolution , and function of genomes , our knowledge of how the genomic landscape influences their integration and fixation is still limited . Gene density and GC content have been shown to be negative predictors of ERV density–for not only full-length elements but also solo-LTRs [31–33] . In contrast , reconstructed HERV-Ks integrate in regions with high numbers of gene transcription units [34] . Similarly , in vitro IAP integrations occurred preferentially in actively transcribed domains of HeLa cells [35] . Interestingly , human and mouse ERVs that are located in introns are mostly present in antisense orientation avoiding gene expression disruption [32 , 36 , 37] . Other important genomic characteristics of ERVs and their genomic neighborhoods are high levels of methylation and epigenetic modifications used by the genome to limit transposition [3 , 38] . ETns , however , show decreased methylation when located in the vicinity of transcription start sites and expressed genes [39] . Chromosome location is another relevant feature of ERV distribution , as illustrated by the description of 100 previously unknown HERV-Ks in the centromeres of 15 chromosomes [40] . It has been suggested [34] that the accumulation of ERVs is the net result of two processes–integration , which can be biased towards certain genomic landscapes , and purifying selection , which removes ERVs disrupting the function of important elements , e . g . of genes . Disentangling these two processes can be challenging and requires the investigation of ERVs that integrated in the genome at different times . Several approaches have been used to elucidate the relationships between genomic features and distribution of TEs . Most studies of the associations between genomic features and TE density , Integration Site ( IntS ) preferences , or neighboring sequences characteristics were performed employing statistical methods such as ROC curves [41] , non-parametric tests [42] , Fisher exact tests [43] , maximum likelihood techniques [33] , MANOVA [44] , and multiple regressions [45 , 46] . The main limitation of many past studies was the low data resolution determined by available technologies . However , resolution has recently improved , e . g . , with the release of ENCODE and ModENCODE consortia data [47] . The application of innovative statistical approaches though has not kept pace with the improvement in data . Statistical methodology should address the fact that many features of the genome act jointly in defining its biological functionality . Being able to consider multiple genomic features simultaneously , e . g . , with multiple regression analyses [45 , 46] , is essential to obtain meaningful biological conclusions . Moreover , with the availability of higher resolution data , it becomes paramount to use statistical techniques capable of detecting and differentiating effects at different scales and locations , e . g . , one genomic feature may be generally enriched or depleted in the broad flanks of a TE , while another may show enrichment or depletion at a specific location in close proximity of the element’s IntS . To perform more powerful and effective analyses , one can view genomic features as “curves” composed of measurements in consecutive genomic intervals . In this framework , Functional Data Analysis ( FDA ) techniques can be exploited to extract signals from these curves , taking advantage of the ordered nature of the measurements and considering different scales and locations , i . e . sizes and positions of genomic intervals ( see [48] and [49] for a comprehensive introduction to FDA ) . This class of techniques includes curve smoothing and registration methods , functional principal component analysis , functional hypothesis testing , functional regression , and functional clustering [50] . In the last decade FDA has been utilized in an increasing number of biomedical applications [51 , 52] , particularly in cardiovascular research [53–55] and kinesiology [56] . Although still limited in number , some applications of FDA also exist in the context of genetics and genomics , e . g . , in genetic association studies [57–59] , epistasis analysis [60] , and ChIP-seq peak shape clustering [61] . Here , applying FDA methodology , we address three questions about the biology of ERVs . First , what genomic features are significant for ERV integration and fixation ? Second , at what genomic scales and locations are these features influential ? Third , and finally , how do genomic features work in concert to provide signals essential for integration and fixation of ERVs ? Using genome-wide data , we applied the recently developed Interval Testing Procedure ( ITP ) [62] to determine the influence of flanking sequence features on integration and fixation of mouse ( polymorphic and fixed ETns and IAPs ) and human ( fixed and in vitro HERV-Ks ) ERVs . As a result , we detected diverse genomic features that affect integration and fixation of these elements ( e . g . gene content , replication timing , AT count , and LINE content ) , and did so differentiating effects at various scales and locations in the flanking regions . Finally , we employed multiple Functional Logistic Regression ( FLR ) models to capture the combined effects of a restricted set of features resulting in a compact group of genome features that define the genomic landscape of integration and/or fixation preferences for ERVs . Importantly , the functional testing procedures and regression techniques we extended , employed and demonstrated in this study can be broadly applied in genomics .
In this study we analyzed in vitro , polymorphic , and fixed ERVs . The distributions of in vitro and polymorphic ERVs are only marginally influenced by selection and thus provide a more accurate view of integration preferences . Fixed ERVs , in contrast , carry information about both integration and fixation . We interrogated the genomic neighborhoods ( 32-kb flanking sequences upstream and 32-kb flanking sequences downstream of each element , so there is no overlap among flanking regions to maximize the number of ERVs in the study ) of one human and two mouse ERV families . In mouse , we considered 1 , 866 ETns ( 242 polymorphic and 1 , 624 fixed ) and 5 , 950 IAPs ( 1 , 986 polymorphic and 3 , 964 fixed ) detected genome-wide by Zhang and colleagues [25]; elements were considered to be fixed if they were shared among four mice strains , and polymorphic if they were present in the C57BL/6J strain but not in the other three strains ( see Methods ) . As control regions , we considered 1 , 379 continuous 64-kb regions of the mouse genome that did not overlap with the flanking sequences of ERVs ( see Methods , Table 1 ) . In human , we considered 826 fixed HERV-Ks ( Table 1 ) annotated by Subramanian and colleagues [63] . We also extracted the genomic locations of 1 , 065 in vitro HERV-K integrations in human embryonic kidney and fibrosarcoma cell lines [34] ( Table 1 ) . A total of 1 , 690 control regions were generated similarly to those in mouse ( see Methods ) . Human and mouse ERVs in our analyses ranged from solo-LTRs ( ~60 bp ) to full-length elements ( ~9 kb ) ( Table A in S1 Text ) . The number of ERVs present on each chromosome correlated with chromosome size ( Fig A in S1 Text ) . Human chromosome 19 was an outlier with an overrepresentation of fixed HERV-Ks ( Fig A in S1 Text ) . We selected a diverse set of genomic features ( Table 2 ) that could be implicated in ERV integration or fixation as reported by previous ERV [31–33] and non-ERV TE studies [45 , 46] . In total , we considered 41 and 43 genomic features in mouse and human ERV flanking regions , respectively ( derived from 43 datasets in mouse and 44 datasets in human ) . These features reflected DNA conformation ( e . g . , G-quadruplex ) , DNA sequence ( e . g . , A/T content ) , position on the chromosome ( e . g . , distance to the closest centromere and telomere ) , recombination ( e . g . , local recombination rates ) , replication ( e . g . , replication timing ) , gene regulation and expression ( e . g . , histone marks and DNase I hypersensitive sites ) , as well as selection ( e . g . , exons and most conserved elements ) . Where possible , we specifically utilized features studied in embryonic stem cells ( ESCs ) or in sperm cells as they most closely proxy characteristics of germ-line and embryonic cells [64] . Four low-resolution features ( replication timing , recombination rates , distance to telomere , and distance to centromere ) were represented by a single value for each 64-kb region . For each high-resolution feature , we measured either its content ( fraction of the genomic window covered by the feature ) , its count or its weighted average ( WA , only for methylation and expression features ) in each of the 64 1-kb windows constituting the flanks of each ERV and each control region ( Fig 1A , see Methods ) . We applied hierarchical clustering to screen out high-resolution genomic features that present strong correlations with each other ( Figs B and C in S1 Text ) . For example , for human , exon content was highly correlated with gene expression in ESCs and thus we removed the latter from the analysis . As a result , a total of 35 mouse and 36 human high-resolution genomic features ( derived from 35 datasets in mouse and 37 datasets in human ) were retained for further analysis ( Figs D and E in S1 Text ) . To identify genomic features significantly affecting the ERV distributions in the human and mouse genomes , we contrasted flanking regions of fixed ERVs ( either mouse ETn and IAP , or human HERV-K ) vs . control regions; such a comparison is expected to reflect both integration and fixation preferences ( Fig 1B ) . In an attempt to disentangle genomic features affecting ERV integration from those affecting their fixation , we conducted additional comparisons; namely , we contrasted flanking regions of polymorphic mouse ERVs ( ETn and IAP ) vs . mouse control regions , and flanking regions of in vitro HERV-K vs . human control regions . In these comparisons integration preferences are expected to be substantially more prominent than fixation preferences because selection had substantially less time to act on polymorphic or in vitro ERVs than on fixed ERVs . Finally , to pinpoint genomic features significant for ERV fixation , we contrasted flanking regions of fixed vs . polymorphic mouse elements ( ETns and IAPs ) , and of fixed vs . in vitro HERV-Ks . In a way , the analysis of fixed ERVs vs . controls can be viewed as “cumulating” that of polymorphic or in vitro ERVs vs . controls , and that of fixed vs . polymorphic or in vitro ERVs . In total , we conducted nine comparisons ( Fig 1B ) , each using four different statistical techniques as described below ( Fig 1C ) . Admittedly , polymorphic integrations are affected by selection to a greater degree than in vitro ones , however we are not in possession of both of these data types for the species in our study; only polymorphic data are available for mouse ERVs and only in vitro data are available for human ERVs . First , we tested whether ERV presence was significantly affected by low-resolution features using a univariate permutation test ( see Methods , Table B in S1 Text; Fig 2 ) ; this is appropriate because these features are represented by a single value for each 64-kb region . Second , for the high-resolution genomic features , we employed the two-population Interval Testing Procedure ( ITP ) for functional data [62] to assess whether each feature , when considered alone , had significantly different content ( or count , or WA ) in a comparison , e . g . in ERV flanking regions vs . controls ( see Methods for details ) . This technique is particularly suitable for our analysis because it considers the data as a curve over the 64 1-kb windows comprising each region , instead of taking one value for the region ( e . g . , an average over the 64 windows ) . ITP combines inference on the whole curve with component-wise inference ( i . e . inference on measurements comprising the curve ) . Thus , it allows us to select relevant genomic features detecting both the scale and the location at which each feature acts ( see Methods for more details ) . From this analysis we expect to detect genomic features that: ( 1 ) show significant enrichment/depletion locally , especially in windows close to the IntS of ERVs or further away from it–we call these localized differential landscape ( LDL ) features ( e . g . Fig 3A ) ; ( 2 ) show a uniform level of significant enrichment/depletion throughout all 64 1-kb windows–we call these invariant differential landscape ( IDL ) features ( e . g . Fig 3B ) ; or ( 3 ) are not significant over the whole 64-kb region examined . In order to capture different nuances of the data , we performed ITP using three test statistics ( mean difference , median difference , and variance ratio; Figs 4–6 and F-T in S1 Text ) , however , below we focus on results concerning mean differences . Third , to determine the individual explanatory power of major predictors , we fitted single Functional Logistic Regressions ( FLRs; Figs 1C and 7 , Table C in S1 Text ) for each feature found to be significant in ITP . This analysis allowed us to summarize and better quantify the results obtained by univariate permutation test and ITP ( Fig 7 ) . Moreover , it allowed us to identify features that , by themselves , explained a percent of deviance in excess of 20% . These are clearly very relevant predictors ( Table C in S1 Text ) but we did not include them in our final multiple FLR models ( see below ) as they would hide the concurrent effects of other potentially relevant predictors . Fourth , we examined the joint effects of the remaining significant predictors ( as determined by ITP and univariate permutation test ) via multiple Functional Logistic Regression ( Fig 1C and Tables 3–11 ) . The multiple FLR models expressed the likelihood of being in the neighborhood of an ERV vs . control ( or of a fixed vs . a polymorphic mouse ERV , or of a fixed vs . in vitro HERV-K ) as a joint function of several predictors . In particular , IDL features and low-resolution features that proved significant in univariate permutation tests were treated as scalar predictors represented by their averages across the 64 windows constituting each region . In contrast , LDL features were treated as functional predictors with curves evaluated at customized scales and intervals to capture the specific behavior of each LDL feature , e . g . around the IntS , as suggested by the ITP . Importantly , the modified ITP we employed ( see Methods ) gave us detailed information on the best scale and location , i . e . on the subregions on which to study the curve , for each of these functional predictors . To identify genomic features affecting the distribution of ETns in the mouse genome–as a result of both integration and fixation preferences of these elements–we contrasted flanking sequences of fixed ETns vs . control regions . Univariate permutation tests applied to the low-resolution features ( Fig 2A and Table B in S1 Text ) indicated that the flanking regions of fixed ETns have lower recombination rates and later replication timing , and are closer to centromeres . ITP indicated ( Fig 4A and Fig F in S1 Text ) that all four microsatellites types , LINEs , L1 target sites , AT count , and the H3K9me3 histone mark are overrepresented , while SINEs , exons , introns , CpG islands , most conserved elements , all features associated with CpG methylation , ESC expression , and two histone marks ( H3K27me3 and H3K36me3 ) are underrepresented , throughout the whole fixed ETn flanking regions . ITP also identified features with interesting localized behaviors: recombination hotspot content and the H3K27ac histone mark are overrepresented immediately next to the IntS , while DNase I hypersensitive sites and three histone marks ( H3K4me1 , H3K4me3 , and H3K9ac ) are underrepresented everywhere except for immediately next to the IntS . We found that also Z-DNA repeats and G-quadruplex repeats are underrepresented , with a yet more complex local behavior . Next , single FLRs revealed that all four microsatellites types , LINEs , introns , most conserved elements , and ESC expression have very strong effects ( relative contribution to the deviance explained , RCDE ≥26%; Fig 7A and Table C in S1 Text ) . We therefore excluded these predictors from the final multiple FLR model , which explained 51 . 1% of the deviance in discriminating fixed ETns from controls ( Table 3 ) . The two strongest scalar predictors in such a model ( i . e . H3K9me3 and L1 target sites , RCDE 20 . 5% and 23 . 5% , respectively ) had positive effects , while AT count and H3K27me3 had negative effects ( RCDE 17 . 8% and 4 . 7% , respectively; see Discussion for the explanation of the negative effect of AT count in multiple FLRs ) . The two strongest functional predictors–CpG methylation and exon content ( RCDE of ~10% ) –had negative effects . Interestingly , the H3K4me1 mark ( RCDE 5 . 1% ) had a strong positive effect strictly localized at the IntS ( -4 to 4 kb ) and a negative effect away from the IntS , while mirror repeats ( RCDE 3 . 1% ) had a positive effect on the whole region . To highlight integration preferences , we contrasted flanking sequences of polymorphic ETns vs . control regions . In this analysis ( Fig 4B and Fig G in S1 Text ) we found many similarities but also a number of noteworthy differences with respect to the analysis of fixed ETns vs . controls ( Fig 4A and Fig F in S1 Text ) . For instance , similar to fixed ETns , the flanking sequences of polymorphic ETns appeared to replicate later than controls ( Table B in S1 Text and Fig 2A ) suggesting that this feature might be important for ETn integration . The underrepresentation of exons , CpG islands , and several histone marks was weaker for polymorphic ETns vs . controls than for fixed ETns vs . controls suggesting that ETns can integrate but tend not to become fixed in such environments ( Fig 4B and 4A ) . Moreover , the content of DNase I hypersensitive sites did not differ significantly from controls in a relatively large region surrounding the IntS for polymorphic ETns ( -6 kb to +16 kb; Fig 4B ) but only at 1 kb upstream from the IntS for fixed ETns ( Fig 4A ) . Single FLRs displayed very similar explained deviances for microsatellites and LINEs in the two comparisons , while the explained deviances for intron content and most conserved elements content were lower for polymorphic ETns vs . controls than for fixed ETns vs . controls ( Fig 7A , Tables 4 and 3 ) –confirming that the former are subject to weaker selection effects . The final multiple FLR model for polymorphic ETns vs . controls explained 48 . 4% of the deviance ( Table 4 ) and was similar to the model for fixed ETns vs . controls ( Table 3 ) . Next , to highlight fixation preferences , we contrasted flanking regions of fixed vs . polymorphic ETns . This analysis , too , revealed some similarities and some important differences relative to that of fixed ETn vs . controls . Fixed ETns were located in regions with lower recombination rates and closer to centromeres compared to polymorphic ETns ( Table B in S1 Text and Fig 2A ) suggesting that these features are important for fixation . The ITP tests contrasting fixed vs . polymorphic ETns ( Fig 4C and Fig H in S1 Text ) did not identify microsatellites as features differentiating fixation and integration propensities . However , they did reveal a more localized underrepresentation of exons and most conserved elements around the IntS ( Fig 4C ) –as compared to the ITP tests contrasting fixed ETns vs . controls ( Fig 4A ) . Recombination hotspots , which were overrepresented in polymorphic ETns vs . controls ( Fig 4B ) and strongly overrepresented near the IntS in fixed ETns vs . controls ( Fig 4A ) , were underrepresented in fixed vs . polymorphic ETns ( Fig 4C ) . Also , several histone marks ( H3K9me3 , H3k4me1 , and H3K9ac ) were overrepresented near the IntS in fixed vs . polymorphic ETns ( Fig 4C ) , but did not show a significant difference near the IntS in fixed ETn vs . controls ( Fig 4A ) . Interestingly , DNase I hypersensitive sites were overrepresented -1 kb upstream and underrepresented up to 2 kb downstream from the IntS ( Fig 4C ) . Single FLRs did not identify features which , individually , had great strength in characterizing fixed vs . polymorphic ETns ( all explained deviances <3 . 5% ) ( Fig 7A and Table C in S1 Text ) . However , taken together in the context of multiple FLR , eight features explained 15% of the deviance in discriminating fixed vs . polymorphic ETns ( Table 5 ) and reiterated most of our observations from the ITP tests ( Fig 4C ) . Interestingly , the genomic features significant in distinguishing between the flanking sequences of fixed IAP and control regions were very similar to those identified in the analogous comparison for ETns . For instance , just as fixed ETns ( Table B in S1 Text and Fig 2A ) , fixed IAPs were found in regions with lower recombination , later replication , and smaller distance to the centromere than controls ( Table B in S1 Text and Fig 2B ) . Most predictors found to be significant by the ITP tests were also shared between the fixed IAPs vs . controls and the fixed ETns vs . controls comparisons ( Figs 5A and 4A , respectively ) . Several noteworthy exceptions included a uniform underrepresentation of H3K27ac throughout the 64 kb region flanking fixed IAPs ( Fig 5A and Fig I in S1 Text ) , while this histone mark was overrepresented near the IntS of fixed ETns ( Fig 4A and Fig H in S1 Text ) . Transcription start sites were underrepresented for fixed IAPs , while they were not significant for fixed ETns . Also , three histone marks ( H3K4me1 , H3K4me3 , and K3K9ac ) , as well as DNase I hypersensitive sites , were uniformly underrepresented throughout the 64 kb region flanking fixed IAPs , but not significantly different from controls immediately next to the IntS for ETns ( Fig 5A , Fig I in S1 Text , Fig 4A , and Fig H in S1 Text ) . Major individual predictors , as identified by single FLRs , were also remarkably similar between fixed IAPs ( Table C in S1 Text and Fig 7B ) and fixed ETns ( Table C in S1 Text and Fig 7A ) –with only one extra predictor for fixed IAPs; H3K9me3 content . The multiple FLR model had a 33 . 3% deviance explained ( Table 6 ) and again showed many similarities to the analogous model for fixed ETns ( Table 3 ) . Next , following a logic similar to the one used above for ETns , we attempted to separate integration vs . fixation preferences for IAPs . Just as was observed for ETns , genomic features significant in distinguishing the flanking sequences of polymorphic IAPs from control regions ( Fig 5B and Fig J in S1 Text ) were very similar to those distinguishing the flanking sequences of fixed IAP from control regions ( Fig 5A and Fig I in S1 Text ) . The underrepresentation of Z-DNA repeats was more localized ( close to the IntS ) , while recombination hotspot content was more uniformly overrepresented in polymorphic IAPs vs . controls than in fixed IAPs vs . controls . Moreover , transcription start sites showed no significant differences and the H3K27ac histone mark was only weakly underrepresented in polymorphic IAPs vs . controls , while these features were more prominent in fixed IAPs vs . controls suggesting their importance for fixation rather than integration . Single FLRs identified the same group of major predictors for polymorphic IAPs vs . controls as were identified for polymorphic ETns vs . controls , with the addition of the functional predictor ESC expression ( explained deviance 22% ) ( Fig 7B and Table C in S1 Text ) . We excluded these features from the final multiple FLR model , which explained 30 . 4% of the deviance in discriminating polymorphic IAPs vs . controls ( Table 7 ) and was similar to the analogous model for ETns ( Table 4 ) . When comparing fixed and polymorphic IAPs ( Fig 5C and Fig K in S1 Text ) , the ITP identified only three features that had lower means throughout the fixed IAP integration regions–SINEs , DNase I hypersensitive sites , and recombination hotspots . However , striking differences were observed for 12 genomic features with localized landscape . For instance , fixed IAPs revealed strong signatures of depressed means within ±4 kb from the IntS for features such as dinucleotide and trinucleotide microsatellites , most conserved elements , unmethylated CpGs , and the histone mark H3K27me3 . Introns and two histone marks ( i . e . H3K4me1 and H3K27ac ) were also underrepresented in a larger area around the IntS . Moreover , we observed overrepresentation of LINEs and H3K9me3 ( the mark of heterochromatic regions ) surrounding the IntS in fixed vs . polymorphic IAPs . Exon content was underrepresented in a few windows downstream of the IntS , while AT count was overrepresented in most windows of both flanks except at the IntS . Single FLRs for fixed vs . polymorphic IAPs identified H3K9me3 as a very strong functional predictor ( explained deviance of 21 . 5% , Fig 7B and Table C in S1 Text ) . All other features , singularly , could explain only a very low portion of the deviance ( <2% ) , and even their concurrent effect was low , producing a multiple FLR model with an explained deviance of 3 . 8% ( Table 8 ) . Flanking sequences of fixed HERV-Ks were characterized by lower recombination rates and later replication timing than controls ( Table B in S1 Text and Fig 2C ) . Based on the ITP ( Fig 6A ) , microsatellites ( of all four types ) , LINEs , recombination hotspots , L1 target sites , AT count , and H3K9me3 marks were overrepresented , while G-quadruplex repeats , SINEs , replication origins , cytosine methylation level features ( i . e . CpG , CHG , and CHH ) , and four histone marks ( H3K36me3 , H3K4me1 , H3K27ac , and H3K9ac ) were underrepresented , throughout these flanking sequences . Exons , introns , and most conserved elements were underrepresented as well , especially near the IntS . Single FLR fits revealed that all four microsatellites types ( scalar predictors ) , introns , and H1-hESC transcript expression , considered as functional predictors , each individually explained 26–76% of the deviance . These features were excluded from the final multiple FLR model ( Fig 7C and Table C in S1 Text ) , which explained 79% of the deviance ( Table 9 ) . The model included LINEs and recombination hotspots as positive scalar predictors ( RCDE of 13% and 9 . 7% , respectively ) . In terms of functional predictors , L1 target sites was the strongest predictor with a positive effect in the whole region from -30 to 30 kb , and stronger away from the IntS ( RCDE 27 . 1% ) . In addition , AT count had a negative effect stronger away from the IntS ( RCDE 12 . 6% ) , hypomethylation in testis and CpG islands had positive and negative effects , respectively , for the whole integration region ( both RCDE of ~5% ) , and G-quadruplex repeats had a negative effect near and upstream of the IntS ( RCDE 1 . 9% ) . To highlight HERV-Ks’ integration preferences , we contrasted flanking sequences of in vitro HERV-Ks vs . control regions . The former replicated later and were more distant from the centromere than the latter ( Table B in S1 Text and Fig 2C ) . The ITP indicated ( Fig 6B ) that G-quadruplex repeats , L1 target sites , recombination hotspots , replication origins , CHH methylation , and four histone marks associated with active transcription or promoters ( H3K36me3 , H3K4me1 , H3K27ac , and H3K9ac ) were underrepresented , while LINEs , AT count , and the H3K9me3 mark were overrepresented , throughout the flanking sequences compared to control regions . Additionally , the ITP indicated that there were fewer CHG methylated sites near the IntS of in vitro HERV-Ks than in control regions . Other genomic features had significant differences further away from the IntS ( e . g . , SINEs , TSS ENCODE , H1-hESC transcript expression , and H3K27me3 ) , or showed more complex localized behaviors ( Fig 6B and Fig M in S1 Text ) . The multiple FLR model for in vitro HERV-Ks vs controls explained 24 . 4% of the deviance ( Table 10 ) . To focus on HERV-K fixation preferences , we contrasted flanking sequences of fixed vs . in vitro HERV-Ks . Recombination rates , replication timing , and distance to centromere were significantly different , with lower means for fixed compared to in vitro HERV-Ks ( Table B in S1 Text and Fig 2C ) . The ITP ( Fig 6C and Fig N in S1 Text ) indicated that microsatellites , L1 target sites , and recombination hotspots were overrepresented throughout the flanking sequences of fixed vs . in vitro HERV-Ks . Introns and most conserved elements were underrepresented throughout the whole region too , but the difference was stronger near the IntS . Exons were underrepresented ±2 kb next to the IntS . Moreover , three histone marks ( H3K4me3 , H3K27ac , and H3K9ac ) were underrepresented upstream of the IntS , while the H3K9me3 mark was overrepresented downstream of the IntS . CpG and CHG methylated sites , as well as H3K36me3 , were generally underrepresented but had more complex localization patterns . Single FLRs identified six positive scalar predictors–all four microsatellites types , recombination hotspots , and L1 target sites–each of which , individually , explained as much as 85% of the deviance ( Fig 7C and Table C in S1 Text ) .
We explored the associations between ERVs and a diverse set of non-B DNA conformation predictors inferred from the primary DNA sequence of the human and mouse genomes . In vivo , such conformations are formed transiently during recombination , repair , transcription , and replication , frequently causing genomic instability [79] and were found to be associated with the presence of DNA transposons [44 , 46] . We observed that mirror repeats and A-phased repeats are overrepresented in the flanking regions of fixed ETns , as well as of fixed and polymorphic IAPs , as compared with control regions ( A-phased repeats are also overrepresented in the vicinity of fixed HERV-Ks ) . The overrepresentation of these repeats in the flanking regions of both fixed and polymorphic IAPs suggests their role in ERV integration–the lack of significant overrepresentation for polymorphic ETns perhaps being due to a more limited statistical power , given the smaller sample size ( Table 1 ) . A subset of mirror repeats–triplex repeats–are thought to bind mismatch and nucleotide excision repair proteins [79] , therefore we propose that these protein complexes might be recognized by the integrase machinery and trigger ERV integration . This hypothesis needs to be evaluated experimentally . Mirror repeats have also been associated with low gene expression levels [79] . In agreement with this we found ERVs enriched in regions with low levels of transcription ( see below ) . A-phased repeats cause double helix bends that have been implicated in nucleosome assembly and expansion of trinucleotide microsatellites [65 , 80] and might be important for the recognition of IntSs by the ERV integrase , as suggested by retroviral studies [81–84] . Moreover , unlike mirror repeats that do not have base composition bias , A-phased repeats are adenine-rich [65] , resonating with the effect of A/T nucleotides ( see below ) . G-quadruplex and Z-DNA repeats displayed negative associations with the ERVs . G-quadruplex repeats are underrepresented in fixed vs . control and in vitro ( or polymorphic ) vs . control contrasts for both HERV-Ks and IAPs and thus likely inhibit ERV integration . Z-DNA repeats might inhibit ERV integration as well , because they are underrepresented in the flanking regions of both fixed and polymorphic IAPs . Importantly , these two types of repeats appear to be inhibitive of ERV integration beyond their GC-rich composition [65] because in several of our models they appear as significant predictors of ERV distributions together with AT-content ( Tables 3 and 9 ) . Both G-quadruplex and Z-DNA repeats are enriched in promoters and in the 5’ and 3’ gene termini [79] , and therefore we cannot exclude the possibility that purifying selection removes ERVs from such regions . We observed a strong overrepresentation of all four types of microsatellites in the fixed and polymorphic mouse ERVs compared to controls , suggesting the importance of microsatellites for ERV integration . Many microsatellites form non-canonical DNA structures–e . g . , ( AG ) n repeats form triplexes , ( AT ) n form four-stranded cruciforms , while ( CA ) n and ( GC ) n form Z-DNA [85 , 86]–which lead to genome instability [87] and may be used by the integrases to recognize potential IntSs . In the human genome , we found an enrichment of microsatellites of all four types for fixed HERV-Ks compared with controls , and for fixed vs . in vitro HERV-Ks . Therefore , microsatellites might not be directly relevant to HERV-K integration , and instead be more relevant for their fixation ( caution should be exercised though when comparing the results from in vitro vs . control and polymorphic vs . control contrasts , as the latter are more influenced by selection ) . The flanking regions of in vitro HERV-Ks actually had an underrepresentation of mononucleotide microsatellites , potentially because such microsatellites are frequently present as ( A/T ) n repeats located at the 3’ ends of retrotransposed genes–where HERV-K integrations might be selected against . Additionally , ( A/T ) n repeats are found in Alus that are also underrepresented in the in vitro HERV-K flanking regions ( see below ) . Corroborating previous studies [24 , 32] , we observed that both mouse and human ERVs integrate and become fixed in AT-rich genomic regions . Indeed , AT-content was a significant predictor in eight out of nine ITP contrasts ( except for in vitro HERV-K ) . Moreover , L1 target sites [88] are overrepresented in the flanking regions of polymorphic and fixed mouse ERVs as compared to controls , suggesting that these sequences correlate with mouse ERV integration . L1 target sites are also overrepresented in the flanking regions of fixed HERV-Ks vs . controls , and of fixed vs . in vitro HERV-Ks , but are underrepresented in the flanking sequences of in vitro HERV-Ks vs . controls , thus suggesting that these sequences correlate with HERV-K fixation . The enrichment of L1 target sites in the vicinity of ERVs can be explained by the high AT-content of these sequences , and thus it might simply reflect the AT-richness of the genomic regions in which ERVs integrate or are fixed , or perhaps it also indicates the enrichment of LINEs in these flanking regions ( see below ) . Both AT nucleotides and L1 target sites have positive effects as single predictors , however , in the context of multiple FLR , when the two predictors are considered jointly , we observed a negative effect of AT nucleotides . Indeed , given the high correlation of the two predictors ( Figs B and C in S1 Text ) , the positive effect shown by AT nucleotides when considered on their own is covered by a statistically dominant effect of L1 target sites in the joint analysis . Mouse and human ERVs tend to be present in genomic regions rich in LINEs , but depleted of SINEs . This trend is significant for the fixed vs . control and polymorphic ( or in vitro ) vs . control contrasts , but not for the fixed vs . polymorphic ( or in vitro ) contrasts , arguing for a link with ERV integration . An association of ERVs with LINEs might reflect a preference towards integration in AT-rich sequences for both types of TEs . SINEs , in contrast , accumulate in GC-rich regions of the genome [12 , 89] . Additionally , a strong positive relationship between LINEs and ERVs , as evidenced by ITP and FLR , could also be explained by H3K9me3 histone marks known to be located in regions rich in these TEs [90] ( see below ) . An important consideration in our analysis is that control regions are almost completely depleted of ERVs ( see Methods ) , as we excluded even older ERVs than the ones we studied here from our control regions; this might underestimate the influence of such ERVs in integration preferences of the studied ERVs . In our study , ERV flanking regions had a late replicating profile . More specifically , the flanking regions of fixed ERVs replicated later than those of polymorphic ERVs , and the flanking regions of polymorphic ERVs replicated later than control regions . This was true for all three types of ERVs studied . Moreover , the flanking regions of both fixed and in vitro HERV-Ks presented a low content of replication origins–also a signature of late replication [91] . We hypothesize that the ERV integrase machinery targets late replicating regions because they are AT-rich [92] or that ERV integration might be coordinated with DNA replication , similar to Tf1 retrotransposon integration at stalled replication forks [93] and as proposed for DNA transposons [46] . Recombination appears to be important for both integration and fixation of ERVs . The flanking sequences of fixed ERVs have lower recombination rates than those of polymorphic ( or in vitro ) ERVs and of controls ( Fig 2 ) , suggesting a signature of fixation preference . This observation is in concordance with the hypothesis that ERVs are removed from highly recombining regions via ectopic recombination [6 , 31 , 94] . Alternatively , drift might fix ERVs in low recombining regions of the genome where selection is weaker . Katzourakis and colleagues [31] did not find a correlation between HERV fixation and recombination rates , but the discrepancy between our results and theirs might be due to the different HERV families investigated in the two studies . We also observed that polymorphic and fixed mouse ERVs are located in genomic neighborhoods with higher content of recombination hotspots than controls , suggesting a role of recombination hotspots in mouse ERV integration . The overrepresentation of hotspots right next to the IntS for polymorphic ETns further supports this observation . In human , Myers and colleagues [68] detected overrepresentation of two retrovirus-like elements ( THE1A and THE1B ) in regions enriched with recombination hotspots . Moreover , for DNA transposons , it has been proposed that recombination hotspots are required by the transposition mechanism [46] , and perhaps a similar interaction is essential for ERVs . In contrast , the comparison of fixed vs . polymorphic mouse ERVs indicated an underrepresentation of recombination hotspots suggesting that their high concentration prevents fixation . HERV-Ks presented an almost opposite pattern: recombination hotspots were overrepresented when contrasting fixed HERV-Ks vs . controls and fixed vs . in vitro HERV-Ks , and were underrepresented when contrasting in vitro HERV-Ks vs . controls , arguing for an association between recombination hotspots and HERV-K fixation but not integration . Note that we found experimentally validated recombination hotspots [67] to be significant in mouse , and predicted ones [66] to be significant in human , potentially explaining some differences in the results ( experimentally validated recombination hotspots in human [68 , 69] ) were not significant in our models ) . We found that fixed mouse ERVs are located closer to centromeres than polymorphic mouse ERVs , and that the latter are located closer to centromeres than control regions . The preferential location of ERVs next to centromeres might be explained by their integration in AT-rich regions and by their fixation in regions with low recombination rates ( see above ) . Indeed , recombination rates and GC-content , which are highly correlated with each other , are markedly decreased near centromeres [95] . In contrast , only small differences in the distance to centromere were observed among fixed HERV-Ks , in vitro HERV-Ks , and control regions . Both methylated and unmethylated CpG sites were underrepresented in the flanking regions of fixed and polymorphic mouse ERVs compared to controls , reflecting integration of these elements in AT-rich areas of the genome . Interestingly , hypomethylated CpGs representing contiguous domains of low methylation measured in sperm ( mean size of 1 . 8 kb ) [75] , are overrepresented in the flanking regions of fixed HERV-Ks compared to controls and to the flanking regions of in vitro HERV-Ks , suggesting a link with HERV-K fixation . Hypomethylated CpGs overlap with promoters and other regulatory regions [75] and thus selection might not tolerate ERVs in such areas of the genome . All three ERV families studied tend to occur in regions with low DNase I hypersensitive sites content , i . e . in areas with closed chromatin . However , our results are conflicting as to whether this is an integration or fixation preference . Indeed , the signal for HERV-Ks is generally weak , in contrast to Brady and colleagues [34] who reported integration close to DNase I cleavage sites for in vitro HERV-K ( i . e . HERV-Kcon ) . There is no significant signal near the IntS in the polymorphic ETns vs . controls contrast , and the signal in the fixed vs . polymorphic ETns comparison is also weak ( Fig 4 ) . For IAPs , the signal is stronger in the fixed and polymorphic vs . control contrasts than in the fixed vs . polymorphic contrast , arguing for a link with integration . Integration in areas with closed chromatin was previously observed for several retroviruses ( e . g . , MLV and HIV ) , and it was proposed that the nucleosomal DNA is targeted for integration by retroviral integrases [81 , 96] . As reported previously for retroviruses [97] , histone marks are important predictors of ERV distributions . Overall , our results corroborate previous findings ( e . g . , [98] ) suggesting that ERVs integrate in areas of repressed chromatin . Consistent with previous studies [38 , 90 , 99–101] , we observed an underrepresentation of histone marks associated with transcribed chromatin ( H3K36me3 and H3K9ac ) , promoters ( H3K4me3 ) , and enhancers ( H3K4me1 and H3K27ac; [102] ) in the flanking regions of fixed and polymorphic ERVs . In agreement with this observation , we observed an overrepresentation of the H3K9me3 mark specific to repressed chromatin [102 , 103] in the flanking regions of both fixed and polymorphic elements . Surprisingly , and at odds with other studies ( e . g . , investigating youngest ETns and IAPs [90] , and ERV-Ls [104] ) , we observed an underrepresentation of the H3K27me3 mark associated with repressed chromatin in the flanking regions of ERVs . What can explain the opposite results we obtained for H3K9me3 and H3K27me3 , two marks of repressed chromatin which are one enriched and the other one depleted in the flanking regions of ERVs ? While both marks signal repressed chromatin in early embryonic development , they are found in different regions of the genome [105] . On the one hand , the H3K9me3 mark is associated with heterochromatin formation due to the presence of tandem repeats [105] , and thus its enrichment in the flanking regions of ERVs supports the strong association we found between ERVs and microsatellites . On the other hand , the H3K27me3 mark is abundant in gene-rich regions [105] , and thus its depletion in the flanking regions of ERVs might reflect purifying selection acting against ERV integration in or around genes . Also , our results might be specific to the ESCs–as they agree with those of Hiratani and colleagues [106] who also studied ESCs . Some differences between our study and that of Brady and colleagues [34] might be explained by the use of histone modification data from different cells ( Brady and colleagues utilized the data generated for CD4+ cells [99] ) . Finally , some of the differences among studies may be due to differences in the protocols used to construct control data sets . Our results concerning ERVs fixation preferences ( from the fixed vs . polymorphic , or in vitro , contrasts ) with respect to histone marks suggest that ETns are fixed in areas rich in enhancer marks ( H3K4me1 and H3K27ac ) and this signal is localized at the IntS , while IAPs show an opposite trend ( depletion of these marks ) that is less localized . ETns bind transcription factors [107] and therefore could act as enhancers attracting H3K27ac , however we would like to see our observation of an association between ETns and enhancer marks validated in subsequent studies including histone marks from different ESC . In our analysis , exons , introns , and most conserved elements were underrepresented in the flanking regions of fixed and polymorphic ( or in vitro ) ERVs vs . controls , as well as in the flanking regions of fixed vs . polymorphic ( or in vitro ) ERVs . This may be evidence of purifying selection acting against integration and fixation of ERVs in areas of the genome rich in genes and most conserved elements . The signal was frequently weaker in the polymorphic ( or in vitro ) vs . control contrast , and localized close to the IntS in the fixed vs . polymorphic contrast . Our observation that mouse polymorphic ERVs are more prevalent in gene-rich regions than mouse fixed ERVs corroborated the findings of Zhang and colleagues [25] . Our results for fixed ERVs are also in agreement with the results of Medstrand and colleagues [32] who observed that older ERVs are underrepresented within 5 kb from genes . A selective purge from areas of the genome that are rich in genes and most conserved elements appears to be a characteristic of most TEs [24 , 32 , 36 , 45 , 46] . Alus , preferentially integrating within genes [42 , 45] , appear to be an exception . We also observed that ERVs are more prevalent in areas with low levels of transcription–likely reflecting both an integration preference for repressed chromatin and a fixation preference for gene-poor neighborhoods . This result contrasts observations made by Brady and colleagues [34] . These differences are probably due to the strict selection of control regions in our analyses and the different methods of analysis . Some of the genomic features analyzed here were included in previous studies of ERV distributions , and our results are largely consistent with such studies [97 , 108–110] . However , our analysis included the most comprehensive list of genomic features to date , with many features not considered in previous studies ( e . g . , replication , most conserved elements , diverse DNA conformation features , L1 target sites , and recombination hotspots for human and mouse ) . Moreover , our use of FDA techniques allowed us to effectively investigate localization and scale at which the effects of these features unfold . As a result , we are in a position to propose a comprehensive model for the integration and fixation preferences of the mouse and human ERVs considered in our study ( Fig 8 ) . ERVs integrate in regions of the genome with high AT-content , enriched in A-phased repeats ( as well as mirror repeats for mouse ERVs ) and microsatellites–the former possessing and the latter frequently presenting non-canonical DNA structure . This highlights the potential importance of unusual DNA bendability in ERV integration , in agreement with previous studies [96 , 111] . Interestingly , some non-canonical DNA structures–e . g . , G-quadruplexes and Z-DNA–appear to inhibit ERV integration ( with Z-DNA avoidance strictly localized at the IntS site for polymorphic IAPs ) . ERV integration regions are rich in LINEs , replicate late , and have low density of replication origins . Replication hotspots might assist integration of mouse ERVs ( with the signal localized close to the IntS for ETns ) . Our results on histone marks and DNase I hypersensitive sites suggest that integration occurs in gene-poor regions with closed chromatin–mimicking the behavior of retroviruses , whose integrases were proposed to target nucleosomal DNA [81 , 96] . Our observations also indicate that ETns and IAPs differ locally in their sensitivity to certain histone marks associated with promoters or expression and open chromatin , with the flanking regions of IAPs completely depleted of these features . However , for ETns these features are not significant next to the IntS . The ERVs in our study are preferentially fixed in areas of the genome where both genes and most conserved elements are scarce ( the signal is frequently stronger next to the IntS ) . There is also a strong preference of ERVs to be fixed in areas with low recombination rates , likely because in such areas they have a lower probability of being removed via ectopic recombination . The initial integration occurs preferentially relatively close to the centromere , likely because of increased AT-richness , and fixed ERVs are located even closer to the centromere , likely because of reduced recombination rates . ETns also show a fixation preference for areas rich in enhancer histone marks ( alternatively they might induce the formation of these marks ) , and this signal is localized at the IntS . Utilizing data on the distributions of fixed , polymorphic , and in vitro ERVs , and comparing their flanking regions with parts of the genome devoid of ERVs ( controls ) and with each other , we were able to separate , at least partially , the integration and fixation preferences of these elements . Ideally , one would like to utilize all three groups of flanking regions plus controls to study each element family . However , we were only in possession of data on polymorphic elements in the mouse genome , and of data on in vitro elements in the human genome . Because of this , observed patterns may not be entirely comparable; in particular , polymorphic mouse ERVs may experience more purifying selection than in vitro human ERVs . Moreover , one drawback of using in vitro integrations is that these are produced in cell lines using ERV constructs . The complexity of our study of ERV biology reflects the intricacy of the genome and of the mechanisms affecting the integration and fixation of these elements . We found evidence of the existence of genomic features associated with both insertion sites and the ability of elements to be fixed in the genome . Our analysis leveraged the availability of data on polymorphic ERVs and in vitro integrations , which we could contrast with fixed elements; we suggest applying a similar strategy to future studies of TE distribution . We expect further refinements of our model as more genomic data become available , e . g . , nuclear lamina interactions and Hi-C profiles data [112] providing additional information on chromatin conformations that may influence DNA accessibility for new integrations . Finally , the FDA statistical methodology we utilized , in particular the ITP and FLR techniques , can be employed to analyze a variety of genomic data . These techniques are very versatile; the ITP does not require assumptions on the distributions underlying the data , and both the ITP and the FLR allow one to effectively characterize location and scale of a phenomenon , linking detected effects to specific intervals . We thus expect them to substantially contribute to numerous applications in genomics research .
We retrieved ETn and IAP elements fixed ( 1 , 868 and 5 , 064 , respectively ) in mouse strain C57BL/6J , and those polymorphic ( 248 and 2 , 224 , respectively ) for C57BL/6J with respect to strains A/J , DBA/2J and 129X1/SvJ ( reference genome mm9 ) [25] ( Table 1 ) . We assume ERVs shared among four mice strains to be fixed , even though these elements are transpositionally active and some of them might be absent in mouse strains not analyzed here . In addition , we chose 1 , 036 full-length and solo-LTR HERV-K elements from the human genome ( hg19 ) [63] ( Table 1 ) . This data set includes all known polymorphic elements ( a total of five at the time of writing this manuscript ) [14] but , because they are also expected to be at least a few hundred years old [113] , we analyzed them together and refer to them all as fixed HERV-Ks . All elements were manually re-annotated , to avoid errors from automatic annotations in the genomes . The data sets were filtered to include only the elements in the 60 bp to 11 kb length range ( Table A in S1 Text ) . We also obtained the IntS of in vitro HERV-K integrations in human embryonic kidney ( 293T , female ) and fibrosarcoma ( HT1080 , male ) cell lines from [34] ( Table 1 ) . Out of the 1 , 565 integrations reported in this study , we located 1 , 208 by alignment of the IntS sequence to hg19 using BLAT [114] . Only matches with 100% identity over the total length of the IntS sequence were considered . In addition , IntS sequences that matched multiple locations ( a total of 31 cases with integration site sequences <100 bp ) were discarded because we could not identify their integration sites definitively . For each element , we considered flanking sequences spanning 32-kb upstream and 32-kb downstream of the element , so that we could maximize the number of ERVs analyzed without overlapping of the flanking regions . This 64-kb region was then divided into 64 1-kb windows on which to quantify genomic features ( Fig 1A ) . We eliminated elements for which the 32-kb flanking sequences overlapped by over 320-bp ( 1% ) with genome gaps ( i . e . sequences with Ns ) or by over 1-bp with other elements’ flanking regions . Similarly , we generated 64-kb control regions , selecting areas that overlapped by at most 2% ( 1% ) with mouse ( human ) LTR elements ( as annotated in the corresponding genome at the UCSC Table browser [115] ) , or with the 32-kb flanking sequences of our ERVs ( Table 1 ) . After this filtering , we were left with 1 , 866 ETns ( 242 polymorphic and 1 , 624 fixed ) , 5 , 950 IAPs ( 1 , 986 polymorphic and 3 , 964 fixed ) , 826 fixed HERV-Ks and 1 , 065 in vitro HERV-Ks , which represent approximately four fifths of the elements analyzed in the original publications , respectively ( Table 1 ) . For ERVs that were annotated in the reverse strand ( i . e . negative orientation with respect to the chromosome orientation ) , we inverted the 64-kb regions to consistently evaluate the effects of the genomic features with respect to the ERVs . Most of the functional datasets used in this study lack data for the Y chromosome , therefore we excluded the ERVs on this chromosome from our analyses . We selected a set of genomic features associated with ERVs by prior studies , or found to be significant in other TE experiments . The data were obtained from the UCSC Table Browser [115] , ENCODE [116] , and previous publications ( Table 2 ) ; when necessary the lift-over tool [117] was used to convert genome locations to mm9 and hg19 . In total , we quantified 38 mouse and 39 human high-resolution genomic features , derived from to 39 mouse and 40 human datasets ( Table 2 ) in 1-kb windows along the flanking sequences of our elements and our control regions ( Fig 1A ) . Genome-wide microsatellite features were extracted from mouse and human genomes requiring a minimum of 9 , 5 , 4 and 4 motif repeats to define mono- , di- , tri- , and tetra-nucleotide microsatellites , respectively [118] . Mono- , di- , tri- , and tetranucleotide microsatellites were analyzed separately due to their unique genome distribution and mutation rates [118 , 119] . High-resolution recombination rates ( in mouse ) , as well as methylation levels and expression levels , were quantified as weighted averages ( average of rates or levels weighted by the total number of base pairs in a window ) in each 1-kb window . All other features were quantified as coverages ( fraction of the genomic window covered by the feature ) in each 1-kb window , except L1 target sites , AT and CG nucleotides for which we used counts . In order to reduce multicollinearity in downstream analyses , we applied a hierarchical clustering with Spearman’s rank correlation ( distance = 1-|correlation| ) and complete linkage [87] , and selected one feature from each cluster above a 80% threshold in mouse and in human ( Figs B and C in S1 Text ) . This reduced the set of features under investigation to 35 and 36 ( 35 and 37 datasets ) in mouse and human , respectively ( Figs D and E in S1 Text ) . In addition to these high-resolution features , we selected two features that were available at low resolution in human and mouse ( recombination rates and replication timing ) and we considered their average value as a single value for each 64-kb region . Replication timing was measured as the log2 of early/late S-phase populations of cells in culture [120]; therefore regions that replicate earlier present high replication timing “values” . Moreover , distances to telomere and centromere ( measured considering the element or the center of the control region ) were included in the study . For each high-resolution genomic feature , we considered curves described by the whole 64-kb signal as the object of our study , embedding the problem in the framework of Functional Data Analysis ( FDA ) [48 , 49] . Importantly , FDA allowed us to naturally incorporate in the analysis the consecutive ordering of the measurements of each feature along the genome . The analysis was divided in two main phases ( see Fig 1A ) . First , we considered each individual genomic feature separately , in order to assess whether it had significant discriminatory power in a given comparison ( e . g . when contrasting ERV flanking sequences vs . controls; Fig 1B ) . We performed this phase using an extended version of the Interval Testing Procedure for functional data ( ITP ) recently introduced by [62] . ITP is able not only to assess whether differences exist between the distributions of the curves in , e . g . , ERV flanking sequences vs . control regions , but also to anchor statistically significant differences to specific locations ( 1-kb windows ) or sub-intervals in the regions , and to specific scales ( e . g . , a specific 5-kb sub-interval near the integration site , comprising a difference that is registered as significant up to a scale of 30-kb ) . Also , in any given comparison ( Fig 1B ) , low-resolution features ( recombination rates , replication timing , distance to telomere and distance to centromere ) were tested using the simple univariate non parametric test employed by the first step of the ITP on each basis component ( see below for more on the ITP ) . In the second phase we evaluated individual features fitting single Functional Logistic Regressions ( FLR ) , and dealt with multiple predictors simultaneously by means of multiple FLR models ( see Fig 1C ) [48 , 49] . Combining the results obtained with these FDA methods we were able to perform an extensive , genome-wide evaluation of the effects of genomic landscape on integration and fixation of the various ERV families considered in this study ( Fig 1C ) . In order to capture the influence of different genomic features on integration and/or fixation of ERVs , we applied the ITP and FLR to nine pair-wise comparisons ( Fig 1B ) : ( 1 ) ETn polymorphic vs . control , ( 2 ) ETn fixed vs . controls , ( 3 ) ETn fixed vs . polymorphic , ( 4 ) IAP polymorphic vs . controls , ( 5 ) IAP fixed vs . controls , ( 6 ) IAP fixed vs . polymorphic , ( 7 ) in vitro HERV-Ks vs . controls , ( 8 ) fixed HERV-Ks vs . controls and ( 9 ) fixed vs . in vitro HERV-Ks . Only non-overlapping regions were employed in these comparisons , leading to the sample sizes indicated in Table 1 . Let XF , R ( w ) w = 1 , … , 64 be the signal ( measured as count , coverage or weighted average ) related to the genomic feature F in the 64-kb region R–this will be one of two types of regions; e . g . , the flanking sequences of an ERV element or a control region . We considered this signal as 64 equidistant pointwise measurements of the curve xF , R ( t ) within the interval [-31 . 5; 31 . 5] surrounding the IntS of an ERV , or the center of a control region ( see Figs D and E in S1 Text ) . In order to determine which are the features that best distinguish between the two types of regions , we employed a functional hypothesis test on the differences between the curves distributions in the two groups . Many FDA methods exist to deal with this inferential problem globally , considering the distribution of the whole curve at the same time [121] , or locally , considering each component of the curve separately . Since we were interested both in global and local significance , we used the ITP [62] , which allowed us to simultaneously test for global differences between the two distributions and to impute observed differences to specific subregions . Because of its non-parametric ( permutational ) nature , this test does not require assumptions on the distributions underlying the data; this too made it highly attractive for our application , since many of the genomic features under consideration have signals that appear to follow complex , non-regular distributions . Importantly , we also introduced some extensions to the original ITP of [62] . In its extended version , the procedure allows us to ( i ) use multiple test statistics–capturing distinct aspects of the curves distributions–in assessing differences , and ( ii ) evaluate multiple scales by changing the maximum length of the intervals on which the global test is performed . Below we give a brief description of the ITP approach ( full detail can be found in [62] ) . Let x1 , i ( t ) i = 1 , … , n1 and x2 , i ( t ) i = 1 , … , n2 be two random samples from two independent stochastic curves . In our application , these are the curves xF , R ( t ) related to a feature F , in the two groups that we want to compare ( e . g . , flanking sequences of ERV vs . control regions ) . We consider the problem of testing the null hypothesis that the distributions of the two stochastic curves are equal , versus the alternative hypothesis that the two distributions differ . The first step of the ITP consists in decomposing the observed curves on a suitable reduced basis {ϕk}k=1K [48 , 49] , i . e . represent each curve xg , i ( t ) as the set {cg , i ( k ) }k=1K of coefficients obtained in the expansion xg , i ( t ) =∑k=1Kcg , i ( k ) ϕk ( t ) . Then , in the second step , a univariate permutation test [122] on each basis component c ( k ) is performed to assess whether its distribution is significantly different for the two stochastic curves ( or underlying “population” of curves ) . Under the null hypothesis that the distributions are the same , all permutations of the n1+n2 “observed” coefficient values c1 , 1 ( k ) , ⋯ , c1 , n1 ( k ) , c2 , 1 ( k ) , ⋯ , c2 , n2 ( k ) have the same chance to occur . Hence , we can compute the empirical null distribution of a test statistics considering the values it assumes on all the different permutations; the p-value is generated dividing the number of permutations with a test statistics more extreme than the one observed on the data by the total number of permutations . The third step is to perform analogous multivariate permutation tests on each possible set of contiguous components ( i . e . on the interval components c ( 1 ) -c ( 2 ) , c ( 2 ) -c ( 3 ) , … , c ( 1 ) -c ( 2 ) -c ( 3 ) , … ) . Here we employ the Nonparametric Combination Procedure developed by Pesarin and Salmaso [121] , that allow us to implement a multivariate test by combining univariate test statistics obtained with synchronized permutations . In detail , the same permutations of the n1+n2 values c1 , 1 ( k ) , ⋯ , c1 , n1 ( k ) , c2 , 1 ( k ) , ⋯ , c2 , n2 ( k ) are applied to all components k = 1 , … , K . Finally , in order to control the familywise error rate over sets of contiguous components , the fourth step computes an adjusted p-value for each component as the maximum among the p-values of all tests whose null hypothesis includes the component under consideration . Such a strategy exploits the ordered nature of the basis components creating a multiple testing correction that is coherent with the structure of data , and allows us to impute observed differences to specific subregions . We chose to represent the curves using B-splines of order 1 ( piecewise constant basis ) with 65 or 33 nodes in the interval [-31 . 5; 31 . 5] and thus retain all the variability of the genomic features observed at 1-kb resolution ( Figs D and E in S1 Text ) . In this way , we were able to test directly the raw data ( 65 nodes choice , one value every 1-kb ) as well as a piecewise constant smoothed version of the raw data ( 33 nodes , one value every two 1-kb windows ) without introducing substantial biases in the ITP . Notably , even if we are considering order 1 B-splines to represent the curves , the functional test statistics of the ITP ( e . g . the mean curves ) as computed on the data are rather smooth and their distributions do not show marked discontinuities ( Figs F-N in S1 Text ) . The test p-values were computed considering 10 , 000 random permutations . In its extended implementation , the ITP can be performed using different test statistics , each highlighting a particular aspect of the curves distributions . In particular , we considered three test statistics: 1 ) the sample mean difference , 2 ) the sample median difference , and 3 ) the sample variance ratio . In addition to introducing different test statistics , we modified the ITP so that differences between the distributions of the two stochastic curves are evaluated at all possible resolutions; that is , we corrected the p-values controlling the familywise error rate on intervals of all possible maximum length . In this way , we were able not only to evaluate the importance of a genomic feature in characterizing ERVs ( globally or at specific locations ) when considering the whole 64-kb flanking region , but also to establish if the feature of interest is able to differentiate between events and controls only at a smaller scale . All genomic features found significant ( p-value<0 . 05 ) , considering at least one test statistics in the ITP , were included in the Functional Logistic Regression ( FLR ) analysis phase ( see Fig 1C and below ) . In particular , features that showed significant differences between the two groups independently of scales and locations ( invariant differential landscape features–IDL ) were represented through their mean values across the 64 1-kb windows and treated as scalar predictors in the FLR . On the other hand , features that were significant only at a particular scale and/or for specific locations , considering at least one test statistics ( localized differential landscape features–LDL ) were treated as functional predictors . Genomic features that were significant on the whole 64-kb region but showed stronger differences ( e . g . , in means ) in a particular location were also treated as functional predictors . To test for significant differences ( e . g . , between ERVs flanking sequences and controls , or other comparisons ) in recombination rate , replication timing , distance from telomere and distance from centromere ( low-resolution features ) , we employed an univariate version of the ITP described above , where a single value is considered for each region . Similarly to the ITP , we tested the null hypothesis that the feature under study has the same distribution in elements and controls ( or other comparisons ) , versus the alternative hypothesis that the two distributions differ , by means of a univariate permutation test . We considered again three test statistics: 1 ) the sample mean difference , 2 ) the sample median difference , and 3 ) the sample variance ratio . For each , we computed the test p-value estimating the distribution of the test statistics under the null hypothesis with 10 , 000 random permutations . The features that resulted significant ( p-value<0 . 05 ) with respect to at least one test statistics were included in the FLR as scalar predictors ( see Fig 1C and below ) . The second phase of our analysis consisted in fitting single and multiple FLR models [48 , 49] using as potential predictors the genomic features selected with the ITP ( Fig 1C ) . The goal of single FLR fits was to identify predictors with very strong predictive power in each comparison . After setting aside these major predictors , whose strength may obscure the role of other features , the goal of multiple FLR was to relate the discrimination ( e . g . , between the flanking sequences of ERVs and control regions ) with all remaining features simultaneously . The details are as follows . For each comparison , we generated a binary response Y encoding events and controls as “1” and “0” , respectively , and considered the functional model logit ( E[Y|ZF1 , ⋯ , ZFr , xFr+1 ( t ) , ⋯ , xFr+s ( t ) ] ) =ln ( p1−p ) =β0+∑i=1rβFiZFi+∑i=r+1r+s1|IFi|∫IFiβFi ( t ) xFi ( t ) dt where p represents the probability of an event conditionally to the observed features . Here ZFi represent the scalar predictors , i . e . the features that emerged as important for all scales and locations ( IDL features ) and the low-resolution features selected through the univariate test , while xFi ( t ) represent the functional predictors ( LDL features ) with support intervals IFi , respectively . In detail , the signal XF , R ( w ) w = 1 , … , 64 in region R for a genomic feature F that was significant independently of scales and locations in the ITP , was summarized by its sample mean over the 64 windows ZF , R=164∑w=164XF , R ( w ) and then included in the FLR model as scalar predictor . Conversely , the curve xF , R ( t ) with pointwise evaluations XF , R ( w ) w = 1 , … , 64 corresponding to a genomic feature F that was significant exclusively in the interval IF was included in the model as functional predictor in the selected interval . For each functional predictor , we expanded predictor curve and related coefficient function on the same reduced basis {ϕF , k}k=1KF , i . e . This basis ( with its specific support ) was chosen separately for each functional predictor , in order to minimize the dimension KF while capturing the scale that emerged as significant in the ITP . To arrive to a meaningful FLR model ( Fig 1C ) we first considered each scalar predictor ZF alone and fitted the single logistic regression model: logit ( E[Y|ZF] ) =ln ( pF1−pF ) =β0+βFZF where pF represents the probability of an event conditionally to the observed feature F . Whenever the distribution of the scalar predictor was skewed , we regularized it with a shifted logarithmic transformation and fit the model with the transformed data . More specifically , we took logs after adding a positive shift parameter s chosen to simultaneously maximize the p-values of the Shapiro-Wilk normality test on the transformed scalar signal in both events and controls . We computed the proportion of deviance explained ( DE ) by the model , or pseudo R-squared , as DE=Rpseudo2=Dnull−DmodelDnull where Dnull is the null deviance and Dmodel the model residual deviance [123] . This measure revealed the discriminatory strength of each individual scalar predictor ( e . g . , to distinguish between flanking sequences of ERVs and control regions ) . Similarly , we fitted a single functional logistic regression model for each individual functional predictor xF ( t ) : logit ( E[Y|xF ( t ) ] ) =ln ( pF1−pF ) =β0+1|IF|∫IFβF ( t ) xF ( t ) dt where pF represents again the probability of an event conditionally to the observed feature F . Similar to scalar predictors , functional predictors were log-transformed as needed , and their strength was measured by means of the DE . Scalar or functional predictors that were very strong ( DE > 20% ) were noted and interpreted ( see Results and Discussion ) , but not included in the multiple FLR–as their inclusion would obscure subtler contributions and co-significance of other , weaker predictors . Next , we chose the best subset among the remaining scalar predictors using a multiple logistic regression model with LASSO regularization after predictors standardization [124] . The optimal regularization coefficient was chosen as the one maximizing the mean 10-fold cross validation misclassification rate . After identifying the best subset of scalar predictors , say {ZFi}i=1r , we fitted the standard additive multiple logistic model ( without LASSO penalization ) logit ( E[Y|ZF1 , ⋯ , ZFr] ) =ln ( pscalar1−pscalar ) =β0+∑i=1rβFiZFi . We then augmented this scalar model attempting to add the remaining functional predictors one at a time in a stepwise forward fashion ( i . e . adding each time the functional predictor inducing the biggest gain in terms of DE ) , as long as the DE increased more than 1% and the AIC decreased . Finally , we excluded scalar predictors that may have been rendered non-significant by the introduction of the functional predictors in a stepwise backward fashion–leading to our final multiple FLR model for the comparison under consideration . The importance of each predictor ( scalar or functional ) in the final multiple FLR model was measured by its relative contribution to the deviance explained ( RCDE ) , defined as RCDE= ( Dnull−Dmodel ) − ( Dnull−Dredmodel ) ( Dnull−Dmodel ) where Dnull is the null deviance , Dmodel the multiple FLR model residual deviance , and Dred model the residual deviance of the multiple FLR model obtained by removing the predictor of interest [87] . All data manipulations were performed with in-house scripts and Galaxy tools ( e . g . lift-over , make windows , assign weighted-average , and feature coverage ) . Statistical analyses were performed in the R environment using the packages fda . usc [125] , car [126] , glmnet [127] , pheatmap [128] and a modified version of the functions in fdatest [129] . | Approximately half of the human genome is composed of repetitive elements . Yet we do not completely understand why certain elements insert in particular genomic locations , and what determines which elements are retained and which are eliminated from the genome . To address these questions we studied endogenous retroviruses , one type of repetitive elements which occupy ~10% of the human and mouse genomes , together with genomic features characterizing various biological processes ( e . g . , recombination and transcription ) in the neighborhoods of these elements . Using statistical techniques , we identified enrichment of genomic features in the vicinity of endogenous retroviruses of different evolutionary ages . Features overrepresented adjacent to young endogenous retroviruses are expected to have facilitated their insertion in the genome . Features overrepresented adjacent to older endogenous retroviruses are expected to have facilitated both their insertion and their chances of being sustained in the genome . Our analyses allowed us to explain the uneven distribution of endogenous retroviruses along the genome , and thus to better understand the interaction of different biological processes in shaping the evolution of genome architecture . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"functional",
"genomics",
"chromosome",
"structure",
"and",
"function",
"centromeres",
"dna-binding",
"proteins",
"human",
"genomics",
"mathematics",
"mammalian",
"genomics",
"epigenetics",
"dna",
"dna",
"methylation",
"chromatin",
"discrete",
"mathematics",
"combinatorics"... | 2016 | Integration and Fixation Preferences of Human and Mouse Endogenous Retroviruses Uncovered with Functional Data Analysis |
Stem cells reside in specialised microenvironments , or niches , which often contain support cells that control stem cell maintenance and proliferation . Hedgehog ( Hh ) proteins mediate homeostasis in several adult niches , but a detailed understanding of Hh signalling in stem cell regulation is lacking . Studying the Drosophila female germline stem cell ( GSC ) niche , we show that Hh acts as a critical juxtacrine signal to maintain the normal GSC population of the ovary . Hh production in cap cells , a type of niche support cells , is regulated by the Engrailed transcription factor . Hh is then secreted to a second , adjacent population of niche cells , the escort cells , where it activates transcription of the GSC essential factors Decapentaplegic ( Dpp ) and Glass bottom boat ( Gbb ) . In wild-type niches , Hh protein decorates short filopodia that originate in the support cap cells and that are functionally relevant , as they are required to transduce the Hh pathway in the escort cells and to maintain a normal population of GSCs . These filopodia , reminiscent of wing disc cytonemes , grow several fold in length if Hh signalling is impaired within the niche . Because these long cytonemes project directionally towards the signalling-deficient region , cap cells sense and react to the strength of Hh pathway transduction in the niche . Thus , the GSC niche responds to insufficient Hh signalling by increasing the range of Hh spreading . Although the signal ( s ) perceived by the cap cells and the receptor ( s ) involved are still unknown , our results emphasise the integration of signals necessary to maintain a functional niche and the plasticity of cellular niches to respond to challenging physiological conditions .
Stem cells are responsible for the integrity of tissues during growth , ageing , and repair . They reside in specialised microenvironments , or niches , which frequently comprise support cells that control stem cell self-renewal , proliferation , and differentiation [1] , [2] . Stem cell niche regulation often involves short-range signalling between stem cells themselves and the surrounding microenvironment . One such short-range signal is the Hedgehog ( Hh ) family of proteins , which mediates homeostasis in several adult tissues , including the gastrointestinal tract , the hematopoietic system , and the vertebrate central nervous system [3]–[7] . In fact , Hh signalling dysfunction can lead to stem cell depletion or proliferative disorders such as tumourigenesis [8] , [9] . However , the detailed mechanisms by which Hh acts in stem cell maintenance remain elusive . In Drosophila females , germline stem cells ( GSCs ) are located at the apex of the ovary , in a structure termed the germarium that constitutes a well-defined stem cell niche . The germarium hosts three types of somatic niche cells: terminal filament cells ( TFCs ) , cap cells ( CpCs ) , and escort cells ( ECs ) , which support two to three GSCs and which can be labelled with specific markers such as the bab1-Gal4 and patched-Gal4 drivers ( Figure 1 ) [10] . The spatial organisation of the GSC niche permits direct contact between two to three CpCs and one GSC , which is anchored to the CpCs by adherens junctions [11] . In addition , approximately two ECs almost completely surround a given GSC [12] . The coordinated action of GSCs and their support cells allows continuous egg production during adulthood . Thus , GSCs normally divide asymmetrically to produce a differentiating cystoblast and a lineage-renewing GSC daughter [13] . Cystoblasts divide four times to give rise to 2- , 4- , 8- , and 16-cell cysts . ECs transfer the differentiating germline cystoblasts and cysts down the germarium using dynamic cytoplasmic processes [14] , [15] . Germline cells in the germarium contain specialised organelles rich in membrane skeletal proteins that adopt a spherical ( called spectrosome ) appearance in GSCs and cystoblasts . Upon germline differentiation , the spectrosome grows in size and becomes a branched structure , termed fusome , characteristic of differentiating cysts . Hence , GSCs can be unambiguously identified by their location within the niche ( in direct contact with CpCs ) and by the presence of spectrosomes ( Figure 1 ) . Reciprocal crosstalk between GSCs and support cells shapes the niche . Firstly , the size and organisation of the CpC cluster depends on proper Notch signalling between GSCs and CpCs [16] . Secondly , both the CpCs and the adjacent ECs play an important role in GSC maintenance , as they transduce the Janus kinase/Signal transducer and activator of transcription ( Jak/Stat ) pathway to induce the production of the Bone Morphogenetic Protein ( BMP ) protein Decapentaplegic ( Dpp ) [12] , [17] , [18] . Thirdly , the germline lineage activates the epidermal growth factor receptor pathway in the ECs to repress dally expression , thus limiting Dpp movement and stability [19] . Because Dpp ( and another BMP homologue called Glass bottom boat [Gbb] ) [20] , [21] act directly on GSCs to repress differentiation and promote self-renewal [22] , [23] , the control of BMP activity is of the utmost importance for correct GSC niche homeostasis . Here , we demonstrate a key role for the Hh pathway in the regulation of BMP signalling in the Drosophila female GSC niche . In addition , we found that wild-type niche support cells grow short Hh-coated filopodia that are functionally relevant for GSC maintenance . Furthermore , support cells sense dysfunctional Hh signalling within the niche and react by growing up to 6-fold longer cytonemes that help increase the range of Hh ligand spreading .
In a number of tissues , the Engrailed ( En ) transcription factor regulates hh expression . Because both en and hh are expressed in TFCs and CpCs ( Figure 1B and 1C ) , and considering the importance of the Hh signalling cascade in stem cell maintenance in insects and vertebrates [24] , [25] , we tested whether the en/hh connection played a role in the GSC niche . To generate en-deficient germaria , we cultured adult females bearing a thermosensitive en allele ( enspt ) in combination with an en deficiency ( enE ) for 7 or 14 d at restrictive temperature ( 28°C; hereafter referred to as ents germaria ) . Compared to control germaria ( enspt/CyO ) , which contained an average of 2 . 3±0 . 8 GSCs and 10 . 2±1 . 3 developing cysts ( n = 62; 7 d ) and 2 . 1±0 . 9 GSCs and 9 . 1±3 . 1 developing cysts ( n = 49; 14 d ) , ents germaria showed a significant decrease in the average number of GSCs and cysts ( 1 . 4±1 . 1 GSCs and 4 . 3±2 . 9 developing cysts , n = 52 , 7 d; 1 . 2±0 . 8 GSCs and 3 . 9±2 . 5 developing cysts , n = 41 , 14 d ) . Interestingly , 28 . 6% of ents germaria analysed after 7 d at restrictive temperature were devoid of germline cells , which emphasised the importance of en gene function in GSC maintenance ( Figure 2A–2D; Table S1 ) . To distinguish between a requirement for en in the germline versus in the niche support cells , we abolished en function from either GSCs or niche cells by utilising two genetically null alleles , enE and en54 . The removal of en from the germline did not affect oogenesis , even 3 wk after gene inactivation ( n>30 for each genotype; Figure 2E ) . To eliminate the activity of en in TFs , CpCs , or ECs we utilised the bab1-Gal4 driver ( Figure 1D ) . Similar to the removal of en from the germline , elimination of en from all ECs in contact with a given GSC did not yield a visible phenotype ( 100% of cases , n = 23; Figure 2F ) . However , in 67 . 7% ( n = 37 ) of mosaic germaria where en function was removed from at least three clustered CpCs , we observed differentiating cysts that contained branched fusomes and showed the accumulation of the differentiation marker Orb in contact with CpCs ( Figures 2G and S1 ) , a phenotype never found in wild-type germaria . Because we did not detect increased apoptosis in mosaic germaria contaning en mutant CpCs and since these mutant cells still expressed CpC markers ( Figures S1 and S2 ) , we conclude that en is required in CpCs to prevent GSC differentiation . The effect on the germline of removal of En from CpCs suggested the existence of one or more En-dependent niche cell signals that act on GSCs to promote their maintenance . Hh expression in TFCs and CpCs has been shown to be required for germline development [26] ( Figures 1 and S3 ) , which made Hh an excellent candidate to mediate En function in GSC maintenance . We examined the distribution of Hh in mosaic germaria that contained en mutant cells and found that en was required in a cell-autonomous fashion for strong membrane accumulation of Hh in TFCs and CpCs ( 81 . 8% of mutant cells , n = 98; Figure 3A and 3B ) . In addition , we established that the removal of Hh from at least three adjacent CpCs induced GSC differentiation ( 51 . 3% of cases , n = 39; Figure 3C ) . It has been shown that the release of the cholesterol-modified form of Hh requires the activity of the dispatched ( disp ) gene [27] . Interestingly , we found that the removal of disp from CpCs was also associated with the appearance of differentiating cysts within the mosaic niche , albeit at a lower frequency ( 31 . 6% of germaria with clusters of ≥3 mutant CpCs , n = 19; Figure 3D ) . The incomplete penetrance of GSC differentiation in en and particularly in hh or disp mosaic niches was most likely due to non-autonomous Hh release from the remaining wild-type cells present in the niche . In fact , the larger the number of hh mutant CpCs , the fewer GSCs remained in the niche ( see below and Table S2 ) . Alternatively ( or in addition ) , disp mutant CpCs may still be able to sustain a certain level of Hh signalling to adjacent ECs , as shown for the wing disc [27] , [28] . Because the absence of either hh or disp from other niche cells , such as TFCs or ECs , did not cause a visible GSC phenotype ( data not shown ) , and considering the requirement for Disp in cholesterol-modified-Hh release , these results strongly suggest that Hh needs to be produced in , and secreted from , CpCs to support a stable GSC population . Hh signalling is transduced intracellularly by Hh ligand binding the Patched ( Ptc ) receptor in receiving cells , allowing the phosphorylation and activation of Smoothened ( Smo ) , a G-protein-coupled receptor normally inhibited by Ptc [29] . In the germarium , Hh ligand produced in the CpCs might act on GSCs directly , indirectly via ECs , or a combination of the two . To distinguish between these possibilities , we studied the expression pattern of ptc , itself a target gene of Hh signalling , as a readout of pathway activation . Analysis of a reporter of ptc expression ( ptc-lacZ ) showed expression in ECs but not in CpCs or TFCs ( Figure 4A ) . To corroborate that activation of the ptc reporter responded to the canonical Hh pathway , we removed smo from ECs to abrogate the Hh response and found that ptc-lacZ expression was largely eliminated ( 100% of cases , n = 43; Figure 4B ) . These results indicate that the Hh pathway is active only in ECs and not in CpCs or TFCs . In fact , the generation of smo− CpC clones showed no effect on GSC loss by differentiation ( 100% of cases , n = 25; Figure 4C ) , whereas the removal of smo function from larval/pupal or adult ECs induced GSC differentiation , as visualised by the appearance of branched fusomes within the mosaic niches ( 69 . 56% of cases , n = 23; Figure 4D ) . Finally , the generation of mutant smo germline clones using two different null alleles did not result in any visible phenotypes 7 , 14 , or 21 d after clone induction ( 100% of cases , n>39 for each genotype and time point; Figure 4E ) . From these observations , we conclude that Hh produced and secreted by CpCs activates Smo in ECs to elicit a response that is responsible for GSC maintenance . In an attempt to identify the nature of this response , we measured the mRNA levels of the essential stem cell factors dpp and gbb in Hh-depleted germaria . Real-time quantitative PCR analysis of ents germaria showed that the levels of dpp , gbb , and hh mRNAs were reduced by more than 60% when compared to control samples ( Figure S4 ) . Because en is not expressed in ECs , and since dpp and gbb are transcribed in CpCs and ECs [17] , [18] , [22] , our data indicate that en could regulate dpp and gbb transcription in ECs via Hh signalling . However , dpp has also been shown to be a target of En [30] . To test the possibility that en is regulating dpp and gbb transcription via hh , we analysed the amounts of dpp and gbb mRNAs in ovaries in which the Hh pathway was blocked specifically in ECs for 7 d ( ptc-Gal4; UAS-smo RNAi/tub-Gal80ts ovaries ) . In this experimental condition , the levels of dpp and gbb mRNAs are diminished by half ( Figure 4F ) . Furthermore , these germaria also show a significant decrease in the number of GSCs per niche ( Figure S5; control , 2 . 7±0 . 5 GSCs/germarium , n = 40; experimental , 1 . 4±0 . 6 GSCs/germarium , n = 44 ) . Finally , in order to demonstrate that the expression of dpp in ECs is essential for GSC maintenance , we analysed niches in which dpp levels were diminished specifically in ECs for 14 d ( ptc-Gal4; UAS-dpp RNAi/tub-Gal80ts ovaries ) . We found a strong reduction in the number of GSCs due to their precocious differentiation ( Figures 4G and S5; control , 2 . 5±0 . 8 GSCs/germarium , n = 28; experimental , 1 . 4±0 . 6 GSCs/germarium , n = 36 ) . Considering that these BMP molecules are essential for GSC survival [22] , [23] and that dpp is a target gene of the Hh pathway [31] , our results support a model in which female GSC self-renewal requires the en-dependent production of Hh in CpCs . Upon secretion by CpCs , Hh juxtacrine signal is transmitted to the adjacent ECs , which in turn control Dpp and Gbb production to sustain GSC maintenance . The fact that the removal of hh from CpCs or smo from ECs induces a decrease in phospho-Mad levels in the germline , a direct reporter of Dpp signalling , supports this hypothesis ( Figure S6 ) . Thus , in addition to the proposed role for CpCs in ovarian niche signalling [32] , ECs emerge as important regulators of niche signalling , as they not only are responsible for controlling the Jak/Stat and the EGFR pathways [12] , [19] but also exert a key role in the regulation of Hh signalling . Morphogens exert their effects over long distances , which , in the case of Hh , can be as long as 300 µm in the vertebrate limb bud [33] . In contrast , in the Drosophila ovarian niche , the Hh-receiving cells adjoin the Hh-producing cells , as ECs directly contact the CpC rosette , which limits the spread of this ligand . To investigate the mechanism by which Hh is transported within the ovarian niche , we analysed in detail the distribution of Hh in the CpCs . The Hh protein is strongly localised to the cell membrane , and in 30 . 1% of germaria analysed ( n = 149; Figure 5A ) , it decorated short cellular projections 0 . 53 to 1 . 11 µm in length ( 0 . 93 µm on average ) and 0 . 1 to 0 . 3 µm in diameter that formed at the CpC–EC boundaries . These narrow , filiform structures were reminiscent of the thin filopodial membranes , called cytonemes , that were initially described in the wing disc . Cytonemes are actin-rich cytoplasmic extensions thought to mediate specific morphogen signalling and to prevent inadequate diffusion of ligands [34]–[36] . In order to test the biological significance of these structures , we analysed two different experimental scenarios . First , we investigated whether these processes would respond to challenging physiological conditions such as deficient Hh niche signalling . To this end , we analysed the distribution of Hh-coated cytonemes in mosaic germaria harbouring en mutant CpCs and found that in 54% of these germaria one or two of the remaining wild-type CpCs displayed thin , Hh-labelled filopodia significantly longer than those of the controls ( average size 3 . 1 µm , n = 50; Figure 5B and 5D ) . We then blocked the ability of adult ECs to respond to Hh signalling by generating smo− ECs , and we looked for long cytonemes in these mosaic niches . We found that the absence of Hh pathway transduction in ECs provoked a response from signalling CpCs in the form of long , Hh-coated cellular extensions detected in 50% of the cases analysed ( average size 3 . 3 µm , n = 28; Figure 5C ) . To discard the possibility that the presence of differentiated cysts within the niche , such as those generated after removing smo from ECs , induces long , Hh-positive cytonemes , we generated CpCs mutant for the Jak/Stat kinase hopscotch , which also causes GSC differentiation [17] , [18] , and measured cytoneme lengths . In this condition , the cellular processes were not significantly different from those of wild-type controls ( average size 1 . 1 µm , n = 20; Figure 5D ) . Altogether , these results clearly show that the GSC niche can react specifically to decreased Hh levels and/or to impaired Hh signalling by increasing the range of ligand spreading . Moreover , the extended cytonemes found in en− or smo− mosaic germaria projected towards the signalling-deficient area of the niche ( Figure 5E ) , demonstrating that niche support cells sense , and respond directionally to , spatial signalling cues . Finally , to determine whether these cytonemes are specialised structures developed to mediate niche signalling , we studied the distribution of the adherens junction components DE-Cadherin and Armadillo in cytonemes . These proteins labelled the periphery of wild-type CpCs , delineating their round , regular shape , but were absent from their short cytonemes . Similarly , in mosaic en− or smo− niches , long cytonemes did not contain DE-Cadherin or Armadillo , which suggests that cytonemes are Hh-coated filopodia grown specifically to deliver a stem cell survival factor rather than a reflection of mere changes in cell shape ( Figure S7 ) . Next , we wished to study cytoneme functionality by affecting their formation . Because cytonemes are rich in actin filaments [34] , we reasoned that disturbing actin polymerisation in adult CpCs could have an effect specifically on cytoneme production and/or kinetics . Thus , we utilised the bab1-Gal4 driver to express modified versions of two known regulators of actin polymerisation in TFCs and CpCs of adult ovaries . We induced the expression of either a constitutively activated form of the Drosophila Formin homologue Diaphanus ( Dia ) , DiaCA [37] , or a myristoylated form of the Arp2/3-complex regulator Wasp , WaspMyr [38] . While interfering with actin polymerisation may affect other cellular processes rather than cytoneme formation , we performed several controls to make sure that the observed results where as specific as possible . First , we measured the mean value of fluorescence intensity per area unit in control ( tub-Gal80ts/+; UAS-diaCA/+ or tub-Gal80ts/+; UAS-waspMyr/+ ) or experimental ovaries ( tub-Gal80ts/+; UAS-diaCA/bab1-Gal4 or tub-Gal80ts/+; UAS-waspMyr/bab1-Gal4 ) kept at 31°C for 5 d upon eclosion to confirm that overexpression of UAS-waspMyr or UAS-diaCA in adult germaria affected significantly neither the overall amounts of Hh protein in the niche cells nor the expression of CpC markers such as bab1 or Lamin C ( Figure 6 and data not shown ) . Second , we manipulated only post-mitotic cells to prevent unwanted effects during mitosis , as we induced ectopic gene expression in adult CpCs . Third , we utilised an experimental setting that did not affect visibly niche morphology or CpC viability . In this scenario , we found that ectopic expression of WaspMyr or DiaCA for 5 d in niche cells halved the number of germaria growing short cytonemes ( from over 30% in controls to 13 . 6% and 15 . 4% , respectively , n>36 for each genotype; Figure 6C ) . Interestingly , this condition also produced a significant decrease in the number of GSCs per niche ( from 2 . 45±0 . 7 in controls to 1 . 8±0 . 55 and 1 . 7±0 . 6 , respectively , n>36 for each genotype; Figure 6D ) . Since we did not observe apoptosis above control levels in germline cells ( data not shown ) and because we could detect differentiating cysts in these experimental niches ( Figure 6B ) , the formation of short Hh-decorated filopodia in CpCs is an essential step to prevent GSC differentiation . We next tested whether diminishing the number of cytonemes per CpC would affect Hh signalling . To this end , we overexpressed DiaCA in adult TFCs and CpCs utilising the bab1-Gal4 driver and monitored the activation of the Hh-signalling reporter ptc-lacZ . We found that , in contrast to controls , experimental females grown for 5 d at 31°C largely failed to activate the ptc-lacZ reporter in the germarium ( Figure 6E and 6F ) . These results , together with our previous finding that Hh is produced in CpCs and received in ECs , strongly suggest that the Hh-coated cytonemes regulate Hh signalling in the germarium by facilitating Hh delivery to the target ECs to ensure that a normal pool of stem cells is maintained .
Niches are dynamic systems often containing stromal cells that provide physical support and survival factors to nurture a population of stem cells . The data presented here demonstrate that the heterotypic association of support cells is crucial for niche function . In the case of the Drosophila ovarian niche , it has been previously described that the Jak/Stat pathway regulates the expression of dpp in CpCs [17] , [18] . Our results show that the maintenance of a stable population of GSCs relies also on the coordinated action of the CpCs and the ECs , which allows the production and release of the GSC survival ligand Hh in the CpCs and its reception in the ECs . As a consequence of the transduction of the Hh pathway , ECs produce the stem cell factors Dpp and Gbb ( see model in Figure 7 ) . The recent finding of a similar partnership between mesenchymal and haematopoietic stem cells that operates in the bone marrow niche [39] indicates that such collective regulatory interactions within support cells may be a common feature of cellular niches . The study of the mechanisms behind Hh signalling in the Drosophila ovary has allowed the identification of Hh-coated cytonemes in a cellular stem cell niche , emphasising the idea that cytonemes mediate spreading of the activating signal from the producing cells . Recently , it has been reported that the Hh protein localises to long , basal cellular extensions in the wing disk [40] . In addition , filopodial extensions in the wing , eye , and tracheal system of Drosophila have been shown to segregate signalling receptors on their surface , thus restricting the activation of signalling pathways in receiving cells [36] . Hence , cytonemes , as conduits for signalling proteins , may be extended by receiving cells—and so are involved in uptake—or may be extended by producing cells—and so are involved in delivery and release . Interfering with actin polymerisation in adult niches leads to a significant reduction in the number of CpCs growing Hh cytonemes , concomitant with precocious stem cell differentiation , demonstrating that these actin-rich structures are required to prevent stem cell loss and thus are functionally relevant . Importantly , because we disturbed actin dynamics in post-mitotic CpCs that still produce wild-type levels of Hh protein and express CpC markers ( but fail to activate the Hh pathway in ECs ) , the observed effects on stem cell maintenance are most likely specific to Hh delivery from CpCs to their target ECs via short cytonemes . This interpretation is further reinforced by the observation that CpCs can sense decreased Hh levels and/or a dysfunction in the transduction of the Hh pathway in the niche and respond to it by growing Hh-rich membrane bridges up to 6-fold longer than in controls . In this regard , it is interesting to note that the two lipid modifications found in mature Hh act as membrane anchors and give secreted Hh a high affinity for membranes and signalling capacities [41] , [42] . In fact , it has been recently described that a lipid-unmodified form of Hh unable to signal does not decorate filopodia-like structures in the wing imaginal disc epithelium , confirming the link between Hh transport along cytonemes and Hh signalling [40] . Thus , cytonemes may ensure specific targeting of the Hh ligand to the receiving germline cells in a context of intense signalling between niche cells and the GSCs . Interestingly , in both en− and smo− mosaic niches , the long processes projected towards the signalling-deficient area of the niche , which showed that competent CpCs sense the strength of Hh signalling activity in the microenvironment . While the nature of the signal perceived by the CpCs or the receptor ( s ) involved in the process are unknown , we postulate that Hh-decorated filopodial extensions represent the cellular synapsis required for signal transmission that is established between the Hh-producing cells ( the CpCs ) and the Hh-receiving cells ( the ECs ) . In this scenario—and because Ptc , the Hh receptor , is a target of the pathway—the membranes of mutant ECs , in which the transduction of the pathway is compromised , contain lower Ptc levels . Thus , longer and perhaps more stable projections ought to be produced to allow proper signalling . In addition , the larger the number of en mutant cells ( and hence the stronger the deficit in Hh ligand concentration or target gene regulation ) , the longer the cellular projections decorated with Hh ( Tables S3 and S4 ) , which indicates that the niche response is graded depending on the degree of signalling shortage . Do the longer cytonemes found in mosaic germaria represent structures created de novo , or do they simply reflect a pre-existing meshwork of thin intercellular bridges that can regulate the amount of Hh protein in transit across them ? Because we utilised an anti-Hh antibody to detect the cytonemes and all of our attempts to identify other markers for these structures have failed , we cannot presently discriminate between these two possibilities . In any case , since we did not detect increased Hh levels in wild-type CpCs that contained cytonemes relative to those that did not , it is clear that long filopodia do not arise solely by augmenting Hh production in the CpCs . Rather , if long cytonemes are not synthesised in response to a Hh signalling shortage and if they already existed in the niche , they ought to restrict Hh spreading independently of significant Hh production . Furthermore , because the strength of Hh signalling in the niche determines the distance of Hh spreading , either cytoneme growth or Hh transport ( or both ) are regulated by the ability of the CpCs to sense the Hh signalling output . Our demonstration that a challenged GSC niche can respond to insufficient signalling by the cytoneme-mediated delivery of the stem cell survival factor Hh over long distances has wider implications . Niche cells have been shown to send cellular processes to their supporting stem cells in several other scenarios: the Drosophila ECs of the ovary and the lymph gland , the ovarian niche of earwigs , and the germline mitotic region in the hermaphrodite Caenorhabditis elegans [5] , [6] , [14] , [43] , [44] . Similarly , wing and eye disc cells project cytonemes to the signalling centre of the disc [34] , [36] . However , definitive proof that the thin filopodia described in the lymph gland , the earwig ovary , or imaginal discs deliver signals from the producing to the effector cells is lacking . Our findings strongly suggest that cytonemes have a role in transmitting niche signals over distance , a feature that may underlie the characteristic response of more complex stem cell niches to challenging physiological conditions . Careful analysis of the architecture of sophisticated niches , such as the bone marrow trabecular zone for mouse haematopoietic stem cells , will be needed to further test this hypothesis and to determine whether it represents a conserved mechanism for stem cell niche signalling .
Flies were grown at 25°C on standard medium for Drosophila . The following genetically null alleles were used: enE , en54 [45] , hh21 , hhAC [46] , dispSH21 [28] , smoD16 [47] , and smo3 [48] . ptcAT96 is a LacZ enhancer trap inserted in the gene [49] . enspt [50] is a temperature-sensitive allele . To express UASt-DsRed and UASt-flp we used the bab1-Gal4 line [51] . The expression of UAS transgenes in ECs was done utilising the ptc-Gal4 driver . In order to generate experimental enspt/enE adult females , flies were shifted from 25°C to 28°C for 7 or 14 d upon eclosion . To obtain adult females overexpressing WaspMyr [38] or DiaCA [37] , w; tub-Gal80ts/CyO; bab1-Gal4/TM2 flies were crossed to w; UAS-waspMyr or w; UAS-diaCA , respectively . To overexpress dpp RNAi ( VDRC ) or smo RNAi ( Bloomington Stock Center ) in ECs , w; ptc-Gal4 , UAS-GFP; tub-Gal80ts/SM6∧TM6B flies were crossed to w; UAS-dpp RNAi , w; UAS-smo RNAi . The offspring were grown at 18°C , and upon eclosion adult F1 flies were shifted to 31°C for 5 , 7 , or 14 d . Ovaries were dissected at room temperature in PBS containing 0 . 1% Tween-20 ( PBT ) , fixed for 20 min with 4% PFA , blocked with PBT+10% BSA for 1 h , and washed in PBT before they were incubated for 15 h with primary antibodies diluted in PBT supplemented with 1% BSA . Primary antibodies were washed three times in PBT containing 1% BSA . Secondary antibodies were diluted in PBT containing 0 . 1% BSA . Primary antibodies were used at the following concentrations: mouse anti-Hts ( 1B1 ) ( Developmental Studies Hybridoma Bank [DSHB] , University of Iowa ) , 1∶50; rabbit anti-Vasa ( a gift from R . Lehmann ) , 1∶1 , 000; mouse anti-En ( 4D9 ) ( DSHB ) , 1∶50; rabbit anti-α-Spectrin ( a gift from R . Dubreuil ) , 1∶400; rabbit anti-GFP ( Molecular Probes ) , 1∶500; mouse anti-GFP ( Molecular Probes ) , 1∶50; mouse anti–Lamin C ( LC28 . 26 ) ( DSHB ) , 1∶50; rabbit anti-Hh ( a gift from S . Eaton [52] ) , 1∶500; rabbit anti-phospho-Mad 1/5/8 ( a gift from E . Laufer ) , 1∶5 , 000; rabbit anti-β-galactosidase ( Cappel ) , 1∶1 , 000; rabbit anti-cleaved Caspase 3 ( BioLabs ) , 1∶50; and mouse anti-Orb ( 6H4+4H8 ) ( DSHB ) , 1∶50 . Secondary antibodies ( Cy2- and Cy3-conjugated , Jackson ImmunoResearch ) were used at 1∶100 . DNA staining was performed using the DNA dye Hoechst ( Sigma ) at 1∶1 , 000 . Images were captured with a Leica SPE confocal microscope and processed using ImageJ , Adobe Photoshop , and Adobe Illustrator . Fluorescence intensity units and cytoneme length were measured using the Leica LAS-AF software . Images were captured with a Leica SPE confocal microscope and processed using ImageJ , Adobe Photoshop , and Adobe Illustrator . To generate mitotic clones we induced the Flipase enzyme using either a heat shock promoter or the bab1-Gal4 driver to activate expression of a UAS-flp construct . en and smo mutant germline clones were induced by giving 3-d-old females two 1-h-long heat shocks at 37°C spaced by 10 h at 25°C . hh , disp , en , and smo mutant somatic clones were induced expressing UAS-flp with the bab1-Gal4 driver . Ovaries were processed 3 d ( for somatic clones ) or 7 , 14 , or 21 d ( for germline clones ) after treatment . To eliminate smo function in adult females , 3-d-old HS-flp1112/+; smoD16 FRT40A/ubi-nls:GFP FRT40A flies were subjected to three 1-h-long heat shocks at 37°C separated by 6-h periods at 25°C . The following chromosomes were used: HS-flp1112 , FRT42D en54 , FRT42D enE , FRT42D ubi-nls:GFP , smoD16 FRT40A , ubi-nls:GFP FRT40A , FRT42D ubi-nls:GFP , hhAC FRT82B , hh21 FRT82B , dispSH21 FRT82B , bab1-Gal4 FRT82B ubi-nls:GFP , UASt-flp , smo3 FRT40A ptcAT96 , and bab1-Gal4 UASt-flp . The relative amounts of hh , dpp , and gbb mRNAs were determined by real-time quantitative PCR using the comparative cycle threshold ( CT ) method [53] , Fam-dye-labelled TaqMan MGB probes ( Applied Biosystems ) , and an ABI-PRISM 7700 Sequence Detection System . RNA polymerase II ( RpII140 ) was used to normalise mRNA levels . hh , dpp , or gbb mRNA relative amount was calculated from the determination of the difference between the CT of the given gene and that of RpII140 . CT values used were the result of three different replicas from three independent experiments . Primers and TaqMan probes for the different cDNAs were obtained from the Assays-by-Design Service ( Applied Biosystems ) with the following sequences ( 5′–3′ ) : RpII140 , forward , ACTGAAATCATGATGTACGACAACGA , reverse , TGAGAGATCTCCTCGGCATTCT , probe , TCCTCGTACAGTTCTTCC; hh , forward , GCAGGCGCCACATCTACT , reverse , GCACGTGGGAACTGATCGA , probe , CCGTCAAGTCAGATTCG; dpp , forward , GCCAACACAGTGCGAAGTTTTA , reverse , TGGTGCGGAAATCGATCGT , probe , CACACAAAGATAGTAAAATC; gbb , forward , CGCTGTCCTCGGTGAACA , reverse , CGGTCACGTTGAGCTCCAA , probe , CCAGCCCACGTAGTCC . cDNA was synthesised from ∼100–200 ovary pairs of the following characteristics: enspt/CyO ( control ) and enspt/enE ( experimental ) females were shifted from 25°C to 28°C for 7 d after eclosion prior to dissection . +; UAS-smo RNAi/SM6∧TM6B ( control ) and ptc-Gal4 , UAS-GFP/+; UAS-smo RNAi/tub-Gal80ts ( experimental ) females were shifted from 25°C to 31°C for 7 d after eclosion prior to dissection . A Student's t test was used to determine whether the following were significantly different between control and experimental samples: ( i ) the mean number of GSCs and differentiated cysts per germarium , ( ii ) the relative levels of hh , dpp , and gbb expression , and ( iii ) the length of cytonemes . To analyse whether the observed differences in the percentages of cytoneme-containing germaria between control ovaries and ovaries overexpressing WaspMyr or DiaCA were significant , we applied the Chi-square test . Differences were considered significant when the p-values were less than 0 . 01 . | The Drosophila ovary contains a well-defined stem cell niche that hosts 2–3 germline stem cells ( GSCs ) . The Hedgehog ( Hh ) family of signalling proteins mediates cellular homeostasis in several adult tissues , and here we decipher the detailed mechanism of action of Hh in the adult female GSC niche . We demonstrate that Hh acts in a juxtacrine manner ( i . e . , it requires physical contact between the cells involved ) to maintain the normal pool of GSCs in the ovarian niche . Hh is produced in one type of niche support cell ( the cap cells ) , and it is received , upon secretion , by a second , neighbouring population of niche cells ( the escort cells ) . In the latter , we show that the Hh signalling pathway regulates the expression of the Drosophila Bone Morphogenetic Protein ( BMP ) homologues and essential stem cell factors decapentaplegic ( dpp ) and glass bottom boat ( gbb ) . We also find that Hh distribution in the GSC niche is mediated by short cellular projections , reminiscent of wing disc cytonemes , although they grow from the ( Hh ) signal-producing cells towards the receiving cells . Under conditions of low levels of Hh protein and/or Hh signalling within the niche , cap cells emit up to 6-fold longer Hh-decorated cytonemes towards the signalling-deficient area of the niche . Our data reveal that stem cell niches are dynamic structures that can sense , and react to , changes in the activity of essential stem cell factors to prevent stem cell differentiation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology",
"stem",
"cells",
"biology"
] | 2012 | Cytoneme-Mediated Delivery of Hedgehog Regulates the Expression of Bone Morphogenetic Proteins to Maintain Germline Stem Cells in Drosophila |
Hepatitis C virus ( HCV ) requires the liver specific micro-RNA ( miRNA ) , miR-122 , to replicate . This was considered unique among RNA viruses until recent discoveries of HCV-related hepaciviruses prompting the question of a more general miR-122 dependence . Among hepaciviruses , the closest known HCV relative is the equine non-primate hepacivirus ( NPHV ) . Here , we used Argonaute cross-linking immunoprecipitation ( AGO-CLIP ) to confirm AGO binding to the single predicted miR-122 site in the NPHV 5’UTR in vivo . To study miR-122 requirements in the absence of NPHV-permissive cell culture systems , we generated infectious NPHV/HCV chimeric viruses with the 5’ end of NPHV replacing orthologous HCV sequences . These chimeras were viable even in cells lacking miR-122 , although miR-122 presence enhanced virus production . No other miRNAs bound this region . By random mutagenesis , we isolated HCV variants partially dependent on miR-122 as well as robustly replicating NPHV/HCV variants completely independent of any miRNAs . These miRNA independent variants even replicate and produce infectious particles in non-hepatic cells after exogenous delivery of apolipoprotein E ( ApoE ) . Our findings suggest that miR-122 independent HCV and NPHV variants have arisen and been sampled during evolution , yet miR-122 dependence has prevailed . We propose that hepaciviruses may use this mechanism to guarantee liver tropism and exploit the tolerogenic liver environment to avoid clearance and promote chronicity .
Chronic HCV infection is one of the most common liver diseases with ~71 million people persistently infected globally; a significant number of those will develop cirrhosis or liver cancer [1] . The binding of liver specific miR-122 to HCV RNA is essential for viral replication [2] . This interaction is unusual in that two molecules of miR-122 bind to the 5’ untranslated region ( 5’UTR ) of HCV using both seed and auxiliary pairing [3 , 4] . It is well established that HCV viral load can be dramatically decreased by inhibiting miR-122 in cell culture , chimpanzees or patients , making antagonists of miR-122 a first-in-class antiviral strategy [5–7] . Several functions have been suggested for the HCV/miR-122 interaction: ( i ) Binding of the AGO/miR-122 complex can protect the uncapped HCV RNA from degradation by cellular exonuclease XRN1 and/or XRN2 [8–10] . ( ii ) AGO/miR-122 binding can increase HCV internal ribosome entry site ( IRES ) -driven translation , thus promoting the HCV replication [11] . This process possibly works by switching the IRES from “closed” to “open” conformation [12 , 13] . ( iii ) Competition between miR-122 and poly ( rC ) -binding protein ( PCBP2 ) that binds and circularizes HCV RNA may act as a switch between translation and replication [14] . ( iv ) In addition , using AGO-CLIP and RNA-seq , we recently showed that HCV RNA can act as a miR-122 “sponge” in a positive feed-back loop to de-repress cellular mRNAs normally targeted by miR-122 , thereby indirectly regulating hundreds of genes [15] . Until recently , the miRNA dependence of HCV was considered unique among viruses , even for viruses possessing similar IRESs . However , we recently demonstrated that pestiviruses , such as the important veterinary pathogens bovine viral diarrhea virus ( BVDV ) and classical swine fever virus ( CSFV ) , similarly are dependent on the cellular miR-17 family , although binding occurs on the viral 3’UTR [16] . The unique GB virus B ( GBV-B ) isolate was the only known HCV-related hepacivirus until 2011 . Using replicon systems , GBV-B appears to be only partially dependent on miR-122 [17] . However , recent discoveries have identified a plethora of HCV-related viruses in horses , rodents , bats , monkeys , and cows [18–25] , most if not all of which contain miR-122 seed sites in the 5’UTR . However , the miR-122 requirement has not been investigated for these viruses , as no cell culture systems have been established [24 , 26] . The equine non-primate hepacivirus ( NPHV ) shares the highest sequence homology to HCV [18 , 27] . The genome structure of NPHV resembles HCV , with a long open reading frame ( ORF ) that can be translated into a 2942 amino acid long polypeptide . The polypeptide is predicted to be cleaved into the ten viral proteins C , E1 , E2 , p7 , NS2 , NS3 , NS4A , NS4B , NS5A , and NS5B . An IRES structure similar to that for HCV was predicted for the NPHV 5’UTR , but with a much longer stem loop 1 ( SL1 ) structure and only one predicted miR-122 binding site . Using NPHV IRES reporter systems , we previously showed a pro-translational role for miR-122 [28] . Here , we aimed to obtain a broader understanding of hepacivirus miRNA dependence and how this might influence tissue tropism . To understand miR-122 requirements for NPHV , we used AGO-CLIP in vivo to pinpoint the miRNA interactome on viral RNA , and developed viable NPHV/HCV 5’UTR chimeras . This approach suggested that NPHV is only partially dependent on miR-122 . Encouraged by this result and to determine whether miRNA independent hepaciviruses are viable , we randomized the miRNA seed site of these chimeras and HCV in an attempt to develop miRNA independent viruses . Our results indicated that hepaciviruses have the potential to escape miRNA requirements and break the liver-specific tissue tropism barrier . It appears that such variants must have been sampled during hepacivirus evolution . Interestingly , miR-122 dependence still has been strongly selected , possibly to leverage the tolerogenic liver environment to establish and maintain chronic infection .
To investigate putative NPHV/miR-122 interactions , we performed AGO-CLIP on liver biopsies from two NPHV infected horses . Although sequence coverage on viral RNA was much lower compared to highly infected HCV cell cultures [15] negating our ability to unambiguously identify interacting miRNAs [16] ( Fig 1A ) , this assay identified regions with AGO/miRNA interactions across the NPHV genome . Interestingly , all significant peaks perfectly overlapped the four conserved miR-122 sites; one in the 5’UTR , one in NS5A and two in NS5B ( Fig 1B ) . These data strongly suggest interaction with miR-122 during NPHV infection in vivo . To functionally characterize NPHV miR-122 requirements in the absence of a cell culture system supporting NPHV replication , we set out to establish NPHV/HCV chimeras based on the HCV genotype 2a recombinant J6/JFH Clone2 [29] . We constructed four different NPHV/HCV chimeras to test regions of the HCV 5’UTR that could be replaced by NPHV ( Fig 2A ) . Following transfection of these chimeric genomes into Huh-7 . 5 cells , we assayed viral replication by staining for NS5A positive cells and measured virus production by limiting dilution ( TCID50; Fig 2B and 2C ) . Replacement of the entire 5’UTR or IRES region ( NPHV-5’UTR and NPHV-IRES ) abolished replication completely , even when followed for 29 days . In contrast , NPHV-SL1 infection spread similarly to the parental HCV construct , although virus production was slightly delayed ( Fig 2B and 2C ) . This indicated that exchange of SL1 had only limited effect on replication efficiency . Although NPHV-SL1/miRBR ( microRNA Binding Region ) was attenuated , it spread to the majority of cells on day 4 with virus production ~10-fold lower than the parental HCV recombinant ( Fig 2B and 2C ) . Since NPHV-SL1/miRBR contains the miR-122 binding site of NPHV , this chimera was of particular interest for studying NPHV miR-122 dependence . We tested the IRES activity of NPHV-5’UTR and NPHV-IRES to distinguish whether the absence of NS5A positive cells was due to block of replication or translation . Compared to HCV , the luciferase signal driven by the NPHV-5’UTR and NPHV-IRES variants was about 3-fold lower , but still much higher than the background ( S1 Fig ) . This suggests that the failure of these viruses to replicate is not due to a block in translation . To determine whether the NPHV-SL1/miRBR recombinant could be further adapted , we took supernatant on day 6 after transfection and infected naïve Huh-7 . 5 cells . Supernatant from newly infected cells was then harvested on day 6 and the 5’ end of the viral genome was sequenced . This analysis revealed a C83A mutation in NPHV-SL1/miRBR upstream of the miR-122 site ( including putative auxiliary pairing ) ( Fig 2D ) . This change did not facilitate binding of a second miR-122 molecule; rather , it changed this region further from mirroring the HCV seed site 1 . To confirm the impact of the C83A mutation , we introduced this nucleotide change into the original NPHV-SL1/miRBR genome and again transfected Huh-7 . 5 cells . The mutant exhibited superior replication and virus production compared to the original NPHV-SL1/miRBR , and was now only slightly attenuated compared to the HCV parent ( J6/JFH1-Clone2 ) , as judged by spread of infection and virus yield ( Fig 2E and 2F ) . Using a CRISPR engineered miR-122 knockout ( KO ) cell line , we next examined the ability of NPHV-SL1/miRBR to replicate in the complete absence of miR-122 . As shown before , HCV replication was dramatically impaired in the miR-122 KO cell line ( Fig 3A and 3B; [15] ) . The number of NS5A positive cells and viral infectivity titers of NPHV-SL1 , NPHV-SL1/miRBR and NPHV-SL1/miRBRC83A were also reduced . Nonetheless , replication and virus production for the NPHV/HCV chimeras was evident in cells lacking miR-122 ( Fig 3A and 3B ) . These results indicated that the loss of miR-122 decreased NPHV replication efficiency , but that replication could still occur in complete absence of miR-122 . Next , we asked whether miRNAs other than miR-122 bind the NPHV miRBR . To this end , we first replaced the miR-122 binding site of NPHV-SL1/miRBR with a miR-15/16-binding sequence . Similar to an HCV variant with both miR-122 sites replaced by miR-15 sites [15] , this virus was viable ( Fig 4A ) . As expected , the HCV and NPHV-SL1/miRBR miR-15 variants were unaffected by the absence of miR-122 ( Fig 4A ) . We then used AGO-CLIP and miRNA-target chimeras [16 , 30] to unambiguously identify interacting miRNAs in wt Huh-7 . 5 cells . For NPHV-SL1/miRBR and NPHV-SL1/miRBR-15 , miR-122 and miR-15/16 , respectively , were the only interacting miRNA species ( Fig 4B ) . This confirmed that no other functional miR-122 site than the single canonical site exists in the NPHV miRBR , and that no other miRNA is binding this region . Given the only partial NPHV miR-122 dependence , we next probed whether miRNA independent hepaciviruses with extrahepatic replication potential could be selected . We exploited the rapid evolutionary capacity of RNA viruses to select for variants with high fitness in the presence ( Huh-7 . 5 ) or absence ( ΔmiR-122 Huh-7 . 5 ) of miR-122 . To increase input diversity , we randomized the miR-122 binding site of NPHV-SL1/miRBR to create NPHV-Rand . In parallel , we randomized the corresponding seed site 2 of HCV to create HCV-Rand ( Fig 5A ) . To enrich for the most efficient variants , we took the supernatant of transfected cells when at least 50% of cells were NS5A positive to inoculate naïve cells ( Fig 5A and 5B ) . After three passages , we sequenced the 5’UTR of the selected viruses in the supernatant . For HCV-Rand in Huh-7 . 5 cells , the majority ( 67% ) contained the wildtype miR-122 binding site , whereas the rest ( 33% ) contained the miR-15 binding site . Interestingly , the latter variant was identical to a synthetic miR-122/15 construct we previously showed to be viable [15] . In ΔmiR-122 Huh-7 . 5 cells , most recovered variants contained a G-rich region; among them 66% had the sequence GGCGNG . Similarly , most recovered NPHV-Rand variants from wild type cells contained the wild type miR-122 binding site . These had also acquired the previously described C83A mutation . Surprisingly , 30% of the recovered strains had a 15-nucleotide deletion in the miRBR region ( NPHV-delta-UUGGCG ) . In ΔmiR-122 Huh-7 . 5 cells , 73% of the recovered NPHV variants had another G-rich GGYAGG motif . One variant from each group was selected for further characterization ( Fig 5C , boxed names ) . We next engineered the selected sequences into the original viral genomes and tested their replicative fitness in Huh-7 . 5 and ΔmiR-122 Huh-7 . 5 cells . For comparison , we also included HCV-U3 in our analysis since this virus , which contains a fraction of the cellular U3 snoRNA sequence in place of the SL1 region , replicates in the absence of miR-122 [31] . Similar to HCV-U3 , the selected variants HCV-S2-GGCGUG , NPHV-delta-UUGGCG and NPHV-GGCAGG all replicated comparably in the presence or absence of miR-122 ( Fig 6A ) . The NPHV-based variants , however , were the most fit . HCV-122/15 replicated and spread comparably to parental HCV in Huh-7 . 5 cells , but was attenuated in the absence of miR-122 . Thus , it appears that the selected NPHV-Rand variants in particular , could replicate with equal efficiently in the absence or presence of miR-122 . The sequences of the selected G-rich random variants did not correspond to known canonical miRNA seed sites . We therefore examined the ability of these variants to replicate in the complete absence of miRNAs . Using CRISPR mutagenesis we ablated DICER to produce ΔDICER Huh-7 . 5 cells . Given the critical role for DICER in cleaving pre-miRNAs [32 , 33] , no mature miRNAs are produced in these cells . As expected , parental HCV and HCV-122/15 were not viable in ΔDICER Huh-7 . 5 cells ( Fig 6B ) . Co-transfection of a synthetic miR-122 mimic ( Fig 6B ) rescued HCV and partially rescued HCV-122/15 . In contrast , NPHV-SL1/miRBRC83A and HCV-S2-GGCGUG replication was only slightly enhanced by miR-122 addition and NPHV-delta-UUGGCG and NPHV-GGCAGG spread with similar efficiency with or without miR-122 . To further confirm the absence of miRNA binding for these viruses , we performed AGO-CLIP of NPHV-delta-UUGGCG in ΔDICER Huh-7 . 5 cells compared to HCV and NPHV-SL1/miRBRC83A . As expected , replication and miRNA binding was observed for HCV only after addition of miR-122 ( Fig 6C ) . The same result was found for NPHV-SL1/miRBRC83A despite replication in miRNA deficient cells . No AGO/miRNA binding was observed for NPHV-delta-UUGGCG . These data prove that these selected NPHV/HCV variants can replicate in the complete absence of mature miRNAs . To determine whether miR-122 independent NPHV variants depend on miR-122 for RNA 5’ end protection , we measured the RNA stability of HCV , NPHV-SL1/miRBR , NPHV-SL1/miRBRC83A , and NPHV-GGCAGG in ΔmiR-122 Huh-7 . 5 cells . miR-122 but not miR-430 supplementation enhanced RNA stability , but only for miR-122 dependent variants ( Fig 6D ) . Thus , although NPHV has a much larger SL1 structure , NPHV may still utilize miR-122 to enhance protection of the RNA from degradation . NPHV-GGCAGG , however , apparently does not need miRNA binding to protect its RNA against 5’ degradation . miR-122 is expressed in Huh-7 . 5 cells [2] , and reports showed that exogenous expression of miR-122 can facilitate efficient replication of HCV in other hepatic cell lines such as Hep3B and HepG2 . Low-level replication was also observed in non-hepatic cells , including 293T kidney cells or engineered immortalized mouse fibroblasts ( iMEF ) [34–37] . We therefore tested whether the selected miR-122 independent strains could replicate in non-hepatic cells . As expected , HCV did not replicate in 293T cells ( Fig 7A ) . In contrast , NS5A positive cells were observed for NPHV-SL1 , NPHV-SL1/miRBRC83A and HCV-S2-GGCGUG , although at very low frequencies . NPHV-delta-UUGGCG and NPHV-GGCAGG infected 2–5% of the cells similar to the frequency of HCV replication in these cells upon miR-122 addition . NPHV-delta-UUGGCG replication in 293T cells was completely abolished after addition of Daclatasvir , a potent NS5A inhibitor [38] , thus confirming its authentic replication in 293T cells ( Fig 7B ) . ApoE is an essential factor for infectious HCV production , including in miR-122-supplemented 293T cells [39] . We therefore asked whether infectious particles could be produced in the presence of exogenous ApoE . In 293T cells transduced with a lentivirus expressing RFP-ApoE , low levels of infectious particles were produced by NPHV-delta-UUGGCG , but not by HCV , HCV-S2-GGCGUG or NPHV-GGCAGG , transfected cells ( Fig 7C ) . No virus production was observed without ApoE expression . Despite the highly attenuated particle production in 293T-ApoE compared to Huh-7 . 5 cells , this proved that miRNA independent NPHV/HCV chimeras could replicate and produce infectious progeny in non-hepatic cells .
miR-122 has attracted great interest as a host requirement for HCV replication and hence a potential antiviral target . Blocking miR-122 leads to prolonged viral inhibition in cell culture , chimpanzees , and patients [5–7] . In addition , due to their long-lasting effects , miR-122 inhibitors are currently being considered for special patient populations where adherence to strict daily therapeutic regimens is problematic [40] . Therefore , it remains important to understand the miR-122 requirement of HCV and to study resistant variants that can replicate in miR-122 depleted environments . Furthermore , both the requirement of a host miRNA for an RNA virus , and the non-classical binding of miR-122 to the HCV 5’UTR are unique . We still have an incomplete understanding of how this interaction evolved and its biological importance . NPHV is the closest homolog of HCV , making it an interesting comparative model . Understanding NPHV is also relevant given its possible association with equine liver disease [26 , 41] . Whereas all HCV genotypes have two conserved miR-122 binding sites , both being essential for replication [2 , 3] , only one binding site was predicted for NPHV [18 , 19] . The length of the NPHV miRBR is comparable to that of HCV , raising the question whether a second miR-122 molecule binds in a non-canonical manner or , alternatively , a different miRNA binds this region . Our in vivo NPHV AGO-CLIP studies confirmed miRNA binding across the miR-122 site . We therefore took advantage of replication-competent NPHV/HCV chimeras engineered to contain parts of the NPHV 5’UTR . NPHV-SL1/miRBR was of particular interest , since it contains the NPHV miR-122 binding site . Although NPHV-SL1/miRBR had partially attenuated kinetics of viral protein accumulation and infectious virus production , more fit variants including NPHV-SL1/miRBRC83A , NPHV-delta-UUGGCG , and NPHV-GGCAGG could be readily selected in cell culture . In our previous studies , we demonstrated that HCV miR-122 tropism could be replaced by other miRNAs , such as miR-15 [15] . It was therefore possible that the increased fitness of NPHV-SL1/miRBRC83A , NPHV-delta-UUGGCG , and NPHV-GGCAGG was acquired by binding to alternative miRNAs . However , we found that NPHV-SL1/miRBRC83A bound only one molecule of miR-122 , and neither NPHV-delta-UUGGCG nor NPHV-GGCAGG bound any miRNA . This was based on two observations: First , in contrast to HCV-m15 , which could not replicate in ΔDrosha Huh-7 . 5 cells supplemented with miR-122 [15] , NPHV-SL1/miRBRC83A was not attenuated in ΔDICER Huh-7 . 5 cells supplemented with miR-122 compared to wt Huh-7 . 5 cells . The NPHV-delta-UUGGCG and NPHV-GGCAGG variants were not attenuated in ΔDICER Huh-7 . 5 cells or in ΔmiR-122 Huh-7 . 5 cells ( Fig 6A and 6B ) . Second , miR-122 and miR-15 , respectively , were the only miRNAs identified from AGO-CLIP chimeras at the miRBR region of NPHV-SL1/miRBRC83A and NPHV-SL1/miRBR-m15 . This was in contrast to the identification of two independent peaks for miR-15 and miR-122 for HCV122/15 and HCV15/122 [16] . Furthermore , no miRNA peak could be observed for NPHV-delta-UUGGCG , even after addition of miR-122 ( Fig 6C ) . Taken together these results strongly suggest that newly adapted NPHV/HCV variants do not bind any other miRNA . The cell culture selected HCV-U3 variant , which is resistant to miR-122 inhibition [31] , contains a large extended SL1 and only one miR-122 site . This 5’UTR is therefore structurally similar to NPHV . A larger 5’-terminal SL1 may at least partially compensate for miR-122 binding , possibly by preventing degradation by exonucleases . This suggests that it is possible to replace one miR-122 seed site and obviate miR-122 dependence ( at least in part ) by combining a larger SL1 with sequence changes in the miRBR region . It is interesting to speculate why HCV requires two molecules of miR-122 , whereas related hepaciviruses such as NPHV , require only one . From the current and previous studies , it is evident that HCV replication is attenuated if only one copy of miR-122 is bound . This includes mutants of individual seed sites [2 , 4 , 11 , 15 , 42] and the current study of HCV-S2-GGCGUG in Huh-7 . 5 cells and HCV-122/15 in ΔmiR-122 Huh-7 . 5 cells . Interestingly , the C83A mutation of NPHV-SL1/miRBR may induce a conformational change in the miRBR region to a more relaxed structure ( S2A and S2B Fig ) . Although it is tempting to speculate that this more relaxed RNA structure might facilitate miR-122 access to the miRBR region , e . g . by providing access to auxiliary pairing with the ACC motif at nt 87–89 , the C83A mutation also facilitated more efficient infection kinetics in the absence of miR-122 ( Fig 3A and 3B ) , indicating that mechanisms beyond miR-122 engagement may be at play . HCV , HCVG28A from a previous study [42] , NHPV-delta-UUGGCG , ( S1D–S1F Fig ) , but not NPHV-GGCAGGG or HCV-S2-GGCGUG ( S1C and S1G Fig ) have similar relaxed structures , suggesting that these predicted secondary structures of miRBR do not necessarily correlate with miR-122 dependency . We identified NPHV-SL1/miRBR variants that are either partially or completely independent of miR-122 yet we could not identify any completely independent HCV-based variants , presumably because seed site 1 of HCV was still intact ( Fig 6A ) . Interestingly , NPHV-delta-UUGGCG , but not HCV variants , could produce infectious viruses in 293T kidney cells expressing ApoE ( Fig 7C ) . Since miR-122 is liver specific and remains an important factor for HCV tropism , the fact that NPHV only requires one molecule of miR-122 could lower the threshold for NPHV to infect other organs . However , except for sporadic evidence of extrahepatic presence of NPHV [18 , 41] , it remains to be determined if replication occurs in other tissues in vivo . Still it remains feasible that low-level replication outside the liver could occur and influence the course of disease or transmission between hosts . It was striking , but not surprising , that the fittest HCV variant emerging from our saturation mutagenesis library in wild type Huh-7 . 5 cells carried the wild type seed sites . This confirms that in the Huh-7 . 5 environment miR-122 binding HCV is indeed the optimal variant ( Fig 5C ) . Another interesting finding was that HCV-122/15 , which we previously predicted and confirmed to be viable , was also selected in Huh-7 . 5 cells . HCV therefore appears to function most efficiently with two miRNA seed sites . While miR-122 could be substituted by other miRNAs , HCV with two miR-122 sites remains superior . In contrast , NPHV either acquired the wild type miR-122 site in combination with an adaptive mutation ( NPHV-SL1/miRBRC83A ) or carried a 15-nt truncation ( NPHV-delta-UUGGCG ) in the miRBR . Thus , miR-122 binding NPHV variants were also selected as long as miR-122 was available . We were able to isolate NPHV and HCV variants in the ΔmiR-122 Huh-7 . 5 cells with similar efficiency ( Fig 6A ) . This makes it likely that during hepacivirus evolution miR-122-indepedent variants have emerged . It is therefore curious why most hepaciviruses isolated thus far retain at least one miR-122 binding site and hence hepatocyte tropism . Clearly , basic replicative functions can ensue in the absence of miR-122 using alternative 5’UTR structures and/or binding sites for different miRNAs . We therefore speculate that sampling of hepaciviruses in nature has been biased to those that can establish chronic infection in their animal hosts . Retaining miR-122 dependence and restricting replication to hepatocytes may allow these viruses to take advantage of the tolerogenic liver environment [43 , 44] to establish and maintain chronicity . Variants with the ability to replicate in other cell types might be selected against if they elicit an adaptive immune response that eliminates infection and prevents chronicity . Consequently , we posit that there may be hepaciviruses in nature that are not strictly hepatotropic and that these viruses will be more likely to cause acute resolving infections with different modes of transmission to ensure their survival . Besides the interaction of miR-122 with the 5’UTR of NPHV , our in vivo AGO-CLIP studies demonstrated miRNA binding at all three other conserved miR-122 sites in the NPHV polyprotein coding region ( Fig 1B ) . Unfortunately , we lack the cell culture systems needed to assess the potential contributions of miRNA binding at these NS5A/5B sites . Thus , it remains possible that these sites also contribute towards NPHV miR-122 dependence . For HCV , however , miR-122 sites in the ORF and 3’UTR do not appear to influence viral replication and/or production [15 , 45] . It will be interesting to explore this further using infectious clones in vivo , or after the development of tractable NPHV culture systems [28] . In conclusion , using a panel of NPHV/HCV chimeras , we were able to test the miRNA dependence of NPHV in cell culture . We demonstrate that one molecule of miR-122 is the only miRNA that binds the NPHV miRBR region , consistent with our in vivo data . Using NPHV/HCV chimeras , miR-122 independent variants could be selected that were capable of extra-hepatic replication . This indicates that the interaction of one molecule of miR-122 with the NPHV miRBR does contribute to its hepatic tropism but that subtle changes in the NPHV 5’UTR have higher potential to break the tissue tropism barrier as compared to HCV . Given that most hepaciviruses observed in nature contain at least one miR-122 binding site despite the fact that minor changes in the miRBR can weaken or even obviate miR-122 dependence indicates a strong selective pressure to maintain hepatotropism . We suggest that this selective force is not for basic hepacivirus replicative functions but rather to exploit the tolerogenic liver environment to orchestrate chronicity .
We used NPHV infected liver biopsies from two horses . This did not require euthanasia of any animal . Ultrasound-guided percutaneous biopsies were taken from horses , using standard procedures at the College of Veterinary Medicine , Cornell University and adhered to the Institutional Animal Care and Use Committee protocol at this institution . NPHV-5’UTR , NPHV-IRES , NPHV-SL1/miRBR , and NPHV-SL1 were constructed in the HCV J6/JFH Clone2 backbone [29] , by replacing the 5’UTR ( nt 1–341 ) , IRES ( 43–341 ) , SL1-miRBR ( 1–42 ) , or SL1 ( 1–20 ) of HCV by the corresponding sequence of the 5’UTR ( 1–384 ) , IRES ( 103–384 ) , SL1-miRBR ( 1–102 ) or SL1 ( 1–74 ) of the NZP1 NPHV isolate [28] . HCV-122/15 and HCV-S2-GGCGUG were constructed by replacing seed site 2 of HCV ( 37–42 ) with the miR-15 binding site ( GCTGCT ) or GGCGTG , respectively . SL1-miRBRC83A , NPHV-GGCAGG , and NPHV-delta-UUGGCG were constructed from NPHV-SL1/miRBR by introducing single mutations or exchanging the miRBR site ( also see Fig 5C ) . NPHV-SL1/miRBR-m15 was constructed by replacing the miR-122 binding site at position of 96–101 of NPHV-SL1/miRBR with the miR-15 binding site ( GCTGCT ) . For TRIP-hApoE3shres-TagRPF , the cDNA clone of human apolipoprotein E ( NM_001302691 ) was amplified from Huh-7 . 5 cDNA using the primers RU-O-19451 and RU-O-19452 that contains MluI and BamHI sites . This was then cloned into pDONOR221 ( Fisher Scientific ) , which was linearized by the same pair of endonucleases . All primer sequences are listed in S1 Table . Huh-7 . 5 [46] and ΔmiR-122 Huh-7 . 5 hepatoma cells [15] , derived previously in our laboratory , were maintained in Dulbecco's Modified Eagle Medium ( DMEM , Invitrogen ) supplemented with 0 . 1 mM nonessential amino acids ( Invitrogen ) and 5% fetal bovine serum ( FBS ) . 293T cells ( ATCC ) were maintained in DMEM supplemented with 10% FBS and 0 . 1 mM NEAA as described [47] . ΔDICER Huh-7 . 5 hepatoma cells were generated as described below and maintained in DMEM supplemented with 10% FBS and 0 . 1 mM NEAA . To produce 293T-ApoE , 293T cells were transduced with lentiviruses produced from pTRIP-hApoE3shres-TagRFP in a 293T producer culture , as previously described [37] . Efficiency of transduction was confirmed by the percentage of RFP positive cells before further analysis . Three different 293T cell lines derived from single colonies of 293T cells were tested to confirm the consistent observation of NS5A positive cells with NPHV-delta-UUGGCG . For Daclatasvir ( DCV ) treatment , 293T cells were pre-incubated with 10 nM of DCV ( 100-fold EC50[48] ) one day before transfection with viral RNA . To make ΔDICER Huh-7 . 5 cells , we deleted exons 2 and 19 in the DICER gene in Huh-7 . 5 cells described above . Guide sequence pairings were as follows: Dicer . sgRNAEX2 . 1a with Dicer . sgRNAEX2 . 1b , Dicer . sgRNAEX2 . 2a with Dicer . sgRNAEX2 . 2b , Dicer . sgRNAEX19 . 1a with Dicer . sgRNAEX19 . 1b , and Dicer . sgRNAEX19 . 2a with Dicer . sgRNAEX19 . 2b . Guide RNAs were cloned into pX458-SpCas9- ( BB ) -2A-GFP ( Addgene , #48138 ) . After sequence confirmation , transfection and single cell dilution , cloning proceeded as previously described [15] . To genotype single cell clones and to approximate editing efficiency in bulk cells , DNA was extracted using QuickExtract ( Epicenter ) and DICER exon 2 and 19 loci were PCR amplified using primers DicerEX2_GenomicF with DicerEX2_GenomicR , and DicerEX19_GenomicF with DicerEX19_GenomicR , respectively . The resulting PCR products underwent gel electrophoresis . The sole surviving homozygous deletion clone was expanded . NPHV/HCV and HCV recombinant RNAs were in vitro transcribed from XbaI linearized DNA plasmids using T7 RiboMAX Express Large Scale RNA Production System ( Promega ) . RNA was treated with RQ1 DNase ( Promega ) at 37°C for 15 min and purified on RNeasy columns ( Qiagen ) . For transfection of Huh-7 . 5 based cell lines , 1 μg RNA was mixed with 5 μL Lipofectamine 2000 ( Life Technologies ) in 500 μL OptiMEM , incubated 10 min and added to 3 . 5x105 cells in 6-well plates where media was changed to DMEM containing 1 . 5% FBS and 1% NEAA before transfection . For transfection of 293T-based cell lines , transfection was performed with 5x105 cells . 10 pmol of miR-122 was co-transfected with viral RNAs where indicated . To increase the transfection efficiency , cells were then spinoculated for 30 min at 37°C with 1000 g . Cells were split every 2 days , 1/3 of cells were seeded for the next time point , 2/3 of cells were pelleted , fixed with 2% PFA , and used for flow cytometry detection of HCV NS5A expression by staining with the 9E10 antibody conjugated with Alexa 647 . Supernatant aliquots were stored at -80°C for virus titration assays ( TCID50 ) . HCV infectious titers were determined by a limiting dilution assay on naïve Huh-7 . 5 cells as previously described [49] . To generate saturated libraries of NPHV-Rand and HCV-Rand containing all possible combinations of nucleotides in the randomized region , we constructed bacterial transformation libraries with at least 10 times more colonies than the combination of all possible nucleotides ( 4^6 = 4096 ) . Briefly , we introduced randomized seed sites into NPHV-SL1/miRBR and HCV by PCR with RU-O-21135 and RU-O-17131 or RU-O-21134 and RU-O-17131 , respectively . After restriction digest with NotI and KpnI , the PCR product was ligated into the parental plasmid and transformed into DH10B with electroporation transformation . After transformation , cells were revived in 500 μl SOC medium . 2 μl were spread on P10 LB medium to count colonies . To minimize potential bias of bacterial growth , bacteria were spread onto 2xP500 LB plates . After overnight incubation , cells were scraped into 20 ml LB medium and plasmids were purified with Plasmid Maxi-prep kit ( Qiagen ) without further expansion . For selected viruses , as indicated in the main text and figures , the 5’UTRs of viruses present in the supernatant was sequenced using 5’RACE according to manufacturers protocol ( Invitrogen ) . The primer RU-O-17102 was used for reverse transcription and RU-17104 and RU-18654 for the subsequent 1st and 2nd PCRs , respectively . PCR products were then cloned into TOPO-TA vectors ( Invitrogen ) for sequencing . Corresponding plasmids were included in the same procedure as negative control . Standard AGO-CLIP was done as described [50] . To enable cost-effective multiplexing , we added sequencing adapters and 5’ indices in the 2nd PCR step using the primers listed in S1 Table . This strategy uses the DP5 and DP3 sequences of the 1st PCR product as priming sites to add 5’ indices and 3’ adapters for a short ( 6–16 cycles ) 2nd PCR step . CLEAR-CLIP is based on standard AGO-CLIP with modifications to enrich for miRNA-target chimeras [30] . Briefly , the following procedures replace the post-immunoprecipitation steps of standard AGO-CLIP: ( i ) 5’-end phosphorylation using PNK ( 3’ phosphatase minus ) , ( ii ) Over-night chimera ligation using T4 RNA ligase 1 , ( iii ) Alkaline phosphatase treatment to remove 3’ phosphate groups , ( iv ) 3’-linker ligation using truncated T4 RNA ligase 2 and pre-adenylated linker ( using a pre-adenylated linker and omitting the enzyme in step ii allows negative controls to distinguish cellular vs . on-bead ligation events ) , and ( v ) Radio labeling using T4 PNK and [γ-32P]-ATP . Here , the ligase-free controls used to establish the method [30] were not needed , and the 3’ linker ligation was therefore done with T4 RNA ligase 1 and a radioactively labeled phosphorylated RNA linker according to the standard AGO-CLIP protocol [50] , and not using truncated T4 RNA ligase 2 and a pre-adenylated linker . Accordingly , the subsequent PNK treatment was done with non-radioactive ATP . For in vivo experiments , NPHV infected liver biopsies from two horses were taken and snap-frozen at the College of Veterinary Medicine , Cornell University adhering to Institutional Animal Care and Use Committee protocols as previously described [28] . Samples were powderized in liquid nitrogen followed by cross-linking with UV 254nm 3x at 400mJ/cm2 . Results were combined from a 54mg biopsy from day 23 of an acutely infected horse ( by intrahepatic RNA inoculation , NZP1 strain ) with a serum titer of 106 . 7 GE/mL and a 137mg biopsy of a chronically NPHV infected horse with a serum titer of 105 GE/mL . Cultured Huh-7 . 5 cells for CLIP were transfected with NPHV-SL1/miRBR or NPHV-SL1/miRB-m15 and cross-linked 4 days post transfection; HCV , NPHV-SL1/miRBRC83A , HCV-S2-GGCGUG , and NPHV-delta-UUGGCG , either in the absence or presence of miR-122 in ΔDICER Huh-7 . 5 cells were cross-linked 3 days post transfection . 1 `μg of HCV-p7ns2Gluc-GNN or NPHV-p7ns2Gluc-GNN RNA , synthesized as previously described [51] , was co-transfected in the presence of miR-122 or miR-430 mimic ( Dharmacon ) using the transfection protocol described above for ΔDICER Huh-7 . 5 cells in 6-well plates . At each time point , cells were washed with PBS 5 times and detached using trypsin . Cells were pelleted at 500 g , 4°C for 5 min and washed once with 1 mL of cold PBS . Cells were then pelleted under the same conditions and resuspended in 200 μL PBS . Cells were stored at -80°C or processed immediately by adding 1 mL Trizol . Samples were then mixed and 200 μL of chloroform were added and processed as described . After phase separation , the upper phase ( approximately 600 μL ) was transferred to another RNase free tube , mixed 1:1 with ethanol ( 100% ) and loaded on RNeasy mini kit column ( Qiagen ) . RNA was further purified as described and quantified by a two-step procedure using the QuantiTect Reverse Transcription kit ( Qiagen ) for cDNA synthesis followed by a qPCR using SYBR Green PCR Master Mix ( Applied Biosystems ) at 95°C for 10 min followed by 40 cycles of 95°C for 15 sec , 60°C for 15 sec and 72°C for 15 sec using primers RU-O-17104 and RU-O-21903 . Standard curves were generated from the same RNAs . RNA transfection was done as described above using NPHV-p7ns2Gluc-GNN and the indicated derivative RNAs . At 6hr post transfection , the supernatant of each well was collected to assay for Renilla luciferase as described [15] and read on a FLUOstar Omega ( BMG Labtech ) . | It has been known for more than 10 years that the hepatitis C virus ( HCV ) genome binds two copies of the liver-specific microRNA ( miRNA ) , miR-122 . But until recently , it was unknown whether this interaction was unique to HCV or also conserved among other hepaciviruses . Now , due to our expanded view of the hepacivirus family , we know that most , if not all , hepaciviruses have at least one predicted miR-122 binding site in their 5’ untranslated region ( 5’UTR ) . In this study , we aimed to obtain a broader understanding of hepacivirus/miR-122 interactions and determine how miR-122 dependence influences tissue tropism . To do this , we chose to study the equine non-primate hepacivius ( NPHV ) , the closest relative of HCV . NPHV has one predicted miR-122 site in its 5’UTR . We show that minor changes in the 5’UTR of HCV and NPHV/HCV chimeras weaken or obviate the need for miR-122 for virus replication . Overall , our data suggest that miR-122-independent hepaciviruses have likely been sampled during evolution , but hepaciviruses may have been selected to utilize miR-122 to restrict replication to the tolerogenic liver environment to help avoid immune clearance . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cell",
"binding",
"transfection",
"cell",
"physiology",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"gene",
"regulation",
"hepacivirus",
"pathogens",
"biological",
"cultures",
"293t",
"cells",
"microbiology",
"viruses",
"micro... | 2017 | miRNA independent hepacivirus variants suggest a strong evolutionary pressure to maintain miR-122 dependence |
Bacterial communities associated with roots impact the health and nutrition of the host plant . The dynamics of these microbial assemblies over the plant life cycle are , however , not well understood . Here , we use dense temporal sampling of 1 , 510 samples from root spatial compartments to characterize the bacterial and archaeal components of the root-associated microbiota of field grown rice ( Oryza sativa ) over the course of 3 consecutive growing seasons , as well as 2 sites in diverse geographic regions . The root microbiota was found to be highly dynamic during the vegetative phase of plant growth and then stabilized compositionally for the remainder of the life cycle . Bacterial and archaeal taxa conserved between field sites were defined as predictive features of rice plant age by modeling using a random forest approach . The age-prediction models revealed that drought-stressed plants have developmentally immature microbiota compared to unstressed plants . Further , by using genotypes with varying developmental rates , we show that shifts in the microbiome are correlated with rates of developmental transitions rather than age alone , such that different microbiota compositions reflect juvenile and adult life stages . These results suggest a model for successional dynamics of the root-associated microbiota over the plant life cycle .
Plants assemble soil-derived root-associated microbial communities referred to as microbiota [1–3] . Taxa within the root-associated microbiota have been found to be beneficial for plant growth and resistance to biotic and abiotic stresses [4 , 5] . Previous studies on plant microbiota have primarily characterized the bacterial and archaeal communities , and in what follows , the term microbiota encompasses these communities . Plants host root-associated microbiota in 3 spatially distinct root compartments with significantly different compositions: the soil adjacent to the root ( the rhizosphere ) , the root surface ( the rhizoplane ) , and the root interior ( the endosphere ) [6–8] . Each root-associated compartment has significantly different compositional profiles from the communities in unplanted soil , thus indicating the roots enrich for subsets of the soil microbiota [6–10] . It has been suggested that enrichment of copiotrophic microbes occurs in the rhizosphere , whereas unplanted soils are enriched for oligotrophic microbes [11] . This enrichment process and how the root-associated microbiota change throughout the life cycle of the plant remain poorly characterized . Limited information is available about the spatiotemporal dynamics of the plant root-associated microbiota , especially for the root endosphere . Previously , we characterized temporal progressions of the microbiota across the rhizosphere-endosphere continuum of greenhouse grown rice seedlings from transplantation to 2 weeks after transplantation [8] . Over this relatively short period of time , the microbiota shifted to become more similar to those hosted by rice plants that were 6 weeks old , suggesting that the endospheric microbiota reaches a steady state . From that study , however , we were unable to formulate how root microbiota change over the full life cycle of rice plants . The most systematic study available on the possible impact of plant developmental stage on the composition of the plant root microbiota has been on the perennial plant Arabis alpina [12] . Analysis of the rhizosphere and root microbiota during 3 time points encompassing a 28-week period under greenhouse conditions showed changes in the microbiota with soil residence time [12] . An early flowering mutant had no significant effect on microbiota compared to the nonflowering wild-type plants , suggesting that the observed changes were due to soil residence time , rather than plant development stage . These data supported a model in which “after microbiota acquisition during vegetative growth , the established root-associated bacterial assemblage is structurally robust to perturbations caused by flowering and drastic changes in plant stature” [12] . However , it is unknown how root-associated microbiota might vary on weekly timescales from seedling to maturity under field conditions for annual plants , what patterns of change can be expected over the annual life cycle , and whether microbiota assemble in a consistent manner across geographically and climatically distinct regions . Given the likely role that root-associated microbiota play in conferring beneficial properties to the host plants [1] , it is important to understand how root microbiota are structured over the life cycle of their crop hosts and whether temporal shifts are consistent between geographic locations . Rice is a staple crop for a large proportion of the world population [13] , and rice cultivation is a major contributor to greenhouse gas emissions arising from microbial activity [14] . Therefore , in this study , we use a high-resolution spatiotemporal approach to detail the successional progression of the microbiota across the rhizosphere , rhizoplane , and endosphere compartments over the life cycle of rice plants grown under field conditions over 3 consecutive growing seasons as well as field sites from 2 geographically distinct regions of the United States . Using related rice varieties with different developmental rates grown under field conditions , we conclude that in contrast to predictions of prevailing models , shifts in the root microbiome are correlated with age and developmental stages and that different microbiota compositions reflect juvenile and adult life stages . Partial validation of these hypotheses was obtained by successfully utilizing computational models to predict plant age .
To investigate the bacterial and archaeal microbiota associated with the rhizosphere , rhizoplane , and endosphere of rice roots as well as those associated with unplanted soil over the growing season , we sequenced the V4 region of the 16S rRNA gene of 1 , 510 samples , revealing 10 , 893 different operational taxonomic units ( OTUs ) . Depending on the season , we collected plants from the field sites either weekly or every other week ( see Materials and methods ) . To visualize the underlying driving forces of microbial community variation in our data , we used principal coordinates analysis ( PCoA ) in combination with permutational multivariate analysis of variance ( PERMANOVA ) on Bray-Curtis dissimilarities between the samples ( S1 Table ) . Because of the relatively large number of samples in this dataset and the use of multiple sequencing runs , we first inspected how much variance was partitioned to sequencing run . We found that samples significantly differed by sequencing run , but this effect was minor compared to the experimental variables tested ( R2 = 0 . 009 , P = 0 . 001 ) . We found that samples from 2014 and 2015 California samples as well as the 2016 Arkansas samples display a spatial pattern of divergence along the first principal coordinate , where communities in bulk soil samples cluster on one end of the axis and endosphere communities cluster on the other ( Fig 1A ) . PERMANOVA of pairwise distances between microbial communities indicated that the microbiota differed significantly between root-associated compartments ( R2 = 0 . 22 , P < 0 . 001 ) , in agreement with prior observations [6–9 , 11 , 15] . The microbiome samples from Arkansas and California separated along the second principal coordinate ( Fig 1B ) . This observation was supported by the PERMANOVA statistic ( R2 = 0 . 13 , P < 0 . 001 ) . The soil from the Arkansas and California field sites differs in that the Arkansas site has a sandy loam soil compared to the California site , which has a silty clay soil ( https://websoilsurvey . sc . egov . usda . gov ) . Chemical analysis showed that the California soil had higher concentrations of potassium , sodium , calcium , magnesium , organic matter , and pH content than the Arkansas soil ( S2 Table ) . Arkansas soil had greater concentrations of nitrate and phosphorus than the California site . We next measured the effect of plant age on the root-associated microbiota . We found that this effect is observable across the third principal coordinate of the PCoA plot and is consistent across the 2 growing regions and across the 2 growing seasons within California ( Fig 1C ) . The amount of variance partitioned to the effect of plant age on the root-associated microbiota was comparable between the 2 growing regions ( California: R2 = 0 . 093 , P < 0 . 001; Arkansas: R2 = 0 . 080 , P < 0 . 001 ) . We observed that the rhizoplane and endosphere microbial communities shift over the first 7 to 8 weeks after germination but stabilize thereafter ( Fig 1D , S1 Fig ) . This pattern was observed over both growing seasons in California and also in the Arkansas field trial . However , a stabilization pattern was not observed in the rhizosphere microbial communities ( Fig 1D , S1 Fig ) despite the rhizosphere microbiota showing distinct shifts over the season ( Fig 1C ) . It is interesting to note that the selected rice varieties grown in California and Arkansas reach panicle initiation ( entry into reproductive growth ) 8–9 weeks after germination in their respective locations , suggesting a correlation between plant developmental stage and root-associated microbiota dynamics . The multiyear sampling scheme of our experimental design allowed us to quantify the effect of year-to-year variation on the root-associated microbiota . Although the root-associated microbiota varied significantly across the 2 different years at the California site , the effect was relatively small ( R2 = 0 . 01 , P = 0 . 001 ) compared to the other factors analyzed within this experiment . Together , these data suggest that the root-associated microbiota shifts in each root-associated compartment during the vegetative growth stage of the season and stabilizes upon entry into reproduction and that these patterns are consistent across different growing seasons . We next sought to characterize the specific phyla responsible for the significant differences between the root-associated compartments and how these various phyla change in abundance over the course of the season . Overall , we noticed similarities in patterns in phyla abundance over time between the 2 growing regions ( Fig 2A ) . Because the Proteobacteria phylum contains a broad phylogenetic makeup and because Proteobacteria make up a large proportion of the rice root-associated microbiota , we further divided the Proteobacteria phylum into its respective classes for this analysis . To model increasing or decreasing relative abundance of individual phyla between the various root-associated compartments , we assigned each compartment a value relative to its spatial position: bulk soil was position 0 , rhizosphere was position 1 , rhizoplane was position 2 , and endosphere was position 3 . We modelled how each phylum either increased or decreased over these positions using beta regression . Beta regression is used for modeling dependent variables that lie in the interval ( 0 , 1 ) and is thus useful for modelling individual taxa as a relative proportion of the total microbial community . Using this method , of the 55 detectable phyla , we identified 42 phyla that significantly differed in spatial distribution from the exterior to the interior of the root ( Fig 2B ) . There were 20 phyla whose relative abundance significantly increased from the soil environment to the root interior and 22 microbial phyla whose relative abundance decreased from the soil environment to the root interior . The absolute values of the regression coefficients were overall greater for root-depleted phyla , suggesting that a smaller proportion of the soil microbiota is enriched by the plant , while more microbial taxa are depleted . We again performed beta-regression within each compartment to identify phyla that significantly increased or decreased in relative abundance as a function of plant age ( Fig 2C ) . In general , trends in abundance over the season were consistent across the rhizocompartments ( excluding the bulk soil ) for each phylum . For instance , in the rhizosphere , rhizoplane , and endosphere , Betaproteobacteria , Verrucomicrobia , and Gemmatimonadetes consistently decreased over the course of the season , while Nitrospirae , WS3 , Deltaproteobacteria , Epsilonproteobacteria , Euryarchaeota , and Spirochaetes all increased . In the rhizoplane and endosphere , most of the taxa with dynamic abundance patterns significantly decreased over the course of the season ( 12/22 and 32/40 , respectively ) , suggesting that the root-associated compartments are initially colonized by a diverse set of microbes from the soil but are either eventually outcompeted by other taxa or are selected against by the host plant . Together , these results indicate that despite large differences in microbiota composition between the growing regions , consistent trends in abundance at the phylum level define microbiota differences between the rhizocompartments and over the course of the growing season . We next sought to identify individual microbial OTUs that could be used to discriminate plant age . We identified 85 of the most important age-discriminant OTUs from the rhizosphere and endosphere from full random forest ( RF ) models and used these OTUs to develop a more accurate sparse RF model for each compartment to predict plant age ( see Materials and methods ) . We found the sparse RF models explained 91 . 5% and 88 . 4% of the variance related to plant age for the rhizosphere and endosphere , respectively . Plant age was accurately predicted by the compartment-specific sparse RF models for the California and Arkansas sites ( Fig 3A ) . Additionally , we note that the ages of plants sampled from California during the 2015 season were able to be accurately predicted , despite the models not being trained on this data . These results indicate that specific sets of microbes in both the endosphere and rhizosphere behave in consistent patterns over the course of the rice plant’s life cycle between seasons and across geographic regions . We next examined the phylogenetic composition of the 85 OTUs used for plant age prediction in the rhizosphere and endosphere sparse models . The OTUs used in the sparse RF models were phylogenetically diverse . In the endosphere model , the model used OTUs belonging to 10 phyla , 19 classes , 29 orders , and 37 families , while the rhizosphere model used OTUs belonging to 11 phyla , 24 classes , 37 orders , and 45 families . Consistent with general trends in alpha diversity between the compartments ( S2 Fig ) , the rhizosphere model contained a more phylogenetically diverse set of microbes for classification . We used a linear model approach to classify the OTUs included in each model as being early ( significantly decreasing slope over the course of the season ) , late ( significantly increasing slope over the course of the season ) , or complex ( slope not significantly different than 0 over the course of the season ) colonizers over the course of the season ( Fig 3B ) . We found that for both the rhizosphere and endosphere sparse RF models , most of the OTUs were late colonizers , while fewer OTUs were classified as complex or early colonizers ( Fig 3B ) . The rhizosphere and endosphere models shared 22 OTUs , 20 of which were in agreement for early/late/complex colonizer classification . Within the OTUs classified as late or early colonizers for the rhizosphere and endosphere models , OTUs belonging to the Betaproteobacteria class were the most represented class ( Fig 3B ) . At the order level , however , the OTUs making up the increasing or decreasing fraction within the sparse models were significantly different ( S3 Fig ) . Betaproteobacteria OTUs classified as early colonizers were mainly composed of Burkholderiales , while the late colonizing OTUs were mainly composed of Rhodocyclales and SBla14 . We also noticed differences between the rhizosphere and endosphere RF models for the taxa they were using for classification . The endosphere model used many more Alphaproteobacteria OTUs than the rhizosphere model , which mirrors differences in trends at the phylum level ( Fig 2 ) . Similarly , the rhizosphere model used OTUs from the phylum Verrucomicrobia , whose relative abundance at the phylum level is significantly higher in the bulk soil and rhizosphere than the endosphere ( Fig 2B ) . We next evaluated the fractional contribution of the age-discriminant OTUs to the total microbiota over the course of the season ( Fig 3C ) . Overall , the age-discriminant OTUs composed a greater proportion of reads in the endosphere compared to the rhizosphere . This was consistent across both the California and Arkansas sites . As expected , the early colonizer OTUs were dominant at the beginning of the season , while those identified as increasing in relative abundance were dominant at the end of the season . Interestingly , we noticed that the switch in dominance between early/late colonizing classified OTUs occurred 8 to 9 weeks after germination , coinciding with the switch from vegetative to reproductive growth for the included varieties . We next asked if the early colonizing age-discriminant OTUs were significantly enriched in their respective root compartments compared to unplanted soil controls . We were only able to use the early time points ( less than or equal to 49 days after germination ) as a comparison because of the inability to gather bulk soil controls in the latter part of the season . We found that the early colonizing microbes in the RF models were predominantly enriched in the rhizosphere and endosphere compared to bulk soil controls ( S4 Fig ) . The late colonizing microbes had more similar relative abundance patterns to the bulk soil control samples . Because this analysis was constrained to using the earlier time points , we expect that the late colonizing OTUs would be predominantly enriched compared to the bulk soil controls in the latter time points . These results suggest that the early colonizing OTUs included in both the rhizosphere and endosphere sparse RF models are selected either actively or passively by the plant roots and their abundance profiles are likely not a product of edaphic processes separate from the plant . Drought is one of the most common and devastating stresses to affect rice production around the world [16] . The activities of soil microorganisms are altered by water status [17 , 18] and have been implicated in alleviating drought symptoms in various plant species under laboratory conditions [19 , 20] . It was unknown until recently how water deprivation affects the rice root-associated microbiota [21] . Drought stress most strongly alters the endosphere microbiota ( relative to the rhizosphere and bulk soil ) , with Actinobacteria and Chloroflexi strongly increasing in abundance and Deltaproteobacteria and Acidobacteria strongly decreasing under drought conditions . Similar patterns have been reported for the root microbiota of other grass species under drought stress [22] . Despite large changes in the rice microbiota in the face of drought stress [21] , it is unknown how these changes reflect on normal development of the microbiota . To study how drought stress may affect normal development of the microbiota , we used our sparse age-predicting RF models in conjunction with the data collected from Santos-Medellín et al . to model plant age as a function of the microbiota ( Fig 4A ) . Santos-Medellín et al . analyzed the rhizosphere and endosphere microbiota of 49-day-old drought-stressed and well-watered rice plants growing in 3 diverse soils under greenhouse settings [21] . When predicting ages from the samples included in this experiment , we found that the soil in which the plants were growing had the largest effect on variation in age predictions ( F = 55 . 69 , P < 2 × 10−16 ANOVA , S4 Table ) . The variation due to soil source was largely caused by samples originating from the Davis soil being predicted as younger compared to samples from plants grown in Arbuckle and Biggs soil . All 3 soils are classified as silty clays , and they were all comparable in chemical composition ( S2 Table ) , but cultivation history differs between the soils: Arbuckle and Biggs soil have cultivated rice every summer season for at least the previous 8 years , while the Davis field has been fallowed for the previous 6 years [21] . Thus , it is likely that soil cultivation history can affect the accuracy of the age-predicting RF models . Watering treatment had the second largest effect on plant age prediction ( F = 32 . 54 , P = 4 . 72 × 10−8 ANOVA , S4 Table ) . This effect was only apparent in the endosphere ( adjusted P = 7 . 9 ×10−9 , Tukey’s post hoc test , S9 Data ) and not in the rhizosphere ( adjusted P = 0 . 34 , Tukey’s post hoc test , S9 Data ) . This finding is consistent with the conclusions by Santos-Medellín et al . that the endosphere microbiota is most affected by water deprivation compared to the other rhizocompartments [21] . In each soil , the drought-stressed plants hosted endosphere microbiota that were consistently less mature than well-watered plants . Because the age-discriminant OTUs are predominantly classified as late colonizers , we considered whether a fundamental shift in the endosphere microbiota could cause a decrease of all age-discriminant OTUs and therefore account for the prediction of RF models that drought-stressed rice plants host immature endosphere microbiota . While we did observe a decrease in the late colonizing age-discriminant microbes , we found that early colonizing age-discriminant microbes have significantly higher relative abundances in the drought-stressed endospheres of each soil ( Fig 4B , P < 0 . 05 Tukey’s Honest Significant Difference Test , S5 Table ) . Together , these data indicate that drought stress is associated with an immature microbiota in the root endosphere , but not rhizosphere . Despite clear distinctions in microbiota composition between the growing regions , PERMANOVA indicated a significant interaction between plant age and the difference in microbiota composition between the sites ( R2 = 0 . 019 , P < 0 . 001 , S1 Table ) . We found that both the endosphere and rhizosphere microbiota became significantly more similar over time between the 2 field sites ( Fig 5A ) . Compared to the rhizosphere , the endosphere microbiota started the season as more dissimilar between the sites but reached comparable levels of similarity by the end of the season . This trend did not hold for the bulk soils ( S5 Fig ) : the bulk soil communities between the sites did not show any change in similarity over time ( P = 0 . 1 ) , although we note that we were only able to sample bulk soil microbiomes over a 5-week span at the beginning of the season due to root proliferation later in the season . Interestingly , the similarity in microbial composition between the plants growing in each site stabilized after the rice plants had entered the reproductive phase . This result suggests that plants growing in disparate field sites initially acquire divergent root-associated microbiota , but the composition of these communities begins to converge throughout vegetative growth and maximizes and stabilizes during the reproductive phase . Together , these results suggest a host selection within the rhizosphere and endosphere that acts on similar microbes within the microbiota provided by the soil community . Additionally , the plant age-predicting RF models were largely composed of OTUs that colonized roots later in the season at each field site . We hypothesized that this was due to the composition of the earlier time points being largely composed of site-specific OTUs . To test this hypothesis , we first identified OTUs in each time point whose abundance was skewed towards a specific site using edgeR , a software package that incorporates empirical Bayes methods to account for overdispersion common across count data [23] . After correcting for multiple comparisons , we identified 3 , 218 OTUs in the rhizosphere and 865 OTUs in the endosphere that were significantly skewed towards one of the field sites . We did not notice a clear trend in the number of OTUs called per time point ( S6 Fig ) . When taking the overall relative abundance of the site-specific OTUs into account , however , we found that the site-specific OTUs made up a significantly greater proportion of the total microbiota of the earlier time points than the later time points in both field sites and in both the endosphere compartment and the rhizosphere compartment ( Fig 5B , S6 Table ) . This effect was different across the rhizosphere and endosphere , with the endosphere having a more pronounced effect . This difference might be due to the rhizosphere hosting plant-responsive microbes as well as soil microbiota that do not respond to plant stimuli . Taken together , these results suggest that the initial colonization of the rhizosphere and endosphere is largely composed of site-specific OTUs; however , 8–9 weeks ( 56–63 days ) after germination , a set of OTUs conserved between the 2 field sites begin to establish in the endosphere and rhizosphere , but not in the bulk soil . Our sparse RF models appeared to detect progressions in the root-associated microbiota that correlate with developmental transitions in the rice plants . However , plant developmental transitions also covary with climatic and edaphic factors , which may have indirect effects on the microbiota , making these factors difficult to uncouple from the direct effect of plant developmental stage on the root-associated microbiota . Because our California study was confined to 1 variety and our Arkansas study was confined to 2 varieties that have very similar developmental progressions , we were unable to unambiguously discriminate the effect of plant development from that of environmental factors on root-associated microbiota assembly . To investigate the direct effect of plant developmental stage on the root-associated microbiota , it is necessary to be able to distinguish the effect of plant age from developmental progression . With this aim , we grew 4 varieties from O . sativa sgp . temperate japonica at the same California field site during the 2016 season: Kitaake , California varieties M206 and M401 , and Nipponbare . These 4 cultivars grow with different developmental progression rates ( Fig 6A ) . The varieties were water seeded in the California field in a complete randomized block design . We sampled plants within each plot every 2 weeks throughout the season , collecting rhizosphere , rhizoplane , and endosphere fractions from the plant roots . We also scored the plants for their developmental stages . Our revised field design allowed us to also collect bulk soil samples throughout the entirety of the season , as compared to the 2014 and 2015 seasons that had restricted soil sampling due to the invasiveness of the rice roots . After sequencing the V4 region of the 16S rRNA gene and filtering out plastid OTUs , we obtained 11 , 986 , 615 total sequences comprising 10 , 547 OTUs from 469 samples . The data from the California 2016 season had similar trends as exhibited in the 2014 and 2015 seasons in California as well as the field trial in Arkansas . We again used PERMANOVA to investigate which experimental factors explained a significant amount of variance in this dataset ( S7 Table ) . The root-associated compartments hosted distinct microbiota ( Fig 6B , R2 = 0 . 328 , P < 0 . 001 ) , and these microbiota varied significantly due to plant age ( Fig 6B , R2 = 0 . 068 , P < 0 . 001 ) . We found that genotype had a very small overall effect on the root-associated microbiota ( R2 = 0 . 010 , P < 0 . 001 ) . This level of variance is smaller than what we have previously detected in rice but is not surprising given that the included varieties were constrained to the O . sativa sgp . temperate japonica subgroup . We observed a significant statistical interaction between plant age and genotype ( R2 = 0 . 008 , P < 0 . 001 ) , suggesting that the trends in microbiota shifts over the season differ depending on the genotype . This result was not observed in the Arkansas field trial , perhaps because the included varieties were very similar in their developmental progression rates . Moreover , we found that plant developmental stage explained more variance in our dataset than plant age ( R2 = 0 . 082 , P < 0 . 001 ) , again suggesting that developmental stage is an important descriptor for root-associated microbiota assembly . To further inspect this observation , we asked whether there was a significant interaction between plant genotype and plant age along the second principal coordinate ( PCo ) of Fig 6B as a function of plant age . We used the second PCo because this axis best differentiated plant age . We hypothesized that if plant developmental rate were to have an effect on the root-associated microbiota , the faster-developing rice varieties would have greater slope estimates than the slower-developing varieties over the second PCo . For each compartment , we fit a linear model to the first 6 collection time points ( corresponding to 84 days after germination , the point at which all varieties had entered the reproductive phase ) . For comparison , a second linear model was fitted to the remainder of the season . These were used to model progression along the second PCo axis as a function of plant age and genotype [15] . We compared models that contained an interaction term between plant genotype and age to models assuming identical slopes between the genotypes and found that for both the rhizosphere and endosphere , models with interaction terms significantly improved the accuracy of the early season models ( P = 0 . 02 , P = 0 . 005 , respectively; S8 Table ) , though a significant improvement was not detectable for the rhizoplane ( P = 0 . 51 ) , possibly reflecting the lower consistency intrinsic to technical limitations of sampling this compartment [8] . However , the late season models did not exhibit similar improvements with inclusion of interaction terms ( P > 0 . 05 for all compartments , S8 Table ) . These results suggest that the genotypes progress at different rates across the second PCo axis during the first 84 days of the season , but not for the remainder of the season . When inspecting the slope estimates for each variety over the second PCo as a function of plant age for the first 84 days of the season , we observed that progression of the microbiota followed similar trends as the plant development rate variation between the cultivars ( Fig 6C ) . Specifically , Kitaake and Nipponbare had significantly different slope estimates from each other in the rhizosphere and endosphere , while M401 and M206 had intermediate slope estimates that were not significantly different to the other varieties . Nevertheless , we did identify a significant difference between M206 and M401 microbiota , both prior to panicle initiation ( R2 = 0 . 015 , P = 0 . 006 ) and after panicle initiation ( R2 = 0 . 013 , P = 0 . 036 ) , suggesting that genotypic differences in microbiota structure do not arise solely from differences in developmental progression . Additionally , there were no significant differences between the genotypes after 84 days on the second PCo axis , with each cultivar having a slope that was not significantly different than 0 . Each compartment had a positive slope estimate over the first 84 days of the growing season , but unplanted soil had a lower slope estimate than the rhizocompartments . Taken together these results indicate that root microbiota shifts throughout the season appear to arise independently of shifts in the bulk soil , and the rate of these shifts in the rhizosphere and endosphere correlate with the developmental progression rate of the host plant . We next used the sparse RF models generated from the California and Arkansas data to predict the age of samples taken from the 2016 season . In general , the models predicted the ages of the plants accurately ( S7A Fig ) , indicating that root microbiota progressed in a similar manner in the California 2016 season as in the 2014 and 2015 seasons . The endosphere predictions showed significant variation due to cultivar ( P = 0 . 001 , ANOVA , S9 Table ) , while the rhizosphere predictions did not ( P = 0 . 191 , ANOVA , S9 Table ) . Furthermore , the endosphere model showed trends in the predictions that correlated with developmental progression rate variation between the cultivars with Nipponbare having predicted ages significantly younger than Kitaake and M206 ( adjusted P = 0 . 0012 and P = 0 . 0021 , respectively , S10 Table ) . When observing the total abundance of age-discriminant OTUs from the models between the cultivars over time , we found that late colonizing OTUs in the endosphere took a significantly longer period of time to establish in Nipponbare compared to Kitaake and M206 ( adjusted P = 0 . 0002 and P = 0 . 0016 , respectively; S7B Fig ) . These data suggest that the rhizosphere age-predicting model is largely robust to differences in phenology due to plant genotype but that the accuracy of the endosphere model is affected by variation in plant developmental rate . To have a better understanding of which microbes are associated with the various developmental stages , we formed new RF models for predicting plant developmental stage rather than plant age . Development in plants is a continuous process marked by numerous stage-specific morphological features . We monitored the developmental stage of the genotypes throughout the season , assigning a numerical value from 1–27 , with a value of 1 corresponding to a recently germinated seeding and 27 corresponding to a senescent plant ( S11 Table ) . These values reflect a numerical value given to the stages used in Counce et al . to describe rice development [24] . Panicle initiation corresponded to a value of 18; thus , a plant that has entered the reproductive phase has a value of 18 or higher , and a plant in the vegetative phase has a value of 17 or lower . We followed the same approach as previously mentioned to develop the RF models . Briefly , we trained full RF models in each compartment where we regressed the full dataset of OTUs against the developmental stage number for a training set of samples . From these full models , we sequentially removed OTUs of lower importance while performing 10-fold cross validation . We found that the models were near peak accuracy when including 54 of the most important OTUs for each compartment . With these 54 most-important OTUs , we developed sparse RF models that model the microbiota as a function of plant developmental stage ( Fig 6D ) . When the predicted values of plant developmental stage were plotted as a function of the plant’s chronological age , we found that the predictions accurately matched the developmental progressions that we witnessed in the field ( see Fig 6A ) . Variation in developmental stage prediction slopes was significantly affected by plant genotype in each compartment ( Rhizosphere P = 8 . 1 × 10−8 , Rhizoplane P = 3 . 3 × 10−5 , Endosphere P = 2 . 3 × 10−2 ) . These results indicate that OTUs within the microbiota in each compartment can be used to predict plant developmental stage , even when developmental progression rates differ between rice cultivars . We again classified the OTUs included in each stage-predicting RF model as early/late/complex colonizers using a linear model to detect whether the OTUs had significantly increasing , significantly decreasing , or complex distributions throughout the season . The phylogenetic classification of early/late colonizing OTUs mimicked those of the age-discriminant models ( S2 and S8 Figs ) . Of the early colonizing OTUs over the season , 3 were shared between the compartment-specific models , while 9 of the late colonizing OTUs were shared . We next sought to address whether OTUs included in the models peak in abundance at different times in the season for the different cultivars . To do this , we calculated the average abundance of each OTU included in each model at each time point for each cultivar . For every compartment , we found that the early colonizing OTUs peaked later in Nipponbare than in the other cultivars . When taking the total abundance of increasing/decreasing OTUs into account , we found that the early colonizing OTUs tended to persist for a longer period of time in Nipponbare than in the other varieties ( Fig 6E ) . Similarly , the OTUs classified as late colonizers took longer to establish in Nipponbare than in the other varieties . This pattern was consistent across each compartment . We note that the developmental stage sparse RF models used fewer OTUs than the plant age models to make accurate predictions . Together , these results indicate that both plant age and developmental stage are important drivers of the root microbiota .
Previous studies conducted under greenhouse conditions have reported that compositions of plant microbiota are dynamic during plant growth [8 , 12 , 28 , 29] . Nevertheless , because of the limited extent of both temporal and spatial sampling in these studies , e . g . , absence of data from the endosphere compartment [28 , 29] , a comprehensive picture of the changes in the microbiome during the life cycle , as has been accomplished for humans , has yet to be formulated for plants . Additionally , the difference between greenhouse and field conditions is important given that nutrient dynamics and plant physiology vary under the 2 settings [30 , 31] , and it has been demonstrated that plant microbiota are also altered by field conditions compared to greenhouse conditions [8] . Similarly , the importance of ecological factors driving community assembly is overestimated if the experimental observations are constrained to one soil type or season [32] . Here we demonstrate spatiotemporal shifts in microbiota composition during the life cycle that are consistent across multiple years of cultivation in the same field ( Fig 1A–1C ) . Moreover , although we found large differences at the OTU level between the 2 tested field sites in California and Arkansas , at the phylum level we found that there were remarkable similarities in spatiotemporal profiles of microbiota abundance between the field sites despite large geographic distances , climatic differences , and cultivation practices ( Fig 2A ) . In general , there were more phyla that were decreasing in relative abundance in the endosphere over the life cycle of the rice plants , while fewer phyla were increasing in relative abundance . In the rhizosphere , we found that more phyla were increasing rather than decreasing over the life cycle ( Fig 2C ) . These data reinforce the exclusionary role of the endosphere compartment compared to the rhizosphere compartment as observed in other studies [6–8] , but indicate that exclusion of microbes in the endosphere is age sensitive . Betaproteobacteria and Deltaproteobacteria were the dominant classes enriched in the root endosphere compared to the bulk soil . These 2 classes showed opposing patterns of abundance over the season: Betaproteobacteria was largely decreasing in abundance , while Deltaproteobacteria increased ( Fig 2C ) . Our high-resolution sampling scheme allowed us to deeply characterize the developmental patterns of microbiota over the growing season . The rhizoplane and endosphere microbiota progressed over the course of the first 7 to 8 weeks after germination but then stabilized in composition thereafter ( Fig 1D ) . This was consistent across the 2 years of sampling from the California field as well as for samples collected from the Arkansas field . By employing a machine learning approach , we were able to model rice plant age as a function of fluctuating relative abundances of OTUs ( Fig 3 ) . Using the RF algorithm , we identified OTUs in the rhizosphere and endosphere compartments that were discriminant of plant age . Using these sets of OTUs , we were able to accurately predict the plant age of samples gathered from the California and Arkansas field sites . We were also able to use these models to accurately predict the age of rice plants grown in the California site in the 2015 season , despite the models not being trained on this data . These results indicate that groups of microbes proliferate predictably between field sites and between years . The sets of OTUs included in these models both increase and decrease in relative abundance over the course of the season ( Fig 3B ) . The early and late colonizing OTUs were distinct at the phylum and order levels , suggesting that functional capabilities encoded by these microbes change throughout the season . These results show that OTUs conserved between 2 diverse field sites can be used infer the age of rice plants . It is unknown whether our models developed for the rhizosphere and endosphere compartments are generalizable to rice plants grown in other regions of the world: climate , geography , and cultivation practices are all factors that contribute to microbiome structure and would likely affect the age-predicting models . The rice rhizosphere microbiota shows similarity at lower-resolution taxonomic levels [33] , even when grown on different continents , so it is likely that RF models built at the order or family level may allow generalization in predictability across continental scales . Diverse plant species host divergent microbiota assemblages , even at the phylum level [34]; thus , it is unlikely that our models could be accurately applied to predicting the age of a set of genetically diverse plant species . To test our hypothesis that environmental variables could affect the accuracy of age prediction , we used the age-predicting models to estimate the age of plants experiencing drought compared to those experiencing well-watered conditions ( Fig 4 ) . We found that the predicted age of drought-stressed plants using the endosphere models was significantly younger compared to well-watered plants , indicating that the water-deprived plants were hosting a developmentally immature endosphere microbiota ( Fig 4A ) . This finding is intriguing given that rice plants experiencing drought stress during the vegetative stage typically have delayed flowering times [35–37] . These data support a model in which an arrest in host plant development caused by exposure to drought stress in turn impacts normal development of the endospheric microbiota . More studies will need to be conducted to confirm this hypothesis . Nevertheless , these results indicate that the models described here can be used to create a baseline for normal microbiota development and to test how environmental or biological perturbations may affect the maturation process . It is unknown how the OTUs included in each model relate to the overall health of the rice plants and assembly of a healthy microbiota . Age-discriminant microbes in the human gut have been found to be important for the health of the host [38–40] . In malnourished Bangladeshi children , the gut microbiota was shown to be underdeveloped compared to healthy children [38] . It was found that age-discriminant OTUs identified through RF models from the healthy human gut microbiota prevented growth defects in mice harboring microbiota from malnourished children [39] . It would be of interest to investigate whether isolated age-discriminant OTUs from plants could also play a role in recovery from biotic or abiotic stresses that delay developmental progression . Over the course of the season , we showed that the rhizosphere and endosphere communities between field sites grow more similar , stabilizing around 8 to 9 weeks after germination ( Fig 5A ) . Similarly , there were more age-discriminant OTUs in the rhizosphere and endosphere RF models that showed increasing trends in relative abundance over the course of the growing season ( Fig 3B ) . These results suggested that there was less conservation between the field sites for early colonizing OTUs . Our data suggests that site-specific OTUs were significantly greater at the beginning of the season in both the rhizosphere and endosphere but were diminished at later time points in the growing season ( Fig 5B ) . This effect was greater in the endosphere than in the rhizosphere , presumably because the rhizosphere is host to both microbes affected by plant processes as well as microbes from the soil biota that are unaffected by signals originating from the host . The increased conservation of the late-emerging microbiota may be due to plant selectivity , while the early colonizing microbiota may be due to opportunistic colonization of plant tissue by the soil microbiota . Nevertheless , through our RF approach , we were able to identify specific early colonizing microbes that are shared between the 2 field sites , and these OTUs were almost unanimously enriched in their respective rhizocompartments compared to bulk soil ( S3 Fig ) , suggesting active or passive recruitment of the conserved early colonizing microbiota . In humans , the early colonizing gut microbiome has been implicated in educating the immune system [41 , 42] . In plants , it is unclear whether the conserved early colonizing microbiota plays a role in conditioning the activity of the plant innate immune system . A recent study in maize using a synthetic community found that certain bacteria , when omitted from the community , drastically disrupted normal microbiota assembly in roots and that disruption of the normal microbiota assembly led to greater susceptibility to fungal pathogens [43] . It may be possible that a portion of the conserved early colonizing taxa may be acting as such keystone species; thus , it is of interest to characterize isolated members from the early colonizing age-discriminant OTUs to understand whether they have a role in immune system function and microbiota assembly . Previous studies have suggested that root microbiota are assembled early in the plant life cycle and are subsequently insensitive to the developmental status of the host plant [28 , 29] . Strong support for this model comes from the absence of significant differences in microbiota structure between an early flowering A . alpina mutant ( pep1 ) and a nonflowering wild-type plant at the same age , implying that developmental stage is not responsible for microbiota compositional changes over time [12] . While this conclusion may hold across other perennial plant species , our results suggest that this conclusion might not be generalizable to all plants . By growing rice varieties with differing developmental trajectories , we were able to quantify the effect of plant development on the root-associated microbiota . We observed that there was a gradient in microbiota maturation across the second PCo in the rhizosphere and endosphere ( Fig 6C ) . By using a RF regression , we were able to identity sets of microbes in each rhizocompartment that are able to distinguish the samples by developmental stage ( Fig 6D ) . These results suggest that root microbiota are affected by both plant age and developmental stage and these effects influence different sets of microbes . There are several likely factors that could cause the observed differences between our study and Dombrowski et al . A . alpina is a wild perennial plant , and rice is a domesticated annual species . One hallmark trait of cereal domestication is the selection for varieties with larger sink sizes in the seed [44] . For instance , in wild species , source carbohydrates are more evenly distributed to various sink tissues in the plant than in domesticated cereals , in which the seeds are a predominant sink [44 , 45] . The discrepancy between sink-source dynamics in these 2 host plants could at least partially explain why our study detected shifts in the microbiota due to developmental stage , while it was undetected in A . alpina . In rice , accumulation and storage of carbohydrates in the stems occur until the onset of reproduction , after which internal signals reprogram the host plant to repartition carbohydrates to the developing panicle and to the filling grain [46] . These host signals , along with the shifting nutritional needs at this stage , could explain the stabilization in the microbiota that occurs at the onset of reproduction . Previous studies have found changes in the root-associated microbiota during the growth of the host plant [12 , 28 , 29 , 47] , but the patterns of compositional changes have not been elucidated . The dense temporal and spatial sampling undertaken here reveals a clear pattern of root microbiome dynamics over the life cycle of an annual plant . The data , from multiple growing seasons and sites , are consistent with a 2-stage model that can be summarized as follows: immediately after germination , a community of early colonizing microbes becomes established inside and in the vicinity of the roots , representing a potential “juvenile microbiome . ” It is unclear to what extent the host actively drives this process , as many of the early acquired microbes are site-specific and may represent opportunistic colonization by a subset of the soil microbiota . Nevertheless , a set of early colonizing microbes conserved between field sites were identified using our RF models and may represent consortia that are responsive to cues from the host plant . At around the time of entering the reproductive phase , a later colonizing set of microbes displaces the early colonizing microbes and then remains stable throughout the remainder of the life cycle of the host plant , potentially an “adult microbiome . ” The rate of these transitions is dependent upon the rate of developmental progression , which in turn is genotype dependent . The microbes colonizing later in development were more conserved between the field sites , suggesting a greater influence of factors from the host . Root exudate composition varies significantly over the life cycle of Arabidopsis and tomato [48 , 49] , and factors in exudates might underlie some of the shifts observed during the vegetative to reproductive transition . A more pronounced stabilization effect of the adult microbiome is observed in the rhizoplane and endosphere as compared to the rhizosphere ( Fig 1D ) , raising the possibility of a more direct control by the host . The activity of the plant immune system can exhibit differences during plant growth , as demonstrated by the decreased susceptibility of some cereal crops to various pathogens during the adult phase [50–52] . Changes in the plant immune system are likely to play a larger role in affecting dynamics of the rhizoplane and endosphere microbiota than the rhizosphere , as microbes on the root surface and interior are in physical contact with the host plant cells . In 1904 , Lorenz Hiltner coined the term “rhizosphere” when he proposed that photosynthates exuded into the soil by plants likely alter the composition and abundance of microorganisms surrounding the roots [53] . Hiltner proposed that the communities of microorganisms in this zone impact plant nutrition and health [53 , 54] and that these communities may be affected by the stage of plant growth [55] . The role of rhizosphere microorganisms in plant nutrition and disease resistance has been confirmed [4 , 5] . Our results provide support for Hiltner’s hypothesis that the stage of plant growth is an important determinant of root-associated microbial composition . Manipulation of the soil microbiota using bacterial isolates to increase crop yield and resistance to pathogens has been proposed as an enhancement to conventional plant breeding [56] . Using monoassociation assays , arrays of plant growth-promoting or disease-suppressing bacteria have been identified [57 , 58] . However , a major problem for translation to field conditions has been the inconsistent persistence of plant beneficial bacteria once inoculated into complex soil microbial communities [59–61] . Our findings that the composition of the root-associated microbiota are sensitive to plant age and developmental stages suggest that a good match with plant age might be a factor for persistence in the face of competing soil microbes . Thus , a plant growth-promoting bacterium that normally establishes during the reproductive phase might not be successful if supplied with seeds . The analysis provided here indicates that consideration of age appropriateness of a microbial inoculum could be utilized to enhance the efficacy of beneficial microbes for agricultural applications .
We collected samples from a commercially cultivated rice field in Arbuckle , California . In both seasons , the field was water seeded in early May . Water seeding is conducted by first preparing the field by removing winter vegetation , disking the soil to produce baseball-sized clods , applying nutrients , and then the field is flooded . The rice seeds are soaked in water overnight and then loaded into an aircraft , where they are then applied aerially to the field in an even density . For this particular field , the farmer grew the M206 cultivar , a medium-grained California variety that has an average heading date of 86 days after germination . We began sampling plants 7 days after the fields were seeded , which coincided with the emergence of the seminal roots . In the 2015 and 2016 seasons , we restricted our area of sampling to a 150×150 foot section of the field . Within this section , we sampled plants from random locations . Our sampling occurred as previously described [8] . Using gloved hands , we would scoop under the root mass to separate the plant from the ground . Grabbing the plant by the base of the stem , we would then shake the plant to remove loosely attached soil from the roots . We would then place the roots with tightly adhering soil into 50 mL Falcon tubes with 15 mL autoclaved phosphate buffered saline ( PBS ) solution . We then brought the samples back to the laboratory at UC Davis for subsequent processing . In the Arkansas 2016 season , we grew an O . sativa sgp . tropical japonica variety , Sabine , and a hybrid variety , CLXL745 , in a split-plot design in a privately owned agricultural field near Jonesboro , Arkansas . Each plot was isolated from other plots via berms , and each plot had its own source of water and drainage ( see design , S9 Fig ) . The roots of the plants were collected over the growing season as described above , placed into 50 mL Falcon tubes , and shipped overnight on ice to UC Davis . At UC Davis , the rhizosphere and endosphere compartments were separated ( see below ) and stored at −80 °C until DNA could be extracted . Due to the extra processing steps required to collect rhizoplane samples and the distance these root samples were shipped , we did not sequence the rhizoplane samples from the Arkansas field trial . We grew closely related O . sativa sgp . temperate japonica cultivars in the same field in Arbuckle , California , for the 2016 season in a Latin square design . Kitaake is a variety typically used under laboratory settings because of its relatively fast life cycle . M206 and M401 are varieties adapted to growing in California with medium times to panicle initiation . M401 , however , requires a longer period of time for flowering . Nipponbare is a variety with a longer time to panicle initiation and flowering . Briefly , we designed a fully randomized block design to grow these varieties in 1×1 m plots in quadruplicate ( see design , S10 Fig ) . We left 0 . 5-m-wide walking lanes between each plot , which subsequently allowed us to sample bulk soil throughout the course of the growing season . To be consistent with the previous years’ methods , we water seeded these varieties . This entailed soaking the seeds in a 2% bleach solution for 4 hours to remove the risk of the field being contaminated with Fusarium moniliforme , a seed-borne fungal pathogen that causes the disease Bakanae . The seeds were then washed 3 times with sterile water and soaked overnight . We then hand seeded each plot at a similar density as what the farmer had applied in previous seasons . At each time point , the plants were sampled as previously described and transported to the lab for further processing . The plants during this season were dissected to ascribe various developmental stages in accordance with Counce et al . [24] . Because the rice plants were water seeded , there was a high chance that each genotype’s plot could be contaminated by seeds drifting in from another plot or seeds that the field manager had planted . The genotypes were confirmed by amplifying SSR marker RM144 [62] using the endosphere DNA samples collected from the plants . From this analysis , we excluded the first week of Nipponbare samples from the analysis because of all of the samples being contaminated by M206 . Similarly , we removed other samples from the analysis in which the SSR marker genotyping did not match the genotype of the plot in the field . All soil chemical analyses were conducted at the University of California , Davis Analytical Laboratory ( S2 Table ) . Soil classifications were obtained from the US Department of Agriculture web soil survey ( https://websoilsurvey . sc . egov . usda . gov/App/HomePage . htm ) . In each instance of sampling roots , we collected material from the first inch of roots just below the root-shoot junction . The root-associated compartments were separated as previously described [8] . The roots with soil attached were vortexed in PBS solution , and 500 uL of the resulting slurry was used for DNA extraction . The roots were cyclically washed in fresh PBS solution until no soil particles were visible in the solution . The roots were then placed into fresh PBS and sonicated for 30 seconds to remove the surface cell layer of the roots . The resulting slurry was centrifuged down to concentrate the biomass and then used as the rhizoplane fraction for DNA extraction . The remaining roots were sonicated twice more , refreshing the PBS solution at each stage , and then ground in a bead beater . The resulting solution was used for DNA extraction as the endosphere fraction . All DNA extraction was performed using the MoBio Powersoil DNA isolation kits . We amplified the V4 region of the 16S ribosomal RNA gene using the universal 515F and 806R PCR primers . Both our forward and reverse primers contained 12 base-pair barcodes , thus allowing us to multiplex our sequencing libraries at over 150 libraries per sequencing run [8] . Each library was accompanied by a negative PCR control to ensure that the reagents were free of contaminant DNA . We purified the PCR products using AMPure beads to remove unused PCR reagents and resulting primer dimers . After purification , we quantified the concentration of our libraries using a Qubit machine . Our libraries were then pooled into equal concentrations into a single library and concentrated using AMPure beads . The pooled library then went through a final gel purification to remove any remaining unwanted PCR products . Pooled libraries were sequenced using the Illumina MiSeq machine with 250×250 paired-end chemistry . The resulting sequences were demultiplexed using the barcode sequences by in house Python scripts . For the case of the drought data , we downloaded the sequences for the Short Read Archive project number PRJNA386367 . The sequences were quality filtered and then assembled into full contigs using the PandaSeq software [63] . Any sequences containing ambiguous bases or having a length of over 275 were discarded from the analysis . The high-quality sequences were clustered into OTUs using the Ninja-OPS pipeline [64] against a “concatesome” composed of the Greengenes 97% OTU representative sequence database ( version 13_8 ) [65] and then assembled into an OTU table . This OTU table was filtered to remove plastidial and mitochondrial OTUs . To account for sequencing depth differences between the samples , each OTU in each sample was divided by the total sequencing depth for the respective sample and multiplied by 1 , 000 , resulting in a relative abundance in units of per mille . OTUs that occurred in less than 5% of the samples were filtered from the table ( S11 Fig ) . This process reduced the total OTU count from 24 , 048 to 10 , 893 taxa . The resulting 10 , 893 taxa were used for the analysis . All statistical analyses were conducted using R version 3 . 1 [66] . Unless otherwise noted , we determined statistical significance at ɑ = 0 . 05 and , where appropriate , corrected for multiple hypothesis testing using the Bonferroni method . Shannon diversity was calculated using the diversity ( ) function , unconstrained PCoA analyses were conducted using the Vegan capscale ( ) function by specifying an intercept-only model ( R Code: capscale ( log2 ( RA ) ~ 1 ) , and PERMANOVA was conducted using the adonis ( ) function from the Vegan package [67] . Linear models were run using the lm ( ) function , and ANOVA was run using the aov ( ) function from the Stats package [66] . Beta regression was performed using the BetaReg package [68] . RF models were generated and analyzed using the randomForest package [69] . Differential OTU abundance was performed using exact tests in the package edgeR [23] . All graphs and plots were generated using the ggplot2 package [70] . R notebooks for the full analyses can be found at https://github . com/bulksoil/LifeCycleManuscript . To model plant age as a function of microbiota composition , we began by developing compartment-specific full RF models for both the endosphere and rhizosphere samples by regressing the relative abundance of all OTUs against the age of the plants from which the samples were collected . For our training data , we selected samples from the California 2014 and Arkansas 2016 data . Half of the samples from each time point and rhizosphere and endosphere compartments were randomly sampled for the training set . From these full models , we were able to rank individual OTUs by their importance in contributing to the accuracy of age prediction by the models . This process is performed by permuting the relative abundance levels for an OTU and calculating the increase in mean squared error of the model . OTUs whose relative abundances when permuted yield increased errors in the model are considered to be important to the accuracy of the model . This step was performed using the importance ( ) command from the randomForest R package . Because not every OTU included in the full RF models will contribute to the accuracy of the models , we next performed 10-fold cross validation while simultaneously removing less important OTUs to evaluate model performance as a function of inclusion of the top age-discriminant OTUs using the rfcv ( ) function in the randomForest R package . We found that there was a minimal increase in accuracy when including more than 85 of the most important OTUs ( S12 Fig ) . The top 85 OTUs from the full RF model of each compartment were then used as inputs for sparse RF models for each compartment with no further parameterization . All raw sequences derived from this project were submitted into the Short Read Archive of NCBI and can be found under the BioProject accession number PRJNA392701 . OTU tables , metadata files , taxonomy files , and R data files are available from the Dryad Digital Repository: https://doi . org/10 . 5061/dryad . 7q7k1 [71] . | Plant roots are colonized by complex communities of bacterial and archaeal microbiota from the soil , with the potential to affect plant nutrition and fitness . Although root-associated microbes are known to have the potential to be utilized to promote crop productivity , their exploitation has been hindered by a lack of understanding of the compositional dynamics of these communities . Here we investigate temporal changes in the root-associated bacterial and archaeal communities throughout the plant life cycle in field-grown rice over multiple seasons and locations . Our results indicate that root microbiota composition varies with both chronological age and the developmental stage of the plants . We find that a major compositional shift correlates with the transition to reproductive growth , suggestive of distinct root microbiota associations for the juvenile and adult plant phases . The results from this study highlight dynamic relationships between plant growth and associated microbiota that should be considered in strategies for the successful manipulation of microbial communities to enhance crop performance . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"ecology",
"and",
"environmental",
"sciences",
"microbiome",
"microbiology",
"seasons",
"developmental",
"biology",
"plant",
"science",
"rice",
"experimental",
"organism",
"systems",
"plant",
"ecology",
"plants",
"flowering",
"plants",
"microbial",
"genomics",
"research",... | 2018 | Compositional shifts in root-associated bacterial and archaeal microbiota track the plant life cycle in field-grown rice |
Extraintestinal pathogenic Escherichia coli colonize the human gut and can spread to other body sites to induce diseases such as urinary tract infections , sepsis , and meningitis . A complete understanding of the infection process is hindered by both the inherent genetic diversity of E . coli and the large number of unstudied genes . Here , we focus on the uncharacterized gene rqlI , which our lab recently uncovered in a Tn-seq screen for bacterial genes required within a zebrafish model of infection . We demonstrate that the ΔrqlI mutant experiences a growth defect and increased DNA stress in low oxygen conditions . In a genetic screen for suppressor mutations in the Δrql strain , we found that the shortcomings of the Δrql mutant are attributable to the activity of RqlH , which is known in other bacteria to be a helicase of the RecQ family that contains a phosphoribosyltransferase ( PRTase ) domain . Disruption of rqlH rescues the ΔrqlI strain in both in vivo and in vitro assays , while the expression of RqlH alone activates the SOS response coincident with bacterial filamentation , heightened sensitivity to DNA damage , and an increased mutation rate . The analysis of truncation mutants indicates that , in the absence of RqlI , RqlH toxicity is due to its PRTase domain . Complementary studies demonstrate that the toxicity of RqlH is modulated in a context-dependent fashion by overlapping domains within RqlI . This regulation is seemingly direct , given that the two proteins physically interact and form an operon . Interestingly , RqlH and RqlI orthologs are encoded by a diverse group of bacteria , but in many of these microbes , and especially in Gram-positive organisms , rqlH is found in the absence of rqlI . In total , this work shows that RqlH and RqlI can act in a strain-specific fashion akin to a toxin-antitoxin system in which toxicity is mediated by an atypical helicase-associated PRTase domain .
Extraintestinal pathogenic Escherichia coli ( ExPEC ) comprise a group of E . coli strains that harmlessly reside within the gastrointestinal tract , but that are capable of infecting extraintestinal niches such as the urinary tract , bloodstream , and meninges [1 , 2] . Urinary tract infections ( UTIs ) alone constitute a major global economic burden due to the high incidence rate of UTIs . Indeed , more than half of all women will experience a UTI during their lifetime [3] , with bladder infections being responsible for more than $2 billion in medical care costs in the United States annually [4] . Additionally , E . coli is the leading cause of bacteremia [5] , and is especially problematic in high risk , immunocompromised populations such as the elderly [6 , 7] and cancer patients undergoing chemotherapy [8] . E . coli is also the leading cause of death in early-onset neonatal sepsis [9] . The burden imposed by ExPEC infection is expanding as the prevalence of antibiotic resistant strains increases [10] , highlighting the need for a deeper understanding of ExPEC-mediated pathogenesis . A major hurdle in deciphering the ExPEC life cycle is the genetic heterogeneity inherent in these strains . Each ExPEC isolate contains ~5 , 000 genes , of which about half comprise a core genome that is shared with most other E . coli strains [11] . The genes outside of the core genome are much less conserved and account for the bulk of the genomic heterogeneity observed among ExPEC isolates . This flexible genome is in large part a consequence of horizontal gene transfer among ExPEC strains and other bacteria . The ExPEC pan-genome , which includes all distinct genes found within ExPEC isolates , contains at least 14 , 877 genes [11] . Due to the dynamic nature of the pan-genome , this estimate is likely to increase as more isolates are sequenced [12] . Because of the genetic differences from one ExPEC isolate to another it is difficult to define a specific set of genes that are universally required for ExPEC fitness . Complicating matters further is the fact that many ExPEC-associated genes are functionally undefined . One method for dealing with the large number of uncharacterized genes is to employ high-throughput analyses that can assay the role of many genes in a given condition simultaneously . One especially useful technique is Tn-seq , in which a pool of transposon mutants is placed under a selective pressure , and the quantity of individual mutant variants before and after the selection are enumerated by next-generation sequencing [13] . Despite their utility , these types of analyses often result in lists of genes that are required for fitness under a given condition without yielding clear insight into the actual functions of individual genes . For example , in a recent Tn-seq screen for factors that affect ExPEC fitness in a zebrafish infection model , many of the 981 genetic elements that were found to be important for host colonization had never been studied [14] . Annotations for 152 of the 981 loci ( 15 . 5% ) that were identified as required for zebrafish colonization include the words “hypothetical” or “putative” . In addition , given the limitations of the current pipelines used to annotate genes , many more of the identified loci are likely misannotated [15] . Thus , understanding why a gene is important for fitness under a given condition can be complicated . This problem can be partially remedied by comparing overlapping datasets from high-throughput experiments carried out in various defined conditions [16 , 17] . This strategy can help uncover roles for previously undefined genes , but additional detailed analysis is still required to validate the high-throughput data and to unambiguously specify gene function . In our recent Tn-seq screen using the ExPEC cystitis isolate F11 , we identified the hypothetical gene EcF11_3933 as an important mediator of pathogen fitness during both localized and systemic infections within the surrogate zebrafish host [14] . This gene , which we propose to rename RecQ-Like Helicase Interactor ( rqlI ) for reasons that will become clear , was also critical to the ability of F11 to colonize the mouse urinary tract and for pathogen survival in a mouse model of sepsis . Considering the importance of rqlI to pathogen survival in diverse hosts and host environments , we set out to define the specific function of RqlI . Here we report that RqlI is especially critical for pathogen growth under low oxygen conditions , as found within many host niches . In the absence of RqlI , the bacteria have higher mutation rates , increased activation of the SOS stress response system , and decreased growth . These effects are remedied by deletion of rqlH , a gene immediately upstream of rqlI that encodes RecQ-like Helicase ( RqlH ) . Deletion of rqlH also rescues the in vivo survival defects observed with the ΔrqlI mutant in several mouse models . Genetic and biochemical analyses indicate that a function of RqlI within ExPEC is to limit the genotoxic effects of RqlH , which we mapped to an unusual phosphoribosyltransferase ( PRTase ) domain localized at the C-terminus of RqlH . Interestingly , we find that RqlH and RqlI homologs are encoded by a diverse collection of bacterial species , though the rqlH and rqlI genes are not always paired as in F11 . These observations suggest that RqlI can have both RqlH-dependent and RqlH-independent functions , and that these two proteins have important roles in the life cycles of a wide range of bacteria .
An initial analysis of the RqlI protein revealed the presence of three putative domains ( Fig 1A ) . First , RqlI is predicted to contain a molybdenum cofactor ( MoCo ) binding domain , which is often found in proteins that facilitate metabolism in low oxygen environments [18] . For example , MoCo facilitates redox reactions critical for using alternative terminal electron acceptors when oxygen is scarce . Second , overlapping the MoCo binding domain is a region with homology to the DNA-protecting protein DprA , which has been shown to be important for the uptake of exogenous DNA in naturally competent bacteria [19] . Third , at the C-terminus of RqlI is a helix-turn-helix motif predicted to bind DNA . Given the involvement of MoCo in anaerobic respiration [18] , and given the putative MoCo-binding domain contained within RqlI , we disrupted the rqlI gene in the reference ExPEC strain F11 and tested the ability of the mutant to multiply in varying oxygen levels . When grown for 24 hours in a modified M9 minimal media exposed to atmospheric oxygen ( ~20% O2 ) , F11ΔrqlI exhibits a slight growth defect in comparison to the wild type ( WT ) strain ( Fig 1B ) . However , in microaerobic culture conditions ( 6–12% O2 ) growth of the ΔrqlI mutant was markedly worse than WT ( Fig 1B ) . This defect was rescued by complementation with the RqlI expression plasmid pCWR16 . Furthermore , we found that the exponential phase generation time of the ΔrqlI mutant was very similar to WT cells under aerobic conditions , but was 3-fold greater than WT bacteria in a microaerobic environment ( Fig 1C ) . Under microaerobic conditions , deletion of rqlI also limited the growth of two other reference ExPEC strains , CFT073 and 536 ( Fig 1D ) . We examined the general morphology of F11ΔrqlI using light microscopy and noticed that the mutant cells were often longer ( more filamentous ) than WT bacteria ( Fig 2A–2D ) . The elongated ΔrqlI mutant cells were especially abundant , and reached greater lengths , under microaerobic or anaerobic conditions ( Fig 2B–2D ) . One reason for bacterial cell length to increase is the induction of the SOS response , which is initiated when DNA damage is sensed [20] . The SOS response involves the upregulation of sulA expression , which functions to temporarily inhibit cell division while DNA repair occurs . Using a PsulA-GFP fluorescent reporter construct , we found that F11ΔrqlI expressed markedly higher levels of sulA than the WT strain in aerobic , microaerobic , and anaerobic conditions ( Fig 2E ) . The differences in sulA expression levels between WT and the ΔrqlI mutant grew more pronounced with decreasing amounts of oxygen in the culture . In control samples , sulA expression levels in the WT and ΔrqlI strains were similarly increased upon treatment with the DNA-damaging agent mitomycin C . These fluorescent measurements were confirmed by western blots probed with GFP-specific antibody ( S1 Fig ) . In total , these data demonstrate that in the absence of RqlI there is increased activation of the SOS response , coincident with an oxygen-sensitive decrease in bacterial replication . In order to understand why sulA expression is increased in the ΔrqlI strain , a screen was designed to search for factors that promote SOS induction in the absence of RqlI ( Fig 3A ) . An F11ΔrqlIΔlacZY strain was transformed with plasmid pCWR2 containing the lacZ gene driven by the sulA promoter , providing a convenient readout of sulA expression levels on tetrazolium lactose plates . On these plates , high LacZ expression—reflecting high sulA promoter activity—results in colonies with pink to white color , while low LacZ activity due to low sulA promoter activity results in a red colony color . F11ΔrqlIΔlacZY/pCWR2 was randomly mutagenized using the mariner transposon from pSAM_Ec [14] , and then plated onto selective tetrazolium lactose plates . Approximately 20 , 000 colonies were screened for darker colony color , indicative of a decreased SOS response . Transposon insertion locations were mapped in potential suppressor mutants , and six independent insertions were identified within rqlH ( Fig 3B ) . This gene encodes a homologue of the Mycobacterium smegmatis RqlH protein ( MSMEG_5935 ) , a RecQ-Like Helicase that is known to unwind dsDNA in a 3’ to 5’ orientation [21] . The F11 and M . smegmatis RqlH proteins align along their entire lengths , with 41 . 2% identity ( 55 . 7% similarity ) overall ( S2 Fig ) . Interestingly—and perhaps , in retrospect , not surprisingly—the F11 rqlH gene is also found adjacent to rqlI on the genome ( Fig 4A ) . To confirm the results of the screen , rqlH and rqlI were deleted from F11 , separately and in combination , and PsulA-GFP expression levels were measured under low oxygen conditions ( Fig 3C ) . As before , GFP expression levels in the ΔrqlI strain were higher than in WT F11 , whereas strains lacking only rqlH or both rqlH and rqlI had GFP levels equal to that of WT . These data indicate that in the absence of RqlI , RqlH can trigger the SOS response , resulting in increased sulA expression . To test whether the reduction in growth of F11ΔrqlI under microaerobic conditions could also be attributed to RqlH activity , WT F11 and the ΔrqlI , ΔrqlH , and ΔrqlHI strains were grown in low oxygen cultures . Relative to the WT strain , only the ΔrqlI mutant grew poorly in these assays ( Fig 3D ) . In addition , we found that overexpression of RqlH caused a marked reduction in bacterial growth under both aerobic and microaerobic conditions ( S3 Fig ) . This effect was seen regardless of genetic background , but was more pronounced in strains that lacked RqlI . Together , these results indicate that RqlH activity in the absence of RqlI both induces the SOS response and inhibits bacterial growth . As the SOS response is activated by DNA damage and is responsible for initiating DNA repair programs , we examined the abilities of the rqlI and rqlH mutants to deal with exogenous DNA damage . WT F11 and mutant derivatives were treated with the DNA crosslinking agent mitomycin C and surviving bacterial cells were quantified . The ΔrqlI mutant had increased sensitivity to mitomycin C , a defect that was again rescued in the ΔrqlHI double mutant strain ( Fig 3E ) and by complementation with the RqlI expression plasmid pCWR16 ( Fig 3F ) . A similar trend was observed when the cells were treated with UV light . The ΔrqlI mutant was more sensitive to both 25 and 50 J/m2 of UV treatment , although this effect was only statistically significant at 25 J/m2 ( Fig 3G ) . In contrast , the ΔrqlH and ΔrqlHI mutants survived UV exposure at levels similar to the WT cells . These data indicate that the ΔrqlI mutant is more sensitive to DNA damage in a manner that is dependent on RqlH . To determine if the increased sensitivity of the ΔrqlI mutant to exogenous DNA damaging agents correlates with a decreased ability to deal with endogenous DNA damage , spontaneous mutation rates that resulted in nalidixic acid resistance were quantified using fluctuation assays [22] . WT F11 and the ΔrqlI , ΔrqlH and ΔrqlHI strains all exhibited low mutation rates of about 3 mutations per 1010 cells per generation when grown under aerobic conditions ( Fig 3H ) . Growth under microaerobic conditions did not change the mutation rates for WT F11 , F11ΔrqlH , or F11ΔrqlHI , but resulted in nearly an 80-fold increase in the mutation rate of the ΔrqlI mutant . These data suggest that in the absence of RqlI , RqlH makes ExPEC more sensitive to DNA damage and the accumulation of spontaneous mutations . Since the rqlH and rqlI genes are located immediately adjacent to each other on the genome , but are separated from upstream and downstream genes , it is likely that they form a two-gene operon ( Fig 4A ) . To test this idea , we performed RT-PCR using primer sets spanning the rqlH and rqlI loci , as well as one set internal to rqlI , as depicted in Fig 4A . Products expected if rqlH and rqlI are co-transcribed as part of the same transcript were obtained from F11 cDNA , but were not seen in control reactions that lacked reverse transcriptase ( Fig 4B ) . These results indicate that rqlH and rqlI form an operon . Given that the ΔrqlI mutant defect is exacerbated by microaerobic conditions , we tested whether rqlHI transcription is modulated by oxygen levels . Using primers that are specific for rqlH , for rqlI , or that span rqlHI , we were unable to detect any changes in rqlHI expression in microaerobic growth as compared to aerobic growth ( Fig 4C ) . In contrast , we observed that transcription of frdA , which is sensitive to oxygen levels [23] , was significantly upregulated in F11 grown under microaerobic conditions . RqlH contains an N-terminal helicase domain as well as a C-terminal phosphoribosyltransferase ( PRTase ) domain that is unique among known helicases [21] . The functional importance of the PRTase domain is unclear , though it is dispensable for the ATPase and helicase activities of RqlH in M . smegmatis [21] . To define domain ( s ) within RqlH that are responsible for the genotoxic effects observed in the absence of RqlI , we expressed epitope-tagged RqlH and several mutant variants in the K-12 strain MG1655 , which lacks both RqlH and RqlI . Expression of each RqlH variant was confirmed by western blot ( S4 Fig ) . The recombinant MG1655 strains were diluted into modified M9 minimal media and grown for 6 h with aeration , at which point the optical density ( OD600 ) of each culture and the numbers of viable bacteria ( CFU/mL ) were measured . Expression of the WT RqlH significantly reduced both the growth of MG1655 and the numbers of CFUs recovered ( Figs 5 and S5 ) . RqlH expression also stimulated the formation of filamentous bacteria and the induction of the SOS response , as determined by microscopy and by use of the SOS reporter strain MG1655 attTn7::PsulA-GFP ( Figs 5 and S6 ) . As seen with F11 , the toxic effects of RqlH expression in MG1655 were negated by co-expression of RqlI . Deletion of the C-terminal PRTase domain ( amino acids 509–690 ) of RqlH abrogated the growth-inhibitory and SOS-inducing effects of RqlH expression in the absence of RqlI ( Figs 5 , S5 and S6 ) . To examine the role of the PRTase domain more specifically , residues within the putative phosphoribosylpyrophosphate binding pocket of RqlH were mutated . Two PRTase domain variants were tested: one having four conservative amino acid substitutions ( DIVD663EVME; [24] ) within the putative binding pocket and one having a single alanine replacement ( D663A ) . These residues were chosen based on alignments of RqlH with two phosphoribosyltransferases , PurF and Hpt , and by overlaying the crystal structure of PurF with the predicted structure of RqlH ( S7 Fig ) . Expression of either the DIVD663EVME or D663A RqlH mutant variants had no notable effects on bacterial growth , activation of the SOS response , or filamentation , and had only modest inhibitory effects on the numbers of viable bacteria recovered from the cultures ( Figs 5 , S5 and S6 ) . These data indicate that the PRTase domain is necessary for RqlH-mediated toxicity , and therefore the defects associated with the ΔrqlI mutant . However , expression of the PRTase domain alone was not sufficient to inhibit bacterial growth and did not trigger activation of the SOS response or filamentation , suggesting that additional elements within RqlH cooperate with the PRTase domain to promote toxicity in the absence of RqlI ( Fig 5 ) . To better understand the role of the helicase domain of RqlH , we tested two mutant proteins with single amino acid substitutions ( K49A and D148A ) that were previously shown to abrogate RqlH helicase activity in M . smegmatis [21] . When these mutant proteins were expressed in MG1655 , the bacteria grew to a normal density , but low CFU counts were recovered from the cultures ( Figs 5 and S5 ) . This reduction in CFUs coincided with increased activation of the SOS response and the development of more filamentous cells by the K49A and D148A mutants ( Figs 5 and S6 ) . The loss of viable CFUs was not rescued by co-expression of RqlI , though RqlI did diminish the effects of the K49A and D148A mutations on activation of the SOS response and filamentation . Titers were restored to WT levels if RqlH carried both the K49A mutation within the helicase domain and the D663A mutation within the PRTase domain ( Figs 5 and S5 ) . These data suggest that both RqlI and the helicase domain of RqlH are required to keep the PRTase domain of RqlH in check . However , it is clear that these regulatory interactions are complex . As a case in point , the combined expression of RqlI with the double mutant RqlH K49A/D663A protein leads to a reduction in bacterial titers without a coordinate decrease in culture density , as seen with the single K49A and D148A RqlH single mutants . ( Figs 5 and S5 ) . These data are further complicated by the fact that mutations within helicases often have dominant-negative effects that disrupt DNA metabolism [25] . The discordance between culture densities and bacterial CFUs observed with recombinant strains that express the K49A or D148A RqlH mutants ± RqlI , or with bacteria that express both RqlI and the RqlH K49A/D663A double mutant ( S5 Fig ) , is likely in part attributable to stress-induced inhibition of septation and the formation of filamentous cells . Filamentous bacteria will increase the optical density of a culture , but may register as fewer CFUs in plating assays . In our experiments , recombinant strains that reached WT culture densities but had much less than WT titers tended to have more filamentous bacteria ( Figs 5 , S5 and S6 ) . For the K49A and D148A mutants , filamentation in the absence of RqlI correlated with activation of the SOS response , but this phenomenon was notably less pronounced when RqlI was also expressed ( S6 Fig ) . The co-expression of RqlI with the K49A/D663A RqlH mutant protein had little effect on SOS activation , even though filamentation levels were elevated and CFU counts were markedly reduced . These data indicate that SOS-dependent and SOS-independent filamentation may contribute to the reduced CFU counts recovered following growth of strains that express RqlH helicase domain mutants . The discrepancies between cell culture densities and titers seen with some of the recombinant strains could also be explained by the generation of anucleate cells or by plating deficiencies , whereby cells may grow well in broth culture but poorly when spread onto agar . To test the former possibility , bacteria were stained using the fluorescent DNA dye Hoechst and imaged . By microscopy , no cells that lack DNA were detected among any of the recombinant strains examined ( see S8 Fig as an example ) . To test for plating deficiencies , bacteria were diluted into fresh LB broth culture at the end of a 6 h growth period and the OD600 was tracked over time . Those strains that had decreased CFU counts at the end of the initial 6 h incubation also had a lag in growth when sub-cultured 1:100 into fresh broth , as indicated in S5 Fig as an increase in the time required to reach an OD600 of 0 . 4 . These results suggest that the recombinant strains that reach WT density levels in culture , but have reduced titers on agar , have fewer viable cells and have no inherent plating deficiencies . As already noted , RqlI contains three putative domains , including a region of homology to DprA , a MoCo binding domain , and a helix-turn-helix ( HTH ) domain ( Fig 1A ) . To test whether these various domains are required for RqlI function , plasmids encoding different truncation variants of RqlI were created and transformed into F11ΔrqlI . Western blots indicated that the recombinant RqlI proteins were expressed , though at varying levels ( S9 Fig ) . As shown in Fig 1B , the growth defect observed with F11ΔrqlI under microaerobic conditions can be partially rescued by expression of the full-length RqlI protein ( 1–400; pCWR16 in Figs 6 and S10A ) . Expression of the N-terminus of RqlI plus the DprA homology region ( 1–250; pCWR6 ) also partially rescued growth of F11ΔrqlI . Of note , the 1–250 RqlI mutant complemented F11ΔrqlI much like the full-length RqlI even though relatively low levels of the truncated protein were detected ( S9 Fig ) . Nearly complete rescue of F11ΔrqlI was attained by expression of an RqlI truncation mutant ( 1–320 ) that lacked the C-terminal 80 residues , including the putative HTH domain ( Figs 6 and S10A ) . These results indicate that the HTH domain is not needed for full RqlI activity under microaerobic conditions , and that the HTH domain may instead actually be somewhat inhibitory to bacterial growth in these assays . In contrast , the expression of RqlI deletion mutants that lack the HTH domain as well as the N-terminal 129 or 84 residues adjacent to the MoCo binding and DprA homology domains did not complement F11ΔrqlI . We also examined the ability of the RqlI variants to abrogate the inhibitory effects of RqlH expression on growth of MG1655 under aerobic conditions ( Figs 6 and S10B ) . Paralleling results obtained with F11ΔrqlI , the expression of RqlI truncation mutants that lack the HTH domain as well as the N-terminal 84 or 129 amino acids were unable to counter RqlH , while the production of full-length RqlI or the 1–320 truncation mutant effectively rescued growth of RqlH-expressing MG1655 . In contrast , the 1–250 RqlI truncation mutant that restored growth of the F11ΔrqlI strain under microaerobic conditions was unable to block RqlH-mediated toxicity in MG1655 under aerobic conditions . In total , these results indicate that the HTH domain is dispensable to the ability of RqlI to interfere with the toxic effects of RqlH under either aerobic or microaerobic conditions . It is also clear that the N-terminus , including both the DprA and MoCo binding domains , is required for full RqlI activity in these assays . However , the C-terminal portion of the putative MoCo binding domain is differentially required , potentially dependent on strain background and/or oxygen levels . To test whether the ability of RqlI to keep RqlH activity in check involves physical interaction between the two proteins , co-immunoprecipitations were performed using MG1655 expressing FLAG-tagged RqlI and HA-tagged variants of either full-length RqlH or a C-terminal truncation mutant ( 1–509 ) that lacks the PRTase domain ( Fig 7 ) . As expected , the RqlH and RqlI proteins were not detected in the control strain carrying the empty plasmid constructs . Full-length RqlH was co-immunoprecipitated with RqlI , and vice-versa , suggesting that these two proteins do associate . This interaction was abrogated in bacteria that express the RqlH 1–509 truncation mutant , indicating that RqlI likely binds the PRTase domain of RqlH . As a control for the specificity of the co-immunoprecipitation procedure , we note that the lac operon repressor protein LacI , which is not expected to interact with either RqlH or RqlI , was not pulled down in these assays . Our finding that RqlI is an especially important regulator of RqlH toxicity under low oxygen conditions ( Fig 3D ) fits well with previous work in which we showed that RqlI is critical to ExPEC survival within the urinary tract and bloodstream [14] . These niches , like many sites of infection , have limiting amounts of oxygen available for use by facultative anaerobes like ExPEC . Since a primary reservoir of ExPEC within vertebrate hosts is the intestinal tract , in which oxygen levels are generally low [26] , we set out to determine if RqlI was also important to the fitness of F11 within the gut . For these studies , we used marked strains that carried chromosomal ClmR or KanR cassettes . Adult Balb/c mice were inoculated via oral gavage with 109 CFU of F11 or F11ΔrqlI , and survival was tracked by titration of fecal pellets on selective plates . WT F11 persisted within the gut at fairly steady levels ( medians ~106 CFU/g feces ) for more than 10 d , while titers of the ΔrqlI mutant were significantly reduced as early as day 1 ( Fig 8A ) . By the end of the experiment , only 2 of the 10 mice that received F11ΔrqlI remained colonized with the mutant . The persistence defect observed with F11ΔrqlI was also evident when the mutant was co-inoculated 1:1 with WT F11 . In these competitive assays , the WT strain outnumbered the ΔrqlI mutant by more than a 1 , 000-fold within 3 d ( Fig 8B ) . To test if RqlI is important for gut colonization by other ExPEC strains , rqlI was deleted from the urosepsis isolate CFT073 and from the pyelonephritis isolate 536 . In competition assays , CFT073ΔrqlI titers within the intestinal tract were greatly diminished within 3 d of inoculation , mirroring the situation seen with F11 ( S11A Fig ) . While 536ΔrqlI was also defective in gut colonization , the decline of this mutant was more gradual and less pronounced , decreasing about a 100-fold relative to the WT 536 strain by day 7 post-inoculation ( S11B Fig ) . These results , coupled with our previous findings [14] , indicate that RqlI promotes the colonization and persistence of various ExPEC strains within distinct host environments , including the urinary tract , the bloodstream , and the intestinal tract . To determine if the in vivo defects observed with the ΔrqlI mutants are attributable to the disregulation of RqlH activities , as seen in our in vitro assays , F11 with a complete deletion of the rqlHI operon was competed 1:1 against the WT F11 strain in our mouse models for gut colonization , UTI , and bacteremia . Following oral inoculation of Balb/c mice with a 1:1 mixture of F11 and F11ΔrqlHI , fecal bacterial titers were tracked over the course of 50 days ( Fig 8C ) . Median competitive indices stayed close to zero throughout the experiment , indicating that the ΔrqlHI mutant does not have a defect in mouse gut colonization , in sharp contrast to the single ΔrqlI mutant . Defects previously observed with F11ΔrqlI survival within the bladder and the bloodstream of mice [14] were also ablated if rqlH was deleted together with rqlI ( Fig 8D and 8E ) . In total , these data indicate that the toxicity associated with RqlH expression in the absence of RqlI in our in vitro assays is also manifest in vivo within diverse host environments . Although the in vivo defects associated with the rqlI mutant are not particularly surprising in light of our in vitro findings , it is noteworthy that the in vivo effects are generally greater in magnitude than those seen in vitro . Given the close relationship between rqlH and rqlI that we observed in F11 , we wondered if these proteins are always present as a pair within bacterial genomes , if they are always adjacent to one another , and if they are widely distributed among bacteria . A search for orthologs of RqlH and RqlI was undertaken using NCBI sequence databases , followed by analysis of the results using custom Python scripts . We defined an ortholog as a hit that aligns with at least 75% of the RqlI or RqlH protein sequence and that has a percent positive score of at least 50 and an E-value of 1e-06 or better . In addition , putative orthologs had to successfully return the original query protein ( RqlI or RqlH ) as the top hit in reciprocal BLAST searches against F11 . For a list of RqlI and RqlH orthologs , and for the raw data used in our analysis , see S1 Dataset . Our results indicate that RqlH or RqlI orthologs are distributed across a phylogenetically diverse range of bacteria , with a total of 240 distinct genera represented in the analysis . Unexpectedly , the majority of bacterial strains ( 55 . 3% ) encoded an ortholog of RqlH , but had no detectable ortholog of RqlI ( Fig 9A ) . There was also a small portion ( 1 . 2% ) of the bacteria that contained only RqlI . 43 . 5% of bacteria within the database encoded both RqlI and RqlH . By analyzing different bacterial subsets , we found that there is a striking difference in the distribution of RqlH and RqlI orthologs in Gram-positive versus Gram-negative . The majority ( 87 . 9% ) of Gram-negative bacteria encode both RqlI and RqlH , with just 9 . 7% carrying only RqlH ( Fig 9B ) . In contrast , nearly all ( 99 . 6% ) Gram-positive bacteria encode only RqlH , while the remaining few have both proteins ( Fig 9C ) . Bacteria with only an RqlI ortholog were found exclusively among the Gram-negative bacteria ( Fig 9B ) . These observations were echoed in our analysis of specific individual genera . For example , 98 . 6% of the Escherichia strains ( Gram-negative ) encode both RqlH and RqlI ( Fig 9D ) , while very few have only RqlH or only RqlI . Streptomyces species—like most other Gram-positive taxa—encode only RqlH , with no detectable orthologs of RqlI in any of the strains ( Fig 9E ) . This is also the case for all strains of Mycobacterium smegmatis , the Gram-positive acid-fast species in which RqlH was first identified [21] . Within strains that have both rqlH and rqlI orthologs , we found that the genes are always oriented with rqlH upstream of rqlI , reminiscent of the operon structure in F11 . In most of these strains , the start of rqlI is close to or overlapping with the terminus of rqlH , though in some strains the two genes are separated by more than 2 kbp . Fig 9F shows examples of the variable arrangements of rqlH and rqlI within divergent species . In Klebsiella oxytoca M5aI and Pseudomonas aeruginosa VRFPA06 , the rqlH and rqlI genes are closely associated , or even slightly overlapping , whereas within the cyanobacterium Trichodesmium erythraeum IMS101 the two genes are separated by an unannotated 2196 bp sequence . Occasionally , we identified other annotated elements inserted between the rqlH and rqlI loci , as in Shigella sonnei 53G where sequences encoding the PRTase domain of RqlH have been separated from the helicase domain by an insertion element and a putative transposase . Together , these observations demonstrate that rqlH and rqlI orthologs , when found together within the same genome , are largely syntenic with the F11 rqlHI operon . The exceptions to this arrangement , as in Shigella sonnei 53G , may reflect alternate ways in which bacteria have evolved to modulate the potentially toxic effects of RqlH expression .
This study was aimed at functionally defining EcF11_3933 , a hypothetical gene that we previously identified as an important facilitator of ExPEC fitness in zebrafish infection models and in mouse models of UTI and sepsis [14] . Homologs of this gene are also expressed by UPEC isolates recovered directly from the urine of women with UTI [27] . Results presented here indicate that the EcF11_3933 gene product , which we have renamed RqlI for RecQ-Like Helicase Interactor , is able to bind to and modulate the activity of the RecQ-Like Helicase RqlH . We propose a model in which RqlI and RqlH act cooperatively to perform an as-yet undefined beneficial function on bacterial DNA ( Fig 10 ) . Although more work is required to further vet this idea , the hypothesis is supported by several pieces of data . First , the RqlH homolog in M . smegmatis is known to be a helicase belonging to the RecQ family [21] , which includes several helicases that perform general maintenance functions on prokaryotic and eukaryotic genomes [28] . Second , RqlH and RqlI physically interact ( Fig 7 ) and are encoded within the same operon ( Fig 4A and 4B ) . Third , the ΔrqlI mutant experiences several defects that are attributable to RqlH activities , all of which involve phenotypes associated with DNA stress . These include induction of the SOS response , increases in cell length ( filamentation ) , higher sensitivity to DNA damaging agents , and elevated mutation rates ( Figs 2 and 3 ) . We suggest that in the presence of a partially functional or disregulated RqlH-RqlI system , some harmful DNA byproduct ( s ) is created that leads to a loss in cellular growth and viability ( Fig 10B ) . Interestingly , the defects associated with disruption of the RqlH-RqlI system are exasperated under low oxygen conditions , suggesting possible links with anaerobic respiration and/or other redox-sensitive processes . Our observations and the model presented in Fig 10 raise a number of questions concerning the RqlH-RqlI system . Chief among these is the question of function . We were unable to detect any decrease in bacterial fitness in ΔrqlHI mutants in our in vitro experiments or in mouse models of infection , using either competitive or non-competitive assays ( Figs 8 and S11 ) . These results may reflect limitations in the resolution of the mouse models , or their inability to recapitulate all aspects of natural colonization and infection processes . For example , our assays employed human ExPEC isolates that may use different strategies to colonize the human versus the murine intestinal tract . The intestinal microbiota associated with humans and mice are often similar at the phyla level , but can be quite different qualitatively and quantitatively at lower taxonomic levels [29 , 30] . Such differences may affect the types of nutrients and stresses encountered by ExPEC within the gut and could potentially influence the requirements for genes like rqlHI . ExPEC are versatile organisms , and so it is also feasible that RqlH-RqlI are important to the survival of these pathogens within environmental reservoirs or at other sites that were not modeled in our study . The key to understanding the role of the RqlH-RqlI system in bacterial physiology may be the PRTase domain of RqlH . Indeed , RqlH is unique among helicases because it possesses a PRTase domain in addition to its helicase domain . PRTase domains can be found in proteins that function in nucleotide salvage pathways by adding phosphoribosyl groups to spent nucleotides . As an example , the guanine-xanthine phosphoribosyltransferase ( Gpt ) of E . coli catalyzes the addition of phosphoribosyl pyrophosphate ( PRPP ) to guanine to make the nucleotide guanosine monophosphate ( GMP ) [31] . Given the combination of a helicase with a PRTase domain in a single protein , it is tempting to speculate that RqlH functions to add PRPP directly to DNA , or that it adds PRPP to nucleobases while in association with DNA . The function of rqlHI may be related to uptake of foreign DNA , as there are at least two connections between the RqlH-RqlI system and natural transformation in competent bacteria . First , RqlI contains a region of homology to DprA , a protein that has been shown to protect ssDNA during the translocation of foreign DNA into naturally competent bacteria [32] . Of note , in addition to the DrpA homology region within RqlI , F11 and other ExPEC isolates encode a DprA protein that is more closely related to the canonical DprA . However , DprA has no detectable effect on transformation in a K-12 E . coli strain [33] . The second connection to natural transformation is the PRTase domain of RqlH , a domain that is found in a number of proteins involved in competence in other bacteria . In Bacillus subtilis , two proteins encoded by the comF operon are suggestive of the functional domains contained within RqlH . Specifically , the comF operon , which B . subtilis requires for competence , codes for a putative helicase and a predicted PRTase separated by an intervening gene with no known domains [34] . The Com cluster of genes that facilitate transformation in Haemophilus influenzae also includes a gene ( ComF ) that is homologous to the PRTase domain of RqlH , but it is seemingly not associated with any nearby helicase [35] . Arguing against a role of the rqlHI operon in natural transformation is the fact that it is generally thought that E . coli is not naturally competent , though it can be made chemically competent or electrocompetent in a laboratory setting . Nonetheless , incidences of natural transformation by E . coli have been reported [36 , 37] , but we have been unable to detect any transformation events using our ExPEC isolates , in agreement with a recent report [38] . In light of our results showing that RqlH is toxic to E . coli strains in the absence of RqlI ( Fig 3 ) , it was surprising to find that many bacteria carry RqlH but do not encode RqlI homologs ( Fig 9 ) . In particular , nearly all Gram-positive organisms that encode an RqlH homolog lack any identifiable RqlI ortholog ( Fig 9C ) . These observations may be explained in several ways . It could be that RqlH on its own is not inherently toxic to some bacteria ( e . g . Gram-positive microbes ) due to differences in their physiology in comparison with ExPEC and the K12 strain MG1655 . It is also possible that bacteria that lack rqlI express different , less toxic rqlH alleles than found in bacteria that encode both genes . However , we did not detect major differences in the sequences of RqlH proteins from bacteria that have just RqlH versus those with both RqlH and RqlI . Finally , RqlH activity could be kept in check by other proteins that are more distantly related to RqlI . In this same vein , it is interesting to note that RqlI orthologs are more prevalent among known pathogens ( 83 . 3% of RqlI homologs are encoded by pathogens in the TEA proteomic database; see [14] ) , while RqlH orthologs are much less often pathogen-associated ( 39 . 1% ) . Altogether , these observations suggest that RqlH function could vary depending on bacterial strain background and the presence or absence of RqlI . The observation that RqlH expression in the absence of RqlI is toxic to E . coli strains is reminiscent of a type II toxin-antitoxin ( TA ) system . These TA systems consist of a protein toxin that is more stable than its cognate antitoxin protein , which binds to and inhibits the toxin [39] . TA systems were originally found encoded on plasmids where they promoted plasmid maintenance . In any cell that spontaneously loses the plasmid , the TA system is no longer transcribed , and the remaining toxin protein outlasts the residual antitoxin , resulting in bacterial cell stasis or death . In this and other similar situations , TA systems can be thought of as selfish genetic elements that function only to proliferate without providing any benefit to the host bacterium . More recently , TA systems have also been shown to have beneficial functions that can protect bacteria from stressors . For example , TA systems can promote the formation of antibiotic-insensitive persister cells [39] and select TA systems in ExPEC can promote stress resistance and colonization of the urinary tract [40] . Results presented here indicate that RqlH and RqlI can in many ways act like a TA system , meshing with data from others showing that an RqlH ortholog ( PsyrT ) in Pseudomonas syringae is toxic to the bacteria in the absence of the RqlI ortholog PsyrA [41] . However , we note that if RqlH and RqlI are to be defined as a TA system , it would be non-canonical for the following reasons: 1 ) RqlH and RqlI are much larger ( 698 and 400 amino acids , respectively ) than typical TA system proteins ( ~100 amino acids ) ; 2 ) rqlH is situated upstream of rqlI , whereas in most type II TA systems the antitoxin is encoded upstream of the toxin; and 3 ) an ΔrqlHI mutant exhibits increased survival ( persisters ) in the face of antibiotic treatments ( S12 Fig ) , in contrast to the results obtained when bona fide TA systems are deleted [40] . Finally , we suggest that the distribution among most bacteria of RqlH orthologs that are encoded in the absence of RqlI argues against the possibility that the RqlH-RqlI system always functions as a TA-like genetic element , although it can resemble one within ExPEC strains . Cumulatively , this work illustrates the complexities of assigning function to hypothetical genes that are identified as fitness determinants in high-throughput screens , such as Tn-seq . The ability of RqlI to counter the oxygen-sensitive genotoxic effects associated with RqlH helps explain why ExPEC strains are dependent on RqlI expression within the microaerobic confines of the gut and other extraintestinal niches . Specifically , our work indicates that RqlI functions to modulate the context-dependent toxicity of the RqlH helicase and its PRTase domain , but opens up many more questions concerning the functional relevance of the RqlH-RqlI system among ExPEC strains and of the numerous other RqlH and RqlI orthologs identified among thousands of phylogenetically diverse species . Future analyses detailing the regulation and functionality of the RqlH-RqlI system and its many homologs may provide the basis for novel anti-bacterial therapeutics that can unleash the toxic potential of RqlH .
Mice were handled in accordance with protocols approved by the Institutional Animal Care and Use Committee at the University of Utah ( Protocol number 10–02014 ) , following US federal guidelines indicated by the Office of Laboratory Animal Welfare ( OLAW ) and described in the Guide for the Care and Use of Laboratory Animals , 8th Edition . The ExPEC strains used in this study included the cystitis isolate F11 , the urosepsis isolate CFT073 and the pyelonephritis isolate 536 , as well as the K-12 strain MG1655 ( S1 Table ) . Manipulations of chromosomal DNA , including both the generation of knockout strains and the knockin strain MG1655 attTn7::PsulA-GFP were achieved with lambda-Red-mediated recombination using plasmid pKM208 [42] . Knockout strains were produced using PCR products containing ~40 bp overhang regions with homology to target loci . To create the MG1655 attTn7::PsulA-GFP reporter strain , a tetA-sacB cassette was amplified from T-SACK cells [43] , and combined via PCR with two 500 bp arms homologous to the attTn7 site downstream of glmS . Next , the tetA-sacB cassette was replaced with the PsulA-GFP reporter from pJLJ3 with selection on Tet/SacB counter-selection agar , as described [43] . Primers used to create knockout and knockin constructs , as well as those used for mutant confirmation , are listed in S1 Table . All PCR products were purified with the DNA Clean & Concentrator-5 kit ( Zymo Research , catalog #D4004 ) . Where indicated , antibiotic resistance cassettes were removed using the flippase-expressing pCP20 plasmid , as described [44] . Bacteria used for genetic manipulations were grown in LB broth , whereas bacteria used for both the in vivo and in vitro experiments were grown in modified M9 media . The modified M9 media contained M9 salts ( 6 g/L Na2HPO4 , 3 g/L KH2PO4 , 1 g/L NH4Cl , and 0 . 5 g/L NaCl ) , as well as 1 mM MgSO4 , 0 . 1 mM CaCl2 , 0 . 1% glucose , 0 . 00125% nicotinic acid , 0 . 00165% thiamine , and 0 . 2% casein amino acids . All plasmids used in this study are listed in S2 Table . For previously unpublished plasmids , the primers utilized in their creation are also indicated . RqlI expression vectors were created by inserting PCR products into the PstI and HindIII sites of pRR48 , under control of an IPTG-inducible promoter [45] . RqlH expression vectors were made by insertion at the SacI and HindIII sites within pBAD33 , downstream of an arabinose-inducible promoter [46] . Plasmids were generated using standard cloning techniques , the QuikChange II XL Site-Directed Mutagenesis Kit by Agilent Technologies , or by overlap extension PCR [47] , as indicated in S2 Table . Female Balb/c mice were purchased from Jackson Labs and used between 7–8 weeks of age , and all experiments were performed in accordance with IACUC-approved protocols . All mouse experiments were repeated at least twice , and the total combined data from six or more animals is presented . Bacteria used to inoculate mice were grown statically at 37°C in 20 mL modified M9 in 250 mL Erlenmeyer flasks for 24 h , pelleted by spinning at 8000 r . c . f . for 8 min , washed , and resuspended in phosphate buffered saline ( PBS ) . For competitive assays , WT and mutant bacteria were mixed in equal parts prior to inoculation . These experiments used bacteria carrying either a chloramphenicol or kanamycin resistance cassette , allowing them to be easily selected and distinguished from other bacteria within the microbiota , and enabling the facile tracking of individual WT and mutant strains in competitive experiments . To assess intestinal colonization by the ExPEC strains and their mutant derivatives , mice were orally gavaged with 50 μl of a bacterial suspension containing 1x109 CFU of bacteria in PBS . Fecal samples were collected at various time points post-gavage , homogenized in 1 mL 0 . 7% NaCl , and briefly spun to pellet any insoluble debris . Supernatants were then serially diluted and plated on LB agar containing chloramphenicol ( 10 μg/ml ) or kanamycin ( 50 μg/ml ) . Fecal samples were also collected and plated prior to gavage to ensure that the resident microbiota did not include any culturable chloramphenicol- or kanamycin-resistant bacteria . For the UTI model , mice were anesthetized using isofluorane inhalation and slowly inoculated via transurethral catheterization with 50 μl of a bacterial suspension containing ~1x108 bacteria in PBS . After 3 d , the bladders were extracted and homogenized in PBS . Homogenates were then serially diluted and plated . To initiate bacteremia/sepsis , mice were injected intraperitoneally with 200 μl of PBS containing ~1x107 bacteria . The animals were monitored for signs of extreme morbidity and euthanized at 12 h post-inoculation . The livers , kidneys , and spleens were then recovered and homogenized in PBS . Bacterial titers present within the various tissues were quantified by plating serial dilutions of the homogenates . To quantify bacterial growth in various oxygen levels , bacteria were grown overnight in M9 in loose-capped tubes , then subcultured the following day 1:100 in M9 . During the subculture stage , the bacteria were grown on a shaking incubator for 24 h in 12-well plates either in aerobic conditions or in a Mitsubishi AnaeroPack 2 . 5 L Rectangular Jar holding a AnaeroPack-Microaero sachet or AnaeroPack-Anaero sachet to model microaerobic and anaerobic conditions , respectively . To calculate generation time , the bacteria were titered at two time points during exponential phase . To measure sulA-GFP expression in F11 bacteria carrying the pJLJ3 plasmid or the pJLJ1 empty vector control , bacteria were grown overnight in loose-capped tubes in M9 with kanamycin , then subcultured 1:100 into fresh M9 with kanamycin . Cells were then grown microaerobically and anaerobically for 24 h in 12-well plates , or aerobically for 4 h in loose-capped tubes . For the control conditions , mitomycin C was added to a final concentration of 0 . 25 μg/mL in aerobic cultures at 3 h , and then incubated for another hour . Both GFP fluorescence and the OD600 were measured using a BioTek Synergy H1 plate reader . For experiments with the MG1655attTn7::PsulA-GFP strain carrying various RqlH mutant proteins , the cells were grown overnight in M9 with ampicillin and chloramphenicol in loose-capped tubes The bacteria were then subcultured 1:100 into M9 with ampicillin , chloramphenicol , IPTG ( 1 mM ) , and arabinose ( 0 . 05% ) , and incubated aerobically in loose-capped tubes for 6 h . A growth curve was performed by again subculturing cells after the 6 h growth period 1:100 into fresh LB . The OD600 was then read every 30 minutes by a Bioscreen C machine . In the assays , 1 mM IPTG or 0 . 05% arabinose was added to the media to induce expression of RqlI , RqlH or their derivatives . Following overnight growth from frozen stocks , F11 , F11ΔrqlH , F11ΔrqlI , and F11ΔrqlHI were diluted 1:100 into modified M9 media and grown with shaking ( aerobically ) for 4 h in loose-capped tubes . After addition of mitomycin C ( 0 . 25 μg/mL ) , the incubations were continued for one more hour . Bacterial titers before and after mitomycin C treatment were determined by plating serial dilutions . Survival rates were calculated as the numbers of viable bacteria present after mitomycin C treatment as a percent of bacteria present prior to addition of the toxin . To assess strain susceptibility to UV irradiation , bacteria were grown aerobically in loose-capped tubes from overnight cultures for 4 h at 37°C , then serially-diluted and spread onto LB plates in order to obtain ~100 cells per plate . Plates were then exposed to 0 , 25 , or 50 J/m2 UV light produced by a Stratalinker 1800 . The plates were then incubated overnight and surviving bacteria quantified . To visualize bacteria , 5 μl of cells ( ~5x107 cells ) resuspended in PBS containing 1 μg/mL Hoechst dye were spread on a glass slide and incubated at room temperature until the PBS had evaporated . A drop of FluorSave Reagent was added to each slide and a coverslip was placed on top . Bacteria were imaged by fluorescence or phase-contrast microscopy using an Olympus BX51 microscope equipped with a 100X oil immersion objective and a QImaging QIClick Cooled CCD camera . To measure cell lengths , straight-line measurements from one tip of a cell to the other were made using ImageJ [48] . For each bacterial strain , 6–10 fields and more than 200 bacteria were examined . A PsulA-lacZ reporter plasmid ( pCWR2 ) was created by overlap extension PCR using a promoterless lacZ plasmid ( pCWR1 ) as a template ( S2 Table ) . The reporter plasmid was introduced into F11ΔrqlIΔlacZY::clm , and the resulting strain was randomly mutagenized by conjugation with EcS17/pSAM-Ec donor bacteria as described [14] . For conjugation , bacteria were grown overnight in LB broth , and 1 mL of donor cells were mixed with 0 . 5 mL of recipient cells , washed once with LB broth , and then spread onto an LB plate . After a 5 h incubation at 37°C , bacteria were recovered from the plate surface , resuspended in 1 mL of LB broth , and diluted 1:100 . Aliquots ( 100 μl ) were plated onto tetrazolium lactose plates containing ampicillin , chloramphenicol , and kanamycin to select for transposon-mutagenized F11ΔrqlIΔlacZY::clm/pCWR2 bacteria . Colonies that were darker than the F11ΔrqlI ΔlacZY::clm/pCWR2 strain were re-streaked onto new tetrazolium lactose plates to verify decreased PsulA-lacZ activity . Transposon mutants that returned sulA expression to WT levels underwent a secondary screen in a broth-based modified Miller assay , as described [49] . For mutants that exhibited decreased sulA expression on both tetrazolium lactose plates and in the Miller assays , transposon insertion sites were mapped by arbitrary PCR . Briefly , 30 μl colony PCR reactions were carried out using the pSAM-EC kan us1 primer paired with each of the arbitrary primers ARB1A , ARB1B , and ARB1C ( S3 Table ) . The thermocycler program for this reaction was as follows: 95°C for 5 min; 5 cycles of 94°C for 30 sec , 30°C for 30 sec , and 72°C for 1:30; 30 cycles of 94°C for 30 sec , 45°C for 30 sec , and 72°C for 2 min; and ending with a 5 min incubation at 72°C . A 2 μl aliquot of this reaction was then used as template in a second PCR reaction using the pSAM-EC kan us2 and ARB2 primers , in a total volume of 50 μl . The second thermocycler program was as follows: 94°C for 2 min; 30 cycles of 94°C for 30 sec , 55°C for 30 sec , 72°C for 1:30; and then 72°C for 5 min . The entire volume of this second reaction was resolved on an agarose gel , and the largest distinct band was gel-purified and sent for Sanger sequencing using the primer pSAM-EC kan us3 . Bacteria were grown overnight in modified M9 media and then sub-cultured 1:100 into several 1 mL cultures . These were grown for either 4 h aerobically in loose-capped tubes or for 24 h microaerobically in 12-well plates . A few cultures were serially diluted and plated onto LB agar to quantify the total number of cells . The entirety of each remaining culture was then spread onto LB plates containing 20 μg/mL nalidixic acid . The following day , colonies were counted to quantify total bacterial titers as well as the numbers of nalidixic acid resistant cells . Results were analyzed by the online FALCOR calculator for fluctuation assays to determine mutation rates [50] . To probe for rqlHI mRNA , WT F11 was grown aerobically for 4 h in loose-capped tubes in modified M9 media after being diluted 1:100 from an overnight culture . RNA was extracted from cells using a Norgen Total RNA Purification Kit , then treated with DNase for 1 . 5 h , precipitated , and resuspended in water . To create cDNA , 0 . 5 μg of RNA was used with the Superscript III First-Strand Synthesis Kit ( Life Technologies ) . Control reactions lacking reverse transcriptase were run concurrently . The cDNA and control reactions were then used as template in PCR reactions using two primer sets that spanned the rqlH and rqlI gene junction , or with one primer set internal to rqlI ( S3 Table ) . For RT-qPCR , RNA was extracted from bacteria grow under aerobic or microaerobic conditions and cDNA produced as described above . A LightCycler 480 instrument was used to run the qPCR reactions using primers specific to frdA , rqlH , or rqlI , or primers that span rqlH and rqlI ( S3 Table ) . Results were normalized to 16S rRNA levels . For analysis of protein complexes , a control MG1655 strain carrying the empty vectors pRR48 and pBAD33 , or recombinant MG1655 strains expressing FLAG-tagged RqlI and HA-tagged RqlH or the RqlH ( 1–509 ) truncation mutant were diluted 1:100 from overnight cultures into 20 mL minimal M9 media with ampicillin and chloramphenicol . After 4 h of growth at 37°C , IPTG ( 1 mM ) and arabinose ( 0 . 05% ) were added to induce expression of RqlI and RqlH , respectively . Cultures were grown for 1 h more and bacteria were then pelleted by centrifugation and frozen . After thawing , bacterial pellets were resuspended in 500 μL B-PER Bacterial Protein Extraction Reagent ( Life Technologies ) containing 1X cOmplete protease inhibitor cocktail ( Roche ) , 1 mM PMSF , and 1 μL/mL Lysonase ( Merck Millipore ) , and then incubated at room temperature for 10 min . A small aliquot of each lysate was set aside for use as input controls for the immunoprecipitations . Lysates were incubated with Dynabeads Protein G ( Life Technologies ) pre-loaded with mouse anti-HA ( Santa Cruz; SC-7392 ) or rabbit anti-FLAG ( Sigma Aldrich; F7425 ) antibodies . After three washes in PBS , bead-protein complexes were resuspended in 30 μL sample buffer ( 4% sodium dodecyl sulfate [SDS] , 20% glycerol , 0 . 02% bromophenol blue , 10% 2-mercaptoethanol , 0 . 125 M Tris-HCl [pH 6 . 8] ) and heated for 5 min at 95°C in preparation for SDS–polyacrylamide gel electrophoresis ( PAGE ) . Input controls ( 5% of initial lysates ) for the co-immunoprecipitation experiments were similarly prepared for SDS-PAGE . Following electrophoresis , proteins were transferred to Immobilon-FL PVDF membranes ( Millipore ) that were then probed using antibodies specific for the FLAG or HA epitope tags or for LacI ( AbCam; ab33832 ) . Blots were developed using HRP-conjugated secondary antibodies with the BM Chemiluminescence Blotting Substrate ( Roche ) and imaged using a BioRad ChemiDoc MP system . To analyze expression of the epitope-tagged RqlH and RqlI proteins and their variants , or to confirm GFP expression levels by the reporter strain MG1655attTn7::PsulA-GFP , bacterial cells were lysed and protein concentrations were determined as described above . Proteins from each sample were resolved by SDS-PAGE , transferred to PVDF membranes , and probed using anti-FLAG , anti-HA , or anti-GFP ( Santa Cruz; SC-9996 ) antibodies . Subsequently , blots were stripped and probed using anti-E . coli antibody ( BioDesign ) to ensure equal protein loading . To find homologs of RqlH and RqlI among other bacteria , the F11 sequence of each protein was used as a query for an online blastp search using the non-redundant protein sequences database . Additionally , tblastn searches were performed against the non-redundant nucleotide and whole genome shotgun databases in order to find unannotated orthologs that would not be present in the non-redundant protein database . All BLAST searches were done with an expect threshold cutoff of 1e-06 , and custom Python scripts utilizing the Biopython module [51] were used to parse the results . The hits were reciprocally blasted against the F11 proteome , and those that brought back the original query protein ( i . e . RqlH or RqlI ) as the top hit , aligned with at least 75% coverage , and had a percent positive score of at least 50 , were defined as homologs . The species that carried homologs of only RqlH , only RqlI , or both were enumerated . To examine synteny of the rqlH and rqlI genes within bacteria that carried both genes , only hits originating from the blastp search were considered . Protein GenBank records were accessed via EFetch with Biopython , and parsed to obtain the accession numbers of the nucleotide sequences containing each rqlH or rqlI ortholog and its location within its respective genome . The gene locations were then compared to determine orientation and proximity of rqlH and rqlI to each other . Persister cell assays were performed as described [40] . Briefly , bacteria were diluted 1:100 from overnight cultures into 5 mL of LB broth and then grown with shaking ( 225 rpm ) at 37°C for 2 h . At this point , an aliquot of each culture was taken for CFU enumeration prior to addition of ampicillin ( 100 μg/mL ) or ciprofloxacin ( 10 μg/mL ) . After a further 5 h incubation , 1 mL of each culture was pelleted , washed once with LB broth , and surviving bacteria were determined by plating serial dilutions . Percent survival , reflecting the numbers of persister cells in each sample , was calculated by dividing the number of viable bacteria present after antibiotic treatment by the number present prior to antibiotic addition . Results from in vivo competition assays were analyzed by one sample T tests . Results from noncompetitive assays in mice were analyzed by Mann-Whitney two-tailed t tests . Results from in vitro assays were analyzed by unpaired Student’s t tests . Prism 5 . 01 ( GraphPad Software , Inc . ) was used for all statistical tests , including Student’s t tests . P values of less than 0 . 05 were defined as significant . | Extraintestinal pathogenic Escherichia coli ( ExPEC ) cause the majority of urinary tract infections , and are also able to infect the bloodstream , meninges , and various other sites within the human host . These infections are becoming increasingly difficult to treat as ExPEC strains gain resistance to many of the antibiotics that are commonly used in the clinic . The development of improved treatment strategies requires a deeper understanding of the factors that promote ExPEC fitness and virulence within the host . In genetic screens , we identified a functionally uncharacterized protein , RqlI , which promotes ExPEC survival within diverse host environments . We find that RqlI binds to and works in tandem with RqlH , a protein that has been shown in other bacteria to unwind DNA . In the absence of RqlI , we found that RqlH can become toxic to ExPEC , causing DNA damage and slower growth . A specific part of RqlH that is predicted to manipulate the nucleotides that make up DNA is responsible for this toxicity . The ability of RqlH to inhibit bacterial growth when not held in check by RqlI suggests that the specific inactivation of RqlI could have therapeutic value in combating ExPEC and other pathogens that express these proteins . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | The Extraintestinal Pathogenic Escherichia coli Factor RqlI Constrains the Genotoxic Effects of the RecQ-Like Helicase RqlH |
The molecular mechanisms by which floral homeotic genes act as major developmental switches to specify the identity of floral organs are still largely unknown . Floral homeotic genes encode transcription factors of the MADS-box family , which are supposed to assemble in a combinatorial fashion into organ-specific multimeric protein complexes . Major mediators of protein interactions are MADS-domain proteins of the SEPALLATA subfamily , which play a crucial role in the development of all types of floral organs . In order to characterize the roles of the SEPALLATA3 transcription factor complexes at the molecular level , we analyzed genome-wide the direct targets of SEPALLATA3 . We used chromatin immunoprecipitation followed by ultrahigh-throughput sequencing or hybridization to whole-genome tiling arrays to obtain genome-wide DNA-binding patterns of SEPALLATA3 . The results demonstrate that SEPALLATA3 binds to thousands of sites in the genome . Most potential target sites that were strongly bound in wild-type inflorescences are also bound in the floral homeotic agamous mutant , which displays only the perianth organs , sepals , and petals . Characterization of the target genes shows that SEPALLATA3 integrates and modulates different growth-related and hormonal pathways in a combinatorial fashion with other MADS-box proteins and possibly with non-MADS transcription factors . In particular , the results suggest multiple links between SEPALLATA3 and auxin signaling pathways . Our gene expression analyses link the genomic binding site data with the phenotype of plants expressing a dominant repressor version of SEPALLATA3 , suggesting that it modulates auxin response to facilitate floral organ outgrowth and morphogenesis . Furthermore , the binding of the SEPALLATA3 protein to cis-regulatory elements of other MADS-box genes and expression analyses reveal that this protein is a key component in the regulatory transcriptional network underlying the formation of floral organs .
In contrast to animals , most developmental processes of plants occur postembryonically and integrate a variety of internal and environmental cues . Modularity of plant development is based on the ability of plants to maintain pools of undifferentiated stem cells throughout the life cycle of the plant . Stem cells near the tip of the growing shoot are located in the shoot apical meristem , where different types of plant organs , such as vegetative leaves or floral organs , can be initiated from the flanks of the meristem . Leaves and floral organs are “variations of a theme”: they arise via modifications of a common basic genetic program [1] . Which type of organ is produced by the meristem depends on the developmental phase of the plant . Initially , leaves are produced during the early vegetative phase of the plant , followed by the transition to reproductive phase , which triggers the transformation of vegetative shoot meristems to inflorescence and floral meristems , giving rise to flowers and floral organs , respectively . Thus , change in the identity of plant organs is initiated by reprogramming within the meristems [2] . Plant developmental biologists have identified a number of key regulatory genes that trigger changes in meristem and organ identity , many of them encoding transcription factors , chromatin remodeling factors , or other signaling molecules like microRNAs ( miRNAs ) . One family of transcription factors that is important in this process is the MADS-box gene family [3] . MADS-box genes play crucial roles in the switches from vegetative to inflorescence and finally to floral meristems . These latter meristems give rise to flowers and floral organs , respectively [4] . Developmental transitions and organ differentiation require global changes in gene expression . The genome of the model flowering plant Arabidopsis thaliana is roughly 20-fold smaller than the human genome; still , it encodes about 27 , 000 protein-coding genes , which is more than found for humans ( http://www . arabidopsis . org; [5] ) . One of the most challenging current questions is how developmental control genes trigger global changes in gene expression during the multiple phase transitions and in organ identity determination , starting from a small pool of undifferentiated cells . In the present study , we focus on the MADS-box transcription factor SEPALLATA3 ( SEP3 ) . SEP3 is a member of the SEP subfamily of MADS-box genes , whose members have nearly redundant functions in the specification of floral meristem identity and in the identity of all types of floral organs: sepals , petals , stamens , and carpels . Triple mutants impaired in SEP1–3 function have flowers with floral organs converted into sepals and display a loss of determinacy in the center of the flower [6] . This phenotype masks the involvement of the SEP genes in processes occurring later in development , e . g . , the formation of the ovules as has been shown by Favaro et al . ( 2003 ) [7] . The SEP3 protein appears to be the central player , since it is part of at least a dozen different MADS domain dimer complexes [8] and it is expressed throughout flower development , from the floral meristem to fully developed floral organs [9] . This suggests that SEP3 is a multifunctional protein controlling a plethora of developmental processes . According to the current model of flower development , the SEP3 protein is proposed to mediate the higher-order protein complex formation between MADS-domain proteins with more specific floral organ identity functions [10] . Furthermore , it may provide the transcriptional activation potential to the floral homeotic protein complexes [10] . More-recent evidence suggests that the SEP3 protein may also recruit transcriptional corepressors , demonstrating that it can modulate the function of the plant protein complexes in a broader sense , depending on the availability of cofactors [11] . However , evidence for higher-order complex formation between MADS-domain proteins comes mostly from protein interaction studies in heterologous systems and genetic data , and there is no indication for the relevance of these interactions in target-gene recognition in planta so far . Another question is how different MADS-domain protein complexes achieve functional specificity , since the in vitro DNA-binding characteristics of MADS-domain proteins appeared rather similar , and the short DNA sequence motifs supposedly bound by MADS-domain proteins are very abundant in the Arabidopsis genome [12] . In order to characterize the mode of action and general downstream pathways of floral homeotic genes , we generated genome-wide DNA-binding profiles of SEP3 in its native context . Chromatin immunoprecipitation ( ChIP ) followed by ultrahigh-throughput Solexa ( Illumina ) sequencing ( ChIP-SEQ ) has been shown recently to be a powerful tool to obtain genome-wide DNA-binding patterns of transcription factors [13 , 14] . The large numbers of short individual sequence reads produced by novel instruments facilitate the digital quantification of DNA sequences that are present in a sample . An alternative method comprises the combination of ChIP and whole-genome microarrays ( ChIP-CHIP ) to map the genomic DNA regions enriched in the immunoprecipitated sample [15 , 16] . These genomic tiling arrays are available for Arabidopsis and have been used to map binding sites for plant transcription factors [17] . We compared the targets of SEP3 in wild-type and the floral homeotic agamous ( ag ) mutant background . In the ag mutant , stamens are replaced by petals , and instead of the carpels in the fourth whorl , a new mutant flower is formed [18] . Accordingly , the analysis of this mutant should reveal SEP3 target genes specifying petal development , whereas targets that are specific to stamens and carpels should be absent . We further studied the function of SEP3 in the regulation of downstream pathways by analyzing the effects of a dominant repressor version of SEP3 in plants . The genome-wide identification of direct target genes of SEP3 provides a framework for a hierarchical transcriptional network underlying the formation of floral organs . SEP3 binds to thousands of genomic regions containing the consensus binding sites for MADS-domain proteins , but it also acts as part of regulatory modules with other transcription factors . These modules link floral homeotic gene functions with organ growth . Our analysis identified multiple links between SEP3 and hormonal pathways , and in particular auxin signaling . Auxin signaling is crucial for the outgrowth and development of lateral organs , and its role in flower development has been suggested previously based on mutant phenotypes [19–21] . Our ChIP-SEQ data and the phenotypes of plants that repress direct SEP3 targets in a dominant-negative fashion suggest cooperation of SEP3 and genes in the auxin pathway in the regulation of floral organ growth and differentiation .
To identify genomic regions bound in vivo by SEPALLATA3 ( SEP3 ) , we used ChIP followed by deep sequencing by ChIP-SEQ [14 , 15] . In a parallel experiment , the ChIP was followed by whole-genome tiling array hybridizations to identify enriched regions ( ChIP-CHIP ) . For the ChIP experiments , we used inflorescences including inflorescence meristems and floral buds of stage 1–12 from Arabidopsis wild-type and agamous ( ag-1 ) mutant plants . Protein–DNA complexes were immunoprecipitated using a peptide antibody specific to SEP3 . As a negative control , we performed ChIP-SEQ using the same antibody on sep3-1 mutant plants . Western blot analysis revealed that the SEP3 antibody reacts exclusively with the SEP3 protein ( Figure S1 ) . From the ChIP-SEQ experiments , we obtained between 3 to 7 . 5 million approximately 35-bp sequence reads after one to three independent rounds of sequencing for each sample , of which 30%–40% were uniquely mapped on the Arabidopsis nuclear genome ( Table S1 ) . The uniquely mapped reads were extended to 300 bp in order to recover the average original DNA fragments that were subjected to sequencing in a similar fashion as described by Robertson et al . ( 2007 ) [14] . This allows positioning of the maximum of enrichment present in the samples at high resolution . The number of mapped reads was counted for every nucleotide position ( defined as “number of hits” ) , and for each strand independently . For some genomic positions , we observed that the number of hits was only due to reads with identical sequence . Although , it is expected that some reads will have an identical sequence , it is also expected that a true peak of enrichment should be represented by several reads with partly overlapping but different sequences . In order to avoid any artifact due to identical sequence reads , we included the requirement that the number of hits at each genomic position should be supported by reads mapped in both DNA strands . To test for enrichment at each nucleotide position in the sample compared to the control , we used a score based on the Poisson distribution , as it is commonly used for statistical modeling of tag counts [22] . For each genomic region representing a candidate peak , the maximum score value was used to test the significance of the peak ( defined as “peak score” ) . We used the false discovery rate ( FDR ) to control the error rate of our testing procedure . The number of significant peaks for the ChIP-SEQ datasets is given in Table 1 . Notably , SEP3 binds to thousands of regions in the Arabidopsis genome . At FDR <0 . 001 , we found 4 , 282 significantly enriched regions for the SEP3 in wild-type plants , and 2 , 828 regions in the ag mutant . Thus , at this level of significance and given our data , SEP3 seems to bind a reduced number of regions in the ag mutant compared to wild-type plants . We used biological replicates and comparison of ChIP-SEQ and ChIP-CHIP data in order to evaluate the reproducibility of the generated genome-wide binding profiles of SEP3 . To simplify the comparison of ChIP-SEQ replicates , the average score in nonoverlapping 5 , 000-bp windows for each replicate was calculated . A high correlation between the two different sequencing rounds and biological replicates was found ( Figure S2E and S2G ) . Peaks positions and ranks from independent biological replicates overlap strongly ( Figure S2H ) . Since more sequences were produced for replicate 1 , we focused in our further bioinformatic analysis on this replicate . Comparison of the ChIP-CHIP and the ChIP-SEQ experiments reveals a good agreement between the results of the two methods , as reflected in the large overlap in peak positions and similar ranking of the peaks ( Figure S2 ) . We were also interested in comparing the positional resolution of the two platforms . The average width of the most strongly significant ChIP-SEQ peaks is around 800 bp . In contrast , the peaks with an equivalent ChIP-CHIP rank have a width of approximately 1 , 300 bp ( Figure S2 ) . This larger window size for the ChIP-CHIP peaks compared to ChIP-SEQ peaks results in a lower positional resolution , which is particularly problematic in regions with multiple binding sites that are close to each other . cis-Regulatory elements controlling gene expression are preferentially found in the promoters of target genes . However , there are also numerous examples of important regulatory sequences in introns , particularly near the 5′ end of genes [23] . A typical example is the second intron in the MADS-box gene AG , which is bound by multiple factors [24–26] . AG , on its turn , binds to the downstream region of the SPOROCYTLESS gene , demonstrating that cis-regulatory elements are not exclusively located in the upstream or intragenic regions of plant genes [27] . We determined the position of the putative binding sites relative to the nearest gene based on our ChIP-SEQ dataset . As evident from our ChIP-SEQ experiments , most in vivo binding sites of SEP3 are close to or within protein-coding genes and only about 6% to 8% of all peaks ( FDR <0 . 001 ) are not located within 3 kb upstream to 1 kb downstream of any genomic locus ( wild-type and ag datasets ) . A total of 3 , 475 genes are targeted by SEP3 in wild type , whereas 2 , 424 genes are putative targets in the ag mutant at FDR <0 . 001 ( Tables S2 and S3 ) . In agreement with its role as transcriptional regulator , DNA-binding sites of SEP3 are predominantly located in the upstream region of genes ( Figure S3; Table S3 ) . Notably , we found the highest enrichment of SEP3 binding sites in a region spanning a few hundred base pairs directly upstream of the annotated transcriptional start of genes ( Figure S3 ) . Surprisingly , binding sites are also enriched in the downstream region of genes , located just downstream of the 3′ UTR , although this enrichment is clearly less pronounced than the one found near the transcriptional start site of genes . Within genes , peaks are preferentially located in introns and UTR regions ( Figure S3 and Table S3 ) . It is known from in vitro–binding studies that MADS-domain proteins bind to specific DNA elements called CArG boxes ( reviewed in [12] ) . Most well known is the serum response element ( SRE or SRF ) -type CArG box , which has the consensus CC[A/T]6GG . A related sequence motif is the MEF2-type CArG box , which has the general consensus C[A/T]8G but is usually more strictly defined as CTA[A/T]4TAG . Plant MADS-domain proteins often show relatively broad DNA-binding preferences , recognizing SRF− and MEF2− , as well as intermediate motifs [28 , 29] . CArG boxes are frequently found in the Arabidopsis genome , so the presence of this motif alone is not sufficient to predict targets of MADS-box transcription factors [12] . Given the large number of MADS-box transcription factor genes present in the Arabidopsis genome and the capacity to form an even larger number of heterodimeric transcription factor complexes [8] , and considering their divergent functions in plant development , it is important to understand how different MADS-domain proteins ( and protein complexes ) achieve target-gene specificity . Also , the relevance of the formation of higher-order protein complexes in DNA binding , associated with the binding to more than one CArG box , has not been demonstrated in planta yet . We considered the 1 , 001 bp surrounding the peak maximum score position , defined as peak area , for the characterization of transcription factor binding sites . The peak areas were searched for the presence of different types of CArG boxes . Our results show that the SRF , MEF2 , and intermediate types are enriched in the genomic regions bound by SEP3 in vivo ( Figure 1A ) . In agreement with a true enrichment of CArG boxes in genomic regions bound by SEP3 , we found that the “background” frequency of CArG boxes of type CC[A/T]6GG in promoter regions ( −1 , 000 bp upstream ) was approximately 7% , whereas in our ChIP-SEQ data at FDR = 0 . 001 , the frequency is approximately 12 . 5%; and in more strongly bound regions , the frequency increases to more than 20% . SRF and intermediate types of CArG boxes showed considerably more enrichment than MEF2-type CArG boxes , suggesting that they are more frequently bound by SEP3 or SEP3-containing protein complexes in planta . The CArG boxes are usually positioned in the center of the peak ( peak maximum score position ) ( Figure 1B ) . The strong preference of CArG boxes for the center of the peak demonstrates the high positional resolution of ChIP-SEQ experiments combined with our method of peak detection . In order to further characterize the features of binding sites recognized by SEP3 and its complex partners in planta , we used the ChIP-SEQ peak score as a measure of affinity for binding sites within enriched genomic regions as implemented in the MatrixREDUCE software [30] . We found that the obtained consensus models tend to be relatively flexible in most nucleotide positions ( Figure 1 ) . This suggests that the in vivo binding data reflect a mix of affinities of different homo- and heterodimeric SEP3 protein complexes . Since MADS-domain proteins bind as dimers to a CArG box , CArG boxes can be considered as composite elements , with each half-site contacted by a different monomer with possibly different binding preferences . We estimated the frequencies of all possible half-sites for the CArG boxes of the general consensus CC[A/T]6GG and CC[A/T]7G in the ChIP-SEQ data . Using a binomial test , we found that three out of eight possible half-sites for the consensus CC[A/T]6GG were significantly overrepresented in the SEP3 ChIP-SEQ data , and four out of 24 for the consensus CC[A/T]7G ( Table S4 ) . Not all possible combinations of the most frequent half-sites are represented among the most strongly overrepresented core CArG boxes ( Table S4 ) , suggesting dependencies between the half-sites . The most frequently represented sequence of type CC[A/T]6GG , which is CCAAAAATGG , is in fact the same or highly similar to the consensus sequences that were identified by MatrixREDUCE ( Figure 1C ) . Based on the combinations of half-site sequences found in the ChIP-SEQ data , we measured the dependencies between single nucleotides using a chi-square test . Here , we found that strong dependencies exist between nucleotides within the [A/T]-rich core of the CArG box ( Table S5 ) . Surprisingly , we also identified dependencies for the nucleotides surrounding the core CArG box , which is in line with experimental evidence suggesting that sites surrounding the core consensus are contacted by MADS-domain proteins and may contribute to DNA-binding specificity as well [31–33] . The dependencies between nucleotide positions in functional CArG boxes could ( at least partly ) explain why only 7 . 7% and 5 . 7% of all CArG boxes perfectly matching the consensus CC[A/T]6GG and C[A/T]7GG , respectively , are bound in vivo by SEP3 . According to the “floral quartet” model [34] , higher-order MADS-domain protein complexes bind to two CArG box–like DNA sequences at short distance from each other . Thus , we would expect an enrichment of ChIP-SEQ peaks with regulatory modules consisting of two CArG box elements . In order to identify regulatory modules composed of more than one binding site , we used the Explain software [35] . The module with the highest fitness score ( 0 . 748 on a scale of 0 to 1 ) was composed of a single pair of CArG boxes separated by a DNA stretch varying between 10 to 200 bp in length . This finding supports the idea that MADS-domain proteins act in complexes composed of two dimers that bind to two adjacent binding sites , as has been predicted by the floral quartet model . Next , we were interested whether CArG boxes within ChIP-SEQ peaks have preferred distances from each other . To test this , we compared the distribution of distances of CArG boxes within peaks to “background distributions” obtained from random sets of Arabidopsis promoters or randomized sequences . As shown in Figure 2 , there is a preference for close distances , with the strongest preference around 42–43 bp ( which corresponds to four helical turns of the DNA ) . Aside from these relatively short distances , the frequent occurrence of multiple peaks in the same genomic regions open the possibility that MADS-domain protein complexes can also bridge and bend larger DNA stretches as was suggested in the floral quartet model [34] . In addition to our targeted screen for MADS-domain binding sites , we used MatrixREDUCE and MEME [36] in order to identify DNA sequence motifs that are abundant in the regions bound by SEP3 . Using these tools , we recovered motifs corresponding to CArG boxes , but interestingly , we also detected sequence motifs potentially bound by non-MADS transcription factors . A motif that was identified using these programs has the consensus sequence CACGTG . This motif has been named “G-box” in the literature and represents a DNA-binding site for bHLH and bZIP transcription factors [37] . We found that this motif is indeed overrepresented in the genomic regions bound by SEP3 ( Figure 3A ) . Similar to CArG boxes , the G-box motif is enriched in the center of the peak ( Figure 3B ) . A second motif that was found using these programs has strong similarity to the DNA-binding consensus of TCP transcription factors with the general consensus motif CCNGGG [38] . We analyzed whether this TCP DNA binding consensus was overrepresented in the regions bound by SEP3 . Indeed , it appeared to be enriched with increasing peak score threshold , and it is enriched in the center of the peak ( Figure 3A and 3B ) . Next , we tested systematically for enrichment of known DNA-binding consensus sequences of transcription factors using information from the Transfac and AGRIS databases . In total , these databases contain information for 105 ( Transfac ) and 72 ( AGRIS ) DNA-binding consensus sequences of plant transcription factors . In addition to confirming the enrichment of MADS- , TCP , and bHLH/bZIP binding sites , we found that also ARF , C2H2 ( ID1 ) DNA recognition motifs , and a bHLH ( MYC ) DNA-binding site similar to the G-box were enriched with increasing peak score threshold , and located in the center of the ChIP-SEQ peaks ( Figures 3 and S4 ) . The ChIP-SEQ experiments were done with samples from wild-type plants and the ag mutant . We were interested in the overlap of DNA-binding sites in these two samples , which could point to target genes involved in the formation of perianth organs . There is clear preference for overlapping genomic positions of SEP3 ChIP-SEQ peak maximum positions in wild type and ag mutant ( Figure 4A ) . The overlap of potential SEP3 targets in wild type and ag mutant is also evident from genes that are targeted by SEP3 ( Figure 4B ) . Whereas the number of peaks in ag is only approximately 65% of the number of peaks in wild type at FDR = 0 . 001 , the overlap in affected target genes is almost 70% at the same FDR level ( Figure 4B ) . Thus , individual peaks are more likely affected by loss of AG than target genes . With increasing significance level , targets of SEP3 in wild type and ag mutant overlap progressively more ( Figure 4B ) , as do individual peaks . Highly enriched target genes are usually common to wild type and ag mutant ( overlap >90% ) . According to these results , only a small fraction of strongly enriched direct target genes is specific to the SEP3 complexes specifying stamen and carpel development ( Table S2 ) . These genes may represent candidate genes determining the specific morphologies of stamens and carpels downstream of the floral homeotic genes . In the ChIP-SEQ approach to identify potential SEP3 targets , different floral tissues corresponding to different developmental stages were used . In order to evaluate the relevance of DNA-binding events in the regulation of the genes corresponding to the ChIP-SEQ peaks , we used comprehensive gene expression array data that are publicly available . Mainly , the collection of AtGenExpress experiments provides information about timing of gene expression and changes in different floral homeotic mutants [39] . We found that about 45% of all genes with significantly enriched peaks in the SEP3 ChIP-SEQ experiment ( FDR <0 . 001 ) were differentially expressed at very young developmental stages in at least one of the homeotic mutants ( lfy-12 , ap1-15 , ap2-6 , ap3-6 , ag-12 ) ( Figure S5 ) . This fraction was higher than the overall genome-wide fraction of differentially expressed genes in these mutants ( 29% ) . Forty-five percent ( 903/2022 ) of the genes with ChIP-SEQ peaks in the ag mutant are differentially expressed in the ag-12 mutant compared to wild type ( up to floral stage 12; 28% ( 5 , 927/21 , 039 ) in the total dataset ) . Considering SEP3 binding sites in genes that are differentially expressed during development , we found the strongest enrichment for genes that change expression in the meristem during the earliest stages of floral development ( FDR <0 . 001 , p-value 4 . 3e−40 , binomial test ) . About 63% of the potential targets of SEP3 are differentially expressed at any stage of reproductive development starting from floral transition to flowers of stage 12 ( Figure S5 ) . In total , 72% of the potential SEP3 targets are differentially expressed during flower development or in any of the homeotic mutants . Although the differential expression can also be due to indirect effects , the data suggest that the majority of potential direct SEP3 targets may also be regulated by SEP3 . We also found an enrichment in frequency of genes that are correlated in expression with SEP3 expression with increasing peak score threshold in ChIP-SEQ ( Figure S5 ) . The fraction of genes with ChIP-SEQ peaks that are positively coexpressed with SEP3 is clearly higher than that of negatively regulated genes , supporting the idea that SEP3 acts mostly as a transcriptional activator . Genetic and gene expression experiments suggest that the SEP genes are required for the up-regulation of floral homeotic genes , and that this up-regulation is crucial for the establishment of the identities of the different floral organs . However , until now , it has not been demonstrated whether this regulation is direct . We analyzed the binding profiles for the genomic loci corresponding to the floral homeotic genes and found that SEP3 binds to nearly all of these loci ( Figure 5 ) . Only SEEDSTICK ( STK ) and CAULIFLOWER ( CAL ) do not have significantly enriched regions . In most cases , the peaks are located in the promoters of the respective genes . In case of the APETALA1 ( AP1 ) , APETALA3 ( AP3 ) , SEP1 , and SEP2 loci , there are also peaks in the 5′ UTR , whereas a SEP3 binding site is present in the second intron of AGAMOUS ( AG ) . Of all homeotic genes , the genomic regulatory sequences controlling the expression of AG and AP3 are best characterized . The spatial expression pattern of AP3 is driven by regulatory elements within approximately 500 bp upstream of the transcriptional start . CArG boxes in this part of the promoter are important for the positive as well as negative regulation of AP3 [40] . Our ChIP-SEQ results demonstrate that SEP3 binds to the genomic region comprising positively and negatively acting CArG boxes in the AP3 promoter , strongly suggesting a direct molecular link between the binding of SEP3 and the regulation of AP3 . Most regulatory sequences controlling the expression of AG are located in its second , 4-kb large intron [24 , 25] . Consistent with this observation , we identified a peak of enrichment of SEP3 in this intron . More specifically , the peak marks a CArG box in the 3′ activation domain located in this intron . The 3′ activation domain functions in the up-regulation of AG in stage 3 floral meristems and is also responsible for maintenance of AG expression in developing carpels [24] . This CArG box was also found to be bound by AG itself in previous experiments [26] . Interestingly , we identified a second peak of enrichment in the upstream region of AG . Consistent with the idea that an AG/SEP heterodimer is responsible for the positive autoregulation of AG [26] , the heights of the peaks in the AG locus are reduced in the ag mutant compared to wild type ( Figure 5 ) . The regulatory sequences controlling the expression of the SEP1–4 genes are still not well characterized . Our ChIP-SEQ results , however , strongly suggest autoregulation of the redundantly acting SEP MADS-box genes . All floral MADS-box genes that are targeted by SEP3 in wild type , are also targeted in the ag mutant , although there is some variation in the heights or presence of individual peaks ( e . g . , AP1 , SEP2 , SHP1; see Figure 5 ) . This raises the possibility that different SEP3 complexes may have different affinities to individual binding sites . In order to characterize the regulatory effects of SEP3 on MADS-box genes that are potential direct targets , we analyzed the expression of these MADS-box genes upon SEP3 induction using a constitutively expressed translational fusion of SEP3 to the rat glucocorticoid receptor hormone binding domain ( GR ) . For this , seedlings expressing the 35S:SEP3-GR construct were treated with dexamethasone ( DEX ) for 8 h , 1 d , or 10 d , and the expression of MADS-box genes was determined by real-time reverse transcriptase ( RT ) -PCR . The relative expression levels in comparison with nontreated plants is shown in Figure 6 . Our results reveal that SEP3 is indeed able to activate the expression of other floral homeotic genes as suggested previously [6] . SEP3 itself is most strongly up-regulated , demonstrating a strong autoregulatory feedback loop . Although some of the tested genes show an early response to SEP3 induction , others are regulated only after prolonged SEP3 induction , suggesting that SEP3 alone is not sufficient to regulate these genes , but needs to interact with partner proteins that are encoded by the induced MADS-box genes . In particular , AP3 , AG , and AP1 are strongly activated by SEP3 . These three genes correspond to the three major classes of floral homeotic genes according to the classical ABC model: class A ( AP1 ) , class B ( AP3 , together with PI ) , and class C ( AG ) . Their gene products also represent major protein interaction partners of SEP3 , suggesting that later induced targets of SEP3 are targets of the corresponding SEP3-containing protein complexes . Thus , SEP3 can activate the flower developmental program by enhancing the expression of its interaction partners as one of its first steps . Induction of the ABC classes of genes is sufficient to form the flower , which explains the very early flowering and the terminal-flower phenotype that we observed upon SEP3 induction . Whereas flower-specific genes are mostly activated , MADS-box genes that are involved in the floral transition ( AGL24 , SOC1 , and SVP ) tend to be down-regulated by SEP3 . Together with the fact that SEP3 binds to the promoters of these genes , our results suggest that SEP3 is involved in the down-regulation of these genes during early flower development , possibly as part of protein complexes together with other flower-specific MADS-domain proteins , such as AP1 . It has been a long-standing question in plant developmental biology whether floral homeotic genes act directly on the structural or metabolic genes that create the final morphology of floral organs , or whether they act via intermediate regulators , i . e . , other transcription factors , which in turn regulate subsets of targets conferring the final organ shape and function . To answer this question , we investigated the enrichment of gene ontology ( GO ) terms [41] among genes that are closest to the peaks as a function of their ChIP-SEQ peak score . In terms of molecular function , genes encoding transcription factors are clearly the most enriched group of genes ( GO:0030528; p-value 3 . 62e−19 ) . When dissecting gene functions according to biological processes , there is a clear enrichment for genes involved in development , in response to hormonal stimuli , and in lipid biosynthesis . Figure 7 presents the top-five most enriched specific GO terms . The SEP3 targets in the GO category “lipid biosynthetic process” include genes involved in hormone biosynthesis ( terpenoid and steroid pathways ) , as well as in sterol and wax synthesis . Next , we were interested in whether some transcription factor families were more frequently represented among potential direct SEP3 target genes than others . The results shown in Table 2 reveal that 15 transcription factor families were significantly overrepresented among SEP3 targets . Interestingly , we found overrepresentation of families for which we also found enrichment of their DNA-binding sites in the ChIP-SEQ data: bHLH , TCP , and ARF families are overrepresented in both the target and the binding site datasets , which points to the existence of autoregulatory feedback loops ( Figure 5 , Table 2 ) . In general , we found overrepresentation of transcription factor families with known functions in the control of flowering time ( SBP and C2C2-Co-like ) , organ growth ( GRF and TCP ) , auxin response ( AUX-IAA and ARF ) , brassinosteroid response ( BES1 ) , and meristem development ( HB and GRAS ) . Sixty-six percent of the loci belonging to these families identified in wild type , were also found in the ag mutant at the same FDR threshold . The characterization of overrepresented GO categories and transcription factor families suggests that SEP3 is involved in the regulation of hormonal signaling . In order to further understand this link , we analyzed the overlap between the potential direct downstream targets of SEP3 and genes regulated by different hormones [42] . We found that genes regulated by auxin , gibberellic acid , and brassinosteroids were most represented among SEP3 targets , but also genes responding to other hormones were enriched ( Figure 8A ) . Among the top 200 genes targeted by SEP3 are several enzymes involved in hormone biosynthesis ( e . g . , GA1 and AOC2 ) , signaling ( e . g . , BRI1 ) , or homeostasis ( e . g . , GH3 . 3 ) ( Figure 8B ) . GH3 . 3 is involved in auxin homeostasis and has been found to be up-regulated as a later response of carpel and stamen induction by AG [26] . Since auxin signaling was consistently found to be overrepresented among our potential direct SEP3 targets , we further analyzed SEP3 binding patterns at auxin-related genes ( Figure 8C ) . Most Aux-IAA genes showed similar SEP3 binding patterns , with a peak close to the transcriptional start site ( similar to GH3 . 3 ) . Next , to genes involved in auxin transport ( e . g . , PIN4 and PID ) and auxin response factors with known roles in flower and fruit development ( e . g . , ARF3 , ARF6 , and ARF8 ) , a miRNA167 locus , which controls ARF6 and ARF8 expression , is also found among the targets . SEPALLATA genes have shown to be crucial for the specification of floral meristem and organ identities [6 , 43] . However , the redundancy among MADS box genes makes the functional characterization of members of this gene family difficult [44] . Indeed , our analysis of potential downstream targets of SEP3 suggests so far unknown links to other developmental and hormonal processes in flower development . Since SEP3 acts mostly as a transcriptional activator , we can alternatively study the functions of this protein by replacing endogenous SEP3 function with that of a chimeric repressor version of SEP3 . For this purpose , we fused the genomic coding region of SEP3 to the EAR ( ERF-associated amphiphilic repression ) domain under the control of a basic SEP3 promoter ( −960 bp ) [45] . This chimeric repressor blocks the activation of direct target genes of the transcription factor complexes in which SEP3 is present . Our transgenic approach indicated novel functions of SEP3 in addition to supporting previously proposed roles: transgenic lines with a strong phenotype showed delayed flowering ( unpublished data ) , reduced number and size of floral organs , as well as defects in organ differentiation ( Figure 9 ) and identity . While petals were mostly absent in these plants , the stamen number was strongly reduced , and the stamens were often reduced to filamentous carpelloid structures or fused with the carpels ( Figures 9C and 9D ) . Plants with strong phenotypes were male and female sterile . The carpel of severely affected lines showed severe growth defects: the size of the ovary was greatly reduced or it was even missing , while the gynophore at the bottom and the style at the top of the gynoecium were enlarged . Ovule placentation was abaxialized in a variable manner ( Figure 9G ) . The aberrant carpel morphology closely resembles the phenotype of mutants impaired in auxin biosynthesis or signaling [18] . The carpel phenotypes are similar to those of pin1 [46] , pinoid ( pid ) [47 , 48] , or arf3/ett mutants [49 , 50 , 51] ( Figure 9B and 9F , versus 9D and 9E ) . In sep1 sep2 sep3 triple-mutant plants , elongated gynophores similar to the pid mutant can be observed ( Figure 9J and 9K ) . The receptor kinase PID determines the polar localization of auxin efflux ( PIN ) proteins , whereas ARF3 , along with other ARF transcription factors , mediates auxin response at the gene regulatory level . Our ChIP-SEQ data suggest multiple links of SEP3 and the auxin signaling pathway: ARF genes and their antagonists , the AUX-IAA factors , as well as ARF-controlling microRNA loci are bound by SEP3 as revealed by the ChIP-SEQ experiments . Also , a genomic region downstream of the PID locus is targeted by SEP3 , both in wild type and in the ag mutant . In agreement with a role of SEP3 regulating these genes , expression microarray data of developmental time series suggest that a majority of ARF transcription factor genes bound by SEP3 , as well as PID and a smaller number of AUX-IAA genes , are up-regulated in reproductive meristems and young floral stages , in a similar fashion to SEP3 itself ( Figure S6 ) . The enrichment of auxin response elements ( ARF binding sites ) in the SEP3 ChIP-SEQ peaks suggests that SEP3 cooperates with ARF proteins in target gene regulation , thus the downstream targets of these complexes could be the target of repression by the SEP3-EAR protein . The need for cofactors in auxin response is illustrated by the finding that ( positive ) auxin response , as measured by the DR5 promoter , is only found in a subset of SEP3-expressing cells . Whereas the expression of SEP3 in undifferentiated meristematic tissues is broader than that of the DR5 marker , the expression domains overlap mostly in growing organs such as the tips of growing sepals ( Figure 9H and 9I ) . DR5 expression has been shown to be also dependent on brassinosteroid signaling , which suggests a complex pattern of upstream regulation of auxin response [52] .
Despite the large number of genomic binding sites , only a small fraction of the potential binding sites that are present in the genome , represented by a CArG box , are indeed bound by SEP3 . de Folter and Angenent ( 2006 ) [12] calculated that the Arabidopsis genome contains far more CArG boxes than the roughly 30 , 000 genes in the genome , whereas our ChIP-SEQ experiments resulted in enriched binding sites in about 3 , 400 genes ( FDR <0 . 001 ) . This indicates that the binding of SEP3 to DNA sequences is highly selective . Most binding sites are located in promoters or other regions with potentially regulatory functions , such as introns . These findings suggest that the majority of significant protein–DNA interactions identified in ChIP-SEQ experiments are likely to be relevant . The functional importance of the DNA-binding events detected in our ChIP-SEQ experiments is also supported by the finding that there is enrichment for specific GO annotations among the targets of SEP3 . The question remains whether all of the enriched DNA regions bound by SEP3 are relevant for gene regulation . Since the plant material used for our ChIP-SEQ experiments comprises different developmental stages , a direct correlation of DNA-binding events and stage-specific changes in gene expression is difficult . However , our comparison of the ChIP-SEQ data to comprehensive gene expression microarray data suggest that the majority of the genes that are bound by SEP3 in planta are also differentially expressed during flower development in a temporal and/or spatial fashion . In contrast to the strongly bound regions , the functional relevance of weakly enriched regions detected in ChIP-SEQ is much less clear . They might represent transient interactions of the transcription factors with DNA [53] , or DNA binding in only a few cells , resulting in a high dilution with tissues that lack the interaction . In a comparable study to decipher binding sites of transcription factors during Drosophila embryo development , Li et al . came to the conclusion that a significant proportion of the poorly bound regions are most likely nonfunctional [15] . These regions corresponded to genes that were poorly modulated in expression , and peaks were often located outside regulatory sequences . It is possible that the weakly enriched regions in our SEP3 ChIP experiment also contain a high proportion of inactive binding sites . Although in vitro studies have revealed consensus binding sites for many transcription factors , including MADS-domain proteins [12 , 32] , it remains unknown what DNA sequence or chromosomal context determines DNA binding site recognition by MADS-box transcription factors in vivo . Our results indicate that the sequence of the cognate binding site , the presence of multiple binding sites , and binding sites for non-MADS cofactors play roles in binding site recognition . The SEP3 ChIP data revealed that not all DNA sequences corresponding to the canonical consensus motifs CC[A/T]6GG and CC[A/T]7G/C[A/T]7GG can serve as functional binding sites for SEP3 and associated MADS-domain proteins in planta . Instead , we found significant enrichment for particular types of half-sites in the SEP3 ChIP-SEQ data . In addition , there is clear interdependence between the sequence of individual half-sites , and between half-sites and surrounding bases . Thus , commonly used consensus sequences and position weight matrices , which assume independency of nucleotide positions of a binding site , are highly oversimplified . The enrichment for certain half-sites may reflect the recognition site for SEP3 , while the other half of the binding site could match the recognition site of the dimer partner of SEP3 . It is known that SEP3 is able to dimerize with many other MADS-domain proteins in yeast assays [8] and in planta ( K . Kaufmann and G . C . Angenent , unpublished data ) , which makes it difficult to determine the consensus CArG box sequence for SEP3 . Combining ChIP-SEQ data obtained from different MADS dimer partners ( e . g . , SEP3 and AP1 ) and determining the overlap in binding sites will elucidate the recognition sites for particular MADS-box dimers . We found that there is a clear overlap of target genes in wild type compared to the ag mutant , although there are differences in number of binding sites and binding affinity at individual sites . Although at this point we cannot exclude that the wild-type binding sites are enriched for perianth-specific targets , these results support the hypothesis that different MADS-domain protein complexes ( e . g . , AG-SEP3 and SEP3-AP1 ) may bind to overlapping sets of target genes , but regulate them in a different way , for instance by recruiting different sets of cofactors leading to differences in activation or repression of genes . In line with the idea that cofactors play a role in differential regulation of downstream targets , it has been shown that SEP3 and AP1 can recruit the corepressor SEUSS , and that this complex acts to repress AG expression in petal development [11] . Our finding that individual peaks are more likely to be affected by loss of AG than target genes makes it also possible that different higher-order complexes may bind to subsets of binding sites present in the promoter . Considering that plant MADS-domain proteins differ in their DNA bending characteristics as determined by gel retardation experiments [28] , different protein complexes may have specific effects on the structural properties of the promoter and by that influence target gene expression . The interplay between different proteins implicates that the dynamics in the floral developmental network is strongly dependent on relative quantities and affinities of individual homeotic proteins competing for common protein interaction partners and key downstream targets . The large number of targets suggests that floral homeotic protein complexes globally control and modify the genetic programs that are active in all plant organs , so that only a limited number of targets would be expected to be unique for the individual floral homeotic protein complexes . In addition to CArG box elements , we found several other transcription factors consensus DNA-binding sequences to be enriched in the center of the peaks , suggesting that these transcription factors act in a combinatorial fashion with MADS-domain proteins in common regulatory modules . Until now , there has been very limited information about interactions between plant MADS-domain proteins and other transcription factors , and this rare information is based on artificial yeast and/or in vitro protein interaction data . The types of transcription factors whose binding sites are enriched in the peaks link MADS-domain proteins with other cellular , developmental , and hormonal pathways . TCP transcription factors have been shown to be important for cell growth [54–56] , and ARF transcription factors are key mediators of auxin response ( reviewed in [57] ) . Whether there are direct protein interactions between these classes of transcription factors still needs to be resolved . In addition to primary sequence characteristics , the accessibility of binding sites , and thus the chromatin structure , can influence the recruitment of MADS-box transcription factors to specific sites in the genome . We observed a strong enrichment of binding sites close to the transcriptional start site of genes , suggesting that the recognition sites are not randomly distributed in the genome . Assuming that any small sequence motif is randomly distributed in the genome , it suggests that not only the primary sequence itself , but also the position of the cis-element is relevant for transcription factor binding . Chromatin remodeling , which is active throughout plant development , is likely affecting the accessibility of transcription factors and transcriptional activity . It occurs during postembryonic developmental transitions ( e . g . , flowering ) and is required for maintenance of meristematic cell identity as well as for the formation of organ primordia , with different chromatin factors acting at different stages ( [58] and references therein ) . Current models of gene regulation suggest that chromatin remodeling and transcription factor binding dynamically alternate due to transient exposure of DNA by displacement of nucleosomes ( [59] and references therein ) . MADS-box transcription factors form large complexes in planta ( K . Kaufmann and G . C . Angenent , unpublished data ) , making it even possible that there is a direct interaction between these transcription factors and chromatin remodeling factors . Interesting in this respect is that we identified four out of five members of ATP-dependent chromatin remodeling complexes as targets of SEP3 in the ChIP experiments , suggesting an interdependence relationship . Our ChIP-SEQ data indicate that there are multiple direct molecular links between floral homeotic genes and hormonal pathways . Interestingly , ARF genes are targets of SEP3 but could also be coregulators , suggesting that autoregulatory circuits exist involving members of unrelated transcription factor families . In addition to specifying floral meristem and organ identity , an important role of SEP and AP1-like proteins is to trigger organ growth . The ap1 cal double mutant produces multiple undifferentiated meristems , in which organ outgrowth and differentiation are impaired [60] . Similar mutant phenotypes arise by combining ap1 and sep mutant alleles , suggesting partial redundancy between members of these subfamilies of MADS-box genes [43] . This hypothesis of a mutual role of AP1/CAL and SEP proteins in organ outgrowth provides a link to auxin-mediated organ development . Interestingly , the expression of SEP3 fused to the EAR suppression domain led to phenotypes that mimicked developmental aberrations observed in pin1 , pid , or ettin ( arf3 ) mutants or plants treated with the auxin transport inhibitor NPA [61] . These plants are characterized by defects in lateral organ outgrowth ( floral buds and floral organs ) and a pistil lacking a functional ovary . Although we observed variable homeotic conversions in SEP3-EAR expressing plants , which would be expected from the down-regulation of SEP3 function , the majority of plants was only affected in outgrowth and differentiation of the floral organs . The mode of action of the SEP3-EAR fusion protein and how it interferes with outgrowth without affecting the homeotic function remain to be studied further . A possible explanation could be that auxin homeostasis is more sensitive to the dominant repression by the SEP3-EAR protein than the floral homeotic functions . Alternatively , SEP3-EAR may suppress ARF function by interacting with ARF proteins in a larger transcriptional complex , a model that is supported by our SEP3-ChIP experiments . Developmental transcription networks are composed of positive and negative regulatory loops , which often unite into larger transcription units . Positive feedback loops that are made up of two transcription factors regulating each other result in a robust expression in response to a transient developmental signal in order to establish and maintain a developmental program [62] . However , negative regulation is also very important in developmental processes to rapidly switch from one program to another ( e . g . , phase transitions ) . The expression of MADS-box genes is often regulated by transcription complexes composed of their own gene products . One of the earliest examples for this phenomenon was the finding that floral homeotic genes responsible for the formation of petals and stamens ( the B class genes in the ABC model ) are up-regulated by dimers consisting of the encoded proteins [63] . More examples of autoregulatory feedback loops in the MADS-box gene family were reported afterwards . Also , SEP genes and their protein products are part of these regulatory networks [6] . The ability of the SEP3 protein to form many different dimer combinations [8] and the interaction of SEP3 protein and other floral homeotic proteins in larger protein complexes [10 , 64] illustrate that SEP3 is a key component in the network . It acts as a hub by linking various developmental programs that occur in the inflorescence and floral meristems and at later stages during organ differentiation . Our ChIP-SEQ data demonstrate that SEP3 is indeed able to bind to the promoters of many floral homeotic genes , supporting the conclusion that SEP3 is a key regulator of flower development ( Figure 10 ) . From Figure 10 , it is apparent that most MADS-box genes are targeted by a combination of several other MADS-domain proteins . The finding that most of these MADS-domain proteins also physically interact with each other suggests that combinatorial interactions of floral homeotic MADS-domain proteins and SEP3 protein are required for multiple positive autoregulatory feedback and feedforward loops . These mutual interdependencies may have evolved in order to enable stable organ-specific expression patterns by attenuating stochastic fluctuations in expression levels that may be more frequent if regulation was mediated by single proteins instead of heteromeric protein complexes . Previously , negative feedback loops were suggested between AP1 , which is a protein interaction partner of SEP3 , and some of the flowering-time MADS-box genes to prevent the expression of these vegetative factors in the flower [65] . Our data also suggest that SEP3 has a role in repression of genes that control flowering time ( e . g . , SOC1 and AGL24 ) . In line with this idea , binding sites of SEP3 and AP1 overlap at the SOC1 promoter ( [65] and our ChIP-SEQ data ) . Interestingly , the binding sites also partially overlap with those of positive regulators at the SOC1 locus ( [66] and our ChIP-SEQ data ) . It is possible that positively and negatively acting factors compete for the same binding sites . Since SEP3 is also able to interact with SOC1 , it is also possible that a SEP3-SOC1 protein dimer is important for negative autoregulation of the SOC1 locus in floral meristems [8] . In general , direct negative ( auto ) regulatory feedback loops may enable an efficient and persistent switch between developmental phases , i . e . , from inflorescence identity to floral identity . Transcription factor complexes and their target genes are major components of transcriptional cascades that drive plant developmental processes . Once we have genome-wide datasets describing more of these interactions , we can integrate the data into transcriptional networks and models describing the interactions of the components of the network and predicting the outcome of modulation of the biological system [67 , 68] . Here , we have shown that ChIP-SEQ is a powerful tool providing these essential datasets to address questions in plant biology . Our findings suggest that many direct regulatory interactions exist in plant developmental networks . Additional information is needed , in particular the dynamics of the interactions in time and space . Furthermore , not only DNA–protein interactions , but also protein–protein interactions of transcription factors need to be resolved for a better understanding of developmental processes in complex organisms .
Arabidopsis thaliana , wild-type ( Col-0 ) , sep3-1 ( Col-0 ) , and ag-1 ( Ler ) mutant plants were grown under standard greenhouse conditions ( 20 °C , long-day light regime: 16-h light , 8-h dark cycle ) . Flower material was harvested from primary and secondary inflorescences of 5–7-wk-old plants . ChIP experiments were performed essentially as described in [69] . For the IP , we used an antibody raised against a C-terminal peptide of SEP3 . The antibody was tested in western analyses on plant extracts of wild-type and sep3 mutant plants . Total extracts were produced using a standard protocol [70]; nuclei extracts were produced following the protocol that was used for the ChIP experiments , only that the flower material was not fixed . Cross-reaction with other SEPALLATA MADS-domain proteins was tested in western blots using proteins produced by in vitro translation . Linker annealing , amplification , and gel purification for the Solexa sequencing were essentially performed as instructed by the Illumina protocol with small modifications . The gel purification was done after the amplification step . We used individual , complete ChIP samples for each amplification reaction for the wild-type and ag-1 mutant samples . For the sep3-1 mutant sample ( negative control ) , we pooled the DNA of three different ChIP experiments to obtain sufficient material for amplification . The amplified material was subjected to Solexa sequencing following Illumina's instructions . The sequence datasets were submitted to Gene Expression Omnibus ( GEO ) ( accession number GSE14600 ) . For the amplification of the ChIP-DNA for ChIP-CHIP , we used a protocol published by [71] with modifications . The amplified DNA was partially digested with DNAse I ( fragments <150 bp ) and labeled with biotin . For the control experiment , we used unamplified , sheared chromatin from the same biological sample as the ChIP-DNA ( “input DNA” ) . The DNA was fragmented and labeled in the same way as the ChIP-DNA . The labeled samples were hybridized to GeneChip Arabidopsis Tiling 1 . 0R Arrays ( Affymetrix ) as biological duplicates . Tiling array data were submitted to GEO ( accession number GSE14635 ) . Enrichments in ChIP sample hybridizations relative to input were calculated from raw intensity ( CEL ) files using a nonparametric statistical method implemented in the Affymetrix Tiling Analysis Software ( TAS ) [72] . Biological replicates were combined in the analysis . Once the significance ( p-value ) was obtained , we define as ChIP-CHIP peaks the genomic regions with a p-value lower than 0 . 05 , and not separated by more than 100 bp with the highest p-value not higher than 0 . 05 . The 35- or 36-nucleotide ( nt ) reads were mapped to the unmasked Arabidopsis reference genome ( ATH1 . 1con . 01222004; ftp://ftp . arabidopsis . org/ ) using the SOAP software [73] , allowing a maximum of two mismatches and no gaps . Iteratively , one base was discarded from the end of the nonmapped reads until the reads were uniquely mapped or fell below a minimum read length of 30 nt . Only uniquely mapped reads were retained . In order to recover the average length of the original DNA fragments that were subjected to Solexa sequencing , the reads were extended directionally to 300 nt . The data analysis was aimed at a comparison of the enrichment found in the wild-type plant and , independently , in the ag-1 mutant , against the enrichment found in the sep3-1 mutant treated as a negative control . For this , data concerning positions of the mapped reads were transformed into numbers characterizing all nucleotide positions in the genome in the following way . Define xis , where i is the nucleotide position and s = 1 , 2 for the examined and control sample respectively , as the minimum of the counts of extended mapped reads that overlap at the position i on the forward and on the reverse strand in the sample s . This value is a conservative estimate of the representation of the nucleotide i in the sequenced samples , supported by both the strands independently . By the transformation observations yis independent of the number of sequenced reads were obtained . The examined sample values yi1 were then normalized with respect to the mean and variance of the distribution of control values yi2 . For the comparison of the observed examined and control counts at position i , the one-sided test based on the Poisson distribution was made according to the probability ( test statistic ) formula where zi2 is the maximum of yi2 and the global coverage obtained for the control whose value is the product of the number and the length of the extended mapped reads divided by the mappable genome length . All genomic regions consisting of nucleotides characterized by calculated probability values smaller than 0 . 05 and not interrupted by a gap of 100 nt or more were identified and assumed to contain a candidate enrichment peak . For presentation purposes , the calculated probabilities were transformed into scores of −loge ( ti ) , and the maximum score value for each candidate peak was used to test the significance . Permutation tests were used to estimate the FDR for the peaks . For this , each mapped read was considered as having the label “sample” when it belonged to the examined sample and “control” if it belonged to the control sample . To obtain the distribution of the test statistic under the null hypothesis of no differences between the examined and the control samples , the labels of the reads were randomly permuted , and for each permutation , the methodology explained above to test differences in distribution was applied . The permutations were run until at least 65 , 000 test statistic values for calculation of the null distribution were obtained . All peaks were characterized by their location with respect to the annotated genes ( as described in the TAIR7_GFF3_genes . gff file , ftp://ftp . arabidopsis . org/ ) . The gene affected by binding at each peak was selected by the following algorithm . First , for all peaks , the affected gene was selected as the one with the peak inside it and the minimum distance to the start . For the peaks outside of the genes , the affected gene was then selected as the one with the peak in its 3 , 000-bp upstream or 1 , 000-bp downstream region and the minimum distance to the start or end , respectively . Thus , it was assumed that a DNA-binding event affects the closest neighboring gene , which results in a conservative estimate of the number of genes controlled by each of the transcription factors . The genes affected by binding inside them were then characterized with respect to the precise annotation of the position of the affecting peak , by considering three categories: 5′ UTR , 3′ UTR , other exon regions ( equivalent to CDS for protein coding genes ) , or “not annotated . ” This characterization was done on the basis of the first available splicing variant for a gene; in practice , this meant using variant 1 for most of the genes and variant 2 for some . Default parameters for MatrixREDUCE [30] and MEME [36] algorithms were used in order to identify motifs de novo . Sequences 500 bp upstream and downstream of the maximum score position for significant peaks ( FDR <0 . 001 ) were obtained from the Arabidopsis reference genome ( ATH1 . 1con . 01222005; ftp://ftp . arabidopsis . org/ ) . Repeat and low-complexity regions were eliminated using RepeatMasker ( A . F . A . Smit , R . Hubley , and P . Green , RepeatMasker at http://repeatmasker . org ) . Sequence affinity logo representations were prepared using AffinityLogo [30] . The MatrixReduce algorithm uses genome-wide occupancy/affinity data for a transcription factor and associated nucleotide sequences to discover the sequence-specific binding affinity of the transcription factor . It utilizes a statistical-mechanical model to describe the relationship between the nucleotide sequences and the occupancy/affinity-related score , therefore avoiding the need of selecting any background sequence model . In contrast , MEME only uses the occupancy/affinity-related score to define a group of nucleotide sequences from which a general binding site consensus will be obtained . Perfect match motif consensus sequences were located in the 1 , 000-bp region around the maximum score position ( defined as “peak area” ) for significant peaks ( FDR <0 . 001 ) using a perl script and the Arabidopsis reference genome . The motif consensus sequences were obtained from AGRIS and Transfac databases [37 , 74 , 75] . Each consensus sequence was associated with the score value of the corresponding ChIP-SEQ peak in order to calculate enrichment . For each significant peak , the nucleotide sequence 500 bp around the position of the maximum peak score location were extracted and associated with the peak score value . To obtain the proportion of peaks with a given DNA binding site consensus at a given peak score threshold level , among the nucleotide sequences associated with a peak score value bigger than the threshold level , the proportion of sequences with at least one DNA binding site consensus was calculated for the sample and control set . Two control sets were generated: ( 1 ) one control set was generated randomly permutating the nucleotide sequences and the peak score value , and therefore destroying any relationship between them , and ( 2 ) another control set was generated permutating the nucleotides within their sequence for each nucleotide sequence independently . Each set was characterized by the number and location ( regarding the center of the sequence ) of the DNA consensus binding sites studied . We considered binding sites to be overrepresented in the ChIP-SEQ data only when they were enriched relative to both controls . For simplicity , only control 2 is shown in Figures 1 and 3 . The proportion of the distance of the DNA binding site consensus to the peak score location was calculated as the distance from the center position of the DNA consensus to the peak score location for each peak with a score bigger than the corresponding threshold at a given FDR level . Nonoverlapping distance ranges of 50 bp were considered to calculate the proportion of distance values within each range among the total number of distance values considered . All graphs were generated with R software . ExPlain [35] was used to identify transcription factor binding sites in the sequence probes generated by the ChIP-SEQ experiment , and to generate promoter models of regulation modules composed of more than one binding site . All CArG box position weight matrices from Transfac [37] were collected into a set , and thresholds were set to minimize false negatives . The algorithm MATCH [37] was used to identify putative sites matching the CArG box matrices in the profile . Site frequencies were compared to a control sequence set consisting of randomly sampled Arabidopsis promoters , assuming a binomial distribution . The algorithm Composite Module Analyst ( CMA , [76] ) was used to identify transcription modules overrepresented in the query dataset , based on the binding site prediction by MATCH . CMA parameters were set to find modules composed of one or more pairs of sites separated by 10 , 11 , . . . , 200 nt . CMA evaluates a fitness score , which includes functions measuring normality , t-test , site orientation and distance , sequence match score , and model complexity . In order to establish whether CArG box pairs presented conserved distances between them , a comparison between the distance distribution between sites found in the real data and those found in two control datasets was performed . The control sets generated consisted of ( 1 ) real promoter sequences from genes chosen at random from the Arabidopsis genome , and ( 2 ) random permutations of the sequences described above . The control or background set using real promoters ( 1 ) makes a more stringent test , because real binding sites , including CArG boxes , can still be found in the sequence . This test shows the specific selection of a given distance between CArG box site pairs in the SEP3-dependent gene's promoters , when compared against other real promoter sequences . The second background set , made out of randomly permuted sequences ( 2 ) , presents sites generated by the random variation of the genome composition , without ( biological ) selection , so that this comparison shows the selection of distances between these sites , when compared with an absolutely random site and distance distribution . The MATCH algorithm was applied to these three sequence sets , using the CArG box profile described above , and the positions of the matching sites collected . The distance between adjacent sites was calculated and the distributions were compared using a binomial test , to identify overrepresented distances . Analysis for GO term enrichment was carried out using the AMIGO server [77] for the top 1 , 000 genes of the ChIP-SEQ dataset ( Database version 2008-05-15 ) . Only GO terms with more than 20 annotated loci were taken into account . We considered the most specific categories , ones that were found to be enriched by AMIGO , for further analyses . To study in more detail the enrichment for the top-five most significant , most specific GO terms ( GO:0008610 “lipid biosynthetic process , ” GO:0045449 “regulation of transcription , ” GO:0009908 “flower development , ” GO:0048513 “organ development , ” and GO:0009733 “response to auxin stimulus” ) and the nonenriched term GO:0051179 “localization , ” each gene was associated with the maximum score value among the peaks that were affecting it . We mapped each gene to a GO term if the gene belongs to one of these GO terms or its children , using the Arabidopsis Information Resource ( TAIR ) ATH-GO-GOSLIM 2008-05-10 . We downloaded log2 gcRMA normalized expression values for experiments of interest from the AtGenExpress developmental atlas [39] . Control probe sets and probe sets matching no or several loci in the Arabidopsis genome were ignored in the analysis . After back-transforming the log2 expression values to original scale , a two-sided Student test statistic was applied to test differential expression between sample and control . These genes with a p-value lower than 0 . 01 were considered as differentially expressed . To generate the dominant repressor , we used the SRDX domain [45] , which is a modified version of the EAR domain of the SUPERMAN protein . The promoter and coding region including introns of SEP3 was amplified from genomic . We added the SRDX domain in two subsequent round of PCR reactions . The primer sequences are available on request . The PCR product recombined into pDONR207 , and subsequently into the destination vector pFP101-35SGa . The destination vector was obtained digesting the plasmid pFP101 with BamHI and HindIII , filled in with Klenow enzyme and blunt-end ligated with Gateway cassetteB . The construct was transformed into the sep3-1 mutant plants and Col-O wild-type plants . SEP3-EAR expression in wild type and sep3 mutant showed similar phenotypes; however . we focused our phenotypic analysis on pSEP3:SEP3-EAR in the sep3 mutant . Transgenic seedlings were germinated and grown on plates without DEX for 10 d and transferred to medium containing 10 mM DEX for 8 h or 1 d , respectively . Alternatively , they were directly germinated on DEX medium and grown for 10 d . cDNA synthesis was produced using the iScript cDNA synthesis kit , and qPCRs were performed using the SYBR green I–based system from BioRad following the manufacturer's instructions . Two technical and two independent biological replicates were analyzed using the MyIQ program . The data were normalized with two reference genes ( TUB and EF ) . All primers are available as supplementary information ( Table S6 ) . The expression values for development stages ( floral transition to flowers stage 12; AtGenExpress experiments 6 , 8 , 29 , 31 , 32 , 33 , and 39 ) were obtained as explained in “Gene expression microarray analysis” above . After back-transforming the log2 expression values to original scale , the Spearman rank correlation test was applied to check the coexpression for each gene on the array with SEP3 . For each peak score threshold , the proportion of genes that correlate with SEP3 expression at different control levels ( 0 . 05 , 0 . 01 , 0 . 005 , and 0 . 001 ) among the genes affected by at least one ChIP-SEQ peak was calculated . In a similar way , the proportion of genes that correlate with SEP3 expression among all the genes was calculated . | Most regulatory genes encode transcription factors , which modulate gene expression by binding to regulatory sequences of their target genes . In plants in particular , which genes are directly controlled by these transcription factors , and the molecular mechanisms of target gene recognition in vivo , are still largely unexplored . One of the best-understood developmental processes in plants is flower development . In different combinations , transcription factors of the MADS-box family control the identities of the different types of floral organs: sepals , petals , stamens , and carpels . Here , we present the first genome-wide analysis of binding sites of a MADS-box transcription factor in plants . We show that the MADS-domain protein SEPALLATA3 ( SEP3 ) binds to the regulatory regions of thousands of potential target genes , many of which are also transcription factors . We provide insight into mechanisms of DNA recognition by SEP3 , and suggest roles for other transcription factor families in SEP3 target gene regulation . In addition to effects on genes involved in floral organ identity , our data suggest that SEP3 binds to , and modulates , the transcription of target genes involved in hormonal signaling pathways . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"plant",
"biology"
] | 2009 | Target Genes of the MADS Transcription Factor SEPALLATA3: Integration of Developmental and Hormonal Pathways in the Arabidopsis Flower |
The Polycomb group ( PcG ) and trithorax group ( trxG ) genes play crucial roles in development by regulating expression of homeotic and other genes controlling cell fate . Both groups catalyse modifications of chromatin , particularly histone methylation , leading to epigenetic changes that affect gene activity . The trxG antagonizes the function of PcG genes by activating PcG target genes , and consequently trxG mutants suppress PcG mutant phenotypes . We previously identified the ANTAGONIST OF LIKE HETEROCHROMATIN PROTEIN1 ( ALP1 ) gene as a genetic suppressor of mutants in the Arabidopsis PcG gene LIKE HETEROCHROMATIN PROTEIN1 ( LHP1 ) . Here , we show that ALP1 interacts genetically with several other PcG and trxG components and that it antagonizes PcG silencing . Transcriptional profiling reveals that when PcG activity is compromised numerous target genes are hyper-activated in seedlings and that in most cases this requires ALP1 . Furthermore , when PcG activity is present ALP1 is needed for full activation of several floral homeotic genes that are repressed by the PcG . Strikingly , ALP1 does not encode a known chromatin protein but rather a protein related to PIF/Harbinger class transposases . Phylogenetic analysis indicates that ALP1 is broadly conserved in land plants and likely lost transposase activity and acquired a novel function during angiosperm evolution . Consistent with this , immunoprecipitation and mass spectrometry ( IP-MS ) show that ALP1 associates , in vivo , with core components of POLYCOMB REPRESSIVE COMPLEX 2 ( PRC2 ) , a widely conserved PcG protein complex which functions as a H3K27me3 histone methyltransferase . Furthermore , in reciprocal pulldowns using the histone methyltransferase CURLY LEAF ( CLF ) , we identify not only ALP1 and the core PRC2 components but also plant-specific accessory components including EMBRYONIC FLOWER 1 ( EMF1 ) , a transcriptional repressor previously associated with PRC1-like complexes . Taken together our data suggest that ALP1 inhibits PcG silencing by blocking the interaction of the core PRC2 with accessory components that promote its HMTase activity or its role in inhibiting transcription . ALP1 is the first example of a domesticated transposase acquiring a novel function as a PcG component . The antagonistic interaction of a modified transposase with the PcG machinery is novel and may have arisen as a means for the cognate transposon to evade host surveillance or for the host to exploit features of the transposition machinery beneficial for epigenetic regulation of gene activity .
The Polycomb group ( PcG ) genes are widely conserved in plants and animals and mediate an epigenetic system for repressing transcription of developmental patterning and other target genes . They were originally identified from genetic studies in Drosophila [2] by virtue of their shared role in repressing homeotic genes and subsequently discovered in other organisms , often through a similar role in controlling developmental patterning and mediating epigenetic transcriptional silencing . Although stable , PcG-mediated silencing can be reversed , most commonly between generations during germline or early embryo development but also during somatic development [3] . Two outstanding questions are how does the PcG mediate transcriptional silencing and how is this overturned ? PcG mediated gene silencing is strongly associated with histone methylation , specifically trimethylation of lysine 27 on the amino tail of histone H3 ( H3K27me3 ) [4] . This modification is catalysed by Polycomb Repressive Complex 2 ( PRC2 ) , that comprises four widely conserved PcG proteins , which in Drosophila are Enhancer of zeste [E ( z ) ] , Extra sex combs ( Esc ) , Suppressor of zeste 12 [Su ( z ) 12] and Nurf55 [5 , 6] . In Arabidopsis the different members are represented by small gene families: for example the catalytic subunit E ( z ) is encoded by the three genes MEDEA ( MEA ) , CURLY LEAF ( CLF ) and SWINGER ( SWN ) ; similarly , the Su ( z ) 12 subunit is encoded by the three genes EMBRYONIC FLOWER2 ( EMF2 ) , VERNALIZATION2 ( VRN2 ) and FERTILIZATION INDEPENDENT SEED DEVELOPMENT2 ( FIS2 ) . MEA and FIS2 act specifically in seed , whereas CLF and SWN show overlapping and partially redundant roles in the plant body as do EMF2 and VRN2 [7 , 8] . Although best known as a histone mark “writer” , it has recently emerged that the PRC2 has other activities towards chromatin including as a “reader” of marks . Thus the Esc component can specifically bind H3K27me3 and when bound it stimulates the histone methyltransferase ( HMTase ) activity of PRC2 [9] . By contrast , the Su ( z ) 12 component can bind the antagonistic marks H3K4me3 and H3K36me3 that are associated with active genes , and this can result in downregulation of the HMTase activity of PRC2 [10] . This interplay between reading and writing activities within a single complex likely helps reinforce alternative stable chromatin states marked by active or repressive marks . Whilst the four core components of PRC2 are very widely conserved throughout metazoans and land plants , various accessory components have been identified that are usually more restricted . For example , in animals the DNA binding protein AEBP2 ( JING in Drosophila ) is a component of PRC2 and may have a role in recruiting PRC2 to chromatin marked with mono-ubiquitinated histone H2A ( H2Aub ) and stimulating HMTase activity [11] . In Arabidopsis , the PHD domain containing protein VERNALIZATION INSENSITIVE 3 ( VIN3 ) , and three related proteins VIN3-like 1–3 ( VIL1-3 , also called VRN5 , VEL1 and VEL2 respectively ) can associate with PRC2 and are thought to upregulate HMTase activity [12] . Although H3K27me3 methylation is necessary for silencing by the PcG , it does not directly inhibit transcription and there are several examples where H3K27me3 decorated targets are activated without removal of this mark [13–15] . This suggests that other PcG proteins have more direct roles in transcriptional silencing . Indeed , a second PcG protein complex , Polycomb Repressive Complex 1 ( PRC1 ) has been shown to have several activities on chromatin that inhibit transcription . Firstly , purified PRC1 has several non-covalent activities towards chromatin templates in vitro including inhibiting chromatin remodeling , promoting chromatin compaction and also inhibiting transcription [16–19]; the role of PRC1 in chromatin compaction has also been demonstrated in vivo [13] . The canonical PRC1 contains four proteins , in Drosophila: Polycomb ( Pc ) , Posterior Sex Comb ( Psc ) , Polyhomeotic ( Ph ) and Sex Combs Extra ( Sce ) [19] . A poorly conserved , C-terminal region of Psc is sufficient for all of these non-covalent activities of the PRC1 in silencing , at least in vitro . Secondly , two PRC1 components—Sce and the N-terminal portion of Psc in Drosophila–have RING finger domains with E3 ubiquitin ligase activity and promote H2Aub ubiquitination most notably when in a variant PRC1 complex termed dRAF [20] . The H2Aub modification may inhibit transcription by blocking the recruitment of factors needed for RNA PolII-dependent transcriptional elongation at target gene promoters [21 , 22] . Genetic analysis in which the E3 ligase activity of Sce orthologues was specifically mutated in mouse embryonic stem cells abolished H2Aub in vivo and caused derepression of many PcG targets , confirming that H2Aub is relevant for PcG silencing [23] . However , chromatin compaction and partial repression was maintained at Hox gene targets , suggesting that the two roles of PRC1 in silencing are partially separable . Furthermore , similar experiments in Drosophila have shown that whilst H2Aub is required for viability , it is dispensable for silencing of canonical PcG targets [24] . The PRC1 members are less well conserved in plants than the PRC2 , however similar proteins and activities have been found in Arabidopsis [25] . For example , LIKE HETEROCHROMATIN PROTEIN 1 ( LHP1 ) is equivalent but not homologous to Pc , and like Pc it can bind H3K27me3 via a chromodomain [26] . The plant specific PcG protein EMBRYONIC FLOWER1 ( EMF1 ) is unrelated to Psc but has similar architectural features to the Psc C-terminal region and likely has a similar role in silencing: like Psc it has been shown to inhibit chromatin remodeling and transcription in vitro and it is required for the silencing of many PcG targets in vivo [27 , 28] . The Arabidopsis AtBMI1 and AtRING1 proteins , orthologues of Psc and Sce , respectively , mediate H2Aub and are needed for the silencing of a subset of PcG target genes [29–31] . Some PcG targets that are heavily H3K27me3 methylated such as FUSCA 3 ( FUS3 ) are strongly dependent on AtBMI1/AtRING1 but not EMF1 for transcriptional repression , whereas others such as AGAMOUS ( AG ) require EMF1 but not AtBMI1/AtRING1 suggesting two partially independent pathways by which H3K27me3 methylated genes are silenced [25] . Whether the plant PRC1 members also co-associate in PRC1-like complexes in vivo is unclear . EMF1 and LHP1 co-immunoprecipitate from plant extracts [32] , and both LHP1 and EMF1 interact with AtRING1 and AtBMI1 proteins in in vitro pull down assays [29] . Additionally , EMF1 is required for AtRING1/AtBMI1 mediated H2Aub in vivo [29] . Together , these observations suggest that a PRC1 like complex containing EMF1/LHP1/AtRING1/AtBMI1 may occur in plants . However , LHP1 was also found to co-purify with MSI1 and other PRC2 components when MSI1 was immunoprecipitated from cross-linked protein extracts [33] and EMF1 also interacts with MSI1 in in vitro pull down assays [28] . Thus some PRC1 members may also have roles in PRC2 complexes in plants , or PRC1 and PRC2 complexes may be less distinct . A second group of genes , termed the trithorax group ( trxG ) , acts as antagonists of PcG silencing and promotes the stable activation of PcG targets . A defining genetic property of trxG mutants is that they suppress PcG mutant phenotypes , as they are required for the transcriptional activation of PcG targets [34] . Although the trxG has been less extensively characterized in plants than the PcG , several members have been identified from forward and reverse genetic screens and in several cases have biochemical activities towards chromatin that are opposite to those of the PcG . For example , RELATIVE OF EARLY FLOWERING 6 ( REF6 ) encodes a jumonji domain protein which demethylates the H3K27me3 and H3K27me2 modifications catalyzed by the PRC2 and genetically acts a suppressor of mutants in CLF , encoding the catalytic subunit of PRC2 [35] . ARABIDOPSIS TRITHORAX LIKE 1 ( ATX1 ) and EARLY FLOWERING IN SHORT DAYS ( EFS ) encode HMTases that deposit H3K4me3 and H3K36me3 , respectively , two marks associated with transcriptional activity that are known to inhibit the H3K27me3 HMTase activity of the PRC2 [36 , 37] . The plant-specific ULTRAPETALA1 ( ULT1 ) gene also antagonizes CLF genetically and can activate CLF target genes when overexpressed . The biochemical function of ULT1 is unclear but it has been found to interact with ATX1 and may therefore have a role in promoting H3K4 methylation [38] . The related BRAHMA and SPLAYED genes act as genetic suppressors of clf mutants and are required to overcome PcG repression of floral homeotic genes during flower development . They encode SWI2/SNF2 chromatin remodelers i . e . an activity opposite to that of EMF1 , which inhibits chromatin remodeling [39] . To further identify genes antagonizing PcG silencing we previously carried out a genetic screen for suppressors of the lhp1 mutant phenotype in Arabidopsis and so identified ANTAGONIST OF LIKE HETEROCHROMATIN PROTEIN1 ( ALP1 ) [1] . Here we perform a detailed genetic , molecular and proteomic characterization . We show genetically that ALP1 interacts with various PcG and trxG members and through transcriptional profiling that it is required for activity of the majority of CLF target genes . The relationship of ALP1 with chromatin was previously uncertain , as it was found to encode a domesticated transposase . Using proteomics we show that CLF is associated not just with the core PRC2 members but also with ALP1 , LHP1 and EMF1 in vivo . By contrast , ALP1 associates with the core components of the CLF and SWN containing PRC2 complexes in vivo but not with EMF1 and LHP1 . This suggests that ALP1 may antagonize PRC2 silencing by inhibiting the interaction with EMF1 and or LHP1 . The association of a domesticated transposase with the PcG machinery is novel and raises the question of whether transposases may more generally have evolved roles in inhibiting epigenetic machinery as a way to evade host surveillance and/or the hosts exploited this as a means to regulate the PcG .
ALP1 was first identified in a genetic screen for suppressors of the Arabidopsis PcG mutant lhp1 [1] . Independently , we identified ALP1 in a second genetic screen [described previously in 40] for suppressors of clf mutants . CLF encodes the catalytic component of the PRC2 and acts largely redundantly with the closely related SWN gene [15 , 41] . Similar to the results for alp1 lhp1 plants , alp1 partially suppressed the clf mutant phenotype . The suppression was clearest in short day conditions , where the clf phenotype is milder than in long days; the alp1-4 clf-50 double mutants closely resembled wild type plants with larger , less curled leaves than clf-50 mutants ( Fig 1A ) and were intermediate in flowering time between clf-50 and wild-type in SD ( Fig 1B ) . Molecular analysis indicated that alp1-4 was caused by a T-DNA insertion in the third exon of ALP1 ( S1A Fig ) . To confirm that the suppression of the clf phenotype was caused by alp1-4 rather than any other mutation in the background , we introduced a transgene ( gALP1 ) containing 2 . 9 kb of genomic DNA spanning the ALP1 locus , into the clf-50 alp1-4 background; this complemented the alp1-4 mutation i . e . clf-50 alp1-4 gALP1 plants , unlike clf-50 alp1-4 , showed a severe clf mutant phenotype ( Fig 1C ) . In addition , we found that an independent alp1 mutation isolated in a different genetic background ( alp1-3 in Ler ) also gave a partial suppression of the clf phenotype ( S1B Fig ) . Our previous analysis showed that many mutants which suppress the clf phenotype are very late flowering in both clf and wild-type backgrounds; their suppression is caused by high levels of the FLOWERING LOCUS ( FLC ) gene which represses FLOWERING LOCUS T ( FT ) and other key targets of CLF [40] . To test whether ALP1 also affected flowering , we characterised alp1 mutants in a wild-type ( CLF+ ) background . The alp1-4 single mutant had normal flowers and showed no aberrant morphological phenotype ( Fig 1A ) ; although leaves of alp1-4 occasionally showed weak downward curling we were not able to reliably distinguish mutants from wild-type siblings in segregating populations . The alp1-4 mutants had normal flowering time in long days , and in short days were on average slightly late flowering ( Fig 1B ) , but there was considerable overlap in flowering time between mutant and wild type . Importantly , the suppression of clf by alp1 was not dependent on FLC activity , as although clf-50 alp1-4 flc-3 triple mutants flowered earlier than clf alp1-4 they nonetheless retained a suppressed clf phenotype ( Fig 1D ) . To test whether ALP1 interacts more generally with the plant PcG , we combined alp1 with other PcG mutants . The emf2 and emf1 mutants have a more severe phenotype than clf or lhp1 , but regulate similar targets in the flowering pathway [33 , 42] . However , alp1 mutations did not suppress either emf2 or emf1 mutants ( Fig 1E and 1F ) . In addition , alp1 did not suppress the severe clf swn double mutant , in which PRC2 activity and H3K27me3 methylation is eliminated from plants ( Fig 1G and 1H ) [43] . The MEA gene is closely related to CLF and SWN and is specifically expressed in the central cell of the female gametophyte and the descendant endosperm; in mea/+ heterozygotes about 50% of seeds abort ( those inheriting the defective allele maternally ) [8] . To test whether alp1 can suppress mea mutations , we made alp1-4 mea-emb173/+ plants . Similar to mea-emb173/+ plants , about 50% of seed on these plants were shrivelled ( 1778:1522 shrivelled:non-shrivelled , 54% , see also S2A Fig ) . Thus the alp1 mutation did not obviously suppress the seed abortion phenotype conferred by mea , although we cannot rule out subtle effects on the mea phenotype . Since the genetic suppression of PcG mutants is a defining property of trxG genes , ALP1 may represent a novel plant trxG member: we therefore tested whether alp1 enhanced plant trxG mutants in double mutant combinations . Double mutants of alp1 with mutants in the ATX1 gene did not enhance the atx1 phenotype ( S2B and S2C Fig ) . By contrast , alp1 enhanced the floral phenotype of mutants in both ULT1 and the closely related ULT2 gene , so that the double mutants had more floral organs than ult1 or ult2 single mutants ( S3A , S3B and S4 Figs ) . Since alp1 single mutants have normal floral organ number , alp1 interacts synergistically with ult1 and ult2 . The increased floral organ number in ult mutants is thought to be due to impaired activation of the floral homeotic gene AG , which results in prolonged activity of the WUSCHEL ( WUS ) gene promoting stem cell activity and increasing meristem size and organ number [44 , 45] . ALP1 antagonises CLF , which is a repressor of AG expression . The enhancement of ult mutations by alp1 might therefore occur if ALP1 and ULT act in parallel in activating AG . In addition , alp1-3 strongly enhanced the dwarf , branched phenotype of mutants in the EFS gene encoding an H3K36me3 histone methyltransferase , although alp1 single mutants had normal height and branching ( Fig 1I ) : for example , whereas Ler and alp1-3 had a similar mean total branch number ( 4 . 0 vs 3 . 18 , P>0 . 05 ) , alp1-3 efs had significantly more branches than efs plants ( 34 . 4 vs 24 . 4 , P<0 . 05 ) . In progeny of an efs alp1-3/+ individual we observed 17 plants with the enhanced phenotype and 39 with the less severe phenotype; genotyping showing that 16 of the 17 plants with the severe phenotype were alp1-3 homozygotes whereas all 39 of the less severe were ALP1+ homozygotes or heterozygotes , consistent with alp1-3 causing the enhancement . Finally , we made double mutants in Col-0 background between alp1-4 and an independent efs allele and observed a similar enhancement ( S3C Fig ) . Together , these data indicate a non-additive genetic interaction , consistent with ALP1 acting in parallel with EFS on common targets . In western blot analysis of total histone extracts we found that although the efs mutation reduced global H3K36me3 levels as previously reported [37] , alp1 mutations did not have any effect ( Fig 1J ) , again suggesting the two genes act independently . Collectively , the fact that the effects of alp1 mutation were most apparent in specific PcG and trxG backgrounds suggested that ALP1 has an activity towards chromatin , and that genetically it behaves as a trxG member . We previously reported that ALP1 is plant-specific , conserved in higher plants ( eudicots ) and encodes a 396 amino acid protein with similarity to a tranposase encoded by the PIF/Harbinger superfamily of transposons in plants and animals [1] . PIF/Harbinger transposons encode two proteins , one with DNA binding activity and the other a transposase with DNA endonuclease activity [46–48] , whereas in most other DNA transposon families both activities are combined in a single protein . ALP1 is related to the transposase component , but is unlikely to retain activity as it has non-conservative substitutions ( DGA in place of DDE ) for two of the three acidic residues that comprise a highly conserved catalytic triad involved in metal ion co-ordination at the active site of transposases and other endonucleases [1 , see also Fig 2 and S5 Fig] . To test whether ALP1 is conserved outside eudicots we queried EST sequence databases from monocots , basal angiosperms , gymnosperms , ferns , bryophtyes and green algae . We identified proteins similar to ALP1 in all major land plant groups . All the ALP1-related proteins retrieved ALP1 as the best hit in reciprocal BLAST searches against the Arabidopsis genome , suggesting they were ALP1 homologues . Alignments between land plant ALP1 proteins revealed blocks of conserved sequence in similar positions to the regions previously found to be conserved between PIF/Harbinger transposases ( S5 Fig ) . We previously identified a potential DNA binding motif within a conserved region of ALP1 [residues 108–138 , see 1] . Structural prediction , using the PHYRE program [49] , suggests that an overlapping region ( residues 96–142 ) has a similar structure to the homeodomain of Centromere Protein B , a DNA binding protein also related to transposases . However the sequence similarity between the two proteins in this region is very low ( 17% identity over 47 amino acids ) . We were unable to identify any other motifs associated with chromatin modification or transcriptional regulation . To further analyse the relationship between ALP1-like proteins and PIF/Harbinger transposases we constructed phylogenetic trees based on the aligned protein sequences . This revealed that the land plant ALP1 proteins form a strongly supported group ( bootstrap value 100 ) distinct from that of PIF/Harbinger transposases ( Fig 2 ) . Several other observations further suggested that the ALP1 homologues are unlikely to be components of functional transposons: firstly , they were single copy in virtually all genomes queried , unlike autonomous PIF/Harbinger transposons which typically occur at much higher copy number in plants [50]; secondly , in cases where flanking genomic sequences were available we found that the ALP1 genes lacked the neighbouring gene encoding a DNA binding protein that is characteristic of PIF/Harbinger transposons; finally , comparison of the genomic sequences flanking ALP1 between Arabidopsis thaliana , Arabidopsis lyrata and Populus trichocarpa ( poplar , phylogenetically close to Brassicaceae ) reveals that ALP1 is in a syntenic region in all three genomes ( S6A Fig ) and is therefore immobile . Collectively , these data suggest that ALP1 arose by domestication of a PIF/Harbinger type transposase gene and was present in the common ancestor of all land plants . We further identified a sequence from the green algae Chara braunii with similarity to ALP1 . This was not well resolved in our tree , but occupies an intermediate position between ALP1 and PIF/Harbinger , and may share a more recent common origin with ALP1 than the transposases . Within angiosperms , ALP1 is represented by an ALP1 clade and a sister clade that includes AT3G55350 , the Arabidopsis protein most similar to ALP1 ( Fig 2 ) . The genes in these two clades contain a single intron which is located at an identical position near the 5’-end of the coding sequence , further supporting that they have a recent common origin ( S6B Fig ) . In both clades , one or two of the three residues in the DDE catalytic triad that is conserved in functional transposases have been mutated ( Fig 2 ) . By contrast , in all the land plant groups basal to the angiosperms the DDE triad is conserved ( Fig 2 ) . This suggests that during angiosperm evolution ALP1 lost endonuclease activity and acquired a novel function . The similarity between ALP1 and At3g55350 raised the possibility that the two genes act redundantly . To test this we made double mutants between alp1-1 and a T-DNA insertion allele of At3g55350 ( Salk 122829 ) , however we did not observe any obvious enhancement of the alp1 mutant phenotype ( S1 Fig ) . The suppression of PcG mutant phenotypes by alp1 suggested that ALP1 is required for the activity of PcG targets . One possibility is that ALP1 acts downstream of targets , for example to mediate their activity in conferring leaf curling . To test whether ALP1 is needed for downstream function of AG , a key target of CLF and LHP1 , we introduced the 35S::AG transgene into wild-type and alp1 mutant backgrounds . The 35S::AG transgene confers a strong leaf curling phenotype , similar to that of clf mutants , due to AG mis-expression in leaves [51] . We observed a similar leaf curling phenotype in both wild-type and alp1 backgrounds ( Fig 3A ) suggesting that ALP1 was not required downstream of AG for its activity . To test whether ALP1 is required upstream of PcG targets for their transcriptional activation we first quantified gene expression of the key CLF targets AG , SEPALLATA3 ( SEP3 ) , FT and FLC by real time RT-PCR ( Fig 3B ) . As previously described , expression of all four genes was strongly increased in clf-50 relative to wild-type seedlings . All four genes were less strongly mis-expressed in clf-50 alp1-4 than in clf-50 consistent with ALP1 acting as a transcriptional activator of PcG targets . To test more globally whether ALP1 was needed for PcG target activity , we compared the transcriptomes of wild-type ( Ws ecotype ) , clf-50 , alp1-4 and clf-50 alp1-4 plants seedlings . In comparison to wild-type , more genes were mis-regulated in clf-50 than in clf-50 alp1-4 or alp1-4 mutants , consistent with the more severe phenotype of clf-50 ( Fig 3C ) . More genes were upregulated than downregulated in clf-50 , consistent with the role of CLF as a repressor , and the up-regulated genes included known CLF targets ( Fig 3D , see also S1 Table ) . Strikingly , of the 331 genes up-regulated in clf-50 , the majority ( 73% ) were no longer up-regulated in clf-50 alp1-4 double mutants ( Fig 3E ) . Therefore , ALP1 is generally required for the activation of PcG targets when CLF activity is lacking . In comparisons of alp1-4 with wild-type , more genes were downregulated than up ( Fig 3D ) , suggesting that ALP1 may also have a role as an activator in CLF+ backgrounds . Furthermore , of the 126 genes that are downregulated in alp1-4 , 57 are enriched for H3K27me3 ( based on [52] ) , a much higher fraction than the genome average ( p<2 . 6 x E-16 , hypergeometric test ) , consistent with a role for ALP1 in activating PcG targets . Gene ontology enrichment analysis suggested that the genes downregulated in alp1-4 were enriched for a wide range of biological processes particularly those involved in stress response and disease resistance , in contrast to the genes upregulated in clf-50 which were enriched for ones involved in flower development ( S2A and S2B Table ) . Indeed , when we compared genes downregulated in an alp1 background with those up-regulated in a clf background ( in order to identify common targets oppositely regulated ) the overlap was small ( 4 genes , S2C Table ) but did include a key PcG target , the floral homeotic gene APETALA3 ( AP3 ) . Since the effects of alp1 mutation are most pronounced in the clf mutant background , we also searched for genes that are oppositely regulated by CLF and ALP1 relative to the clf alp1-4 double mutant background . We identified a small but significant overlap of 12 genes which included the floral homeotic genes SHATTERPROOF2 ( SHP2 ) and APETALA3 ( AP3 ) ( S2D Table ) . Although these results suggested that ALP1 might play a role in activation of floral homeotic gene expression , we did not observe floral homeotic defects in alp1 single mutant flowers ( Fig 3F ) . To reveal subtle defects , we removed ALP1 activity in the weak leafy-5 ( lfy-5 ) mutant background , which has reduced transcriptional activation of floral homeotic genes and is especially sensitive to any mutation that further weakens activation [53] . Indeed , alp1-4 strongly enhanced lfy-5 mutations , such that double mutant flowers lacked petals ( Fig 3F and 3G ) ; consistent with the enhanced floral phenotype , transcription of AP3 and PISTILLATA ( PI , like AP3 is required for petal and stamen specification ) was severely reduced in alp1-4 lfy-5 double mutant inflorescences compared to lfy-5 single mutants ( Fig 3H ) . RT PCR suggested that ALP1 was expressed broadly in plants ( S1D Fig ) . To characterise expression further , we made reporters that expressed in-frame fusions of ALP1 with GFP or GUS proteins under control of the native ALP1 promoter . The GFP fusion construct fully complemented the alp1-4 mutation whereas the GUS fusion construct only gave a partial complementation ( Fig 4A ) . Since both constructs had the same ALP1 regulatory sequences , including introns , this suggested that the differences were not in expression but rather in the extent to which the fusions impaired ALP1 protein activity . The pALP1::ALP1-GUS reporter was expressed broadly in leaves , stems , flowers and roots ( Fig 4B–4E ) , with strongest expression in meristems and young leaves similar to the expression of many other plant chromatin regulators ( e . g CLF and SWN ) . The ALP1-GFP fusion was nuclear localised in transgenic plants , consistent with ALP1 functioning as a transcriptional regulator ( Fig 4F and 4G ) , and was also widely expressed ( S1E Fig ) . We also made 35S::ALP1-GFP constructs which complemented alp1-4 , however we did not observe ectopic activation of PcG targets suggesting that ALP1 activity is insufficient to overcome PcG repression . To test whether ALP1 was part of a chromatin-related protein complex , we immunoprecipitated GFP-tagged ALP1 from transgenic plants expressing pALP1::ALP1-GFP or p35S::ALP1-GFP constructs and identified co-purifying proteins by mass spectrometry ( IP-MS ) . Strikingly , in both ALP1-GFP lines the core PRC2 components SWN , CLF , EMF2 , FIE and MSI1 were identified , but not in extracts from 35S::GFP control plants ( Table 1 ) . No trxG components were identified in any of the extracts ( S3 Table ) . We also performed IP-MS on extracts from 35S::GFP-CLF plants ( Table 1 ) . Consistent with the presence of CLF in the ALP1-GFP IP , we found ALP1 in the reciprocal GFP-CLF IP , although we also detected a few ALP1 peptides in 35S::GFP controls in two of three replicates . Furthermore , we identified the core PRC2 complex members FIE , MSI1 , EMF2 , VRN2 and the plant-specific PRC2 accessory components VERNALIZATION5 ( VRN5 ) /VIN3-LIKE1 ( VIL1 ) and VEL1/VIL2 which are thought to boost activity of the HMTase complex [12] . We also found LHP1 which has been variously associated both with PRC1 components and with PRC2 complexes in IP-MS experiments [32 , 33] . We did not identify SWN , suggesting that PRC2 complexes contain either SWN or CLF as the catalytic component but not both together , consistent with the 1:1 stoichiometry of PRC2 components in structural models [54] . Lastly , we identified EMF1 , which has not previously been shown to associate with the PRC2 in vivo but has been strongly implicated as a PcG component based on interaction of EMF1 and MSI1 in vitro and effects of emf1 mutation on H3K27me3 levels in vivo [28 , 42] . Strikingly , neither the PRC2 activators VRN5/VIL1 and VEL1/VIL2 nor the alleged PRC1 components LHP1 and EMF1 were present in either the 35S::ALP1-GFP or the pALP1::ALP1-GFP pull-downs suggesting that their presence is mutually exclusive . To verify the association of ALP1 with the PRC2 in vivo , we performed co-immunoprecipitation ( co-IP ) experiments . To make the co-IP assays and IP-MS independent , we immunoprecipitated extracts from 35S::ALP1-GFP plants using different anti-GFP antibodies from those used in IP-MS and analysed the proteins coimmunoprecipitated with ALP1 using Western blotting . To identify CLF we first generated antibodies to an amino-terminal portion of CLF ( see S1 File ) . The antibodies recognised both CLF-GFP and native CLF in western blots of plant protein extracts although they also cross-reacted with other proteins ( Fig 5A ) . Using these antibodies , we confirmed that CLF was co-immunoprecipitated with ALP1-GFP whereas the cross reacting proteins were not ( Fig 5B ) . In addition we verified that MSI1 is co-immunoprecipitated with ALP1 using a well characterised antibody to MSI1 [55] ( Fig 5C ) . Collectively , these results indicate that ALP1 associates with the PRC2 complex in vivo . Since ALP1 is an activator of PcG targets , it presumably antagonises the function of SWN-PRC2 and/or CLF-PRC2 . The clf-50 mutation is a null allele that carries a deletion of the CLF locus so that clf-50 plants have no CLF protein ( see Fig 5A ) or CLF-PRC2 . The most straightforward explanation for the suppression of the clf phenotype by alp1 mutants is therefore that ALP1 normally inhibits SWN-PRC2 HMTase activity , so that in clf alp1 mutants SWN-PRC2 inhibition is alleviated allowing it to repress key targets such as SEP3 . To test this we performed ChIP assays on chromatin at the SEP3 , AG , AP3 and FLC loci ( Fig 5D ) . In clf-50 H3K27me3 levels were significantly reduced at AG and FLC intron 1 , whereas SEP3 and AP3 were less affected . H3K27me3 levels seemed to be increased in alp1-4 compared to Ws and in alp1-4 clf-50 compared to clf-50 , but the differences were not statistically significant . Hence , the alp1 mutation does not alleviate the clf phenotype by restoring H3K27me3 levels . In addition , alp1 did not affect H3K36me3 , consistent with the immunoblot results ( Figs 1J and 4E ) .
Previous IP-MS experiments using tagged versions of the core PRC2 components EMF2 or MSI1 , or the accessory component VRN5/VIL1 identified PRC2 complexes containing SWN but not CLF as the catalytic unit [12 , 33] . Using tagged CLF we identified EMF2 and VRN2 [Su ( z ) 12 homologues] , MSI1 ( Nurf55 homologue ) , and FIE ( Esc homologue ) , confirming that CLF occurs in both VRN2-PRC2 and EMF2-PRC2 complexes in vivo . The discrepancy between these results from reciprocal pull down experiments might be explained if SWN is more abundant or more stable than CLF in vivo , so that the bulk of EMF2/MSI1 containing complexes have SWN rather than CLF . In addition , we found VRN5/VIL1 and VEL1/VIL2 , two related PHD domain proteins that have also been shown to associate with the core PRC2 in pull downs of MSI1 or of VIL1 itself [12 , 33] . This suggests that VIL1 and/or VIL2 are components of most CLF-PRC2 complexes . No mutant phenotype has been reported for vil2 mutants , whereas vil1 mutants have an impaired vernalization response broadly similar to that of vrn2 mutants indicating that VIL1 is needed for full activity of VRN2-PRC2 [56 , 57] . In the absence of vernalization , vil1 mutants have a very weak phenotype relative to clf mutants , but it is possible that more severe defects are masked by redundancy between VIL1 and VIL2 [56] . Thus VIL1 and VIL2 are likely to be required for the full activity of the CLF-PRC2 . Additional to the core and accessory PRC2 components , CLF also pulled down two proteins—LHP1 and EMF1—generally thought to be in plant PRC1-like complexes [25] . LHP1 and EMF1 are functional equivalents of Drosophila Pc and Psc , and have been found to interact with each other as well as with the plant homologues of the other Drosphila core PRC1 components , namely the AtBMI1 and AtRING1 proteins [29 , 31 , 42] . The fact that CLF pulls down the core PRC2 together with EMF1 and LHP1 does not prove that all are in the same complex , as similar results would be obtained if there are distinct CLF-PRC2 and CLF/EMF1/LHP1 complexes . However , there is additional evidence to support EMF1 and LHP1 associating with the other PRC2 components . Notably , IP-MS experiments using MSI1 identified LHP1 ( but only when cross-linked protein extracts were used ) and LHP1 was found to co-immunoprecipitate both with MSI1 and also with EMF2 [33] . Furthermore , both EMF1 and LHP1 directly interact with MSI1 in in vitro pull down assays [28 , 33] . One possibility is that in plants a PRC1-like complex ( AtRING1/AtBMI1/LHP1/EMF1 ) interacts with a CLF-PRC2 complex via its MSI1 component . This would be consistent with recent proteomic studies using cross-linked extracts , which suggest that in Drosophila the PRC1 and PRC2 complexes can interact via a common bridging component , Sex Combs on Midleg ( Scm ) [58] . Alternatively , EMF1 and LHP1 may participate in distinct complexes , namely a PRC1-like complex ( AtRING1/AtBMI1/EMF1/LHP1 ) with a role in histone ubiquitination ( via AtRING1/AtBMI1 ) and in a PRC2/EMF1/LHP1 complex with a role in histone methylation and transcriptional silencing ( via the EMF1 component ) . The latter scenario is more consistent with the fact that no AtBMI1 or AtRING1 proteins were found in the CLF IP-MS and also with genetic data suggesting that AtRING1 and AtBMI1 genes regulate only a subset of PcG targets in Arabidopsis . Further biochemical purification of plant PcG complexes , together with in vitro reconstitution experiments should help distinguish between these alternatives . Using two different transgenic lines ( expressing ALP1-GFP from the native or the 35S promoter ) and three independent experiments ( effectively six replicates ) we unequivocally identify the core PRC2 components FIE , MSI1 , EMF2 , SWN and CLF as ALP1 partners in vivo . PRC2 complexes contain a single catalytic unit , here either SWN or CLF but never both proteins , as can be seen from the fact that SWN was not found in the CLF IP-MS . Since ALP1 IP-MS retrieves SWN and CLF , ALP1 interacts both with SWN-PRC2 and CLF-PRC2 complexes in vivo . We identified EMF2 but not VRN2 , which may indicate a preference for EMF2 over VRN2 containing complexes however genetic data suggests EMF2-PRC2 is more abundant in the absence of vernalization treatment . Notably , we did not identify any peptides from VIL1 , VIL2 , EMF1 or LHP1 in any of these experiments . Given that all of these were identified with high confidence in all three IP-MS experiments using CLF , we conclude that ALP1 associates with a subset of CLF and SWN-PRC2 complexes that lack VIL1/VIL2/EMF1/LHP1 . Genetically ALP1 has all the hallmarks of a trxG gene . Firstly , alp1 mutants suppresses the phenotype of several PcG mutants . Transcriptional profiling showed that this is because when PcG activity is impaired , ALP1 activity is needed to activate the bulk of the target genes that are normally de-repressed . Secondly , even when PcG are fully active , ALP1 has a role in overcoming PcG repression at some PcG targets . This is revealed by subtle defects in the transcriptional activation of the floral homeotic genes AP3 and PI in alp1 mutants , but also in that a significantly higher proportion of the genes downregulated in the alp1 background are PcG targets than in the genome as a whole . Thirdly , alp1 mutants enhance the phenotype of several trxG mutants including ult1 , ult2 and efs . Interpretation of these synergistic interactions is complicated as there may be substantial genetic redundancy ( for example , EFS is not the only Arabidopsis H3K36 HMTase ) , but the simplest explanation is that ALP1 acts in parallel to ULT1/2 and EFS in opposing PcG repression . The finding that a protein inhibiting PcG silencing is actually a component of CLF and SWN-PRC2 complexes is counter-intuitive . One possibility is that the ALP1 containing PRC2 complexes constitute a small specific fraction of the total PRC2 and occur at situations where PcG repression is being downregulated or over-turned . This is supported by the finding that whereas CLF and SWN are readily detected in ALP1 IP-MS , in the reciprocal experiment involving CLF IP-MS , ALP1 is not greatly enriched over background—in other words , most or all ALP1 occurs in PRC2 complexes , whereas a much smaller fraction of CLF-PRC2 contain ALP1 . A comparable example of an inhibitor interacting with PRC2 was recently described , in which the tumor suppressor BRCA1 interacts with PRC2 in mouse embryonic stem cells and inhibits PRC2 binding to genes involved in cell differentiation , promoting their expression [59] . Under our growth conditions , alp1 single mutants did not show major developmental phenotypes , and did not affect the expression of most of the genes mis-regulated in clf mutants . Thus ALP1 regulates a small subset of PcG targets under laboratory conditions . However , it was notable that the genes that were downregulated in alp1 were enriched for functions in disease resistance and stress response . An intriguing possibility is that ALP1 may be required to overcome PcG silencing of genes involved in stress or disease , and therefore alp1 mutants may show more severe mutant phenotypes under other growth conditions closer to natural environments . It is notable that alp1 mutations can only suppress relatively weak PcG mutants ( lhp1 and clf ) in which PRC2 activity is impaired but not abolished . clf swn mutants , which lack all sporophytic PRC2 activity , were not rescued by alp1 , implying that PRC2 activity is needed for rescue . The simplest explanation for the suppression of clf and lhp1 by ALP1 is that ALP1 inhibits the HMTase activity of the CLF-PRC2 and SWN-PRC2 . Indeed , by blocking the association with accessory components such as VIL1 and VIL2 , ALP1 is likely to reduce HMTase activity . In clf mutants H3K27me3 levels are reduced at some targets , but if the HMTase activity of SWN-PRC2 ( or in lhp1 mutant backgrounds , both CLF-PRC2 and SWN-PRC2 ) is upregulated when ALP1 activity is withdrawn normal H3K27me3 levels and silencing might be restored ( Fig 6A–6C ) . Additionally , if ALP1 possesses DNA binding activity it may inhibit silencing indirectly by luring the PRC2 away from PcG targets to other sites in the genome . Against these scenarios , our H3K27me3 ChIP experiments did not support an increase in H3K27me3 at PcG targets in alp1 mutants . Given that alp1 mutants give a weak rescue of PcG mutant phenotypes , and that subtle effects on H3K27me3 levels may only be visible in dividing cells rather than whole seedlings [e . g . see 33] we can’t exclude that ALP1 inhibits PRC2 HMTase activity and it will be important to test the effects of ALP1 on PRC2 catalytic activity in vitro . An alternative possibility is that ALP1 acts by inhibiting a function of the PRC2 independent of its H3K27me3 HMTase activity , for example a direct role in silencing transcription . Notably , we found that CLF-PRC2 but not ALP1-PRC2 associates with EMF1 , a protein playing a similar role to the Drosophila PcG protein Psc in inhibiting chromatin remodeling [27] and transcription in vitro . If ALP1 competes with EMF1 for CLF- and SWN-PRC2 , then removing ALP1 activity might restore silencing by increasing EMF1 occupancy at CLF targets ( Fig 6D ) . Indeed , alp1 did not rescue emf1-1 mutants , suggesting that EMF1 activity is important for ALP1 to rescue PcG . Although there are numerous examples of genes which have arisen by domestication of transposases [60] , to our knowledge this is the first case where a domesticated transposase has become an inhibitory component of the host core epigenetic machinery . Autonomous PIF/Harbinger transposons are known to mobilise a group of small non-autonomous transposable elements ( specifically the Tourist class of Miniature Inverted repeat Transposable Elements [MITEs] ) that have proliferated massively within plant genomes—for example , there are around 90 , 000 MITEs in the rice genome , comprising the bulk of the transposon content [50] . A further characteristic feature of MITEs is that they have a strong preference to insert into single copy , euchromatic regions of the genome [46 , 50] . Plant hosts typically inactivate transposons by siRNA mediated DNA methylation , which silences expression of their transposase [for review see 61] . In several cases , transposons have been shown to encode proteins that inhibit the host machinery mediating their methylation [62 , 63] , rather as plant viruses encode anti-silencing proteins that interfere with the siRNA machinery . Although PcG silencing is typically thought of in terms of developmental target genes , it also serves a backup function in silencing transposons when DNA methylation is compromised . Thus , in met1 mutants where CG DNA methylation is severely reduced , there is a massive relocation of H3K27me3 onto transposons [64] . Similarly , in endosperm tissue where DNA methylation levels are generally low , transposons are frequently H3K27me3 methylated and this contributes to their transcriptional silencing [65] . Furthermore , studies using the unicellular green alga Chlamydomonas rheinhardtii suggest that the ancestral role of PRC2 in may have been in silencing transposons and other repetitive elements [66] . One possibility therefore is that the association of a transposase with the PRC2 originally evolved as a way for PIF/Harbinger transposons to evade host surveillance and promote their own proliferation . This would be particularly effective if PIF/Harbinger transposons have also evolved means to inhibit host RNA-directed DNA methylation systems . An association with PcG would also benefit the tranposons by targetting them to euchromatic gene rich regions of the genome , where it may be difficult for the host to permanently silence the transpsoson due to effects on expression of neighbouring genes . Alternatively , the association of a domesticated transposase with the PRC2 has arisen because it benefits the host . Given that ALP1 is an ancient gene in land plants , it is highly unlikely that it would have been conserved if it functioned solely to promote PIF/Harbinger transposon proliferation . This would require ALP1 to be part of an active transposon , able to proliferate faster than its hosts could eliminate it , whereas ALP1 is 1–2 copy and immobile . In many cases where transposases have been domesticated , the DNA binding property of the transposon has been conserved , rather than the endonuclease activity [60] . However , in PIF/Harbinger this activity is encoded by a second gene which encodes a Myb class DNA binding protein that is necessary for transposition and has been shown to bind DNA sequences at the tranposon ends and to interact with the nuclease protein to form a functional transposase [67 , 68] . ALP1 is unlikely to retain nuclease activity , as studies expressing the rice PIF/Harbinger class transposon PING in a heterologous system have demonstrated that mutating just one of the three residues in the DDE triad drastically reduces its ability to catalyse transposition [69] . However , it is possible that it retains the ability to interact with a Myb class DNA binding protein and this is useful for targetting PcG to its targets . This would be comparable to vertebrates , where the nuclease of Harbinger has been domesticated to produce the Harbi1 gene and the Myb gene to produce the Naif1 . Although the biological function of these genes is unknown , the HARBI1 and NAIF1 proteins are able to interact [67] . A role for Myb proteins in PcG recruitment to targets in plants has also been demonstrated [70] . A role for ALP1 in recruitment does not however explain why it antagonises PcG silencing . A notable feature of transposons is that they are often activated during stress—for example , in Arabidopsis several retrotransposons are activated by heat shock treatments[71 , 72] , and in blood orange varieties , anthocyanin production is stimulated by a cold-inducible retrotransposon inserted upstream of the RUBY gene [73] . The inhibitory interaction of ALP1 with the PcG might have arisen as a way for the plant to promote the activation of PcG target genes involved in stress response . This would be consistent with the fact that ALP1 targets are enriched for genes involved in biotic and abiotic stress response . By disabling the nuclease activity of the transposase , the plant host may also have limited the side effect of promoting transposon proliferation . It will be interesting in future to test the role of ALP1 in transposon mobilisation and stress response , and also to see whether the vertebrate HARBI1 and/or NAIF domesticates have any role in epigenetic control by PcG or DNA methylation . | Transposons are parasitic genetic elements that proliferate within their hosts’ genomes . Because rampant transposition is usually deleterious , hosts have evolved ways to inhibit the activity of transposons . In plants , this genome defence is provided by the Polycomb group ( PcG ) proteins and/or the DNA methylation machinery , which repress the transcription of transposase genes . We identified the Arabidopsis ALP1 gene through its role in opposing gene silencing mediated by PcG genes . ALP1 is an ancient gene in land plants and has evolved from a domesticated transposase . Unexpectedly , we find that the ALP1 protein is present in a conserved complex of PcG proteins that inhibit transcription by methylating the histone proteins that package DNA . ALP1 likely inhibits the activity of this PcG complex by blocking its interaction with accessory proteins that stimulate its activity . We suggest that the inhibition of the PcG by a transposase may originally have evolved as a means for transposons to evade surveillance by their hosts , and that subsequently hosts may have exploited this as a means to regulate PcG activity . Our work illustrates how transposons can be friend or fiend , and raises the question of whether other transposases will also be found to inhibit their host’s regulatory machinery . | [
"Abstract",
"Introduction",
"Results",
"Discussion"
] | [] | 2015 | Kicking against the PRCs – A Domesticated Transposase Antagonises Silencing Mediated by Polycomb Group Proteins and Is an Accessory Component of Polycomb Repressive Complex 2 |
Animal coordinated movement interactions are commonly explained by assuming unspecified social forces of attraction , repulsion and alignment with parameters drawn from observed movement data . Here we propose and test a biologically realistic and quantifiable biosonar movement interaction mechanism for echolocating bats based on spatial perceptual bias , i . e . actual sound field , a reaction delay , and observed motor constraints in speed and acceleration . We found that foraging pairs of bats flying over a water surface swapped leader-follower roles and performed chases or coordinated manoeuvres by copying the heading a nearby individual has had up to 500 ms earlier . Our proposed mechanism based on the interplay between sensory-motor constraints and delayed alignment was able to recreate the observed spatial actor-reactor patterns . Remarkably , when we varied model parameters ( response delay , hearing threshold and echolocation directionality ) beyond those observed in nature , the spatio-temporal interaction patterns created by the model only recreated the observed interactions , i . e . chases , and best matched the observed spatial patterns for just those response delays , hearing thresholds and echolocation directionalities found to be used by bats . This supports the validity of our sensory ecology approach of movement coordination , where interacting bats localise each other by active echolocation rather than eavesdropping .
Group movement patterns are dependent on perceptual inputs , governed by cognitive mechanisms , and constrained by the motor abilities of the interacting agents [1–4] . The intrinsic sequential occurrence of these three stages gives rise to delayed movement responses [5] . Current collective movement interpretations acknowledge such delays [6–9] , but to explain observed patterns the existence of generic , albeit plausible , interaction rules are postulated . Since individual decisions depend on the information about the actions of other conspecifics [10–14] and movement response are shaped by an animal’s sensory-motor abilities , it is however biologically most realistic to explain interaction exclusively from perception biases and delayed responses without inferring social forces from the observed movements . Even though various models are capable of capturing many of the emerging features of coordinated movement patterns , the individual behavioural rules are heuristics of the underlying mechanisms . There is awareness that more biologically oriented experimental analysis and modelling is necessary [15] . Recent experimental studies on golden shiners ( Notemigonus crysoleucas ) [16] have indicated that speed regulation along the direction of travel appears a more dominant component of the interaction compared to alignment , whereas studies in mosquitofish reported the lack of any justifiable alignment rule [17] . From three-dimensional stereographic images of starling flocks [18] , evidence of a topological rather than a metric interaction rule was brought to the fore , i . e . individuals interacted with an average of six or seven neighbours rather than every conspecific within a certain distance . Other network topologies have also been proposed recently based on animal communication across a ‘Voronoi shell’ of nearest-neighbours [19] or the network of visible neighbours estimated from a ray-casting algorithm around a fish eye [20] . Other investigations have also questioned the assumptions about how individuals perceive and integrate information from their neighbours . Experiments on saithe ( Pollachius virens ) pointed to the importance of the use of lateral line compared to vision on school structure and dynamics [21] . All these recent efforts point to the importance of sensory inputs that shape an animal’s movement decisions . Animal species that use an active sensory modality , such as echolocation , offer an opportunity to tap directly into the source of this sensory information . Active-sensing , compared to passive , requires the emission of a signal whose spatial propagation is governed by quantifiable physical laws [22] . Therefore , once the call intensity and echolocation directionality of an echolocating species are known , it is possible to infer with sufficient accuracy over what spatial range individuals can detect and track each other , either by active sensing or , over greater distances , by eavesdropping on the others’ signals [23–25] . Here we introduce and test the biosonar movement interaction ( BSMI ) hypothesis , whereby bats interact with conspecifics , to perform coordinated flight , based on the information collected by their biosonar and not by eavesdropping . While foraging , bats send out sonar search pulses several times per second . Whenever a sound pulse hits an object an echo is generated , which then travels back to the bat who interprets the echo for target detection , recognition and localisation . Because real objects do not reflect all the impinging sound back to the echolocator ( target strength < 1 ) and because the sound has to travel the bat-object distance twice incurring double propagation losses , the biosonar survey range is limited to several meters at best ( see e . g . [24 , 26 , 27] ) . The spatial profile of biosonar is therefore determined by the amplitude of the biosonar pulses and by the directionality of sound emission as well as the directionality of bat hearing , where the biosonar detection range is greatest on axis and drops off towards either side . The core of the BSMI hypothesis is thus that the directional spatial profiles of the individuals’ biosonar fields determine the animals’ mutual positions and headings during movement interactions , with the information obtained via eavesdropping having a minor role in movement coordination . Note that by eavesdropping we mean locating conspecifics by listening to their echolocation calls [23] and not locating oneself by trailing a conspecific and extracting information from the echoes resulting from their calls [28] . Sound field emission profiles of echolocating bats have been documented in various species [29] including Daubenton’s bat , Myotis daubentonii , our study animal . This widespread medium-sized species [30] has been selected for two reasons . On one hand , the animal’s cognitive abilities allow for rich behavioural patterns . On the other hand , its habit to forage low over still water surfaces to glean arthropod prey [31] reduces the spatio-temporal complexity of the interactions to just a two-dimensional plane . To test our BSMI hypothesis we collected movement trajectories of Daubenton’s bat pairs foraging freely in the wild . Based on these observations , we proceeded in three steps: ( i ) , ( ii ) , and ( iii ) . First , we built a behavioural classifier ( i ) based solely on the paired movement paths , which segments these movement trajectories into different behaviours , and determines actor-reactor roles by extracting response delays . We then tested the BSMI hypothesis by comparing the relative positions of the interacting bats with the calculated biosonar perception field ( ii ) . Finally , we constructed and tested a BSMI model of interacting bats ( iii ) whereby the mechanisms of interaction are based exclusively on aerodynamic constraints . i . e . actual speed and acceleration , and sensory inputs , i . e . the known spatial extent of the sonar perception field . The quality of the BSMI model is judged against its ability to reproduce the observed spatio-temporal interaction dynamics of the bats .
We recorded flight interactions of foraging Daubenton’s bats flying in the South-West corner of Barrow Tanks reservoir number three near Barrow Gurney in England . The experimental set-up allowed for a 20 ms temporal resolution of flight tracking ( see Methods section ) . The recordings were taken over 10 evenings in the summer of 2009: June 15th , 22nd , 29th; July 1st , 7th , 8th , 9th , 15th; and August 19th and 25th . Overall we collected nearly 70 , 000 individual data points ( fixes ) including 70 interactions as displayed in Table 1 . To identify interacting flight behaviour we used two mathematical tools: the time-dependent delayed correlation ( TDDC ) and the time-dependent delayed separation ( TDDS ) . For a given point in time , the TDDC compares the heading of one of the bats with the heading of the other at all other times , thus allowing us to establish whether one bat was copying the direction of the other ( coordinated flight ) . Similarly the TDDS compares the position of one of the bats at one time with the position of the other bat at all other times , thus allowing us to specify if one bat was copying the position of the other ( chase flight ) . The remaining data , when no indication of interaction emerged , were deemed unclassified . In Fig . 1A we show a sample trajectory ( shown also in S1 Video ) , whose reaction delays have been extracted from the time-dependent delay directional correlation ( TDDC ) plot displayed in Fig . 1B—the corresponding time-dependent delay separation ( TDDS ) plot is not shown in this example as it reveals absence of chase behaviour . The presence of a black dot with coordinates ( t , τ ) in the Ci , j ( t , τ ) plot shows that bat j has copied the heading of bat i with a delay τ at time t . A straight black line segment indicates that the bats have maintained a constant reaction delay over a certain period of time , whereas an inclined line , e . g . in the intervals 3 . 14 s < t < 3 . 20 s and 3 . 50 s < t < 3 . 88 s , illustrates an interaction with changing delay . If the extracted delay turns negative , bat i now copies the heading of bat j and actor-reactor roles have swapped . In this sample trajectory , this swap has occurred in two stages . At t = 1 . 4 s bat i first terminates the nearly parallel flight it was maintaining with bat j , and subsequently , at t = 1 . 8 s , tries to align its heading with that of bat j ( see caption of Fig . 1 for more details ) . We analysed those flights which had more than one bat in the field of view of the camera at one time . Our analysis allowed us to classify each recorded data point from all seventy paired trajectories as non-interacting ( 74% ) and interacting behaviour ( 26% ) , with the latter further subdivided into chases ( 4% ) and coordinated flights ( 22% ) . Non-interacting is the definition given to flight segments where there is low correlation between the headings of the individuals . Coordinated flight is defined as a flight segment with high correlation between the headings of the individuals , and chase flight is a flight segment where one individual occupies the previous positions of another . Note that chase flights are a subset of coordinated flights . As shown in Fig . 1C chase flights and most coordinated flights were characterized by similar response delay values of up to 500 ms , which corresponds to 5–7 biosonar calls , i . e . information updates , at the known average call interval of 70–100 ms [32] . One marked distinction between chases and coordinated manoeuvres was in their leader-follower dynamics , synonym here of actor-reactor . During chase flights , the individual ahead was clearly the leader [33] , as the follower was retracing the movement path of the individual in front . This leader-follower hierarchy was not equally prominent when bats were performing coordinated flights , for example the reactor was not necessarily behind the actor ( see e . g . S1 Fig . panel ( f ) ) . In summary to realise step ( i ) in the validation of our BSMI hypothesis , we reconstructed movement trajectories from videogrammetry data , analysed the corresponding TDDC and TDDS plots , and extracted the associated delay values . This allowed us to distinguish movement paths among non-interacting flights , chases and coordinated flights . For the last two categories , the sign of the extracted delay values provided the sought actor-reactor roles as a function of time for each interacting paired trajectory ( see S1 Fig . panels ( a ) - ( f ) for a graphical representation of the 6 steps of this procedure ) . To quantify alignment dynamics of the interacting bats , in Fig . 2 we plot bat separation distance , relative heading and relative position [19] . Chases were characterised by low separation distances ( Fig . 2B ) , high alignment ( Fig . 2C ) and narrow frontal exposure angles ( Fig . 2D ) , while coordinated flights occurred over larger separations , wider range of relative headings and only a moderate preference for frontal exposure angles ( see S2 Fig . for the difference in the TDDS and TDDC plots between a coordinated flight and a chase ) . Unclassified data showed no preference in relative headings and exposure angles , yet comparison with randomised pairing of trajectories ( dashed lines ) revealed that unclassified data points were overrepresented at separations below 3 . 5 m and underrepresented above approximately 7 m . This indicated that even non-interacting bats tended to fly closer than random . With the identified actor-reactor roles , it became possible to superimpose the bat’s echolocation field and test whether a plausible explanation for the observed interacting flights is the existence of an acoustic response threshold—step ( ii ) in our BSMI hypothesis . A match between the actual spatial patterns of interactions with these superimposed sound fields would indicate a potential functional link . In Fig . 2A , where we plotted the actor’s position relative to the heading and location of the reactor , we added isocontours of the calculated biosonar field , i . e . the amplitude of the echo reflected from the actor’s body returning to the reactor . We noticed that 99% of observations were above 0 dB and 97% above 10 dB , i . e . almost no reaction occurred when the actor’s echo was below the likely range for the reactor’s hearing threshold of 0–10 dB [34] . In short , there was no reaction when the actor was outside the reactor’s echolocation field . Step ( iii ) in our BSMI hypothesis consists of investigating whether alignment dynamics with delay in combination with the perceptual bias created by the shape of the echolocation field suffice to produce the observed behavioural interactions . For that purpose we constructed a mathematical movement model ( BSMI model ) of echolocating bat pairs capable of generating interacting trajectories that mimic the patterns of chases and coordinated flights observed in the field data ( see Methods ) . From the spatio-temporal patterns emerging in the BSMI model , we generated relative position plots in Fig . 3 for different delay response τ , directionalities of the echolocation field—asymmetry parameter A in Equation ( 11 ) below—and hearing threshold—parameter B in Equation ( 12 ) below . By varying these parameters in the model we produced actor-reactor and chaser-chasee spatial relationships that can be compared to the observations in Fig . 2A . We found that the best fit was achieved for response delays between 100 ms and 500 ms , and with B = 10 ± 5 dB and A = 16 ± 1 ( root mean standard error RMS = 0 . 0179 ) , see Fig . 4 . We used this latter value to calculate sound fields displayed in Figs . 2A and 3 .
Movement analysis has a long tradition in animal ecology [35 , 36] , however , the segmentation and classification of trajectories of co-moving individuals is an area of research still in its infancy [37] . Here we have shown that a use of the velocity and separation delayed correlation maps , respectively the TDDC and TDDS functions , allowed us to classify interaction patterns in the movement trajectories of coflying bat pairs . Considering only those reactions that are strictly ordered in time , it is possible to reconstruct the delay with which one bat copies the heading of the other bat as a function of time . With our automatic classifier we have found that bats foraging in the field move in coordination five times more often than chasing the bat in front . Although a chaser has a spatial advantage when either the leader fails to capture a prey item ( 2nd attacker advantage ) or when a patch of resources is found ( shared exploitation ) , a chase might also become an aggressive behaviour if the chaser directs unpleasant or even harmful loud signals at the chasee . On the other hand , coordinated flight could improve search efficiency because both individuals hear when and where the other has searched for and found prey , which can allow each member of the pair to survey larger areas of the water surface per unit of time through eavesdropping . Empirical evidence strongly suggests that active echolocation was used by the reacting individual for alignment during interactions: 99% of all interactions were within the echolocation field and the entire echolocation field down to echo amplitudes of 0–10 dB was used ( Fig . 2A ) . The possibility that eavesdropping per se might have produced the observed patterns was tenuous . We reached that conclusion by drawing the reactor’s position relative to the actor ( see S3 Fig . ) . In that case no match of spatial interaction patterns with the received sound appeared obvious as interactions did not follow any sound exposure isocontour and most interactions happened in a direction opposite to the main lobe of emission . However , the fact that unclassified , i . e . non-interacting , bats flew at closer than random distances ( Fig . 2B ) , often outside the maximum calculated echolocation range but within eavesdropping range ( S3 Fig . ) , suggests the idea that it is beneficial for a bat to remain in the eavesdropping region of others , thereby possibly profiting from overhearing the other’s search success . Eavesdropping might also account for the low number of observations above -∣135∣° relative heading ( Fig . 2C ) in unclassified flights , which indicates that even bats outside mutual echolocation range avoid head-on flights . In contrast to other species , where alignment occurs by copying the heading of the individual in front , e . g . in Surf scoters [38] , we observed several coordinated flights where the reactor was in front of the actor ( Fig . 2A and 2D when ∣θij∣ > 90° ) . Although our dataset is limited in size , these large exposure angles indicate that bats may be able to respond to an individual behind them because echolocation can provide information from rear angles e . g . in Fig . 2A an interaction , represented by a row of green points , tracing just outside the -40 dB contour from a position of 3 o’clock ( lateral ) to 6 o’clock ( directly behind ) . Also note that in contrast to other movement coordination studies foraging Daubenton’s bats search for ephemeral and depletable prey and therefore their direction of motion is not targeted at one specific location of a shareable resource . As a result their respective position and orientation could change considerably facilitating actor-reactor swaps within an interaction . Our BSMI model was able to create interactive behaviours ( chases and coordinated flights ) given that the model is based on biosonar perception with no input from movement interaction data other than observed flight speeds and turning angle distributions . Most importantly , the model output can reproduce the observed spatial interaction patterns only when the emission directionality A , hearing threshold B and delay were at their natural values . The best fit directionality value A = 16 is in the middle of the known range of emission directionality for this species ( A = 7 . 3 for an independent measurement shown in S4 Fig . , and A = 25 . 7 [29] ) , and the hearing threshold of B = 10 dB is also within the expected range [34] . In this study we have provided an answer to how individuals react to each other while they are foraging together , but it was not possible to determine why . As at least some data were collected at dates after juvenile bats had taken flight , it is not entirely impossible that some flight interactions were training flights between mothers and their offspring . Increased foraging success might be a driver for coordinating flight manoeuvres . This might be possible because bats that coordinate their movement at short distance have each other survey zones overlap and may thus enter into coordinated exploitations in presence of locally rich prey patches . Alternatively , with small prey density a bat may enter agonistic interactions forcing conspecifics away to monopolise a diminishing resource patch . The chases we observed might well contain such an agonistic component [39] . With the help of collective movement models of delayed interaction [40] future research could address some of these issues by testing responses of conspecifics to the locations and timings of foraging success of others , which they inadvertently indicate by their feeding buzzes . In conclusion , we corroborate our BSMI hypothesis of foraging bats by ( i ) classifying bat interactions in the wild from highly resolved trajectory data , ( ii ) explaining the individuals’ relative positions based on the calculated biosonar perception field , and ( iii ) fitting the experimental observations with a bio-inspired movement model of interacting bats . In the process we have constructed a procedure to identify actor and reactor based not on their relative distance but on the delay dynamics of their relative orientation ( defined through relative headings and exposure angles ) , which is independent of the animal species studied and can be applied to all types of coordinated movement data . The emerging actor-reactor relations and the associated delays with which an animal copies the heading of a conspecific provides a moving reference frame with which to determine when , where and how individuals respond to the actions of a conspecific and represents a powerful tool in the arsenal of computational techniques for movement ecology [41 , 42] .
Animal movement trajectories were obtained by videogrammetry using two 525EX CCIR ( Watec Inc . , Newburgh , NY ) with Cinegon 1 . 8/4 . 8 IR lens ( Jos . Schneider Optische Werke GmbH , Bad Kreuznach , Germany ) on tripods at heights between 4 . 13 and 5 . 66 m above the water surface as measured with Laser rangefinder DLE50 professional ( Robert Bosch GmbH , Gerlingen , Germany ) . Pitch towards the water surface and lateral roll angles of the cameras were measured using Clinometer GeoMaxiclin ( Geo Supplies Ltd , Sheffield , UK ) to the nearest half degree . The recording area of approximately 900 m2 was illuminated with 850 nm IR light by a Raymax 200 ( Raytec , Ashington , UK ) positioned between the cameras at the shoreline . The video outputs were recorded synchronously on one XM-DVR PRO S153 ( Datatoys , Mequon , MI ) solid state dual-channel video recorder . Videos with bat flights were selected manually . Bat positions in the videos were selected by hand in every frame and the corresponding spatial positions calculated in custom written scripts ( M . Holderied and H . Goerlitz ) in MATLAB ( The MathWorks Inc . , Natick , MA ) . Each camera was calibrated and pixel based ray bundles for each camera were obtained using VMS software ( Mark Shortis and Stuart Robson ) . Fields of view of both cameras did overlap only slightly and bat positions were not triangulated by stereo-imaging but rather by intersection of the pixel bundles originating at position , height and orientation of the single respective camera with a 2D plane 20 cm above the water surface , which was taken as the average flight height of this species [43] . The slight left-right asymmetry in the dataset indicates a preference for clockwise turns and a slight tendency to move parallel to the shoreline ( S5 Fig . ) . This resulted in fewer trajectories moving away from the cameras , and thus making the ( projected ) reconstruction of the flight paths with similar level of errors . Videos were recorded in VGA format at 25 Hz frame rate . Each VGA frame is based on two half frames ( odd and even lines taken in alternation ) . We analysed each half frame independently and interpolated the respectively missing odd or even lines such that the effective rate was 50 half frames per second , allowing for a temporal resolution of flight tracking equal to Δt = 20 ms . This is about a factor four higher than the average call interval , i . e . the information update rate of Daubenton’s bats [32] . The dataset of paired bat trajectories was analysed to determine the fraction of data where interactions had occurred . For those data deemed to represent interaction we identified leader and follower dynamics in pairs of co-flying bats based on the sign of the automatically extracted delay values as a function of time with the help of the time-dependent directional correlation ( TDDC ) function , shown in Equation ( 2 ) below , and the time-dependent distance separation ( TDDS ) function , displayed below in Equation ( 4 ) . Analysis of the TDDC and TDDS plots allowed us to segment paired movement trajectories into different behaviours . An agent based model was created to demonstrate the effects of asymmetric perception and reaction delays on the spatio-temporal trajectories of simulated bat pairs . The model contains the minimal ingredients necessary to account for the salient sensory biology features of the bats and their interaction . The individual movement statistics of a bat follows that of a correlated random walker [46 , 47] with turning angles and speed values obtained from the observed ones . Each animal is limited in its manoeuvrability by aerodynamic constraints and tries to align to a conspecific present within its interaction field . Each bat echolocates every 100 ms with a certain hearing and emission directionality and detects the presence of the other animal only when above a certain threshold . When another bat is detected , an individual responds by copying the other’s heading with some delay ( see Equations ( 8–12 ) and S9 Fig . for further details ) . The simulations were run in discrete time , for 10 s and Δt = 20 ms corresponding to the resolution of the observations , and continuous space by generating initial random positions and headings inside a region of 30 × 30 m2 . | Collective movements of flocking birds or shoaling fish are amongst the most fascinating natural phenomena , and everyone has experienced the challenges of walking through a moving crowd . What information individuals use for movement coordination is , however , very difficult to know , except for echolocating bats . These flying mammals perceive their surroundings by emitting loud and high-pitched biosonar calls and listening for the returning echoes . Because bat biosonar imaging is much sparser in information than vision , we can accurately measure the biosonar calls of interacting bats with a group of microphones and then calculate what each of the individuals perceived . When observing pairs of Daubenton’s bats foraging low over water for stranded insects , we found they have intriguing ‘traffic rules’—they chase each other , perform tandem turns and even slow down to avoid collision . When we then modelled their biosonar view of the surroundings during these interactions we discovered that one simple trick suffices to create all their interactive behaviours: once another individual is close enough for your biosonar to pick up its echo , copy this individual’s flight direction within 4–5 of your own wingbeats . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2015 | Delayed Response and Biosonar Perception Explain Movement Coordination in Trawling Bats |
Trachoma is chronic kerato conjunctivitis , which is caused by repeated infection with Chlamydia trachomatis bacterium . It is hyper endemic in many rural areas of Ethiopia . The objective of this study was to measure the effect of water , sanitation and hygiene interventions on active trachoma in selected woredas of North and South Wollo zones of Amhara Region , Ethiopia . A community based quasi-experimental study was conducted from October 2014 to December 2015 among children aged 1–8 years at baseline and among one year older same children after intervention . A four-stage random cluster-sampling technique was employed to select study participants . From each selected household , one child was clinically assessed for active trachoma . Structured questionnaire was used to collect socio demographic and behavioral data . MacNemar test was applied to compare the prevalence of active trachoma between baseline and after the intervention period at both intervention and non-intervention study areas . The prevalence of active trachoma was reduced from baseline prevalence of 26% to 18% after one-year intervention period in the intervention woredas ( P≤0 . 001 ) . MacNemar test result showed significant reduction of active trachoma prevalence after the intervention period in the intervention woredas compared to the non-intervention woredas ( P≤0 . 001 ) . Water , sanitation and hygiene related activities were significantly improved after the intervention period in the intervention woredas ( P<0 . 05 ) . There was a significant reduction of active trachoma prevalence between the baseline and after the intervention period in the intervention woredas , but not in the non-intervention ones . Improved water , sanitation and hygiene interventions contributed to the reduction of active trachoma . However , the magnitude of active trachoma prevalence observed after the intervention is still very high in the studied areas of North and South Wollo Zones communities . To achieve the global trachoma elimination target by the year 2020 as set by the WHO , continued WaSH interventions and periodic monitoring , evaluation and reporting of the impact of WaSH on active trachoma is warranted .
Trachoma is a chronic kerato conjunctivitis , which is caused by repeated infection with Chlamydia trachomatis bacterium . The disease is often transmitted by fomites ( objects or materials that are likely to carry infectious organisms ) , contaminated fingers with discharge from the eyes or nose of an infected person , and through eye-seeking fly Musca sorbens [1] . Trachoma is the leading cause of infectious blindness in the world [2] . Globally , trachoma is responsible for the visual impairment of about 2 . 2 million people , of whom 1 . 2 million became irreversibly blind [3] . Ethiopia is one of the five countries in the world where half of the global burden of active trachoma is concentrated [4] . Of the ten National Regional States in Ethiopia , the Amhara National Regional State ( ANRS ) is disproportionately affected by trachoma [4] . A former study indicated a 52% and 13% overall prevalence of trachomatis inflammation-follicular ( TF ) among children aged 1–9 years in North and South Wollo zones of Ethiopia , respectively [5] . Evidence show that some improvements in the control of trachoma have been brought through socioeconomic development and control programs along with other disease control programs in many countries . However , trachoma continues to be hyperendemic in many of the poorest rural areas of the world , especially in areas that have limited access to water and sanitation [2] . The prevention , control and eventual elimination of trachoma depend heavily on the availability of improved Wash programs in endemic communities [6] . As antibiotic treatment alone will not break the cycle of transmission , improvements of WaSH infrastructure and appropriate health seeking behavior of patients are essential to achieving sustained control , elimination , or eradication of trachoma [7–9] . To achieve the elimination goal of blinding trachoma as a public health problem , the World Health Organization ( WHO ) endorsed the SAFE strategy ( Surgery for correction of trichiasis; Antibiotics to treat infection; Facial cleanliness and Environmental improvement to reduce chlamydia transmission ) have been implemented by different stakeholders in ANRS for the last 10 years . As reported by the Carter Center , Ethiopia , the “S” and “A” components of the SAFE strategy were the most implemented ones in the region [10] . Majority of the previous trachoma studies conducted in Ethiopia focused on analyzing the magnitude and associated factors of the disease . Consequently , programs were designed and predominantly implemented in the most severely trachoma affected areas based on SAFE strategy . However , there are limited studies regarding the effect of each component of the SAFE strategy on trachoma control . Therefore , this study aimed at examining the effect of WaSH interventions on active trachoma elimination in North and South Wollo Zones of Amhara Region , Ethiopia . We conducted a baseline active trachoma prevalence survey in selected woredas in 2014 . The objective of the current study was to measure the effect of WaSH program implementation on active trachoma burden after one year intervention period in the same woredas where the base line survey was conducted .
This study was approved by Ethical Review Committee of Ethiopian Institute of Water Resources . Permission was also obtained from the ANRS and the District Health Offices to conduct the study . In addition , Kebele Administrators were informed about the purpose of the study . Before the commencement of data collection , caregivers of a child were adequately informed by data collectors about the objectives and importance of their involvement in the study , the confidentiality of the information they provided , the time that the interview will take and other relevant information . Finally , caregivers of a child who were willing to take part in the study provided written informed consent . The study was conducted in North and South Wollo Zones of ANRS , Ethiopia . Based on the Federal Democratic Republic of Ethiopia Central Statistical Agency population projections , North Wollo Zone has a total estimated population of 1 , 764 , 655 . Majority ( 86% ) of the population of the zone are rural inhabitants . Eighty three percent of the population practiced Orthodox Christianity and the remaining 17% were Muslims . South Wollo Zone has a total estimated population of 2 , 980 , 618 , and ( 84% ) of the people reside in rural areas . Muslims and Orthodox Christians comprised 71% and 29% of the total population of the zone , respectively [11] . The WHO recommended WaSH program is being implemented in selected woredas of ANRS . Considerable proportion of woredas in North and South Wollo Zones are included in this initiative . The main objective of the WaSH program is to increase access to improved water supply and sanitation services for inhabitants in participating woredas/towns in ANRS and Ethiopia at large . Similarly , the Amhara Trachoma Control Program ( ATCP ) has been operating in Dessie Zuriya and Raya Kobo woredas of North and South Wollo zones . The program’s overall objective complements the WHO campaign for the global elimination of blinding trachoma by the year 2020 and The ATCP adopts multi-sectoral comprehensive approach . Provision of water and hygiene and sanitation education have been the focus areas of the program . A total of 177 improved water supply schemes were constructed in Raya Kobo and Dessie Zuriya Woredas after the WaSH program was introduced . The water supply schemes were thus increased to 508 and 387 in Raya Kobo and Dessie Zuriya Woredas from the previous numbers of 406 and 312 , respectively [12] . Among the 177 water schemes developed , 37 were constructed by the ATCP within the one-year project period ( May 2014-July 2015 ) [13] . Improved latrine coverage has increased to 70% and 79% in Raya Kobo and Dessie Zuriya Woredas from the previous percentage of 64% and 67% , respectively after the intervention period [14] . In addition , within the one-year project period of ATCP , two model latrines were constructed in intervention Woredas in public places close to health centers . Moreover , the community sensitization activities by ATCP workers have improved the construction and utilization of latrines [15] . Health education was provided through the establishment of Anti-Trachoma School Clubs in primary schools and through sensitization of the community by village hygiene educators . There were classroom trachoma ambassadors and sanitation corners in school compounds . These ambassadors check the personal hygiene situation of each student every morning at the time of flag ceremony . Students found with poor personal hygiene were taken to the sanitation corner to wash his/her hand/face and get health education about the role of hygiene on health , specifically on trachoma . The Anti-Trachoma School Clubs conducted trachoma prevention sensitization activities in the community at public gatherings . This helped to raise awareness of trachoma prevention and to promote the construction and use of latrines at household level . Community level uptake of latrines and other environmental improvements was monitored at school level with students’ under taking hygiene and sanitation audits . Village hygiene educators as part of the WaSH committee promoted the effective use of water for face washing and other hygiene practices , and for the development and uptake of improved sanitation through house to house visits and participating community meetings . Health extension workers and village hygiene educators worked together on the construction and utilization of improved latrines , solid and liquid waste disposals and other hygiene and sanitation related activities . In the non-intervention woredas ( Gubalafto and Tehuledere ) a total of additional 101 improved water supply schemes were constructed from October 2014 to September 2015 [12] . The latrine coverage was improved to 30% and 88% in Gubalafto and Tehuledere woredas from the previous percent of 25% and 84% , respectively [14] . In both intervention and non-intervention woredas , all the respective organizations were implementing the same activities however , additional water supply schemes and sanitation services were constructed by the ATCP in the intervention woredas . This was a quasi-experimental study . Sample size was calculated using Epi-Info software package . Therefore , by considering a 95% confidence interval ( Zα/2 = 1 . 96 ) , 5% for type I error , 80% power , design effect of 1 . 5 , 52% prevalence of active trachoma from a previous study [6] , a 42% proportion of active trachoma at some future date such that the quantity of ( p2-p1 ) would be the size of the magnitude of change ( 10% difference ) , and a 10% non-response rate , the total sample size was calculated to be 1358 participants . A four-stage random cluster-sampling technique was employed for selecting the study units and participants . In the first stage , four woredas were purposively selected by taking Raya Kobo and Gubalafto woredas from North Wollo , and Dessie Zuriya and Tehuledere woredas from South Wollo Zones . The presence or absence of ATCP water provision and health education on hygiene and sanitation intervention was used as criteria for selecting and including woredas in the study . Accordingly , Raya Kobo and Dessie Zuriya woredas were already selected as intervention woredas by the WaSH program implementers . Simultaneously , non-WaSH intervention woredas ( Gubalafto and Tehuledere woredas ) were selected as control group by the research team for comparison purposes . In the second stage , six kebeles ( smallest administrative units ) were randomly selected from each intervention and non-intervention woredas . The six kebeles in the intervention woredas were randomly selected out of the ten intervention kebeles . In total , 24 rural kebeles were included in the study . In the third stage of the selection process , households were randomly selected from each selected nominated kebele . The numbers of households in each woreda were decided based on the woreda level population proportion of children . In the fourth stage of sampling , only one child was randomly selected from each selected household to participate in the study . Those households that did not have eligible children were excluded from the study and were replaced by another household [Fig 1 . ] . At baseline , children aged 1–8 years were randomly selected from each randomly selected household . This survey was conducted from October to December 2014 and recently published [16] . This survey took place from October to December 2015 , 12 months after the baseline survey was conducted . It was done among the same children examined at baseline following similar survey methods . A semi-structured questionnaire was used to collect data . The data collectors were four university graduates and were adequately trained regarding the data collection process . The questionnaire was pre–tested in a non-study area and the necessary corrections were made before the actual data collection was commenced . Collected data included socio demographic , environmental and behavioral factors of active trachoma . Behavioral factors such as face and hand washing practices were measured by asking the caregivers of a child using a structured questionnaire . Information on variables such as primary source of water , amount of water consumed per day , distance to water source and availability of latrine facilities were collected from caregivers of diagnosed children for trachoma . The volume of water consumed per day was calculated by estimating the volume of fetched water per day and dividing it by the number of family members . A round trip distance to water source was measured in hour . In addition , data for WaSH practice were collected from caregivers of diagnosed children for trachoma . Clinical eye examination for the presence or absence of active trachoma was performed for each child participating in the study . Two nurses who participated as eye examiner in the previously conducted zonal trachoma survey took part in clinical diagnosis of active trachoma among the study participants . Simultaneously each child’s face was examined for facial cleanliness by the nurses and the presence or absence of ocular/nasal discharge was properly recorded . The clinical examiners didn’t have any information about those intervention and control woredas ( they were blinded ) . In addition , they were not informed about the follow up survey . Binocular loupes manufactured by Donegan optical company , Inc . at USA ( ×2 . 5 magnifications ) and penlight torches were used during eye examinations . The right eyes followed by the left were examined to avoid failure to recall in which eye the examiner saw an abnormality . Then , the diagnosed children were classified according to WHO simplified trachoma-grading card as trachomatous inflammation-follicular ( TF ) , trachomatous inflammation-intense ( TI ) , and trachomatous scarring ( TS ) or free from trachoma [17] . All trachoma-positive individuals were treated with topical tetracycline eye ointment and were advised to consult the nearby health center for further eye health follow-ups . The principal investigator and the supervisors closely monitored the entire data collection processes . The filled-out questionnaires and eye examination results were collected and delivered to supervisors after checking for consistency and completeness on daily bases . Improved water source-is a water source protected by construction from outside contamination Improved latrine-is well functioning latrine with locally made roofs , walls , floors and toilet seats . Unimproved water source-is a water source unprotected by construction from outside contamination . Woreda-refers to the third-level administrative entity of Ethiopia . In this study , several data quality assurance methods were used . Trainings and field guides were given to data collectors and supervisors . Pretests were made in a non-study area before the actual data collection was started . Intensive supervision was conducted throughout the survey period by trained supervisors and the principal investigator . Questionnaires were checked for consistency and completeness by supervisors at the end of each day . The raw data were entered using Epi Info Version 3 . 5 . 1 and exported to Statistical Package for Social Sciences ( SPSS ) IBM Version 20 ( SPSS Inc . Chicago , IL , USA ) for analysis . The analysis part contains descriptive statistics ( frequency , percentage , mean and standard deviation ) . The prevalence’s of active trachoma between baseline and follow up surveys were compared both in the intervention and non-intervention groups using the MacNemar test . In addition , MacNemar test was used to compare the status of water , sanitation and hygiene related activities between baseline and follow up surveys .
At baseline , 1 , 358 children ( 1–8 years of age ) were examined for active trachoma and were followed for one year . After one-year intervention period , 1 , 353 of the same children who were included in the baseline survey were examined for active trachoma . The socio demographic characteristics of the study participants were almost the same before and after the intervention period . Majority ( 96% ) of the participants in both intervention and non-intervention woredas were married . Regarding their religion , almost half of them were Orthodox Christians . On average , 55% of heads of the households in both intervention and control woredas were illiterate , while 45% attended primary school and above ( Table 1 ) . The availability and use of water in the study areas were compared between baseline and after WaSH interventions among intervention and non-intervention woredas . The results showed that except for the time required to fetch water ( P>0 . 05 , MacNemar test ) , type of water source and face washing habits showed significant improvement after the intervention period in intervention woredas ( P<0 . 05 , MacNemar test ) . However , in non-intervention woredas , time required to fetch water showed significant improvement after the one-year follow up period ( P<0 . 05 , MacNemar test ) ( Table 2 ) . According to the results of the MacNemar analysis , presence of ocular and nasal discharges significantly decreased after the one year intervention period in the intervention woredas ( P≤0 . 001 , MacNemar test ) . In addition , number of households with latrine and hand washing container near latrine , were increased . After the one-year intervention period in intervention woredas ( P<0 . 05 , MacNemar test ) ; while in the non-intervention woredas , presence of nasal discharge didn’t decrease after the intervention period ( P>0 . 05 ) . ( Tables 3 and 4 ) . There was a significant reduction in the prevalence of active trachoma after one-year WaSH intervention in intervention woredas ( P<0 . 05 , MacNemar test ) . However , non-intervention woredas didn’t show significant reduction after one-year study period compared to the baseline survey ( Table 5 ) . The current quasi-experimental study in North and South Wollo Zones of ANRS , Ethiopia measured the effect of WaSH intervention after one-year follow up period . The follow up survey was carried out using the same methods in the same communities as the baseline survey . Our study showed a remarkable improvement in the status and practice of water supply , sanitation and hygiene situation after the intervention period in the intervention woredas as compared to the non-intervention ones . Households from the intervention woredas had better access to improved water sources and they washed their hands and face more than two times a day . The frequency of latrine use was improved significantly in the intervention woredas than the non-intervention ones and more households were free from open defecation . Compared to the non-intervention woredas , presence of ocular and nasal discharges significantly decreased after the intervention period in the intervention woredas . Thus , many children were observed to have clean faces after the intervention . This might be due to the coordinated approach of ORDA and ATCP in the community that aimed at developing water schemes , latrines as well as health education activities on hygiene and sanitation at community as well as school levels . The establishment of Anti-Trachoma School Clubs in primary schools may also have a contribution for the improvement of the current hygiene and sanitation practices of the community at school and community level . There were sanitation corners and trachoma ambassadors in schools . These ambassadors were checking the personal hygiene of students , and when they observed poor personal hygiene , the students were taken to the sanitation corner so that he/she could wash his/her face and get education about personal hygiene . This approach may have contributed for the improvement of the current hygiene and sanitation situation in the community . The prevalence of active trachoma observed among children in both surveys among intervention and non-intervention woredas in the current study was above the 2020 global trachoma elimination target as set by the WHO [9] . However , after one-year WaSH intervention , there was a significant reduction in the prevalence of active trachoma in all communities in the intervention woredas , but not in the non-intervention woredas . This might be related to the effectiveness of the trachoma control program conducted in the intervention woredas , Several factors may have contributed to improved trachoma control during the one year ATCP intervention period . Thirty-seven water supply schemes were developed by ORDA in addition to the development of improved water supply schemes by the Regional Water , Irrigation and Energy Development Bureau in both intervention and non-intervention woredas . Similarly , the community sensitization by ATCP workers has improved the construction and utilization of latrines . Therefore , these activities may have contributed to the significant reduction of the prevalence of active trachoma in the intervention woredas as compared to the non-intervention ones . Our finding is consistent with other study results , for instance , a systematic review and meta-analysis done in October 2013 revealed a strong association between improved WaSH conditions and reduced trachoma [18 , 19] . A study conducted in December 2012 also confirmed the significant contribution of WaSH implementation for sustained control , elimination , or eradication of trachoma [6] . A population-based survey conducted in Malawi showed that sustained reductions in active trachoma could be achieved without community-based antibiotic distribution through health , water and hygiene programs [20 , 21] . In line with the current study , a review of 19 studies selected from different parts of the world showed the sustainable role of hygiene and environmental improvements on trachoma control in a population where trachoma is endemic rather than the short-term impact of antibiotic treatment [22–24] . The absence of significant reduction in the prevalence of active trachoma observed in the non-intervention woredas of the current study also strengthens the effect of WaSH intervention on trachoma control . Overall , endemic countries strategy to control and eliminate trachoma needs improved access to , and use of , water , sanitation and hygiene . Otherwise , the administration of antibiotics and correction of trichiasis cases alone to at-risk populations may not eliminate trachoma , due to its high re-occurrence nature [25] . Therefore , there should be a collaborative programming on WaSH practice to control and eliminate trachoma from its endemic areas . This study has both strengths and limitations . The use of large sample size , the presence of control woredas for comparison purpose and the internal consistency of our data indicate strength of the study . Limitations of the study are that the estimation of household fetched water volume per day and time taken to fetch water were merely based on respondents’ response to the interviewer questions , which may be uncertain . The errors due to this were minimized by asking different individuals from the households to answer the same question and taking the average value by the interviewer . In addition , the overlap in the use of facilities and education regarding hygiene practices may affect the internal validity of the study
The study showed a statistically significant reduction of active trachoma prevalence in the intervention woredas compared to the base line study . The non-intervention woredas did not show significant reduction of prevalence after the intervention period compared to the baseline survey . The finding also supports the association between active trachoma and WaSH intervention . It is therefore likely that the observed reductions in active trachoma prevalence after the intervention period may be linked to the effect WaSH interventions . However , even though we observed significant reduction of prevalence after the one-year intervention period , the magnitude of the trachoma prevalence in the study area is still very high . Therefore , to achieve trachoma elimination target by the year 2020 as set by the WHO , continued WaSH interventions and periodic monitoring , evaluation and reporting of the impact of WaSH on active trachoma is warranted in the study area . | Trachoma is an infectious disease of the eye , which is caused by repeated infection with Chlamydia trachomatis bacterium . The disease is the leading cause of preventable blindness . Ethiopia is the most trachoma affected country in the world . The World Health Organization ( WHO ) recommends Water , Sanitation and Hygiene ( WaSH ) interventions to control and eliminate blinding trachoma . With the aim of assessing the effect of WaSH intervention on active trachoma in selected woredas/districts of North and South Wollo zones of Amhara Region , Ethiopia , we selected intervention and control woredas , and WaSH program was implemented in the intervention woredas for one year . Baseline trachoma prevalence was conducted in both intervention and control areas before WaSH program was implemented . Prevalence was determined in both study areas after one-year intervention period and was compared with the baseline . The finding showed that the prevalence of active trachoma was reduced from baseline prevalence of 26% to 18% after one-year WaSH program implementation in the intervention woredas compared to the non-intervention woredas . However , as the magnitude of active trachoma prevalence observed after the intervention is still very high . Continued WaSH interventions and periodic monitoring , evaluation and reporting of the impact of WaSH on active trachoma is warranted in the studied areas of Ethiopia . | [
"Abstract",
"Introduction",
"Methods",
"and",
"materials",
"Results",
"Conclusions"
] | [
"medicine",
"and",
"health",
"sciences",
"ecology",
"and",
"environmental",
"sciences",
"water",
"resources",
"education",
"sociology",
"tropical",
"diseases",
"geographical",
"locations",
"social",
"sciences",
"health",
"care",
"bacterial",
"diseases",
"eye",
"diseases... | 2017 | Effect of water, sanitation and hygiene interventions on active trachoma in North and South Wollo zones of Amhara Region, Ethiopia: A Quasi-experimental study |
Finding bacterial cellular targets for developing novel antibiotics has become a major challenge in fighting resistant pathogenic bacteria . We present a novel compound , Relacin , designed to inhibit ( p ) ppGpp production by the ubiquitous bacterial enzyme RelA that triggers the Stringent Response . Relacin inhibits RelA in vitro and reduces ( p ) ppGpp production in vivo . Moreover , Relacin affects entry into stationary phase in Gram positive bacteria , leading to a dramatic reduction in cell viability . When Relacin is added to sporulating Bacillus subtilis cells , it strongly perturbs spore formation regardless of the time of addition . Spore formation is also impeded in the pathogenic bacterium Bacillus anthracis that causes the acute anthrax disease . Finally , the formation of multicellular biofilms is markedly disrupted by Relacin . Thus , we establish that Relacin , a novel ppGpp analogue , interferes with bacterial long term survival strategies , placing it as an attractive new antibacterial agent .
The emergence of multi drug resistant bacteria dictates the need to develop novel antibiotics that target key components of essential bacterial processes . The pleiotropic response to starvation , known as the Stringent Response , provides a potential target , as it is crucial for activation of survival strategies such as stationary phase , sporulation and biofilm formation [1]–[4] . Further , the Stringent Response has been recently shown to mediate antibiotic tolerance in nutrient-limited bacteria [5] . The Stringent Response is induced by the accumulation of the bacterial signaling molecules 5′-triphosphate-3′-diphosphate and 5′-3′-bis-diphosphate , collectively called ( p ) ppGpp [6] . Synthesis of ( p ) ppGpp has been characterized as a ribosome-dependent pyrophosphate transfer of the β and γ phosphates from an ATP donor to the 3′ hydroxyl group of GTP or GDP [7] . In Gram negative bacteria ( p ) ppGpp is mostly synthesized by RelA and hydrolyzed by SpoT , while in Gram positive bacteria a bifunctional enzyme , Rel/Spo , both synthesizes and hydrolyses ( p ) ppGpp [8] , [9] . Upon nutrient deprivation , Rel proteins bind to ribosomes blocked by uncharged tRNA and catalyze the synthesis of ( p ) ppGpp [10] . It has been proposed that Rel proteins hop between stalled ribosomes in order to achieve the ( p ) ppGpp concentration required to induce the Stringent Response [10] . A recent report , however , proposes that RelA actually synthesizes ppGpp only after it is dissociated from the ribosome [11] . The Rel proteins comprise two major domains: a catalytic domain located in the N-terminus and a regulatory domain in the C-terminus [12] . Crystal structure analysis of the N-terminal domain of Rel/Spo from Streptococcus equisimilis ( S . equisimilis ) revealed two conformations with opposing hydrolase and synthetase states [13] . Further , the N-terminal domain was found to harbor two catalytic subdomains: N-terminal which hydrolyses ( p ) ppGpp and C-terminal that catalyzes its synthesis [12] . When ppGpp accumulates within the bacterial cell it affects transcription and a plethora of physiological activities [14]–[16] . Indeed , the activation of many stress-induced genes is partially or totally dependent on ppGpp [17] , [18] . The importance of ( p ) ppGpp as a regulator of bacterial survival prompted us to develop a series of non-hydrolyzable ppGpp analogues [19] potentially targeting Rel proteins . Here we present Relacin , a potent inhibitor of Rel proteins . We demonstrate that Relacin inhibits RelA and Rel/Spo in vitro and impairs growth , sporulation and biofilm formation in Gram positive bacteria .
Based on the Rel/Spo crystal structure [13] , we designed Relacin ( Figure 1A ) , a 2′-deoxyguanosine-based analogue of ppGpp , in which the original pyrophosphate moieties at positions 5′ and 3′ were replaced by glycyl-glycine dipeptides linked to the sugar ring by a carbamate bridge ( see Text S1; Figure S1A and S1B ) . Modeling the binding of Relacin to the Rel/Spo synthetase site shows that it occupies a considerable volume of the binding pocket and forms a range of hydrogen bonds and hydrophobic interactions ( Figure S1C ) , providing a structural basis for the inhibitory effect of Relacin . To investigate the biological activity of Relacin , we first evaluated its inhibitory potential on the ( p ) ppGpp synthetase activity of RelA and Rel/Spo purified from Escherichia coli ( E . coli ) and Deinococcus radiodurans ( D . radiodurans ) , respectively . Relacin was shown to inhibit both Rel proteins in a dose-dependent manner . Remarkably , at the highest Relacin concentration , the Rel enzymes from Gram negative and positive bacteria were inhibited by approximately 100% and 80% , respectively ( Figure 1B and 1C ) . Notably , the synthesis of ppGpp and pppGpp by both Rel proteins was similarly inhibited ( Figure S2A and S2B ) . Next , we examined the effect of Relacin on the interaction between Rel/Spo and stalled ribosomes . Ribosomes purified from D . radiodurans were incubated with Rel/Spo in the presence of increasing concentrations of Relacin , and the relative amount of Rel/Spo molecules associated with 70S complexes was examined . Western blot analysis revealed that Relacin increases the levels of Rel/Spo locked on the ribosomes ( Figure 2A ) , suggesting that the presence of Relacin reduces the pool of protein molecules available for ( p ) ppGpp re-synthesis [10] . To further investigate whether ribosomes are actually required for Relacin activity , we took advantage of a RelA mutant protein ( RelAC638F ) , which exerts its activity in a ribosome-independent manner . Relacin was equally able to inhibit the mutant protein ( Figure 2B ) , indicating a direct Relacin-RelA interaction . We then examined the influence of Relacin on ( p ) ppGpp production in living cells upon induction of the Stringent Response . To this end , cells of the Gram positive spore forming bacterium Bacillus subtilis ( B . subtilis ) were incubated with Relacin and treated with serine hydroxamate ( SHX ) to simulate amino acid starvation [20] , [21] . Subsequently , the accumulated levels of ( p ) ppGpp were monitored from cell extracts . In line with the inhibitory activity observed in vitro , Relacin markedly reduced ( p ) ppGpp production in vivo ( Figure 1D ) . Interestingly , although Relacin was found to completely inhibit the activity of purified RelA from the Gram negative bacterium E . coli ( Figure 1B ) , no obvious effect of the compound on bacterial ( p ) ppGpp synthesis was observed ( Figure S2C ) . This is most likely due to the inability of Relacin to penetrate the E . coli cell and reach its target . Having ascertained that Relacin affects the production of ( p ) ppGpp in vivo by B . subtilis cells and given the vital role of the Stringent Response in bacterial survival , we investigated the impact of Relacin on cell growth and viability . Interestingly , in the presence of Relacin , cells exhibited an extended logarithmic phase as indicated by the increase in OD600 values , implying that they failed to properly enter into stationary phase ( Figure 3A ) . Of note , a similar phenomenon was observed for spoT null mutant of Helicobacter pylori [22] . This failure led to substantial dose-dependent cell death after 24 hours , with an estimated IC50 of 200 µM , as monitored by the reduction in colony forming units ( Figures 3B and S3 ) . Moreover , after 48 hours the deleterious effect of Relacin persisted , reducing the number of colonies by approximately five orders of magnitude relative to untreated cultures ( Figure 3C ) . A similar viability pattern was observed in untreated B . subtilis cells lacking Rel/Spo ( Figure 3C ) , suggesting that this enzyme is indeed the main target for Relacin action . Consistent with this observation , the survival of the mutant strain was not affected by the addition of Relacin ( Figure 3C ) . Notably , the effect of Relacin on survival is not likely to be dependent on spore formation , as only few spores , if any , were present in untreated cultures . On the other hand , the appearance of dead cells as well as disintegrated cells was largely increased within the treated population over time ( Figure S4 ) . Consistent with the inability of Relacin to perturb ( p ) ppGpp production in E . coli ( Figure S2C ) , no effect on growth and viability was detected in these cells . The biological activity of Relacin was further explored in non-spore-forming Gram positive bacteria . Treating the Group A streptococcus ( GAS ) with Relacin revealed that , although growth rate was only slightly affected , cell viability was largely reduced after 24 hours ( Figure S5A and S5B ) . This toxic effect was enhanced after 48 hours ( Figure 3D ) and was associated with the formation of very small colonies . Additionally , as observed for B . subtilis , entering stationary phase was perturbed by Relacin in the extremely slow growing bacterium D . radiodurans ( Figure S5C ) . Furthermore , the addition of Relacin to D . radiodurans cells diminished bacterial viability , as indicated by plating efficiency assay carried out after 56 and 72 hours of incubation ( Figure S5D ) . Thus , we established that Relacin functions as an antibacterial agent that impairs entry into stationary phase and causes bacterial death . In addition to switching into stationary phase some bacteria , such as Bacilli , respond to nutrient limitation by producing highly resilient dormant spores as a strategy for long term survival [23]–[25] . Entry into sporulation is triggered by a decrease in the intracellular GTP pools , in part due to conversion of GTP into ( p ) ppGpp by RelA [26] . At the onset of sporulation , an asymmetric septum is generated , dividing the cell into a nurturing mother cell and a smaller forespore compartment that develops into a spore . Subsequently , the forespore is engulfed by the mother cell to form a fully mature spore . Remarkably , when nutrients become available the spore can rapidly convert into an actively growing cell [23]–[25] . To explore whether Relacin affects sporulation , B . subtilis cells were induced to sporulate in the presence or absence of Relacin and sporulation progression was monitored by observing polar septa formation . Indeed , sporulation was largely inhibited , with asymmetric septa exhibited by only 8% and 0 . 5% of the cells treated with 200 µM and 1 mM of Relacin , respectively . In comparison , 47% of untreated cells displayed polar septa at the same time point ( Figure 4A ) . In line with these observations , Relacin lowered the number of cells expressing early ( SpoIIE ) , middle ( SpoIIQ ) and late ( SspE ) sporulation-specific proteins along the process [25] ( Figures 4B and S6 ) . Subsequently , a fivefold drop in the formation of mature heat resistant spores was measured at the highest Relacin concentration ( Figure 4A and 4C ) . Remarkably , adding Relacin to sporulating cells at different time points , up to six hours after the induction of sporulation , strongly inhibited spore formation regardless of the time of addition ( Figure 4E ) . These findings indicate that the ppGpp signal is crucial throughout the entire pathway of sporulation , and demonstrate the potency of Relacin to impede this process . Importantly , spore formation in the pathogenic bacterium Bacillus anthracis , the causative agent of anthrax disease , was inhibited by Relacin in a similar fashion ( Figure 4D ) , establishing the compound as a general inhibitor of the Bacilli sporulation process . Since it has been reported that relA mutant cells fail to properly form multicellular biofilm structures [2] , the effect of Relacin on the ability of B . subtilis cells to produce biofilms was evaluated . Indeed , a disrupted pellicle was visualized at the air/liquid interface of standing cell cultures grown in the presence of the compound ( Figure 5A ) . Importantly , the effect on biofilm formation was found to be dose-dependent ( Figure 5A ) . Consistent with this observation , Relacin also inhibited the development of biofilm on solid medium , leading to the formation of colonies with altered morphology that were smaller in size than the untreated ones ( Figure 5B ) . To visualize cell assembly within the biofilm pellicle in higher resolution upon Relacin treatment , we took advantage of a strain harboring the rrnE promoter fused to gfp . This promoter was found to be constitutively active [27] , and therefore reports cell viability and localization . Observing biofilm pellicles by confocal laser scanning microscopy revealed that the untreated cells formed homogeneous biofilm layers , while the treated cell pellicles contained large gaps , indicating their disintegrated state ( Figure 5C ) . Moreover , staining the biofilm with propidium iodide ( PI ) , indicative of unviable cells , showed the signal to be higher within the treated biofilm ( Figure 5E ) . Finally , quantifying GFP fluorescence from recovered pellicles revealed a clear reduction in the viable biomass upon Relacin treatment , as the measured fluorescence level was significantly reduced ( Figure 5D ) . Taken together , we conclude that Relacin interferes with biofilm formation , an alternative bacterial developmental pathway .
In this report , we established Relacin as a novel antibacterial agent . By specifically interfering with the activation of the Stringent Response , Relacin perturbs the switch into stationary phase in several tested Gram positive bacteria and leads to bacterial death . Although Relacin did not affect growth and survival of the Gram negative E . coli , it was found to effectively inhibit the E . coli RelA in vitro , implying that improving the delivery of Relacin to Gram negative bacteria may lead to an effective outcome . Relacin was found to block every tested stage of B . subtilis sporulation , proving the essentiality of the Stringent Response throughout this process . Finally , we demonstrate that Relacin affects the production of multicellular biofilm communities , formed in response to challenging conditions . Taken together , we present evidence that Relacin impedes bacterial long term survival pathways , placing the compound as a new promising antibacterial agent . By utilizing the crystal structure of Rel/Spo from the S . equisimilis , we were able to model the interaction of Relacin with amino acid residues located within the Rel/Spo synthetase site . This analysis yielded the identification of a putative binding mode of Relacin , presumably adopting the conformation shown in Figure S1C . In this conformation , Relacin forms a net of hydrogen bonds and hydrophobic interactions that are most likely to provide a more efficient binding in comparison to previously identified inhibitors exhibiting lower activity [19] . Relacin appears to specifically target Rel proteins , as the effect of the compound was nearly undetectable when tested on Rel/Spo mutant cells . Consistently , Relacin activity in vivo resulted in a sharp decrease in ( p ) ppGpp synthesis . Since ppGpp inhibits the enzyme inosine monophosphate dehydrogenase , it causes the cellular GTP levels to decrease [28] . The intracellular levels of GDP/GTP are known to determine the initiation of several developmental pathways such as sporulation and biofilm formation [26] , [29] , [30] that were indeed shown to be influenced by Relacin . Interestingly , we also observed that Relacin treatment resulted in a large decrease in Rel/Spo ability to dissociate from ribosomes in vitro . This deficiency could be explained by the model proposed by Wendrich et al . , [10] in which the rapid accumulation of ppGpp during amino acid starvation is attributed to the ability of RelA to ‘hop’ between ribosomes . This potential hopping is probably a consequence of the synthesis of ( p ) ppGpp that releases RelA from the ribosome , liberating it for another synthesis cycle . The emergence of bacterial resistance to the current array of antimicrobial agents demands the development of novel strategies to eradicate pathogenic bacteria . The traditional cellular antibiotic targets include ribosomes , cell wall constituents and components of nucleic acids synthesis [31] . These cellular targets are mainly active during the bacterial vegetative phase , making the available antibiotics effective mostly during growth . However , the ability of bacteria to reside in nature within biofilm communities or as durable spores , as well as to become persistent to antibiotic treatment [32] , sets the need to tackle these alternative modes . In this regard , Relacin affects specifically the Stringent Response , a pathway crucial for the activation of bacterial survival strategies . Since Relacin can persist for a relatively long period of time , and exert its effect even a few days post addition , it might become a valuable antagonist of these long term survival approaches . Taken together , Relacin may be combined with antibiotics currently in use , to eradicate non-homogenous bacterial populations with cells residing in diverse life cycles . Cellular components , which are conserved throughout the bacterial kingdom and crucial for cellular survival , provide attractive antimicrobial targets as long as they lack eukaryotic counterparts . One of such targets is the highly conserved bacterial tubulin-like cell division protein FtsZ , which provides the basis for the assembly of the division machinery [33] . Indeed , a promising inhibitor of FtsZ with potent and selective activity against Staphylococci has been described [34] . In a similar fashion , the ubiquity of Rel enzymes among bacteria , combined with the absence of known ( p ) ppGpp synthetases in mammalian cells [35] , [36] , strengthen the potential of Relacin to turn into a therapeutic antibiotic . The profound influence of Relacin on long term bacterial survival makes it an attractive compound to serve as a scaffold for generating an array of new antibacterial agents .
Synthesis of Relacin and a structural model for its interaction with Rel/Spo ( p ) ppGpp synthetase binding pocket are described in details Text S1 . Bacterial strains used in this study are described in Table S1 . Plasmid construction is described in Text S1 . All general methods for B . subtilis were carried out as described previously [37] . B . subtilis cells were grown in hydrolyzed casein ( CH ) at 37°C [37] , unless indicated differently . GAS strain was grown at 37°C without shaking in Todd-Hewitt medium supplemented with 0 . 2% yeast extract ( THY ) [38] . D . radiodurans R1 cells were grown in TYG which contains: 0 . 5% tryptone , 0 . 3% yeast extract and 0 . 1% glucose at 30°C with shaking . E . coli cells were grown at 37°C in LB medium . Cultures were inoculated to an OD600 of 0 . 05 using an overnight culture grown in the same medium , unless indicated differently . Sporulation conditions and biofilm colony and pellicle formation are described in Text S1 . Purification of RelA or RelA-C638F from E . coli ( CF9467 ) harboring ΔrelA and over-expressing pQE30-RelA or pQE30-RelA-C638F respectively , was carried out as described previously [19] . Purification of Rel/Spo from D . Radiodurans R1 was performed under identical conditions; however , the protein was expressed in E . coli BL21 CodonPlus ( Stratagene ) cells . Of note , Rel/Spo from D . Radiodurans R1 , is the only known full length active protein purified from Gram positive bacteria . Isolation of crude ribosomes ( containing 70S , mRNA , tRNA ) from E . coli ( CF9467 ) was carried out as described previously [19] . Isolation of crude ribosomes from D . Radiodurans was carried out in a similar fashion with the following modifications: D . radiodurans R1 cells were grown in LB ( + ) over night at 30°C , cells were diluted 1∶100 in LB ( + ) medium and incubated at 30°C for additional 48 hours . For measuring ( p ) ppGpp synthesis by RelA , RelA-C638F or Rel/Spo proteins in vitro: 1 µg of purified Rel protein , 20 µg of isolated ribosomes and 10 µCi of α-32P labeled GTP , were mixed in a buffer [0 . 5 mM GTP , 4 mM ATP , 50 mM Tris-HCl ( pH 7 . 4 ) , 1 mM DTT , 10 mM MgCl2 , 10 mM KCl and 27 mM ( NH4 ) 2SO4] to a final volume of 20 µL without or with increasing amounts of Relacin as indicated . Reactions were stopped by the addition of 5 µL formic acid . Each reaction was loaded ( 5 µL ) and separated on Cellulose PEI TLC plates ( Merck ) using 1 . 5 M KH2PO4 as mobile phase . Plates were autoradiographed using the Fijix Bas100 PhosphorImager ( Japan ) . ( p ) ppGpp signal was measured using TINA 2 . 0 software ( Raytest , Strauben-Hardt ) . The total amount of ( p ) ppGpp was the sum of signals from ppGpp and pppGpp . B . subtilis ( PY79 ) or E . coli ( W3110 ) cells were grown in MOPS glucose minimal medium [39] supplemented with all amino acids except glutamine and glutamate . At OD600 0 . 1 , cells were supplemented with H332PO4 and incubated for 45 minutes , after which Relacin was added at the indicated concentrations . Cells were incubated for additional 15 minutes . Next , amino acid starvation was induced by adding serine-hydroxamate ( SHX , Sigma ) 1 mg/mL [20] . Samples were withdrawn ten minutes after addition of SHX and analyzed for their ( p ) ppGpp content as described above ( Measuring ( p ) ppGpp synthesis in vitro ) . The reaction was carried out as described above for measuring ( p ) ppGpp synthesis in vitro , without the addition of radiolabeled GTP , with or without increasing amounts of Relacin as indicated . Reactions were centrifuged for 4 hours at 35 , 000 g ( 4°C ) , ribosomal fractions were separated by 12% SDS-polyacrylamide gel electrophoresis , transferred to PVDF membrane ( Millipore Bedford ) and processed for immunoreaction using mouse anti-His antibody ( 1∶10 , 000; Amersham ) . Immunoreactive proteins were detected using a chemiluminescence kit ( Biological Industries ) according to the manufacturer's protocol . Fluorescence microscopy was carried out as previously described [40] . Samples ( 0 . 5 mL ) of a given culture were removed , centrifuged briefly , and resuspended in 10 µL of PBS×1 ( Phosphate-Buffered Saline ) supplemented with 1 µg/mL membrane stain FM1–43 or FM4–64 ( Molecular Probes , Invitrogen ) . Cells were visualized and photographed using an Axioplan2 microscope ( Zeiss ) equipped with CoolSnap HQ camera ( Photometrics , Roper Scientific ) or an Axioobserver Z1 microscope ( Zeiss ) equipped with a CoolSnap HQII camera ( Photometrics , Roper Scientific ) . System control and image processing were performed using MetaMorph 7 . 2r4 software ( Molecular Devices ) . For observing biofilm pellicles , the medium of 3 day-old pellicles grown in microplates was gently removed and minimal volume of PBS ×1 with or without 10 µg/ml PI ( Molecular Probes , Invitrogen ) was added . Cells were visualized and photographed using a confocal laser scanning fluorescence microscope LSM700 ( Zeiss ) . System control and image processing were performed using Zen 2009 ( Zeiss ) and MetaMorph 7 . 2r4 ( Molecular Devices ) softwares . | The development of new antibacterial agents has become the major demand for fighting against pathogenic bacteria . The identification of new unexplored cellular targets in this combat is essential to prevent a possible return to the pre-antibiotic era . Traditional antibiotics target essential cellular components such as ribosomes and cell wall constituents , making them effective mostly during bacterial growth . However , the ability of bacteria to reside in nature at durable stages sets the need to cope with these alternative survival strategies . In this report , we present a novel antibiotic , termed Relacin , which targets the Stringent Response , a process required for the transition into stationary phase , crucial for bacterial survival . Relacin inhibits the abundant bacterial Rel enzymes that synthesize the signaling molecules required to activate the Stringent Response . We found that Relacin perturbs the switch into stationary phase in Gram positive bacteria and leads to cell death . Further , Relacin inhibits sporulation and biofilm formation , additional bacterial long term survival strategies . The ubiquity of Rel enzymes among bacteria , combined with the absence of known homologues in mammalian cells , strengthen the potential of Relacin to turn into a therapeutic antibiotic . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicinal",
"chemistry",
"chemical",
"biology",
"microbial",
"pathogens",
"microbial",
"control",
"chemistry",
"biology",
"microbiology",
"microbial",
"growth",
"and",
"development"
] | 2012 | Relacin, a Novel Antibacterial Agent Targeting the Stringent Response |
Infection by enveloped coronaviruses ( CoVs ) initiates with viral spike ( S ) proteins binding to cellular receptors , and is followed by proteolytic cleavage of receptor-bound S proteins , which prompts S protein-mediated virus-cell membrane fusion . Infection therefore requires close proximity of receptors and proteases . We considered whether tetraspanins , scaffolding proteins known to facilitate CoV infections , hold receptors and proteases together on cell membranes . Using knockout cell lines , we found that the tetraspanin CD9 , but not the tetraspanin CD81 , formed cell-surface complexes of dipeptidyl peptidase 4 ( DPP4 ) , the MERS-CoV receptor , and the type II transmembrane serine protease ( TTSP ) member TMPRSS2 , a CoV-activating protease . This CD9-facilitated condensation of receptors and proteases allowed MERS-CoV pseudoviruses to enter cells rapidly and efficiently . Without CD9 , MERS-CoV viruses were not activated by TTSPs , and they trafficked into endosomes to be cleaved much later and less efficiently by cathepsins . Thus , we identified DPP4:CD9:TTSP as the protein complexes necessary for early , efficient MERS-CoV entry . To evaluate the importance of these complexes in an in vivo CoV infection model , we used recombinant Adenovirus 5 ( rAd5 ) vectors to express human DPP4 in mouse lungs , thereby sensitizing the animals to MERS-CoV infection . When the rAd5-hDPP4 vectors co-expressed small RNAs silencing Cd9 or Tmprss2 , the animals were significantly less susceptible , indicating that CD9 and TMPRSS2 facilitated robust in vivo MERS-CoV infection of mouse lungs . Furthermore , the S proteins of virulent mouse-adapted MERS-CoVs acquired a CD9-dependent cell entry character , suggesting that CD9 is a selective agent in the evolution of CoV virulence .
Enveloped virus-cell entry requires glycoprotein-catalyzed fusion of viral and host cell membranes . These viral fusion glycoproteins are catalytically-inactive on virus particles and become triggered to mediate membrane mergers only in response to cellular and environmental factors . This triggering process ensures that virus-cell entry occurs at the appropriate time and place . The triggering factors include host cell receptors , endosomal acids , and proteases . Many viruses require a single , soluble trigger , for example , influenza A virus fusion proteins are triggered by protons within the target-cell endosome [1] . Other viruses require two triggering agents , for example , avian sarcoma leukosis virus fusion proteins are partially advanced into fusion-catalyzing forms after binding to host cell receptors , and then fully execute fusion after being exposed to endosomal protons [2] . Most CoVs also require two triggering agents , receptor binding and proteolytic cleavage , with the proteolysis taking place on receptor-bound viral ligands [3] . As many of the CoV-cleaving proteases are transmembrane-anchored , it follows that CoV-susceptible cells might have the two triggering agents , receptors and proteases , in close proximity . Here we considered whether the two CoV entry factors are coalesced on cell surfaces to facilitate infection , and whether particular host cell features are required to juxtapose the two entry factors . The CoV receptors are all transmembrane glycoproteins . Their presence is a defining feature of host cell susceptibility to infection [4–7] . Proteases , the second required determinants of host susceptibility , are variable in type and subcellular location [8] , with proteases in the type II transmembrane serine protease ( TTSP ) family figuring prominently [8–10] . TTSP family members , most notably the transmembrane protease serine type 2 ( TMPRSS2 ) , can cleave CoV fusion glycoproteins ( termed spike [S] proteins ) , into unlocked , fusion-catalyzing forms [8 , 9 , 11] at the cell surface and facilitate a rapid , “early” entry . Studies examining HIV [12] and influenza [13] glycoproteins indicate that multiple adjacent fusion glycoproteins must be activated in order to successfully complete the fusion reaction . Assuming similar requirements for CoV fusion , it is likely that multiple S proteins need simultaneous receptor engagement and sufficient proteolytic cleavage to form an activated cluster that can pull opposing membranes together . Thus , fusion likely occurs in regions of the cell membrane with a relatively high local concentration of these entry factors . Recent studies have confirmed that TTSPs are concentrated into punctate locations on the cell surface , in association with tetraspanin scaffolding proteins [14] . Tetraspanins comprise a family of proteins with four transmembrane spans and two extracellular loops [15] . Tetraspanins interact with other tetraspanins [16] and with other membrane-associated proteins [17 , 18] , including transmembrane proteases [19 , 20] , to form “webs” of interacting proteins [15] . There is evidence that these tetraspanin webs are locus points for CoV-cell entry , as tetraspanin-specific antibodies protect several cell types from CoV infection [14] . However , it remains unclear if individual tetraspanin proteins facilitate CoV entry and what function they have in determining viral entry routes . As there are demonstrations that the tetraspanin CD9 interacts with the MERS-CoV receptor dipeptidyl peptidase 4 ( DPP4 ) [21 , 22] and hints of similar CD9 interactions with the HCoV-229E receptor aminopeptidase N ( APN ) [23] , we hypothesized that CD9 is necessary to bring these virus receptors to TTSP-enriched regions on the cell surface . No study to date has determined the relative importance of individual tetraspanins and TTSPs to MERS-CoV infection in the lung environment . Indeed , there are 34 human tetraspanins and at least 17 members of the TTSP protease family [24] as well as several soluble extracellular proteases , such as elastases [25] , that may be expressed in the lung parenchyma . While studies suggest that TMPRSS2 can trigger MERS-CoV in cell culture [9 , 25] , it is unclear whether CD9 or TMPRSS2 stand out in vivo as single proviral members of their respective protein families . Therefore , we set out to determine whether , and to what extent , MERS-CoV utilizes CD9 and TMPRSS2 during in vivo infection . To this end , we established a mouse model in which virus-resistant mice are rendered susceptible to MERS-CoV infection by expression of human DPP4 ( hDPP4 ) . The system utilizes a recombinant adenovirus type 5 ( rAd5 ) to transduce the hDPP4 gene , thereby sensitizing only the Ad5-transduced lung cells to subsequent MERS-CoV infection [26] . The rAd5-hDPP4 vectors were engineered to include additional genes encoding the potential virus-promoting factor human TMPRSS2 [9] or potential virus-restricting factors , in the form of shRNAs targeting murine Tmprss2 and Cd9 . We considered the rAd5-hDPP4 system to be especially valuable , as MERS-CoV infection can only occur in cells expressing hDPP4 and , thus , only in cells simultaneously expressing the putative virus-promoting or virus-restricting factors . Using the dual-expressing rAd5 vectors , as well as tetraspanin knock-out cell lines , we evaluated the roles for CD9 and another related tetraspanin , CD81 , in dictating receptor clustering with proteases and in promoting CoV infection . Our results indicate that a CoV-cell entry portal is a multipartite complex that minimally includes the virus receptor , a virus-activating protease , and one or more tetraspanins . These complexes are responsible for the majority of MERS-CoV entry in lung cells . Furthermore , CD9 facilitated cell entry by MERS-CoV spikes that were adapted for lung virulence , but CD9 provided no support to cell culture-derived , avirulent spike-mediated cell entry . These data establish tetraspanins as factors controlling early entry events in pathogenic CoV infections .
Tetraspanins CD9 and CD81 are known to influence enveloped virus entry [14 , 27 , 28] . Therefore , we used CRISPR/Cas9 technology [29] to eliminate these tetraspanins from cells , with the expectation that this would affect cell susceptibility to CoVs . 293T and HeLa cells were transfected with Cas9/guide RNAs targeting CD9 or CD81 , selected for puromycin resistance , and cloned by endpoint dilution . All KO cell lines grew equivalent to parallel “WT” control clones , and the only observable distinctions were with the CD9KO cells , which adhered less tightly to plastic than WT or CD81KO cells . Western blot analyses of the WT and KO clones confirmed the absence of CD9 or CD81 , with maintenance of a control tetraspanin CD63 ( Fig 1A ) . Interestingly , CD81 levels were highest in CD9KO cells and CD9 levels were low in CD81KO cells , possibly due to heterotypic CD9-CD81 interactions influencing tetraspanin stability . Lower-resolution immunofluorescent assays ( IFAs ) of umpermeabilized cells showed similar cell-surface CD9 in WT and CD81KO cells , confirmed the absence of the respective tetraspanins in KO cells , and demonstrated that CD63 distribution remained unchanged in all cell lines ( Fig 1B ) . To determine whether the tetraspanins CD9 or CD81 operate in CoV entry , we utilized HIV pseudoparticle ( pp ) transduction methodologies , which allow for a specific focus on the virus-cell entry stage . We first sensitized the cells to transduction by overexpressing CoV receptors , then transduced cells with the respective CoVpps . Relative to WT cells , CD9KO cells were 94% less susceptible to MERS ( EMC strain ) pp transduction ( Fig 1C ) , and 80% less susceptible to 229Epp transduction ( Fig 1D ) . However , CD9KO cells remained fully susceptible to SARSpp or MHVpp transduction ( Fig 1E and 1F ) . CD9 complementation restored susceptibility to MERSpp and 229Epp transductions ( Fig 1C and 1D ) . CD81 KO cells were fully susceptible to all four of the CoVpps ( S1 Fig ) . These data identify an individual tetraspanin , CD9 , as an entry factor for a CoV . To determine whether receptor overexpression might have contributed to CD9 dependence , MERSpps were also transduced into cells containing endogenous CoV receptor levels . Consistently , CD9 was necessary to fully sensitize cells to MERSpps ( S2 Fig ) , indicating that CD9 proviral activity was independent of hDPP4 receptor levels . However , CD9 was not necessary for MERSpp transduction into cells overexpressing TMPRSS2 , a MERS-CoV activating protease ( S2 Fig ) . The fact that TMPRSS2 obviated the CD9 requirement indicated a role for CD9 in proteolytic activation of CoV entry . The observation that a single tetraspanin family member , CD9 , promoted cell entry for some , but not all CoVs , suggested that CD9 interacts with one or more MERS-CoV and 229E-CoV entry factors . We considered whether CD9 associates with DPP4 and APN , the MERS-CoV and 229E-CoV receptors , or with TMPRSS2 . Furthermore , we considered whether CD9 does not interact with ACE2 and CEACAM , the receptors for CD9-independent SARS and MHV-CoVs . This was first investigated through biochemical isolation of tetraspanin-enriched membrane fractions , and detection of tetraspanin-associated receptors and proteases . To this end , CD9 or CD81 KO cells overexpressing CoV receptors or TMPRSS2 were surface-biotinylated , and tetraspanins were liberated from cells using zwitterionic CHAPS detergent , which solubilizes cell membranes while leaving tetraspanin-mediated protein interactions largely intact [30] . Low-Density ( LD ) fractions , with ρ<1 . 13 g/ml , were then separated from High-Density ( HD ) CHAPS-solubilized proteins on sucrose density gradients [31] . As evaluated by streptavidin pull-down and western immunoblotting , the LD sucrose gradient fractions from CHAPS-solubilized cells contained nearly 100% of cell-surface tetraspanins ( S3 Fig ) , but only ~ 20% of the surface-biotinylated plasma membrane proteins [14] , indicating efficient tetraspanin segregation into LD fractions . Strikingly , the LD fractions from WT control cells contained ~60% of cell-surface DPP4 , while LD fractions from CD9 KO cells completely lacked DPP4 ( Fig 2A , rows 1 and 2 ) . Complementing CD9 back into CD9KO cells restored LD-associated DPP4 ( Fig 2A , row 3 ) . The presence or absence of CD81 had no effect on DPP4 distribution between HD and LD fractions ( Fig 2A , rows 4 and 5 ) . Similar results were observed with the 229E receptor APN ( Fig 2B ) . By contrast , CD9 and CD81 expression had little effect on the distribution of ACE2 , CEACAM , or TMPRSS2 , all of which distributed about equally between LD and HD fractions ( Fig 2C–2E ) . These data indicated that DPP4 and APN positioning into tetraspanin-enriched membranes required CD9 . The fact that CD9 repositioned DPP4 and APN , but not ACE2 or CEACAM , correlated with the fact that CD9 promoted the entry of DPP4- and APN-utilizing MERS and 229E viruses , but not ACE2- or CEACAM-utilizing SARS and MHV viruses ( Fig 1 ) . The evaluations of membrane fractions suggested that CD9 might localize the MERS-CoV and 229E-CoV receptors close to virus-activating TMPRSS2 . To determine whether CD9 facilitates specific interactions between DPP4 and TMPRSS2 , we analyzed intact tetraspanin microdomains in situ . We performed proximity ligation assays ( PLAs ) , which can determine whether two or more transmembrane proteins are adjacent [32] . In PLAs , antibodies differentially tagged with oligonucleotide probes are applied to cells , and their close spacing ( <40 nm ) allows for probe hybridization into DNA polymerization templates , which provide a locus point for fluorescent DNA synthesis [33] . PLAs have been used to identify interactions between tetraspanins and their partner proteins [34 , 35] and we used this method to analyze clustering of two tetraspanin partner proteins . HeLa cells were chosen for PLAs because their relatively flat morphology facilitated quantification of fluorescent foci , and because our quantitative reverse transcriptase–PCR measurements revealed endogenous expression of CD9 , DPP4 and TMPRSS2 ( S1 Table ) . Notably , CD9 transcripts were plentiful in the HeLa cells ( ~10 times more abundant than the reporter gene HPRT ) , while DPP4 and TMPRSS2 were scarce ( ~50 times less abundant than HPRT , and 5- to 100 times less than that found in several human airway epithelia-derived cell cultures ( see S1 Table ) ) . Thus , we presumed that , with HeLa cells , we could readily detect a CD9-directed coalescence of sparse DPP4 and TMPRSS2 . We performed PLAs on unpermeabilized CD9KO HeLa cells , using primary antibodies to CD9 , DPP4 , and/or TMPRSS2 . Following secondary antibody incubation and amplification of ligated oligonucleotide templates , punctate fluorescent DNAs were detected by confocal microscopy and counted using Imaris version 6 . 3 . 1 software . Using hDPP4 and hTMPRSS2 antibodies , fluorescent foci were rarely observed on the HeLa-CD9KO cells ( Fig 3A ) , and the cells were only modestly susceptible to MERSpp transduction ( Fig 3H , leftmost bar ) . When CD9 was replenished in the CD9KO cells , foci were ~10-fold more abundant ( Fig 3D ) , and these increased foci correlated with a greater cell susceptibility to MERSpp transduction ( Fig 3H ) . These findings argue that CD9 sensitizes cells to MERS-CoV entry by bringing DPP4 and TMPRSS2 into proximity . We considered whether this role for CD9 applied only when DPP4 and TMPRSS2 levels were low , i . e . , at endogenous HeLa-cell levels . Thus , hDPP4 and hTMPRSS2 were forcibly overexpressed; with overexpression , ~ 30 foci/cell were observed ( Fig 3E ) , and this increased to ~ 80 foci/cell in the presence of CD9 ( Fig 3F ) . MERSpp entry into cells correlated with the number of foci present , at least for values up to ~ 30 foci/cell ( Fig 3H ) . Overall , these results indicated that CD9 connects DPP4 and TMPRSS2 entry factors , and is necessary for their proximity when they are sparse on cell surfaces . The CD9:DPP4:TMPRSS2 complexes then function as MERS-CoV entry portals . CD9 also helped to connect overexpressed DPP4 and TMPRSS2 together , but in this overexpression condition , CD9 did not increase MERSpp transduction , perhaps because other tetraspanins come in to bridge the abundant receptors and proteases . These results also revealed CD9-directed DPP4:TMPRSS2 complexes on intact cells in the absence of virus , suggesting that the CoVs infect through pre-existing complexes . Because CD9 brought DPP4 in proximity with TTSPs , we hypothesized that CD9 facilitates TTSP-mediated early cell entry at or near plasma membranes , but does nothing to support the late , endosomal route that is enabled by cathepsin proteases . To test this , we inactivated cellular TTSPs using camostat [25] and found that camostat suppressed MERSpp transduction into WT cells by ~50% , but did not affect transduction into CD9KO cells ( Fig 4A ) . CD9 complementation modestly restored MERSpp sensitivity to camostat . Furthermore , CD81 had no effect , as MERSpp entry into CD81KO and CD81-positive cells were equally suppressed by camostat ( Fig 4A ) . These data were consistent with CD9 specifically enabling TTSP-directed , early virus entry . Without CD9 , the MERSpp entry route may be directed to a late , endosomal stage in which cathepsins provide fusion-activating triggers . To test this , we blocked late entry in WT and CD9KO with 100 μM bafilomycin A ( Baf ) , an inhibitor of endosome acidification , or with 10 μM E64D , a cysteine protease inhibitor . In WT cells , Baf did not significantly decrease MERSpp entry , while E64D decreased entry ~4-fold ( Fig 4B ) . However , in CD9KO cells , these inhibitors were far more antiviral , decreasing entry 20- and 100-fold , respectively . Complementing CD9 back into the CD9KO cells restored the WT phenotype in which the inhibitors were only weakly antiviral ( Fig 4B ) . These differential effects of the inhibitors were not observed in CD81KO or CD81-overexpressing cells ( Fig 4B ) . We conclude that CD9 is necessary for TTSP-mediated MERS early entry . We advanced to evaluating MERS-CoV entry factors in vivo . Of note , a previous study has demonstrated that camostat inhibits SARS-CoV spread in mouse lungs [36] , suggesting that the virus exhibits dependence on serine proteases , probably TTSPs , for its entry in vivo . However , the importance of specific TTSPs , or for tetraspanins , is unknown for any in vivo CoV infection . Here we established infections in the mouse lung under conditions in which putative CoV entry factors were reduced . To do this , we developed dual-expressing recombinant adenovirus 5 ( rAd5 ) vectors expressing both human DPP4 , which sensitizes mouse cells to MERS-CoV infection [26 , 37 , 38] , and shRNAs that knock down Tmprss2 or Cd9 mRNAs . In initial experiments , these rAd5 vectors were transduced into mouse Lung Epithelial Type 1 ( LET-1 ) cells , a line derived from C57/Bl6 mouse alveolar type 1 cells [39] . After 3-days , the cells were analyzed for the presence of hDPP4 , TMPRSS2 , and CD9 by western blot ( Fig 5A ) . Relative to the control rAd5-GFP transductions , all single and dual-expressing rAd5-hDPP4 transductants contained recognizable DPP4 and TMPRSS2 , and those Ad5 vectors expressing shRNAs reduced the levels of endogenous CD9 proteins ( Fig 5A ) . Due to endogenous TMPRSS2 protein levels being too low for detection on immunoblots , we used qRT-PCR to quantify TMPRSS2 transcripts . LET-1 cells transduced with rAd5-hDPP4-shTmprss2 had only 25% of the transcripts of cells transduced with rAd5-hDPP4-empty vector ( Fig 5B ) . This level of Tmprss2 transcripts indicated an efficient knockdown of TMPRSS2 in the approximately 75% of cells that were successfully transduced . These results indicate that the different rAd5 vectors , transduced into cells derived from mouse alveolar epithelia , consistently express equivalent levels of hDPP4 , while simultaneously increasing or decreasing TMPRSS2 or CD9 . To determine whether the rAd5-transduced LET-1 cells were susceptible to MERS-CoV S protein-directed virus entry , the cells were inoculated with recombinant VSVs encoding firefly luciferase [40] and pseudotyped with MERS-CoV S proteins . As expected , hDPP4 expression established susceptibility to VSV-MERSpp transduction ( Fig 5C ) . TMPRSS2 co-expression from the Ad5 vectors increased susceptibility to MERSpps by ~ 4-fold , while shTmprss2 and shCd9 both restricted MERSpps by ~3 fold ( Fig 5C ) . These results indicated that CD9 and TMPRSS2 act as entry factors in mouse lung-derived LET-1 cells , and suggested that the dual-expressing Ad5 vectors might be effective tools for identifying viral entry factors in the mouse lung . To identify the role of CD9 and TMPRSS2 in vivo , the Ad5 vectors were instilled intranasally into mice which were , after 5 days , challenged with MERS-CoV . Lungs were harvested 2 days post-infection ( d . p . i . ) and MERS-CoV titers were measured as PFU/gram of tissue . Relative to MERS-CoV titers in rAd5-hDPP4 transduced animals , the MERS-CoV titers in rAd5-hDPP4-shCd9 transduced animals were ~20-fold lower ( Fig 5D ) . Furthermore , the MERS-CoV titers in rAd5-hDPP4-shTmprss2 transduced mice were reduced by ~10-fold . Interestingly , overexpression of TMPRSS2 by the rAd5-hDPP4-TMPRSS2 vector had no effect on MERS-CoV titers in the lungs , presumably because the lung environment has sufficient endogenous murine TMPRSS2 to facilitate efficient MERS-CoV infection . These data indicate that CD9 and TMPRSS2 act as MERS-CoV susceptibility factors in the lung parenchyma and that their role in entry is slightly more pronounced in vivo than in in vitro LET-1 mouse alveolar cell cultures . Indeed , these data show that CD9 and TMPRSS2 are responsible for ~90% of MERS-CoV titers in vivo . MERS-CoV , a camel and human virus [41 , 42] , has recently been adapted for robust growth and virulence in hDPP4+ mouse lungs [43 , 44] . This adaptation process was initiated by intranasally infecting mice with avirulent , Vero Cell Culture-Adapted ( CCA ) MERS-CoVs and then serially passaging viruses through hDPP4+ mouse lungs . Relative to CCA MERS-CoVs , the Mouse-Adapted ( MA ) viruses have distinct S protein changes [44] ( S2 Table ) . We considered whether these MA changes fixed into S proteins adapt viruses to utilize CD9-facilitated early entry . To address this question , we produced VSV-based MERSpps , pseudotyped with the CCA or MA S proteins . These CCA and MA MERSpps were transduced into CD9-replete or CD9-knocked down ( CD9KD ) LET-1 cells . The CD9-replete and CD9KD cells were equally susceptible to CCA S-mediated pp entry . However , the same CD9KD cells had 90% and >95% reduced susceptibility to MA1 and MA2 S-driven pp entry , respectively ( Fig 6A ) . Thus , it appears that in vivo passage in mouse lungs adapts MERS-CoVs to a CD9-facilitated cell entry pathway . The MA viruses’ utilization of CD9 for entry correlated with their relatively rapid entry kinetics [44] . Furthermore , CD9-facilitated entry correlates with TTSP utilization ( Fig 4 ) , and TTSP utilization correlates with rapid CoV entry into cells [45] . Therefore , we hypothesized that CD9 is a determining factor in CoV-cell entry kinetics . To test this , CCA and MA MERSpps were transduced into CD9-replete or CD9KD LET-1 cells . To measure pp entry kinetics , the transduction process was abruptly halted at defined time points with a nontoxic protease inhibitor cocktail that prevents S-directed fusion , but has no effect on transduced reporter gene expression [45] . This strategy allows Fluc reporter accumulations to indicate the extents of virus entry taking place within timed intervals . We found that CD9 had no influence on the rate of CCA S-directed virus entry . In both CD9-replete and CD9-KD cells , half-maximal entry was complete within 45 min ( Fig 6B ) . A 30–45 min half time for virus entry is found for several viruses requiring endocytosis prior to genome delivery [46] . However , CD9 strongly influenced MA S-mediated virus entry . The MA1 and MA2 pps reached 50% entry in CD9-replete cells in 20 and 19 minutes , but were delayed to 34 and 30 minutes , respectively , in CD9KD cells ( Fig 6C and 6D ) . These data were corroborated with the tetraspanin KO cell lines ( S4 Fig ) . Thus , we conclude that CD9 utilization and rapid cell entry correlate with mouse adaptation and MERS virulence in mouse lung infections .
In this study , we demonstrated that the MERS coronavirus enters cells using an entry complex that includes a receptor , a protease and a tetraspanin . The tetraspanin operates by bringing the receptor and protease ( s ) into proximity , such that viral spikes , once attached to receptors , are quickly and efficiently cleaved into fusion-activated forms . These ternary complexes pre-exist on virus-target cells and can theoretically have highly variable subunit composition . Presently , there are three known human CoV receptors , 17 human transmembrane serine proteases , and 34 human tetraspanins , and therefore there are thousands of potential combinations of receptor , protease and tetraspanin that might provide a coronavirus entry platform . For the MERS coronavirus , a particularly effective complex includes the hDPP4 receptor , the protease TMPRSS2 , and the tetraspanin CD9 . This was discovered , in large part , through creative use of recombinant adenoviruses ( rAds ) . The approaches used here extended from the finding that a transducing rAd5 expressing the MERS-CoV receptor hDPP4 sensitized laboratory mice to MERS-CoV infection [26] . By incorporating RNA silencing genes into rAd5-hDPP4 , we were able to simultaneously establish MERS-CoV susceptibility , through hDPP4 expression , and potentially restrict MERS-CoV , through shRNA-mediated suppression of candidate proviral factors . Thus , the dual-expressing rAd5 vectors revealed CD9 and TMPRSS2 as relevant proviral factors , operating to support the primary hDPP4 susceptibility factor . It is notable that dual-expressing adenovirus vectors can potentially be utilized to identify any MERS-CoV host factor . Indeed , they can be utilized to identify host factors in any virus-host system that requires an exogenously supplied susceptibility determinant . Furthermore , the adenovirus transduction process bypasses the need to establish partially humanized mice for studies of human coronavirus infections , and actually has distinct advantages over transgenic animals , in that shRNAs reduce pro- or anti-viral factors solely in coronavirus-infectable cells , making for reliable measurements of changes in virus susceptibility . We expect that dual-expressing rAd5 vectors will be excellent general tools to rapidly identify in vivo pro- and anti-viral host factors . The hDPP4:CD9:TMPRSS2 complexes promoted an “early” MERS-CoV entry . CoV S proteins require simultaneous receptor engagement and proteolysis before catalyzing virus-cell membrane fusion [3 , 47 , 48] , a process demanding that TMPRSS2 be closely juxtaposed to hDPP4 . We suggest that CD9 tetraspanins position TMPRSS2 next to the receptor-bound S proteins , perhaps in association with cholesterol , a lipid having profound effects on both tetraspanin structural interactions [49] and CoV entry [50–54] . Precisely how TMPRSS2 abuts against hDPP4 to access S proteins is not clear , although the structures of three CoV S proteins [55 , 56] and hDPP4 [57] indicate that the proteolytic cleavage sites would be displayed at the outer edges of each S trimer . Furthermore , it is likely that proteolytic cleavage of several adjacent S proteins is needed to activate membrane fusion , as cooperative “pulling” by several viral fusion proteins is frequently required for virus entry processes [12 , 13 , 58] . Therefore , the tetraspanin-enriched environment , in which DPP4 and TMPRSS2 are collected together , likely permits rapid and simultaneous cleavage of multiple , closely-spaced virion S proteins , generating clusters of activated S proteins that drive the membrane fusion process . Without CD9 , the hDPP4 and TMPRSS2 are not held closely together ( Fig 3 ) . In this condition , MERS-CoV still infects hDPP4-positive target cells ( Fig 3 ) , but it takes a slower “late” endosomal route , which we and others find to be around 90% less efficient than early entry ( Figs 1 , 5 and 6 ) [25] . In the late entry route , virus-associated S proteins are first endocytosed and then cleavage-activated by furin proprotein convertases [59 , 60] , cathepsin L [8 , 25 , 61 , 62] , and/or cathepsin B [63] . However , the protease-enriched endo-lysosomal environment [64] can also generate inactivating CoV S protein cleavages , as evidenced by C-terminal S protein fragments , 40 kDa and smaller , that must be inactivated fusion domain fragments [48 , 65] . Therefore , in the late entry route , there may be a short time span between a cathepsin-activated fusogenic state and a permanently inactivated , excessively proteolyzed state , accounting for inefficient entry . Inefficient late entry may also be explained by differences in lysosomal and plasma membranes , which have unique lipid profiles [66] and therefore may be differentially susceptible to S -catalyzed fusion . Finally , late entry is restricted by interferon-induced gene products , notably interferon-induced transmembrane ( IFITM ) proteins [67 , 68] , but early TTSP-facilitated entry is not [8] . All of these virus-restricting conditions may combine in vivo to make CD9-facilitated “early” cell entry the predominant route for MERS-CoV infections . That the TTSP-facilitated entry route is predominant in vivo is supported by the recent finding that serine protease inhibitors reduce SARS-CoV infection in mouse lungs [36] . Additionally , clinical HCoV-229E isolates use a rapid TTSP-facilitated entry route , unlike lab-adapted HCoV-229Es [45] . More recently , similar patterns were observed for MERS-CoVs . Mouse lung-adapted MERS-CoVs take a rapid TMPRSS2-mediated cell entry , while cell culture-adapted ( CCA ) MERS-CoVs are avirulent and enter cells through the slower and less efficient endocytic route [44] . Here we demonstrated that the virulent MA MERS-CoV S proteins utilized CD9 during cell entry , while avirulent CCA viruses did not . This new finding suggests that CoV receptors and proteases alone are not the selective agents in CoV adaptation . Rather , the CoVs adapt to the ternary receptor-tetraspanin-protease complexes . In the case of MERS-CoV , the key adaptive S mutation facilitating the usage of the ternary complexes was at position 1015 ( S2 Table ) . In the MERS-CoV S protein cryo-EM structure [55 , 56 , 69] , this residue 1015 is part of a peptide that connects two of the helices comprising the fusion domain . The change from N1015 to T may ease restrictions to conformational change in the S trimer , thereby exposing cleavage sites to the nearby CD9-associated transmembrane protease , with cleaved spikes then converting to fusion-active forms . Finally , these findings may shed light on general roles for tetraspanins in virology . Four CoV receptors ( DPP4 , APN , ACE2 , and CEACAM ) were found in tetraspanin-rich membrane fractions ( Fig 4 ) , and our previous report indicated that tetraspanin antibodies block several CoV infections by interfering with receptor-associated CoV access to surface proteases [14] . Even antibodies binding to CD81 suppressed MERS S-mediated entry [14] , indicating that several tetraspanins , including those that are not required per se for clustering hDPP4 and TMPRSS2 , organize into cell-surface “webs” [15] and enclose the CoV entry factors . Here , there may be parallels with several tetraspanin-facilitated viruses , including influenza A ( IAV ) [28] and canine distemper ( CDV ) [70]; the retroviruses HIV [71 , 72] , feline immunodeficiency virus ( FIV ) [73] , and human T-lymphocytic virus 1 ( HTLV-1 ) [74]; herpes simplex virus 1 ( HSV-1 ) [75]; hepatitis C virus [76]; and several human papillomaviruses ( HPVs ) [77] . For these viruses , tetraspanins facilitate viral entry ( CoVs , IAVs , HCV , HPVs ) syncytia formation ( CDV , HIV , FIV , HTLV-1 ) , or promote viral exit ( IAVs , HSV-1 and HIV ) , by unclear mechanisms . Conceivably , a common mechanism may involve tetraspanin-mediated clustering of host factors . For example , tetraspanin CD81 is both an HCV receptor [78] and a linker of the HCV co-receptors scavenger receptor class B I ( SR-BI ) [76] and claudin-1 [79] , whose complexing promotes viral endocytosis ( reviewed in [80] ) . Another example is with the tetraspanins CD151 and CD63 , which do not directly interact with HPVs , but rather hold several co-receptors together to permit HPV binding and endocytosis ( reviewed in [81] . Therefore , many of the proviral activities ascribed to tetraspanins may relate to their ability to cluster transmembrane proteins , as we have found for the pro-MERS-CoV activity of CD9 . Given that several viruses depend on tetraspanin webs , it may be useful to consider ways to target entry-blocking drugs to these locations and thereby increase their antiviral efficacy .
C57BL/6 mice were purchased from the National Cancer Institute and housed in the animal care facility at the University of Iowa . The MERS-CoV ( EMC2012 strain ) was provided by Drs . Bart Haagmans and Ron Fouchier ( Erasmus Medical Center ) . HEK293T and HeLa cells were obtained from Dr . Edward Campbell ( Loyola University Chicago ) and maintained in Dulbecco's Modified Eagle Media ( DMEM ) supplemented with 10% fetal bovine serum ( FBS , Atlanta Biologicals ) , 10 mM HEPES , 100 mM sodium pyruvate , 0 . 1 mM non-essential amino acids , 100 U/ml penicillin G , and 100 μg/ml streptomycin . LET-1 cells were obtained from BEI Resources and were maintained in DMEM supplemented with 10% FBS , 100 U/ml penicillin G , and 100 μg/ml streptomycin . Cells were maintained in a humidified environment at 37°C and 5% CO2 . HAE cultures were isolated and maintained as described previously [48] . This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . Animal experiments were approved by the Institutional Animal Care and Use Committee at the University of Iowa ( Protocol #4041009 ) . Codon-optimized MERS-CoV S containing a C9 tag was purchased from Genscript and subsequently cloned into pcDNA3 . 1+ between the EcoRI and NotI restriction sites . pcDNA3 . 1-229E-Spike-C9 and pcDNA3 . 1-hAPN plasmids were provided by Dr . Fang Li , ( University of Minnesota ) . pcDNA3 . 1-SARS-Spike-C9 and pcDNA3 . 1-ACE2-C9 plasmids were provided by Dr . Michael Farzan ( Scripps Research Institute ) . C-terminal FLAG-tagged human DPP4 plasmid pCMV6-Entry-hDPP4 ( NCBI Reference Sequence NM_001935 ) was purchased from OriGene . pCAGGS-TMPRSS2-FLAG was previously constructed [82] . The pNL4 . 3-HIVluc plasmid was provided by the NIH AIDS Research and Reference library . pCMVSport6-hCD9 was purchased from Open Biosystems . pSpCas9-BB-2A-puro was a gift from Feng Zhang ( Addgene plasmid # 52961 ) . psPAX2 was a provided by Dr . Ed Campbell ( Loyola University Chicago ) . For transfections , plasmid DNAs were incubated with polyethylenimine ( PEI , Polysciences Inc . , Warrington , PA ) , at 1:3 DNA:PEI mass ratio , in Opti-MEM ( Life Technologies , Carlsbad , CA ) for 15 min at room temperature ( RT ) , then added dropwise to adherent cells ( 2 μg DNA per 106 cells ) . Monoclonal mouse antibodies against CD9 ( clone M-L13 ) , CD63 ( clone H5C6 ) , and CD81 ( clone JS-81 ) were obtained from BD Pharmingen . Rabbit anti-FLAG was obtained from Sigma Aldrich . Mouse anti-rhodopsin antibodies were obtained from Millipore . Rabbit anti-CD13 ( APN ) antibodies were obtained from Abcam . Mouse anti-CD26 ( clone M-A261 ) was obtained from BD Biosciences . Rabbit anti-TMPRSS2 ( clone EPR3681 ) was obtained from Abcam . Secondary antibodies were purchased from Invitrogen and include goat-anti-rabbit-AlexaFluor 488 , goat-anti-mouse-AlexaFluor 488 , and goat-anti-mouse-AlexaFluor 568 . Donkey-anti-goat , goat-anti-mouse , and goat anti-rabbit HRP conjugated antibodies were purchased from Thermo Scientific . Recombinant adenovirus vectors were produced as previously described by the University of Iowa Gene Transfer Vector Core [83] . To generate TMPRSS2-expressing adenoviruses , hTMPRSS2 containing a C-terminal FLAG tag was cloned into the pAd5CMV shuttle vector between XhoI and EcoRI restriction sites . To generate shRNA—expressing adenoviruses , gene blocks containing an shRNA targeting either the coding region of CD9 ( target sequence: CCGATTCGACTCTCAGACCAA ) or the 3' UTR of TMPRSS2 ( target sequence: ACACTAGAGTGGATGAATGTCTGGA ) , flanked by the U6 promoter and RNApolIII termination sequence , were purchased from GenScript . These gene blocks were subcloned into the pacAd5k-NpA E1 shuttle vector between the KpnI and EcoRI restriction sites . Shuttle vectors were linearized and transfected into HEK 293 cells along with a linearized Ad5 backbone containing an RSV promoter -driven hDPP4 in the E3 region . Homologous recombination in HEK 293 cells yielded recombinant adenovirus encoding both the shRNA and hDPP4 . Titers of purified recombinant adenoviruses ranged from 1010−1011 pfu/ml . Isoflurane-anesthetized mice were transduced intranasally with 2 . 5 x 108 pfu of the indicated Ad5 virus in 75 μl of DMEM . 5 days posttransduction , mice were infected intranasally with 105 pfu of MERS-CoV in a total volume of 50 μl DMEM . At 2 d . p . i . , mice were euthanized by isoflurate inhalation followed by cervical dislocation . Lungs were removed into PBS and manually homogenized . Virus was plaqued on Vero81 cells . Cells were fixed with 10% formaldehyde and stained with crystal violet 3 d . p . i . All work was performed in the University of Iowa Biosafety Level 3 ( BSL3 ) Laboratory . HIV pseudoviruses were produced as previously described [84] . Briefly , 293T cells were co-transfected with pNL4 . 3-HIV-luc and pcDNAs encoding appropriate glycoproteins . After two days , supernatants were collected , centrifuged at 10 , 000 x g at 4°C for 10 minutes to remove debris , and stored in aliquots at -80°C . VSV pseudoviruses were produced by the methods of Whitt , 2010 [40] . Briefly , 293T cells were transfected with plasmids encoding viral glycoproteins . Two days later , cells were inoculated for 2h with VSVΔG-luciferase [40] , rinsed extensively and incubated for one day . Supernatants were collected , centrifuged at 10 , 000 x g at 4°C for 10 minutes to remove cellular debris , and stored in aliquots at -80°C . Pseudovirus transductions were carried out by incubating target cells with pseudoviruses for 1h at 37°C . Following initial incubation , unadsorbed viruses were removed by washing thrice with PBS . Complete media was placed on the cells and incubated for 18h for VSV or 48h for HIV at 37°C . At the end of transduction periods , cells were dissolved into cell culture lysis buffer ( 25 mM Tris-phosphate [pH 7 . 8] , 2 mM DTT , 2 mM 1 , 2-diaminocyclohexane-N , N , N ′ , N ′-tetraacetic acid , 10% glycerol , 1% Triton X-100 ) and luciferase levels were measured by addition of firefly luciferase substrate ( 1 mM D-luciferin , 3 mM ATP , 15 mM MgSO4·H2O , 30 mM HEPES [pH 7 . 8] ) using a Veritas microplate luminometer ( Turner BioSystems , Sunnyvale , CA ) . pSpCas9-BB-2A-puro was digested with Esp3I ( Fermentas ) for 4h at 37°C . The digested plasmid was purified and ligated with annealed guide DNAs specific for CD9 or CD81 . Tetraspanin-specific pSpCas9-BB-2A-puro plasmids were transfected into 293T cells . After 72h , cells were selected with 4 μg/ml puromycin for 96h . Selected cells were serially-diluted to isolate clonal populations and clones were selected by western blot . Adherent 293T cells ( ~105 / cm2 ) were rinsed with ice-cold PBS , incubated for 30 min at 4°C with PBS-1 mg/ml EZ-Link Sulfo-NHS-LC-Biotin ( Pierce ) , then for 20 min at 4°C with PBS-100 mM glycine . Cells were rinsed with PBS , then incubated for 20 min at 4°C in MES buffer ( 25 mM MES [pH 6 . 0] , 125 mM NaCl , 1 mM CaCl2 , 1 mM MgCl2 ) containing 1% 3-[ ( 3-Cholamidopropyl ) dimethylammonio]-1-propanesulfonate ( CHAPS ) detergent ( Calbiochem Cat # 220201 ) or 1% Triton X-100 detergent ( Sigma ) . Cell lysates ( 107/ml ) were removed from plates and emulsified by 20 cycles of extrusion through 27G needles . Nuclei were removed by centrifugation , lysates mixed with equal volumes of 80% w/v sucrose in MES buffer , placed into Beckman SW60 tubes , and overlaid with 3 ml of 30% w/v sucrose , then with 0 . 5-ml of 5% w/v sucrose , both in MES buffer . Samples were centrifuged with a Beckman SW60 rotor at 370 K x g for 18 h at 4°C . Fractions were collected from air-gradient interfaces . Biotinylated proteins in gradient fractions were bound to streptavidin agarose beads ( Pierce ) . Non-reducing western-blotting procedures were used to identify the distributions of proteins in gradient fractions , as described previously [38] . HeLa cells were transfected with indicated plasmid DNAs and a GFP reporter , incubated for two days , and then lifted from tissue culture plates using 0 . 05% trypsin . Cells were transferred to microscope coverslips coated with fibronectin . Cells were allowed to adhere for 24h . Cells were then fixed for 30 minute at 37°C with 3 . 7% paraformaldehyde in 0 . 1 M piperazine-N , N′-bis ( 2-ethanesulfonic acid ) buffer ( pH 6 . 8 ) . Coverslips were washed with PBS and PLA was performed using DuoLink Proximity Ligation Assay ( Sigma-Aldrich ) using primary antibodies against TMPRSS2 and CD26 . Images were captured using a DeltaVision microscope ( Applied Precision ) equipped with a digital camera ( CoolSNAP HQ; Photometrics ) , using a 1 . 4-numerical aperture 60X objective lens . Images were deconvoluted with SoftWoRx deconvolution software ( Applied Precision ) . PLA foci were detected and quantified using Imaris version 6 . 3 . 1 ( Bitplane Scientific Solutions ) . 293T cells were transfected with DPP4 and either an empty vector or complementing tetraspanin . 24h after transfection , cells were plated in a 96-well plate . MERSpps were added to cells at 4°C for 1 hour to allow viral binding . Media was removed and replaced with 37°C media and the plates were moved to an incubator . At sequential time points following the shift to 37°C , a protease inhibitor cocktail was added to cells such that the final concentration was 100 μM camostat , 10 μM proprotein convertase inhibitor , and 10 μM E64d . These drugs were left on cells overnight before cells were lysed and luciferase was measured as described above . Luciferase levels were compared to that of cells treated only with DMSO control . 293T cells were transfected with DPP4 and an empty vector or the complementing tetraspanin . Cells were pre-treated for 1h with 100 μM camostat , 100 μM bafilomycin , or 10 μM E64D before transduction with MERSpps in the presence of the inhibitors . After 2h , cells were washed to remove drugs and unadsorbed virus . Luciferase assays were performed as described above . Cellular RNA was isolated using the RNeasy Mini Kit ( Qiagen ) and 100 or 500 ng was reverse transcribed using an iScript cDNA synthesis kit ( Bio-Rad ) . Quantitative PCR was performed using Power SYBR Green ( Thermo Fisher ) and primers specific to human CD9 , DPP4 , TMPRSS2 , or HPRT . Statistical comparisons were made by two-tailed Student’s t-test . Error bars in the figures indicate the standard error of the data . Non-linear regression analysis was used to fit a curve to the entry kinetics data and obtain the time of 50% infection . This analysis was performed using Minitab 17 software . | Enveloped viruses rank among the most dangerous zoonotically emerging pathogens . Their cell entry often requires multiple transmembrane proteins in the target cell , which may interact with each other to promote viral-cell membrane fusion . Susceptibility to virus infection may correlate with these transmembrane protein interactions . Here we report that the scaffolding tetraspanin protein CD9 links the receptor for MERS-CoV to a membrane fusion-activating protease called TMPRSS2 , forming a complex that promotes rapid and efficient infection . The related human CoV strain 229E was also facilitated by CD9 , indicating that multiple CoVs depend on tetraspanin-directed clustering of receptors and proteases for efficient cell entry . Reliance on CD9 specifically applied to virulent , in vivo mouse lung-adapted MERS-CoVs , suggesting that the most efficient virus entry pathways in natural respiratory CoV infections are facilitated by tetraspanins . This suggestion was reinforced by selectively regulating gene expression in vivo , using recombinant adenovirus transducing vectors . The findings demonstrated that CD9 facilitated MERS-CoV infections in mice . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"coronaviruses",
"pathology",
"and",
"laboratory",
"medicine",
"viral",
"transmission",
"and",
"infection",
"enzymes",
"pathogens",
"293t",
"cells",
"biological",
"cultures",
"microbiology",
"enzymology",
"viruses",
"membrane",
"... | 2017 | The tetraspanin CD9 facilitates MERS-coronavirus entry by scaffolding host cell receptors and proteases |
Gene regulatory networks consist of direct interactions but also include indirect interactions mediated by metabolites and signaling molecules . We describe how these indirect interactions can be derived from a model of the underlying biochemical reaction network , using weak time-scale assumptions in combination with sensitivity criteria from metabolic control analysis . We apply this approach to a model of the carbon assimilation network in Escherichia coli . Our results show that the derived gene regulatory network is densely connected , contrary to what is usually assumed . Moreover , the network is largely sign-determined , meaning that the signs of the indirect interactions are fixed by the flux directions of biochemical reactions , independently of specific parameter values and rate laws . An inversion of the fluxes following a change in growth conditions may affect the signs of the indirect interactions though . This leads to a feedback structure that is at the same time robust to changes in the kinetic properties of enzymes and that has the flexibility to accommodate radical changes in the environment .
The adaptation of bacteria to changes in their environment involves adjustments in the expression of genes coding for enzymes , regulators , membrane transporters , etc . [1]–[3] . These adjustments are controlled by gene regulatory networks ensuring the coordinated expression of clusters of functionally related genes . The interactions in the network may be direct , as in the case of a gene coding for a transcription factor regulating the expression of another gene . Most of the time , however , regulatory interactions are indirect , e . g . when a gene encodes an enzyme producing a transcriptional effector [4] . A gene regulatory network can thus not be reduced to its transcriptional regulatory interactions: by ignoring indirect interactions mediated by metabolic and signaling pathways we may miss crucial feedback loops in the system . The network controlling carbon uptake in the bacterium Escherichia coli is a good example because it integrates metabolism , signal transduction , and gene expression . At the level of gene expression , the network includes intricate feedback loops that arise from indirect interactions between the subsystems . Global regulators like Crp control expression of enzymes in carbon metabolism [5]–[8] , while intermediates of the latter pathways control the expression of global regulators . For instance , the phosphorylation of EIIA activates adenylate cyclase ( Cya ) to produce cAMP which is required for the activation of Crp [9] , [10] . The aim of this paper is to develop a method for the systematic derivation of direct and indirect interactions in a gene regulatory network from the underlying biochemical reaction network . Due to the complexity of the intermediate metabolic and signaling networks , determining indirect interactions is difficult in general . We show that model reduction based on quasi-steady-state ( QSS ) approximations expressing weak assumptions on time-scale hierarchies in the system [11]–[13] , together with sensitivity criteria from metabolic control analysis ( MCA ) [12] , [14] , are able to uncover such interactions . Indeed the MCA formalism uniquely allows to relate systemic sensitivities ( ‘control coefficients’ ) with the sensitivities of individual reactions to reactants and effectors [12] , [15] . It therefore provides a proper framework for investigating metabolic effects in gene regulation . We apply our approach to a model of the upper part of the carbon assimilation network in E . coli , consisting of the glycolysis and gluconeogenesis pathways and their genetic and metabolic regulation . The analysis of the derived gene regulatory network leads to three new insights . First , contrary to what is often assumed , the network is densely connected due to numerous feedback loops resulting from indirect interactions . This additional complexity is an important issue for the correct interpretation of data from genome-wide transcriptome studies . Second , the derived gene regulatory network for carbon assimilation in E . coli is sign-determined , in the sense that the signs of interactions are essentially fixed by weak information on flux directions of biochemical reactions , without explicit specification of kinetic rate laws or parameter values . Therefore the feedback structure is robust to changes in kinetic properties of enzymes and other biochemical reactions species . Third , a change in environmental conditions may invert fluxes , and thus the signs of indirect interactions , resulting in a dynamic rewiring of the regulatory network .
We used standard approaches from biochemistry to build a kinetic model of the network of glucose assimilation in E . coli . The model describes the genetic and metabolic regulation of glycolysis and gluconeogenesis . The model takes the form of a system of ordinary differential equations ( ODEs ) , describing the rate of change of the concentrations of proteins , RNAs and metabolites: ( 1 ) denotes the vector of concentrations and the vector of reaction rates . is a stoichiometry matrix . In the presence of conserved quantities , is reformulated in such a way that the dependencies between variables are eliminated [16] . In the following , we assume that is such a reduced matrix . Eq . 1 can be simplified by applying the QSS approximation [12] . Two different time-scales are distinguished , one corresponding to the slow processes ( protein synthesis and degradation ) and one to the fast processes ( complex formation and enzymatic reactions ) . Considering metabolic processes as fast is justified when metabolic pools undergo turnover times in the range of seconds , as is the case for the very active glycolysis in E . coli [17] . Therefore , we introduce vectors of slow and fast variables , and , respectively ( ) , defined as linear combinations of the original variables : ( 2 ) with . The slow variables typically correspond to total protein concentrations , whereas the fast variables include concentrations of metabolites and biochemical complexes: ( 3 ) ( 4 ) where and are stoichiometry matrices for the slow and fast part , respectively , and and the corresponding reaction rates ( see Sec . 1 of Supporting Information Text S1 for details ) . The QSS hypothesis states that at the time-scale of the slow processes , the fast part of the system can be assumed to be at steady state , instantly adapting to the dynamics of the slow variables , i . e . . The conditions for the applicability of this approximation are given by the Tikhonov theorem , which imposes exponential stability of the fast system [12] . The stability of metabolism in its normal range of operation is a reasonable assumption in most situations . The QSS approximation implicitly relates the steady-state values of the fast variables to the concentrations of the slow variables , i . e . , , if such a function can be found . The resulting system at the slow time-scale has the following form ( 5 ) This reduced model makes explicit the fact that the biochemical reactions in the fast subsystem induce additional interactions between the slow variables . For metabolic systems the QSS equation is nonlinear in terms of and it is generally impossible to obtain a closed-form expression for the function . We therefore follow another strategy to characterize the indirect interactions between the slow variables , that is , the regulation of gene expression via metabolic intermediates . We study the Jacobian matrix of the system in Eq . 5 , which captures the interaction structure of the gene regulatory network: ( 6 ) The Jacobian matrix includes the direct effect of each slow variable on the others ( first term ) and the indirect effect via the coupling through the fast system ( second term ) . It accounts for direct regulation of gene expression by transcription factors as well as indirect regulation through metabolism . The indirect regulation involves both the effect of changes in fast variables on the rates of slow variables ( ) and the effect of changes in slow variables on QSS values of fast variables ( ) . The former effect can be directly determined from the rate equations , as it describes , for instance , the regulation of a gene by a metabolic effector . The latter effect expresses the sensitivity of the metabolic state to changes in the slow variables , which corresponds to concentration control coefficients in the framework of MCA [12] , [16] . Implicit differentiation of the QSS equation results in ( 7 ) which describes the response of the fast system around its steady state to changes in the slow variables . Notice that corresponds to the Jacobian matrix of the fast system . The reduction of for conserved quantities assures that is not singular ( see Sec . 1 of Text S1 ) . Therefore , if the steady state is stable then , using Eq . 7 and the definition of , concentration control can be expressed [16] as ( 8 ) The latter formula can then be substituted into Eq . 6 , the expression of the Jacobian matrix of the slow system . is the matrix of non-normalized concentration control coefficients [12] , [16] . The computation of as described above requires the manipulation of complex algebraic expressions . As this is too cumbersome and error-prone to do by hand , the process has been implemented by means of the Symbolic Math Toolbox of MATLAB ( MathWorks ) . Inversion of large symbolic matrices like is a computationally challenging task , but the matrices considered in the E . coli example are within the reach of state-of-the-art computer algebra tools . The computations take a few seconds to complete on a PC ( Intel Core 2 , 1 . 86 GhZ , 2 Gb of RAM ) . The rate vectors and are typically nonlinear functions involving many parameters with unknown values . However , since and are usually monotonic functions of the variables , the signs of the partial derivatives in Eqs . 6 and 8 are fixed over the entire state space . This information can be used to evaluate the sign of the elements of . This argument can be clarified by considering the partial derivatives of the rates occurring in Eqs . 6 and 8 ( see Fig . 1 for a schematic illustration ) . describes the direct interactions between slow variables , typically the control of gene expression by a transcriptional regulator . The signs of these interactions are in general unambiguously given by the literature [5] . We omit the special case of non-specific degradation and growth dilution , which are not usually interpreted as regulatory interactions [18] . describes the direct relations between the fast and slow parts of the system through fast coupling species , e . g . , a transcriptional regulator whose activity is modified by a metabolite: their signs are known . accounts for the direct influence of fast variables on the fast dynamics , typically the variation of enzyme rates with a change in concentration of substrate , product , or effector . Given a convention on the positive flux direction , the signs of these elasticities are usually unambiguously defined , except in rare cases of substrate inhibition or product activation . In such a case our analysis pertains provided the ranges of concentrations are restricted so that enzyme rates remain monotonic functions of concentrations . Finally , describes the direct influence of slow variables on the fast dynamics , typically the variation of a reaction rate with a change in enzyme concentration . In this case is positive because absolute values of reaction rates increase with enzyme concentration , so that the sign of this effect is solely determined by the direction of the flux ( equal to its sign ) . Therefore a change in growth conditions implies a switch of the signs of some interactions , whenever there is a change in flux direction . For instance in the carbon assimilation model , different regulatory patterns will emerge depending on whether the bacteria grow on glycolytic or gluconeogenic substrates . When do the signs of the partial derivatives of the rates unambiguously fix the signs of the structure of interactions between the slow variables ? Analysis of the Jacobian matrix in Eq . 6 reveals that the following four conditions are sufficient to obtain what we call a sign-determined network ( see Sec . 2 of Text S1 ) . ( C1 ) A slow variable acts directly either on the slow system or on the fast system , but not on both simultaneously . In practice this excludes enzymes as transcriptional regulators or moiety conserved species as transcriptional effectors . Under this condition at most one of the terms in Eq . 6 is non-zero for each element of . ( C2 ) No variable has direct antagonistic ( i . e . , both activating and inhibiting ) effects on a slow variable . This means , for example , that a transcription factor cannot both activate and inhibit the expression of the same gene ( no mixed regulation ) , although it may activate one gene and inhibit another . ( C3 ) The concentration control coefficients of the fast coupling species with respect to the slow variables have a determinate sign . ( C4 ) If a slow variable contributes to the concentration control of several fast coupling species , the latter do not simultaneously regulate any of the slow variables ( no concerted regulation ) . Together C3 and C4 guarantee that the second term in the right-hand side of Eq . 6 is unequivocally defined . Fig . 2 illustrates the four conditions C1–C4 in terms of allowed and forbidden patterns in the biochemical reaction system . Notice that these conditions do not give the actual signs of the elements of , but they help in relating the sign- ( un ) determinedness of the network to specific features of the underlying biochemical reaction system . Whereas C1 and C2 are not very restrictive , the satisfaction of especially C3 is not evident in practice . In the case of a metabolic network with a complex structure , involving substrate cycles or allosteric regulation , antagonistic effects may compete in the control of concentration . Such situations were analyzed previously in the framework of MCA [19] . For instance , the signs of concentration control coefficients are frequently undetermined for metabolites on the path between an allosteric effector and its target . Another case of undeterminedness concerns substrate cycles . Whenever such antagonistic effects arise , additional information will be required on the relative magnitudes of opposing effects . The stability underlying the QSS approximation imposes additional constraints that can be exploited to resolve ambiguities . A classical result in linear system theory [20] states a necessary condition for the stability of the fast system , namely that the coefficients of the characteristic polynomial ( 9 ) all have the same sign . This provides an independent set of inequalities between partial derivatives that can be used to estimate the signs of control coefficients in Eq . 8 and thus satisfy C3 . Fig . 3 shows the network of direct and indirect gene regulatory interactions computed for a simplified model of the carbon assimilation network . The model describes the main reactions involved in the control of the glycolysis pathway , during growth on glucose ( Fig . 3A ) . In particular it accounts for the genetic regulation of enzymes levels , and thus provides an interesting example for the analysis of indirect interactions arising from the coupling between gene expression and metabolism . The corresponding ODE system , written in the form ( 3 ) – ( 4 ) , is shown in Fig . 3B . Application of the method explained above results in the appearance of novel interactions between genes fbaA and pykF , mediated by the fast coupling species free FruR ( see Fig . 3C ) . These interactions are not expected on the basis of a purely transcriptional control . The derivation of the interactions from the model are described in detail in Text S2 . In this case , the stability condition is sufficient to satisfy all conditions and make the network sign-determined .
Glucose is the preferred carbon source of E . coli and its assimilation is tightly regulated in the cell . This control involves a signaling pathway and transport system ( PTS ) , a modification of metabolic activities ( glycolysis , TCA cycle , pentose-phosphate pathway , gluconeogenesis ) , and the regulation of gene expression ( glycolytic and gluconeogenic enzymes , global regulators ) . These different modes of control have mostly been studied in isolation , whereas in fact they are interwoven and form a large and complex regulatory network . In this study we focus on the part of the regulatory network controlling glycolysis and gluconeogenesis . Briefly , this network accounts for the sensing and uptake of glucose via the PTS , its conversion to pyruvate , as well as the regeneration of more complex sugars from pyruvate when the latter is used as a carbon source ( Fig . 4 ) . At the level of gene expression we consider genes coding for metabolic enzymes and their key regulators , fis , crp , and fruR [6] , [8] . In addition , we include the general stress factor RpoS and the regulators of DNA topology ( GyrAB , TopA , … ) , as changes in the superhelicity of DNA affect the expression of many of the above-mentioned genes [7] . Changes in gene expression modify the concentrations of enzymes , and thus of intracellular fluxes and metabolite concentrations . A critical point in the regulation of carbon assimilation is the pair of reactions interconverting PEP and Pyr , involving the differentially regulated enzymes PykF and PpsA , required respectively for glycolysis and gluconeogenesis [21] , [22] . Metabolism also acts back on gene expression . For instance , FBP and cAMP are two key metabolites that modulate the activity of the transcription regulators FruR and Crp , respectively [6] , [10] , [23] , [24] . The PTS plays a special role in this context by converting information on glucose availability into an activation signal for cAMP synthesis , thus inducing a reorganization of global gene expression by CrpcAMP [10] , [25]–[27] . We have developed a model that describes the coupling between metabolism and gene expression , consisting of 66 reactions and involving 40 species . The model is based on existing models of carbon metabolism [27] , [28] and global regulators of gene expression [29] , which include the experimentally validated interactions reported in the literature ( see Text S3 ) . However , contrary to these models , we do not specify kinetic rate laws , as only the signs of the partial derivatives are used for reconstructing the ( signs of ) indirect interactions . We apply the QSS approximation by distinguishing two distinct time-scales in the system: a fast time-scale for complex formation , DNA supercoiling and all reactions involved in glycolysis , gluconeogenesis , PTS signaling , and cAMP production , and a slow time-scale for the synthesis and degradation of global regulators , enzymes and stable RNAs . The equations of the original and the reduced model , as well as the different approximation steps , are described in detail in Secs . 1 and 2 of Text S3 . For analytical purposes , four variants of the model are analyzed below , accounting for differences in growth conditions and regulatory effects . The differences concern only a few of the 66 reactions . We consider two possible carbon sources , glucose or pyruvate , thus imposing a fixed direction on reactions . Some reactions have negligible flux , such as the PEP synthase during glycolysis [21] . Glycolysis and gluconeogenesis are therefore treated separately by two distinct models and . For each of these we define two variants that do not or do include allosteric regulation of enzyme activities: , , respectively , for glycolysis and , , respectively , for gluconeogenesis . The coupling between metabolism and gene regulation leads to additional , indirect dependencies between genes . We first focus on the networks obtained in the absence of allosteric regulation , using models and . Application of the method introduced above to the glycolytic model , as described in Sec . 3 of Text S3 , results in the sign pattern of the Jacobian matrix in Table 1 . Several novel indirect interactions appear , some of which are straightforward , like the inhibitory effect of Crp on cya through CrpcAMP . Others , however , are less evident or even counter-intuitive such as the predicted negative control of the expression of the global regulator FruR by enolase ( Eno ) during growth on glucose . This effect is explained by the fact that an increase in eno expression leads to a reduced FBP concentration , and thus to an increased fruR downregulation . The most striking result of our analysis is that the signs of the indirect interactions are uniquely defined , that is , during growth on glucose , the proteins exert an unambiguous effect ( zero , positive or negative ) on their target genes . The signs of these indirect interactions are therefore a structural property of the underlying system of biochemical reactions . The same result is observed in the case of growth on pyruvate , for the gluconeogenic model ( Sec . 3 of Text S3 ) . The sign-determinedness of the network can be analyzed by means of the conditions C1–C4 . satisfies all sufficient conditions for sign-determinedness . In particular , the concentration control coefficients acting on coupling species have a unique sign , as requested by C3 . satisfies C1–C3 , but not C4 . The concerted regulation excluded by C4 does not pose a problem for sign-determinedness in this particular case , however , because PykF has the same effect through both fast coupling species CrpcAMP and free FruR . Allosteric regulation is important for metabolism , but adds a level of complexity that may affect the sign-determinedness of the network . We verified this by applying the method to the glycolytic model with allosteric regulation , . The latter model notably includes the positive regulation of PykF activity by FBP [30] , [31] and the inhibitory effect of PEP on PfkA [27] . As a consequence of the feedforward loop from FBP to PykF , C3 and C4 do no longer hold for , and in fact the network becomes partially sign-undetermined . In particular , the glycolytic enzymes FbaA , GapA , Pgk , and Eno exert antagonistic effects on the control of the concentration of free FruR , thus invalidating C3 . Moreover , the presence of allosteric regulation results in a denser Jacobian matrix of the fast system . This causes some of the glycolytic enzymes to contribute to the control of both CrpcAMP and free FruR . Contrary to C4 , these fast coupling species simultaneously regulate the genes coding for three of the enzymes , in antagonistic ways . By means of conditions C1–C4 the partial sign-undeterminedness can thus be related to specific network features . Interestingly , it also enables one to identify experiments that would resolve sign ambiguities: indeed , a single observation , measuring the response of the FBP concentration to an increased expression of FbaA , would allow us to unequivocally determine the signs of all control coefficients ( Sec . 3 of Text S3 ) . Such an observation has been reported in the literature [32] and makes condition C3 true . The resulting signs of the control coefficients are the same as for the model without allosteric effects , thus indicating that the regulation of PykF activity by FBP finetunes rather than inverses the concentration control of the system . The derived gene regulatory network during glycolysis , after disambiguation of the concentration control coefficients , is shown in Fig . 5A . The experimental data do not resolve the ambiguities invalidating C4 . In particular , the regulation mediated by CrpcAMP leads to an activation whereas free FruR is responsible for a negative control . In this situation , the resulting net effect of these regulators on their targets cannot be predicted without information on the parameters or gene expression patterns under glycolytic growth conditions , and a double sign appears in Table 1 . Notice however that this concerns only 12 out of 256 entries in the Jacobian matrix describing the interaction structure . The network is found to be completely sign-determined in gluconeogenesis , even when taking into account allosteric regulation ( Sec . 3 of Text S3 ) . The above analysis is based on the assumption that the net flux direction is fixed , which means that the obtained network is growth-condition specific: some indirect interactions appear under one growth condition and are absent in the other ( Fig . 5 ) . Moreover , the same interaction may have an opposite sign in the two cases , for instance the effect of Eno on the concentration of free FruR . This context-dependency of the regulatory structure is due to the fact that the concentration control exerted by the glycolytic enzymes on free FruR and CrpcAMP , the two main connections between carbon metabolism and gene regulation , changes sign depending on whether the bacteria grow on glucose or pyruvate . More generally , it can be shown with MCA that concentration control coefficients change sign upon flux inversion , resulting in an inversion of the corresponding gene interactions . This shows that the structure of regulatory interactions may be dynamically rewired by the environment , which potentially enhances the adaptive capacity of the system . Classically , gene regulatory networks are considered to be sparsely connected , with only a few regulators per gene [33]–[36] . Most of these studies , however , have focused on direct transcriptional regulations , without considering the indirect interactions arising from the coupling between metabolism and gene expression . As these indirect interactions are operative on the time-scale of the slow variables , they can not be ignored when studying the dynamics of the gene regulatory network , for instance in the context of transcriptome studies . In order to assess the effect of including indirect interactions in the E . coli network , we have counted the average connectivity per gene and the number and the length of the feedback loops in the system ( Sec . 4 of Text S3 ) . We compare the results with a baseline model that only considers classical , direct interactions . The carbon assimilation network of the baseline model has an average connectivity of 1 . 4 regulatory proteins per gene . These values are in agreement with estimations made for E . coli and other organisms at the genomic scale [33]–[36] . Only four feedback loops are detected , most of which ( 3 out of 4 cases ) are cases of direct autoregulation . The addition of indirect interactions changes the picture completely ( Table 2 ) . The average connectivity rises to over 4 and the number and length of feedback loops increases dramatically . Some feedback loops involve 12 elements , that is , 75% of the genes in the network . The influence of metabolism on gene expression is channeled through a small number of intermediates , essentially CrpcAMP and free FruR . Leaving out one of these coupling species immediately reduces the number and length of the feedback loops . For instance , eliminating the indirect interactions associated with CrpcAMP reduces the number of feedback loops to a mere 20% of those present in Table 1 , and the maximal loop length drops from 12 to 6 . This agrees with the central role of CrpcAMP in the control of carbon assimilation in E . coli [5] , [6] . The effect of eliminating the interactions mediated by FruR is less dramatic , consistent with its more local role [5] , [6] , [37] . The comparison of the models with and without allosteric regulation ( vs , vs ) , shows a large increase in the number of feedback loops in the former ( Table 2 ) . This is intuitively expected from the fact that allosteric regulation allows a local perturbation to spread to remote parts of the network . As a consequence , it has a higher chance of affecting a fast coupling species . This increases the number of non-zero elements in , and thus on average the number of feedback loops .
The regulation of gene expression is tightly interwoven with metabolism and signal transduction . A realistic view of gene regulatory networks should therefore not only include direct interactions resulting from transcription regulation , but also indirect regulatory interactions mediated by metabolic effectors , as in the classical example of the lac operon [38] , [39] . We show here how such a regulatory network can be derived from the network of biochemical reactions in a mathematically rigorous way . Our approach starts from a model of the biochemical reaction system in the form of Eq . 1 . We reformulate this system into coupled fast and slow subsystems , by distinguishing between reactions that are fast and slow in the physiological range of interest , and by redefining fast and slow variables accordingly ( Sec . 1 of Text S1 ) . This is rather straightforward to achieve for the types of systems considered here , as enzymatic and complex formation reactions are typically fast on the time-scale of protein synthesis and degradation . Assuming that the fast subsystem is at quasi-steady state , the indirect interactions between genes are now defined by the Jacobian matrix in Eq . 6 . In order to derive the indirect interactions between genes by means of this matrix , the rate laws defining the reaction rates do not need to be specified: the dependencies of the reaction rates on metabolite and enzyme concentrations are sufficient . The signs of these partial derivatives are usually unambiguously defined once the metabolic flux directions are fixed . Their substitution into the symbolic expressions of the Jacobian matrix allows the computation of the global effect of a change in gene expression , if such an effect can be unambiguously determined . The advantage of this approach is that it does not require fully specified kinetic models with numerical values for the parameters , instead of weaker information on the signs of the partial derivatives ( see [40] for related ideas in a different context ) . This information may not be available and the results would be less generic , that is , only hold for these specific kinetic mechanisms and parameter values . Moreover , numerical calculation of requires the state space of the system to be sampled . For larger models with many variables , this may become very costly . For systems of the size studied in this paper , the derivation of the symbolic expressions does not pose computational problems , although this may change if still larger systems are considered . An interesting topic for further research would be the development of methods that combine symbolic and numerical computations in a clever way . The derivation of direct and indirect interactions between genes has been addressed before , notably by methods for the inference of networks from transcriptome and other high-throughput data ( see [1] , [41] , [42] for representative examples ) . Our approach is different from these methods in that it does not infer the interactions from experimental data , but rather starts with available knowledge on the underlying biochemical reaction system . The results are complementary , in the sense that we present a principled way to obtain a core structure of the network that can be completed or refined through data-driven inference procedures . Other related approaches are extensions of flux balance analysis ( FBA ) that aim at integrating gene regulation with metabolism ( e . g . , [43]–[45] ) . Gene regulation is modeled by Boolean rules and , like in our approach , the kinetic rate laws are not specified . The two approaches are quite different though . We do not aim at predicting flux distributions under different environmental conditions , but rather at eliciting indirect interactions between genes mediated by metabolism and to identify modifications of the interaction structure following changes in flux directions . Our approach can thus be seen as a model reduction that uncovers the effective network structure on the time-scale of gene expression . The indirect interactions are expected to have important consequences for the network dynamics , but we leave an analysis of these aspects for further work . Applied to the carbon assimilation network in E . coli our method shows that the resulting gene regulatory network is much more densely connected than the purely transcriptional regulatory network . We notably observe a strong increase of the average connectivity of the network and the number of feedback loops . The indirect interactions revealed by our analysis are operative on the time-scale of gene expression and therefore cannot be ignored . However , some of these may be too weak to be physiologically important , so the actual connectivity may be lower than predicted . In order to decide on the relative strength of the interactions , additional quantitative information is required . We are not aware of any systematic experimental studies to test the predicted indirect regulatory interactions , with the exception of transcriptome studies using deletion mutants . Notice that these results should be taken with some care for the validation of the derived indirect interactions , as the deletion of a mutant may change the direction of the fluxes and thus the sign of the interactions . In this case , the data agree well with the interaction matrix in Table 1 . For instance , our method correctly predicts that a pykF deletion leads to increased expression of fruR and decreased expression of cya during glycolysis [46] . Moreover , in a ppsA strain the expression of crp is lower during gluconeogenesis [47] , in agreement with the interaction matrix ( Sec . 3 of Text S3 ) . The most remarkable conclusion of our study of the E . coli network is that for given growth conditions , the signs of the indirect interactions are largely independent of the exact form of kinetic rate laws and precise parameter values . The fact that most interactions have an unequivocal sign was not expected on the basis of results obtained with similar approaches for the qualitative analysis of ecological and economic systems [48]–[50] . We have interpreted this surprising finding in terms of sufficient conditions for sign-determinedness . The conditions help us understand what causes most of the interactions in the E . coli network to be sign-determined and some of them to be sign-undetermined . The most important of these conditions is the requirement that the concentration control coefficients of the fast coupling species are unambiguously defined . This condition is indeed satisfied by three of the four models studied , but violated by the glycolysis model with allosteric effects , due to the regulation of PykF activity by FBP . The determinate sign of most of the indirect interactions is interesting , because it points at the robustness of the effective structure of this network to changes in the kinetic properties of enzymes and other biochemical reaction species . Another interesting finding is that radical changes in the environment , e . g . , the exhaustion of glucose , may invert the signs of indirect interactions , resulting in a complete rearrangement of the feedback structure of the E . coli gene regulatory network . The change in growth conditions affects the direction of the metabolic fluxes , which translates into a switch of the sign of some of the concentration control coefficients . Such an overall modification of the control architecture in response to environmental perturbations may be beneficial to the cell , as it increases its adaptive flexibility . Related to this , radical changes in the genetic background , e . g . , the knock-out of a particular gene , may also invert metabolic fluxes and thus change the sign or even the existence of indirect interactions . This may have important consequences for the interpretation of transcriptome data , which often take the form of knock-out datasets [1] . The approach described in this paper provides a sound methodological basis for investigating gene regulatory networks . Its application to E . coli carbon assimilation leads to novel insights into the structure of this network . How much of these carry over to other organisms ? While the increased connection density and the dependency of the interaction signs on the environmental conditions follow rather straightforwardly from the theory , there is no a priori reason why a network should be sign-determined . However , since sign-determinedness confers robustness to the regulatory structure of the system , an important functional requirement [51] , it may be more common than expected on purely mathematical grounds . | The regulation of gene expression is tightly interwoven with metabolism and signal transduction . A realistic view of gene regulatory networks should therefore not only include direct interactions resulting from transcription regulation , but also indirect regulatory interactions mediated by metabolic effectors and signaling molecules . Ignoring these indirect interactions during the analysis of the network dynamics may lead crucial feedback loops to be missed . We present a method for systematically deriving indirect interactions from a model of the underlying biochemical reaction network , using weak time-scale assumptions in combination with sensitivity criteria from metabolic control analysis . This approach leads to novel insights as exemplified here on the carbon assimilation network of E . coli . We show that the derived gene regulatory network is densely connected , that the signs of the indirect interactions are largely fixed by the direction of metabolic fluxes , and that a change in flux direction may invert the sign of indirect interactions . Therefore the feedback structure of the network is much more complex than usually assumed; it appears robust to changes in the kinetic properties of its components and it can be flexibly rewired when the environment changes . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"microbiology/microbial",
"physiology",
"and",
"metabolism",
"computational",
"biology/metabolic",
"networks",
"computational",
"biology/systems",
"biology",
"computational",
"biology/transcriptional",
"regulation"
] | 2010 | The Carbon Assimilation Network in Escherichia coli Is Densely Connected and Largely Sign-Determined by Directions of Metabolic Fluxes |
Chromosome segregation in bacteria occurs concomitantly with DNA replication , and the duplicated regions containing the replication origin oriC are generally the first to separate and migrate to their final specific location inside the cell . In numerous bacterial species , a three-component partition machinery called the ParABS system is crucial for chromosome segregation . This is the case in the gammaproteobacterium Pseudomonas aeruginosa , where impairing the ParABS system is very detrimental for growth , as it increases the generation time and leads to the formation of anucleate cells and to oriC mispositioning inside the cell . In this study , we investigate in vivo the ParABS system in P . aeruginosa . Using chromatin immuno-precipitation coupled with high throughput sequencing , we show that ParB binds to four parS site located within 15 kb of oriC in vivo , and that this binding promotes the formation of a high order nucleoprotein complex . We show that one parS site is enough to prevent anucleate cell formation , therefore for correct chromosome segregation . By displacing the parS site from its native position on the chromosome , we demonstrate that parS is the first chromosomal locus to be separated upon DNA replication , which indicates that it is the site of force exertion of the segregation process . We identify a region of approximatively 650 kb surrounding oriC in which the parS site must be positioned for chromosome segregation to proceed correctly , and we called it “competence zone” of the parS site . Mutant strains that have undergone specific genetic rearrangements allow us to propose that the distance between oriC and parS defines this “competence zone” . Implications for the control of chromosome segregation in P . aeruginosa are discussed .
Most bacteria possess a single chromosome , circular and replicated bi-directionally from a specific sequence called oriC . It must be highly compacted to fit inside the bacterial cell , which is typically one thousand times smaller than it is when extended . Despite recent advances , the mechanisms by which the two copies of the chromosome are segregated in daughter cells concomitantly with replication—allowing faithful transmission of genetic material—remains mostly a mystery . Diverse forces might be involved , including DNA replication [1] , intranucleoid pushing forces resulting from radial confinement [2] , or forces originating from polymer properties ( that vary according to different models that consider the chromosome as either a self-avoiding [3] or a self-adherent polymer [4] ) . In bacteria , two main molecular actors that are involved in chromosome segregation have been characterized: the structural maintenance of chromosome ( SMC ) complex and the ParABS system . SMC proteins are ubiquitous from eukaryotes to prokaryotes , and are only absent in a few bacterial species [5] . In contrast , the ParABS system is specific to bacteria , and was found in 70% of the sequenced species in 2007 [6] . It comprises three components . The ParB protein binds specifically to parS sites ( sometimes compared to centromeric sequences ) to form a nucleoprotein complex . ParA is a Walker A-type ATPase thought as the “motor” that provides the force for the segregation of the nucleoprotein complex [7] . Briefly , interaction of ParA-ATP dimer bound aspecifically to DNA with the ParB-parS nucleoprotein complex induces its ATPase activity and its release from DNA , which is thought to pull the ParB/parS complex via a diffusion-Ratchet mechanism [8] . This model was further elaborated by the recent proposition that the elasticity of chromosomal DNA could contribute to the directional transport of the ParB-parS nucleoprotein complex across a ParA-ATP gradient [9] . Moreover , two different models have been proposed regarding the molecular basis involved in the ParB-parS nucleoprotein complex formation , either by ParB from Bacillus subtilis ( the “spreading and bridging” model , [10 , 11] ) or by ParB from the F plasmid ( the “nucleation and caging” model , [12] ) . Both models require the ability of ParB to also bind DNA non-specifically [13] . Although the parS sites exhibit an unusually high degree of sequence conservation and close proximity to oriC in the vast majority of bacteria , the copy number varies among species from 1 to more than 20 [6] . A functional link between the ParABS system and the SMC complex has been demonstrated in B . subtilis and Streptococcus pneumoniae , where the ParB-parS complex recruits SMC to the oriC region , thereby allowing correct chromosome segregation [14–16] . This recruitment is thought to depend on ParB ability to bridge DNA , a phenomenon previously described as “spreading” [10] , but may also require a specific interaction of SMC with ParB [17 , 18] . The gamma-proteobacterium Pseudomonas aeruginosa is an ubiquitous opportunistic pathogen responsible for nosocomial infections and for the morbidity of Cystic Fibrosis patients . The large size of its genome ( 6 . 3 Mb ) results from genetic complexity rather than gene duplication , and allows this bacterium to colonize diverse niches [19 , 20] . It was previously shown that a ParABS system and an SMC complex participate in chromosome segregation in P . aeruginosa [21–26] . Ten parS sites scattered along the chromosome have been proposed , based on sequence homology and in vitro binding experiments [21 , 27]; however , another bioinformatics study predicted only 4 parS close to oriC in P . aeruginosa [6] . We previously showed that P . aeruginosa PAO1 chromosome is globally oriented from the old pole of the cell to the division plane/new pole along the oriC-dif axis , with the oriC region positioned around the 0 . 2/0 . 8 relative cell length in a ParA and ParB dependent manner [26] . To better understand the molecular function of the ParABS system in P . aeruginosa , we used here an in vivo approach to identify and characterize the activity of its different determinants . We show that ParB binds in vivo to 4 parS sites located close to oriC , and that one of these parS is sufficient for proper chromosome segregation . Using cells that carry only one parS site , and displacing this parS site from its native position , we show that parS is bound by ParB independently of its location on the chromosome , and positioned near the 0 . 2/0 . 8 relative cell length in a ParA dependent manner . It is the first chromosomal locus to be separated after replication , indicating that it is the site of force exertion of the segregation process . Moreover , we identify a region of approximately 650 kb surrounding oriC in which the parS site must be located for chromosome segregation to proceed correctly ( as assessed by quantification of anucleate cell in growing cultures ) , suggesting a regional control of chromosome segregation in P . aeruginosa . We further provide evidence that efficient chromosome segregation requires proximity between parS and oriC , suggesting coordination between ParABS mediated chromosome segregation and the replication process .
It was previously reported that the P . aeruginosa ParABS system plays a role in chromosome segregation [26 , 27] . However , there has been controversy about number of parS sites on the genome . To identify ParB binding sites in vivo , we replaced the chromosomal copy of the parB gene by a gene encoding a 3xFLAG tagged version of ParB ( S1A Fig ) . The resulting strain behaves like the wild type strain ( same generation time of 47 minutes and same amount of anucleate cells , less than 1% , S1B Fig ) . We then analyzed the positioning of two chromosomal loci inside growing cells using fluorescent microscopy . In minimal medium supplemented with citrate , chromosomal tags located near oriC and dif present the same localization pattern in the PAO1 ParB-3xFLAG strain and in the wild type PAO1 strain ( S1C Fig ) , indicating that the ParB-3xFLAG is fully functional . We performed Chromatin Immunoprecipitation of ParB-3xFLAG followed by high throughput sequencing ( ChIP-seq , see methods ) . Results presented in Fig 1A and S1 Table show that in vivo , ParB is mostly bound to four of the ten parS sites previously proposed ( parS1 and parS4 ( TGTTCCACGTGGAACC ) , and parS2 and parS3 ( TGTTCCACGTGGAACA ) from [21] , which were also predicted as parS sites by [6] ) . Strong enrichment was found in more than 20 kb surrounding these 4 parS sites , which is consistent with the “spreading” phenomenon described for ParB binding to parS sites in other bacteria [10 , 12 , 13] . This suggests that these four parS site are included in a large nucleoprotein complex formed by ParB . In contrast with previous in vitro results , ParB does not bind significantly in vivo to the other proposed parS sequences , which present two mismatches compared to the parS1/parS2 sequences . We thus propose the TGTTCCACGTGGAACM sequence as P . aeruginosa parS site . This sequence is found in 4 occurrences in P . aeruginosa genome , and no additional occurrence is found when only one mismatch is allowed . It is interesting to note that ChIP-seq analysis revealed eight additional secondary ParB binding sites that do not contain putative parS sequences ( Fig 1 and S1 Table ) . However , the enrichment profile at these sites is different from that obtained with the four parS sites involved in chromosome segregation , as no “spreading” was observed ( S2A Fig ) . As a control , we performed ChIP-seq experiments in a strain lacking the ParB-3xFLAG protein , and no enrichment was observed around the parS sites and the ParB secondary binding sites mentioned above ( S2B , S2C and S2D Fig ) . To study the role of ParB binding sites in chromosome segregation and positioning inside the cell , we inactivated the parS sites bound by ParB . Several point mutations were introduced in parS1 and parS2 and unmarked deletion of parS3 and parS4 were performed ( see S1 Text ) . No obvious growth defect could be observed when parS1 , parS2 and parS3 were inactivated ( alone or in combination ) , in contrast to when the four parS sites were inactivated , which is consistent with results from Jecz and colleagues [27] . Therefore , we focused our study on the mutant presenting only one functional parS site ( ΔparS123 ) and the mutant defective in all four parS sites ( ΔparS1234 ) . We first measured their growth ability in minimal medium supplemented with glucose and casamino acids . The generation time of the ΔparS123 mutant was similar to that of the wild type strain ( 50 and 47 minutes , respectively , Fig 2 and S3 Fig ) , whereas the generation time of the ΔparS1234 mutant was considerably increased ( 81 minutes ) , as observed for the ΔparB mutant [26] . The amount of anucleate cells was also low for the wild type strain and the ΔparS123 mutant ( 0 . 7 and 1% respectively , Fig 2 ) , in contrast to the ΔparS1234 mutant ( approximately 25% ) . In agreement with these results , the positioning of chromosomal tags located near oriC and dif was the same in the wild type strain and in the ΔparS123 mutant ( S4 Fig ) . In contrast , it was strongly affected in the ΔparS1234 mutant ( S4 Fig ) , like in the ΔparB and ΔparA mutants ( previously described in [26] ) . To further confirm that a single parS site bound by ParB is sufficient to sustain proper chromosome segregation in P . aeruginosa , we re-introduced a parS2 site 6 . 5 kb away from oriC ( downstream of gyrB ) in the ΔparS1234 mutant . The resulting strain ( ΔparS1234 parS2 +6 . 5 ) presents a generation time and an amount of anucleate cells only slightly higher than the wild type strain ( Fig 2 ) . We also introduced a sequence that is not efficiently bound by ParB in vivo ( TCTTCCTCGTGGAACA , referred as parS9 in [21] and [27] ) at the exact same location . The resulting strain ( ΔparS1234 parS9 +6 . 5 ) presents a generation time and an amount of anucleate cells similar to the ΔparS1234 mutant ( Fig 2 ) . These results indicate that a single parS site bound by ParB is enough to sustain proper chromosome segregation , and that a parS-derived sequence containing 2 mismatches from the consensus sequence TGTTCCACGTGGAACM is not , probably due to the fact that it is not recognized by ParB in vivo . To characterize ParB binding to P . aeruginosa chromosome in the ΔparS123 and ΔparS1234 mutants , we replaced the chromosomal copy of the parB gene by a gene encoding the 3xFLAG tagged version of ParB and performed ChIP-seq experiments . Results presented in Fig 1B and 1C and S1 Table indicate that more than 8 kb surrounding the parS4 site are enriched in the ΔparS123 mutant , whereas no enrichment could be detected around the localization of the parS sites in the ΔparS1234 mutant . This suggests that a nucleoprotein complex is formed by ParB from a single parS . Moreover , the secondary sites identified in the wild type strain were also bound by ParB in the ΔparS123 and ΔparS1234 mutants . An additional secondary site upstream of dnaA was identified . This site is also present in the wild type strain , but it is included in the large region enriched around the parS sites ( Fig 1A ) . Considering that the chromosome segregation defect of the ΔparS1234 mutant is similar to that of the ΔparB mutant , the fact that ParB is bound to the secondary sites in the ΔparS1234 mutant suggests that they are not sufficient for correct chromosome segregation . Moreover , no binding of ParB to the parS putative sequences identified in vitro in [21] and [27] was observed in the ΔparS123 and ΔparS1234 mutants , indicating that even in the absence of the major ParB binding sites , these sequences are not bound by ParB in vivo . To investigate the impact of the parS site position on chromosome segregation , we reintroduced a parS site at different locations in a ΔparS1234 mutant . We could not obtain and/or propagate strains with a parS site located 1 or 1 . 5 Mb from oriC on the right replichore ( in contrast to what was described with the PAO1161 strain [27] ) . However , we could generate strains with a parS site located between 898 kb from oriC on the left replichore to 550 kb from oriC on the right one ( Fig 2A ) . We analyzed growth rate and chromosome segregation in these strains , by measuring the generation time and the amount of anucleate cells present in liquid culture in minimal medium supplemented with glucose and casamino acids . Results are shown in Fig 2B . Compared to the ΔparS1234 mutant , strains ΔparS1234 parS -898 , ΔparS1234 parS -440 and ΔparS1234 parS -330 ( which carry a parS site 898 , 440 and 330 kb from oriC on the left replichore , respectively , Fig 2A ) present a similar generation time ( more than 80 minutes ) , and a similar amount of anucleate cells ( more than 20% ) . In contrast , generation times of strains ΔparS1234 parS +347 , ΔparS1234 parS +449 , ΔparS1234 parS +545 and ΔparS1234 parS +552 ( which carries a parS site on the right replichore ) were significantly different from the generation time of the ΔparS1234 mutant , closer to the generation time of the ΔparS123 mutant ( 56 , 48 , 58 and 65 minutes respectively ) . In addition , the amount of anucleate cells are significantly reduced in these strains compared to the ΔparS1234 mutant ( 4 , 2 , 11 and 9% respectively , compared to 25% for the ΔparS1234 mutant ) , even if these amounts are significantly higher than what is observed in the ΔparS123 mutant ( less than 1% ) . These results indicate that the efficiency of chromosome segregation depends on the location of the parS site on the chromosome , suggesting a regional control of this process . To investigate the apparent difference in functionality of the parS site depending of its location on the chromosome , we first tested its ability to be bound by ParB . We built a functional GFP-ParB fusion ( see S1 Text ) that we used to visualize ParB in different genetic backgrounds . The gene encoding this fusion is expressed from a plasmid ( called pPSV38-NGFP-ParB ) . It is under the control of an IPTG-inducible lacUV5 promoter flanked by two lac operators , strictly repressed in P . aeruginosa in absence of IPTG ( [28] and S . L . Dove personal communication ) . This pPSV38-NGFP-ParB plasmid was introduced in different strains containing a native copy of parB . We used the minimal amount of IPTG to visualize this fusion without interfering with the cell physiology ( overexpression of parB is toxic in P . aeruginosa , [21] ) . The ability of GFP-ParB to form foci in these strains was assessed using fluorescent microscopy to visualize the GFP fusion . In the ΔparS123 mutant , more than 90% of the cells contained 2 foci ( S2 Table ) , whereas no visible focus could be observed in the ΔparS1234 mutant . This indicates that foci formation depends on ParB binding to a parS site , and probably of the formation of a nucleoprotein complex ( illustrated by spreading of ParB around the parS site , see above ) . Indeed , secondary sites are bound by ParB in the ΔparS1234 mutant ( see above ) , but this binding in the absence of spreading does not result in visible foci . Also , no focus could be observed when the pPSV38-NGFP-ParB plasmid was introduced in the ΔparS1234 parS9 +6 . 5 mutant ( containing only a parS-derived sequence with 2 mismatches from the consensus sequence ) . Strikingly , when the pPSV38-NGFP-ParB plasmid was introduced in strains containing a bona fide parS site at different locations on the chromosome ( ΔparS1234 parS +347 , ΔparS1234 parS +552 and ΔparS1234 parS -898 ) or in the ΔparA mutant , a majority of cells contained foci when observed using fluorescent microscopy ( S2 Table ) . We noticed that the proportion of cells without visible fluorescent focus was higher in the ΔparS1234 parS -898 and in the ΔparA mutant than in the wild type strain , which is in agreement with the higher proportion of anucleate cells observed with these strains . These results indicate that , in contrast with efficient chromosome segregation , ParB binding to parS is not dependent on ParA and on the parS position on the chromosome . The 2 fluorescent foci observed in the ΔparS123 mutant containing the pPSV38-NGFP-ParB plasmid are positioned close to the 0 . 2/0 . 8 relative cell length , with an average interfocal distance of 0 . 6 relative cell length ( Fig 3 ) . Strikingly , when the pPSV38-NGFP-ParB plasmid was introduced in the ΔparS1234 parS +347 , ΔparS1234 parS +552 and ΔparS1234 parS -898 strains ( presenting 4% , 9% and 25% of anucleate cells , respectively ) , ParB positioning was not affected , remaining close to the 0 . 2/0 . 8 relative cell length independently of the parS site position on the chromosome ( Fig 3 ) . In contrast , when the pPSV38-NGFP-ParB plasmid was introduced in a ΔparA mutant , foci were mispositionned and their interfocal distance was strongly reduced ( Fig 3 ) . These results indicate that ParB bound to parS is positioned near the 0 . 2/0 . 8 relative cell length in a ParA dependent manner , but independently of the location of parS on the chromosome . To determine whether the ParB/parS complex positioning at the 0 . 2/0 . 8 relative cell length impacts chromosomal loci adjacent to parS , we analyzed the localization of chromosomal tags in different strains . We used three different tags , located 327 kb or 628 kb from oriC on the right replichore , or 851 kb from oriC on the left replichore ( 327-R , 628-R and 851-L , respectively ) . We compared their positioning inside the cell in the ΔparS1234 mutant with their positioning when a parS site was located close by ( in the ΔparS1234 parS +347 , ΔparS1234 parS +552 and ΔparS1234 parS -898 strains , respectively ) . As a control , we used a chromosomal tag located 82 kb from oriC on the right replichore . As previously described , this tag is positioned at the 0 . 2/0 . 8 relative cell length in the wild type strain ( with an interfocal distance of 0 . 6 relative cell length ) , and this positioning is lost in a ΔparS1234 mutant . Results presented in Fig 4 indicate that all three chromosomal tags are positioned near the 0 . 2/0 . 8 relative cell length when a parS site is located nearby ( at 20 kb , 76 kb and 47 kb from the observed tag , respectively ) , in contrast to what happen in a ΔparS1234 mutant . This is surprisingly also the case for the ΔparS1234 parS -898 strain , which presents 25% of anucleate cells , a level similar to that of the ΔparS1234 mutant . We analyzed the impact of the parS site chromosomal location on chromosomal loci separation after replication . We used strains with two chromosomal tags allowing to visualize two chromosomal loci in the same cells: one close to oriC , and another farther away . We compared the number of foci visible in the wild type strain ( in which the four parS located between 3 and 15 kb of oriC are assimilated to one parS ) with the number of foci visible in a strain containing an ectopic parS . The rational was that to appear as 2 foci , a chromosomal locus must be replicated and separated upon replication . We used three combinations of chromosomal loci to analyze the impact of three parS positions that lead to different amount of anucleate cells: parS +347 ( less than 5% ) ; parS -898 ( around 25% ) ; and parS +552 ( around 10% ) ( see Fig 2 ) . Complete foci repartitions for each strain are shown in S6 Fig . Fig 5 considers only the cells for which 3 foci are visible , i . e . cells in which one chromosomal locus was separated after replication but not the other . In the wild type background , the chromosomal locus close to oriC was separated first in more than 90% of the 3 foci cells , which is expected considering that it is replicated first and also close to the parS sites . Strikingly , in the ectopic parS backgrounds , more than 80% of the 3 foci cells contained 1 focus for the locus proximal to oriC ( replicated first ) and 2 foci for the locus located proximal to the parS site , ( replicated last ) . This indicate that an active process is required to separate replicated chromosomal loci , and that it originates from the parS site , which thus appear to be the site of force exertion of the segregation process . Altogether , these results show that inserting a parS site at an ectopic location induces a repositioning of this parS site and of the adjacent chromosomal loci , and that parS are the first sequences to be separated after replication . This is strikingly the case for all three tested locations of parS , despite the fact that the ΔparS1234 parS +347 , ΔparS1234 parS +552 and ΔparS1234 parS -898 strains produce different amount of anucleate cells , as observed upon nucleoid staining . This suggests that the defect in chromosome segregation observed in the ΔparS1234 parS -898 strain does not originate from the inability of the ParABS system to position chromosomal loci flanking the parS site close to the 0 . 2/0 . 8 relative cell length . To better characterize the position effect of the parS site , independently of putative local effects due to parS insertion , we used a high-throughput approach . We inserted a parS site in a mariner transposon ( modified from the pSC189 vector , [29] ) , and used it to generate a randomly inserted transposon library in the ΔparS1234 mutant . A parS site inserted at a position for which chromosome segregation occurs correctly ( few anucleate cells ) is expected to have a fitness advantage . Therefore , although the transposon library would have parS inserted at random locations over the genome , the ones having a parS in position allowing correct chromosome segregation would be enriched during the propagation . As a control , an insertion library build with a transposon without parS was carried out in parallel and insertions sites of these 2 libraries were then identified according to the protocol described in [30] ( see Methods for details ) . Ratios between the number of insertions using the two transposons are presented in Fig 6 . A clear enrichment of parS insertions is observed in the region between approximately -200 kb to +450 kb from oriC . It is due to a very strong bias of insertions obtained with the mariner transposon containing a parS site , but not the control transposon . This indicates that insertion of a parS site in this region provides a growth advantage compared to the ΔparS1234 mutant , whereas an insertion of a parS site outside of this region does not . We will refer to this region as the «competence zone» of the parS site . In agreement with our previous results , parS sites inserted on the left of oriC in the ΔparS1234 ( respectively -898 , -440 and -330 ) are located outside of the «competence zone» , whereas the parS sites inserted on the right of oriC are inside ( +347 and +449 ) or at the extreme limit of it ( +545 and +552 ) . Remarkably , we noticed that the «competence zone» is not centered on the parS native position , in the vicinity of oriC , and that the left border might coincide with a ribosomal operon ( located at -220kb on the left of oriC ) . We deleted this ribosomal operon in the ΔparS1234 parS -330 and ΔparS1234 parS -898 strains ( generating strains ΔparS1234 parS -330 ΔrrnD and ΔparS1234 parS -898ΔrrnD ) , and analyzed the efficiency of chromosome segregation in the resulting strains by nucleoid staining . The proportion of anucleate cells in the ΔparS1234 parS -330 ΔrrnD strain was strongly reduced compared to the ΔparS1234 parS -330 strain ( 2% compared to 20% ) , indicating that the rrnD operon , which is located between parS and oriC , impairs chromosome segregation ( Fig 7 ) . We note however that the proportion of anucleate cells in the ΔparS1234 parS -898 ΔrrnD strain was close to 20% , indicating that deleting the rrnD operon does not restore efficient chromosome segregation in a strain containing a parS site located 898 kb on the left of oriC . Altogether the observed asymmetry of the competence zone is likely due to the presence of the rrnD operon between oriC and parS . To identify genetic determinants of the parS «competence zone» , we used the λ derived site-specific recombination system [31] to invert chromosomal fragments and bring specific sequences closer to parS sites located outside of the «competence zone» . The impact of these programmed chromosome rearrangements on anucleate cell formation was then analyzed by nucleoid staining . More specifically , we inserted a parS site next to an attL site at position 851-L ( 851 kb on the left of oriC , see methods ) , and attR sites at three positions: 92-L ( -92 kb from oriC ) , 82-R ( +82 kb from oriC ) or 327-R ( +327 kb from oriC ) . Recombination between attL and attR sites upon Int and Xis action led to the inversion of the chromosomal region between these sites , whereas the position of the parS site remains unchanged . The amount of anucleate cells observed in the inverted strains was compared to the one in the parental strains ( Fig 8A ) . In the case of attR 92-L , no difference in anucleate cells could be detected whereas with attR 82-R and 327-R , a significant decrease was observed in the inverted strains compared to the non-inverted strains . Therefore , when a parS is located at position 851-L , efficient chromosome segregation can occur when oriC is moved to a closer position ( only 83 kb or 328 kb away from parS , respectively ) . It is noteworthy to mention that in these strains , rrnD is not located between parS and oriC ( Fig 8A ) . The same approach was used to analyze the right side of the «competence zone»: an attL site next to a parS site was inserted at position 628-R ( 628 kb on the right of oriC ) , and attR sites were inserted at positions 327-R , 82-R or 92-L respectively . Remarkably , a significant decrease in anucleate cells production was only observed in the inverted strain containing the attR 92-L , i . e . when oriC was translocated 93 kb away from parS ( Fig 8B ) . Altogether , these results demonstrate that genetic determinant ( s ) of the parS «competence zone» is located between -92 kb and +82 kb from oriC . This region encompasses a number of essential genes , but its most striking feature is that it contains oriC itself . Therefore , we propose that the «competence zone» is defined by the distance between oriC and parS , and that proximity between these two sequences is critical for efficient chromosome segregation .
Using a ChIP-seq approach , we identified 4 strong ParB binding sites in vivo , which allow us to define the TGTTCCACGTGGAACM sequence as the P . aeruginosa parS site . Although in vitro 2 mismatches did not affect ParB binding to this sequence drastically , [21 , 27] , no significant enrichment of such degenerated sequences was detected in vivo , either in the wild type strain or in the strain deprived of the 4 bona fide parS sites . We also demonstrate that re-positioning a 2 mismatches sequence closer to oriC does not allow ParB binding . Interestingly , we identified nine secondary binding sites for ParB . These sites are less efficiently bound than the parS sites , no major “spreading” phenomenon could be observed , and these sites are not involved in chromosome segregation . They might be involved in transcriptional regulation by ParB , and this is of particular interest in the case of the secondary site identified in the promoter region of dnaA , a gene encoding the replication initiation protein in bacteria . A link between the ParABS system and the regulation of replication initiation has been characterized in B . subtilis and V . cholerae chromosome I . However , it most probably occurs post-transcriptionally , through the interaction of ParA with DnaA , and not at the transcriptional level [32] . The mechanism of ParB binding to these secondary sites and their biological significance remain to be characterized . It is noteworthy to mention that the enrichment profiles surrounding the parS sites encompasses more than twenty kb and is compatible with the presence of ParB molecules involved either in a “spreading and bridging” mechanism [11] , or a “nucleation and caging” mechanism [12] of binding . This is consistent with the fact that it was previously shown that P . aeruginosa ParB was able to bridge DNA in vitro [10] . We show that in our growth conditions , one parS site is sufficient to promote proper chromosome segregation , as was described in the study from Jecz and colleagues [27] . However , the amount of anucleate cells in a strain deprived of parS sites differs between the 2 studies ( more than 20% in our case , in contrast to 2–3% in the work of Jecz and colleagues [27] ) . We use a PAO1 isolate that presents differences from the sequenced isolate [33] , whereas they use the PAO1161 strain , which is a derivative of the sequenced PAO1 [19] , which might account for this difference . In this study , we also demonstrate that the location of the parS site on the chromosome is critical for chromosome segregation ( as assessed by measuring growth rate and anucleate cells in liquid cultures ) , and identify what we called a «competence zone» for parS , which ranges roughly from 200 kb on the left of oriC ( where it is limited by rrnD ) to 450 kb on its right . Strikingly , we were unable to construct and propagate strains containing a parS site located at either 1 . 5 Mb or 1 Mb on the right of oriC , suggesting that the introduction of a parS site far from oriC can be more detrimental for growth than the absence of parS site . We also determined that ParB binds to parS independently of ParA and of the parS location on the chromosome , and that ParB bound to parS is positioned to the 0 . 2/0 . 8 relative cell length . This localization depends on ParA , but not on the parS location on the chromosome . These results suggest a repositioning of the chromosomal region containing the parS site due to the ParB and ParA proteins , which could be linked to an “anchorage process” through ParA . Indeed , a repositioning of chromosomal loci containing ectopic parS has been previously described in Caulobacter crescentus [34] and V . cholerae chromosome I [35] , two bacteria that possess specific proteins allowing the anchorage of the parS sites to the cell pole [36–38] . Interestingly , no anchorage mechanism to cellular positions other than cell poles has been described so far . In contrast , introduction of parS arrays at different locations on the B . subtilis chromosome does not induce a repositioning of the chromosomal loci in which they are inserted [39] , and no anchorage process was described in this bacterium . Subpolar positioning of parS sites have also been described in Myxococcus xanthus , however the impact of their displacement on chromosomal loci positioning has not been assessed yet [40] . By engineering chromosome rearrangements using a lambda-based recombination system , we demonstrate that the «competence zone» is linked to the distance between parS and oriC . Strikingly , this is in perfect agreement with the observation that parS sites are found near oriC in the vast majority of bacterial species [6] , although a functional link was not demonstrated previously . In contrast , it was reported that parS sites could be moved 650 kb away from oriC in the 3 Mb long V . cholerae chromosome I without impacting the efficiency of chromosome segregation [35] . Similarly , no defect was described when parS sites where displaced 400 kb away of oriC in the 4 Mb long C . crescentus chromosome [41] . To our knowledge , the only previous hint of functional coupling between parS site and oriC has been described in B . subtilis . In this bacterium , the chromosome segregation defect observed when parS sites were inserted near the terminus of replication was mainly due to the recruitment of the SMC condensin away from oriC . SMC is the prominent factor involved in B . subtilis chromosome segregation [14] . In P . aeruginosa , a Δsmc mutant presents only a slight defect in chromosome segregation [25] . However , we were not able to delete smc in the ΔparS1234 parS -898 strain , and deletion of smc in the ΔparS1234 parS -330 ΔrrnD strain leads to an increase in anucleate cells ( from approximately 1% to approximately 7% ) , which indicates that unlike the situation described in B . subtilis , the segregation defect observed when the parS site is distant from oriC is not linked to SMC function . In this study , we establish that the parS site is the site of force exertion of the segregation process , ( as described in C . crescentus [41] ) ; indeed , chromosomal loci close to the parS site are the first to be separated after replication . Moreover , they reach the 0 . 2/0 . 8 relative cell length upon segregation , independently of parS location on the chromosome . Strikingly , this is also the case in the ΔparS1234 parS -898 strain for which more than 20% of anucleate cells are observed . This suggests that chromosome segregation in this strain does not originate from a defect of positioning of the chromosomal loci surrounding the parS site , or from a defect in the “segregation force” itself . Our demonstration that bringing oriC closer to the parS site restores proper chromosome segregation indicate that the timing of separation of the chromosomal loci surrounding oriC after replication initiation might be critical , and that it is normally determined by the distance between oriC and parS . When this distance is too large , or when a rrn operon is located between parS and oriC , the “segregation force” is still applied but the timing of separation of oriC is lost , and this is detrimental for the segregation process . Interestingly , two studies in B . subtilis also indicate that a failure in separating newly replicated origins is detrimental for the segregation process [42 , 43] . However , the specific problems arising from the delayed separation of the replicated oriC remains to be characterized . It was previously shown that rrn operon can interfere with chromosome organization and form conformational barriers [17 , 44] . It was proposed that these barriers serve as flexible tethers creating a spatial gap between domains , which might be consistent with our results: when such a barrier is inserted between parS and oriC in P . aeruginosa , the timing of separation of replicated oriC is lost , even if the parS site is not too far from oriC . An alternative explanation might be that rrn operons are transcribed altogether , in a nucleolus-like structure , which would interfere with the segregation process originating from parS , and impair once again the timing of oriC separation . However , even if clusters of RNA polymerase have been observed [45] , the existence of nucleolus-like structure in bacteria remains to be proven . The presence of a ribosomal operon between oriC and the parS site is not always detrimental to chromosome segregation . Indeed , this is the case for Vibrio cholera chromosome I ( rDNA found at + 53 kb from oriC when the 3 parS sites are at + 63 , + 66 and + 69 kb from oriC ) , and several ribosomal operons alternate with parS sites on the oriC proximal part of the right chromosomal arm in B . subtilis . In both cases however , the impact of the ParABS system impairment on chromosome segregation is marginal . Overall , this study demonstrates a functional link between oriC and parS in P . aeruginosa . It suggests that the timing of separation of the chromosomal loci surrounding oriC after replication is critical , and that it could be an important role of the ParABS system to keep this timing right , which would explain the proximity of parS and oriC in most bacterial species .
P . aeruginosa strain PAO1 was initially provided by Arne Rietsch ( Case Western Reserve University ) . This PAO1 isolate does not present the inversion described for the sequenced PAO1-UW subclone resulting from homologous recombination between the rrnA and rrnB loci , which are orientated in opposite directions and separated by 2 . 2 Mbp [19] . It also contains the 12 kb insertion and 1006 bp deletion described in [33] . Escherichia coli DH5α ( Invitrogen ) and DH5α λpir was used as the recipient strain for all plasmid constructions , whereas E . coli strains β2163 [46] and MPFpir [47] were used to mate plasmids into P . aeruginosa . Details of plasmid and strain constructions are provided in S1 Text and S3 Table . For growth rate and anucleate cells analysis , overnight cultures grown in Lysogeny broth ( LB ) at 37°C were diluted 300 times in Minimal Medium A ( Miller 1992 ) supplemented with 0 . 12% casamino acids and 0 . 5% glucose , and strains were grown at 30°C until they reach an OD of approximately 0 . 15 . For fluorescent microscopy analysis of chromosomal loci and for NGFP-ParB localization , Minimal Medium A supplemented with 0 . 25% citrate was used . IPTG was added to growth medium at 0 . 5 mM for observation of chromosomal tags and 0 . 1 mM for observation of NGFP-ParB until they reach an OD of approximately 0 . 1 . Cells were grown until OD600 0 . 1 , and fixed with an equal volume of a 1×PBS solution containing 5% paraformaldehyde and 0 . 06% glutaraldehyde . After overnight incubation at 4°C , the cells were washed twice in PBS and then incubated in a solution of 1 μg ml−1 HOESCHT 33258 ( Thermofisher ) . After 20 min incubation , the cells were washed in 1×PBS , spread out on agarose pads and observed immediately using a Leica DM6000 microscope , a coolsnap HQ CCD camera ( Roper ) and Metamorph software . Chromosomal loci and NGFP-ParB were observed when cultures reached an OD600 between 0 . 05 and 0 . 1 . Cells were then spread out on agarose pads and observed immediately using a Leica DM6000 microscope , a coolsnap HQ CCD camera ( Roper ) and Metamorph software . Image analysis was performed using the MATLAB-based software MicrobeTracker Suite [48] . Briefly , The MicrobeTracker program was used to identify cell outlines , and SpotFinderZ to detect fluorescent spots inside the cells . Spots and cell outlines were then manually validated using homemade matlab functions [35] . Spot numbers and positions were then analyzed according to cell length . Formaldehyde was added to 150 ml of culture ( Minimal Medium A supplemented with 0 . 25% citrate , OD600 approximately 0 . 1 ) to a final concentration of 1% and samples were incubated at room temperature for 30 min . To quench cross-linking reaction , glycine was added to a final concentration of 125 mM and followed by a 10 min incubation at room temperature . Cell pellets were washed three times with 10ml of PBS and then resuspended in 1 ml of lysis buffer ( 50 mM Tris-HCl pH 7 . 4; 150 mM NaCl; 1 mM EDTA; Triton X-100 1%; and Roche Protease Inhibitor Cocktail ) . Chromosomal DNA was sheared by sonication to an average size of 0 . 5–1 kb . After the removal of cell debris by centrifugation , 50 μl of each sample was removed to serve as an input control . The remaining samples were added to 50 μl of ANTI-FLAG M2 affinity resin ( Sigma A2220 ) previously washed twice with TBS and twice with lysis buffer . After incubation at 4°C overnight , beads were pelleted and washed twice with TBS Tween 0 . 5% and three times with TBS . Elution was performed using the 3X FLAG peptide ( Sigma F4799 ) , as recommended . The recovered supernatants were placed at 65°C overnight to reverse the cross-links . The input samples were also incubated at 65°C overnight after the addition of 200 μl of TBS . Library preparation and sequencing were performed by the IMAGIF facility ( I2BC , Gif sur Yvette ) . Sequences were aligned against the reconstituted genome of our PAO1 strain ( available on request ) . The sequencing results were analyzed as described in [49] . Briefly , the number of reads for the input and IP data was smoothed over a 200 bp window ( the estimated size of the fractionated DNA ) , normalized to the total number of reads , and enrichments fold were calculated for each base as the ratio of the number of reads in the IP fraction and the number of reads in the input fraction . Peak calling was done using Matlab functions . Regions presenting a tenfold enrichment were selected , except in the Δpar123 strain , for which an enrichment of 15 was preferred . The rational was to select for the approximatively 10 most abundant regions . Considering that only the 4 parS sites are involved in chromosome segregation , further investigation of smaller peaks was not undertaken . Results are presented in S1 Table . Sequencing data are available from the GEO database via accession number GSE87409 ( http://www . ncbi . nlm . nih . gov/geo/ ) . The pSC189 vector , described in [29] , was modified for use in P . aeruginosa . The aacC1 gene from pEXG2 [50] was PCR amplified and cloned on a XhoI/NcoI fragment into the pSC189 vector , giving rise to pSC189Gm . Then , 129 pb containing the parS3 site were PCR amplified from PAO1 chromosome and cloned on a XhoI fragment into the pSC189Gm digested with SalI , giving rise to pSC189Gm-parS . pSC189Gm and pSC189Gm-parS were then tranformed into strain MFPpir for further mating into the ΔparS1234 mutant . Mutant libraries were generated according to [51] . Briefly , donor strains ( MFPpir containing whether pSC189Gm or pSC189Gm-parS ) and recipient strain ( Δpar1234 ) were scraped from overnight plates grown at 37°C and 42°C , respectively . Optical densities were adjusted to 40 for the donor strain and 20 for the recipient strain . Equivalent volumes were mixed , 50 microliter spotted on dried LB plates supplemented with 0 . 3 mM of DAP ( Sigma D1377 ) , and incubated overnight for transposition and selection . Mating mixtures were scraped and plated on Pseudomonas Isolation Agar ( Sigma 17208 ) containing 60 μg mL-1 of Gentamicin ( Sigma G1397 ) . Sixty matings were done , giving rise to approximately 80 , 000 colonies per mutant library . Colonies were scraped from plates , washed once in LB and froze in 10% DMSO at -80°C in 5 aliquots . Genomic DNA was prepared from one aliquot using the GenElute™ Bacterial Genomic DNA Kit from Sigma , and sequencing libraries were prepared as described in [30] . Sequencing was performed at the IMAGIF facility ( I2BC , Gif sur Yvette ) . Sequences were aligned against the reconstituted genome of our PAO1 strain to determine the insertion locations of transposons . ( Raw data are available from the SRA database ( https://trace . ncbi . nlm . nih . gov/Traces/sra/ ) under accession number SRP090425 ) . Ratio between numbers of insertion of the pSC189Gm-parS and pSC189Gm were calculated , log2 calculated to improve the representation of the data , and results were binned over 10 kb . | Accurate transmission of the genetic information relies on replication and segregation , two processes essential to all living organisms . In bacteria , these processes occur concomitantly . Replication of the bacterial circular chromosome initiates at a single specific sequence called oriC , and proceed bi-directionally along the chromosome arms . A partition system called ParABS is involved in chromosome segregation in many bacteria . It involves the binding of the ParB protein to parS sequences , which are often found in the close vicinity of oriC . The importance of this system for chromosome segregation varies according to species , ranging from essential to dispensable . In Pseudomonas aeruginosa , an important opportunistic pathogen , the ParABS system plays an important role in chromosome segregation , as mutants affected in this system present a severe growth defect as well as anucleate cells formation , but is not essential . In this study , we characterize the activity of the different determinants of the ParABS system in P . aeruginosa and demonstrate that it is critical for the parS site to be located close to oriC , which suggest that the timing of separation of regions close to oriC after replication is important , and that it could be a function of the ParABS system to keep this timing . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"bacteriology",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"microbiology",
"operons",
"light",
"microscopy",
"pseudomonas",
"aeruginosa",
"nucleoproteins",
"microscopy",
"genetic",
"elements",
"bacterial",
"genetics",... | 2016 | Regional Control of Chromosome Segregation in Pseudomonas aeruginosa |
MicroRNAs ( miRNAs ) are small , non-coding RNAs that play essential roles in plant growth , development , and stress response . We conducted a genome-wide survey of maize miRNA genes , characterizing their structure , expression , and evolution . Computational approaches based on homology and secondary structure modeling identified 150 high-confidence genes within 26 miRNA families . For 25 families , expression was verified by deep-sequencing of small RNA libraries that were prepared from an assortment of maize tissues . PCR–RACE amplification of 68 miRNA transcript precursors , representing 18 families conserved across several plant species , showed that splice variation and the use of alternative transcriptional start and stop sites is common within this class of genes . Comparison of sequence variation data from diverse maize inbred lines versus teosinte accessions suggest that the mature miRNAs are under strong purifying selection while the flanking sequences evolve equivalently to other genes . Since maize is derived from an ancient tetraploid , the effect of whole-genome duplication on miRNA evolution was examined . We found that , like protein-coding genes , duplicated miRNA genes underwent extensive gene-loss , with ∼35% of ancestral sites retained as duplicate homoeologous miRNA genes . This number is higher than that observed with protein-coding genes . A search for putative miRNA targets indicated bias towards genes in regulatory and metabolic pathways . As maize is one of the principal models for plant growth and development , this study will serve as a foundation for future research into the functional roles of miRNA genes .
The last decade has witnessed remarkable progress in our knowledge of the biogenesis and activity of diverse classes of small non-coding RNAs ( sRNA ) . These include microRNAs ( miRNA ) [1] , small interfering RNAs ( siRNA ) [2] , trans-acting siRNAs ( ta-siRNA ) [3] , and others [4] . While the majority of plant sRNAs are transcribed from siRNA genes residing in repetitive and transposon-rich regions , and regulate chromatin silencing [4] , a great deal of interest has been placed on miRNAs due to their ability to post-transcriptionally regulate gene expression [5] . This is exemplified by the critical regulatory behavior of miRNAs at key positions in a variety of pathways , such as root [6] , [7] , shoot [8] , leaf [9]–[11] and flower [12] , [13] development and cell fate [14] , [15] . Additionally , they also include responses to phytohormones [16] , nutrient [17]–[19] and other environmental stresses [20]–[23] . As a genetic model system , maize has contributed significantly to our understanding of plant development and evolution , and more recently this knowledge has been employed to elucidate the regulatory functions of miRNA genes . For instance , teosinte glume architecture 1 ( tga1 ) is one of the major genes responsible for the evolution of maize from its ancestor teosinte and has also been identified as a target for miR156 [24] , [25] . Intriguingly , mutations in Corngrass1 ( Cg1 ) result in the over-expression of miR156 and decreased miR172 levels , resulting in alterations of the juvenile to adult phase transition [24] , [25] . Taken together , it is apparent that miRNA regulation is intertwined with key plant development processes . Thus , there is considerable interest in taking advantage of the complete genome sequence of maize B73 reference genome version 1 ( B73 RefGen_v1 ) [26] to systematically identify miRNA genes , their corresponding targets , and to decipher their regulatory roles . The mature biologically active products of miRNA genes define miRNA gene families . This , along with the characteristic ‘hairpin’ structure of its precursor ( pre-miRNA ) , allows computational detection and annotation of miRNA genes [1] , [4] . Post-transcriptional regulation is accomplished by the RNA-induced silencing complex ( RISC complex ) which directs complementary binding of the mature miRNA product to mRNA transcripts , usually resulting in target cleavage or inhibition of translation [27] . In plants , the near perfect complementarity between the miRNA and its substrate mRNA has permitted computational methods for the specific identification of target genes [5] . In this study , we systematically annotated miRNA genes based on the first nearly complete assembly of the maize B73 RefGen_v1 [26] . This analysis includes experimental validations for many of these genes , as well as expression studies using deep small RNA transcriptome sequencing across a broad panel of maize tissues . We also conducted one of the first comprehensive characterizations of maize pri-miRNA transcripts , providing a deeper understanding of their transcription and regulation . Prediction of protein-coding targets confirmed many known regulatory substrates and provided a window into many more potentially novel pathways . We also describe the first analysis of allelic diversity of miRNAs in maize using a large panel of highly diverse inbred lines as well as its wild relative teosinte . A comparative genomic analysis with sorghum provided insight into the evolutionary dynamics of miRNA family expansions and will serve as basis for future comparative functional genomic analyses using syntenic orthologs . We believe the multi-faceted nature of this study will help accelerate our understanding of miRNAs , their regulatory roles in critical biological processes , as well as offer the community detailed annotations applicable to their own research .
Starting with a set of 1822 plant mature miRNA sequences available in miRBase [28] , we predicted putative pre-miRNA structures at more than 4 , 000 loci in the maize B73 RefGen_v1 [26] . Of these , more than 300 passed initial filters for positional overlaps , secondary structure and orientation of mature miRNA sequences within the respective stem-loop structures . After additional screening for overlaps with transposable elements ( Table S1 ) and removal of families not well characterized in current literature , we annotated 150 miRNAs from 26 families with high confidence . Since our search space for pre-miRNA structure included only 250nt surrounding the aligned mature miRNA , our method may have missed families that have introns in their precursors . One example is miR444 , which has been reported as conserved in rice , maize , sorghum , and sugarcane [29]–[31] , but was missed by our pipeline due to a large intron in its pre-miRNA structure . There are a total of 98 maize miRNA genes deposited in miRBase [32]–[34] , which were predicted based on the MAGI ( Maize Assembled Genomic Islands: maizegdb . org ) and TIGR EST collections . Of these , 89 were detected in our set and are now annotated on the B73 reference genome assembly . We failed to detect nine miRNA genes in our pipeline: miR160e , miR166e , miR166f , miR169e , miR169g , miR169h , miR172a , miR172e and miR408 . It is possible that these are located in regions that are not captured in the current genome assembly . In addition to those previously characterized , we found 61 putative miRNA genes representing new members of conserved families . A summary of known and previously identified miRNA gene families is provided in Table 1 . Detailed information about the genomic location of each miRNA , as well as the mature miRNA and miRNA* for each miRNA gene , is listed in Table S2 . Annotated miRNA genes were found on all maize chromosomes , having a distribution pattern similar to those of protein-coding genes ( Table S2 ) [26] . No miRNA genes were found in unanchored portions of the assembly . Like protein coding genes , a proportion of the miRNA genes are organized as tandem paralog clusters ( see section “Genome Organization and Conservation with Sorghum” ) . However , a subset of these was found as unusually compact clusters , with less than 2000nt separating adjacent genes , as shown in Table 2 . Two members of the miRNA156 family ( miR156b and miR156c ) on chromosome 3 are separated by less than 200nt , as are two members of the miRNA166 family on chromosome 5 . The close distance of miR156b and miR156c is observed in several monocots and they are transcribed as one transcript ( polycistronic ) in maize [24] and rice [35] . Clusters of two members of miR166 are also observed in several plant species , including moss [33]–[37] . Compact clustering was particularly prevalent in the miR395 family . Four such clusters were found on chromosomes 2 and 10 , comprising 16 genes in total . In rice , compact clusters of miR395 genes have also been observed with each forming a single polycistronic transcription unit [38] . Primary transcripts of miRNA genes are known to have features typical of transcription by RNA polymerase II , including 5′ capping , 3′ polyadenylation , and intron splicing [39] . To characterize the pri-miRNA transcripts in maize , we designed gene-specific primers for each of the 89 previously identified miRNA genes and conducted 5′ RACE and 3′ RACE using a template containing mixed tissues of seedling , immature tassel and immature ear . Overall , we were able to capture the upstream transcribed regions ( 5′ region ) of 55 miRNA genes and the downstream transcribed regions ( 3′ region ) of 51 miRNA genes ( Table S2 ) . Among these , we obtained both 5′ and 3′ RACE products for 40 miRNA genes , producing full-length transcript sequences . Failure to amplify some miRNA gene products might have been a limitation of the tissues sampled , as some miRNA genes might be expressed in highly specific tissue/cell types , developmental stages , or environmental conditions . From our RACE data , we were able to confirm the clusters of miR156b/c and miR166k/m , but failed to detect miR395a/b clusters . Genes of miR395 family are known to be up-regulated in response to low-sulfate conditions [40] and it is possible that transcript abundance was below the detection threshold given the normal nutrient status in which our plants were grown . While our work was underway , the data for maize full-length complementary DNA ( FLcDNA ) became available [41] , [42] . To identify pri-miRNAs in this FLcDNA set , we mapped 150 miRNA precursors against 63 , 000 FLcDNAs and found 33 transcripts that harbor 27 pre-miRNA sequences . After excluding those pri-miRNAs that were mapped with RACE , we identified 10 additional pri-miRNAs using FLcDNA analysis ( Table S3 ) . The overall lengths for pri-miRNA transcripts ranged from 250nt to nearly 2000nt , with an average size of 810nt , far less than the 1 , 433nt average for maize protein-coding transcripts [26] . Figure 1 shows that the 3′ region of pri-miRNAs ( measured from the stem-loop to the transcriptional stop site ) is generally longer than the 5′ region ( measured from the transcriptional start site ( TSS ) to the stem-loop ) , having mean lengths of 953nt and 523nt respectively . We observed that 83% of maize pri-miRNAs have TATA-box like motifs in the −23 to −28 positions relative to TSS ( data not shown ) . Previous studies using rice and Arabidopsis FLcDNAs and Arabidopsis pri-miRNAs showed a high prevalence of an adenine at the TSS and a cytosine at the −1 position [43] . Of the available 76 miRNA gene transcripts with TSS sites , most have A at the TSS and C at the −1 position ( Figure 2 ) . We also observed evidence for alternative TSS . Among the 55 miRNA transcripts with 5′ region information , 10 had two TSSs that are 3nt to 9nt apart . This is similar to FLcDNAs studies of Arabidopsis and rice showing two TSSs per locus , with an average distance of 4 . 2nt and 8 . 7nt apart , respectively [44]; however , the distance between various stop sites can be up to several hundred nucleotides . Examples of polycistronic transcripts and multiple TSS and stop sites are shown in Figure 3 . In addition we found that introns were more abundant in the 3′ region than in the 5′ region . From RACE reaction products , we found only one gene ( miR164c ) with an intron in the 5′ region , whereas 15 genes showed one or two introns in the 3′ region ( Table S4 ) . Intron length varied from 71nt to 2196nt , and the canonical splice motif GU…AG was found for all but one intron ( Table S4 ) . Alternative non-spliced transcripts were found for three genes: miR159a , miR166g , and miR169i . Using Illumina's sequencing-by-synthesis technology [45] , we profiled the genome-wide transcript profiles of miRNA abundance across five different maize tissue types ( root , seedling , tassel , ear , and pollen ) . Small RNA libraries were generated for each tissue and sequenced using the Illumina 1G Sequencer . The small RNAs ( 18–22nt ) were mapped to the predicted pre-miRNAs . The number of reads mapping to each pre-miRNA were enumerated and normalized against the total count of 18 to 22 nucleotide reads , reported as reads per million ( RPM ) , for each respective library . A summary of the quantitative expression profiles for each of the 26 miRNA families is shown in Table 3 . Of the five tissues surveyed , the seedling samples showed the highest expression levels while the pollen samples showed the lowest expression levels , with RPM counts an order of magnitude larger in the seedling . We noted that miR482 lacks an expression signature in any of the five tissues , while miR162 miR394 , miR395 , miR398 , miR399 , miR408 , miR528 and miR1432 had low expression counts ( less than 200 RPM ) . In contrast , miR156 , miR159 , miR167 , miR168 , miR169 , miR171 , miR319 , and miR529 had high expression counts ( slightly over 3 , 000 RPM , on average ) . In our samples , miR529 was over-expressed in the tassel ( almost 16 , 000 RPM ) but under-expressed in the other four tissues ( slightly over 300RPM , on average ) . Several families showed higher expression levels in juvenile tissues . The miR156 , miR164 , miR168 , miR393 , miR395 , miR396 , miR398 , and miR399 families had higher signatures in juvenile root and seedling tissues while miR172 demonstrated a higher expression level in reproductive tissues ( tassel and ear ) . Within miRNA families , there are examples of tissue-specific expression differences among individual miRNAs . For example , the miR396 family is overall highly expressed in the juvenile root and seedling samples but only certain individual family members ( miR396a , miR396b , and miR396g ) are highly expressed in pollen ( Figure 4 ) . We computationally predicted potential maize targets using the Filtered Gene Set transcripts ( Release 4a . 53 ) of the B73 maize genome sequence [26] . The filtered set of transcripts contains about 32 , 500 entries , and is comprised of the longest representative cDNA transcripts . These are presumed free of pseudogenes and low-complexity repetitive elements such as short-tandem repeats and transposons . Our computational pipeline predicted 247 unique putative genes targeted by 150 miRNA sequences belonging to 26 miRNA families . Around 85% of the predicted targets have functional InterPro annotations [46] . We observed that 12 out of 26 miRNA families are predicted to target transcription factors , as shown in Table S6 , suggesting roles of these miRNA families in post-transcriptional regulation and transcription networks . In addition to transcription factors , other predicted target genes are involved in diverse physiological and metabolic processes . Such targets include protein kinases , signal transduction histidine kinases , antifreeze proteins , F-box proteins , cytochrome P450 , cupredoxin , peroxidases , multicopper oxidases , transporters , ATP sulfurylases , and cell division proteins ( Table S6 ) . To gain a better understanding of the functional roles of the predicted miRNA target genes in maize , we looked for target enrichment in Gene Ontology ( GO ) molecular function and biological process categories [47] . The targets were annotated by using the GO annotations available from the B73 RefGen_v1 . Of the predicted targets , 76% had GO assignments whereas only 53% of the genes in the entire refined set were associated with GO terms . BiNGO ( Biological Networks Gene Ontology ) [48] was used to study targets enrichment and to construct a hierarchical ontology tree in Cytoscape [49] , as shown in Figure 5 . We found that miRNA families preferentially target genes involved in a wide spectrum of regulatory functions and selected biological processes including gene expression/transcription , metabolism , catalysis , transport , and response to stimuli ( Table S7 ) . The genes targeted by miRNA families showed a strong affinity for binding activity ( 82 . 3% ) , transcription factor activity ( 10 . 6% ) and transcription regulator activity ( 11 . 2% ) . In the GO biological process enrichment analysis , targeted genes were found to be involved in biological regulation of cellular processes ( 35 . 2% ) , biosynthetic processes ( 32% ) , and metabolic processes ( 32% ) . Approximately 32% of the target genes were involved in transcription/gene expression while about 19% of targets were classified in the response to stimulus category . The latter were found to be involved in the response to abiotic stimulus , endogenous stimulus , and hormone stimulus ( Table S7 ) . Since one gene can be related to multiple GO terms , the sum of the percentages of GO terms is not a relevant parameter . We also computationally evaluated whether specific miRNA families were preferentially enriched in certain GO categories . Target genes of the following 8 miRNA families were overrepresented in both of the biological processes i . e . the regulation of transcription and the response to stimulus . These families are: miR156 , miR160 , miR164 , miR166 , miR167 , miR172 , miR396 , and miR528 . In addition to these families , target genes of miR169 , miR319 , miR408 and miR529 were involved in transcription regulation , whereas miR159 , miR397 , and miR399 target genes were involved in response to stimulus . Targets of only miR395 family showed involvement in sulfate assimilation pathway whereas targets of both miR395 and miR399 families showed enrichment in transmembrane transport activities . Target genes of miRNA families , miR164 , miR397 , miR408 , and miR528 showed enrichment in laccase and oxidoreductase activities and were found to be involved in secondary metabolic processes such as phenylpropanoid , amino acids , aromatic compounds and lignin catabolic processes ( Table S7 ) . Previous studies in Arabidopsis [37] , [50] and rice [38] have shown that miRNA gene families evolved from a combination of tandem , segmental , and whole-genome duplication events [51] . Maize was derived from an ancient allotetraploid , and it is estimated that the two maize progenitor genomes diverged from the ancestor of sorghum ∼12 MYA [52] . Comparative mapping between maize and sorghum can therefore reveal the fate of miRNA genes after whole genome duplication . In addition , synteny can be used to help infer orthologous relationships between these species . Based on the maize high confidence set , we filtered previously annotated Sorghum bicolor miRNA genes [29] . The distribution of these genes by family is shown in Table S8 , along with corresponding information for maize . Synteny was examined in the context of orthologous protein coding genes which numbered 25 , 216 in maize and 20 , 408 in sorghum [26] ( See Materials and Methods ) . In total , we found 136 maize and 106 sorghum miRNA genes within syntenic regions , corresponding to 91% and 79% of their respective totals . These values are similar to the percentages of syntenic protein-coding orthologs , 85% in maize and 89% in sorghum [47] . The lower percentage of syntenic sorghum miRNA genes may be indicative of false positives within this set , as these did not undergo the same rigorous screening process as for maize . Synteny was found amongst all families except miR827 and miR482 ( Table S8 ) . The former has a single representative in each genome , located in non-syntenic regions; the latter has one member in maize but none annotated in sorghum . As shown in Figure 6 , conserved synteny among miRNA genes was detected on all chromosomes of maize and sorghum . This figure also shows that many miRNA genes in sorghum map to both sister sites created after the genome-wide duplication event in maize . Many miRNA genes are organized within paralog clusters , defined as family members having no more than two intervening genes . Some of these are comprised of compact clusters , as described above . In maize , we found 13 paralog clusters containing 40 genes in total , while sorghum has 15 clusters containing 47 genes . All of the 13 paralog clusters in maize are at syntenic positions . In sorghum , 14 of the clusters ( 43 genes ) are in syntenic regions . Since the relationship between paralog clusters is not known between species ( some may be orthologs and some paralogs depending on the timing of tandem duplications relative to speciation ) we collapsed these into single sites . This led to 104 sites in sorghum and 123 sites in maize . Of these , 81 sites are syntenic in sorghum , therefore representing ancestral locations of these genes in the common ancestor of maize and sorghum . In maize , 28 of these sites are retained in sister duplicate positions . Thus , almost 35% of ancestral miRNA positions have been retained at sister homoeologous positions in maize , as detailed in Table S9 . This is higher than the retention rate of protein-coding genes , measured at ∼21% . Maize had a net gain of 30 syntenic genes compared to sorghum . All of these gains occurred at retained duplicate sites . Amongst singleton sites ( those harboring miRNA genes on just one sister region ) , there was a deficit of five maize genes due to the reduced sizes of four paralog clusters relative to sorghum . Whether this was caused by the expansion of clusters in sorghum or the contraction of clusters in maize is not known . In contrast , the combined gene counts at the 28 retained duplicate sites in maize exceed gene counts at the corresponding sorghum sites in all but two cases . This resulted in a net gain in maize of 35 genes at retained duplicate sites . There were twenty-two sites harboring single genes in sorghum with corresponding duplicate sites in maize also harboring single genes . An additional site contains exactly two genes at each corresponding region . Assuming these gene counts were stable through evolution , then a gain of 24 genes can be directly attributed to the whole genome duplication in maize . The remaining five retained sites have clusters with varying numbers of genes in sorghum and corresponding maize sister regions . These sites contributed a net gain of eleven maize genes , the combined result of whole-genome duplication and differential expansion/contraction of paralog clusters . The evolutionary dynamics influencing miRNA family size is illustrated by closer examination of the miR159 family . This family has the largest difference in membership , 11 genes in maize compared to 3 genes in sorghum ( Table S8 ) . This excess of eight genes includes three ( miR159d/e/g ) that have no syntenic counterparts in sorghum , suggesting long-distance movement in maize or loss of ancestral sites in the sorghum lineage . Another four exist within a paralog cluster that is expanded in maize relative to sorghum . In maize this cluster contains six genes ( miR159a/b/h/i/j/k ) on chromosome 8 while the corresponding cluster on sorghum chromosome 3 has only two genes . Finally there is a single gene ( miR159f ) at the homoeologous site on maize chromosome 3 which resulted from the whole genome duplication . Because of the characteristic structure and important regulatory function of miRNAs , it would be expected that the mature miRNAs be highly conserved across maize accessions . Sequence diversity in the flanking regions of the mature miRNA can be used to imply natural or artificial selection on the miRNA loci . To explore the genetic diversity of miRNAs in maize , we selected 28 maize miRNA genes ( Table S10 ) based on their availability in the initial reduced representation libraries of maize [53] , [54] . We designed sequencing primers for ∼400nt surrounding each mature miRNA sequence . The 28 maize miRNA genes and flanking regions were sequenced in 28 maize inbred lines chosen to maximize genetic diversity [55] . Additionally , 16 partially inbred teosinte lines [56] ( Zea mays ssp . parviglumis ) were selected to represent the diversity in the progenitor of cultivated maize . As expected , there was no polymorphism detected within the mature miRNA sequences in either the maize inbred or teosinte lines . However , the level of nucleotide diversity flanking the mature miRNA sequence is similar to the genic level of diversity: the average proportion of pairwise nucleotide differences per nucleotide site ( π ) is 0 . 0065 for miRNA genes in the maize inbred lines versus 0 . 0067 for a random collection of 1 , 095 maize genes[57] . By comparing genetic diversity in inbred versus teosinte accessions , we tested if any of the miRNAs had a reduction in polymorphisms consistent with the selective sweep during maize domestication or improvement . We found π = 0 . 0105 for the miRNA genes among the teosinte lines compared to π = 0 . 0095 for a large collection of teosinte protein coding genes [56] . We found no evidence of such a selective sweep ( Table S11 ) suggesting that mature miRNA sequences are under strong purifying selection due to their functional importance , making them highly conserved over long evolutionary periods , while flanking regions evolve equivalent to other genes .
We have systematically annotated miRNA genes on the first complete assembly of the maize genome . Our computational methods for detection were stringent by design , distinguishing 150 genes in 26 miRNA families from many potentially spurious predictions . This high confidence set represents genuine miRNA genes is reinforced by our characterization of pre-miRNA structures , promoter features , expression profiles , genetic diversity , and identification of orthologous genes in sorghum ( see further discussion below ) . Plant miRNA families can be historically divided into two classes: the highly conserved families , and the emerging class of lineage-specific families . Recent work suggests that a subset of the predicted lineage-specific families have characteristics of siRNA genes , associated with transposon-related repeats and showing little or no expression at the pre-miRNA level [58] . One such family is miR414 , which is currently under consideration for removal from miRBase [5] , [28] . Another example is the miR437 family , which has expanded in the sorghum and maize lineages relative to rice . In maize , more than 90% of miR437 pre-miRNA structures show the presence of the Stowaway miniature inverted repeat transposonable element ( MITE ) . Interestingly , these MITEs account for more than 87% of the repeats found in miRNA genes . In our analysis , maize transposon-related repeats were found in 11 miRNA families ( miR437 , miR854 , miR1128 , miR1132 , miR1133 , miR1320 , miR1435 , miR1436 , miR1439 , miR1884 , and miR2102 ) , thus making them suspect and supporting their removal from the high-confidence set ( Table S1 ) . In [27] , Voinnet also noted the association of recently evolved families with MITEs and raised similar questions as to whether they represent true miRNAs versus siRNAs . In animals , approximately 80% of miRNAs are found within introns of either protein-coding or non-coding genes [59] . In contrast , most of the annotated miRNAs in plants are located in intergenic regions , with some exceptions [60] . In the current work , 87% of the miRNAs were found in intergenic regions . Exceptions were 19 genes whose mature miRNA coordinates lie within exons of predicted protein coding genes . In six of these cases the host genes were also the predicted target of the embedded miRNA gene , oriented either on the same strand ( miR159c , miR319b , miR396a and miR397a ) or opposite strand ( miR169j and miR399d ) . Of the remaining 13 miRNA genes , all except miR159e were located on the same strand as the protein-coding genes they are embedded in . miR159e was not expressed in any of the tissues sampled and has no syntenic relationship with any of the sorghum miR159 genes ( see sections below ) , suggesting that miR159e might not be functional . The protein-coding genes with embedded miRNAs are small ( encoding proteins of less than 120aa compared to the average of 358aa [26] ) and have no identifiable InterPro domains . We speculate that a proportion of these may not be bona fide protein-coding genes but likely the result of misannotation based on populations of non-coding transcripts used in the evidence-based prediction method [26] . MicroRNA tissue-specificity is known to play a role in plant development , an example of which is the regulation of juvenile-to-adult vegetative phase transition [61] . To assess miRNA tissue specificity , we surveyed miRNA expression levels in five maize tissue types ( root , seedling , tassel , ear , and pollen ) . In our data , we found that the expression of miR156 and miR172 families are anti-correlated ( Table S4 ) ; miR156 is expressed higher in young roots and seedlings but lower in adult tissues ( tassel , ear , and pollen ) , while miR172 expression has an opposite trend , albeit with a lower overall expression level . These results fit well with the current models of phase transition , whereby opposing gradients of miR156 and miR172 are responsible for the transition from juvenile to adult . This converse regulatory relationship between miR156 and miR172 has also been reported recently [62] . The miR529 family is expressed conspicuously higher in the tassel as compared with the other surveyed tissues . We note that the mature miRNA products of families miR529 and miR156 are identical from positions 8–14; which are key residues in miRNA-target recognition . Consequently both miRNA families have similar predicted targets consisting mainly of SQUAMOSA-promoter binding proteins ( SBP ) . It is likely that miR529 is either related to , or is a sub-grouping of , the miR156 family , but with a distinct tissue expression profile . As seen in the miR396 gene family , tissue expression profiles are not necessarily conserved amongst miRNA genes of the same family . This family consists of 7 members that exhibit distinct expression profiles ( Figure 4 , Table S2 ) – miR396a/b/g shows elevated expression in the pollen , miR396c/d is highly expressed in the juvenile tissues but not adult tissues , while miR396e/f expression profile shows a peak in seedling tissues . miR396c/d mature miRNA sequences carry an additional guanine residue between positions 8 and 9 as compared to the rest of the miR396 family . Other members of the miR396 family are conserved in both monocots and dicots , but miR396c/d appear to be monocot specific ( equivalent to osa-miR396d/e found in rice ) [63] . The latter are also the only members of the miR396 family that target QLQ ( glutamine-leucine-glutamine ) and WRC ( tryptophan-arginine-cysteine ) domains that define the Growth-Regulating Factor ( GRF ) family of transcription factors [64] . The GRF transcription factors are involved in leaf and cotyledon growth and expressed most abundantly in active developing tissues [65] . QLQ has been found to be involved in mediating protein interactions whereas WRC plays a role in DNA binding [64] . Therefore , miR396c/d could be acting as regulatorsr of GRF genes in juvenile tissues independently from the rest of the miR396 family . There are also miRNA gene families that appear to be constitutively expressed in all 5 tissues . For example , miR168 family has a read count of at least 2400 RPM in all tissues surveyed , reflecting its role in maintaining a steady-state balance of the RNA silencing machinery by targeting the slicer AGO1 ( ARGONAUTE1 ) of the RISC complex [66]–[69] . Recent transcriptome analysis of Arabidopsis indicates that 15 genes involved in the miRNA pathway ( including AGO1 , AGO2 , AGO4 AGO7 , and DCL1–3 ) are absent in pollen [70] . In the current study , we observed that expression of most families are lower in pollen as compared with other tissues , suggesting that the miRNA biogenesis machinery is similarly down-regulated in maize pollen tissues . Of the 26 miRNA families on our refined list , only miR482 ( consisting of a single member ) failed to register any evidence of expression in the 5 tissues sampled . While miR482 had previously been annotated in poplar , pine , soybean and grape , it has not been identified in any monocots . Our analysis of the pri-miRNA transcripts shows that these genes exhibit many of the conserved signatures of RNA polymerase II transcripts . These include TATA boxes in conserved locations , alternative TSS and polyadenylation sites , and introns [39] . Using the stems of the pre-miRNA for orientation , we found that , on average , the 5′ region is shorter than the 3′ region . The average length of the 3′ region is likely an underestimation due to ascertainment biases inherent in the RACE protocol , where smaller transcripts are more likely to be amplified than longer ones . This may also explain the shorter average length of pri-miRNAs amplified by RACE as compared with the average length of FL-cDNA that harbor pre-miRNA sequences ( Table S3 ) . The frequency of TATA-box appearance ( 83% ) is similar to that of Arabidopsis miRNA transcripts ( 83% ) [39] and much higher than average for protein promoters ( 50% ) [43] . As in [39] , we also found a high occurrence of cytosine at the −1 position and adenine at the +1 position of TSSs ( Figure 2 ) . Together , such characteristics as the TATA box , the nucleotide frequency surrounding the TSSs and expected size of 5′ transcript regions , will help in the development of new methods to predict proximal promoters of miRNA genes . Our target prediction method is stringent , but still allows us to capture most miRNA targets that are conserved across several plant species , including Arabidopsis [4] , [71]–[73] , poplar [21] , rice [67] , [74] , [75] , wheat [76] , soybean [77] , mustard [78] , and grape [79] . For example , miR156 targets SBP transcription factors [80] , [81] , while miR159 targets the MYB family [82] . Both miR160 and miR167 target Auxin Response Factor ( ARF ) transcription factors in Arabidopsis and maize [83] , [84]; and are also captured by our predictions . The same trend is observed in many other miRNA families including miR164 , miR166 , miR169 , miR171 , miR172 , miR319 and miR396 as they target various families of transcription factors such as NAM ( No Apical Meristem ) proteins , bZIP ( basic-leucine Zipper ) genes , CBF ( CCAAT binding factor ) , GRAS transcription factor , AP2 ( APETALA2 ) -EREBP ( Ethylene-Responsive Element Binding Proteins ) , CCCH type zinc finger protein and TCP ( Teosinite branched , Cycloidea , and PCF ) , GRF transcription factor families respectively [12] , [71] , [85]–[87] . These transcription factors are known to regulate plant development . MiRNA families miR393 and miR394 target F-box protein and are known to play a role in the expression control of genes involved in regulation of metabolic processes [88] . It has been reported that in Arabidopsis , rice , and wheat , miR168 targets AGO1 ( ARGONAUTE1 ) [66]–[69] . The prediction pipeline captured three AGO1 orthologs in maize [89] . Our target predictions are further consistent with the literature in the case of miR395 , which regulates sulfate metabolism by targeting sulphate transporter genes and ATP sulfurylase ( APS ) proteins , and whose expression increases in response to low sulfate growing conditions [73] . We also predicted potential targets of miRNA families not previously identified in maize ( miR482 , miR528 , miR529 , miR827 and miR1432 ) . Trehalose phosphatase , cytochrome P450 , pentatricopeptide , and CCHC type zinc finger protein were predicted as targets of miR482 , suggesting the involvement of miR482 in a wide range of biosynthetic reactions . Similar to miR397 and miR408 , miR528 also targets copper proteins cupredoxin , multicopper oxidase and laccase genes and thus might play a critical role in regulating physiological processes ( photosynthetic and respiratory electron transport ) and stress responses . Both miR156 and miR529 were predicted to target genes encoding the SBP box . miR827 was predicted to target NAD ( P ) -binding and SPX ( SYG1/Pho81/XPR1 ) proteins , whereas miR1432 was predicted to target Poly ( ADP-ribose ) polymerase , catalytic region functions and calcium binding EF hand domains . Both SPX and calcium binding EF hand domains are associated with proteins that are active in signal transduction pathways . The stringent criteria used to predict targets could potentially reduce false positive rates at the cost of missing several authentic targets . For example , miR162 targets DCL1 [90] , but this target ( GRMZM2G040762 in maize ) was excluded by our pipeline due to the presence of a 1nt bulge in the alignment with miR162 . In addition , for about 15% of the predicted targets , functional information is not available at this time . We hypothesize that some of these targets may be novel but this needs to be further verified based on experimental evidence . The maize lineage tetraploidy event occurred an estimated 5 to 15 MYA [91]–[94] , and precipitated large-scale chromosomal rearrangements as well as massive losses of duplicate genes in the process of returning to a genetically diploid state [26] , [94]–[97] . Our analysis showed that greater than 1/3 of ancestral miRNA positions were retained at both homoeologous sites and these accounted for the majority of gene family expansions in maize relative to sorghum . Interestingly this proportion of retained sites was greater than that seen for protein-coding genes . Gene duplication , especially by polyploidization , has long been thought to provide raw material for the evolution of functional novelty [98] . The initial redundancy of duplicate genes is followed by a period of relaxed selection with gene loss being the ultimate fate for most duplicates [99] , [100] . Classical models for retention of duplicate homoeologs include neofunctionialization , subfunctionalization , and maintenance of stochiometry among interacting components of pathways [101] . Evidence for subfunctionalization has been reported in tandem duplicate paralogs of miR168 in the Brassicaceae[50] . However , more recent hypotheses on the causes of retention following genome-wide duplication events are based on common characteristics of retained genes . For example , studies of the after effects of polyploidy in Arabidopsis [102] , rice [103] and most recently maize [26] show bias for retention of transcription and other regulatory factors . Such classes may have indispensable functions such that their retention provides buffering against gene loss [104] . As shown here and elsewhere , miRNA genes are largely associated with regulatory functions and thus mutations in miRNAs can be expected to have profound deleterious effects on growth and development [62] . Thus the observed preferential retention of miRNA genes compared to protein coding genes is consistent with both the genetic buffering hypothesis and with previous observations of biased retention of regulatory genes . The majority of maize miRNA genes ( 91% ) lie within ancestral locations that existed prior to the speciation of maize and sorghum . Those no longer at syntenic positions could represent lineage-specific movements , duplications , or losses . Identification of ortholog genes has important practical application , allowing inferences to be drawn between functional studies in each species . Conservation of gene position has proven useful for determining orthologous relationships , particularly in cases where multiple highly related paralogs exist at different genome locations [105] , [106] . Such is the case for miRNA genes , which tend to occur in large yet highly conserved families . Although synteny provides a strong indication of orthology , potentially complex relationships are suggested in the case of paralog clusters , in which closer phylogenetic analysis may be required to sort out duplication histories [50] . Nevertheless , we found simple synteny relationships for 69 sorghum genes and 81 maize genes , allowing presumptive assignment of orthology among these genes . To investigate the evolution of miRNA loci we sequenced 28 loci and flanking regions in panels of inbred and teosinte lines . This germplasm set was selected as it provides an excellent representation of the allelic diversity in maize and wild relatives . As expected , there was no polymorphism detected within the mature miRNA sequences . Their conservation within maize throughout its evolution is expected given the importance of miRNA genes in suppressing target gene expression during development and stress . The flanking regions displayed diversity levels similar to protein coding genes indicating that purifying selection is limited to mature miRNAs . None of the 28 loci tested exhibited the extreme reductions in diversity in inbreds relative to teosinte accessions that would be indicative of artificial selection during domestication or crop improvement . MicroRNA loci may control such fundamental processes in development that alterations of sequence or expression are not tolerated . Our results are similar to studies in Arabidopsis [107] , [108] and rice [35] where strong purifying selection has acted on the mature miRNA , whereas flanking regions generally reflect species wide diversity levels . In summary , we have investigated genome-wide maize miRNA genes from several aspects: their pre- and pri- structure , their expression level , their targets , their conservation , and their evolution , providing evidence of the important function of miRNAs in regulating metabolic and developmental process and adaptation to stress . Identification and characterization of this important class of regulatory genes in maize may enable breeders to engineer crops with improved architecture and stress responses necessary to increase yields for food and fuel , while addressing environmental concerns in our changing climate and ever-increasing human population .
Maize ( Zea mays ) inbred line B73 was used in this study . The seedlings were collected from plants grown in soil in a green house at 22°C with a 16 hr light cycle and harvested 7 days after germination . Roots were collected from B73 seeds grown in sterile water without light at 30°C for three days . Immature ear ( size 0 . 5–2cm ) and immature tassel ( size 0 . 5–2 . 5cm ) were harvested from plants grown in the field . All the tissues were harvested and immediately frozen in liquid nitrogen and stored at −80°C . Tissues used in the RACE analysis included seedlings and immature inflorescence ( tassel and ear ) . A list of plant mature miRNAs were obtained from miRBase ( version 13 . 0 ) and aligned using Vmatch , an alignment algorithm that employs suffix arrays ( www . vmatch . de ) , against a suffix array index of the maize genome sequence built using mkvtree from the vmatch package . Up to 2 mismatches were allowed in the alignment . A 250-nt contextual sequence surrounding each aligned mature miRNA was extracted and secondary structure modeling of the DNA fragment was performed using Mfold [109] . Segments that could potentially form the canonical stem-loop signature of a precursor miRNA were tagged as putative miRNAs . This putative list was further trimmed based on the following criteria: ( 1 ) overlapping loci were reduced to non-redundant representations , ( 2 ) loci that overlapped transposable and other repetitive elements were removed , ( 3 ) loci in which the matched mature miRNA was in the wrong orientation were removed , ( 4 ) each predicted secondary structure was curated manually and those with spurious stem-loop structures were removed [73] , and ( 5 ) miRNA families that had no evidence of expression in the small RNA libraries or RACE were removed ( for more details see Figure S2 ) . The final refined list of predicted miRNAs consisted of 150 individual miRNAs from 26 miRNA families . Based on the hairpin structure of the pre-miRNA , the mature miRNA and corresponding miRNA* sequence were identified for each . We used a modified method based on Llave et al to prepare small RNA libraries [2] . Total RNA from different tissues was extracted with TRIzol solution ( Invitrogen ) . 100µg of total RNA were separated on 15% Acrylamide/8M Urea gel ( Sequagel , National Diagnostic ) along with 19 and 24nt RNA samples as marker . Small RNAs were extracted from the gel slices corresponding to 19 and 24nt RNA using high salt . Small RNAs were ligated with Modban ( Linker 1 from IDT ) . Ligated samples were separated again on 15% Acrylamide/8M Urea gel . The gel fragments corresponding to 37–42nt were excised . Small RNA was purified from the gel fragments , and 5′ sequencing adaptors were added using T4 RNA ligase . This RNA was amplified using Superscript III Reverse transcriptase ( Invitrogen ) and PCR amplification ( Phusion HF polymerase , NEB ) . Amplified cDNA were separated on 3% MetaPhor Agarose ( VWR ) and the bands corresponding to correct size ( 108–115nt ) were cut and purified with QIAGENE gel purification kit and sent for Illumina sequencing . Small RNA libraries were sequenced on an Illumina Genome Analyzer using the 36-cycle Solexa Sequencing Kit ( Illumina ) . The Illumina Gerald pipeline was used to process and extract the first 36 bases of the runs . Adaptor sequences were identified and trimmed from each read using a customized Perl script . Reads in which the adaptor could not be identified were discarded . Novoalign ( version 2 . 03 , http://www . novocraft . com/ ) was used to align the trimmed reads to the high-confidence set of 150 pre-miRNAs . For each library , we counted the number of trimmed reads within the 18–22nt range that were mapping to each pre-miRNA and normalized by the total number of 18–22nt trimmed reads in the library . Trimmed reads that were <18nt or > = 23nt were not considered in this analysis . Total RNA was extracted from maize seedlings and immature inflorescence using Spectrum Plant Total RNA kit ( Sigma ) and treated with RNase Free DNase ( QIAGEN ) . cDNA templates were prepared by following the instruction of FirstChoice RLM-RACE Kit ( Ambion ) . For each miRNA precursor , two sets of gene-specific primers , mostly immediate upstream and downstream of predicted hairpin structures ( for 5′ RACE , for 3′ RACE , respectively ) , were designed using Primer 3 [110] ( Table S5 ) . These primers were used for two rounds of PCR amplification . Nested PCR products were analyzed on agarose gel . Positive PCR products were cloned into pCR2 . 1-TOPO vector using TOPO TA cloning kit ( Invitrogen ) . Each transcript sequence was confirmed by at least 8 unique clones . Sequences corresponding to the transcripts were mapped to the genome and listed in Table S3 . The miRNA target transcript prediction pipeline was developed using Vmatch . For the maize protein coding transcripts , the predicted cDNA of the longest consensus maize transcript from the filtered gene sets were used in this analysis and indexed by using mkvtree ( Vmatch ) [26] . The mature miRNA sequences from the refined miRNA set were reverse complemented and matched against the indexed maize transcript database , with the parameters relaxed to allow up to six mismatches . The matched miRNA target transcripts were further filtered by applying empirical rules defined by Schwab et al [111] . Briefly , we scored each miRNA complementary site . Perfect matches were given a score of 0 , and all other mismatches were scored 1 . Only 1 mismatch score was allowed between positions 2 to 12 inclusive . However , no mismatches were allowed at position 10 and 11 and no more than 2 consecutive mismatches were allowed after position 12 . A maximum of three mismatches ( excluding mismatch at position 1 ) was allowed across the length of the mature miRNA . We subjected potential miRNA targets for functional enrichment analysis against GO molecular function and GO biological process terms database by using BiNGO [48] which is a Cytoscape [49] plugin that maps over-represented functional themes present in a given gene-set onto the GO hierarchy . For enrichment P-value calculation ( at a significance level of 0 . 05 or better ) , a hypergeometric distribution statistical testing method was selected to ensure that target genes are not hitting their corresponding biological function/process classes purely by random chance . For multiple hypotheses testing , the Benjamini and Hochberg false discovery rate ( FDR ) correction [112] was applied to reduce false negatives at the cost of a few more false positives . In order to fully understand the function of each target , we also extracted information based on phylogenetic trees of putative genes from maize , sorghum , Arabidopsis , and rice genomes . The trees were computed using the Ensembl Compara GeneTree method [113] . The input sequences used for the GeneTree analysis were the longest translation at a given gene locus , filtered for transposons and other low-confidence genes , from the following genome annotation resources: the maize genome ( www . maizesequence . org ) release 4a from June 2009 [26] , the Sorghum genome [29] JGI release Sbi 1 . 4 from March 2008 , The Arabidopsis Information Resource [114] release 8 from April 2008 , and the MSU/TIGR Rice Genome Annotation Resource [115] release 5 from January 2007 . Shared synteny among homologous maize and sorghum miRNA genes was evaluated in the context of orthologous protein-coding genes . Orthologous protein-coding genes were identified using the Ensembl Compara pipeline , which is based on a phylogenetic analysis [116] . In all , the orthologue sets included 25 , 216 maize genes and 20 , 408 sorghum genes , which participated in 27 , 275 relationships . DAGchainer [117] was used to identify colinear chains amongst these orthologs , together with all pairwise combinations of within-family miRNA genes . Such chains were required to have at least five colinear genes with no more than ten intervening genes between neighbors . Additional syntenies among non-colinear genes were searched based on distance to colinear anchors that flank the orthologue of the gene in question in the other genome . The gene in question was considered syntenic if positioned within five genes of its nearest anchor . Paralogous clusters of miRNA genes were identified as family members separated by no more than two intervening genes . Two diverse sets of maize materials were used for DNA sequence analysis: maize inbred lines , and partially inbred teosinte accessions ( Table S11 ) . The 28 maize inbred lines ( B73 , B97 , CML103 , CML228 , CML247 , CML277 , CML322 , CML333 , CML52 , CML69 , Hp301 , IL14H , Ki11 , Ki3 , Ky21 , M162W , M37W , Mo17 , Mo18W , MS71 , NC350 , NC358 , Oh43 , Oh7B , P39 , Tx303 , Tzi8 and W22 ) were chosen to maximize allelic diversity . The 16 partial inbred teosinte ( Zea mays ssp . parviglumis ) accessions are identical to those used previously [56] , [57] . We conducted extended DNA sequencing of the known 28 maize miRNAs . Primer 3 was used to design the sets of PCR primers for analysis of the selected candidate miRNA genes . The genomic PCRs were performed using PCR master Mix I ( Promega ) or Takara LA Taq polymerase . Unincorporated primers and dNTPs were removed by exonuclease I ( NEB ) and shrimp alkaline phosphatase ( USB ) . The PCR products were ethanol precipitated and sequenced with forward , reverse , and internal primers using Illumina machines . Base calling , quality assessment , and trimming of trace files were conducted with PHRED and sequence assembly was performed by PHRAP . The multiple sequences for each gene were aligned with ClustalW . We used DNAAlign Editor [118] for aligning the amplicon sequences across the germplasm set to accurately identify nucleotide variations , insertions , and deletions . The alignments are available to the community from the www . panzea . org [119] . The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus [120] and are accessible through GEO Series accession number GSE17943 ( http://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE17943 ) . | MicroRNAs are non-coding RNAs that regulate gene expression post-transcriptionally and play roles in diverse pathways including those acting on development and responses to stress . Here , we describe a genome-wide computational prediction of maize miRNA genes and their characterization with respect to expression , putative targets , evolution following whole genome duplication , and allelic diversity . The structures of unprocessed primary miRNA transcripts were determined by 5′ RACE and 3′ RACE . Expression profiles were surveyed in five tissue types by deep-sequencing of small RNA libraries . We predicted miRNA targets computationally based on the most recent maize protein annotations . Analysis of the predicted functions of target genes , on the basis of gene ontology , supported their roles in regulatory processes . We identified putative orthologs in Sorghum based on an analysis of synteny and found that maize-homoeologous miRNA genes were retained more frequently than expected . We also explored miRNA nucleotide diversity among many maize inbred lines and partially inbred teosinte lines . The results indicated that mature miRNA genes were highly conserved during their evolution . This preliminary characterization based on our findings provides a framework for future analysis of miRNA genes and their roles in key traits of maize as feed , fodder , and biofuel . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/genomics",
"computational",
"biology/comparative",
"sequence",
"analysis",
"plant",
"biology/plant",
"genetics",
"and",
"gene",
"expression",
"genetics",
"and",
"genomics/plant",
"genetics",
"and",
"gene",
"expression"
] | 2009 | A Genome-Wide Characterization of MicroRNA Genes in Maize |
The macaque parasite Plasmodium knowlesi is a significant concern in Malaysia where cases of human infection are increasing . Parasites infecting humans originate from genetically distinct subpopulations associated with the long-tailed ( Macaca fascicularis ( Mf ) ) or pig-tailed macaques ( Macaca nemestrina ( Mn ) ) . We used a new high-quality reference genome to re-evaluate previously described subpopulations among human and macaque isolates from Malaysian-Borneo and Peninsular-Malaysia . Nuclear genomes were dimorphic , as expected , but new evidence of chromosomal-segment exchanges between subpopulations was found . A large segment on chromosome 8 originating from the Mn subpopulation and containing genes encoding proteins expressed in mosquito-borne parasite stages , was found in Mf genotypes . By contrast , non-recombining organelle genomes partitioned into 3 deeply branched lineages , unlinked with nuclear genomic dimorphism . Subpopulations which diverged in isolation have re-connected , possibly due to deforestation and disruption of wild macaque habitats . The resulting genomic mosaics reveal traits selected by host-vector-parasite interactions in a setting of ecological transition .
Plasmodium knowlesi , a common malaria parasite of long-tailed Macaca fascicularis ( Mf ) and pig-tailed M . nemestrina ( Mn ) macaques in Southeast Asia , is now recognized as a significant cause of human malaria . A cluster of human P . knowlesi cases were reported from Malaysian Borneo in 2004 [1] , but now human infections are known to be widespread in Southeast Asia [2 , 3] , and have been reported in travellers from outside the region [2 , 4] . Clinical symptoms range from asymptomatic carriage to high parasitaemia with severe complications including death [5 , 6] . As rapid human population growth , deforestation and encroachment on remaining wild macaque habitats potentially increases contact with humans [7] , in Southeast Asian countries P . knowlesi is now coming to the attention of national malaria control and elimination programmes that have hitherto focused on P . vivax and P . falciparum [2] . P . knowlesi commonly displays multi-clonality in humans and macaques , and analysis of microsatellite markers , csp , 18S rRNA , and mtDNA sequences indicates no systematic differences between human and macaque isolates from Malaysian Borneo [8] . Whole genome-level genetic diversity among P . knowlesi from human infections in Sarikei in Sarawak demonstrates substantial dimorphism extending over at least 50% of the genome [9] . This finding is supported by analysis of microsatellite diversity in parasites from Mf , Mn and human infections across Peninsular and Borneo Malaysia [10] . It also provides evidence that the two distinct genome dimorphs reflect adaptation to either of the two host macaque species , although no evidence of a complete barrier in primate host susceptibility was found [10] . A third genome cluster has been described from geographically distinct Peninsular Malaysia [11 , 12 , 13 , 14] . Studies of mtDNA have revealed that ancestral P . knowlesi predates the settlement of Homo sapiens in Southeast Asia , the evolutionary emergence of P . falciparum and P . vivax , and underwent population expansion 30–40 thousand years ago [8] . Diversity at the genomic level is thus likely to reflect host- and geography-related partitioning during this expansion , as well as additional recent complexity due to contemporary changes in host and vector distributions during ongoing ecological transition in the region [15] . Several Anopheles species , all from the Leuchosphyrus group , are capable of transmitting P . knowlesi malaria , including A . latens and A . balbacensis in Malaysian Borneo [16 , 17 , 18] , A . hackeri and A . cracens in Peninsular Malaysia [19] and A . dirus in southern Vietnam [20] . It is thus likely that patterns of genome diversity in natural populations of P . knowlesi reflect partitioning among both Dipteran and primate hosts occurring on varying time-scales through the evolutionary history of the species . Such partitioning can plausibly prevent or reduce panmictic genetic exchange . Genomic studies of P . knowlesi to date have considered nuclear gene diversity and dimorphism among naturally-infected human hosts , and macaque-derived laboratory-maintained isolates from the 1960s [10 , 12] . However , these studies did not consider non-nuclear organellar genomes in the mitochondrion and apicoplast of malaria parasites , which are non-recombinant and uniparentally inherited , and can provide evidence of genome evolution on a longer timescale [21] . Recombination barriers among insect and primate hosts may have less impact on sequence diversity in the organellar genomes of P . knowlesi . Utilising a new P . knowlesi reference genome generated using long-read technology [22] we performed a new analysis of all available nuclear and non-nuclear genome sequences . Patterns of polymorphisms were analysed to identify evolutionary signals of both recent and ancient events associated with the partitioning of the di- or tri-morphic genomes previously reported .
Raw short-read sequence data from all available P . knowlesi isolates ( S1 Fig ) were mapped to a new reference genome [22] from the human-adapted P . knowlesi line A1-H . 1 genome [23] , yielding an average coverage of ~120-fold across 99% of the reference genome ( S1 Table ) , and 1 , 632 , 024 high quality SNPs . The high density of point mutations ( 1 every 15bp ) in P . knowlesi compared to P . vivax and P . falciparum has been previously noted [10] . Seven macaque-derived isolates were found to have high multiplicity of infection ( S2 Fig ) , and were excluded , leaving an analysis set of 60 isolates . SNP-based neighbour-joining tree analysis revealed three subpopulation groups that coincide with isolates presenting the Mf-associated P . knowlesi genotype ( Mf-Pk , Borneo Malaysia , Cluster 1 ) , the Mn-associated P . knowlesi genotype ( Mn-Pk , Borneo Malaysia , Cluster 2 ) [10 , 11 , 12 , 14] , and older Peninsular Malaysia strains ( Cluster 3 ) ( Fig 1A ) . Within Cluster 1 we observed two geographic sub-groups that coincide with Kapit and Betong regions in Malaysian Borneo . The samples from Sarikei region ( DIM prefix ) , geographically located equidistant between Kapit and Betong , fall into either cluster ( S3 Fig ) . Overall , the regional clusters from Kapit and Betong were more genetically similar to each other ( mean fixation index FST 0 . 03 , S4 Fig ) than were the host-associated clusters ( Cluster 1 vs . 2 , mean FST 0 . 21 ) . However , a significant chromosomal anomaly was identified that differentiated the Kapit and Betong Mf-Pk subgroups; this occurred in a multi-gene region on chromosome 8 ( ~500 SNPs with FST values >0 . 4; Fig 1B; S4 Fig ) . To explore the anomaly in chromosome 8 , individual haplotypes and neighbour-joining trees were constructed across several loci ( Fig 1C and Fig 1D ) revealing two very distinct patterns . The first pattern was observed in the chromosomal sections with low genetic diversity between the two Mf-Pk regional clusters ( FST < 0 . 2 , Fig 1B ) . The tree structure for these genomic regions ( Fig 1D , 1st tree ) mimics that of the genome-wide tree in Fig 1A . Strong haplotype differentiation between the host-associated Clusters 1 ( Mf-Pk ) and 2 ( Mn-Pk ) was confirmed in the SNP-based profiles ( Fig 1C , 1st column ) . A second pattern was observed in regions of chromosome 8 with distinct genetic differentiation between Kapit and Betong subgroups ( FST > 0 . 4 ) . Many Mf-Pk Betong subgroup isolates presented segments almost identical to chromosome 8 sequences of the Mn-Pk genotype from Cluster 2 ( Fig 1D , 2nd , 3rd and 4th trees ) . This exchange is supported by the SNP-based haplotype patterns , where a distinct haplotype in the Betong samples is Cluster 2-like ( Fig 1C , 2nd , 3rd and 4th columns , black arrows ) , suggesting the introgression of large chromosomal regions ( up to 200Kb ) between Mf-Pk ( Cluster 1 ) and Mn-Pk ( Cluster 2 ) . This is consistent with a very recent event of natural genetic exchange between these subgroups of P . knowlesi recently isolated from human infections . The high frequency of the new haplotype ( 73% ) in the Betong subgroup suggests that it is under ( recent ) strong selection pressure in this region . The presence of differences in extended haplotype homozygosity between the recombinant and non-recombinant regional Mf-Pk subpopulations provides additional evidence of recent positive selection ( XP-EHH peak , P<0 . 0001 ) in a region of increased population differentiation ( FST > 0 . 4 , Fig 1B ) . The functional nature of genes in chromosome 8 involved in these putative introgression events was investigated ( FST > 0 . 4 , Table 1 ) , and found to include loci that are important in the vector component of the Plasmodium life cycle . For example , cap380 ( PKNH_0820800 , 101 SNPs with FST > 0 . 4 ) encodes a protein expressed in the external capsule of the oocyst . This gene is essential in the maturation from ookinete into oocyst in P . berghei , and is assumed to assist in evasion of mosquito immune mechanisms [24] . Another gene , PKNH_0826900 ( 19 SNPs ) encodes for the circumsporozoite- and TRAP-related protein ( CTRP ) , which has an established role in ookinete motility in P . berghei and is essential for binding to and invading the mosquito midgut [25] . Further , homologues of PKNH_0826400 ( 21 SNPs ) display increased transcription levels in ookinete and gametocyte V sexual stages in both P . falciparum [26] and P . berghei [27] compared to the asexual ring stage ( fold change of at least 2 ) . The transcriptomic profiles of these strongly selected genes are shown in S5 Fig . By applying a combination of neighbour joining trees and SNP diversity analysis across 50 Kbp windows , we identified that 33/60 isolates show clear evidence of genetic exchange between Clusters 1 and 2 ( S2 Table ) . Regions involved in exchange ( recombination ) ( 137/494 regions , 86% contained an ookinete related gene ) showed evidence of enrichment for ookinete-expressed genes compared to other ( non-recombinant ) chromosome regions ( 357/494 regions , 77% contained an ookinete related gene ) ( Chi Square P = 0 . 03 ) . One such region in chromosome 12 included the Pf47-like ( PKH_120710 ) gene , where the orthologue in P . falciparum is a known mediator of the evasion of the mosquito immune system [28] . Furthermore , it has been shown that a change in haplotype in this gene in a P . falciparum isolate is sufficient to make it compatible to a different mosquito species [28] . Nearly half ( 45% ) of isolates from Betong presented with a recombinant profile in PKH_120710 . In general , the genetic exchanges generated differing levels of mosaicism in each population and among individual isolates across all chromosomes ( S6 Fig ) . One isolate from Sarikei with the Mf genome dimorph type ( DIM2 ) appeared to harbour Mn-type introgressed sequences in 8% of the genome , occurring across 6 chromosomes ( 6 , 7 , 8 , 9 , 11 and 12 ) , including an almost complete Mn-type chromosome 8 . Of the 33 samples with evidence of exchanges , 13 were from the Betong region , 14 from Kapit and 6 from Sarikei , which indicates that the events are not geographically restricted . Although , the majority of genetic exchange events involve the integration of Mn-type motifs into Mf-type genomes , introgression in the opposite direction was also observed , but on a smaller scale and at lower frequency . The mitochondrial and apicoplast genomes of each P . knowlesi isolate was interrogated for signals of evolutionary history over longer time-scales , as in previous studies [21 , 29 , 30] . Combining the mitochondrial sequence data from the 60 P . knowlesi isolates from this study together with 54 previously published mitochondrial sequences including human and both Mn and Mf samples [9] , we generated a phylogenetic tree ( Fig 2 ) . This tree shows four clades ( shown in purple , red , blue and green ) . To interpret these clades , they were cross-referenced to the previously defined 3 nuclear genotypes ( Clusters 1 to 3 ) and the host contributing the sample ( human , macaque-type ) . The red and purple clades possess similar mitochondrial haplotypes as highlighted by their inter-cluster average FST ( red vs . purple: average FST = 0 . 16 ) , which is lower than comparisons including the other two clusters ( red or purple vs . blue or green: average FST > 0 . 18 ) . The purple clade consists of cultured isolates from Peninsular Malaysia , and is associated with the Peninsular nuclear genotype ( Cluster 3 ) . The red and green clades each contain a mixture of Borneo Malaysia samples from both humans and macaques with nuclear genotypes from Clusters 1 and 2 . The green clade also includes the only sequence sourced from a M . nemestrina host . The blue clade contains samples from humans and macaques , all with Cluster 1 nuclear genotypes . The divergence of these mitochondrial clades from their common ancestor was estimated to be 72k years ago , and younger than the previous the estimate of 257k but within error [8] . Furthermore , the presence of monkey-derived sequences spread across the tree seems to indicate that none of the mitochondrial genotypic groups found is human-specific as all have also been observed in macaques , also consistent with previous findings [9] . Using the common SNPs ( 280/425 with MAF > 5%: apicoplast 252 , mitochondria 28 SNPs ) in the 60 isolates with the sequence data we confirmed that the organellar genomes are co-inherited ( mean pairwise organellar linkage disequilibrium D’ = 0 . 99 ) . SNP-based haplotype profile analysis ( S7A Fig ) revealed clustering that is consistent with the three main clusters seen in Fig 2 . Similarly , a phylogenetic tree constructed using only apicoplast SNPs ( S7B Fig ) is congruent with the mitochondrial based tree ( Fig 2 ) . The presence of mismatched nuclear and organellar type genomes in two of the three clusters ( black arrows in Fig 2 ) and the presence of such mismatched samples with little or no evidence of nuclear genome recombination suggests ancient genetic exchange events between distinct lineages . The nuclear footprints of such exchanges are likely to have been broken down by recombination over time . We observed a significant incongruence between the robust phylogenetic tree topologies based on organellar and nuclear genome SNPs ( Shimodaira-Hasegawa test P = 0 . 001; Templeton test P = 0 . 003 ) ( Fig 2 ) . These results from organellar and nuclear genomes , in a small but geographically diverse set of P . knowlesi , indicate that there have been several genetic exchanges between the host-associated clusters in Malaysian Borneo .
P . knowlesi is now the major cause of malaria in Malaysian Borneo , but the biology of the parasite [15 , 22 , 23] , host and vector interactions , and disease distribution and epidemiology [19 , 31 , 32] are not well understood . The availability of a new high-quality reference sequence and a more robust approach to MOI were used to re-evaluate the previously described peninsular and macaque-associated subpopulations of P . knowlesi parasites . We report two major new findings . First , clear evidence of natural genetic exchanges between the divergent Mf- and Mn-associated subpopulations of P . knowlesi , including a major segment of introgression on chromosome 8 , is presented . Second , the presence of haplotype sub-divisions in the organellar genomes that do not map onto the subpopulations implied by nuclear genome analysis indicate that exchange events have previously occurred in non-recent history . A similar multi-tiered pattern of evolution among nuclear and organellar genomes has been found in Trypanosoma cruzi , an unrelated protozoan parasite with a mammalian host-insect vector life cycle [29 , 30] . Unexpectedly , observed mosaicism and population differentiation signals were not encountered equally across the P . knowlesi nuclear genome , but were particularly prominent on chromosome 8 , with genes expressed in mosquito stages over-represented . For example , the majority ( 73% ) of Mf-associated isolates from Betong harboured the Mn-associated allele of the oocyst-expressed cap380 gene , which differs at 101 positions from the allele found in the Mf-associated cluster . This is essential for ookinete to oocyst maturation and therefore for the transmission of the parasite during the vector stage [24 , 25]; here , we identify signals of recent selective pressure on this locus ( Fig 1B ) . Other vector-related genes were identified within the introgressed segment , and point towards strong evolutionary selection pressure on the parasites driven by the transmitting Anopheles vector species . Such effects have been found in P . falciparum [28] and P . vivax genomes [33] , and highlight the importance of understanding the distribution of the different Anopheles vector species , their host feeding preferences , and their interactions with the parasite in highly dynamic and complex environments such as the ecological niche of P . knowlesi . Nearly 80% of Malaysian Borneo has undergone deforestation or agricultural expansion , which have driven habitat modification affecting both macaque and Anopheles host species , and the proximity of humans to both [8 , 31] . Furthermore , studies have predicted that Mn predominantly inhabits forested areas while Mf reside in more cosmopolitan areas , which include croplands , vegetation mosaics , rubber plantations and forested areas [8 , 34] . The main genomic exchange event on chromosome 8 involves essential vector-related genes and is pin-pointed geographically to the Betong area . This region has undergone significant forest degradation due to expansion of industrial plantations in the recent years [35] . These types of environmental changes have been previously related to alterations in the vector species distribution in Malaysia , leading to malaria epidemics [36] . Environmental changes also affect macaque habitats , and increase the opportunities for human-macaque interaction [31] , but selection events highlighted in this study seem to primarily reflect adaptation of the parasite to changes in mosquito distribution or to recent changes in the vectorial capacity of the existing vectors . The depth , breadth and spread of the genetic exchanges observed in three different areas ( Betong , Kapit and Sarikei ) in Sarawak highlight the potential importance of these events for parasite adaptation in both vertebrate and invertebrate species . Although , the level of genetic diversity between Mf- and Mn-associated P . knowlesi has some similarity to that observed between P . ovale curtisi and P . o . wallikeri , now considered separate species [37] , the evidence of recombination and genetic exchanges observed in this study precludes species designation , as reproductive isolation is not complete . Nevertheless , better understanding of P . knowlesi population structure could aid future studies across the regions where human populations have been identified at risk of infection including both symptomatic and asymptomatic cases [4 , 38 , 39] . This would assist with characterising and tracking subpopulations and genetic exchanges , and provide a flexible framework for better understanding P . knowlesi diversity across the region . Our work has provided insight into Plasmodium parasite evolution . It has been suggested that malaria parasites have survived using either adaptive radiation where host switching plays a key role [40] , or alternatively adaptation to complex historical and geographical environments leading to speciation [41] . Plasmodium species in non-human natural conditions in the absence of drug selection pressure have a wide range of possible hosts [41 , 42] . The P . knowlesi data has shown that geographical or ecological isolation of the different hosts over an extended time can generate subgroups of parasites with substantial genetic differentiation , but capable of recombining when in contact [12 , 30 , 31] . This pattern has a major impact on the parasite genome , as illustrated by the profound chromosome mosaicism observed among our study isolates . Our data suggest that the broad host specificity of some of the Plasmodium species are important drivers of parasite genomic diversity . In P . knowlesi this means that genetic divergence is enabled not only by long-term geographic isolation , as is the case between Peninsular and Bornean isolates , but also via the isolation afforded by extended transmission cycles within different primate hosts . The genetic trimorphism suggests that the separate macaque hosts provides sufficient genetic isolation to allow for host specific adaptations to occur , even within relatively small geographic areas . Furthermore , the possibility of recombination between partially differentiated parasite genomes increases opportunities for new adaptation , including further host transitions , and can only make malaria control more difficult . Genome-level studies on P . knowlesi isolates from Mf and Mn across the parasite’s geographic range are now needed to test the generalizability of this remarkable conclusion .
Raw sequence data were downloaded for 48 isolates from Kapit and Betong in Malaysian Borneo [11] , 6 isolates from Sairikei in Malaysian Borneo ( S1 Fig ) [9] and 6 long-time isolated lines , maintained in rhesus monkeys sourced originally from Peninsular Malaysia and Philippines [11] . The sequence data accession numbers can be found in S1 Table . The samples were aligned against the new reference for the human-adapted line A1-H . 1 ( pathogenseq . lshtm . ac . uk/knowlesi_1 , accession number ERZ389239 , [22] ) using bwa-mem [43] and SNPs were called using the Samtools suite [44] , and filtered for high quality SNPs using previously described methods [45 , 46] . In particular , the SNP calling pipeline generated a total of 2 , 020 , 452 SNP positions , which were reduced to 1 , 632 , 024 high quality SNPs after removing those in non-unique regions , and in low quality and coverage positions . Samples were individually assessed for detecting multiplicity of infection ( MOI ) using: ( i ) estMOI [47] software , and ( ii ) quantifying the number of positions with mixed genotypes ( if more than one allele at a specific position have been found in at least 20% of the reads [46] ) . The measures led to correlated results ( r2 = 0 . 8 ) , which highlighted the robustness of these two methods . Samples were classified into three subcategories: ( i ) single infections ( > = 98% genome showing no evidence of MOI and < = 1/10 , 000 SNP positions with mixed genotypes ) , ( ii ) low MOI ( >85% genome showing no evidence of MOI and < = 4/10 , 000 SNPs positions with mixed genotypes ) ; ( iii ) high MOI ( <85% genome showing no evidence of MOI , and > 4/10 , 000 SNPs positions with mixed genotypes ) . Samples with high MOI were removed from subsequent analyses . For comparisons between populations , we first applied the principal component analysis ( PCA ) and neighbourhood joining tree clustering based on a matrix of pairwise identity by state values calculated from the SNPs . We used the ranked FST statistics to identify the informative polymorphism driving the clustering observed in the PCA [48] . Finally , we created haplotype plots using only SNP positions with MAF > 0 . 05 over all the populations , and displayed each sample as a row to allow closer inspection of the chromosome regions where interesting recombination events are observed . The XP-EHH metric [49] implemented within the rehh R package was used to assess evidence of recent relative positive selection between regional clusters from Kapit and Betong . The results were smoothed by calculating means in 1 Kbp windows , where windows overlapped by 250bp . The raXML software ( v . 8 . 0 . 3 , 1000 bootstrap samples ) was used to construct robust phylogenetic trees ( 90% bootstrap values > 95 ) for nuclear and organellar SNPs . Estimates of divergence times for subpopulations was based on a Bayesian Markov Chain Monte Carlo ( MCMC ) ( BEAST , v . 1 . 8 . 1 ) approach applied to mitochondrial sequences , with identical parameters settings to those described elsewhere [8] . The Shimodaira-Hasegawa [50] and the Templeton [51] tests were used to detect incongruence between the tree topologies . In order to identify regions that have undergone introgression we calculated the pairwise SNP diversity ( π ) of each sample against all the Borneo samples using a 50 Kbp sliding window . This window size was sufficient to include the required number of SNPs for the robust identification of introgression events . The average π in the M . nemestrina associated ( Mn-Pk ) and M . fascicularis associated ( Mf-Pk ) clusters was calculated , leading to two diversity values for each sample ( Mfπ and Mnπ ) and thereby a measure of genetic distance to the average of the two clusters . For Mf samples , an increase in the Mfπ and a decrease in Mnπ would mean the sample is more similar to the Mn-Pk cluster than the average; vice versa for the Mf samples . In order to avoid the identification of spurious events , we applied a threshold of a 0 . 001 increase in the deviation from the original cluster . For P . knowlesi genes of interest , orthologues in P . falciparum and P . berghei genomes were identified using PlasmoDB ( plasmodb . org ) . Gene expression data ( including from the RNAseq platform ) for these genes across different stages of the life cycle of the parasite were considered [26 , 27] . In particular , we compared the average of the asexual blood stages and the sexual ookinete stage , highlighting the genes upregulated with a two-fold change ( P<0 . 000001 ) , for P . falciparum [26] and P . berghei [27] . | Plasmodium knowlesi , a common malaria parasite of long-tailed and pig-tailed macaques , is now recognized as a significant cause of human malaria , accounting for up to 70% of malaria cases in certain areas in Southeast Asia including Malaysian Borneo . Rapid human population growth , deforestation and encroachment on wild macaque habitats potentially increase contact with humans and drive up the prevalence of human Plasmodium knowlesi infections . Appropriate molecular tools and sampling are needed to assist surveillance by malaria control programmes , and to understand the genetics underpinning Plasmodium knowlesi transmission and switching of hosts from macaques to humans . We report a comprehensive analysis of the largest assembled set of Plasmodium knowlesi genome sequences from Malaysia . It reveals genetic regions that have been recently exchanged between long-tailed and pig-tailed macaques , which contain genes with signals indicative of rapid contemporary ecological change , including deforestation . Additional analyses partition Plasmodium knowlesi infections in Borneo into 3 deeply branched lineages of ancient origin , which founded the two divergent populations associated with long-tailed and pig-tailed macaques and a third , highly diverse population , on the Peninsular mainland . Overall , the complex Plasmodium parasite evolution observed and likelihood of further host transitions are potential challenges to malaria control in Malaysia . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"taxonomy",
"parasite",
"groups",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"plasmodium",
"dna-binding",
"proteins",
"vertebrates",
"animals",
"mammals",
"parasitology",
"parasitic",
"protozoans",
"genetic",
"mapping",
"primate... | 2017 | Analysis of nuclear and organellar genomes of Plasmodium knowlesi in humans reveals ancient population structure and recent recombination among host-specific subpopulations |
Pathogens switching to new hosts can result in the emergence of new infectious diseases , and determining which species are likely to be sources of such host shifts is essential to understanding disease threats to both humans and wildlife . However , the factors that determine whether a pathogen can infect a novel host are poorly understood . We have examined the ability of three host-specific RNA-viruses ( Drosophila sigma viruses from the family Rhabdoviridae ) to persist and replicate in 51 different species of Drosophilidae . Using a novel analytical approach we found that the host phylogeny could explain most of the variation in viral replication and persistence between different host species . This effect is partly driven by viruses reaching a higher titre in those novel hosts most closely related to the original host . However , there is also a strong effect of host phylogeny that is independent of the distance from the original host , with viral titres being similar in groups of related hosts . Most of this effect could be explained by variation in general susceptibility to all three sigma viruses , as there is a strong phylogenetic correlation in the titres of the three viruses . These results suggest that the source of new emerging diseases may often be predictable from the host phylogeny , but that the effect may be more complex than simply causing most host shifts to occur between closely related hosts .
A major source of emerging infectious diseases are host shifts , where the parasite originates from a different host species . In humans , HIV [1] , influenza [2] and Plasmodium [3] have all been recently acquired from other species . Host shifts can also have devastating effects on wildlife; for example Ebola epidemics have resulted in marked declines in some primate populations [4] and canine distemper virus has jumped from dogs into Serengeti lions and caused considerable mortality [5] . As we have come to realise that the sources of human , domestic animal or crop pathogens are likely to be from wild species [6] , [7] , understanding what causes these parasite host shifts to occur has become increasingly important . For a host shift to occur , the new host must first be exposed to the parasite , the parasite must then be able to replicate in the new host , and finally there must be sufficient onward transmission in the new host for the infection to spread in the population [6] . Exposure is clearly important in determining whether a host shift occurs , and some cases of disease emergence have followed changes in the geographic range of species that have brought parasites in contact with new hosts [5] , [8] , [9] , [10] . However , once exposure has occurred , the factors that determine whether the pathogen can replicate in a new host are poorly understood . One factor that can potentially affect whether a parasite can replicate in a new host species is host relatedness — parasites may be more likely to replicate in species closely related to the original host [11] , [12] , because closely related hosts will tend to present a more similar environment to the parasite . Parasites must evade an elaborate array of host defences and rely on the host for their physiological needs , and this will result in specialised adaptations [13] , [14] . These adaptations have in turn resulted in some extremely specialised parasites that are only able to survive in a narrow range of similar host species [15] . If this is the case , host shifts may occur most frequently between closely related species . Here we use a new analytical approach to analyse host shifts , which allows us to separate two different ways in which the host phylogeny might affect the ability of a parasite to infect a new host species . The first of these , which we term the ‘distance effect’ , reflects the fact that the chances of successful infection may be higher in species that are more closely related to the natural host . However , it is also likely that related species share similar levels of susceptibility independently of how related they are to the natural host , a process that we term the ‘phylogenetic effect’ . These are statistically and biologically distinct phenomena . The distance effect will result in the expected susceptibility of new hosts declining as they become less related to the natural host . In contrast , the phylogenetic effect will have no effect on the expected susceptibility with distance from the natural host . However , it will result in distantly related species often having very different levels of susceptibility from the natural host , as it results in the variance in susceptibility increasing among more distantly related species . The two effects may generate very different patterns of host switching . The distance effect would result in most host shifts infecting species closely related to the natural host . In contrast , the phylogenetic effect might mean that host clades distantly related to the natural host are susceptible to a parasite , and this could cause parasites to jump between distantly related species . Previous research has examined the distance effect only . While there is evidence that parasites most often shift between related hosts from correlative studies of parasite-incidence in wild animals ( e . g . [16] ) , experimental evidence has been surprisingly rare . Cross-infection experiments using plants and fungi [17] , [18] , Drosophila and nematode worms [19] , and beetles and Spiroplasma bacteria [20] have all found that the ability of a parasite to establish an infection declines as a novel host's relatedness to the natural host declines . The extent to which host relatedness influences host switching varies between different groups of parasites , and it has been suggested that RNA viruses may be particularly prone to jump between distantly related hosts [21] . Reviewing emerging viral diseases in vertebrates , Parrish et al [21] observed that “Spillover or epidemic infections have occurred between hosts that are closely or distantly related , and no rule appears to predict the susceptibility of a new host . ” Viruses are more likely than other groups of parasites to be shared between distantly related primates [16] , and many human diseases that have been recently acquired from other species are RNA viruses [22] . The ability of certain viruses to infect distantly related hosts may result from the use of conserved host receptors to enter cells [23] , [24] , or the existence of hosts that do not posses broad resistance mechanisms to that type of parasite [25] , [26] . However , some studies have found evidence for the importance of the host phylogeny; rabies virus strains have higher rates of cross species transmission between closely related host species in the wild [27] and primate lentivirus phylogenies show signs of preferential switching between closely related hosts [28] . To explore this question we have conducted a large cross-infection experiment in which three sigma viruses were injected into 51 different species of Drosophilidae . Sigma viruses are a clade of rhabdoviruses ( RNA viruses with single-stranded negative-sense genomes ) , which infect various species of Diptera [29] , [30] . They are normally vertically transmitted [31] , [32] , leading to extreme specialisation on just a single host species . However , the sigma virus of Drosophila melanogaster ( DMelSV ) will replicate in a range of different dipteran hosts [33] , and differences between the host and virus phylogenies show that sigma viruses have switched between distantly related host lineages during their evolution [30] . Here we find that the host phylogeny explains most of the variation in the ability of sigma viruses to replicate in novel hosts , with both the distance and phylogenetic effects being large . These results not only allow us to explore the different ways in which the host phylogeny may affect host switching , but they are also , to our knowledge , the first study to experimentally test the effect of host genetic distance on infection success in RNA viruses — the most important source of emerging diseases .
We extracted DAffSV , DMelSV and DObsSV from infected stocks of D . affinis , D . melanogaster and D . obscura . To clear these stocks of any bacterial or other viral infections , they were aged for at least 20 days , before collecting embryos [32] and de-chorionating them in ∼2 . 5% w/v sodium hypochlorite solution for one minute [31] . The embryos were then rinsed in distilled water and placed onto clean food . To collect flies infected with a sigma virus , the adults were exposed to 100% CO2 at 12°C for 15 mins and the paralysed individuals were retained [31] , [32] , [34] . These were frozen at −80°C to rupture cells , homogenised in Ringer's solution [35] ( 2 . 5 µl/fly ) , and then briefly centrifuged twice , each time retaining the supernatant . This was passed through Millex PVDF 0 . 45 µM and 0 . 22 µM syringe filters ( Millipore , Billerica , MA , USA ) to remove any remaining host cells or bacteria , before being stored in aliquots at −80°C . Stocks of each fly species were kept in half pint bottles of staggered ages , and each day freshly eclosed flies were sexed , males were removed , and females were aged at 18°C for 3 days on agar medium ( recipe in Text S1 ) before injection . At the same time we stored remaining flies in ethanol for wing size measurements . The food medium , rearing temperature and whether each species was composed of single or multiple lines can be found in Table S1 . Female flies were injected with 69 nl of the virus extract intra-abdominally using a Nanoject II micro-injector ( Drummond scientific , Bromall , PA , USA ) . Half the flies were frozen immediately in liquid nitrogen as a reference sample to control for relative dose size , and the rest were kept on agar medium at 18°C for 15 days to allow the virus to replicate before being frozen in liquid nitrogen . The day 15 time-point was chosen based on pilot time-course data , and we note that the change in viral titre includes a decline in the virus following injection , followed by a growth/replication phase ( Figure S1 ) . Frozen flies were then homogenised in Trizol reagent ( Invitrogen Corp , San Diego , CA , USA ) . Based on quantitative reverse-transcription PCR ( qRT-PCR ) , the dose of the three viruses was similar ( with a maximum of a 1 . 6x difference between viruses ) . The injections were carried out over a period of 18 days , with the aim of completing 3 biological replicates for each virus per fly species ( 3 replicates each of the day 0 and day 15 treatments ) . The virus ( DAffSV , DMelSV or DObsSV ) was rotated on a daily basis , whilst treatment ( frozen immediately or on day 15 ) and the injection order of fly species were randomised each day . On average we injected and quantified viral titre in a pool of 10 flies per replicate ( range of across species means = 5–15 ) . Out of the 153 fly-virus combinations , 126 had 3 biological replicates , 24 had 2 biological replicates and 3 had 1 biological replicate . Wolbachia endosymbionts have recently been shown to provide resistance to a range of positive sense RNA viruses [36] , [37] , [38] , [39] . Although it does not affect the replication of DMelSV ( L . Wilfert and M . Magwire , unpublished data ) , we nonetheless tested each species for Wolbachia using PCR primers that amplify the wsp gene [40] . We also checked that the body size of the different species did not affect our results . To do this , we measured wing length , which is commonly used as a body size measure in Drosophila and strongly correlates with thorax length [41] , [42] . Wings were removed from ethanol-stored flies , photographed under a dissecting microscope and the length of the IV longitudinal vein from the tip of the proximal segment to where the distal segment joins vein V [43] was measured ( relative to a standard measurement ) using ImageJ software ( v1 . 43u ) [44] . Viral titres were estimated using qRT-PCR . To ensure that we only amplified viral genomic RNA and not messenger RNA , the PCR primers were designed to amplify a region spanning two different genes . The copy-number of viral genomic RNA was expressed relative to the endogenous control housekeeping gene RpL32 ( Rp49 ) . We designed different RpL32 primers specific for each species . First , we sequenced the RpL32 gene from all of the species ( we were not able to amplify RpL32 from Drosophila busckii , see Table S2 ) . We then designed species-specific primers in two conserved regions ( Table S3 ) . Total RNA was extracted from our samples using Trizol reagent , reverse-transcribed with Promega GoScript reverse transcriptase ( Promega Corp , Madison , WI , USA ) and random hexamer primers , and then diluted 1∶4 with DEPC treated water . The qRT-PCR was performed on an Applied Biosystems StepOnePlus system using a Power SYBR Green PCR Master-Mix ( Applied Biosystems , CA , USA ) and 40 PCR cycles ( 95°C for 15 sec followed by 60°C for 1 min ) . Two qRT-PCR reactions ( technical replicates ) were carried out per sample with both the viral and endogenous control primers . Each qRT-PCR plate contained a standard sample , and all experimental samples were split across plates in a randomised block design . A linear model was used to correct for the effect of plate . We repeated any samples where the two technical replicates had cycle threshold ( Ct ) values more than 1 . 5 cycles apart after the plate correction . To estimate the change in viral titre , we first calculated ΔCt as the difference between the cycle thresholds of the sigma virus qRT-PCR and the endogenous control . The viral titre of day 15 flies relative to day 0 flies was then calculated as 2−ΔΔCt , where ΔΔCt = ΔCtday0 – ΔCtday15 , where ΔCtday0 and ΔCtday15 are a pair of ΔCt values from a day 0 biological replicate and a day 15 biological replicate for a particular species-virus combination . We used a dilution series to calculate the PCR efficiency of the three sets of viral primers and thirteen of the RpL32 primer combinations ( covering 40 of the 51 Drosophila species ) . The efficiencies of the three virus primers were 95% , 97% , and 100% , ( DAffSV , DMelSV and DObsSV ) and the average efficiency of RpL32 primers across species was 106% , with all being within a range of 98–112% . The host phylogeny was inferred using the COI , COII , 28S rDNA , Adh , SOD , Amyrel and RpL32 genes . We downloaded all the available sequences from Genbank , and attempted to sequence COI , COII , 28S rDNA , Adh and Amyrel in those species from which they were missing ( details in Table S4 ) . This resulted in sequence for all species for COI , COII and 28S and partial coverage for the other genes ( 50 out of 357 species-locus combinations were missing from the data matrix ) . The sequences of each gene were aligned using ClustalW ( alignments and accession numbers are Datasets S1-S8 in supporting information ) . To reconstruct the phylogeny we used BEAST [45] , as this allows construction of an ultrametric ( time-based ) tree using a relaxed molecular clock model . The genes were partitioned into 3 groups each with their own substitution and molecular clock models . The three partitions were: mitochondrial ( COI , COII ) ; ribosomal ( 28S ) ; and nuclear ( Adh , SOD , Amyrel , RpL32 ) . Each of the partitions used a HKY substitution model ( which allows transitions and transversions to occur at different rates ) with a gamma distribution of rate variation with 4 categories and estimated base frequencies . Additionally the mitochondrial and nuclear data sets were partitioned into codon positions 1+2 and 3 , with unlinked substitution rates and base frequencies across codon positions . Empirical studies suggest that HKY models with codon partitions are a good fit for most protein coding data sets [46] . A random starting tree was used , with a relaxed uncorrelated lognormal molecular clock and we used no external temporal information , so all dates are relative to the root age . The tree-shape prior was set to a speciation-extinction ( birth-death ) process . The BEAST analysis was run for 100 million MCMC generations sampled every 1000 steps ( additionally a second run was carried out to ensure convergence ) . The MCMC process was examined using the program Tracer ( v1 . 4 ) [47] to ensure convergence and adequate sampling . Trees were visualised using FigTree ( v . 1 . 3 ) [48] . We used a phylogenetic mixed model to examine the effects of host relatedness on viral persistence and replication in a new host [49] , [50] , [51] . This framework allows ( random ) phylogenetic effects to be included in the model , with the correlation in phylogenetic effects between two host species being inversely proportional to the time since those two host species shared a common ancestor ( following a Brownian model of evolution ) . In general , conclusions drawn from phylogenetic comparative methods that include a species term in the model seem to be robust to alternative ( non-Brownian ) evolutionary models [52] . We fitted the model using a Bayesian approach in the R package MCMCglmm [53 , R Foundation for Statistical Computing , Vienna , Austria] and REML in ASReml [54] . The two methods gave similar results so we only report the Bayesian analysis ( Figure S2 ) . The model had the form:where yvhi is the viral titre of the ith biological replicate of host species h infected with virus v . is the intercept term for virus v , and can be interpreted as the viral replication rate in the species at the root of the phylogeny . dvh is the phylogenetic ( patristic ) distance between the original host of virus v and species h , and the associated regression coefficient ( ) determines the degree to which viral replication rate of virus v changes as the phylogenetic distance increases . The random effect up:vh is the deviation from the expected viral replication rate for virus v in host h due to historical processes ( i . e . the host phylogeny ) . The species random effect us:vh is the deviation from the expected viral replication rate of virus v in host h that is not accounted for by the host phylogeny . The residual is evhi , which included within-species genetic effects , individual and micro-environment effects and measurement/experimental error . The random effects ( including the residual ) are assumed to come from multivariate normal distributions with zero mean vectors ( because they are deviations ) and structured covariance matrices . Denoting as a vector of phylogenetic effects across species for virus v , and A as a matrix with elements ajk representing the proportion of time that species j and k have had shared ancestry since the root of the phylogeny:where is the variance of phylogenetic effects for DAffSV , and is the covariance between phylogenetic effects for DAffSV and DMelSV . Similar distributions are assumed for species effects:where I is an identity matrix indicating that species effects are independent of each other . The posterior modes for were close to zero for viruses DAffSV and DObsSV and these were omitted from the model ( except for the calculation of σ2p/ ( σ2p+ σ2s ) , see below ) . The residuals are distributed as:The off-diagonal elements of Ve ( i . e . the covariances ) were set to zero since viruses were not replicated within biological replicates . In a Bayesian analysis prior probability distributions have to be specified for the fixed effects and the covariance matrices . As described in detail in the supporting materials ( Text S1 ) we used several different priors to check if the results are sensitive to the choice of prior . The results presented were obtained using parameter expanded priors for the Vp and Vs matrices [53] . The P-values reported ( PMCMC ) correspond to 2pmin , where pmin is the smaller of the two quantities a ) the proportion of iterations in which the posterior distribution is positive or b ) the proportion of iterations in which the posterior distribution is negative . The 95% credible intervals ( CI ) were taken to be the 95% highest posterior density intervals . Marginal means of the posterior distribution are used as summaries of central tendency . Significance of the fixed effects was inferred if the 95% CI of the posterior distribution did not cross zero , and the P-values were equal to or less than 0 . 05 . We also checked whether several additional factors affected viral replication by repeating the analysis with these factors included in the model as fixed effects . There was no significant effect of wing size ( an average of 33 measured per species , PMCMC = 0 . 50 ) , the presence of the bacterial endosymbiont Wolbachia ( Table S4 , PMCMC = 0 . 51 ) or rearing temperature ( PMCMC = 0 . 55 ) . We also repeated the analysis with three outliers removed , so that the distribution of the residuals was not significantly different from normal according to an Anderson-Darling test ( A = 0 . 61 , P = 0 . 11 ) . The parameter estimates were very similar to those obtained when including all the taxa ( as reported in the results ) .
We measured the change in viral titre over 15 days for three sigma viruses each injected into 51 species of Drosophila , including their natural hosts ( see Figure 1 ) . In total we injected and quantified viral titre in 887 biological replicates ( a total of 8762 flies ) . To investigate how the host phylogeny affects the ability of the virus to persist and replicate in the different species , we reconstructed the phylogeny of all 51 species using the sequences of seven different genes . The resulting tree broadly corresponds to previous studies [55] , [56] , with the close phylogenetic relationships being generally well supported and more ancient nodes were less well supported ( Figure 1 ) . There are two ways in which the host phylogeny could affect the ability of the three viruses to infect new host species . First , the chances of successful infection may be higher in species that are more closely related to the natural host ( the ‘distance effect’ ) . Second , related species may share similar levels of susceptibility independently of how related they are to the natural host — an effect that we refer to as the ‘phylogenetic effect’ . To separate these two processes we fitted a phylogenetic mixed model to our data . All three viruses have greater viral titres in fly species that are more closely related to their natural host ( Figure 2 ) . If we assume that titres of all three viruses decline with genetic distance from their natural host at the same rate , then there is a significant negative relationship between titre and distance ( slope: γ = −1 . 96; 95% CI = −3 . 66 , −0 . 43; PMCMC = 0 . 022 ) . If we instead allow the effect to differ between viruses , the negative effect of genetic distance from the natural host on replication is greatest for DObsSV ( Figure 2; slope: γO = −4 . 03; 95% CI = −6 . 11 , −0 . 94; PMCMC = 0 . 005 ) , is smaller and only marginally non-significant for DAffSV ( Figure 2; slope: γA = −1 . 82; 95% CI = −3 . 99 , 0 . 37; PMCMC = 0 . 095 ) , and not significant for DMelSV ( Figure 2; slope: γM = −0 . 47; 95% CI = −3 . 06 , 1 . 94; PMCMC = 0 . 692 ) . These effects were still present when the natural host species were removed from the analysis ( data not shown ) . Therefore , the rate at which viral titres decline with genetic distance of the new host from the natural host differs between the individual viruses . There is also a strong influence of host phylogeny on viral replication that could not be explained by the distance of the novel host from the original host . The between-species variance consists of two components; σ2p , which is the variance that can be explained by the host phylogeny , and a species-specific component σ2s which cannot be explained by a Brownian-motion model of evolution on the host phylogeny . These statistics do not include the effects of the distance from the natural host , as this was included as a fixed effect in the model [57] . To assess the importance of the host phylogeny , we calculated the proportion of the between-species variance that can be explained by the phylogeny ( σ2p/ ( σ2p+ σ2s ) , which is similar to Pagel's λ [58] , [59] or phylogenetic heritability [50] , [51] ) . The phylogeny explained almost all of the between-species variance in viral titre for DAffSV and DMelSV ( σ2p/ ( σ2p+ σ2s ) = 0 . 86 , 95% CI = 0 . 53–1 and 0 . 91 , 95% CI = 0 . 74–1 , respectively ) , and most of the between-species variation for DObsSV ( σ2p/ ( σ2p+ σ2s ) = 0 . 72 , 95% CI = 0 . 43–0 . 98 ) . Therefore , most of the differences between species in viral titres can be explained either by the host phylogeny or the distance from the natural host . Is it the distance from the natural host , or host phylogeny per se , that is most important in determining viral replication and persistence in a new host ? To allow a direct comparison of these two effects , we calculated the expected amount of change in viral titre from the root to the tips of the tree that will result from the phylogenetic effect . This was done by taking the product of the standard deviation of the phylogenetic effect and , which is the mean of a folded zero-centred normal distribution , and is the predicted change under a Brownian model . This gave values of 2 . 15 , 3 . 28 and 2 . 69 viral-titre-units for DAffSV , DMelSV and DObsSV respectively . These can be compared directly to the estimates described above of the amount of change in viral titre as the genetic distance from the natural host increases ( −1 . 82 , −0 . 47 and −3 . 70 viral-titre-units for DAffSV , DMelSV and DObsSV respectively ) . The time from the root to tip of the phylogeny has been estimated as ∼40 million years [60] , so for every ∼40 million years travelled along the phylogeny , or from the natural host , we expect to see the above changes in viral titre . From these estimates it is clear that over this timescale the two processes are of similar importance for DAffSV and DObsSV , but that the host-phylogeny is more important than distance-from-the-original-host in determining the replication and persistence of DMelSV in a new host . Differences between hosts in viral replication and persistence could either reflect differences in susceptibility to all three viruses ( ‘general susceptibility’ ) , or the effects on the three viruses could be independent ( ‘specific susceptibility’ ) . We found that most of the phylogenetic effect was caused by species differing in their level of general susceptibility , as there were strong phylogenetic correlations between viruses ( Table 1 ) . Furthermore , the correlation is not greater between the two viruses that naturally infect closely related hosts ( DAffSV and DObsSV ) . Therefore , the phylogenetic effects mean that a given host species' susceptibility to one virus is strongly correlated to its susceptibility to another sigma virus , regardless of whether the virus originated from a closely or distantly related host . The analysis above assumes that we have the correct phylogeny , but some of the relationships are poorly resolved ( Figure 1 ) . To check whether this affected our results , we repeated the analysis integrating over the posterior sample of trees generated during the phylogenetic analysis [61] . This was achieved by fitting the phylogenetic mixed model to 2000 different trees from the posterior sample ( from 100 , 000 trees we used a burn-in of 30 , 000 trees and then used every 35th tree ) . This gave very similar results to our main analysis , suggesting that phylogenetic uncertainty does not affect our conclusions . We would note however , that σ2p is biased downwards whenever the tree is incorrect , and this bias is not removed by this procedure .
We found that the ability of three sigma viruses to persist and replicate in 51 different species of Drosophila is largely explained by the host phylogeny . The effect of phylogeny can be broken down into two components; not only did viral titres tend to decline with increasing genetic distance from the natural host , but there is also a tendency for related hosts to have similar titres , independent of the distance effect . The decline in viral titres with increasing distance from the natural host suggests that the greater the change in the cellular environment , the less well adapted the virus is . This might be caused by changes in the cellular machinery used by the virus in its replication cycle , or the virus being less adept at avoiding or suppressing the immune response . Regardless of the causes of this effect , it suggests that successful host shifts may be more likely between closely related hosts [6] . A host shift requires the new host to be exposed to the pathogen , the virus to replicate sufficiently for an individual to become infected , and finally for there to be sufficient onward transmission for the infection to become established in the population . Our data suggests that the second step is most likely to occur between closely related hosts . It is possible that higher titres may also lead to greater onward transmission , as the titre of DMelSV in D . melanogaster correlates with the rate at which the virus is transmitted [31] , [62] . Furthermore , it has also been reported that although DMelSV will replicate in a range of Drosophila , but it was stably transmitted only in the closely related Drosophila simulans and not the more distantly related Drosophila funebris [63] . However , viral titres should only be used with caution as a proxy for transmission rates , as many other factors may affect transmission rates , including trade-offs between replication and virulence [64] . There is tentative evidence that host shifts of sigma viruses occur most often between closely related species in natural populations . Although comparisons of Drosophila and sigma virus phylogenies show evidence of past host shifts , the host and virus phylogenies are more similar than expected by chance [30] . This may be the result of more frequent host switches between closely related species , as would be predicted by our results ( although cospeciation would produce the same pattern and more data is required to confirm these findings ) . This result is interesting because it has previously been questioned whether the genetic distance between host species plays an important role in predicting the source of host shifts , especially for RNA viruses [6] , [22] . Indeed , some plant viruses can replicate in an enormous range of species; Cucumber mosaic virus can infect 1 , 300 plant species in over 100 families and Tomato spotted wilt virus can infect 800 plant species in 80 families [65] . The use of conserved receptors to enter host cells may be key to large potential host ranges in animals [23] , [24] , [66] . However , although a virus may be able to enter the cells of many different species , it then relies on numerous different components of the cellular machinery to replicate effectively , and this may make shifts to hosts that are distant from the natural host unlikely . A factor that could lead to changes in host suitability across the phylogeny is selection for resistance to viruses . One reason to suspect that this may be important is that genes involved in antiviral immunity often evolve exceptionally rapidly in Drosophila [67] , [68] , [69] , [70] , and this may translate into rapid phenotypic changes in host susceptibility . If this process is driving the patterns that we see , then the observation that natural host-parasite combinations tend to be more susceptible would suggest that the viruses have been able to overcome these host defences , resulting in viruses that are well adapted to their natural hosts , rather than vice versa . After accounting for the effect of distance from the natural host , the host phylogeny still explains most of the remaining variation in viral titre between species . This ‘phylogenetic effect’ means that that closely related host species have similar levels of resistance due to their non-independence as a result of common ancestry . Indeed , the most distantly related clade to all of the natural hosts examined ( the Scaptodrosophila ) have one of the highest viral titres ( Figure 1 ) . For two of the viruses ( DAffSV and DObsSV ) , we found that this phylogenetic effect was of comparable importance to the effect of genetic distance from the natural host , and for the third virus ( DMelSV ) it was more important . The phylogenetic effect and distance effects are statistically ( and biologically ) distinct phenomena . If we imagine two sister species ( A and B ) and an out-group ( C ) are infected with a virus originally from species A , there are two ways in which the host phylogeny could affect the ability of the viruses to infect the three species . Under a Brownian motion model of evolution we expect viral titre in species A to be more different to that in C than B . Importantly , however , we do not expect this difference to have a consistent sign , as it is only the magnitude of the difference that should be larger for species C . A second process is that as we move away from species A we may expect a systematic change in viral titre – either that the viral titre increases as we move to species B and then to species C , or alternatively a systematic decrease . We call this first effect – where the change does not have a predictable sign – a phylogenetic effect , and the second effect - where change does have a predictable sign – a distance effect . The phylogenetic and distance effects may also generate distinct patterns of host switching ( see Introduction ) . For example , our data regarding the phylogenetic effect imply that sigma viruses may more easily switch between infecting flies in the subgenus Sophophora and the distantly–related , but highly susceptible , Scaptodrosophila . However , the two distinct patterns may emerge from the same underlying evolutionary process . If related hosts have similar levels of susceptibility ( i . e . the phylogenetic effect ) , and pathogens can only become established in the most susceptible hosts , then we would expect to see a decline in viral titre in species distantly related to the natural hosts ( i . e . the distance effect ) . The phylogenetic effect is mostly caused by variation in susceptibility to all three viruses ( there is a strong phylogenetic correlation in the titres of the three viruses ) . Such patterns may arise if the common ancestors of different host clades have acquired or lost immune or cellular components that affect susceptibility to all sigma viruses . The frequent gain and loss of immune components is well-established , for example , Drosophila species in the obscura group have lost a type of blood cell ( lamellocytes ) that are found in other Drosophila , which means they are particularly susceptible to parasitoid wasps [26] . Similarly a class of antifungal peptides ( drosomycins ) are found only in the melanogaster group of Drosophila [71] , [72] and components of antiviral RNAi pathways have lineage-specific distributions [73] , [74] . Part of the phylogenetic effect could be explained by the evolutionary history of the viruses , for example if they have recently switched between host species and are still well-adapted to a previous host . The strong phylogenetic correlation between the three viruses we studied might seem surprising as these viruses are very different to one another at the sequence level ( amino-acid identities are ∼20%–40% [29] , [30] ) . However , even viruses which show no similarities at the sequence level often share elements of protein structure [75] , [76] , [77] , and different rhabdoviruses are known to have similar modes of action ( for example , infecting nervous tissue [31] , [78] ) . The strong phylogenetic effect that we found also has practical implications for comparative studies of resistance in different species . It means that observations on related species will not be independent , so it is essential to account for these effects in the analysis of comparative data [79] . For example , the decline in the resistance of novel hosts with genetic distance from the natural hosts that has been observed in some previous studies may be attributable to a phylogenetic effect , rather than distance itself . In conclusion , our results show that the host phylogeny is an important determinant of viral persistence and replication in novel hosts , and therefore may also be an important influence on the source of new emerging diseases . The effect is more subtle than simply leading to a decline in infection success with genetic distance from the original host , because the strong phylogenetic effect may sometimes result in susceptible hosts being grouped in phylogenetically distant clades , allowing parasites to jump great phylogenetic distances . The importance of these phylogenetic effects on replication and persistence relative to factors affecting exposure and onward transmission requires further study if we are to understand how they affect a parasites ability to host shift in nature . | Emerging infectious diseases such as SARS , HIV and swine-origin influenza have all been recently acquired by humans from other species . Understanding the reasons why parasites jump between different host species is essential to allow us to predict future threats and understand the causes of disease emergence . Here we ask how host-relatedness might determine when host-shifts can occur in the most important group of emerging diseases—RNA viruses . We show that the relationship between host species is the primary factor in determining a virus's ability to persist and replicate in a novel host following exposure . This can be broken down into two components . Firstly , species closely related to the virus's natural host are more susceptible than distantly related species . Secondly , independent of the distance effect , groups of closely related host species have similar levels of susceptibility . This has important implications for our understanding of disease-emergence , and until now the only large-scale studies of viruses have been correlative rather than experimental . We also found groups of related species that are susceptible to these viruses but are distantly related to the natural hosts , which may explain why viruses sometimes jump between distantly related species . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"public",
"health",
"and",
"epidemiology",
"epidemiology",
"biology",
"evolutionary",
"biology"
] | 2011 | Host Phylogeny Determines Viral Persistence and Replication in Novel Hosts |
Dihydrodipicolinate synthase ( DHDPS ) catalyzes the first committed step in the diaminopimelate pathway of bacteria , yielding amino acids required for cell wall and protein biosyntheses . The essentiality of the enzyme to bacteria , coupled with its absence in humans , validates DHDPS as an antibacterial drug target . Conventional drug design efforts have thus far been unsuccessful in identifying potent DHDPS inhibitors . Here , we make use of contemporary molecular dynamics simulation and Markov state models to explore the interactions between DHDPS from the human pathogen Staphylococcus aureus and its cognate substrate , pyruvate . Our simulations recover the crystallographic DHDPS-pyruvate complex without a priori knowledge of the final bound structure . The highly conserved residue Arg140 was found to have a pivotal role in coordinating the entry of pyruvate into the active site from bulk solvent , consistent with previous kinetic reports , indicating an indirect role for the residue in DHDPS catalysis . A metastable binding intermediate characterized by multiple points of intermolecular interaction between pyruvate and key DHDPS residue Arg140 was found to be a highly conserved feature of the binding trajectory when comparing alternative binding pathways . By means of umbrella sampling we show that these binding intermediates are thermodynamically metastable , consistent with both the available experimental data and the substrate binding model presented in this study . Our results provide insight into an important enzyme-substrate interaction in atomistic detail that offers the potential to be exploited for the discovery of more effective DHDPS inhibitors and , in a broader sense , dynamic protein-drug interactions .
Dihydrodipicolinate synthase ( DHDPS ) catalyzes the first and committed step in the diaminopimelate ( DAP ) biosynthesis pathway of bacteria and plants , namely the condensation of pyruvate and ( S ) -aspartate-β-semialdehyde ( ASA ) to ( 4S ) -4-hydroxy-2 , 3 , 4 , 5-tetrahydro- ( 2S ) -dipicolinic acid ( HTPA ) [1–4] ( Fig 1A ) . ( 4S ) -4-hydroxy-2 , 3 , 4 , 5-tetrahydro- ( 2S ) -dipicolinic acid ( HTPA ) is then converted via a series of enzyme-catalyzed reactions to yield meso-diaminopimelate ( meso-DAP ) and ( S ) -lysine , which are important metabolites for cell wall and protein biosyntheses . Gene knock-out studies demonstrate that dapA , which encodes dihydrodipicolinate synthase ( DHDPS ) , is an essential gene [5–9] . Given its essentiality to bacteria and plants , but absence in humans , DHDPS has gained considerable traction as a promising target for both antimicrobial drugs and herbicides [3 , 10 , 11] . Despite sustained interest in DHDPS , potent inhibitors of the enzyme have not yet been realized [3 , 10 , 11] . The most effective inhibitors discovered to date are predominantly derived by analogy to HTPA , which nonetheless show poor ( low millimolar ) inhibitory potency . There is thus a need to consider other factors in the rational design of DHDPS inhibitors . One such factor is the dynamics of the DHDPS-substrate interaction . Herein , we explore this phenomenon with a view to providing insight into the design of more effective DHDPS inhibitors . Functionally , DHDPS exists as a homo-oligomer ( Fig 1B ) [12] . Two DHDPS monomers self-associate to form a ‘tight-dimer’ ( Fig 1B ) [12] , with the articulating surface between monomers capturing a significant proportion ( 10–13% ) of the total subunit surface area [12] . Dimerization of DHDPS is crucial to enzyme function since monomeric DHDPS mutants demonstrate attenuated enzyme activity and decreased substrate binding affinity , particularly for pyruvate , which is the first substrate to bind DHDPS [13–15] . Each DHDPS monomer adopts a classical TIM-barrel fold that encloses a 30 Å-long cavity where the enzyme active site is encapsulated [16 , 17] . Several key catalytic residues have been identified using X-ray crystallography and site-directed mutagenesis [16 , 18 , 19] . A catalytic triad of two tyrosines ( Tyr109 , Tyr135 , S . aureus numbering ) and a threonine ( Thr46 , S . aureus numbering ) function as a proton relay during catalysis ( Fig 1C ) [20] . Dimerization of DHDPS allows for Tyr109 to interdigitate across the dimer interface , completing the catalytic triad of the adjacent monomeric unit and concomitantly creating two equivalent active sites per DHDPS dimer . Another residue , Lys163 ( S . aureus numbering ) , forms a Schiff-base with pyruvate during catalysis [18] . These residues are virtually indispensable for enzyme function [20–22] . Crossing the lip of the active site cavity is the solvent-exposed residue Arg140 ( S . aureus numbering ) , which has been implicated in the role of stabilizing the catalytic triad and presumably binding of substrates , particularly ASA [16 , 23] . However the precise mechanism of Arg140’s role in DHDPS remains poorly understood . Molecular dynamics ( MD ) simulation has emerged as a useful tool for gaining insight into various biological phenomena , such as enzyme allostery [24] , protein dynamics [25] , and binding of small molecules to their cognate protein receptors [26–28] . Multiple independent simulations can be performed that follow the unbiased motion of ligands in and around the binding site , allowing for identification of various factors that contribute to the interaction , such as residue bonding networks and protein conformational change [26] . Large ensembles of trajectories from such simulations can be pooled and clustered into microstates based on criteria such as root-mean square deviation ( RMSD ) . The statistics of transitions between microstates across the trajectory ensemble can then be used to create a Markov state model ( MSM ) that captures the essential dynamics of the process [29 , 30] . These models can capture the kinetics of binding events and allow for thermodynamic quantities to be calculated . In this work , we present a detailed description of the binding dynamics of the substrate pyruvate and of the enzyme DHDPS . We use all-atom MD simulations to completely recapitulate the entire pyruvate binding process from bulk solvent to the crystallographic bound pose . By means of MSMs , we find that there are several key and metastable intermediates in this pathway defined as ‘hotspots’ . Addressing the results from this study , the long-term goal of this project is to design DHDPS inhibitors that incorporate the targeting of this binding intermediate .
The choice of starting structure is an important decision in any MD simulation study . We employed the structure of DHDPS from the bacterium S . aureus ( PDB ID 3DAQ ) [12] . The rationale for selecting S . aureus DHDPS as the subject of this study was made on the basis of two considerations . Firstly , it adopts a non-canonical dimeric assembly that is in contrast to tetrameric orthologs from the majority of other plant and bacterial species [12 , 31 , 32] . This provides a considerable cost-saving in computational time ( i . e . approximately two-fold reduction in the number of protein non-hydrogen atoms required ) . Secondly , DHDPS from S . aureus has been well-characterized structurally , providing both pyruvate-bound and apo structures as reference points [12 , 33] . The availability of these structures provides both start and end configurations of DHDPS against which we can make direct comparisons with our own computational results . We investigated pyruvate binding to DHDPS using all-atom MD simulation . The process of ligand binding to an enzyme active site must inherently be dynamic , requiring small-scale ( e . g . residue side chain motion ) and potentially large-scale receptor motions ( e . g . conformational dynamics ) . Sampling the entirety of these motions within a single extended trajectory is both impractical and computationally inefficient . Thus , to broaden our coverage of the conformational landscape of DHDPS we generated an ensemble of different DHDPS configurations by sampling from a 100 ns MD simulation . Each of these configurations were used as starting points for new rounds of minimization . For each of our systems we simulated DHDPS with two molecules of the substrate pyruvate . We placed each of these ligands at locations distal to the crystallographic binding site ( i . e . the DHDPS active site ) using a randomized placement protocol ( see Methods for details ) . This step was included to alleviate imposing a bias on the binding pathway . We first conducted pilot MD simulations of the system and observed evidence of pyruvate binding into the DHDPS active within 10–100 ns . Thus we extended the simulation data set by performing 80 MD simulations ranging between 10 ns and 100 ns for a total simulation time of 8 . 4 μs per molecule of pyruvate , yielding a mean simulation length of 52 . 4 ns ( ± 30 ns ) . Many of our simulations capture ligand binding events . We track binding by calculating the root-mean square deviation ( RMSD ) of pyruvate heavy atoms ( i . e non-hydrogen atoms ) against a reference structure of pyruvate in complex with DHDPS ( PDB ID 3DI1 ) [33] , correcting first for translational and rotational protein motions ( Fig 2A ) . In 5 of our simulations the RMSD converges upon a value of less than 2 . 3 Å ( Fig 2A ) . In some simulations this is achieved after as little as 12 ns of simulation time . Transitions from this converged position are rare ( i . e . the vast majority of simulations remain in this state for the remainder of the simulation ) . Some simulations show remarkable agreement with the reference structure , achieving conformations that deviate by as little as 1 . 85 Å . These findings suggest that our simulations are able to recover the crystallographic ‘bound pose’ with reasonable resolution in the absence of a priori knowledge of the binding trajectory . Furthermore , our MD simulations identify several pyruvate interaction ‘hotspots’ ( assessed by spatial residence of pyruvate; Fig 2B ) . Mapping out the regions of high pyruvate density ( ‘hotspots’ ) relative to DHDPS , we find that many of these appear spatially proximal to key active site residues such as Thr46 , Tyr109 , Tyr135 , Arg140 , and Lys163 ( Fig 2B ) . Similar clustering patterns are observed when comparing other 2-dimensional planes ( S1 Fig ) . For reference , we provide an example of a single , 75 ns-long binding trajectory in S1 Video . The majority of simulations , however , do not reach the bound pose even up to simulation time-scales of 75 ns ( Fig 2A ) . Furthermore , no complete dissociation events to solvent from the bound pose occur . Nonetheless , the simulation data available is sufficient for a comprehensive analysis of binding kinetics . We submitted our MD simulation data to more exhaustive analysis of pyruvate binding kinetics by building a MSM . The first step in this process involved assigning structurally-related pyruvate conformations into clusters . In brief , we found that partitioning our simulation data set ( over 5 . 5 million conformations ) into 738 clusters provided sufficient resolution for the resulting model , with the mean cluster radius calculated as 2 . 6 ± 0 . 4 Å . By measuring the time-dependence of relaxation time-scales in our model ( S2 Fig ) , we found that the Markov property was satisfied using lag times at and beyond 3 ns . Thus , we built an initial fine-grained MSM using a lag time of 4 ns . S3 Fig shows that the states comprising this fine-grained model are densely interconnected . Such fine-grained models , while a resource of quantitative information , suffer from a high degree of complexity that makes analysis difficult . With this in mind , and with a view to providing a more readily-interpretable representation of our model , we grouped states from our fine-grained model into a series of long-lived and metastable states ( S ) ( Fig 3 ) . We used a Bayesian approach [34] to assign kinetically-related states into a metastable MSM ( Fig 3 ) . The resulting metastable MSM identifies 17 states ( S0–S16 ) , each describing components of the pyruvate binding pathway . In Fig 3C we make use of a network diagram to show the most highly populated of these states ( i . e . those with equilibrium probabilities in excess of 4% ) . These states correspond to S2 , S6 , S7 , S9 , S10 , S12 , S13 , and S14 . Here , representatives from each macrostate are depicted structurally as nodes with interstate transition probabilities indicated by the shade of connecting arrows ( darker arrows indicate greater probability ) . The starting point for pyruvate binding in the coarse-grained model is depicted in Fig 3A and 3C as two states: S2 and S6 . These represent unbound states of pyruvate , given that ligand conformations within S2 and S6 are distal to the active site and lack any substantial points of interaction with DHDPS residues . We next evaluated the relevance of other metastable states in our model ( S0–S1 , S3–S5 , and S7–S17 ) to pyruvate binding by making comparisons with available experimental data , using structural similarity to the pyruvate-bound DHDPS crystal structure ( PDB ID 3DI1 ) . For this we used two measures: ( i ) the average RMSD of pyruvate heavy atoms relative to key active site residues , and ( ii ) intermolecular bonding patterns between protein and ligand . Under these conditions we find that S13 best captures the bound pose ( Fig 3B ) , since this state demonstrates the best agreement with the pyruvate-bound enzyme structure ( RMSD 3 . 69 ± 0 . 72 Å ) ( Fig 3B ) . Moreover , S13 includes the formation of a hydrogen bond network with key active site residues , namely Thr46 , Tyr109 , and Tyr135 , in a remarkably similar geometry to the X-ray structure [33] . This is exemplified in Fig 3B . From our metastable MSM we made use of transition path theory ( TPT ) [35 , 36] to investigate pathways of pyruvate binding . The output of such analyses is a collection of paths that describe the routes by which the system can unidirectionally traverse from unbound states , S2 and S6 ( Fig 3A and 3C ) , to the bound state S13 ( Fig 3B and 3C ) , which are then ranked according to reactive flux . In S4 Fig we show that the top six pathways capture almost three quarters of the total binding flux . We first examine the dominant binding pathway , comprising four transitions . We label each of these transitions sequentially from T1 to T4 ( Fig 4A ) . In the first transition ( T1 ) pyruvate partitions from bulk solvent ( S6 ) to an interacting surface of DHDPS ( S7 ) . Here , pyruvate is stabilized by the establishment of a salt bridge between the carboxylate moiety of pyruvate and the guanidinium group of the solvent-exposed Arg140 residue of DHDPS . In the second transition ( T2 ) , the side chain of the interacting Arg140 demonstrates a high degree of flexibility ( Fig 4A ) . Specifically , Arg140 flips away from the bulk solvent towards the active site cavity , carrying pyruvate deeper into the entryway of the active site ( S12 ) . The carboxylate-guanidinium salt bridge between pyruvate and Arg140 is sacrificed in the third transition step ( T3 ) as pyruvate migrates further into the active site cavity ( Fig 4A ) . The loss of this interaction with Arg140 in the penultimate step is balanced by the establishment of a new hydrogen bond network with several active site residues such as Thr46 , Tyr135 , and Lys163 , allowing the substrate to enter into a ‘pre-bound’ pose ( S9 ) . It is from this ‘pre-bound’ pose that pyruvate undergoes a twisting and flipping motion ( T4 ) to assume the pyruvate-DHDPS complex , or bound pose ( S13 ) [33] . Alternative pathways show noteworthy deviations from the dominant pathway shown in Fig 4A . Many of these alternative paths allow for pyruvate to transition to the bound state whilst boycotting the ‘pre-bound’ intermediate S9 , suggesting that this may not be an essential requirement for pyruvate binding . Likewise some of the highest flux pathways boycott the binding intermediate S12 . Other paths allow for greater flexibility at the entryway to the active site , instead relying on transient non-covalent interactions with other residues proximal to the active site ( e . g . Ile248 ) . Compellingly however , the six highest flux pathways have a strict requirement for the formation of the pyruvate-Arg140 salt bridge intermediate ( S7 ) , suggesting the S7-intermediate is a highly-favorable event towards pyruvate binding . Consistent with these findings , an Escherichia coli DHDPS mutant incorporating alanine at the equivalent position of Arg140 demonstrates ~1000-fold reduced enzyme activity and decreased affinity for substrate [23] . This may be expected , given the absolute conservation of Arg140 , and other key catalytic residues ( Thr46 , Tyr109 , Tyr135 , Lys163 ) in bacterial DHDPS sequences ( Fig 4B ) . Umbrella sampling was used to determine the relative free energies of metastable binding states using a simple one-dimensional coordinate along the axis of the pyruvate binding pathway ( depicted in S1 Fig ) . This coordinate was mostly sampled in windows of 1 Å , while windows centered on 8 , 9 , 14 and 19 Å experienced unacceptably long de-correlation times ( ≈1 ns ) . Thus , these windows required a finer sampling of 0 . 5 Å with a greater biasing force to reduce this the de-correlation time to no greater than 200 ps . Fig 5 shows that 5 ns of simulation time for each window was sufficient to converge a potential of mean force ( PMF ) curve along this coordinate . Error estimates , de-correlation times , and force constants for each window are included in S1 Table for reference , and histograms shown in S6 Fig . The PMF profile in Fig 5 shows that the local energy minimum corresponding to S10 has an associated free energy difference of ( -0 . 5 ± 0 . 2 kcal mol-1 ) , while a similar energy minimum is observed for the Z-coordinate equivalent to S7 ( -0 . 2 ± 0 . 2 kcal mol-1 ) which , as Fig 4 shows , is en route to the binding intermediate S12 ( -3 . 0 ± 0 . 2 kcal mol-1 ) . S12 presents as a relatively higher-energy intermediate between states S4 ( -8 . 6 ± 0 . 2 kcal mol-1 ) and S5/S9 ( -9 ± 0 . 2 kcal mol-1 ) , the latter of which forms the global minimum of the PMF profile . Surprisingly , the bound state S13 ( -3 . 0 ± 0 . 2 kcal mol-1 ) , which was bimodal with respect to the Z-coordinate ( Fig 5 , gray box ) , appears as an energetic intermediate between S5/S9 and the least energetically favorable state along the PMF profile , S0 ( 0 . 5 ± 0 . 2 kcal mol-1 ) . These umbrella sampling data support that the observed states in the model show in Fig 4 are indeed thermodynamically metastable . Large-scale changes to protein structure and dynamics upon ligand binding are not uncommon [38 , 39] . Thus , we considered whether similar large-scale dynamic motions were at play in the case of DHDPS pre- and post-binding pyruvate . Comparison of apo and pyruvate-bound crystal structures of several bacterial DHDPS enzymes reveals only minor structural differences ( . 12– . 19 Å; S2 Table ) . Thus , it was not anticipated that DHDPS would undergo substantial structural changes after transitioning from the apo to pyruvate-bound form . Indeed , there remained no substantial perturbation of DHDPS conformation ( S7A Fig ) or residue fluctuations ( S7B Fig ) up to 60 ns after transitioning into the bound form ( S13 ) . Previous MD simulations performed using the crystal structure of DHDPS ( PDB ID 3DAQ ) over longer time-scales indicate that while the enzyme is afforded a degree of inter-subunit flexibility ( so called “enzyme-breathing” ) , intra-subunit motions remain minimal [40] . Thus , our results are consistent with these earlier reports . We reason that these data provide compelling evidence that macroscopic , intra-subunit protein motions play a limited role in pyruvate binding . It remains to be seen whether this is the case for subsequent catalytic steps to product formation , such as binding of the second DHDPS substrate , ASA ( Fig 1A ) .
DHDPS is a promising antibacterial drug target . However , despite decades of conventional rational drug design and the determination of more than 80 DHDPS structures , potent inhibitors remain elusive [3 , 10 , 11] . We suggest that a major shortcoming of previous studies is a lack of dynamic information describing ligand binding to druggable sites of DHDPS . Thus , the primary goal of this study was to identify and characterize the dynamics of ligand binding to the potentially druggable active site of DHDPS . More specifically , we set out to map the binding pathway for the DHDPS substrate pyruvate in atomistic detail using MD simulation . Importantly , our simulations recover the bound pose observed in the crystallographic structure with remarkable precision , allowing for robust conclusions to be articulated concerning the binding trajectory . We find that ( i ) Arg140 side chain motion correlates with the recruitment of pyruvate from solvent in a role we define here as a ‘gateway’ residue; ( ii ) the majority of binding passes through a transient state involving a salt bridge between the carboxylate and guanidinium moieties of pyruvate and Arg140 , respectively; ( iii ) pyruvate binding , independent of catalysis , is a dynamic phenomenon with several distinct metastable states described here as ‘hotspots’; and ( iv ) pyruvate binding is an energetically favorable event with discrete thermodynamic intermediates . Firstly , with regards to Arg140’s role as a ‘gateway’ residue , it is interesting to compare our in silico findings with previous structural [16] and kinetic [23] reports in vitro . Structural studies by Mirwaldt et al . [16] put forward a potential role for Arg140 in substrate recruitment given the positioning this residue at the ‘gateway’ to the active site . Whereas mutation of the Arg140 equivalent in E . coli DHDPS was shown to markedly reduce catalytic activity ( i . e . decreased kcat ) and increase the apparent Michaelis constant ( KM ) for ASA by ~50-fold , only a subtle increase in the apparent KM was found for pyruvate [23] . However , what remains to be explained is the molecular mechanism that underpins the ~50-fold increase in the apparent KM for ASA . We propose a similar , but more pronounced , ‘gateway’ role for Arg140 during the subsequent step of the DHDPS-catalyzed reaction , when ASA is recruited ( Fig 1 ) . It would thus be of interest in future studies to explore the binding dynamics of DHDPS with ASA using similar in silico approaches to those reported here for DHDPS-pyruvate interactions . Secondly , to validate our second conclusion regarding the crucial transition intermediate involving the carboxylate of pyruvate and the guanidinium of Arg140 , future studies could explore the importance of this phenomenon by substituting the side chain and/or ligand moieties with uncharged equivalents . For example , MD simulations could be performed on a mutant S . aureus DHDPS structure incorporating alanine at position 140 , which would replace the positively-charged guanidinium moiety with a shorter-chain methyl group . Indeed , this mutation has already been explored kinetically in vitro[23] . Alternatively , the carboxylate group of pyruvate could be replaced by an electrostatically neutral aldehyde group . Either of these chemical transmutations would provide further insight into the chemical requirements of this transition . Thirdly , we recover similar binding energies for pyruvate in our in silico model compared to those experimentally determined in vitro . Moreover , we calculate that the relative energy difference between the apo- and pyruvate-bound state correlates to approximately -10 kcal mol-1 ( Fig 5 ) . This agrees well with isothermal titration calorimetry measurements [22] . Importantly , these data imply that intermediate states observed in our model are indeed thermodynamically metastable and are not likely attributable to short-lived artifacts . Defined thermodynamically-metastable ligand binding/unbinding intermediates appear to be broadly applicable to many protein-small molecule interactions [26–28 , 41] . For example , in a thematically related computational study Da et al [41] indicated that egress of inorganic pyrophosphate from the active site of yeast RNA polymerase II adheres to a four-state model . In this model Da et al [41] showed that transitions between these kinetically metastable ‘hotspots’ is principally mediated by mutable electrostatic interactions between protein lysine or histidine residues and the ligand pyrophosphate . Furthermore , in silico mutation of these residues was found to retard product release from the RNA polymerase II active site [41] . Relating this back to our current study , it would be a worthwhile endeavour for future studies to measure whether transmuting residues involved in the pyruvate binding transition states , such as Thr46 , Tyr109 , Tyr135 , Arg140 , Lys163 , or Ile248 , alter the kinetics of substrate entry to the DHDPS active site . In conclusion , this study sheds light on the binding pathways of an important enzyme-substrate interaction that has identified several metastable binding intermediates with distinct protein-ligand interaction profiles ( i . e . ‘hotspots’ ) . We suggest that rational drug design can be augmented for next-generation DHDPS-inhibitors by considering the outcomes of this study .
The structure of pyruvate was drawn and optimized using Avogadro 1 . 1 . 0 [42] . Topology and parameter files for pyruvate were generated with SwissParam [43] . The protonation state was determined using Marvin Sketch 5 . 11 . 5 ( ChemAxon , Budapest , Hungary; http://www . chemaxon . com/ ) , appropriate to a solution pH of 7 . 0 . The high-resolution ( 1 . 45 Å ) crystal structure of DHDPS from the bacterium S . aureus ( PDB ID 3DAQ ) [12] was utilized for MD simulations . Protein chains C and D , both artefacts of crystal packing , were removed . Crystallographic waters were discarded and missing hydrogens added using VMD [44] . The structure was solvated in a rhombic dodecahedron using TIP3P waters [45] extending at least 12 Å from any protein atom . Na+ and Cl- were added at a set concentration of 150 mM , with additional Na+ added to neutralize the total charge of the system . This process required a total of 18 , 975 waters , 80 Na+ , and 54 Cl- ions . Molecular dynamics data were generated with NAMD 2 . 9 [46] , using the CHARMM22 force field with CMAP corrections [47 , 48] . Temperature was maintained at 310 K using a Langevin thermostat ( damping coefficient 5 ps-1 ) , and system pressure was adjusted to 1 atm by use of a Langevin piston barostat . Periodic boundary conditions were implemented . Long range electrostatics were computed using the Particle Mesh Ewald Method [49] , applying a nonbonded distance cut-off of 12 Å . Covalent bonds associated with hydrogen atoms were restrained using the SHAKE algorithm [50] , allowing for an integration time-step of 2 fs . Trajectory frames were recorded at simulation time intervals of 10 ps . VMD 1 . 9 . 2 [44] was used for trajectory analysis unless otherwise specified . Systems were first minimized for 5000 steps of steepest descent followed immediately by 8 ns of simulation during which the global protein backbone RMSD was allowed to equilibrate , priming the system for production runs . The final frame of this equilibration run was used as a starting reference for a longer 100 ns production run . 80 replicas of the DHDPS system were seeded by sampling from the 100 ns production run at evenly spaced time intervals . This provided an ensemble of different starting protein configurations . Two ligand molecules were randomly placed in each of these systems using Packmol [51] , imposing a distance limit of 6 Å from any protein atom and at least 15 Å from the ζ-nitrogen atom of Lys163 of protein chain B . Overlapping waters were removed . A restraint was placed on the carbon atom of the ligand carboxylate group such that the upper limit of searchable space in the system was restricted to a radius of less than 26 Å from Lys163-Nζ . A schematic of the simulation set-up is given in S8 Fig . Simulations were run for between 10 ns to 100 ns as computing resources allowed . 553 , 200 conformations were obtained from the simulation data set . These conformations were clustered using the hybrid k-centers k-medoids algorithm implemented within MSMBuilder2 2 . 7 [52] , using a reduced selection of the total simulation data set sub-sampled at intervals of 250 ps . Clustering was performed based upon the RMSD of pyruvate heavy atoms to produce a fine-grained , 738-state model . Protein backbone alignment using both protein chains was carried out prior to clustering . Clustering was performed using MSMBuilder2 2 . 7 [52] . Following clustering , the remaining conformations from the simulation data set were assigned to these clusters . The average cluster radius in the fine-grained model ( ± standard deviation ) was found to be 2 . 6 ± 0 . 4 Å , indicating that clusters are sufficiently tight to provide a high degree of resolution . MSMs in this study were constructed based upon methods described elsewhere [26] , using a lag time of 4 ns ( see S1 Text for details ) , allowing us to construct a highly detailed microstate representation of the binding pathway . Such microstate representations are useful for examining kinetic properties of MSMs , but are often too complex to be readily understandable . To address this shortcoming we coarse-grained the microstate model into a simpler , 17-state model . The number of states was selected by examination of the Bayes factor using similar principles reported by [53] . Observations made from coarse-grained MSMs are useful in that they can allow for comparison with experiment . For this study , we used a Bayesian method [34] as implemented in MSMBuilder2 2 . 7 [52] to perform coarse-graining . The major flux pathways connecting the unbound state to the bound state were investigated using TPT [35 , 36] as implemented within MSMBuilder2 2 . 7 [52] . In our calculations , we considered S2 and S6 as the unbound states , and S13 as the bound state . PMF profiles were obtained using umbrella sampling methods with the collective variable module within NAMD [46] . A one-dimensional binding coordinate ( Z-coordinate ) was defined using two arbitrary reference points designed to create an axis that encompassed the binding pathway depicted in Fig 4 . The binding , or Z-coordinate , was sampled in windows of 0 . 5 Å or 1 Å , using harmonic force constraints of 10 . 0 kcal mol-1 and 7 . 5 kcal mol-1 , respectively . The geometric center of non-hydrogen pyruvate atoms was restricted to within a radius of 7 . 5 Å by a boundary force of 10 . 0 kcal mol-1 from the Z-axis . This was imposed to reduce the degrees of freedom . Additionally , the Cα of the protein were constrained to within 0 . 5 Å of their starting positions by a 10 . 0 kcal mol-1 force constraint to prevent rotational and translation drifts that may affect the definition of the Z-coordinate . PMF curves were calculated using the weighted histogram analysis method ( WHAM ) [54] ) . Z-coordinate measurements were captured every 0 . 2 ps . Convergence of the dataset was assessed by segmentally truncating the simulation data within each window to reduce the sample size . Reported free energies were calculated from the final 2 . 5 ns of simulation time and errors attributed using 100 Monte Carlo bootstraps with time correlation values equal to the time taken for an autocorrelation to decay to 1/e . | Interactions between proteins and ligands underpin many important biological processes , such as binding of substrates to their cognate enzymes in the process of catalysis . These interactions are complex , often requiring several intermediate steps to fully transition into the bound state . Here , we have used computational simulation to study binding of pyruvate to Dihydrodipicolinate synthase ( DHDPS ) , an enzyme in the bacterial diaminopimelate pathway . In bacteria , such as the human pathogen S . aureus , DHDPS functions to make building blocks necessary for protein and bacterial cell wall biosyntheses . As the enzyme is absent in humans , yet essential for bacterial growth , DHDPS is a valid target for broad-range antibiotics . However , known DHDPS inhibitors show poor potency . One avenue that has not yet been taken into consideration for inhibitor design is the dynamics of DHDPS’s interaction with its reaction substrates ( e . g . pyruvate ) . Using molecular dynamics simulation , we find that pyruvate binding to DHDPS must pass through a transition intermediate ‘hotspot’ in which the substrate is held in place by a dense network of noncovalent bonds . Given that many of the protein residues involved in this interaction are also shared by DHDPS from many pathogenic bacteria , this binding intermediate ‘hotspot’ may help in development of better broad-range DHDPS inhibitors . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"crystal",
"structure",
"chemical",
"compounds",
"pathology",
"and",
"laboratory",
"medicine",
"protein",
"interactions",
"markov",
"models",
"pathogens",
"ketones",
"condensed",
"matter",
"physics",
"microbiology",
"pyruvate",
"... | 2016 | Dynamic Modelling Reveals ‘Hotspots’ on the Pathway to Enzyme-Substrate Complex Formation |
Glutamate homeostasis in the brain is maintained by glutamate transporter mediated accumulation . Impaired transport is associated with several neurological disorders , including stroke and amyotrophic lateral sclerosis . Crystal structures of the homolog transporter GltPh from Pyrococcus horikoshii revealed large structural changes . Substrate uptake at the atomic level and the mechanism of ion gradient conversion into directional transport remained enigmatic . We observed in repeated simulations that two local structural changes regulated transport . The first change led to formation of the transient Na2 sodium binding site , triggered by side chain rotation of T308 . The second change destabilized cytoplasmic ionic interactions . We found that sodium binding to the transiently formed Na2 site energized substrate uptake through reshaping of the energy hypersurface . Uptake experiments in reconstituted proteoliposomes confirmed the proposed mechanism . We reproduced the results in the human glutamate transporter EAAT3 indicating a conserved mechanics from archaea to humans .
Glutamate is the primary excitatory neurotransmitter in the central nervous system . Excitatory amino acid transporters ( EAAT ) are membrane proteins , which remove released glutamate from the synaptic cleft [1] . Substrate transport by the human EAATs is coupled to transport of three sodium ions and one proton followed by the counter-transport of one potassium ion [2–4] . Persistent elevation of extracellular glutamate can lead to excitotoxicity through neuronal cell death . Dysfunction in EAATs has been implicated in several neurological diseases , which are associated with loss of neurons ( e . g . , amyotrophic lateral sclerosis , Alzheimer’s disease , stroke , cerebral ischemia , traumatic brain injury , epilepsy and Huntington’s disease ) [5–9] . Development of drugs , which could enhance EAAT transporter activity , would be beneficial for patients . The aspartate transporter from Pyrococcus horikoshii ( GltPh ) [10–13] is a homolog of the mammalian EAATs . GltPh and human EAATs share ~36% sequence identity . Sequence conservation is higher for residues implicated in substrate and ion binding [14–18] . The first crystal structure of the trimeric GltPh was solved in the outward-occluded state [10] , followed by structures determined in three additional conformations [11–13 , 19]: outward-open , inward-occluded and intermediate , revealing a large scale translational motion during substrate transport [12] . The transport domain contains substrate and ion binding sites , while the trimerization domain mediates the contact between the protomers . Functional studies indicated that transport of one substrate molecule is coupled to the co-transport of three sodium ions [20 , 21] . The respective sodium binding sites were termed Na1 , Na2 and Na3 . Simulations of the outward-occluded and the inward-occluded conformation have been used to study binding of substrate and co-transported ions [22–24] . The importance of the HP1 and HP2 gates for restricting the accessibility to the substrate-binding site has been studied using molecular dynamics ( MD ) simulation [23 , 25–28] . The Na3 sodium-binding site [23] has been predicted by MD simulations and was experimentally validated in GltPh and EAATs [24 , 29–31] . The conformational changes of substrate translocation have been studied by motion planning , elastic network model and metadynamics [32–34] . A comparision of the outward and inward facing crystal structures ( PDB ID: 2NWX , and 3KBC ) suggested a nearly rigid body movement of the transport domain vs . the trimerization domain . We used steered molecular dynamics ( SMD ) simulations to investigate the substrate translocation step from the outward-facing to the inward-facing conformation . We identified local events that are essential for coupling of the ion gradient to substrate transport and verified our observations by experiments . Repeated SMD simulations of 0 . 3 μs length revealed the existence of an intermediate state , which is characterized by maturation of the Na2 binding site . Maturation required changes in the local environment , which are triggered by rotation of the side chain of T308 . We provide biochemical evidence for the localization and the functional role of the Na2 site: ( i ) mutation of residue T308 increased Km for substrate and sodium in GltPh and in human EAAT3 . ( ii ) The Hill coefficient of sodium dependence was reduced from 2 . 5 for wild type to 1 . 7 for T308V and 1 . 1 for T308A confirming the location of the Na2 site and its role in coupling of the sodium gradient to substrate uptake .
Substrate uptake by GltPh involves a large conformational change [12] , which exposes the substrate binding site alternatingly to the extracellular or intracellular side of the membrane , consistent with the alternating access models [38] . The equilibrium simulations indicated that the transport process is too slow to be investigated directly . We therefore used SMD simulations to accelerate the transition . Each SMD simulation was repeated 3 times . The crystal structures ( PDB ID: 2NWX and 3KBC ) provided the framework for defining the reaction path [11–13] . They show that transport and trimerization domain undergo a translational motion relative to each other . The reaction path is almost perpendicular with respect to the membrane . We therefore applied a force on both domains in opposing directions , oriented perpendicular to the membrane plane . The force was applied to the backbone atoms of residues in the region that forms the domain interface , selecting residues K266-V274 , L282-K290 , I298-L305 of the transport domain and residues L54-G69 , P153-Y167 , K178-A214 of the trimerization domain ( Fig 1A ) . Selection of the interface region was motivated by the minimal torque induced in case of side chain entanglements at the domain interface . Trial simulations with several different pull group selections supported this choice . The pulling velocity applied was set to 0 . 01 pm ps-1 . Faster pulling rates with the same and with stronger spring constants were also tested , but resulted in increasingly unstable behavior ( S1 Fig ) . Our simulations are well behaved , as evident from the force vs . time and distance vs . time plots ( S2 Fig ) . We monitored the time evolution of each protomer ( chain ) by measuring the RMSD with respect to the inward-occluded state as observed in the crystal structure ( PDB ID: 3KBC ) [12] ( Fig 1B , 1C and 1D ) . The RMSD between the outward-occluded and the inward-occluded state is 0 . 90 nm . A decrease to 0 . 3–0 . 5 nm indicated a successful transition to the inward-occluded state with essentially overlapping secondary structure elements ( Fig 1E ) . The RMSD between the inward facing crystal structure ( PDB ID: 3KBC ) and the overlaid MD snapshot was 0 . 35 nm . Structure and conformation of the individual domains remained close to the initial conformation ( S3 Fig ) , indicating that the structure of the domains remained intact . The final rise of the RMSD towards the end of the simulations , after reaching the inward-occluded state , reflects pulling of the transporter beyond the inward-facing state , while reaching the 300 ns defined as simulation length . It marks the transition from conformational changes as part of the transport cycle to distortions ( S3 Fig ) induced by the SMD protocol . The caesura observed in the structural stability of the individual domains by pulling beyond the inward-facing state also indicates that we were able to move the transporter along the low energy path for the investigated transition from the outward-facing to the inward-facing state . The inward-occluded state was crystallized with the help of a conformation stabilizing double cysteine mutant ( K55C and A364C ) . We observed a transition to the inward-facing state of at least one protomer in each simulation . Thereby residues K55 and A364 were found to come into close proximity ( See S4 Fig ) . A second chain reached the inward-facing state in two instances . The third chain showed only once a transition to the inward facing state . The extracellular gate of the A chain is more open in the crystal structure ( PDB ID: 2NWX ) ( See also S3A–S3C Fig ) . For simplicity , below we discuss results of only one simulation . Analyses of the other two simulations lead to the same conclusions ( S5 and S6 Figs ) . We observed a progression towards the inward-occluded state in two global steps . Representative snapshots ( Fig 2A , 2B and 2C ) are depicted at three time-points , representing the outward-occluded ( at 0 ns ) , the intermediate ( at 152 ns ) and the inward-occluded state ( at 272 ns ) . The structural overlap with the crystallized intermediate structure ( PDB ID: 3V8G ) was largest in the intermediate conformation of our trajectories ( S7 Fig ) . The observed structural rearrangements showed a movement of the transport domain ( orange ) relative to the trimerization domain ( blue ) . The transport domain is in direct contact with the membrane; it moved as a rigid body ( S3 Fig ) , and translated only minimally relative to the membrane . In contrast , the core of GltPh ( TM2 , TM4 and TM5 ) , which showed minimal contacts with the membrane , displayed a large movement towards the extracellular side . This was accompanied by a switch in accessibility of the substrate binding site from extracellular to intracellular . We identified two distinct dynamic subdomains within the trimerization domain . A comparison between available crystal structures ( PDB ID: 2NWX , 3V8G and 3KBC ) suggested that the trimerization domain could bend within helices TM2 and TM5 . We observed that the central part of the trimerization domain moved relative to the transport domain . In contrast , the membrane-exposed helices TM1 and the membrane adjacent parts of TM2 and TM5 did not translocate . Instead , the trimerization domain bent at residues P45-G47 , P60 in TM2 and P206-G208 in TM5 to allow the membrane-exposed part to maintain membrane contacts . We observed that sodium bound to the Na2 site was unstable in the crystallographically observed position of the outward-occluded state ( Fig 3A ) . The side chain of T308 rotated by 120 degrees in the course of the SMD simulations ( Fig 3B ) . The same re-rotation did not occur in equilibrium simulations . Residue T308 is adjacent to the crystallographically observed Na2 site and located on the last turn of the helix TM7A . The contiguous NMDG motif was implicated in substrate and ion binding [15 , 39] . The hydroxyl group of the T308 side chain formed a hydrogen bond with the backbone carbonyl of P304 on the preceding helical turn . Breaking of the P304-T308 hydrogen bond triggered rotation of the T308 side chain ( Fig 3C ) . The transition correlated with a reduced hydration of T308 ( Fig 3D ) . We extracted closely spaced snapshots from before and from after rotation of the T308 side chain to investigate , if partial dehydration and rotation of the T308 side chain improved coordination of sodium at the Na2 site . We re-inserted sodium into its original position in the Na2 site and carried out 50 ns long equilibrium MD simulations . The protein was stable in all simulations: we observed Cα RMSD values between 0 . 12 and 0 . 16 nm for the TM helices . The low RMSD showed that the pulling rate of the SMD simulation was slow enough to allow the transporter to proceed along the low energy path without significant force induced distortions . Sodium bound to the Na2 site was unstable before rotation of the T308 side chain and dissociated from the Na2 binding site in simulations using starting conformations extracted before rotation of T308 ( Fig 4A ) . In contrast , sodium was stable after rotation . The structural adjustments , which were only possible after breaking of the hydrogen bond between T308 and P304 , needed a few nanoseconds to propagate through TM7A to form the Na2 sodium binding site , because sodium was found to remain stably bound to the position in the crystal structure ( of Tl+ ) only 6 ns after rotation of T308 . We observed that backbone carbonyl oxygen atoms of the last turn of helix TM7A improved coordination of the sodium ion . The fully assembled Na2 binding site was formed by the carbonyl oxygen atoms at the C-terminal ends of helix TM7A , the first helix of the HP2 loop and by the sulfur atom of the M311 side chain . Na2 remained stably coordinated in the inserted position in the simulations started from the last two tested time points ( See Fig 4B and 4C ) , indicative of a completed Na2 sodium binding site . We also found ( using APBS [40] ) a considerable increase in the negative potential at the Na2 site . A slightly different binding site for sodium 2 had been proposed from 2 ns equilibrium simulations , followed by free energy calculations [22] . T308 was found in the rotated state after 2 ns [22] and sodium 2 had moved from the crystallographically observed Tl+ position and interacted with the side chain of T308 . Thereby , sodium had moved away from the center of the region with negative electrostatic potential created by the helix dipoles of helix TM7A and the first helix of the HP2 loop . We carried out three independent simulations to investigate the possibility that a rotated T308 in the start structure could stabilize the sodium in the Na2 binding site . Sodium shifted to the same position with direct contact to the hydroxyl group of the T308 side chain as proposed earlier [22] in the beginning of each simulation . This position resulted metastable . In two simulations sodium dissociated from this intermediate site after 5 ns and diffused into the water phase . In the third simulation , sodium detached within the first nanosecond , suggesting that this could be a intermediate state on the path of sodium binding to the Na2 site . The outward-facing transporter is stabilized at the cytosolic side by an interaction network between the trimerization domain and the transport domain . The network includes residues E192 , Y195 and N199 on helix TM5 ( trimerization domain ) and residues R287 and K290 on the HP1 loop ( transport domain ) ( Fig 5A and 5B ) . We observed salt bridges between residues E192 and K290 and between residues N199 and R287 . The salt bridge between E192 and K290 was further stabilized by a cation-π interaction of K290 with the aromatic ring of Y195 . The distance of the salt bridge between atoms Cδ of E192 and atom Nζ of K290 increased from direct residue contact in the outward-facing conformation to a broad distribution around 1 . 5 nm in the intermediate state ( Fig 5C ) . Opening of the salt bridge ( E192-K290 ) was accompanied by rotation of Y195 . The distance increased further , when the inward facing state was reached . The second salt bridge between residues N199 and R287 showed similar behavior . Our simulations suggested that ( i ) the Na2 sodium-binding site forms during substrate translocation and that ( ii ) the side chain of residue T308 is a key player of its maturation . We challenged this hypothesis by experimentally measuring the involvement of T308 in binding and transport . We created three types of mutations: ( i ) the T308W mutation introduced an amino acid that would be too large to allow for transport; ( ii ) the smaller T308A and the T308V mutations were designed to remove the hydroxyl group; ( iii ) the T308S mutation retained the hydroxyl group ( Fig 6 and Table 1 ) . The T308W mutation abolished transport as predicted ( Fig 6A ) . The T308A and T308V mutations showed increased Michaelis-Menten constant ( Km ) values for the substrate L-Asp and for the co-transported sodium ions ( Fig 6B and 6C ) . The transport stoichiometry of GltPh predicts that three sodium ions are co-transported with substrate . A Hill coefficient of ~2 . 5 for sodium indicates strong cooperativity [20] . This was recapitulated in our experiments: we observed for wild type GltPh a Hill coefficient of 2 . 5 ± 0 . 3 for sodium-supported substrate uptake ( Fig 6C ) . The cooperativity was reduced in both mutants , which removed the hydroxyl group , i . e . to 1 . 1 ± 0 . 1 and 1 . 7 ± 0 . 3 in T308A and T308V , respectively . The drop in cooperativity is consistent with the prediction that the mutations affected binding of sodium to the Na2 site . The Km of substrate uptake was increased from 119 ± 18 nM in wild type GltPh to 183 ± 26 nM and 220 ± 24 nM for T308A and T308V , respectively . The T308S mutation maintained the hydroxyl group at the γ position . We observed an almost 3 fold decrease in the maximal transport rate ( Vmax ) , but only a very small change in the Hill coefficient ( 2 . 3 ± 0 . 7 ) . The Km for sodium ( 9 . 2 ± 2 . 3 mM ) and substrate ( 134 ± 27 ) were unchanged . GltPh is predicted to share its fold with the human EAATs . Sequence alignments showed that T308 and P304 are conserved between the archeal and the human transporters . We therefore reasoned that mutations in human EAATs would recapitulate the effects seen in GltPh , if the structure-function relationship was shared . The same mutations ( tryptophan , alanine , valine and serine ) of corresponding residue T364 were introduced to the human EAAT3 . Experiments were carried out in transiently transfected HEK293 cells ( Fig 7 ) . Mutations of EAAT3 reproduced the findings seen in GltPh: the Km for transport was not changed in EAAT3-T364S ( 37 ± 7 μM ) as compared to wild type EAAT3 ( 41 ± 4 μM ) , but increased in EAAT3-T364V ( 79 ± 20 μM ) , in EAAT-T364A ( 159 ± 48 μM ) and EAAT3-T364W ( 341 ± 82 μM ) . In contrast to GltPh , we found residual transport in EAAT3-T364W , indicating that EAAT3 might be more tolerant .
During substrate translocation , accessibility of the substrate binding site changed from the extracellular to the intracellular side . We used SMD simulations to investigate the structural details of this large scale conformational change . Analysis of the conformational landscape revealed the existence of an intermediate state , which is similar to the conformation observed in the crystal structure ( PDB ID: 3V8G ) [13] , showing an RMSD of 0 . 3–0 . 4 nm . Binding of substrate and sodium has been shown to induce the closure of the HP2 gate [25 , 26 , 36] . We found that the Na2 sodium-binding site is not yet fully formed in the outward-occluded state [12] . The sodium ion escaped in spite of weak restraints , which were imposed to keep it in place . The Na2 site was found to mature during the transition to the intermediate state . The side chain of residue T308 played a pivotal role: breaking of the P304-T308 hydrogen bond weakened the structural stability of the last turn of the TM7A helix ( Fig 3 ) . The backbone carbonyl oxygen atoms could therefore adjust and optimally coordinate the sodium ion in the Na2 site . The structural adaptation increased the negative electrostatic potential at the Na2 site , which became more attractive for the positively charged sodium ion . Sodium binding can neutralize the negative potential , thereby removing an energy barrier for the transition to the intermediate state . Therefore , we carried out additional MD simulations to confirm that the Na2 site matured during the transition to the intermediate state . Simulations starting from closely spaced snapshots showed that the Na2 site became competent for sodium binding only after breaking of the hydrogen bond between the side chain of T308 and the backbone carbonyl oxygen of P304 . Our results indicate that breaking of this hydrogen bond is the key event that triggers the structural adjustments that lead to maturation of the Na2 site . This observation is in line with the predicted sequence of binding events in GltPh: sodium 3 , followed by sodium 1 , substrate and sodium 2 [22 , 25 , 26] . Electrophysiological measurements determined that binding to the outward-facing EAAT3 transporter occurred in at least two steps: ( i ) fast binding of substrate and sodium ion ( s ) followed by ( ii ) slow binding of one additional sodium ion [41] . The transporter translocated substrate only after the slow sodium binding event . We used SMD simulations to study substrate uptake . A drop in the RMSD of each chain reflects a movement towards the inward-occluded state . Analyses of the trajectories revealed two crucial local events that accompanied the transition through the intermediate state: ( i ) formation of the Na2 site ( Fig 3 ) and ( ii ) opening of the interaction network at the intracellular side ( Fig 5 ) . We applied correlation analyses independently to each protomer ( Fig 8 ) to determine , if these local changes occurred first and thereby triggered the global changes ( See S8 Fig for analyses of all 3 simulations ) . The two largest global motions consisted of the translation of the transport domain vs . the trimerization domain ( principal component 1 ) , and a rotation of the transport vs . the trimerization domain ( principal component 2 ) . Only protomer B and C reached the inward-occluded state . The time points of local changes were color-code mapped onto the 2d projection . The analyses revealed that ( i ) formation of the Na2 site occurred before the transition to the intermediate state . ( ii ) Opening of the interaction network was also required prior to formation of the intermediate state . It followed after maturation of the Na2 site . Mutations of residues contributing to the intracellular interaction network ( specifically K290A ) were shown to reduce substrate transport [42] , most likely by destabilizing the outward-facing state . We found the conformation of the Y195 side chain to be a sensor for early onset of conformational changes at the inner site of the transporter . The long side chains of E192 and K290 allowed for limited structural adjustment before rupture . Only once the salt bridge opened could substrate translocation proceed . The same results were obtained in the analyses of the other two simulations ( S7 Fig ) . Our correlation analyses therefore suggest that the local changes are the key switches , which prepare GltPh for the global changes and should therefore control the transition through the transport cycle . These observations have important implications for the transport mechanism . They predict that the local events represent essential shifts in the energy hypersurface along the transport cycle: ( i ) binding of substrate and sodium to the Na1 and Na3 site stabilizes the outer gate HP2 loop in the closed state [36] . ( ii ) Occupation of the Na2 site reduces the energy barrier for the transition to the intermediate state . ( iii ) Opening of the interaction network at the cytosolic site allows transport and trimerization domain to move relative to each other . Accessibility of the substrate binding site changes from extracellular to intracellular . ( iv ) Opening of the intracellular gate allows for substrate release and ion dissociation [28] . These events indicate that the role of sodium binding is to modify the energy hypersurface of the transporter and , as a consequence , to allow for directional transport . Directionality can be maintained as long as the internal sodium concentration remains low to reset the transporter [43] . We confirmed this mechanism by performing radiotracer flux experiments of wild type and mutant GltPh . Residues P304 and T308 in TM7A are conserved from archaea to humans , preserving the hydrogen bonding property . The conservative T308S mutation , which maintains the hydroxyl group , had little effect on Hill coefficient and Km . Removal of the hydroxyl group ( mutated to alanine and valine ) was predicted to alter the Km and the Hill coefficient . We observed a larger effect for the T308A mutant as compared to the T308V mutant . Valine and threonine are both β-branched amino acids and have helix destabilizing properties , while alanine does not have the same property , suggesting that an imperfectly stabilized TM7A helix is important for transporter function . These experimental results support the transport mechanism model developed from our simulations . Our results revealed mechanistic insights into the mechanics of substrate translocation , which are conserved between GltPh and human EAATs . Controlled glutamate uptake by the EAATs is essential for brain function: a reduced uptake is associated with excess glutamate in the synaptic cleft that can lead to excitotoxicity and neurodegenerative diseases including amyotrophic lateral sclerosis , stroke , epilepsy or Huntington’s disease . Currently no drug on the market can enhance EAAT transporter function . It should be of interest to identify compounds , which could provide pharmacological leads that apply such a new strategy for patient treatment . The interactions and structural changes , which we identified , represent checkpoints to ensure correct binding of substrate and sodium ions before the conformational transition . We speculate that the Na2 site fulfills three functions: ( i ) it allows the transporter to reach an intermediate state only if substrate and three sodium ions are bound . ( ii ) It prevents the escape of bound substrate to the extracellular environment . ( iii ) Futile cycling events of sodium ions bound transporter are precluded in the absence of substrate , because closure of the HP2 loop and the presence of substrate and sodium in the Na2 site are needed . This arrangement is an elegant solution to harness the energy stored in the sodium gradient and to prevent unproductive cycling , which only translocates sodium ions . Accumulation of internal sodium would be harmful to neurons and glia cells , because the electrochemical gradient across the cell membrane is eventually dissipated by futile cycling .
Simulations were carried out using the GROMACS 4 . 5 . 4 package [44] . The OPLS [45] all atom force field was applied to GltPh , the POPC membrane was described using the Berger lipids [46] force field . The inward-facing ( PDB-ID: 3KBC ) [12] and the outward-occluded crystal structure of GltPh ( PDB-ID: 2NWX ) [11] were selected; missing residues ( side chains and residues ) were added using MODELLER version 9 . 8 [47] . The third sodium ion was placed into the previously reported ion binding site [23] . GltPh was embedded in a pre-equilibrated POPC bilayer using the g_membed [48] procedure and solvated with water , electro-neutralized and a physiological concentration of NaCl ( 150 mM ) . The SPC water model [49] was used . Lennard-Jones interactions were calculated until a cutoff of 1 . 0 nm . Long range electrostatic interactions were treated using PME [50] . Bond lengths were constrained using LINCS [51] . The temperature was maintained at 310 K applying the v-rescale thermostat [52]; pressure was set to 1 bar and controlled by the Berendsen barostat [53] , coupling protein , membrane , and water/ions separately . The assembled system was equilibrated by first relaxing membrane and solvent . The protein was then gradually relaxed by reducing the restraints on the protein atoms in five steps [54] ( each 2 . 5 ns simulation ) applying a restraining force on the protein of 1000 , 100 , 10 , 1 , 0 . 1 kJ mol-1 nm-2 . Production runs were then started . We carried out equilibrium simulations starting from two independently prepared inward and the outward-occluded state , and nine simulations starting from snapshots extracted from one SMD simulation . We used SMD simulations [55] to induce a conformational change from the outward- to the inward-facing conformation . We selected the same number of Cα atoms from both the trimerisation and the transport domain . The spring force constant was set to 1000 . 0 kJ mol-1 nm-2 , applied only along the membrane normal ( Z-direction ) . Pilot simulations with different force constants indicated that higher force constants would lead to a force profile that showed an increasing level of fluctuations and a larger number of force peaks in the force profile ( See S1 Fig ) , indicative for entanglement and forced barrier-crossing events . These force peaks seemed to depend on the applied sprint constant , because they gradually disappeared with decreasing restraining force . Repeated simulations showed that the peaks did not always appear at the same time points . Correct behavior of the protein could therefore only be expected with a force constant of 1000 . 0 kJ mol-1 nm-2 . All simulations were repeated three times . Pilot simulations were also performed to identify the optimal pulling velocity . We selected a pulling rate of 0 . 00001 nm ps-1 . This velocity is slow enough to give GltPh time to relax all not coupled degrees of freedom and fast enough to be still accessible to MD simulations . Higher pulling speed resulted in unphysical structural deformations and increasingly larger peaks in the force time contour . We carried out three SMD simulations starting from independently prepared systems . Eric Gouaux ( Oregon Health & Science University ) generously donated wild type GltPh . Site-directed mutagenesis was performed using Quick Change Lightning site directed mutagenesis kit ( Agilent Technologies ) . All mutations were confirmed using DNA sequencing . GltPh was purified as described previously [56] . Briefly , membranes were isolated , solubilized with ( 1% ) n-dodecyl-β-D-maltopyranoside ( DDM ) ( Anatrace Detergents ) , and the protein was purified using Ni-NTA resin . The detergent was then exchanged to ( 0 . 15% ) n-decyl-β-D-maltopyranoside ( DM ) ( Anatrace Detergents ) . The protein was reconstituted into liposomes using the method described previously with slight modifications [57] . Escherichia coli total lipids ( Avanti Polar lipids Inc . ) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine ( Avanti Polar Lipids Inc . ) were mixed at a ratio of 3:1 , dried under nitrogen and re-suspended in internal buffer ( 100 mM KCl , 20 mM HEPES , pH 7 . 5 ) . The lipid mixture was briefly sonicated using a point tip sonicator . The lipid suspension was flash frozen in liquid nitrogen and thawed at room temperature multiple times . Liposomes were formed by extrusion through 400 nm polycarbonate membranes ( Avanti Polar Lipids Inc . ) and destabilized with Triton X-100 prior to addition of protein at a ratio of 0 . 25 μg protein per mg lipid . The protein-lipid mixture was kept at room temperature for 30 minutes . Detergent was then removed using SM2 biobeads ( Biorad ) . The protein-lipid mixture was incubated and gently agitated with three consecutive batches of biobeads ( 20 mg ml-1 ) . The biobeads were subsequently removed by filtration . The resulting proteoliposomes were concentrated by centrifugation at 150 , 000 g for 30 min in a Beckman centrifuge , re-suspended at 100 mg lipid mL-1 and either used immediately or flash frozen in liquid nitrogen and stored at −80°C . Transport of [3H] L-aspartate by wild type and mutant GltPh was assayed by applying the protocol modified from Gaillard et al . , 1996 [56] . Briefly , proteoliposomes were loaded with buffer A ( 100 mM KCl , 20 mM HEPES , pH 7 . 5 ) by repeated freeze and thaw cycles followed by extrusion . The uptake was initiated by diluting the proteoliposomes into buffer B ( 100 mM NaCl , 20 mM HEPES , pH 7 . 5 , 1 μM valinomycin and the indicated concentrations of [3H] L-aspartate ) pre-warmed to 30°C . At each time point , a 200 μL aliquot was removed and diluted 10 fold into ice cold stop buffer C ( 100 mM LiCl , 20 mM HEPES , pH 7 . 5 ) , followed by filtration over nitrocellulose filters ( 0 . 22 μm pore size , Merck Millipore ) . The filters were washed three times with 2 mL of ice-cold quench buffer and assayed for radioactivity . Sodium dependence of [3H] L-aspartate transport was measured by varying extraliposomal Na+ concentration from 0 . 1 to 300 mM . Osmolarity was balanced using choline chloride ( Sigma Aldrich ) . [3H] L-aspartate concentrations used for determining the sodium dependence were 100 nM for wild type and around Km for other mutants . Background levels of uptake were measured by diluting proteoliposomes into buffer A containing 1 μM valinomycin and the indicated concentrations of [3H] L-aspartate . The plasmid PL28 encoding for the human excitatory amino acid transporter 3 ( EAAT3 ) was a kind gift by Peter Larsson ( Oregon Health & Science University ) . The EAAT3 cDNA was cloned into pmGFP-C1 . The point mutations T364A , T364V , T364S , and T364W were generated using Quick Change Lightning site-directed mutagenesis kit II ( Agilent technologies ) and confirmed by DNA sequencing . Human embryonic kidney 293 ( HEK293 ) cells were cultured in Dulbecco’s Modified Eagle’s Medium ( DMEM ) , which was supplemented with 10% fetal bovine serum at 37°C and 5% CO2 . Cells were transiently transfected with wild type mGFP-EAAT3 with the EAAT3 mutants T364A , T364V , T364S , and T364W using the calcium phosphate method . Cells were seeded ( 0 . 1*105 cells per well ) on 48-well plates pre-coated with poly-D-lysine . The transport assay was performed 48 hours after transfection . Cell were washing with 500 μl of Krebs HEPES buffer ( 10 mM HEPES , 130 mM NaCl , 1 . 3 mM KH2PO4 , 1 . 5 mM CaCl2 , 0 . 5 mM MgSO4 , pH adjusted to 7 . 4 using NaOH ) . Uptake was initiated by incubating cells with increasing concentrations ( 10 , 30 , 100 , 300 , 600 and 1000 μM ) of unlabeled substrate L-glutamate ( Sigma Aldrich ) containing tracer amount ( 100 nM ) of radiolabeled [3H] L-glutamate ( Perkin Elmer ) . The uptake was terminated after 10 min by washing with ice cold Krebs HEPES buffer . Cells were lysed with 500 μl of 1% sodium-dodecyl-sulfate and radioactivity was measured with the Liquid Scintillation Analyzer . Non-specific uptake was determined by pre-incubation with the inhibitor L-trans-pyrrolidine-2 , 4-dicarboxylate ( PDC ) ( Sigma Aldrich ) 10 min prior and during incubation ( 100μM ) and subtracted from the total counts . | We used the archaeal homolog GltPh of the human glutamate transporters to refine our understanding how large scale conformational changes are translated into substrate translocation . We identified the structural changes that accompany substrate transport and convert the energy stored in the ion gradient into a directional transport . Insights into the mechanics of these transporters are likely to increase our understanding of how they contribute to excitotoxicity and to develop drugs , which preclude the underlying accumulation of glutamate in the synaptic cleft . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
] | [] | 2015 | Refinement of the Central Steps of Substrate Transport by the Aspartate Transporter GltPh: Elucidating the Role of the Na2 Sodium Binding Site |
Quiescence and gametogenesis represent two distinct survival strategies in response to nutrient starvation in budding yeast . Precisely how environmental signals are sensed by yeast cells to trigger quiescence and gametogenesis is not fully understood . A conserved signalling module consisting of Greatwall kinase , Endosulfine and Protein Phosphatase PP2ACdc55 proteins regulates entry into mitosis in Xenopus egg extracts and meiotic maturation in flies . We report here that an analogous signalling module consisting of the serine-threonine kinase Rim15 , the Endosulfines Igo1 and Igo2 and the Protein Phosphatase PP2ACdc55 , regulates entry into both quiescence and gametogenesis in budding yeast . PP2ACdc55 inhibits entry into gametogenesis and quiescence . Rim15 promotes entry into gametogenesis and quiescence by converting Igo1 into an inhibitor of PP2ACdc55 by phosphorylating at a conserved serine residue . Moreover , we show that the Rim15-Endosulfine-PP2ACdc55 pathway regulates entry into quiescence and gametogenesis by distinct mechanisms . In addition , we show that Igo1 and Igo2 are required for pre-meiotic autophagy but the lack of pre-meiotic autophagy is insufficient to explain the sporulation defect of igo1Δ igo2Δ cells . We propose that the Rim15-Endosulfine-PP2ACdc55 signalling module triggers entry into quiescence and gametogenesis by regulating dephosphorylation of distinct substrates .
The ability of cells to sense deleterious changes in environment and mount an appropriate physiological and metabolic response is essential for cellular survival . Response to nutrition starvation in budding yeast has been an extremely powerful model to study this biological trait [1] . Upon complete nutrient starvation , yeast cells enter either gametogenesis or quiescence . Diploid yeast cells undergo gametogenesis when subjected to nitrogen starvation in the absence of glucose and in the presence of a non-fermentable carbon source . They undergo one round of DNA replication followed by two rounds of nuclear divisions to form 4 haploid spores which can stay dormant for long periods of time . Haploid and diploid cells enter quiescence when subjected to nutrient starvation or when treated with a drug called rapamycin , an inhibitor of the TOR ( Target of Rapamycin ) signalling pathway . Quiescence ( -also referred to as G0 ) is a reversible non-proliferative state characterized by low rates of transcription and translation , increased stress-tolerance , elevated rate of macroautophagy and synthesis of storage carbohydrates ( trehalose and glycogen ) . Many of the G0-features like increased macroautophagy , low rates of transcription and translation are also characteristic of quiescent mammalian cells suggesting that the core features of quiescence program are conserved [2] , [3] . Ablation of G0-entry/exit control mechanisms is frequently linked to either reduced life span ( especially in unicellular organisms ) or cellular transformation ( in multi-cellular organisms ) [4] , [5] . In budding yeast , entry into quiescence is controlled by the master regulator Rim15 , a member of the AGC ( named after protein kinase A , G and C families ) group of serine-threonine kinases [6] . Activity of Rim15 is controlled by two nutrient signalling pathways namely the Ras/Protein Kinase A ( Ras/PKA ) and the Target of Rapamycin Complex 1 ( TORC1 ) pathways . The TORC1 pathway responds to the availability of nitrogen source in the growth medium [2] , [5] . In contrast , the Ras/PKA pathway responds to levels of glucose in the growth medium [2] , [5] . Both pathways positively regulate cell proliferation in response to nutrient availability and thereby inhibit entry into G0 . PKA phosphorylates Rim15 at five consensus PKA phosphorylation sites to inhibit its kinase activity and promote its retention in the cytosol [6] . Apart from PKA and TORC1 pathways , Rim15 also integrates signalling from Sch9 kinase ( ortholog of mammalian Akt/S6 kinase ) and Pho85-Pho80 kinase ( phosphate-sensing ) pathways [7] . Signalling through TORC1 , Sch9 and Pho85/Pho80 pathways phosphorylate Rim15 at Thr-1075 and inhibit its nuclear localization . Nutrient deprivation inhibits signalling through these four pathways which results in dephosphorylation of Rim15 at its five PKA sites and Thr-1075 leading to its activation and translocation to the nucleus . Activated Rim15 stimulates stress-responsive transcription factors Msn2/4 and post-diauxic shift transcription factor Gis1 which in turn activate transcription of several genes required for survival in G0 . Rim15 phosphorylates endosulfines , a highly conserved family of cAMP regulated phosphoproteins to promote entry into quiescence [8] . Endosulfines , following phosphorylation by Rim15 , protect mRNA , which are transcriptionally controlled by Msn2/4 and Gis1 , from degradation via the 5′-3′ mRNA decay pathway by inhibiting Dhh1 ( decapping activator ) and Ccr4 ( deadenylation factor ) [8] . Entry into gametogenesis in yeast is mainly regulated at the level of transcription of IME1 which encodes the master transcription factor for early-meiosis genes ( EMGs ) [9] . Like G0 , entry into gametogenesis is negatively regulated by TORC1 and Ras/PKA pathways . Ime1 is recruited to promoters of EMGs and activates their transcription [10] . During vegetative growth , a DNA-binding protein Ume6 binds to promoters of EMGs and represses their expression by associating with the Sin3/Rpd3 histone deacetylase and Isw2 chromatin remodelling complexes [11] , [12] . Absence of glucose and nitrogen in the medium results in the replacement of Sin3/Rpd3 and Isw2 by Ime1 at the EMG promoter regions [13] . It was proposed that Ime1 activates transcription of EMGs by converting Ume6 from a repressor into an activator [10] . This model was consistent with the observations that Ime1 physically interacts with Ume6 and that cells lacking Ume6 fail to sporulate efficiently [10] , [14] , [15] . However , this model was disputed by subsequent studies which showed that interaction of Ime1 with Ume6 facilitates Ume6 destruction and meiotic gene induction [16] . Rim15 has been implicated in the removal of histone deacetylase complex from the promoters of EMGs [13] but precisely how this is achieved is not known . In this paper , we demonstrate that endosulfines are required for entry into gametogenesis and quiescence in budding yeast . Phosphorylation of endosulfine by Rim15 results in its association with the protein phosphatase PP2ACdc55 and inhibition of its phosphatase activity . We show that the Rim15-endosulfine-PP2ACdc55 signalling module regulates entry into quiescence and gametogenesis by distinct mechanisms . We also demonstrate that this signalling module is required for pre-meiotic autophagy which is necessary for gametogenesis in budding yeast . Remarkably a similar signalling module regulates M-phase progression during mitosis and meiosis in higher eukaryotes . In Xenopus egg extracts , the Greatwall kinase phosphorylates α-endosulfine ( ENSA ) and Arpp19 at a conserved serine residue , which then inhibits PP2A-B55δ to promote entry into mitosis [17] , [18] . Depletion of Greatwall kinase and endosulfine in Drosophila leads to mitotic defects suggests that the module regulates entry into mitosis in flies [19] , [20] . Inactivation of endosulfine in flies causes a failure in oocyte progression from prophase I to metaphase I indicating that this module regulates entry into M-phase during meiosis [21] . Our results therefore expand the repertoire of functions for this highly conserved signalling module that regulates distinct biological processes in different systems .
Since Rim15 is required for expression of early meiotic genes [22] we examined the function of endosulfine in gametogenesis . Budding yeast has two endosulfines Igo1 and Igo2 . We first assessed the ability of wild type , igo1Δ , igo2Δ and igo1Δ igo2Δ strains to sporulate . While wild type , igo1Δ and igo2Δ strains sporulated with an efficiency of ≥65% , only 3% of igo1Δ igo2Δ cells formed spores ( Figure 1A ) . To determine the precise function of endosulfines in spore formation , we induced wild type and igo1Δ igo2Δ cells to enter meiosis by transferring them to Sporulation medium ( SPM ) . We examined expression of early meiotic proteins Ime1 and Rec8 ( by Western blotting ) , pre-meiotic DNA replication ( flow cytometry ) and nuclear division ( DAPI staining ) . Wild type cells replicated their DNA after 5 hours into SPM ( Figure 1B ) , expressed Ime1 and Rec8 ( Figure 1D ) , and underwent two rounds of nuclear division to form tetranucleate spores ( Figure 1C ) . However igo1Δ igo2Δ cells failed to express both Rec8 and Ime1 , did not undergo pre-meiotic DNA replication and remained mononucleate even after 12 hours into SPM ( Figure 1B–1D ) . These results indicate that endosulfines are required for entry into gametogenesis in budding yeast . Induction of sporulation [23] involves arresting cells in stationary phase by growth in nutrient medium contacting acetate as a carbon source for 16 hours . To rule out the possibility that the failure of endosulfine mutant cells to sporulate was due to their inability to exit from stationary phase , we induced logarithmically growing wild type and igo1Δ igo2Δ cells to enter gametogenesis . Wild type but not igo1Δ igo2Δ cells underwent pre-meiotic DNA replication and spore formation ( Figure S1 ) , indicating that endosulfines are required for entry into gametogenesis . The spores formed in igo1Δ igo2Δ cells at a low frequency had viabilities similar to wild type spores ( data not shown ) suggesting that endosulfines are required for efficient entry into gametogenesis but not for rest of the sporulation program . Phosphorylation of endosulfine at a conserved serine residue ( Figure 1E ) by Greatwall kinase is required for entry into mitosis in Xenopus egg extracts [17] , [18] . Phosphorylation at the corresponding Serine residue ( Serine-64 ) in budding yeast Igo1 by Rim15 kinase is required for entry into G0 [8] . To test whether phosphorylation at S-64 also regulates entry into gametogenesis , we tested the ability of phospho-inhibitory igo1-S64A mutant to sporulate . About 50% of igo1Δ igo2Δ cells expressing wild type Igo1 sporulated in comparison to just 2% of control igo1Δ igo2Δ cells . In contrast , only 10% of igo1Δ igo2Δ cells expressing Igo1-S64A sporulated ( Figure 1F ) . The sporulation efficiency of igo1Δ igo2Δ cells expressing the phospho-mimetic mutant Igo1-S64D was 1 . 7 fold more than that of Igo1-S64A expressing igo1Δ igo2Δ cells ( Figure 1F ) . This effect of S64D mutation on sporulation efficiency was independent of Rim15 function ( Figure S2 ) . These results suggest that phosphorylation of Igo1 at Serine-64 by Rim15 is required for efficient entry into gametogenesis . Phosphorylation of Igo1 at Serine-64 occurs at a constant level during the mitotic cell cycle [24] . To examine the phosphorylation of Igo1 at Serine-64 during entry into gametogenesis , we induced IGO1-myc8 and igo1-S64A-myc8 cells to enter gametogenesis by transferring them to SPM . Analysis of DNA content by flow cytometry indicated that pre-meiotic DNA replication was initiated after 3 hours into SPM and completed by 5 hours in both strains ( Figure S3B ) . We prepared whole cell extracts and analysed electrophoretic mobility of Igo1 by Phos-tag affinity gel electrophoresis and SDS-PAGE . Phos-tag specifically retards the mobility of phosphoproteins [25] . We observed a phos-tag dependent mobility shift of wild type Igo1 but not Igo1-S64A . This upshifted band in wild type cells was present before transfer to SPM and was detectable up to 2 hours after transfer ( Figure S3A ) . As expression of early meiotic genes like Ime1 and Rec8 is detectable even after 1 h in SPM ( Figure 1C ) , we conclude that Igo1 is phosphorylated at S-64 during entry into gametogenesis but dephosphorylated subsequently . Endosulfine contains a conserved protein kinase A site RK/RXS/T at its C-terminus ( Figure S4A ) . Since PKA inhibits entry into gametogenesis , we reasoned that phosphorylation at this site might have an opposite effect to that mediated by Rim15 phosphorylation of S-64 . However replacement of the Serine-105 in Igo1 with alanine or aspartate did not affect sporulation ( Figure S4B ) . Phosphorylated endosulfine promotes entry into mitosis in Xenopus egg extracts by inhibiting the Cdk-antagonizing protein phosphatase PP2A-B55δ [17] , [18] . We have demonstrated that PCLB2CDC55 cells which express CDC55 from the mitosis-specific promoter PCLB2 , fail to undergo meiotic nuclear divisions and form monads [23] . The meiotic nuclear division defect of PCLB2CDC55 cells can be suppressed by net1-6Cdk , a mutant allele encoding the nucleolar protein Net1 lacking 6 Cdk recognition sites [23] . We also noted that PCLB2CDC55 cells underwent pre-meiotic DNA replication earlier than wild type cells [23] suggesting that PP2ACdc55 might negatively regulate entry into gametogenesis . We therefore investigated whether budding yeast proteins Rim15 , endosulfine and PP2ACdc55 regulate entry into gametogenesis and G0 . If PP2ACdc55 and Rim15/endosulfines play opposing roles in entry into gametogenesis and endosulfines promote entry into gametogenesis only by antagonising PP2ACdc55 , we reasoned that inactivation of PP2ACdc55 might suppress the sporulation defect of igo1Δ igo2Δ and rim15Δ cells . While 80% of wild type cells formed spores , only about 10% and 18% of igo1Δ igo2Δ and rim15Δ cells respectively , did . Remarkably igo1Δ igo2Δ and rim15Δ cells carrying a meiotic-null allele of CDC55 ( PCLB2CDC55 ) formed monads ( 75% ) like PCLB2CDC55 cells ( Figure 2A ) . Crucially , combining net1-6Cdk with PCLB2CDC55 igo1Δ igo2Δ and PCLB2CDC55 rim15Δ cells resulted in efficient formation of tetrads ( Figure 2A ) . The ability of PCLB2CDC55 to suppress igo1Δ igo2Δ was specific as deletion of a gene encoding an alternative PP2A regulatory subunit Rts1 had no effect on sporulation efficiency of igo1Δ igo2Δ cells ( Figure 2A ) . To confirm suppression of igo1Δ igo2Δ by PCLB2CDC55 we induced wild type , igo1Δ igo2Δ and igo1Δ igo2Δ PCLB2CDC55 cells to enter meiosis by transferring them to SPM . Wild type cells completed pre-meiotic DNA replication after 4 hours ( Figure 2B ) , and expressed Cdc5 ( a marker for mid-meiosis ) after 7 hours in SPM ( Figure 2C ) . In contrast , igo1Δ igo2Δ cells did not initiate DNA replication ( Figure 2B ) and failed to express Cdc5 even after 12 hours in SPM ( Figure 2C ) . Crucially igo1Δ igo2Δ PCLB2CDC55 cells completed pre-meiotic DNA replication ( 3–4 hours ) and expressed Cdc5 ( 5–6 hours ) ( Figure 2B–C ) . These results indicate that PP2ACdc55 and Rim15/endosulfine play opposing roles in regulating entry into gametogenesis . We then determined whether PP2ACdc55 also negatively regulates entry into quiescence . Wild type , igo1Δ igo2Δ , cdc55Δ and igo1Δ igo2Δ cdc55Δ cells were treated with rapamycin and entry into G0 was monitored by assaying expression of Hsp26 , a gene that is specifically induced during entry into G0 [8] . While wild type cells induced expression of Hsp26 after 2 hours following rapamycin treatment , the igo1Δ igo2Δ cells failed to express Hsp26 ( Figure 2D ) . Crucially , both cdc55Δ cells and igo1Δ igo2Δ cdc55Δ cells expressed Hsp26 even in the absence of rapamycin treatment . These results indicate that the Rim15-endosulfine-PP2ACdc55 pathway regulates entry into gametogenesis and quiescence in budding yeast . To test whether phosphorylation of Igo1 at S64 results in increased association with PP2ACdc55 , we performed an in vitro binding assay . We purified wild type Igo1 , Igo1-S64A and Igo1-S64D from bacterial cells by attaching a Maltose Binding Peptide ( MBP ) to their N-termini . We then incubated endosulfine ( and its variants ) bound to amylose resin via the MBP with yeast extracts containing Cdc55-TAP ( Tandem Affinity Purification ) . Specifically Igo1-S64D but not WT Igo1/Igo1-S64A physically interacted with Cdc55 in vitro ( Figure 3A ) . We then tested whether phosphorylation of wild type endosulfine by Rim15 results in increased association with Cdc55 . For this , we purified either wild type or a Kinase-Dead ( kd ) version of Rim15 ( Rim15-C115A ) from yeast cells using a GST affinity tag . We then incubated either wild type or Igo1-S64A or Igo1-S64D bound to amylose resin with Rim15/Rim15-kd for 45 minutes in the presence of ATP . The phospho-mimetic mutant Igo1-S64D interacted with Cdc55 regardless of whether it was incubated with Rim15/Rim15-kd . Wild type Igo1 , but not Igo1-S64A , incubated with catalytically active Rim15 interacted with Cdc55 ( Figure 3B ) . These results indicate that phosphorylation of Igo1 at S64 promote its association with PP2ACdc55 . We confirmed the phosphorylation of Igo1 at S-64 by Rim15 using a phospho-specific antibody directed against S-64-P which recognized Igo1 incubated with wild-type but not a catalytically dead version of Rim15 ( Figure 3B ) . To test whether phosphorylation of endosulfine at S64 converts it into an inhibitor of PP2ACdc55 , we measured the phosphatase activity of PP2ACdc55 in the presence of Igo1 , Igo1-S64A and Igo1-S64D proteins . We purified PP2ACdc55 by attaching a TAP ( Tandem Affinity Purification ) tag to the C-terminus of Cdc55 ( Figure 3C ) . At first , we measured the phosphatase activity of purified Cdc55 using a phosphorylated peptide as a substrate . Crucially , the TAP eluates from Cdc55-TAP tagged strain but not from an untagged strain , had phosphatase activity ( Figure 3D ) . We then purified MBP-fused versions of Igo1/Igo1-S64D/Igo1-S64A from bacteria and tested their effect on PP2ACdc55 phosphatase activity . Only Igo1-S64D but not Igo1/Igo1-S64A inhibited the phosphatase activity of PP2ACdc55 ( Figure 3E ) . Purified endosulfines had little or no phosphatase activity on their own ( Figure S5 ) . These results show that phosphorylation of endosulfine Igo1 at S64 by Rim15 converts it into an inhibitor of PP2ACdc55 . Endosulfines activated by Rim15 were proposed to protect mRNA involved in stress response from the 5′ to 3′ mRNA decay pathway by direct inhibition of decapping enzyme Dhh1 [8] . Consistent with this possibility , deletion of genes DHH1 and CCR4 , which are required for 5′ to 3′ decay were reported to suppress entry into quiescence defect of igo1Δ igo2Δ cells [8] . We therefore tested whether dhh1Δ and ccr4Δ also suppress the sporulation defect of igo1Δ igo2Δ cells . While PCLB2CDC55 suppressed the sporulation defect of igo1Δ igo2Δ cells , both dhh1Δ and ccr4Δ did not ( Figure 4A ) . We also found that dhh1Δ and ccr4Δ did not suppress the G0 entry defect of igo1Δ igo2Δ cells ( Figure 4B ) contrary to what was previously reported [8] . We do not know the reason for this discrepancy but differences in the strain background used ( BY4741 vs . SK1 ) for the experiments could be an explanation . However , our data are consistent with a simple model which posits that endosulfines regulate entry into quiescence and gametogenesis only through inhibition of PP2ACdc55 . Entry into quiescence is mediated by activation of three master transcription factors namely Msn2 , Msn4 and Gis1 [5] . If the roles of Rim15-Endosulfine-PP2ACdc55 module during entries into quiescence and gametogenesis were identical then one would predict that cells lacking the three G0–specific transcription factors to be also defective in entry into gametogenesis . However we found that msn2Δ msn4Δ gis1Δ cells formed spores ( 60% ) although with defective spore walls ( Figure 5A and data not shown ) . To confirm that the tetrad formation in msn2Δ msn4Δ gis1Δ cells was dependent on endosulfines , we deleted IGO1 and IGO2 in msn2Δ msn4Δ gis1Δ cells . The quintuple mutant cells were defective in forming tetrads like igo1Δ igo2Δ cells ( Figure 5A ) . We confirmed that msn2Δ msn4Δ gis1Δ cells were defective in entry into quiescence ( Figure 5B ) . These results indicate that Rim15-Endosulfine-PP2ACdc55 pathway regulates entry into gametogenesis independently of activation of G0 transcription factors . Ime1 is a master transcription factor for expression of early meiotic genes [9] . As indicated above ( Figure 1D ) Ime1 was not strongly expressed in igo1Δ igo2Δ cells after transfer to SPM . In wild type cells , IME1 is not expressed in glucose-containing nutrient medium but is transcribed at low levels in pre-sporulation medium ( which contains acetate as a carbon source ) and induced further following transfer to SPM [26]–[28] . We tested whether endosulfines are required for this transcriptional induction of IME1 by assaying IME1 mRNA levels by quantitative RT-PCR . In wild type and igo1Δ igo2Δ cells grown in pre-sporulation medium ( which contains acetate as the carbon source ) , the levels of IME1 transcript were around 500-fold higher than in log-phase cells grown in glucose-containing nutrient medium ( Figure 6A ) . However upon transfer to SPM , the IME1 mRNA levels increased further by about 8-fold after 2 hours in wild type but not in igo1Δ igo2Δ cells ( Figure 6A ) . This suggests that endosulfines are required for transcriptional induction of IME1 caused by transfer to SPM . If the only role of endosulfines in entry into gametogenesis was to activate transcription of IME1 , then ectopic expression of IME1 should bypass the sporulation defect of igo1Δ igo2Δ cells . To test this , we constructed wild type and igo1Δ igo2Δ strains in which IME1 expression can be induced by addition of β-estradiol to the medium using the PGAL/Gal4-ER system [29] . We transferred wild type and igo1Δ igo2Δ cells to SPM in the presence or absence of β-estradiol . While wild type cells sporulated in the presence of β-estradiol , igo1Δ igo2Δ cells failed to do so ( Figure 6B ) . Ime1 was expressed in β-estradiol treated igo1Δ igo2Δ cells although at a lower level compared to wild type cells ( Figure 6C ) . Since endosulfines have been implicated in mRNA stability [8] , we tested whether the difference in the Ime1 levels in the two strains was due to difference in the IME1 transcript levels . Quantitative RT-PCR analyses revealed that the IME1 transcript levels were induced to similar extent in wild type and igo1Δ igo2Δ strains and remained relatively unchanged up to 8 hours following induction ( Figure 6D ) . This suggests that endosulfines are not required for regulating IME1 mRNA stability . Decreased Ime1 levels in igo1Δ igo2Δ cells could be caused by either decreased translational efficiency of IME1 mRNA or decreased Ime1 stability . These results indicate that endosulfines promote entry into gametogenesis independently of regulating IME1 expression . Rim15 is required for autophagy induced by inhibition of PKA and Sch9 but not for autophagy induced by rapamycin treatment [30] . Since autophagy is required for spore formation in yeast [31] , we tested whether endosulfines are required for autophagy during entry into gametogenesis . Autophagy can be assayed by following proteolytic cleavage of GFP-Atg8 , which is a N-terminal fusion of GFP to Atg8 ( a ubiquitin-like protein required for formation of autophagosomal membranes ) [32] . We induced wild type , PCLB2CDC55 , igo1Δ igo2Δ and igo1Δ igo2Δ PCLB2CDC55 cells to enter meiosis by transferring them to SPM and assayed autophagy . In wild type cells , GFP-Atg8 underwent proteolytic cleavage after 2 hours into SPM ( Figure 7A ) . In contrast , GFP-Atg8 remained intact in igo1Δ igo2Δ cells even after 12 hours in SPM ( Figure 7A ) . Strikingly , GFP-Atg8 was cleaved earlier in PCLB2CDC55 and igo1Δ igo2Δ PCLB2CDC55 cells in comparison to wild type cells . These results are consistent with the hypothesis that PP2ACdc55 inhibits pre-meiotic autophagy and that this inhibition is overcome by endosulfines after transfer to SPM . We then tested whether endosulfines are required for autophagy induced by rapamycin treatment . We treated wild type , cdc55Δ , igo1Δ igo2Δ and igo1Δ igo2Δ cdc55Δ cells with rapamycin and assayed autophagy by western analysis . While autophagy in cdc55Δ cells was slightly advanced in comparison to wild type cells , endosulfine mutant cells underwent autophagy as efficiently as wild type cells ( Figure 7B ) . We also found that endosulfines were not required for autophagy triggered by nitrogen starvation ( Figure S6 ) . Since rapamycin treatment and nitrogen starvation trigger autophagy by inhibiting TORC1 , our results indicate that endosulfines are not required for autophagy induced by inhibition of TORC1 signalling . Rapamycin treatment of diploid cells induces sporulation [33] . While rapamycin–treated wild type cells formed tetrads after 24 hours , igo1Δ igo2Δ cells did not ( Figure 7C ) . This suggests that induction of autophagy per se is insufficient for rescuing the sporulation defect of endosulfine mutant cells . To determine the role of autophagy in sporulation , we induced wild type and atg1Δ cells ( ATG1 encodes a serine-threonine kinase required for autophagy ) to enter meiosis by transferring them to SPM . Wild type cells completed pre-meiotic DNA replication after 4–5 hours and underwent two rounds of nuclear division to form 45% tetrads ( Figure S7A–B ) . Although the kinetics of Rec8 expression in wild type and atg1Δ cells were similar ( Figure S7B ) , atg1Δ cells were delayed in initiation of pre-meiotic DNA replication by about 1–2 hours in comparison to wild type cells . Expression of Cdc5 ( marker for mid-meiosis ) in atg1Δ cells was delayed by about 3 hours relative to wild type cells ( Figure S7C ) . However atg1Δ cells failed to undergo nuclear divisions and remained largely mononucleate with prophase I spindles after 10 hours in SPM ( Figure S7B and data not shown ) . Since the phenotype of atg1Δ cells is distinct from that of igo1Δ igo2Δ cells ( which fail to enter gametogenesis as indicated in Figure 1 ) , we conclude that endosulfines regulate entry into gametogenesis independently of controlling pre-meiotic autophagy . Ume6 associates the histone deacetylase Sin3/Rpd3 to negatively regulate entry into gametogenesis [11] . Interestingly , Ume6 is phosphorylated during sporulation in a Rim15-dependent manner [34] . If endosulfines promote entry into gametogenesis through inhibition of Ume6 and Sin3/Rpd3 , then ume6Δ and rpd3Δ should suppress igo1Δ igo2Δ . However ume6Δ and rpd3Δ did not suppress the poor sporulation efficiency of igo1Δ igo2Δ cells ( Figure S8 ) . Surprisingly , both ume6Δ and rpd3Δ completely abolished the ability of igo1Δ igo2Δ cells to form tetrads ( Figure S8 ) . This suggests that endosulfines and Ume6/Rpd3 regulate spore formation via independent pathways .
We have shown that a signalling module consisting of a serine-threonine kinase Rim15 , endosulfine Igo1/2 and PP2ACdc55 regulates entry into gametogenesis and quiescence in budding yeast ( Figure 8 ) . While our manuscript was in preparation another group reported that Rim15-Endosulfine-PP2ACdc55 pathway is required for onset of quiescence in yeast cells [35] . We show that the signalling module also regulates entry into gametogenesis via a mechanism that is independent of entry into quiescence . Remarkably both studies show that Rim15-Endosulfine-PP2ACdc55 signalling module in budding yeast mechanistically works like the Greatwall Kinase-endosulfine-PP2A-B55δ pathway that regulates mitotic entry in Xenopus egg extracts [17] , [18] . Entry into quiescence is controlled by activation of the G0 transcription factors Msn2 , Msn4 and Gis1 . We show that entry into gametogenesis is not dependent on the G0 transcription factors suggesting that the Rim15-Endosulfine-PP2ACdc55 regulates entry into G0 and gametogenesis by distinct mechanisms . Precisely how PP2ACdc55 prevents entry into gametogenesis and G0 remains unknown . Although the stress-responsive transcription factor Gis1 is hyperphosphorylated in cdc55 mutant cells [35] , it is not known whether it is a direct substrate of PP2ACdc55 . It is possible that PP2ACdc55 inhibits a factor that is a positive regulator of entry into both gametogenesis and G0 ( Figure 8 ) . Alternatively , PP2ACdc55 might inhibit entry into gametogenesis and G0 by dephosphorylation of distinct substrates ( Figure 8 ) . Comparing the phosphoproteomes of wild type , cdc55 and igo1Δ igo2Δ cells during entry into gametogenesis and quiescence would be illuminating . Contrary to previous observations [8] , we did not find any genetic evidence for endosulfine function in controlling 5′ to 3′ mRNA decay pathway . We suggest that endosulfines regulate entry into both gametogenesis and G0 only via inhibition of PP2ACdc55 ( Figure 8 ) . We have shown that the Rim15-Endosulfine-PP2ACdc55 is required for pre-meiotic autophagy . However the inability to undergo autophagy does not account for the meiotic phenotype of igo1Δ igo2Δ cells as atg1Δ cells ( defective in autophagy ) enter gametogenesis but fail to undergo any nuclear divisions . Induction of autophagy in igo1Δ igo2Δ cells by rapamycin treatment did not rescue the sporulation defect . Expression of Ime1 , the master transcription factor for early meiotic genes , also did not rescue the sporulation defect of igo1Δ igo2Δ cells . These results indicate that the Rim15-Endosulfine-PP2ACdc55 module regulates entry into gametogenesis independently of controlling pre-meiotic autophagy and Ime1 expression . Interestingly Cdc55 has been found to physically interact with Atg1 and Atg18 , two proteins required for autophagy , in interactome screens [36] , [37] . It will be informative to test whether these interactions are altered during entry into gametogenesis . Precisely how Rim15/Gwl phosphorylated endosulfine inhibits PP2ACdc55 activity is not known . Structural analyses of PP2ACdc55-endosulfine complex would be illuminating in this respect . It is also important to determine whether endosulfine inhibits PP2ACdc55 activity towards all or only a specific subset of its physiological substrates . Hypomorphic mutations in CDC55 suppress the dyad phenotype of spo12Δ strains ( Gary William Kerr and Prakash Arumugam , unpublished observations ) . This is consistent with antagonistic roles of Spo12 and PP2ACdc55 in FEAR pathway and exit from meiosis I [23] , [38] , [39] . In contrast to cdc55 hypomorphic alleles , igo1-S64D did not suppress the spo12Δ dyad phenotype ( data not shown ) suggesting that phosphorylated endosulfine inhibits PP2ACdc55 activity towards only some of its cellular substrates . Testing whether endosulfines are required for quiescence and gametogenesis in mammalian cells would be very interesting . Notably , expression of endosulfines was first noted in brains [40] and was decreased several fold in patients with neurodegenerative diseases [41] . Given the high conservation of this signalling module , deconstructing its mechanism in budding yeast might give insights into regulation of mitosis in human cells , and vice versa .
A complete list of yeast strains and their genotypes can be found in Table S1 . The MBP fused wild-type and mutant forms ( S64A and S64D respectively ) of Igo1 were expressed and purified from bacteria using the amylose resin ( NEB ) according to the manufacturer's instructions . Briefly , E . coli cells expressing the MBP fusion proteins were grown overnight and sub-cultured in 2× TY medium containing 0 . 2% glucose and grown at 37°C to an OD600 nm of ∼0 . 5 . IPTG ( Isopropyl β-D-1-thiogalactopyranoside ) was added to the culture to a final concentration of 0 . 2 mM and cells were allowed to grow for another 3 hours at 37°C . Cells were harvested at 4000 rpm for 10 minutes at room temperature and resuspended in 5 ml of buffer A ( 20 mM Tris-Cl , pH 7 . 5 , 250 mM NaCl , 1 mM EDTA , 5 mM β-mercaptoethanol , Roche Complete EDTA-free Protease Inhibitors and 100 mM PMSF ) and stored at −20°C after freezing it in liquid N2 . Cells were thawed in cold water and lysed by sonication ( 40 Amp , 5×15 seconds , 1–2 minutes interval between each pulse ) . Cells were centrifuged at 13 , 200 rpm for 20 minutes at 4°C and supernatant was transferred to separate tubes . Total amount of protein was measured using the Bradford assay [42] . Equal amount of 50% slurry of Amylose resin ( pre-equilibrated in buffer A ) was added to the cell lysate . The mixture was incubated for 20 minutes on ice . Beads were collected at low speed ( 2000 rpm , 1 min , 4°C ) , washed thrice with 1 ml of buffer A . Proteins bound to beads were recovered by elution with maltose ( 10 mM ) or by adding 2× SDS sample buffer to the beads followed by incubation at 95°C for 5 minutes . GST-tagged wild-type or mutant forms of Rim15 was purified from yeast cells . Briefly , cells carrying the plasmids ( encoding either wild-type or mutant Rim15 ) were grown to log phase in SD –URA medium at 30°C containing 2% raffinose . After allowing the cultures to reach an OD600 nm ∼1 . 0 , YPG ( 1% Yeast extract , 2% bactopeptone and 2% galactose ) was added to the culture and grown for another 4 hours at 30°C . Cells were collected , washed in cold water and frozen in liquid N2 and stored at −80°C . Cells were thawed , resuspended in lysis buffer ( 50 mM Tris-Cl , pH 7 . 5 , 100 mM NaCl , 1 mM EDTA , 1% NP-40 , 1 mM PMSF and Roche Complete EDTA-free Protease Inhibitors ) and lysed by using glass beads . Total amount of protein was measured; equal amount of protein was mixed with 150 µl of 50% slurry of GST beads ( pre-equilibrated in lysis buffer ) and rotated at 4°C for 1 hour . Beads were collected and washed once with lysis buffer , twice with lysis buffer +250 mM NaCl and twice with lysis buffer +500 mM NaCl . The GST-fused proteins were eluted using 10 mM reduced glutathione . Yeast cells expressing Cdc55-TAP were grown to log phase at 30°C in YEPD medium , harvested at 4000 rpm for 5 minutes at 4°C . The cell pellets were stored at −80°C after freezing it in liquid N2 . The pellet was thawed , resuspended in yeast lysis buffer ( 50 mM Tris-Cl , pH 7 . 5 , 100 mM NaCl , 1 mM EDTA , 1% NP-40 , 1 mM PMSF and Roche Complete EDTA-free Protease Inhibitors ) and lysed by using glass beads . Protein concentration was measured by Bradford method . MBP fused wild-type and mutant Igo1 proteins were purified as described above . Equal amounts of bead bound proteins were added to equal amounts of total yeast cell extract . The mixture was incubated on ice for 20 minutes . The beads were collected by centrifugation , washed three times with lysis buffer , resuspended in SDS sample buffer , boiled and run on 10% SDS-PAGE . GST fused Rim15 and Rim15-kd proteins were purified as described above and the purified protein was used to phosphorylate purified MBP-fused Igo1 . The reaction was carried out in kinase buffer ( 50 mM Tris-Cl , pH7 . 5 , 20 mM MgCl2 , 1 mM DTT ) containing 1 mM ATP at room temperature for 45 minutes . Beads were collected , mixed with equal amount of yeast cell extract containing Cdc55-TAP and incubated on ice for 20 minutes . The beads were washed three times with lysis buffer , resuspended in 2× SDS sample buffer , boiled and analyzed by Western blotting following SDS-PAGE . Phosphatase assay was carried out using the Ser/Thr phosphatase assay kit containing a phospho-peptide as a substrate ( from Millipore ) . Briefly , strain expressing TAP-tagged Cdc55 was grown in 1 litre of YEPD medium to log-phase . Cells were harvested , resuspended in 5 ml of yeast lysis buffer and soluble extracts were prepared by bead beating . The extract was mixed with 0 . 2 ml of IgG sepharose beads ( pre-equilibrated in lysis buffer ) and the mixture was incubated for 2 hours on a rotary wheel at 4°C . The beads were precipitated , washed 4 times with lysis buffer , once with TEV cleavage buffer ( 10 mM Tris-Cl , pH 7 . 5 , 150 mM NaCl , 0 . 5 mM EDTA , 0 . 1% NP-40 and 1 mM DTT ) and resuspended in 350 µl of TEV cleavage buffer . The bound protein was cleaved and eluted from the beads after incubating overnight at 4°C with 15 U of TEV protease ( Invitrogen ) . The eluted protein was then used for phosphatase assay . Purified Cdc55 was mixed with equal amount ( 25 µg ) of MBP-fused Igo1 or Igo1-S64A or Igo1-S64D or MBP alone and incubated for 20 minutes on ice . The mixture was then incubated with 500 µM of phospho-peptide for 1 hour at 30°C . The reaction was terminated by addition of malachite green solution provided with the kit and absorbance was measured at 620 nm . For in situs , cells from 1 ml of yeast culture were fixed for 15 minutes with 3 . 7% formaldehyde , pelleted and resuspended in 100 mM K-phosphate buffer ( pH 6 . 4 ) containing 3 . 7% formaldehyde and kept overnight on ice . Immunostaining was performed as previously described [23] . The following primary antibodies were used: monoclonal rat anti-α-tubulin 1∶500 ( Serotec ) , monoclonal mouse anti-HA 1∶500 ( Covance ) . Secondary antibodies , pre-absorbed against sera from other species used in labeling , were conjugated with Cy3 or Cy5 ( Chemicon ) and diluted 1∶500 ( Cy3 ) or 1∶50 ( Cy5 ) . DNA was visualized by staining with DAPI . Images were acquired using a Nikon TE-2000 inverted microscope with a 100×1 . 49 N . A . objective lens equipped with a Photometrics Coolsnap-HQ2 liquid cooled CCD camera ( Photometrics , Tucson , AZ ) . 16 Z-stacks ( spacing = 0 . 2 µm ) Exposure times of 1 second were used for both Cy3 and Cy5 , and 0 . 25 seconds for DAPI . Images were analysed using Metamorph ( version 7 . 5 . 2 . 0 MAG Biosystems Software ) . Whole cell extracts were prepared by cell breakage with glass beads in 20% Trichloroacetic acid . Cell pellets were resuspended in 2× SDS sample buffer , neutralized with 2M Tris base and proteins were denatured by heating the samples at 95°C for 5′ . After centrifugation , protein samples were electrophoresed on 10% SDS-PAGE gels . The HA epitope was detected by mouse monoclonal antibody 16B12 at 1∶1000 . Goat anti-Cdc5 ( Santa Cruz SC-6733 ) antibody , Goat anti-Cdc28 ( Santa Cruz-6708 ) antibody , mouse anti-Pk ( Serotec ) antibody , mouse anti-GFP ( Roche ) antibody and rabbit anti-TAP antibody ( Pierce ) were all used at 1∶1000 dilution . Myc epitope was detected using the 9E10 antibody ( Cambridge Biosciences ) at 1∶1000 dilution . Phospho-specific antibody was raised against the phosphorylated synthetic peptide KRKYFDpSGDYALC ( pS indicates phosphoserine ) by Eurogentec . For phos-tag gels , TCA extracts were prepared as above and analysed as previously described [25] with a few modifications . Briefly 12 . 5% polyacrylamide gels were prepared and Phos-Tag ( Wako ) was added at its final concentration of 50 µM to the separating gel mixture before polymerization . Electrophoresis was performed at a constant current of 30 mA at room temp . After electrophoresis , gels were first soaked in Transfer Buffer ( 25 mM Tris , 192 mM Glycine , 10% methanol ) containing 1 mM EDTA for 20 minutes ( 2×10 minutes ) and then in Transfer Buffer for 30 minutes ( 3×10 minutes ) . Electrotransfer onto PVDF membrane was done at a constant voltage of 36 V for 16 hours at 4°C . After running the protein sample on 10% SDS-PAGE , the gel was washed once with water and then fixed with 100 ml of fixative ( 50% methanol and 5% acetic acid ) for 2 hours . The gel was washed once with 100 ml of 20% ethanol and twice with water . The gel was sensitized with 100 ml of 0 . 02% sodium thiosulfate for 1 minute and washed immediately with water . The gel was incubated with 100 ml of silver nitrate ( 0 . 1% in water ) solution containing 20 µl of 37% formaldehyde and kept for 20 minutes at 4°C in dark . The gel was then washed again with water and 100 ml of developing solution ( 2 . 5% sodium carbonate 0 . 0185% formaldehyde ) was added . After the bands were visible , 5% acetic acid was added to terminate the reaction . Total RNA was extracted from yeast cell pellets using the MasterPure Yeast RNA purification kit ( Epicentre ) . RNA integrity was confirmed by agarose gel electrophoretic analysis after denaturation with formamide . Reverse transcription reactions were performed on 0 . 5 µg of DNAase I-treated RNA with Oligo-dT , using the GoScript Reverse Transcription System ( Promega ) . Quantitative real-time PCR primers for analyzing IME1 and ACT1 were designed as previously described [43] and their specificity was confirmed by melt curve analyses . cDNA reactions were diluted 100-fold , and triplicate quantitative real-time PCRs were performed in a Rotor-Gene Q ( Qiagen ) using the 2× Rotor-Gene SYBR Green PCR kit ( Qiagen ) . Reactions were analyzed using RotorGene Q software by the comparative CT method , normalizing IME1 mRNA levels against the ACT1 reference gene . Induction of sporulation was carried out as previously described [44] . To measure sporulation efficiency of yeast strains on solid media , cells were streak purified on YEPD plates . Three single colonies were patched onto YEPD plates . After 24 h of growth at 30°C , cells were patched onto Sporulation plates ( 0 . 82% Sodium acetate , 0 . 19% Potassium chloride , 0 . 035% Magnesium sulphate , 0 . 12% Sodium chloride and 1 . 5% Agar ) and incubated at 30°C for 24 h . Sporulation efficiency was assayed using a light microscope . To induce GAL1-IME1 expression , β-estradiol was added to the cultures at the final concentration of 1 µM . The DNA content of sporulating cells was measured by flow cytometry as previously described [45] . | The fundamental property of a cell is to sense changes in the environment and then respond in a way that maximizes its chances of survival . When diploid budding yeast cells are subjected to complete nutrient starvation they have two possible fates , namely quiescence and gametogenesis . Quiescent cells have reduced rates of transcription and translation and increased stress tolerance . Gametogenesis results in production of haploid spores that can survive for long periods of time . In this paper , we report a signalling module that regulates entry into both quiescence and gametogenesis in budding yeast . The module consists of three molecular components namely a serine-threonine kinase Rim15 , a phosphatase PP2ACdc55 and a conserved protein called as endosulfine . PP2ACdc55 negatively regulates entry into gametogenesis and quiescence . Upon nutrient starvation , Rim15 becomes active and phosphorylates endosulfine . This converts endosulfine to an inhibitor of PP2ACdc55 and thereby leading to entry into quiescence and gametogenesis . Remarkably , an analogous module consisting of Greatwall kinase , PP2A-B55δ and endosulfine regulates entry into mitosis in frog egg extracts and meiotic maturation in flies suggesting that this signalling module is highly conserved and co-opted during evolution to control distinct biological processes in different organisms . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biology",
"and",
"life",
"sciences"
] | 2014 | The Rim15-Endosulfine-PP2ACdc55 Signalling Module Regulates Entry into Gametogenesis and Quiescence via Distinct Mechanisms in Budding Yeast |
The Escherichia coli chemotaxis network is a model system for biological signal processing . In E . coli , transmembrane receptors responsible for signal transduction assemble into large clusters containing several thousand proteins . These sensory clusters have been observed at cell poles and future division sites . Despite extensive study , it remains unclear how chemotaxis clusters form , what controls cluster size and density , and how the cellular location of clusters is robustly maintained in growing and dividing cells . Here , we use photoactivated localization microscopy ( PALM ) to map the cellular locations of three proteins central to bacterial chemotaxis ( the Tar receptor , CheY , and CheW ) with a precision of 15 nm . We find that cluster sizes are approximately exponentially distributed , with no characteristic cluster size . One-third of Tar receptors are part of smaller lateral clusters and not of the large polar clusters . Analysis of the relative cellular locations of 1 . 1 million individual proteins ( from 326 cells ) suggests that clusters form via stochastic self-assembly . The super-resolution PALM maps of E . coli receptors support the notion that stochastic self-assembly can create and maintain approximately periodic structures in biological membranes , without direct cytoskeletal involvement or active transport .
Efficient biological signal processing often requires complex spatial organization of the signaling machinery . Understanding how this spatial organization is generated , maintained , and repaired inside cells is a fundamental theme of biology . A well-understood signaling network with complex spatial organization is the bacterial chemotaxis system , which directs the movement of cells towards or away from sugars , amino acids , and many other soluble molecules [1] . In Escherichia coli , five types of transmembrane chemoreceptors form trimers of dimers [2] , [3] , which cluster into large complexes containing tens of thousands of proteins [4]–[7] . Receptor clustering enables cooperative interactions between receptors [8]–[11] , contributing to a bacterium's ability to sense nanomolar concentrations of chemicals and small fractional changes in chemical concentrations over a wide range [12]–[14] . Chemotaxis clusters are stabilized by the adaptor protein CheW and the histidine kinase CheA , which bind receptors in a ternary complex . CheA transduces signals from membrane receptors to the cytoplasmic response regulator CheY , which diffuses to flagellar motors and modulates their direction of rotation ( Figure 1A; for review see [5] ) . A variety of imaging studies have advanced our understanding of how the spatial organization of the chemotaxis network arises and contributes to function [15] . Time-lapse fluorescence microscopy suggests that receptors are inserted randomly into the lateral membrane via the general protein translocation machinery and then diffuse to existing clusters [16] . Immunoelectron and fluorescence microscopy have shown that receptor clusters are found at the cell poles [4] and future division sites [17] . Despite much research , the fundamental mechanisms responsible for positioning chemotaxis clusters at specific sites in the membrane remain unclear [15] . Perhaps cells possess intracellular structures that anchor clusters to periodic sites along cell length [17] . However , fluorescence microscopy of cells overexpressing all chemotaxis proteins showed that the number of clusters per cell saturates well below the number of proposed cluster anchoring sites . Furthermore , the distance between chemotaxis clusters varies broadly within cells [18] . Based on those observations , Thiem and Sourjik [18] proposed that cluster nucleation and growth is a stochastic self-assembly process in which receptors freely diffuse in the membrane and then join existing clusters or nucleate new clusters . In their model , clusters nucleate anywhere in the membrane and later become attached to anchoring sites . Shortly thereafter , it was reported that anchoring sites may not be required for periodic positioning; surprisingly , simulations reveal that periodic positioning of clusters can emerge spontaneously in growing cells [19] . Direct tests of these stochastic nucleation models involve measuring , as accurately as possible , the relative spatial positioning of clusters and the distribution of cluster sizes . This requires ( 1 ) the high specificity of genetically encoded fluorescent tags and ( 2 ) spatial resolutions sufficient to count and localize single proteins , even when these proteins are densely packed . Electron microscopy has the required spatial resolution , but the density of immunogold labeling is too low to visualize a significant fraction of receptors [4] . Cryo-electron microscopy tomography has provided detailed information on large polar clusters [20] , [21] , but identification of individual receptors is not yet possible . Fluorescence microscopy does not have the required spatial resolution to observe individual receptors in dense clusters . Single-cell Förster resonance energy transfer ( FRET ) studies have been instrumental in measuring the dynamics of signaling within the chemotaxis network [12] , [13] , but cannot obtain the distribution of receptors inside cells . The optical super-resolution technique photoactivated localization microscopy ( PALM ) combines high specificity with high resolution . In PALM , target proteins are genetically labeled with photoactivatable proteins , thus rendering them nonfluorescent until activated by near-UV light . By employing near-UV light of sufficiently low intensity , only one protein per diffraction-limited region ( ∼250 nm ) is activated at a time . Following activation , each individual protein is then excited and imaged . Since only one protein is imaged at a time in each diffraction-limited region , the center of each molecular point spread function indicates the location of each protein [22] . Serial cycles of activation and excitation are repeated until all fusion proteins are bleached . Since individual proteins are imaged , we can count the number of proteins and computationally assemble the locations of all proteins into a composite , high-resolution image . The location of each protein can be determined to a precision of 2–25 nm , or approximately 10–100× better than the diffraction limit [23]–[26] . The localization error in each protein location depends on the number of photons collected for that protein , as well as background noise , pixel size , sample drift , and whether cells are live or chemically fixed [22] , [23] , [26] . The highest optical resolution is obtained with chemically fixed cells [23] . Several other optical techniques , including FPALM [27] , STORM [28] , [29] , STED [30]–[32] , and SSIM [33] , also image below the diffraction limit . Here , we use PALM images to directly test stochastic nucleation models of chemotaxis cluster self-assembly in E . coli . We show that many receptors are part of small clusters not previously observed in electron microscopy or fluorescence microscopy , and that these small clusters provide direct evidence for a stochastic nucleation mechanism without anchoring sites .
Three main components of the bacterial chemotaxis network ( Figure 1A ) were visualized by constructing photoactivatable fluorescent protein fusions to Tar , CheW , and CheY ( Figure 1A , zoom ) . Tar is the high-abundance aspartate receptor and makes up 30%–45% of all receptors [34] . CheW is the adaptor protein , which binds all five types of chemotaxis receptors with variable stoichiometry . CheY is the chemotactic response regulator , which transduces signals from the receptors to flagellar motors . All fusion proteins were expressed from plasmids in strains lacking a genomic copy of the protein ( Δtar cells , ΔcheW cells , or ΔcheY cells ) and are therefore nonchemotactic unless complemented ( with Tar , CheW , or CheY , respectively ) . Labeling several distinct components of the network and comparing their localization patterns ensures that there are no confounding effects of our tags on clustering . All cells were cultured in H1 , which is a defined minimal salts medium [35] that has been extensively characterized for its effects on chemotaxis protein expression [34] . To understand how the distribution of cluster sizes may arise from a simple stochastic nucleation mechanism , we extended the cluster growth model of Wang et al . [19] . According to their model , receptors are inserted into the membrane at random locations and then diffuse until they are captured by a preexisting cluster or they nucleate a new cluster . The growth of a specific cluster depends on competition for receptors with nearby clusters . In our model , we treat the competing clusters as an absorbing barrier a distance R away from a preexisting cluster of radius a , which is also absorbing ( Figure S9 ) . The rate of growth of a cluster is given by , which depends only on R , a , and γ , the deposition rate of the receptors into the membrane . Integrating relates the size of a cluster with its age tage . In an exponentially growing population of cells , the ages of the clusters will be exponentially distributed according to , where 1/τ is the growth rate of the cells . Assuming that receptors diffuse freely in the membrane , but clusters are stationary , we predict that the probability of a cell containing a cluster of size N is ( 1 ) where we have defined the constants α = πγR2 and β = 2ln ( R ) −ln ( ΔA/π ) , and where N is the number of receptors ( or receptor dimers ) , N0 is the number of receptors at nucleation , and ΔA is the area per receptor ( Text S1 ) . In the cell membrane , small clusters would be expected to diffuse and occasionally combine with other clusters . To account for this attrition of small clusters , we modify P ( N ) by multiplying it by a survival probability , such that the total probability of observing a cluster of size N receptors is Ptot ( N ) = P ( N ) Psurv ( N ) . We calculate the survival probability to be: ( 2 ) where η is the viscosity of the membrane , h is the thickness of the membrane , and c is a constant set by the dimensions of the cell and the area per receptor ( Text S1 ) . Combining Equations 1 and 2 results in an expression with the functional form: ( 3 ) which we use to fit our cluster-size distributions with free parameters c2 , c3 , and c4 . Normalizing each cluster-size distribution fixes the constant c1 . Equation 3 fits our observed cluster-size distribution well ( Figure 4A and 4B , red line ) , consistent with a stochastic cluster growth and nucleation mechanism . To evaluate the fit of our model to our data in a bin-independent manner , we compared the cumulative distribution function ( CDF ) of our cluster-size distribution with the CDF of our model ( Figure 4A and 4B , insets ) . Importantly , our cluster growth model does not invoke cluster anchoring to cytoskeletal or predivisional structures , nor does it require active transport of receptors or clusters . To provide further , independent , support that receptors stochastically self-assemble into clusters , we analyzed another aspect of the data . In our model , proteins that happen to be inserted close to existing clusters will be absorbed by them , whereas those inserted far from existing clusters will nucleate new clusters [19] . Thus , our model predicts that the highest density of small clusters will be found predominantly at sites that are furthest from all existing large clusters . We identified cells with one or two large polar clusters ( ≥400 proteins ) and measured the locations of small clusters ( <400 proteins ) within these cells . As predicted by our model , cells with one large polar cluster have the highest remaining cluster density at the opposite end of the cell ( Figure 4C ) . Moreover , cells with two polar clusters have the highest cluster density in the middle of the cell , furthest from the two large clusters ( Figure 4D ) . To ensure these results were not affected by our definition of clusters , we performed a similar analysis with receptor density . Cells with two polar clusters have significantly higher receptor density in the middle of the cells ( Figure 4F , arrow ) in comparison to cells with only one polar cluster ( Figure 4E ) ( two-sample Kolmogorov-Smirnov test , p = 0 . 00013 , n = 9 , 115 , 19 , 967 proteins , middle 25% of cell length ) . These results are robust to changes in the specific size cutoff for large polar clusters . Our data and modeling ( Figure 4A–4F ) directly support a stochastic nucleation mechanism of cluster assembly and positioning . In addition to explaining how the exponential distribution of cluster sizes arises , the model also sheds light on the mechanism for spatial self-organization along cell length , in the particular manner shown in Figure 4G and also detected by diffraction-limited imaging [17] . As cells grow , new clusters form primarily at locations that are furthest from large existing clusters . This is because the density of solitary receptors ( or receptor dimers ) is highest in regions furthest from existing clusters . A cell with one polar cluster will tend to form the next large cluster at the opposite pole , yielding a cell with clusters at both poles . A cell with clusters at both poles will tend to form new clusters at the cell midline , the location furthest from both poles . In addition to generating clusters , receptor self-assembly may maintain and repair the location of clusters inside cells . In the event that a daughter cell begins without a large cluster , the first new cluster will form at a random location , but subsequent clusters nucleate furthest from that first cluster , at one of the cell poles . Furthermore , new membrane and cell wall are inserted into lateral regions of the cell [44] , so that cell growth and division will eventually reposition lateral clusters at the cell poles . In this way , cells that begin without clusters will generate periodic positioning of new clusters along cell length as well as the particular exponential distribution of cluster sizes detected by PALM ( Figure 4A and 4B ) . The mechanism of stochastic cluster formation allows cells to recover from the loss of all clusters , as well as begin to correctly position clusters soon after growth in new media . We note that our model does not address the reported difference in diffusion rates between polar and lateral clusters [17] . It is possible that the difference in membrane curvature or membrane composition between polar and lateral regions affects cluster diffusion or cluster dynamics . There may be multiple advantages to arranging a fixed number of receptors among a variety of cluster sizes , such as fine-tuning of signal processing [45] . Our PALM images of receptors are reminiscent of the model of Berg and Purcell [46] , who theorized that for optimum detection sensitivity , membrane receptors should be dispersed widely over the surface of the cell rather than concentrated in one location . In addition , recent in vitro data suggest that different densities of receptors have different kinase and methylation rates [47] , suggesting that the chemotaxis network may adjust its kinase activity based on the local concentration of receptors . Recent in vitro evidence shows that purified membrane-associated proteins can spontaneously self-assemble into complex , dynamic structures [48] , [49] . Our super-resolution PALM maps of E . coli receptors support this notion that stochastic self-assembly can create and maintain dynamic patterns in biological membranes , without direct cytoskeletal involvement or active transport . Perhaps stochastic self-assembly is the simplest mechanism to produce robust patterns in membranes without additional machinery . Our model may apply to clustering of other proteins and to chemotaxis receptors in other organisms; however , many details are expected to be organism-specific . Analysis of super-resolution images similar to those presented here will allow counting of proteins and complexes in individual cells , reveal new levels of cell organization , and allow mechanistic hypotheses to be directly tested .
All strains are derivatives of RP437 , a chemotactic wild-type E . coli K-12 strain . Each chemotaxis protein was expressed in a strain lacking the genomic copy of that protein . All proteins were expressed from the inducible trc promoter on the medium-copy plasmid pTrc-His2 ( Invitrogen ) containing a pBR322-derived origin , the ampicillin resistance gene ( bla ) , and the lac repressor gene ( lacIq ) . The receptor knockout strain is HCB436 [50] , which lacks all chemoreceptors except Aer and also lacks the adaptation enzymes CheR and CheB . pALM5000 contains the tandem dimer Eos ( tdEos ) gene only , and pALM6000 contains the monomer Eos ( mEos ) gene only . pALM5001 , pALM5003 , and pALM6001 contain tdEos-cheW , cheY-tdEos , and tar-mEos gene fusions , respectively . Eos is a photoconvertible protein that irreversibly switches its peak emission from green ( 516 nm ) to red ( 581 nm ) upon exposure to near-UV light [36] . Eos consists of 226 amino acids with a molecular mass of 26 kDa . Tandem dimer Eos ( tdEos ) consists of two copies of wild-type Eos [36] connected by a 15-residue linker SRGHGTGSTGSGSSE ( nucleotide sequence TCTCGAGGTCACGGTACTGGTTCTACTGGTTCTGGTTCTTCTGAG ) . Monomer Eos ( mEos ) is the improved monomeric photostable “mEos2” from McKinney et al . [38] . The tandem dimer Eos ( tdEos ) gene on plasmid pALM5000 is followed by the residues ENSGS ( nucleotides GAGAATTCGGGATCC ) containing a BamHI site . The tdEos-cheW gene on plasmid pALM5001 consists of tdEos , a five-residue linker ( ENSGS ) , the entire cheW gene ( residues 1–167 ) , and a terminal Gly-Ser encoding a BamHI site . The cheY-tdEos gene on plasmid pALM5003 consists of the entire cheY gene ( residues 1–129 ) , a one-residue linker Ala encoding part of a NcoI site , tdEos , and ENSGS . The monomer Eos gene ( mEos ) on plasmid pALM6000 is followed by Gly-Ser . The tar-mEos gene on plasmid pALM6001 consists of the entire tar gene ( residues 1–553 ) joined to mEos with no linker , followed by Gly-Ser . The tar gene second codon was mutated from ATT ( Ile ) to GTA ( Val ) to introduce a NcoI site . Plasmid pALM5000 was constructed by PCR amplification of the tandem dimer Eos gene from the plasmid ptdEos-Vinculin [26] using primers 5′-ACCATGGTGGCGATTAAGC-3′ and 5′-TTAGGATCCCGAATTCTCTCGTCTGGCATTGTC-3′ containing underlined NcoI and BamHI sites , respectively . This PCR product was inserted into plasmid pTrc-His2 ( Invitrogen ) according to the manufacturer's instructions . The N-terminal plasmid leader sequence was removed by digestion with NcoI and religation . pALM5001 ( tdEos-CheW ) was constructed by PCR amplification of cheW from strain RP437 using primers 5′-AAAGGTGGATCCATGACCGGTATGACGAATGTAAC-3′ and 5′-TCGGGAGGATCCCGCCACTTCTGACG-3′ , and cloned into the BamHI site of pALM5000 , immediately after the tdEos gene . pALM5003 ( CheY-tdEos ) was constructed by PCR amplification of cheY from strain RP437 using primers 5′-AGTGTGCCATGGCGGATAAAG-3′ and 5′-AGTCGCCCATGGCCATGCCCAGTTTC-3′ , and cloned into the NcoI site in pALM5000 , immediately before the tdEos gene . pALM6000 was constructed by PCR amplification of the monomeric Eos gene from plasmid pRSETa_mEos2 ( Addgene plasmid 20341 ) using primers 5′-GGATCCATGGGGGCGATTAAGCCAGAC-3′ and 5′-CAAGCTTCTTAGGATCCTCGTCTGGCATTGTCAGGC-3′ containing underlined NcoI and BamHI sites , respectively . This PCR product was cloned into pALM5000 , replacing tdEos with mEos . pALM6001 ( Tar-mEos ) was constructed by cloning a 2 , 345-bp synthesized DNA ( DNA 2 . 0 ) into the NcoI and BamHI sites of pALM5000 , replacing tdEos with tar-mEos . The synthetic DNA coded for the tar gene of wild-type strain MG1655 immediately followed by the monomer Eos gene , and the entire sequence was flanked by appropriate restriction sites . These restriction sites added a terminal Gly-Ser to the Eos gene . RP437 ΔcheW and RP437 Δtar were made by P1 transduction from the Keio collection strains JW1876 ( ΔcheW::kan ) and JW1875 ( Δtar::kan ) , respectively . The deletions in these strains were constructed to minimize polar effects on downstream gene expression by retaining the native start codon and the last 18 C-terminal nucleotides [51] . When cured of kanamycin resistance , the Keio deletion strains retain a translatable scar sequence in-frame with the deleted gene initiation codon and its C-terminal 18-nucleotide coding region . This scar sequence is expected to produce a 34-residue scar peptide with an N-terminal Met , 27 scar-specific residues , and six C-terminal gene-specific residues . RP437 ΔcheY was made according to Datsenko and Wanner [52] , using primers that exactly removed the entire cheY gene and replaced it with a 1 . 1-kb DNA from pKD3 encoding the chloramphenicol resistance gene . Strains were cured of resistances using plasmid pCP20 as described in Cherepanov and Wackernagel [53] . Tryptone broth ( T-broth ) contains 1% w/v Difco Bacto-Tryptone ( Becton Dickinson and Company ) , and 0 . 5% w/v NaCl ( Fisher-Scientific ) ( pH 7 . 0 ) . H1 minimal medium [35] contains 100 mM potassium phosphate ( pH 7 . 0 ) ( 11 . 2 g/l K2HPO4 anhydrous , 4 . 8 g/l KH2PO4 ) , 15 mM ( NH4 ) 2SO4 , 1 mM MgSO4 , 2 µM Fe2 ( SO4 ) 3 , with 0 . 5% glycerol and 1 mM required amino acids ( histidine , leucine , methionine , and threonine ) . Minimal phosphate medium [54] contains 10 mM potassium phosphate ( pH 7 . 0 ) , 1 mM ( NH4 ) 2SO4 , 1 mM MgSO4 , 1 mM glycerol , and 0 . 1 mM required amino acids . Media were supplemented with 50 µg/ml ampicillin ( Shelton Scientific ) . Cultures were grown overnight in T-broth at 30°C with aeration . Day cultures were inoculated to an optical density at 600 nm ( OD600 ) of approximately 0 . 01 into H1 minimal medium with appropriate antibiotics at 30°C with aeration until they reached an OD600 0 . 1–0 . 3 . Protein expression , when indicated , was induced by adding 10 µM IPTG for 3 h . Media and temperature were chosen to obtain the highest expression levels of properly folded proteins [36] , [55] . To determine functionality of chemotaxis fusion proteins , 2 µl of stationary-phase cells were spotted on soft-agar swarm plates and incubated at 30°C for 16–18 h . Wild-type cells containing cytoplasmic Eos ( positive control ) were compared with appropriate deletion strains containing cytoplasmic Eos ( negative control ) and deletion strains with Eos-tagged chemotaxis fusions ( cells used for imaging ) . All cells contain plasmids derived from pTrc-His2 , which confers ampicillin resistance . Swarm plates contain 0 . 3% agar ( Becton-Dickinson ) in 10 mM minimal phosphate medium ( or H1 medium ) supplemented with 100 µM aspartate , 50 µg/ml ampicillin , and varying concentrations of IPTG . Aspartate was added to the plates to ensure that complemented mutants display chemotaxis toward aspartate , since RP437 Δtar still contains the remaining four receptors that are capable of chemotaxis toward serine and oxygen . Cells were grown in tryptone broth with ampicillin at 30°C prior to spotting on swarm plates . Sapphire coverslips , used for their high refractive index ( Olympus APO100X-CG ) , were placed in a 5∶1∶1 solution of Milli-Q filtered water , ammonium hydroxide , and hydrogen peroxide overnight at 75°C . The coverslips were subsequently rinsed with filtered water , sonicated in acetone for 20 min , rinsed again with water , rinsed with methanol , dried quickly under air flow , passed through a flame , and then stored dry until use . Clean sapphire coverslips were covered in 0 . 05% w/v poly-l-lysine for 30 min then rinsed with water . Cells were added and allowed to settle for 30 min at room temperature in the dark or spun onto coverslips at 2 , 000g for 10 min . Cells were fixed with 4% paraformaldehyde in 10 mM PBS ( pH 7 . 4 ) for 15 min at room temperature . Fixative solution was prepared daily by mixing 0 . 8 g of paraformaldehyde , 18 ml of water , and 20 µl of 10 N NaOH , then dissolved by heating to approximately 50°C for several minutes with stirring , buffered to pH 7 . 4 with the addition of 2 ml of 10× PBS solution and 140 µl of 1 N HCl , and finally filtered . After fixation , cells were rinsed with PBS . To compensate for drift during imaging [26] , a 40× dilution of 40-nm and 100-nm Au beads ( Microspheres-Nanospheres , 790114-010 and 790122-010 ) were added . PALM imaging was performed according to Shroff et al . [24] on an Olympus IX81 inverted microscope equipped with DIC optics and a 100× , 1 . 65 NA objective . Laser light was delivered to the microscope through free space from a platform where 405-nm , 488-nm , and 561-nm lasers were combined . Single-molecule tdEos and mEos fluorescence signals generated during acquisition were separated from the activation and excitation light using appropriate filter sets [24] within the microscope and passed to an electron-multiplying charge-coupled device ( CCD ) camera running at approximately 20 Hz ( 50-ms exposures ) . Movie acquisition times were dependent on the regions of highest labeled-protein density , which are the largest chemotaxis clusters . Activation intensity was increased slowly such that a given diffraction-limited spot contained no activated proteins >90% of the time . This is necessary to ensure that only one protein is activated at a time in a single diffraction-limited spot . Image generation and data analysis were done using custom Matlab scripts ( Mathworks ) . Acquisition times were 30–180 min for TIR , and 90–240 min for epi-illumination . Localization and image-rendering algorithms have been described [26] . Briefly , images were filtered and proteins were identified as signals that contained counts larger than four standard deviations above background . Proteins that became dark , but reappeared within five frames , were counted as the same protein . Only proteins that emitted at least 100 photons and had localization errors less than 40 nm were counted , and these thresholds were chosen to maximize the signal to noise for our images and minimize false positives ( Figure S7 ) . Sample drift was corrected by tracking the motion of fiducial nanoparticles , which were localized at approximately 1 Hz to better than 1 nm precision ( Figure S2 ) . Images from the TIR , epi , DIC , and brightfield channels were aligned by recording the position of fiducial nanoparticles common to all channels . All epi-PALM images were rendered with the “hot” colormap in Matlab that varies smoothly from black through shades of red , orange , and yellow to white , and TIR-PALM images were rendered with a variation of the same colormap with red and blue channels switched . Parameters used to acquire PALM data and render images are shown in Table S1 . | Cells arrange their components—proteins , lipids , and nucleic acids—in organized and reproducible ways to optimize the activities of these components and , therefore , to improve cell efficiency and survival . Eukaryotic cells have a complex arrangement of subcellular structures such as membrane-bound organelles and cytoskeletal transport systems . However , subcellular organization is also important in prokaryotic cells , including rod-shaped bacteria such as E . coli , most of which lack such well-developed systems of organelles and motor proteins for transporting cellular cargoes . In fact , it has remained somewhat mysterious how bacteria are able to organize and spatially segregate their interiors . The E . coli chemotaxis network , a system important for the bacterial response to environmental cues , is one of the best-understood biological signal transduction pathways and serves as a useful model for studying bacterial spatial organization because its components display a nonrandom , periodic distribution in mature cells . Chemotaxis receptors aggregate and cluster into large sensory complexes that localize to the poles of bacteria . To understand how these clusters form and what controls their size and density , we use ultrahigh-resolution light microscopy , called photoactivated localization microscopy ( PALM ) , to visualize individual chemoreceptors in single E . coli cells . From these high-resolution images , we determined that receptors are not actively distributed or attached to specific locations in cells . Instead , we show that random receptor diffusion and receptor–receptor interactions are sufficient to generate the observed complex , ordered pattern . This simple mechanism , termed stochastic self-assembly , may prove to be widespread in both prokaryotic and eukaryotic cells . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"microbiology/microbial",
"growth",
"and",
"development",
"biophysics/macromolecular",
"assemblies",
"and",
"machines"
] | 2009 | Self-Organization of the Escherichia coli Chemotaxis Network Imaged with Super-Resolution Light Microscopy |
Cellular efficiency in protein translation is an important fitness determinant in rapidly growing organisms . It is widely believed that synonymous codons are translated with unequal speeds and that translational efficiency is maximized by the exclusive use of rapidly translated codons . Here we estimate the in vivo translational speeds of all sense codons from the budding yeast Saccharomyces cerevisiae . Surprisingly , preferentially used codons are not translated faster than unpreferred ones . We hypothesize that this phenomenon is a result of codon usage in proportion to cognate tRNA concentrations , the optimal strategy in enhancing translational efficiency under tRNA shortage . Our predicted codon–tRNA balance is indeed observed from all model eukaryotes examined , and its impact on translational efficiency is further validated experimentally . Our study reveals a previously unsuspected mechanism by which unequal codon usage increases translational efficiency , demonstrates widespread natural selection for translational efficiency , and offers new strategies to improve synthetic biology .
Eighteen of the 20 amino acids are each encoded by two or more synonymous codons in the standard genetic code , yet the synonymous codons are often used unequally in a genome . Such codon usage bias ( CUB ) has been extensively documented in all three domains of life [1]–[3] . Within a genome , highly expressed genes tend to have stronger CUB than lowly expressed ones [4] , and the codons preferentially used in highly expressed genes of a species are referred to as preferred codons . Although codon usage is clearly determined by the joint actions of mutation , drift , and selection [5]–[6] , the fitness benefit of CUB is less clear . There are two prevailing , non-mutually exclusive , hypotheses on the selective utility of CUB: accuracy and efficiency of protein translation [6] . The translational accuracy hypothesis asserts that different synonymous codons have different probabilities of mistranslation , and that the use of accurately translated codons is beneficial because mistranslation reduces the number of functional molecules , wastes energy , and/or induces cytotoxic protein misfolding . Unequivocal evidence for this hypothesis exists [7]–[10] . By contrast , the translational efficiency hypothesis lacks direct evidence . This hypothesis holds that different synonymous codons are translated at different speeds , and that faster translation is beneficial because it minimizes ribosome sequestering and so helps alleviate ribosome shortage [5] , [11]–[12] . The relevance of ribosome shortage is evident from the findings that most ribosomes are actively engaged in translation during rapid cell growth [13]–[14] and that ribosome concentration increases with the rate of cell growth [15] . An important observation invoked to support the efficiency hypothesis is that cognate tRNAs of preferred codons tend to have higher cellular concentrations ( or more gene copies ) than those of unpreferred codons [4] , [16] , which may allow faster translation of preferred codons than unpreferred codons . While results from several earlier studies are consistent with this hypothesis [12] , [17] , these studies do not exclude the possibility that the observed differences in activity or fitness caused by synonymous mutations are entirely due to CUB's influence on translational accuracy ( see Discussion ) . Here we directly test the efficiency hypothesis and its presumed underlying mechanism .
The translational efficiency hypothesis assumes that synonymous codons have different translational speeds , caused by disparities in codon selection time ( CST ) , the time needed for ribosomal A site to find the cognate ternary complex of aminoacylated tRNA+eEF-1α+GTP . To test this proposition , we took advantage of a genome-wide ribosome profiling study of Saccharomyces cerevisiae that surveyed ribosome-protected mRNA fragments at a nucleotide resolution in a cell population at a given moment by Illumina deep sequencing [18] . Because the probability that a codon is docked at the A site is proportional to its CST , we estimated the relative CSTs of all 61 sense codons ( Figure 1A ) by the ratio of the observed codon frequencies at the A site in the ribosome profiling data and the expected codon frequencies estimated from mRNA-Seq data generated under the same condition in the same experiment ( Figures S1 , S2 , S3; see Materials and Methods ) . The standard errors of the CST estimates , measured by bootstrapping genes from the original datasets , are on average 12% of the CST estimates ( Figure 1A ) , indicating that our CST estimates are overall quite precise . CUB is commonly measured by the relative synonymous codon usage ( RSCU ) , defined by the frequency of a codon relative to the average frequency of all of its synonymous codons in a set of highly expressed genes [19] . To compare the usage of all 61 sense codons , we also use RSCU' , which is the proportion of use of a given codon among synonymous choices in a set of highly expressed genes ( see Materials and Methods ) . Another commonly used measure of CUB is the codon adaptation index ( CAI ) [20] , which is calculated for a gene , and measures its usage of high-RSCU codons ( see Materials and Methods ) . The greater the CAI , the more prevalent are preferred codons in the gene . Contrary to the widely held presumption that preferred codons are translated faster than unpreferred codons , no significant negative correlation between RSCU' and CST was observed among the 61 sense codons ( Figure 1B ) . It is also believed that codons with abundant cognate tRNAs tend to have low CSTs . Because tRNA gene copy number and tRNA concentration are highly positively correlated [21]–[22] , the former is often used as a proxy of the latter . However , neither tRNA gene copy number ( Figure 1C ) nor tRNA concentration ( Figure 1D ) correlates negatively with CST . Because codons and tRNAs do not have one-to-one correspondence , in the foregoing analysis , we considered the best-matching tRNA species for each codon . This codon-tRNA relationship has been shown to be more accurate than the wobble rule , at least in yeast [22] . We also examined each amino acid separately . Among the 18 amino acids with at least two codons , 12 ( Ala , Asn , Cys , Gln , Glu , Gly , Ile , Lys , Ser , Thr , Tyr , and Val ) showed a negative correlation between RSCU' and CST , while 6 ( Arg , Asp , His , Leu Phe , and Pro ) showed a positive correlation , when statistical significance of the correlation was not required ( Figure 1A ) . The number of negative correlations is not significantly more than the chance expectation of 9 ( P = 0 . 12 , one-tail sign test ) . Using the standard errors of the CST estimates for the foregoing 18 amino acids ( Figure 1A ) , we tested whether the CSTs are significantly different between the synonymous codon with the highest RSCU' and that with the lowest RSCU' . After the control for multiple testing by the Bonferroni correction , only two amino acids showed significant differences . The highest-RSCU' codon has a lower CST than the lowest-RSCU' codon for glycine ( nominal P = 0 . 002 ) , while the opposite is true for arginine ( nominal P<0 . 001 ) . Our results are robust to different multiple-testing corrections , as no other amino acids show a nominal P<0 . 01 . Furthermore , when RSCU' is not considered , arginine is the only amino acid for which synonymous codons show significant heterogeneity in CST at the 5% significance level after the correction for multiple testing . Following an earlier study [1] , we also tried defining preferred codons without using gene expression data , but the results are not different ( Figure S4 ) . The overall lack of a significant negative correlation between CST and synonymous codon usage is real rather than an artifact of imprecise CST estimation , because the standard errors of CSTs are quite small ( Figure 1A ) and CSTs of several nonsynonymous codons differ significantly from one another ( see below ) . To validate the above findings , we also directly compared RSCU' values of individual codon positions of Illumina reads from the ribosome profiling data , without estimating CSTs . If unpreferred codons are translated more slowly and therefore stay at the ribosomal A site longer than preferred codons , codons at the A site should have a lower RSCU' on average than its neighboring sites of the same read , after the correction of sequencing bias by mRNA-Seq data . However , we observed no dip in RSCU' at the A site ( Figure 1E ) . We further calculated , within each gene , the ratio between the frequency of preferred codons and that of unpreferred codons at the ribosome A site of Illumina reads from the ribosome profiling data , after correction by mRNA-Seq . This ratio is expected to be 1 if preferred and unpreferred codons are translated equally fast . Indeed , after combining the ratio for all amino acids and all genes using the Mantel-Haenszel procedure [23] , we found the overall ratio to be 0 . 984 , not significantly different from 1 ( P = 0 . 21 , two-tail χ2 test ) . The above findings are puzzling , because the first step in the interaction between tRNA and mRNA is non-specific [24] and the relative waiting time for the cognate tRNA to arrive at the ribosome A site is expected to be inversely proportional to the relative concentration of the cognate tRNA . It was also reported that CST is the rate-limiting step in translational elongation [25] . The only plausible explanation of similar CSTs among synonymous codons is that , in wild-type yeast cells for which the ribosome profiling was conducted , available cognate tRNAs for translating synonymous codons have effectively the same concentration . In rapidly growing yeast , ∼80% of total RNA is rRNA and ∼15% is tRNA [15] . The mean length of yeast tRNAs is ∼72 nucleotides and the total length of rRNAs per ribosome is 5469 nucleotides [15] . Thus , the number of tRNA molecules per cell is approximately ( 15%/72 ) / ( 80%/5469 ) = 14 . 2 times the number of ribosomes per cell , substantially exceeding the expected ratio of two tRNAs per active ribosome ( at A and P sites , respectively ) if tRNA recharging and diffusion is instantaneous . In reality , however , tRNA recycling takes time and thus cannot be ignored . Each tRNA , after completing its job of transferring an amino acid to the elongating peptide and then exiting the ribosomal E site , needs to be recharged with the cognate amino acid and then with eEF-1α+GTP to form a ternary complex before it can be reused in translation . It has been estimated that each ribosome translates ∼32 . 6 codons per second in yeast [26] . This implies that on average a tRNA molecule needs to be used 32 . 6/14 . 2 = 2 . 3 times per second , or once every 0 . 44 second . It is possible that the time for ternary complexes to form and diffuse to ribosomal A site is a substantial fraction of 0 . 44 second , so that the local concentration of ternary complexes is much lower than the total tRNA concentration . A recent study reported that consecutive synonymous codons in an mRNA tend to use the same tRNA and proposed that this codon choice is beneficial because a tRNA does not diffuse far from the ribosome after exiting its E site and is reused for translating the next synonymous codon when the ternary complex is formed again [27] . This observation and its explanation strongly implies that the local concentration of ternary complexes is low; otherwise , the addition of one cognate tRNA molecule among on average 20 tRNAs ( because identical amino acids are expected to be on average 20 residues apart ) cannot significantly increase the relative concentration of the cognate tRNA around the ribosome . Based on available information in E . coli , we calculated that the physiological concentration of ternary complexes is only ∼4 . 3% of the total concentration of tRNAs and ∼22% of the concentration of ribosomes ( see Materials and Methods ) . These observations strongly support our hypothesis that available tRNA is in shortage during translation . Consistent with our hypothesis , total tRNA concentrations increase with the rate of cell growth in E . coli [28] and tRNA gene copy number increases with the shortening of the minimal generation time across species [29] . Under tRNA shortage , the optimal usage of synonymous codons in minimizing the total CST ( i . e . , maximizing translational efficiency ) is to use isoaccepting tRNAs in proportion to their concentrations ( see Materials and Methods ) . That is , pi = qi , where pi is the relative usage of the ith synonymous codon of an amino acid ( Σpi = 1 ) and qi is the relative concentration of the corresponding tRNA ( Σqi = 1 ) . Under this codon usage , available cognate tRNAs of synonymous codons have equal concentrations and synonymous codon selection times become identical ( see Materials and Methods ) . We will refer to this theoretical optimal codon usage under tRNA shortage as the proportional rule . The proportional rule is not predicted by other models . For example , without tRNA shortage , two optimal solutions in minimizing the total CST exist . When codon usage is fixed , isoaccepting tRNA concentrations should follow , which is referred to as the square rule [30]–[31] . When tRNA concentrations are fixed , only the codon corresponding to the most abundant tRNA species should be used [30] , which is referred to as the truncation rule . To test if the actual codon usage of yeast follows the proportional rule , we examined the 12 amino acids that are each translated by at least two tRNA species in yeast . For each amino acid , the relative transcriptomic usage of a codon among synonymous codons ( i . e . , pi ) is quite close to the relative gene copy number of its cognate tRNA among isoaccepting tRNAs ( i . e . , qi ) , as predicted by the proportional rule ( Figure 2A ) . We measured the Euclidian ( Figure 2B ) and Manhattan ( Figure 2C ) distances in synonymous codon usage from the observed values to those predicted by the proportional rule , and found these distances significantly shorter than expected by chance ( Figure 2B–2D; Table S1; see Materials and Methods ) . Not surprisingly , genomic codon usage fits the proportional rule less well than the transcriptomic codon usage ( Figure 2A ) , reflected by greater distances from the predicted values ( Figure 2B , 2C ) . The better fitting of the transcriptomic codon usage to the proportional rule than to the square rule and truncation rule can be seen from a comparison of the distances under these three models ( Figure 2D ) . We also compared the likelihood of the three models , given the observed codon usage ( Figure 2D ) . The proportional model has a much higher log10 ( likelihood ) than the square model . Because the likelihood of the truncation model is 0 , this model is much worse than the other two models . The same conclusions are reached for the transcriptomic codon usage of all other model eukaryotes we examined ( Figure 2A , 2D ) . In the above analysis , we combined synonymous codons that are recognized by the same tRNA species ( referred to as iso-synonymous codons ) . Because the relative usage of such iso-synonymous codons does not affect the relative usage of isoaccepting tRNAs , it presumably does not affect translational efficiency . Nonetheless , iso-synonymous codons are not used equally , and factors other than translational efficiency ( e . g . , translational accuracy ) may be at work ( Table S2 ) . The observation of similar CSTs among synonymous codons and the empirical validation of the proportional rule strongly support the following model that includes three elements: ( 1 ) available tRNAs are in shortage during translation , ( 2 ) translational efficiency is optimized in nature by balanced codon usage according to tRNA concentrations , and ( 3 ) synonymous codons are translated with similar speeds under the codon-tRNA balance . Our model predicts reduced translational efficiency due to ribosome sequestering when the codon-tRNA balance is broken . It further predicts lower efficiency under exclusive use of preferred codons than balanced use of preferred and unpreferred codons . We experimentally tested the above predictions by quantifying the cellular efficiency in translation , represented by the protein expression of a reporter gene , under different levels of codon-tRNA imbalance induced by the expression of another gene . Unlike previous studies [12] , [17] , our separation of the inducer and reporter allows the distinction among several potential mechanisms of CUB's impact on protein expression . We inserted our reporter gene , the Venus yellow fluorescent protein ( vYFP ) gene controlled by the GPD promoter , into Chromosome XII of a haploid strain of S . cerevisiae ( Figure 3A ) . We then designed four synonymous sequences encoding another fluorescent protein , mCherry , as our inducer ( Figure S5 ) . The four mCherry sequences , named mCherry-1 , 2 , 3 , and 4 , cover the entire range of CAI of native yeast genes ( Figure 3B ) . We developed an index , distance to native codon usage ( Dncu ) , to measure the difference between the codon usage of a ( heterologous ) gene and the overall codon usage of the host cell , which is proportional to tRNA concentrations ( see Materials and Methods ) . The four mCherry versions also span a large range of Dncu ( Figure 3C ) and show different degrees of codon-tRNA imbalance for individual amino acids ( Figure S6 ) . Other than synonymous codon usage , the four mCherry versions are nearly identical: they encode the same protein sequence , have similar G+C content ( 42–44% ) , and have identical sequences in the first 56 nucleotides of the coding region , because this region may affect the level of protein expression [12] , [32]–[33] . Each mCherry gene is expressed from a constitutive and strong promoter on a high-copy-number plasmid ( see Materials and Methods ) . The four plasmids were separately transformed to yeast cells carrying the vYFP reporter gene ( Figure 3A ) . Our model predicts that the higher the Dncu of mCherry , the lower the vYFP expression . The four yeast strains were grown in rich media to the log phase , and the expression levels of vYFP and mCherry proteins were inferred from their fluorescent signals , which were simultaneously measured for each cell by fluorescence-activated cell scanning of at least 300 , 000 cells . We found mCherry expression levels to be significantly different among the four strains ( see Materials and Methods ) . Within each strain , expression levels of mCherry and vYFP are negatively correlated among cells ( see Materials and Methods ) . Hence , the expressions of vYFP cannot be directly compared among strains . Instead , we separated the cells of each strain into three bins on the basis of mCherry expression and then compared vYFP expressions among the four strains for cells with similar mCherry expressions ( Figure 3D ) . We found that , across the range of mCherry expressions shared by the four strains , the higher the Dncu of mCherry , the lower the expression of vYFP ( Figure 3D ) . Furthermore , the vYFP expression-level difference among the strains increases with the mCherry expression level ( Figure 3D ) . Of special interest is the comparison between mCherry-3 and mCherry-4 , which clearly shows that it is a low Dncu rather than a high CAI that enhances translational efficiency ( Figure 3D ) . A multivariate regression analysis of all cells from the four strains further demonstrated that Dncu is significantly more important than CAI in explaining the variation of the vYFP signal ( P<0 . 001 ) . The above results were not due to different random mutations fixed in the genomes of the four strains during our experiments , because the vYFP signals were not significantly different among the strains upon removal of the plasmids ( Figure 3E ) . We also sequenced the entire plasmid DNA from each strain and found no mutation . Using quantitative polymerase chain reaction , we further verified that the vYFP mRNA abundance is not different among the four strains ( Figure 3F ) . Thus , the among-strain variation in vYFP signal must be due to a variation in translation . We also confirmed our results by a finer control of mCherry expression and ruled out the possibility that our observation is a byproduct of potential differences in translational accuracy among different mCherry versions ( 7; see Materials and Methods ) . Furthermore , because the accuracy hypothesis is based on CAI and thus predicts a higher vYFP expression in the strain carrying mCherry-4 than that carrying mCherry-3 , our results ( Figure 3D ) are inexplicable by this hypothesis . Similarly , mechanisms resulting from translational errors , such as protein misfolding or aggregation , cannot explain our observation either . In the experiment , we used vYFP to represent native genes in the yeast genome . However , because vYFP and mCherry have 71/220 = 32% of protein sequence identity , one might ask whether our observation can be generalized . Specifically , could the negative influence of mCherry expression on vYFP expression be caused entirely by the similarity in codon usage between mCherry and vYFP ? We measured the codon usage dissimilarity between a pair of genes by a Euclidian distance and examined the distribution of this distance between each mCherry version and all yeast genes ( Figure S8 ) . The distribution is approximately bell shaped and the distance between mCherry and vYFP falls in the central part of the bell , suggesting that mCherry is no more similar to vYFP in overall codon usage than to average yeast genes . Furthermore , our results cannot be explained by amino acid similarity between mCherry and vYFP , because all mCherry versions have the same amino acid sequence and should not differentially affect vYFP expression through amino acid usage . Thus , our observation from vYFP can be extrapolated to native genes in the yeast genome . If translational efficiency is maximized when the cellular codon usage follows the proportional rule , why do highly expressed genes necessarily prefer codons with highly abundant cognate tRNAs and have stronger CUB than lowly expressed genes ? We hypothesize that these phenomena are due to differential selective coefficients associated with synonymous mutations occurring in highly expressed and lowly expressed genes in the regain of the codon-tRNA balance upon a genetic perturbation . Let us imagine an amino acid with two synonymous codons ( codon1 and codon2 ) that each uses a distinct tRNA species ( tRNA1 and tRNA2 ) and assume that the present codon usage follows the proportional rule . Now , if the proportion of tRNA1 rises due to a mutation , natural selection will promote the fixations of synonymous mutations from codon2 to codon1 to reestablish the codon-tRNA balance . Such advantageous mutations occurring in highly expressed genes affect tRNA usage more than those occurring in lowly expressed genes and hence have a greater selective advantage and are fixed faster . This difference becomes even bigger when clonal interference [34] is considered . As a result , highly expressed genes use more codon1 and fewer codon2 than before and show stronger CUB . The contrasting scenario , in which the tRNA usage is rebalanced by frequent use of codon1 in lowly expressed genes , requires many synonymous substitutions in many lowly expressed genes , which will not happen because it takes much longer than rebalancing the tRNA usage by increasing codon1 frequency in highly expressed genes . Indeed , in a computer simulation of codon usage evolution that starts from the equal usage of 4 synonymous codons whose cognate tRNAs have different concentrations , the final usage of the codons , after 500 generations of random mutation , genetic drift , and natural selection for translational efficiency , follows the proportional rule ( Figure 4A ) . More importantly , the preferential use of high-concentration tRNA species and strong CUB in highly expressed genes are seen from both the average of 1000 simulation replications ( Figure 4B ) and any one replication ( Figure 4C ) . The standard deviations presented in Figure 4B indicate an extremely low probability for CUB to be stronger or a preferred codon to be used more frequently in lowly expressed genes than highly expressed genes . As expected , the phenomena in Figure 4 disappear when the natural selection for translational efficiency is removed in the simulation ( Figure S9 ) . These observations support our model that the high CAI of highly expressed genes is a byproduct of natural selection for an overall cellular efficiency in translation , rather than the direct product of stronger selection for translation efficiency in more highly expressed genes [6] . Analogous to synonymous codon usage , we predict that the optimal amino acid ( or nonsynonymous codon ) usage in speeding up translation is in proportion to the corresponding tRNA concentrations . Indeed , amino acid frequencies inferred from transcriptome data were reported to correlate positively with the corresponding tRNA gene copy numbers in yeast [35] and C . elegans [36] . More importantly , actual amino acid usage is significantly closer than random usage to our predicted optimal ( i . e . , the diagonal line in Figure 5A; P<10−6 , simulation test ) . This phenomenon is also true in all other model eukaryotes examined , although the level of match between the observation and prediction varies among species ( Figure 5A ) . Transcriptomic amino acid usages instead of proteomic amino acid usages are plotted here because the latter are unavailable for most species . Nevertheless , S . cerevisiae data showed an almost perfect correlation between transcriptomic and proteomic amino acid usages ( Figure S10 ) , indicating that the former is a good proxy for the latter . We also predict a positive correlation between aminoacyl tRNA synthetase concentration and corresponding tRNA concentration to enhance the efficiency of amino acid charging . Such a correlation is indeed found in S . cerevisiae ( r = 0 . 45 , P = 0 . 03; Figure S11 ) . If amino acid frequencies are in perfect proportion to tRNA concentrations , the mean CST for an amino acid should not vary among amino acids . This uniformity , however , is not observed in yeast ( Figure S12 ) , suggesting that amino acid usage is only roughly proportional to tRNA concentrations ( Figure 5A ) , which may be due to mutational bias [37] or antagonistic selective pressures from factors such as physiochemical properties [38] and synthetic costs [39] of various amino acids . Our model predicts that the average CST of an amino acid increases with the decrease of the relative availability of tRNAs for the amino acid . Indeed , a negative correlation exists between the tRNA availability and CST for the 20 amino acids ( Pearson's r = −0 . 40 , P = 0 . 03 , permutation test; Figure 5B ) . This finding reconfirms tRNA shortage in translation , explains in part why CSTs of nonsynonymous codons vary , and indicates compromised translational efficiency due to other fitness effects of amino acid usage .
Results from several earlier experiments are consistent with the role of CUB in enhancing translational efficiency or reducing ribosome sequestering [12] , [17] . For example , when expressing many synonymous versions of a green fluorescent protein ( GFP ) gene in E . coli , Kudla and colleagues reported that strains harboring high-CAI GFP genes tend to grow faster than those harboring low-CAI GFP genes , despite the lack of a correlation between the GFP protein expression level and its CAI [12] . Although these authors found no correlation between CAI and protein misfolding , their experiment was unlikely to be sensitive enough for quantifying GFP misfolding [12] . Thus , it could not rule out the possibility that the observed variation in fitness was entirely caused by CUB's influence on translational accuracy . By contrast , we were able to demonstrate CUB's impact on translational efficiency after excluding its impact on translational accuracy . A recent study in E . coli showed that the ribosome shortage induced by over-expression of unneeded proteins can be alleviated by physiological adaptation in 30 to 40 generations , owing to the manufacture of additional ribosomes [40] . This finding suggests that the disadvantage of suboptimal codon usage may also be mitigated by physiological adaptation . Nevertheless , physiological adaptation takes time . If the growth rate fluctuates rapidly due to frequent environmental changes , the fitness of the individual with suboptimal codon usage is expected to be much lower than the individual with balanced codon usage . We hypothesized and demonstrated that translational efficiency is optimized by codon-tRNA balance . This new model of translational efficiency by unequal codon usage differs substantially from the prevailing model ( Table 1 ) . One critical piece of evidence for our model is similar CSTs of synonymous codons in wild-type yeast . Our CST estimation is based on the assumption that the time a codon occupies the ribosomal A site equals the waiting time for the cognate tRNA . Our estimates of all CSTs would be biased upward to a similar level if downstream “traffic jams” happen during translational elongation . However , a recent study suggested that downstream traffic jams are unlikely , due to slow “ramps” at the beginning of an mRNA [21] . Furthermore , even if downstream traffic jams occur , it should affect synonymous codons as well as nonsynonymous codons and thus cannot explain why only synonymous codons but not nonsynonymous codons have similar CSTs . Over two decades ago , Curran and Yarus indirectly estimated relative CSTs for 29 sense codons in E . coli , under the assumption that the probability of a frame shift in the translation of a codon is proportional to the CST of the codon [41] . They reported that only codons of very low CSTs tend to be preferentially used [41] . However , because their fundamental assumption about the frame-shift rate is incorrect [42] , their CST estimates are unlikely to be correct . It is also possible that prokaryotes and eukaryotes have some differences in using CUB to regulate translational efficiency ( e . g . , translational attenuation in prokaryotes ) . In another E . coli study , Sorensen and colleagues reported faster translation of a multicopy-plasmid-borne lacZ gene when a segment of the gene comprises mainly preferred codons than when it comprises mainly unpreferred codons [43] . This result cannot be used to infer relative CSTs of synonymous codons in wild-type cells , because the extremely high expression of synonymous versions of the endogenous lacZ gene from plasmids potentially breaks the codon-tRNA balance and alters CSTs . Nevertheless , their observation is fully compatible with our finding of different levels of translational efficiency induced by the expressions of different synonymous versions of mCherry . Several other studies reported similar findings [25] , [44] . Recently , some authors calculated CSTs by assuming that the CST of a codon is determined by the relative concentrations of its cognate , nearly cognate , and non-cognate tRNAs without considering tRNA shortage or using ribosome profiling data [45] . Because of the violation of the fundamental assumption they made , their estimates are likely to be incorrect . Indeed , their estimated CSTs would predict a slower translation of mCherry version 3 than 4 , contradictory to our experimental result ( Figure 3D ) . While the present work was under review , Ingolia and colleagues reported estimates of translational elongation speeds in mouse embryonic stem cells using a pulse-chase strategy that does not involve expressions of heterologous genes [46] . Although their method is different from ours , their finding of similar elongation speeds among synonymous codons is highly consistent with our results from yeast . Our discoveries require reinterpretation of several earlier observations . For example , higher prevalence of codons with abundant cognate tRNAs in genes with higher expressions is often interpreted as a result of a stronger demand for fast translation of more abundant proteins [19]–[20] . This interpretation is not supported by our results . Rather , we suggested and demonstrated by simulation that , the selection coefficient for synonymous mutations that help achieve the codon-tRNA balance is greater in highly expressed genes than in lowly expressed genes , leading to quicker and more acquisitions of codons with abundant cognate tRNAs in the former than in the latter . In this regard , our results support that CUB serves as a global strategy to enhance the efficiency of the translation system [12] , [47] . Within an organism , the transcriptome can vary among cell cycle stages , developmental stages , and tissues . How do such variations affect the codon-tRNA balance ? We found pairwise Pearson's correlations in transcriptomic usage of all 61 sense codons to be nearly 1 among different time points in the S . cerevisiae mitotic cell cycle ( Figure 6 ) . We further analyzed the transcriptomic usage of all 61 codons across tissues and/or developmental stages in the worm , fruitfly , and human . If multiple replications of the same cell type exist in a dataset , we randomly chose one replication in our analysis . Similarly high correlations were observed among different cell types within species ( Figure 6 ) . By contrast , the correlation is generally below 0 . 5 between any pair of the four species examined here . The high correlation in codon usage across cell cycle stages , developmental stages , and tissues of the same species is likely due to house-keeping genes , which are always highly expressed . Thus , within-organism gene expression variations have little impact on the maintenance of the codon-tRNA balance . Further , tRNA concentrations may covary with the transcriptomic codon usage to maintain the codon-tRNA balance across tissues [48] . A byproduct of our CST estimation is the translational initiation rate of each gene . We found that the translational initiation rate is significantly positively correlated with the mRNA concentration ( ρ = 0 . 34 , P = 6×10−81 ) , suggesting a coordinated regulation of gene expression at the transcriptional and translational levels . We also observed a strong positive correlation between the translational initiation rate and CAI ( ρ = 0 . 51 , P<10−196 ) , suggesting that CAI provides a moderate amount of information about the translational initiation rate . This may explain why the protein concentration correlates with the product of mRNA concentration and CAI better than with the mRNA concentration alone [49] . Several studies revealed reduced mRNA stability near the translation initiation site , suggesting that the reduced stability may enhance the translational initiation rate [12] , [32]–[33] . Indeed , we found a weak but significant positive correlation between the reduction in mRNA stability [32] and our estimated translational initiation rate ( ρ = 0 . 08 , P = 1×10−5 ) . Given that CUB improves both translational efficiency and accuracy , one wonders whether one of these effects is a side-effect of the other . For instance , it was previously suggested that the variation in translational accuracy among synonymous codons may be a byproduct of the variation in translational efficiency , because ( i ) most translational errors are believed to occur during codon selection , ( ii ) codon selection has been assumed to be faster for preferred codons than unpreferred codons , and ( iii ) faster codon selection is thought to result in fewer errors [50] . Because our result invalidates assumption ( ii ) for wild-type cells , the above argument no longer holds . Thus , even though translational accuracy may be affected by relative concentrations of tRNAs in engineered yeast cells with grossly imbalanced codon-tRNA usage [51] , this impact is not expected in wild-type cells because our results strongly suggest that isoaccepting tRNA species have effectively the same concentrations in wild-type cells . In addition , the enrichment of preferred codons at evolutionarily conserved amino acid residues cannot be explained by the translational efficiency hypothesis [7]–[10] . Furthermore , experimental data showed that translational accuracies of iso-synonymous codons vary [52] , suggesting that the variation in accuracy cannot be entirely caused by the variation in cognate tRNA concentration , because iso-synonymous codons use the same cognate tRNA . Rather , comparative genomic analyses strongly suggest that translational accuracy is likely to be intrinsically different among synonymous codons [1] , [53] . Further , we were able to establish CUB's impact on translational efficiency even after we controlled its impact on translational accuracy ( Figure 3 , Figure S7 ) . In addition , because translational accuracy is not entirely determined by translational efficiency [7]–[10] , the proportional rule , which is predicted from selection for efficiency , is not predicted from selection for accuracy , especially because translational errors at different residues have different fitness effects . Thus , the impact on efficiency cannot be a byproduct of the impact on accuracy . Taken together , we conclude that translational accuracy and efficiency are two separable benefits of CUB . Let us compare three evolutionary models of CUB that differ in the roles of translational accuracy and efficiency as the selecting agent . We also consider mutational bias and genetic drift , two known factors in the evolution of CUB , in these models . In model I , translational efficiency is the sole selecting force ( Figure 7 ) . This model predicts co-evolution of codon usage and cognate tRNA concentrations and a codon-tRNA balance at which the relative frequency of a synonymous codon ( pi ) equals the relative abundance of its cognate tRNA ( qi ) . The expected values of pi = qi are determined by the mutational bias , which directly affects codon usage and indirectly affects tRNA concentrations . However , this model cannot explain the observation that , although preferred codons of an amino acid vary among species , this variation decreases substantially ( but does not disappear ) after the control of genomic GC content [1] . For example , GTT and GTA both code for valine and have the same GC content , but GTT is frequently used as the preferred codon when the genomic intergenic GC content is below 50% [1] . When the GC content exceeds 50% , GTG rather than GTC is often used as the preferred codon for valine [1] . This observation suggests that , in addition to translational efficiency , there is a separate selecting force with a relatively constant direction . In model II , translational accuracy is the sole selecting agent on CUB ( Figure 7 ) . The demand for translational accuracy , coupled with the mutational bias , determines the expected CUB , whereas selection for translational efficiency determines tRNA concentrations based on codon frequencies . The phenomenon of stronger CUB in more highly expressed genes is explainable by the protein-misfolding-avoidance hypothesis which predicts that highly expressed genes are translated more accurately by using accurate codons more frequently [7] , [54] . Model II predicts that , after the control for the mutational bias , accurate codons are always the preferred codons in a species . If the translational accuracy of a codon is an intrinsic property of the codon and does not vary among species [29] , we should observe no variation in the choice of preferred codons , after the control of mutational bias . This prediction , however , is incorrect , because preferred codons are not always the same in different species with the same mutational bias [1] , [29] . A more rigorous test of this model is to compare the accurate and preferred codons of each amino acid in a species , because model II predicts a complete match between them . For each codon , we calculated an odds ratio by the relative use of the codon over other synonymous codons at conserved amino acid positions divided by that at non-conserved amino acid positions; the synonymous codon with the highest odds ratio is regarded as the most accurate codon because it is most preferentially used at important amino acid positions [7]–[10] . By comparing S . cerevisiae with its relative S . bayanus , we identified conserved and non-conserved amino acid positions . We calculated the odds ratio for each codon in each gene and then combined the odds ratios from all genes using the Mantel-Haenszel procedure [23] . By definition , the preferred codon of an amino acid is the one with the highest RSCU' . We found that , in 6 ( Ala , Asp , Gly , His , Thr , and Val ) of the 18 amino acids that have at least two synonymous codons , the codon with the highest odds ratio is different from the codon with the highest RSCU' ( Figure 8 ) . Furthermore , for three amino acids ( Asp , His , and Thr ) , the codon with the highest RSCU' has an odds ratio significantly lower than 1 ( Figure 8 ) . We also used the 10% most highly expressed genes to calculate odd ratios; 8 ( Ala , Arg , Asp , Cys , Ile , Leu , Thr , and Val ) of the 18 amino acids show mismatches between the codon with the highest RSCU' and the codon with the highest odds ratio ( Figure 8 ) . These results provide unambiguous evidence for the inadequacy of model II . In model III , selections for translational accuracy and efficiency jointly determine CUB ( Figure 7 ) . Let us consider three types of synonymous mutations with regard to their impacts on translational accuracy and efficiency . First , a synonymous mutation is likely to be fixed when it enhances both translational accuracy and efficiency , but is likely to be lost when it decreases both . Second , a synonymous mutation may increase the accuracy but reduce the efficiency . One possible outcome is that selection for higher accuracy will gradually alter the codon usage , which is followed by tRNA concentration changes that recover the loss of efficiency . Eventually , accurate codons will be the preferred codons . Alternatively , selection for higher accuracy may not be able to alter the codon usage permanently if the loss of efficiency is either too large or cannot be recovered by a corresponding tRNA change as quickly as the switch back of the codon usage . Consequently , accurate codons cannot become the preferred codons and the system is trapped in a local fitness peak that is the maximum for efficiency but not accuracy . For example , while codon CCA is more accurate than CCT for proline ( Figure 8 ) , there are still about a quarter of bacterial species with GC%<40 that use CCT as their preferred proline codon [1] , suggesting that it is not rare for codon usage to be trapped in a local fitness peak . Third , a synonymous mutation may increase the efficiency but reduce the accuracy when the system is at a codon-tRNA imbalance . Although the fate of this mutation is determined by the relative strengths of the two forces , changes of tRNA concentrations could resolve the conflict better because they can increase efficiency without reducing accuracy . So , the final codon usage pattern will also depend on the rate of mutations that alter tRNA concentrations . While the quantitative aspects of model III require further exploration , it is clear that the model is able to explain , at least qualitatively , both the matches and mismatches between the accurate and preferred codons ( Figure 8 ) . It is also able to explain the codon-tRNA balance and the phenomenon of stronger CUB in genes with higher expressions . Thus , model III is most compatible with and best supported by available data . In addition to translational accuracy and efficiency , synonymous codon usage of individual genes may also be shaped by other forces , for example , those related to RNA splicing and stability [55] . But these forces are gene-specific and do not create genomic patterns of CUB . Synthetic biology designs and constructs novel biological functions not found in nature . It has long been known that , in many but not all cases , increasing the CAI of a transgene boosts its protein expression [12] , [56]–[57] . Different protein expression levels of synonymous transgenes are likely caused by CST differences created by various degrees of codon-tRNA imbalance induced by transgene expressions . Consistent with this idea , overexpression of rare tRNAs of E . coli ( the bio-reactor ) can rescue the tRNA depletion when heterologous human genes are expressed in E . coli [56] . When an artificially designed gene is added to a host cell , the potential imbalance between the overall cellular codon usage and the tRNA pool also affects the expressions of native genes and hence the growth of the host cell . We showed that Dncu , a newly devised index measuring the distance in codon usage between the transgene and the host cell , is an accurate indicator of the impact of per transgene protein molecule production on the expressions of native genes . We demonstrated that it is the Dncu rather than CAI of the transgene that predicts its impact on the host protein expression . Therefore , Dncu should be considered in synthetic biology when the impact of transgene expression on host gene expressions is a concern . Further , when genes from multiple species are assembled into a synthetic genome , designing tRNA gene numbers in proportion to the usage of their cognate codons will likely make protein expressions in the entire cell most efficient .
The yeast ribosome profiling data [18] were downloaded from Gene Expression Omnibus ( www . ncbi . nlm . nih . gov/geo/ ) under accession number GSE13750 . Gene expression and protein expression levels were from http://web . wi . mit . edu/young/expression/ [58] , http://www . imb-jena . de/tsb/yeast_proteome/ [59] , and the supplementary data of a previous study [60] . Transcriptomic data for the yeast mitotic cell cycle were from a previous study [61] . Gene sequences and reading frames were downloaded from Saccharomyces Genome Database ( SGD , www . yeastgenome . org ) . Numbers of tRNA gene copies were retrieved from an earlier study [22] . Gene expression levels in A . thaliana , D . melanogaster , M . musculus , and H . sapiens were downloaded from Gene Expression Omnibus ( GDS416 , GDS2784 , GDS592 and GDS596 , respectively ) . Gene expression levels in S . pombe and C . elegans were retrieved from two earlier studies [62]–[63] , respectively . Peptide and cDNA sequences of S . pombe , A . thaliana , C . elegans , D . melanogaster , M . musculus , and H . sapiens were from Ensembl ( www . ensembl . org/ ) . Numbers of tRNA gene copies in the above species were obtained from the genomic tRNA database ( http://lowelab . ucsc . edu/GtRNAdb/ ) . Using the S . cerevisiae ribosome profiling data [18] , we identified codons docked at the ribosomal A site , from the Illumina Genome Analyzer sequencing reads . By comparing the observed codon frequencies in the ribosome profiling data with the expected codon frequencies estimated from mRNA-Seq data generated under the same condition in the same experiment , we calculated the relative CSTs of all 61 sense codons . Although Illumina sequencing may be biased toward certain sequences or nucleotides [64] , this bias affects the mRNA-Seq and ribosome profiling data equally and thus will not affect our estimation of CST . For a sequencing read from the ribosome profiling data , nucleotide positions 16–18 were considered to be at the ribosomal A site where codon selection occurs [18] . Only those reads with exactly 28 nucleotides and 0 ambiguous sites were used to ensure the accurate determination of positions 16–18 . We calculated the fraction of in-frame codons by comparing the read sequences with annotated yeast coding sequences . Consistent with what was previously reported [18] , the majority of codons at positions 16–18 were in-frame in the ribosome profiling data . In the mRNA-Seq data , the fraction of each phase was close to one third , as expected . All out-of-frame codons were excluded . The probability of incorrect codon assignment was low , because only codons misaligned by at least 3 nucleotides may be assigned incorrectly . Transposons and uncharacterized genes were removed . Our CST estimation procedure ( Figure S1 ) is as follows . We first calculated fi , the observed frequency of codon i , in the ribosome profiling data by ( 1 ) where cij is the count of codon i in mRNA j positioned at the ribosomal A site measured by ribosome profiling and N is the number of genes with ribosome profiling data ( N>3000 for both rich and starvation conditions ) . The expected ribosome footprint frequencies of codon i ( Fi ) when all codons have equal CST can be calculated based on the frequency of the codon in the mRNA-Seq data using ( 2 ) where Rj is the translational initiation rate of mRNA j and Cij is the count of codon i in mRNA j measured by mRNA-Seq . Then , the relative codon selection time for codon i is calculated by ( 3 ) We used an iterative approach to estimate the translational initiation rates that appear in Eq . 2 . We first used Rj = 1 for all j . After the CST is calculated for each codon , the elongation rate ej of mRNA j ( i . e . , the number of codons translated per unit time ) is calculated by ( 4 ) where Lj is the number of codons in each molecule of mRNA j and Dij is the number of codon i in each molecule of mRNA j . The translational initiation rate Rj can be estimated from ( 5 ) where dj is the ribosome density on mRNA j ( i . e . , the number of ribosomes per codon ) and can be estimated by ( 6 ) We then used the newly estimated translational initiation rates to calculate CSTs . After 10 iterations , CST estimates converge ( Figure S2 ) and are considered as our final estimates . Because our estimates of CSTs are relative values , we rescaled them by setting the maximal observed value at 1 . CST estimates from different experimental replicates were highly correlated ( r = 0 . 79 , P = 6×10−14 ) and were thus pooled for the rest of the analysis . Three different sets of initial values of translational initiation rates ( uniform , proportional to CAI of each gene , inversely proportional to CAI ) were used in CST estimation and they resulted in identical estimates of CSTs ( Figure S3A , S3B ) . Thus , CST estimation does not depend on the initial values of R . The standard errors of the CST estimates were estimated by bootstrapping genes present in the ribosomal profiling data 1000 times . The CST estimates from two different media ( rich and starvation ) are also very similar ( Figure S3C ) . To ensure no mistake in the estimation of CST , the first two authors of this paper independently derived the formulas , wrote the computer programs , and estimated the CSTs , and their results were virtually identical . There are two commonly used measures of synonymous codon usage bias . The first is the relative synonymous codon usage ( RSCU ) , defined by the frequency of a codon relative to the average frequency of all of its synonymous codons in a set of highly expressed genes [19] . Codons with RSCU>1 are preferred and those with RSCU<1 are unpreferred . To compare the usage of all 61 sense codons , we also used RSCU' = RSCU/n , where n is the number of synonymous codons of an amino acid . RSCU' of a codon is the proportion of use of a given codon among synonymous choices in a set of highly expressed genes . The second commonly used measure of synonymous codon usage bias is the codon adaptation index ( CAI ) , which is calculated for a gene , and measures its usage of high-RSCU codons [20] . Briefly , CAI of a gene is the geometric mean of RSCU divided by the highest possible geometric mean of RSCU given the same amino acid sequence . CAI is a positive number no greater than 1 . The greater the CAI , the more prevalent are preferred codons in the gene . We first selected 200 most highly expressed genes based on a previous study [59] . Sixteen of these genes did not have expression information in another study [58] and 4 had expression levels lower than 4 times the genomic average ( 2 . 7 mRNA/cell reported in an earlier study [58] ) . The remaining 180 highly expressed genes were used to calculate RSCU and RSCU' for each codon . Our RSCU estimates were highly correlated with those previously reported [20] ( r = 0 . 995 , P<0 . 001 , permutation test ) . CAI was calculated for each yeast gene and for each version of mCherry based on the RSCU values obtained above , following a previous study [20] . We also estimated the effective number of codons ( Ncp ) for each gene , after controlling the GC content of the gene [65]–[66] . We separately estimated the frequency ( f ) of each of the 61 sense codons in each gene . We then estimated Spearman's rank correlation ( ρ ) between Ncp and f among all genes for each codon . Among synonymous codons , those with more negative ρ values are considered to be more preferred [1] . This dataset was used in Figure S4 only . It has been reported that the physiological concentration of the ternary complex is ∼200 nM for Phe tRNA and Lys tRNAs in E . coli [67] . Because the number of Phe tRNA and Lys tRNA molecules per cell is 1830 and 4300 , respectively [68] , we calculated that the Phe tRNA concentration is 1830/ ( 6 . 02×1023 ) / ( 1 . 1×10−15 ) = 2 . 8×10−6 M = 2800 nM , where 6 . 02×1023 is the number of molecules per mole and 1 . 1×10−15 liter is the average volume of an E . coli cell . Similarly , Lys tRNA concentration is estimated to be 6500 nM . Thus , about 200/[ ( 2800+6500 ) /2] = 4 . 3% of tRNAs are in ternary complexes . Because there are ∼1 . 2×104 ribosomes per E . coli cell [68] , ribosome concentration is ∼18 , 000 nM . Thus , the ratio in the concentration of ternary complexes to that of ribosomes is expected to be 200×20/18000 = 0 . 22 , if Lys and Phe can represent all 20 amino acids in ternary complex concentration . Without loss of generality , we assume that an amino acid is encoded by synonymous codons 1 and 2 , which are respectively recognized by isoaccepting tRNAs 1 and 2 . Let the relative usage of the two codons be p1 and p2 = 1−p1 and the relative concentrations of the two tRNAs be q1 and q2 = 1−q1 , respectively . Let the codon selection time for the two synonymous codons be t1 and t2 , respectively . Thus , the expected codon selection time for the amino acid concerned is t = p1t1+p2t2 . When tRNAs are in shortage , the local concentrations of tRNA 1 and 2 are aq1/p1 and aq2/p2 , where a is a constant . Because codon selection time is proportional to the inverse of the local tRNA concentration , we have , where b is another constant . The above formula can be simplified to . It is easy to find that t reaches its minimal value of b/a when and . In other words , the expected codon selection time is minimized and thus translational efficiency is maximized when relative synonymous codon frequencies equal relative tRNA concentrations . Under this condition , codon selection time equals b/a for both codons and local tRNA concentration equals a for both tRNAs . A full treatment considering tRNA cycle and kinetics gave the same result [31] . We measured the Euclidian distance and Manhattan distance in synonymous codon usage from the observed values to the values predicted from the observed tRNA fractions using the proportional rule . To evaluate whether the observed distances are shorter than expected by chance , we conducted a computer simulation with 106 replications under random codon usage . That is , the frequency of a synonymous codon is uniformly distributed between 0 and 1 with the constraint of the total frequency of all synonymous codons being 1 . We then obtained the distribution of the distance between a random codon usage and the codon usage predicted from the observed tRNA fractions . We also conducted a second simulation with 106 replications , in which tRNA factions vary randomly according to the above uniform distribution . We then obtained the distribution of the distance between the observed codon usage and that predicted from random tRNA fractions . This way , the potential confounding effect of genomic GC content on the assumed null distribution of codon usage becomes irrelevant to the test . We similarly tested the square rule and the truncation rule . Results from the first simulation are presented in Figure 2D , while those from the second simulation are in Table S1 . We developed an index , distance to native codon usage ( Dncu ) , to measure how different the codon usage of a ( heterologous ) gene is from the overall codon usage of the host cell , which is presumably balanced with tRNA concentrations . First , the Euclidean distance in synonymous codon usage between the heterologous gene and the host is calculated for each of the 18 amino acids with at least two synonymous codons by ( 7 ) where Yij is the fraction of codon j among the synonymous codons of amino acid i for the heterologous gene and Xij is the fraction of codon j among the synonymous codons in the host transcriptome , ni is the number of synonymous codons for amino acid i . Dncu of the gene is defined as the weighted geometric mean of Di , or ( 8 ) where k≤18 is the number of amino acid types encoded by the gene excluding Met and Trp , which have no synonymous codons , mi is the number of amino acid i found in the protein , and l is the protein length excluding Met and Trp residues . By definition , Dncu is between 0 and 1 . The mCherry gene sequence was obtained from a previous study [69] . We designed four synonymous DNA sequences encoding the same mCherry peptide ( Figure S5 ) . The first 56 nucleotides were the same for all four sequences to avoid potential effects on the mRNA secondary structure , which affects protein translation [12] , [32]–[33] . The GC contents of the four sequences ( 42–44% ) were also made similar to each other and to the average value in yeast coding sequences ( 40% ) . In all sequences , synonymous codons were randomized in order and thus were unlikely to cause differences in order-related effects [27] . The different versions of mCherry DNA sequences were synthesized by Blue Heron Biotechnology . They were cloned into p426GPD [70] at SpeI and XhoI ( New England Biolabs; Promega ) and are under the control of the GPD promoter . The plasmids were subsequently transformed individually into a haploid yeast cell ( BY4742 ) with vYFP [71] inserted into Chr XII [72] . The genotype of the cell is MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 hoΔ0::PGPD-Venus . We measured the expressions of mCherry and vYFP in log growth phase in Yeast extract/Peptone/Dextrose ( YPD ) media by florescence-activated cell scanning ( FACSCalibur , BD ) . Fluorescence of mCherry was measured from FL4 with a 670 nm pass filter and fluorescence of vYFP was measure from FL1 with a filter having a 30 nm bandpass centered on 530 nm . Yeast cells with mCherry fluorescence signals greater than the BY4742 negative control cells ( i . e . , mCherry fluorescence signals >10 ) were gated . We retrieved the forward scatter ( FSC , which is proportional to cell size ) and mCherry and vYFP fluorescence signals for all gated cells . The expression levels of fluorescent proteins were defined as their fluorescence signals divided by FSC . The mean mCherry expression level is 3 . 388±0 . 002 , 6 . 468±0 . 007 , 14 . 003±0 . 032 , and 14 . 544±0 . 022 for the strains carrying mCherry-1 , 2 , 3 , and 4 , respectively . Expression levels of mCherry and vYFP were negatively correlated for each strain ( mCherry-1: r = −0 . 22; mCherry-2: r = −0 . 57; mCherry-3: r = −0 . 60; mCherry-4: r = −0 . 62; P<2 . 2×10−16 in all cases ) . All gated cells were then grouped into 3 ( Figure 3D ) or 15 ( Figure S7 ) bins with equal mCherry expression ranges . For each genotype , multiple independently transformed strains were examined on different days , but the results were highly similar . We thus combined all results obtained from different strains of the same genotype . The total numbers of cells measured were 456333 , 648792 , 352863 , and 793832 , respectively , for the strains carrying mCherry-1 , 2 , 3 and 4 ( Figure 3B ) . To confirm that our results were not due to random secondary mutations , we removed the plasmids from each strain by using 5′-FOA media to select against the plasmids , and then measured the vYFP fluorescence intensities . We also sequenced the entire plasmid DNA from each of the four strains . To compare the vYFP mRNA levels among strains , we extracted the total RNA ( RiboPure-Yeast Kit , Ambion ) from three independently transformed strains of each genotype . The total RNA was reversely transcribed into cDNA ( Moloney Murine Leukemia Virus Reverse Transcriptase , Invitrogen ) with random hexamer primers . The vYFP mRNA level was measured by quantitative polymerase chain reaction ( 7300 Real-Time PCR System , Applied Biosystems ) with ACT1 as an internal control . The primers for vYFP are 5′ – CATGGCCAACACTTGTCACT– 3′ and 5′ –TACATAACCTTCGGGCATGG– 3 , while the primers for ACT1 are 5′ - CTGCCGGTATTGACCAAACT - 3′ and 5′ – CGGTGATTTCCTTTTGCATT – 3′ . The software package RELAIMPO ( http://cran . r-project . org/web/packages/relaimpo/ ) was used for a multivariate regression analysis of the yeast experimental data from all cells of the four strains . We compared the relative importance of Dncu and CAI in explaining the among-cell variation in vYFP signal by the LMG method and used 1000 bootstrap replications to determine the statistical significance . Use of other methods ( LAST , FIRST , and PRATT ) implemented in RELAIMPO gave similar results . Proponents of the translational accuracy hypothesis might argue that , because different synonymous codons have different mistranslation rates [52] , [73] and preferred codons are considered to be more accurately translated than unpreferred codons [7] , the mCherry with a low CAI is expected to produce fewer functional protein molecules than the mCherry with a high CAI even when the same numbers of protein molecules are produced . In other words , using red florescent signals may have led to a more severe underestimation of protein expression for the mCherry with a low CAI than for that with a high CAI . The average mistranslation rate has been estimated to be ∼5×10−4 per codon , and unpreferred codons have been posited to undergo mistranslation five times as often as preferred codons [7] . Based on these numbers and the CAIs of the four mCherry versions ( Figure 3B ) , we assume that the mistranslation rate is 10×10−4 , 8×10−4 , 5×10−4 , and 2×10−4 per codon for mCherry-1 to mCherry-4 , respectively . Let us further assume that no mistranslated protein fluoresces . Given the length of mCherry ( 236 amino acids ) , we expect that 11 . 8% , 9 . 44% , 5 . 9% , and 2 . 36% of mCherry-1 to mCherry-4 proteins respectively fail to fluoresce due to mistranslation . On this assumption , we corrected mCherry expression levels from the observed florescent signals . We also conducted a better control of mCherry expression among strains by dividing cells of each strain into 15 bins based on the above corrected mCherry expression ( Figure S7 ) . Again , we observed a lower vYFP expression when the Dncu of the mCherry gene is higher , across the range of mCherry expressions shared by the three strains ( Figure S7 ) . This result is conservative , because only a minority of mistranslations are expected to prevent fluorescence , and it is likely that we have overcorrected the effect of mistranslation . We simulated the evolution of synonymous codon usage in an asexual haploid unicellular digital organism . In this organism , we focused on a single amino acid with four synonymous codons ( codon1 to codon4 ) that are respectively recognized by four distinct tRNA species ( tRNA1 to tRNA4 ) . We assume that the relative concentrations of the four tRNA species are 20 , 21 , 22 , and 23 , respectively . The digital organism has ten genes with relative ( mRNA and protein ) expression levels from 20 to 29 , respectively . These genes each have 12 codons that are sampled from the four synonymous codons . We started the simulation with exactly the same usage of the four synonymous codons in each gene . Synonymous mutations among codons all have the same rates and the total mutation rate per genome is assumed to be one synonymous change per generation . The relative CST for a codon is assumed to equal the number of times the codon is used in translation divided by the number of corresponding tRNA molecules . The total time ( T ) required for translating all the proteins can be considered as the generation time . T can be calculated by summing up the CSTs of all codons in all transcripts if there is only one ribosome in the cell . If there are m ribosomes in the cell , the time required would simply be m times shorter . Thus , without loss of generality , we assume m = 1 . A strain with a shorter generation has a higher fitness and will spread in the population . Genetic drift is simulated by random sampling of cells for the next generation . The population size is 104 individuals and the simulation lasts for 500 generations . We repeated the simulation 1000 times . Our results did not change when we simulated the evolution for more generations . By contrast , when we removed the natural selection for translational efficiency in simulation , the phenomena observed in Figure 4 disappeared ( Figure S9 ) . Note that , in the simulation , we allow codon usage to evolve while fixing tRNA concentrations . If tRNA concentrations evolve while the codon usage is fixed , we also expect to observe the rebalance of codon-tRNA usage , but the correlation ( or the lack of ) between CUB and gene expression level will not change during this evolutionary process . In reality , tRNA concentrations and synonymous codon usage likely co-evolve to regain the balance . As long as codon usage is allowed to evolve , we expect stronger CUB to appear in more highly expressed genes , as demonstrated in Figure 4 . | Although an amino acid can be encoded by multiple synonymous codons , these codons are not used equally frequently in a genome . Biased codon usage is believed to improve translational efficiency because it is thought that preferentially used codons are translated faster than unpreferred ones . Surprisingly , we find similar translational speeds among synonymous codons . We show that translational efficiency is optimized by a previously unknown mechanism that relies on proportional use of codons according to their cognate tRNA concentrations . Our results provide important molecular details of protein translation , answer why codon usage is unequal , demonstrate widespread natural selection for translational efficiency , and can guide designs of synthetic genomes and cells with efficient translation systems . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genomics",
"molecular",
"cell",
"biology",
"genetics",
"biology",
"computational",
"biology",
"evolutionary",
"biology",
"genetics",
"and",
"genomics"
] | 2012 | Balanced Codon Usage Optimizes Eukaryotic Translational Efficiency |
Depending on the epidemiological setting , a variable proportion of leprosy patients will suffer from excessive pro-inflammatory responses , termed type-1 reactions ( T1R ) . The LRRK2 gene encodes a multi-functional protein that has been shown to modulate pro-inflammatory responses . Variants near the LRRK2 gene have been associated with leprosy in some but not in other studies . We hypothesized that LRRK2 was a T1R susceptibility gene and that inconsistent association results might reflect different proportions of patients with T1R in the different sample settings . Hence , we evaluated the association of LRRK2 variants with T1R susceptibility . An association scan of the LRRK2 locus was performed using 156 single-nucleotide polymorphisms ( SNPs ) . Evidence of association was evaluated in two family-based samples: A set of T1R-affected and a second set of T1R-free families . Only SNPs significant for T1R-affected families with significant evidence of heterogeneity relative to T1R-free families were considered T1R-specific . An expression quantitative trait locus ( eQTL ) analysis was applied to evaluate the impact of T1R-specific SNPs on LRRK2 gene transcriptional levels . A total of 18 T1R-specific variants organized in four bins were detected . The core SNP capturing the T1R association was the LRRK2 missense variant M2397T ( rs3761863 ) that affects LRRK2 protein turnover . Additionally , a bin of nine SNPs associated with T1R were eQTLs for LRRK2 in unstimulated whole blood cells but not after exposure to Mycobacterium leprae antigen . The results support a preferential association of LRRK2 variants with T1R . LRRK2 involvement in T1R is likely due to a pathological pro-inflammatory loop modulated by LRRK2 availability . Interestingly , the M2397T variant was reported in association with Crohn’s disease with the same risk allele as in T1R suggesting common inflammatory mechanism in these two distinct diseases .
Leprosy is a chronic dermato-neurological infectious disease caused by M . leprae . Leprosy irrespective of its clinical presentation is curable by multi-drug therapy . The current global effort in leprosy control is focused on early recognition of leprosy cases and prevention of permanent disabilities [1] . A common complication of leprosy are excessive pro-inflammatory episodes termed type-1 reactions ( T1R ) [2] . If untreated , T1R can lead to irreversible nerve function impairment due to a pathological cellular immune response directed against host peripheral nerve cells [3] . Up to 50% of all leprosy cases can undergo T1R with the incidence varying according to endemic settings and criteria for case definition [4] . Why only a proportion of leprosy patients undergo T1R is not known . However , clinical and environmental factors have been associated with T1R outcome [5 , 6] . Individuals categorized as borderline in the Ridley and Jopling clinical spectrum of leprosy are at increased risk to develop T1R while patients with tuberculoid or lepromatous polar leprosy forms rarely develop T1R [6] . Positive bacillary index , PCR detection of M . leprae , and increased age at leprosy diagnosis are other factors associated with T1R-risk [6–8] . A number of studies have shown a consistent upregulation of pro-inflammatory cytokines , i . e . TNF , IFNγ and the chemokine IP10 , in the blood of T1R patients [4 , 9–12] . In a prospective study , the transcriptional profile of leprosy cases destined for T1R displayed a distinct signature from leprosy patients that remained T1R-free [11] . A dysregulated balance between innate pro- and anti-inflammatory responses emerged from this study as a key factor in T1R outcome [11] . A number of genes have been shown to be associated with T1Rincluding TNFSF15 and the pathogen recognition genes TLR1 , TLR2 and NOD2 [3 , 13] . All of these genes had also been found to be associated with leprosy per se in other studies [14–16] . Since leprosy patients are usually not stratified by their T1R status , it is possible that some of the leprosy per se associations were caused by T1R patient subgroups . For example , the TNFSF15 gene had initially been shown to be associated with leprosy per se by a genome wide association study ( GWAS ) in a Chinese population [14] . However , in a Vietnamese sample stratified by T1R status the association signal could be unambiguously assigned to the T1R group [13] . Of the genes reported by the GWAS , the TNFSF15 and LRRK2 were the only leprosy susceptibility genes not validated for association with leprosy per se in a Vietnamese population [17] . Like TNFSF15 , LRRK2 is a gene with an uncertain role in leprosy per se susceptibility . Several groups have evaluated the association of the LRRK2 gene with leprosy susceptibility but results were inconsistent [17–20] . Given that T1R affects different proportions of leprosy cases according to the studied population , we wondered if inconsistencies in LRRK2 association with leprosy per se were due to different proportions of T1R in each setting . Here , we evaluated a possible role for LRRK2 in T1R-affected families and contrasted the results with T1R-free families . We identified a set of 18 SNPs in LRRK2 preferentially associated with T1R . These variants overlapped with previous associations reported for Crohn’s Disease ( CD ) , Ulcerative Colitis ( UC ) and Inflammatory Bowel Disease ( IBD ) [21 , 22] .
For the LRRK2 study , a total of 1372 individuals were selected [13] . These individuals were divided in two family-based samples . The first set of families contained 229 leprosy affected offspring that underwent T1R ( T1R-affected ) and their respective parents . The T1R-free subset was matched to the T1R-affected subset by leprosy clinical subtype of the offspring ( Fig 1 ) . Consequently , the second set of families included 229 leprosy affected offspring and their parents in which the offspring had no signs of leprosy reaction ( T1R-free ) . There was no difference in gender and age at leprosy onset regarding T1R outcome between both subsets ( S1 Table ) . The subjects included in the eQTL analyses were part of a study evaluating the transcriptional profile of leprosy patients prior to T1R onset [23] . Briefly , 53 newly diagnosed leprosy cases in the borderline spectrum ( 19 BT , 30 BB and 4 BL ) were enrolled . A blood sample was collected from each subject within 3 months of leprosy diagnosis and none of the subjects suffered T1R at enrolment . The samples used in the current study were selected from our records at the Dermato Venerology Hospital , Ho Chi Ming City , Vietnam , as described previously [13] . Written informed consent was obtained from all subjects enrolled in the study and all subjects were anonymized . This study was approved by the regulatory authorities in Ho Chi Minh City , Vietnam , and the Research Ethics Board at the Research Institute of the McGill University Health Centre , Montreal , Canada . The investigation have been conducted according to the principles expressed in the Declaration of Helsinki . Genotypes for156 SNPs mapping to a 500 kb window overlapping the LRRK2 and MUC19 genes were obtained via the 660W-quad v1 Illumina array [24] . The variants selected covered 89% of the SNPs with a MAF > 5% at a r2 > 0 . 5 for the Vietnamese ( KHV ) and Chinese ( CHB ) and 84% for Caucasians ( CEU ) populations from the 1000 genomes project [25] . All genotypes passed standard quality-control presenting call rates greater than 98% , less than 2 Mendelian Errors ( ME ) and were in Hardy-Weinberg Equilibrium ( HWE ) with P > 0 . 05 in 763 leprosy unaffected parents from both T1R-affecte and T1R-free subsets . LRRK2 expression levels were obtained using Illumina HumanHT12 v4 BeadChips as previously described [23] . Briefly , whole blood from the 53 leprosy patients was divided in two aliquots . One aliquot was stimulated with M . leprae sonicate for 24hrs to 30 hrs while the second aliquot was incubated for the same time interval in the absence of M . leprae sonicate ( non-stimulated ) . Total RNA was extracted from all aliquots and used for LRRK2 quantification . Family based SNP and haplotype association tests were performed using Transmission Disequilibrium Test ( TDT ) as implemented in FBAT 2 . 0 . 4 [26] . Association testing was carried out under the same genetic model in T1R-affected and T1R-free families and the P values for the best genetic model ( additive or dominant ) were displayed ( Fig 1 ) . Due to the highly correlated nature of the genotyped SNPs , we did not perform a correction for multiple testing . Subsequently , a formal heterogeneity test was performed to evaluate preferential association of genetic variants with T1R ( Fig 1 ) by using a modified version of the FBAT statistics ( FBATHet ) as described by Gaschignard , J . et al [27] . A multivariate analysis was performed to test for independence of T1R-specific associations . For each SNP bin ( r2 > 0 . 5 ) , SNP with the most significant evidence for association was included in the multivariate model . Multivariate analyses were done by stepwise conditional logistic regression ( SAS v . 9 . 3 ) . Logistic regression was also used to estimate the odds ratio for each individual SNP in the T1R-affected subset . Briefly , the TDT evaluates the non-random transmission of alleles from heterozygote parents to affected offspring . We used the non-transmitted alleles from the TDT to create up to three unaffected pseudo-sibs per family , one for each possible genotype . We compared the original T1R-affected offspring with T1R-unaffected pseudo sibs in a matched case-control design as described in [28] . Under the additive model , the TDT and the conditional logistic regression statistics result in the same P values . The recombination rate in centimorgan by mega base ( cM/Mb ) according to the 1000 genomes and the LD structure of the LRRK2/MUC19 locus were obtained with the R packages Locuszoom v . 1 . 3 [29] and snp . ploter v . 0 . 5 . 1 [30] , respectively . The minor allele frequencies ( MAF ) and HWE for each SNP were estimated using Haploview 4 . 2 [31] . For the eQTL analyses , the correlation between genotypes and gene expression levels was performed with a simple linear regression under both stimulated and non-stimulated conditions using R 3 . 2 . 0 . Genotype associations of LRRK2 and PD were obtained from the PDgene database ( www . pdgene . org ) . LRRK2 variants associated with PD were obtained from a GWAS meta-analysis of 14 studies with a total population sample of 12 , 771 PD cases and 93 , 386 controls [32] . Genotype associations of LRRK2 variants with IBD and CD were obtained from the IBDgenetics database ( www . ibdgenetics . org ) . The IBDgenetics population sample consisted of 42 , 950 IBD cases ( 22 , 575 CD and 20 , 417 UC patients ) and 53 , 536 controls [21 , 22] .
A genome-wide association ( GWAS ) in a Chinese sample provided the most comprehensive study of LRRK2 in leprosy . Hence , we first analyzed six LRRK2 SNPs ( rs1873613 , rs10878220 , rs1491938 , rs12820920 , rs11174812 and rs11173979 ) described in association with leprosy per se by a GWAS in a Chinese population . None of the six variants presented significant evidence of association with leprosy in the T1R-free subset ( Table 1 ) . In contrast , SNPs rs10878220 , rs1491938 and rs12820920 that belonged to the same r2 < 0 . 5 SNP bin were nominally associated with disease in the T1R-affected subset ( Table 1 ) . These SNPs were preferentially associated with T1R when a formal heterogeneity test was performed ( Table 1 ) . Moreover , the direction of association was the same as previously reported for leprosy per se . While SNP rs10878220 is located in the core promoter region of the LRRK2 gene , SNPs rs1491938 and rs12820920 are LRRK2 intronic variants ( Table 1 ) . The remaining three SNPs that did not show association with T1R are located 66kb to 367kp upstream of LRRK2 transcription start site ( Table 1 ) . To further evaluate the association of LRRK2 with T1R , an additional 150 SNP were selected from the LRRK2 gene region . Of all 156 SNPs evaluated for association in the T1R-affected subset , 34 showed P < 0 . 05 including the initial three SNPs associated with T1R ( Fig 2 ) . In the T1R-free family subset , the SNP rs7972711 situated near the 3’ end of the MUC19 gene and rs10878434 , rs7303525 and rs11564172 located near 3`end of LRRK2 provided evidence of association ( P < 0 . 05; Fig 2 ) . None of the 34 SNPs associated with T1R showed evidence of association with leprosy in the T1R-free families ( Fig 2 ) . When formally tested for heterogeneity of association , 18 out of the 34 SNPs were preferentially associated with T1R ( Fig 2 and Table 1 ) . These 18 SNPs preferentially associated with T1R belong to two extended SNP bins ( r2 > 0 . 5 ) and two single SNP bins ( Table 1 ) . Notably , although not nominally significant these 18 SNPs showed the opposite allelic enriched in T1R-free subset relative to the T1R-affected subset ( Table 1 ) . The bin tagged by the missense M2397T variant ( rs3761863 ) situated in the WD40 domain of LRRK2 displayed the strongest preferential association with T1R ( P = 0 . 003; odds ratio ( OR ) = 1 . 49; 95% confidence interval ( CI ) = 1 . 12–1 . 97 and P Het = 0 . 008 for M2397 allele under an additive model ) in the same direction as previously reported for CD ( Table 2 ) . To test if the r2 > 0 . 5 SNP bins were independently associated with T1R we performed a multivariate analysis including the most significantly associated SNP for each bin ( Table 2 ) . The multivariate analysis identified the missense M2397T variant as the main association with T1R ( P multi = 0 . 01; Table 2 ) . However , a trend for independent association was observed for the SNP bin tagged by rs1031996 ( P multi = 0 . 05 ) and the single SNP bin rs1463739 ( P multi = 0 . 13; Table 2 ) while the association of T1R and rs1427271 disappeared . We further investigated the combined effect of M2397T and rs1031996 by conducting a haplotype analysis . We found that the M2397 allele was mostly observed in the presence of the rs1031996 C allele ( 48% of the T1R-affected offspring ) and this haplotype displayed a strong risk effect ( Table 3 ) . Importantly , the trend towards protection observed on the T2397 background was not modulated by rs1031996 alleles . Hence , the haplotype analysis confirmed that the main effect of the LRRK2 gene on T1R susceptibility was driven by the M2397T variant . No haplotype association was observed when leprosy per se was considered as phenotype . To investigate if SNP alleles associated with T1R were correlated with LRRK2 transcriptional levels we performed an eQTL analysis . Of the 18 SNPs preferentially associated with T1R , nine variants belonging to the same SNP bin as the missense M2397T variant ( r2 = 0 . 5 ) were eQTLs for LRRK2 in non-stimulated whole blood of 53 individuals ( Table 1 , S1 Fig ) . When a tighter LD threshold ( r2 = 0 . 8 ) was considered , the eQTL variants were separated from the bin tagged by M2397T . This observation suggested that the modest eQTL effect of M2397T was due to the linkage disequilibrium with the causal eQTL . The strongest eQTL effect was observed for SNP rs2404580 with the T1R-risk “T” allele being associated with higher LRRK2 expression in unstimulated cells ( P = 5 . 1E-05; Fig 3 ) . Following stimulation with M . leprae sonicate , an abrogation of the eQTL effect was observed for all nine SNPs ( Figs 3 and S1 ) . Clinical subtypes of leprosy had no detectable impact on the eQTL effect of LRRK2 genotypes .
We identified an amino acid change M2397T in the WD40 domain of LRRK2 and a bin tagged by the variant the variant rs1031996 as being associated specifically with T1R . LRRK2 is a protein that exerts a diverse set of functions . LRRK2 mediates catalytic processes through its enzymatic ROC/COR domain; facilitates signal transduction through a MAPK domain and interacts with other proteins through three scaffold domains , an ankylin repeat ( ANK ) , a leucine rich repeat ( LRR ) and a WD40 repeat domain [33] . Scaffold domains are also responsible for protein conformation and stability [33] . Although , LRRK2 displays multiple functions , the association of T1R with a variant in the WD40 domain of LRRK2 suggests that protein conformation and/or stability are key factors in T1R . Indeed , the M2397T variant in LRRK2 has previously been shown to impact on LRRK2 protein turnover [34] . The half-life of LRRK2 with the T1R-risk allele M2397 had been estimated at approximately 8 hrs which is substantially shorter than the estimated 18 hrs half-life of the T2397 allele of LRRK2 ( Fig 4 ) [34] . Cytoplasmic LRRK2 forms a complex that arrests nuclear factor of activated T-cells ( NFAT ) in the cytoplasm [34] . A consequence of LRRK2 deficit in the cytoplasm is the translocation of NFAT to the nucleus , which strongly induces the transcription of pro-inflammatory cytokines ( Fig 4 ) [35 , 36] . Thus , the association of the M2397 allele with T1R-risk is in agreement with an exacerbated pro-inflammatory response in T1R cases ( Fig 4 ) . While the M2397T variant is a strong candidate for functional impact in T1R , the association of the bin tagged by rs1031996 requires further investigations . A set of eQTLs for LRRK2 was observed in non-stimulated cell from leprosy patients . Due to LD the M2397 allele and the eQTL alleles that correlated with increased LRRK2 expression were preferentially observed on the same haplotype . This suggests that the effect of M2397 on faster LRRK2 turn-over is mitigated by higher levels of LRRK2 message . Interestingly , in the presence of mycobacterial antigen the compensatory effect of higher LRRK2 message for the more rapid turnover of the M2397 protein is strongly abrogated . This implies that the genetic effect of the M2397 allele will be more pronounced in the presence of mycobacterial antigen while it may be largely abolished without such antigen exposure . While these in-vitro findings in whole blood will need to be validated employing purified cell types and in-vivo studies in leprosy lesion , the results obtained provide an example of how environmental stimuli can modulate germline encoded genetic risk factors ( Figs 3 and 4 ) . LRRK2 variants associated with T1R overlapped previous reported LRRK2 associations with CD and PD [21 , 22 , 32] . In PD , rare coding variants in the enzymatic and kinase domains of LRRK2 were shown to be causally linked to PD [33] . In addition , common LRRK2 variants were shown in association with PD [32] . Two of these variants , rs1491932 and rs7970326 , were observed in association with T1R and borderline evidence T1R specificity for the alleles opposite to PD . Common variants may tag rare variants with stronger effects . However , given our sample size we were unable to evaluate the role of rare variants in T1R . Of 18 SNPs preferentially associated with T1R , 17 were nominally associated with IBD with the same risk allele observed for T1R [22] . The only exception was rs1463739 a SNP located outside of LRRK2 in the 3’ region of MUC19 ( Table 2 and Fig 2 ) . When considering IBD subtypes , 14 T1R-risk variants were associated with risk of CD and 11 with risk of UC [22] . The M2397 allele is a risk factor for T1R , CD and UC , suggesting that a faster LRRK2 turnover leads to an increased pro-inflammatory response that is common to these diseases . T1R and CD may share susceptibility to mycobacterial species as common etiology [37 , 38] . In PD , pathogen involvement is controversial although some studies suggested that Helicobacter pylori and prions might play a role in disease susceptibility [39–41] . Further studies will be needed to understand the precise role of LRRK2 in the pathogenesis of these three diseases . | A major challenge of current leprosy control is the management of host pathological immune responses coined Type-1 Reactions ( T1R ) . T1R are characterized by acute inflammatory episodes whereby cellular immune responses are directed against host peripheral nerve cells . T1R affects up half of all leprosy patients and are a major cause of leprosy-associated disabilities . Since there is evidence that host genetic factors predispose leprosy patients to T1R , we have conducted a candidate gene study to test if LRRK2 gene variants are T1R risk factors . The choice of LRRK2 was motivated by the fact that LRRK2 was associated with leprosy per se in some but not in other studies . We reasoned that this may reflect different proportions of leprosy patients with T1R in the different samples and that LRRK2 may in truth be a T1R susceptibility gene . Here , we show that variants overlapping the LRRK2 gene , reported as suggestive leprosy per se susceptibility factors in a previous genome-wide association study , are preferentially associated with T1R . The main SNP carrying most of the association signal is the amino-acid change M2397T ( rs3761863 ) which is known to impact LRRK2 turnover . Interestingly , eQTL SNPs counterbalanced the effect of the M2397T variant but this compensatory mechanism was abrogated by Mycobacterium leprae antigen stimulation . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"genome-wide",
"association",
"studies",
"mycobacterium",
"leprae",
"medicine",
"and",
"health",
"sciences",
"population",
"genetics",
"tropical",
"diseases",
"alleles",
"multivariate",
"analysis",
"bacterial",
"diseases",
"linkage",
"disequilibrium",
"mathematics",
"statis... | 2016 | A Missense LRRK2 Variant Is a Risk Factor for Excessive Inflammatory Responses in Leprosy |
Successful embryogenesis is a critical rate limiting step for the survival and transmission of parasitic worms as well as pathology mediated by them . Hence , blockage of this important process through therapeutic induction of apoptosis in their embryonic stages offers promise for developing effective anti-parasitic measures against these extra cellular parasites . However , unlike in the case of protozoan parasites , induction of apoptosis as a therapeutic approach is yet to be explored against metazoan helminth parasites . For the first time , here we developed and evaluated flow cytometry based assays to assess several conserved features of apoptosis in developing embryos of a pathogenic filarial nematode Setaria digitata , in-vitro as well as ex-vivo . We validated programmed cell death in developing embryos by using immuno-fluorescence microscopy and scoring expression profile of nematode specific proteins related to apoptosis [e . g . CED-3 , CED-4 and CED-9] . Mechanistically , apoptotic death of embryonic stages was found to be a caspase dependent phenomenon mediated primarily through induction of intracellular ROS . The apoptogenicity of some pharmacological compounds viz . DEC , Chloroquine , Primaquine and Curcumin were also evaluated . Curcumin was found to be the most effective pharmacological agent followed by Primaquine while Chloroquine displayed minimal effect and DEC had no demonstrable effect . Further , demonstration of induction of apoptosis in embryonic stages by lipid peroxidation products [molecules commonly associated with inflammatory responses in filarial disease] and demonstration of in-situ apoptosis of developing embryos in adult parasites in a natural bovine model of filariasis have offered a framework to understand anti-fecundity host immunity operational against parasitic helminths . Our observations have revealed for the first time , that induction of apoptosis in developing embryos can be a potential approach for therapeutic intervention against pathogenic nematodes and flow cytometry can be used to address different issues of biological importance during embryogenesis of parasitic worms .
Helminth infections account for the highest burden of neglected tropical diseases [NTD] , which afflict the most impoverished population of the world . About two billion people are presently infected with these parasites while many more people living in endemic areas are at risk of acquiring these infections [1]-[3] . Chronic diseases caused by such metazoan parasites often inflict crippling morbidity and debilitating disability with profound economic , social and political consequences [1] . Accumulating evidence in literature suggests that coinfection with these helminth parasites increases susceptibility to or worsens progression of three major infectious diseases- HIV/AIDS , tuberculosis and malaria [4] , [5] . The above facts accentuate the need for launching a global assault on parasitic worms . However , our ability to control diseases caused by these class of parasites is constrained by several factors including a limited repertoire of sub-optimal drugs and paucity of robust tools to investigate biology of nematode parasites [6] . The burden of human helminthiasis is mostly attributed to high prevalence diseases caused by pathogenic nematodes i . e . by intestinal nematodes - Ascariasis [807 million] , Trichuriasis [604 million] , Hook worm infections [574 million] and filarial nematodes - Lymphatic filariasis [120 million] and Onchocerciasis [37 million] respectively[1] , [7] . Preventive chemotherapy through MDA programs is the mainstay for treatment and control of diseases caused by nematode pathogens , at present . However , constraints of currently available therapies including high cost , low therapeutic efficacy , rapid reinfection after treatment , poor safety profiles and patient compliance and emerging or existing drug resistance coupled with lack of robust biomarkers for detection of resistance of nematode parasites to the mainstay drugs of MDA programs etc . limit the utility of existing drugs [4] , [8]–[12] . The problem is further compounded by the fact that none of the existing drugs are effective against all the life stages of the parasitic worms and almost all of them are remarkably ineffective against adult stage parasites [6] , [8] . Additionally , poor understanding of the mode of action , pharmacology [6] , [8] , [13] and adverse side reactions associated with the front line antifilarial drugs e . g . Diethylcarbamazine [DEC] and Ivermectin [13] , [14] etc . are issues of great concern . On the other hand , a number of studies on human immune responses to helminthic infections have been carried out so far , but no clear immune effector mechanism has emerged on which an effective vaccine could be designed against these parasites [2] , [15] . Hence , control of these major tropical diseases warrants radically new approaches to chemotherapy as well as vaccine design . Induction of apoptosis in parasites for drug development is a novel possibility that has been explored in great detail in several unicellular pathogens [16] but has not been evaluated so far , against metazoan helminth parasites . As a form of programmed cell death , apoptosis was initially described and extensively studied in the free living soil nematode C . elegans . Homologs of mammalian apoptosis related proteins have been identified in a trematode parasitic worm Schistosoma and function of some of these proteins has been studied by over-expressing them in artificial systems viz . , mammalian cells [17] . However , no attempts have been made so far , to demonstrate functional apoptosis in parasitic worms . Presence of a thin and transparent cuticle , availability of complete genetic database with detailed molecular analysis and standardized protocols for prolonged maintenance in artificial culture medium [without requirement of susceptible animal hosts , as in case of pathogenic nematodes] , have made dissection of pathways of cell death possible in free living nematode C . elegans . Pathogenic nematodes however , don't offer such a luxury to study apoptosis in them . Unlike free living nematodes , pathogenic nematodes of vertebrates have evolved distinct life cycles and engrossed numerous adaptations [e . g . modification of cuticle , elaborate reproductive capacity with enormous fecundity , optimum temperature for growth matching their vertebrate hosts i . e . close to 37°C as opposed to 15–25°C for C . elegans etc . ] to suit their parasitic mode of life . Further , pathogenic nematodes very often establish chronic infections and are exposed to a volley of stress inducing factors/molecules inside their vertebrate hosts which can potentially trigger distinct pathway of apoptosis in them . Thus , induction of apoptosis in metazoan parasites might be an important arm of host immunity operational in helminth infections which is yet to be established in literature . Successful embryogenesis is a critical rate limiting step for the survival and transmission of parasitic worms as well as pathology mediated by them [18] . Hence , blockage of this important process through therapeutic induction of apoptosis in their developing embryos offers a promising approach for developing intervention strategies against these extra cellular parasites and underscores the need for studying apoptotic pathways in pathogenic nematodes . There have been no attempts so far , to study programmed cell death in nematode parasites , primarily due to lack of sensitive assay systems and the present study attempts to fill this lacuna . We have used here Setaria digitata , a bovine filarial parasite as a model to demonstrate induction of apoptosis in developing embryos i . e . eggs and larvae [L-1] of a pathogenic nematode , using confocal microscopy and flow cytometry based quantitative assays , for the first time . Through these assays we have convincingly demonstrated for the first time , multiple conserved features of apoptosis in embryonic stages of a pathogenic filarial nematode . Further , in the present study we have demonstrated characteristic features of apoptosis e . g . mitochondrial depolarization , redistribution of Cytochrome-c , formation of hypodiploid nuclei etc . which has not been demonstrated in the model nematode C . elegans , so far . We have gone one step ahead to evaluate the effect of some of the existing pharmacological compounds for their apoptogenicity which has opened up avenues for development of embryogenesis blocking drugs against nematode pathogens . Finally , the induction of apoptosis in embryonic stages by lipid peroxidation product molecules , commonly associated with inflammatory responses in filarial disease [19] and demonstration of in-situ apoptosis of developing embryos in adult female parasites in a natural bovine model of filariasis [20] have offered a framework to understand anti fecundity host immunity against parasitic worms .
No investigations were undertaken using humans/human samples in this study . No experimental animals were used to conduct any of the experiments reported in this manuscript . Adult Setaria digitata worms from peritoneal cavities and blood samples from slaughtered cattle were collected from the abattoir attached to a local zoo after obtaining approval from zoo authorities . The animals are slaughtered in the abattoir regularly for feeding the wild cats and no animals were slaughtered specifically for the purpose of our study . Hanks Balanced Salt Solution [HBSS] medium , RPMI-1640 medium , glucose , Penicillin , Streptomycin , Amphotericin – B , pancaspase inhibitor Z-VAD–FMK [N-Benzyloxycarbonyl- Val-Ala-Asp[O-Me] fluromethyl ketone] , Colorimetric caspase substrate Ac-DEVD-pNA [N-Acetyl- Asp-Glu-Val-Asp p-nitroanilide] , Fluorogenic substrate for reactive oxygen species H2-DCFDA[2′–7′-Dichloro dihydro fluorescein diacetate] and anti- goat IgG –FITC antibodies were purchased from Sigma . Annexin-V- PE apoptosis detection kit-I , Mitoscreen JC-1 kit , Apo-Direct apoptosis detection kit and fluorochrome conjugated antibodies for cleaved PARP ( Poly [ADP-ribose] polymerase ) were purchased from BD Biosciences . Fluorochrome conjugated antibodies to Cytochrome-c was procured from eBioscience . Primary goat antibodies to CED-3 , CED-4 and CED-9 and anti-goat IgG-PE were purchased from Santacruz . Mito Tracker red CMX Ros was purchased from Invitrogen Molecular Probes . Lipid Peroxidation Products were purchased from Sigma Aldrich . Adult female filarial worms Setaria digitata were collected from the peritoneum of cattle , slaughtered at a nearby abattoir and transported to the laboratory in sterile Hanks Balanced Salt Solution [HBSS] medium [Sigma H 2387] . The medium containing 1% glucose [Sigma G 7525] , Penicillin 100 units/ml , Streptomycin 100 µg/ml [Sigma P 4333] , and Amphotericin - B 0 . 25 µg/ml [Sigma A2942] was buffered with NaHCO3 [Sigma S 5761] . About 5 to 7 worms were taken in a petridish , washed three times in medium , dissected into small pieces in 10 ml of medium under sterile conditions and incubated at 37°C for 30 minutes to allow the release of embryonic stages [eggs and microfilariae] into the medium . The embryonic stages were harvested into sterile 15 ml centrifuge tubes and washed three times by centrifuging at 300 g for 10 minutes each with medium and the final pellet was suspended in 1 ml of RPMI-1640 medium [Sigma R 8005] supplemented with 10% FBS [Sigma F 2442] . One ml each from a suspension containing 1×105 embryonic stages/ml was dispensed into individual wells of a 24 well tissue culture plate . Control set of wells were left untreated . Other sets of wells containing suspension of developing embryos were subjected to treatment with various agents separately at a concentration ranging from 10–100 µM for 24–48 hr . at 37°C in a 5% CO2 incubator . Motility of microfilariae as a marker of viability of embryonic stages , after treatment with different agents was scored under an inverted microscope [Video S1 , S2 and S3] . Viability of embryonic stages was also checked by propidium iodide staining . In the present study 3 replicates of each treatment type were used except for the lipid peroxidation products for which 5 replicates each were used . Following treatment and incubation , embryonic stages of S . digitata were analyzed either live or after fixation with 1% paraformaldehyde in a flow cytometer [BD FACS Calibur , Becten and Dickinson , USA] . Apoptosis of different developmental stages were studied by gating the respective population in the Dot Plots . Embryonic stages were stained for externalization of phosphatidylserine , mitochondrial depolarization and fragmentation of chromosomal DNA using Annexin-V: PE apoptosis detection kit-I [BD Biosciences 559763] , Mitoscreen JC-1 kit [BD Biosciences 551302] , Apo-Direct apoptosis detection kit [BD Biosciences 556381] respectively , as per the manufacturer's instructions . Intra cellular profile of apoptosis related proteins was studied using either fluorochrome conjugated antibodies directly viz , . anti-Cytochrome c-FITC [e Bioscience 11-6601-82]/ anti cleaved PARP–PE [BD Biosciences 552933] , or primary goat antibodies [1: 100 diluted] such as anti CED-3 [Santacruz sc- 9192] /anti CED-4 [Santacruz sc-9193] /anti CED-9 [Santacruz sc- 9202] followed by probing with fluorochrome conjugated secondary antibodies [1∶100 diluted] viz . , anti-goat IgG-PE [Santacruz sc-3747]/ anti- goat IgG -FITC [Sigma F7367] . Incubation for both primary and secondary antibodies was performed for 1 hr at 37°C in dark . Analysis was performed on 10 , 000 acquired events in flow cytometer using Cell quest Pro software . Subcellular distribution and colocalization profile of Cytochrome-c and CED-4 were studied by analyzing the immuno-stained embryonic stages under confocal microscope [Leica , Germany] equipped with a 63X objective after mounting them in 30% glycerol on glass slides . Sub cellular localization of CED-4 in developing embryos was confirmed by colocalization experiments with Mito Tracker red CMX Ros [Invitrogen Molecular Probes M 7512] , used at a final concentration of 50 nM [incubation was done for 45 minutes on ice] . Activity of caspase family of cysteine proteases was confirmed by detecting the presence of cleaved PARP through inta cellular staining , using fluorochrome conjugated antibodies to cleaved PARP [BD Biosciences 552933] , studying the inhibition of externalization of phosphatidylserine using the cell permeable pan caspase inhibitor Z-VAD–FMK [Sigma V 116] , used at a concentration of 50 µM [preincubation was done 1 hr before treatment] by flow cytometry and a colorimetric assay by incubating the cell lysate of the in-vitro cultured and treated embryonic stages with the colorimetric Caspase substrate Ac-DEVD-pNA [Sigma A 2559] for 2 hrs at 37°C at a final concentration of 200 µM . The intracellular accumulation of ROS was measured in terms of DCF fluorescence by using the fluorescent probe [2 , 7 , Dichloro Dihydro Fluorescein Diacetate] H2-DCFDA as described previously [21] , [22] with little modification . Briefly , the cultured developing embryos of S . digitata were pre incubated with 5 µM H2-DCFDA [Sigma D 6883] before treatment with different pharmacological agents mentioned earlier . The control set of embryonic stages were also subjected to the same manipulation except for treatment with any of the agents . After 24 hrs of incubation , the developing embryos were harvested washed with PBS twice and analyzed by flow cytometer . Inhibition of intracellular ROS was studied using antioxidant N-Acetyl-L-Cysteine [Sigma A 7250] . The amino acid sequence of Cytochrome-c of C . elegans [target] was retrieved from the sequence database of NCBI [P19974] and its 3-D structure was generated by homology modeling , using the academic version of MODELLER9v6 software [23] . The crystal structure of Cytochrome-c from bovine heart [PDB code: 2B4Z] [24] which shares sequence identity and similarity of 61% and 74% respectively , with the target was taken as a template for the modeling . Out of 20 models generated by MODELLER , one with the best G-score of PROCHECK [25] and VERIFY -3 D [26] profiles was selected and subjected to energy minimization using CHARMM [27] force field . The final stable 3-D structure of Cytochrome-c of C . elegans was used for molecular docking with CED-4 , using Patch Dock algorithm . The top 20 solutions , out of about 60 predicted Cytochrome-c-CED4 complexes were sorted according to their geometric shape , followed by refinement and ranking of the models using Fire Dock server . The Fire Dock top rank was sent to PDB sum server to predict the protein-protein interactions . We have also retrieved the amino acid sequence of Cytochrome-c of human filarial parasite Brugia malayi [target] from the sequence database of NCBI [Accession NO . XP_001897096] and generated it's 3-D structure by homology modeling , using the academic version of MODELLER9v6 software [23] as described above . 1CCR was taken as a template for the modeling . To ascertain whether apoptogenic compounds present in the external milieu surrounding live worms can exert their effect on embryonic stages present in the uterine cavity of intact adult female parasites , we designed a simulated experiment where we incubated , randomly selected adult female worms of comparable size [4 each] in two different 50ml sterile vials containing RPMI-1640 medium with 10% FBS . One set of worms was treated with 10 µM of Plumbagin while the other set was left untreated and regarded as control . After overnight incubation , developing embryos harvested separately from each set of cultured worms was subjected to ex-vivo analysis for apoptosis in terms of expression of caspase homologue , CED-3; cleavage of intracellular caspase substrate PARP; and fragmentation of chromosomal DNA . Female adult worms were collected from the peritoneum of infected cattle , slaughtered in a nearby abattoir . Along with worms , 10 ml of blood was also collected in separate sterile acid citrate dextrose [ACD] vial from each animal . After lysing the RBCs in blood by distilled water washing , the final cell pellet was resuspended in TEBS buffer and passed through 5 µM polycarbonate membrane [Nucleopore , USA] . The membrane filter was then checked for retention of microfilariae under light microscope . The result was further confirmed by Giemsa stained thick blood smears . Depending upon the presence or absence of microfilariae , the naturally infected bovine hosts were categorized into microfilaraemic and amicrofilaraemic groups . The suspension of developing embryos harvested from adult female worms of each animal was divided into two fractions , followed by staining of one fraction with dUTP – FITC in the presence of TdT enzyme[experimental fraction] and the other fraction with dUTP – FITC in the absence of TdT enzyme[control fraction] ex-vivo . The GMI ( Geometric mean intensity ) of dUTP-FITC fluorescence for endogenous fragmentation of DNA as a measure of apoptosis in the developing embryos was obtained after subtracting the GMI of dUTP-FITC fluorescence of control fraction [representing back ground fluorescence] from that of experimental fraction . Apoptosis in terms of extent of DNA fragmentation in microfilariae , early and late embryonic stages was scored after gating respective population in the Dot Plots as described earlier in this study . Paired comparison were conducted using paired t – test , and all data are presented as mean value ± SEM . Differences were considered significant at 95% confidence levels . All statistical analysis was performed with Graph Pad Prism , version 5 . 0 .
In the absence of suitable quantitative assay systems to score apoptosis in nematode embryos we developed flow cytometry based assays to initially demonstrate multiple conserved features of apoptosis in developing embryos of S . digitata [Table 1] . Apoptosis of developing embryos , which includes microfilariae /larval stage-1 and eggs , was scored by gating the respective populations [Figure 1A] . The identity of microfilariae and egg populations in dot plots for developing embryos of S . digitata was previously reported by us [28] . After induction of apoptosis using a known apoptogenic compound Plumbagin [29] , [30] surface and intracellular staining of developing embryos were performed to demonstrate conserved apoptotic features . Plumbagin induces intracellular ROS mediated mitochondrial pathway of apoptosis in different cellular systems [29] , [30] . Since we are addressing intrinsic pathway of apoptosis in developing embryos of the filarial nematode , we chose to use Plumbagin in this study . Annexin-V-PE staining revealed dose dependent externalization of phosphatidyl serine [Figure 1B and Table 1] and Mitoscreen JC-1 staining demonstrated mitochondrial depolarization [Figure 1C and Table 1] in embryonic stages undergoing apoptosis . The diffusely cytoplasmic staining pattern for evolutionarily conserved molecule - Cytochrome-c in apoptotic embryos as opposed to punctate staining in normal embryos [Figure 1D] and overlaid histogram for intra cellular staining of Cytochrome-c [Figure 1E and Table 1] suggested redistribution of Cytochrome-c during apoptosis . Enhanced intracellular expression of nematode apoptosis related proteins CED-3 , CED-4 and CED-9 [Table 1] in embryonic stages of S . digitata were also observed upon stimulation with Plumbagin . Intracellular staining with antibodies to CED-4 revealed a distinct web like staining pattern similar to that of Mitotracker Red which specifically labels mitochondria [Figure 1F] in untreated normal embryos indicating mitochondrial localization of CED-4 . Whereas , diffusely cytoplasmic staining pattern of CED-4 in apoptotic embryos of S . digitata [Figure 1G] suggests redistribution of CED-4 in response to apoptotic stimuli . These observations for CED-4 in apoptotic embryos of pathogenic nematode S . digitata were comparable with observations made with apoptotic embryos of free living nematode C . elegans [31] . Intracellular staining with fluorochrome conjugated antibodies to cleaved PARP demonstrated cleavage of highly conserved intra cellular caspase substrate-PARP [Figure 2A and Table 1] indicating activation of caspase family of cysteine proteases during apoptosis in developing embryos of S . digitata . Further , inhibition of apoptosis by pan caspase inhibitor , Z-VAD – FMK [Figure 2B and C] and quantitative assay for caspase activity using colorimetric caspase substrate Ac-DEVD-pNA [Figure 2D] corroborated the role of caspase activity in embryonic stages during apoptosis . Analysis of nuclear features of apoptosis revealed fragmentation of chromosomal DNA as shown by TUNEL assay [Figure 2E and Table 1] and formation of sub diploid nuclei by PI [propidium iodide]/RNase staining [Figure 2F] in developing embryos of S . digitata . Taken together , these results represent the first ever definitive demonstration of conserved cytoplasmic and nuclear features of apoptosis in a pathogenic nematode . Following demonstration of canonical features characteristic of apoptosis using Plumbagin , we evaluated apoptogenicity of some pharmacological agents viz . , Curcumin , Primaquine , Chloroquine and Diethylcarbamazine [DEC] and three lipid peroxidation products [LPPs] namely TTH [Trans-Trans 2-4 Heptadienal] , TTN [Trans-Trans 2-4 Nonadienal] and TTD [Trans-Trans-2–4 Decadienal] against developing embryos of S . digitata . Curcumin , which activates ROS mediated apoptosis in mammalian systems [32] was found to be the most effective pharmacological agent followed by Primaquine as shown by externalization of phosphatidyl serine , mitochondrial depolarization and fragmentation of chromosomal DNA while Chloroquine displayed minimal effect [Figure 3 A–C] and DEC had no demonstrable effect [Data not shown] . Amongst lipid peroxidation products TTN was found to be very efficient in inducing apoptosis of embryonic stages followed by TTH and TTD as shown by TUNEL assay [Figure 3 D] . Several proapoptotic stimuli including ROS are known to mediate MOMP [Mitochondrial outer membrane permeabilization] and release Cytochrome-c from IMS [Inter membrane space] of mitochondria leading to apoptosis in mammalian systems [33] . Screening of pharmacological agents in the present study revealed that Curcumin and Primaquine induce intra cellular ROS [Figure 4A] in embryonic stages similar to their potential to induce externalization of phosphatidyl serine , mitochondrial depolarization and fragmentation of chromosomal DNA [Figure 3 A-C] i . e . induction of ROS matched with degree of apoptosis mediated by Curcumin and Primaquine . Further , N-Acetyl-L-Cysteine [NAC] , a known scavenger of ROS [21] , [22] not only inhibited ROS generation [Figure 4B , C] but also reversed apoptotic death of developing embryos , as shown by significant decrease in propidium iodide positivity among embryonic stages [Figure 4 D] , reversal of several features of apoptosis [e . g . mitochondrial depolarization [Figure 1C and Figure 4E] , redistribution of Cytochrome-c [Figure 4F] and PARP cleavage [Figure 4G]] and restoration of microfilarial motility [Video S1 , S2 and S3] scored by microscopy [data not shown] . Taken together , these results implicate ROS in mediating apoptosis of developing embryos in this study and suggest possible existence of mitochondrial cell death pathway in parasitic worms . Cytosolic release of Cytochrome-c and its role in caspase activation during apoptosis is yet to be demonstrated in nematodes [34]–[37] . Hence , we examined the status of Cytochrome-c during apoptosis in developing embryos of S . digitata . Observations in the present study have revealed redistribution of Cytochrome-c during caspase dependent apoptotic death in developing embryos [Figure 1D and Table 1] . These findings indirectly suggest a possible role of Cytochrome-c during caspase activation in developing embryos of S . digitata . In mammalian system of apoptosis , caspase activation usually involves cytosolic release of Cytochrome-c and its subsequent interaction with Apaf-1 leading to formation of apoptosome [35] . However , it is not known whether similar interaction between Cytochrome-c and Apaf-1 homologue-CED-4 occurs in nematode system of apoptosis . Hence , demonstration of redistribution of Cytochrome-c and CED-4 in this study [described earlier] was followed by analysis of the sub cellular localization of these two proteins in apoptotic embryos of S . digitata . Results of immuno-localization study revealed significant degree of colocalization between CED-4 and Cytochrome-c in Plumbagin treated apoptotic embryos [Figure 5A] suggesting a possible interaction between these two proteins during activation of apoptosis . This finding was further confirmed by performing fluorescence intensity line profile/colocalization profile analysis for both the labeled proteins in control as well as apoptotic embryos [Figure 5A] . The only nematode for which complete genome data base is available is C . elegans . Hence , to substantiate our experimental finding , molecular docking studies between CED-4 and Cytochrome-c was undertaken , using protein data base of C . elegans . Since the crystal structure for Cytochrome-c of C . elegans was not available , homology modeling for three dimensional structure of Cytochrome-c followed by molecular docking with CED-4 of C . elegans was performed [Figure 5B , Figure S1 and S2] which revealed significant interaction between these two proteins of C . elegans involving 5 hydrogen bonds and 262 hydrophobic interactions [Figure S2] . The Cytochrome-c molecules are known to be highly conserved . The molecular docking of Brugia malayi Cytochrome-c with CED-4 also revealed significant interaction between these two molecules involving 11 hydrogen bonds and 232 hydrophobic interactions [Figure S3] . In mammalian systems , Cytochrome-c has been reported to enhance binding as well as hydrolysis of dATP by Apaf-1 during apoptosome formation [38] , [39] . Hence , the predicted interaction of Cytochrome-c with the evolutionarily conserved α/β [P-loop] – ATP binding domain of CED-4 , that leaves the CARD domain of the later , free for interaction with its target proteins [Figure 5B] suggests that such interactions between Cytochrome-c and CED-4 could be functionally relevant during activation of caspase homologue - CED-3 in nematodes . However , the above observations provide only preliminary evidence for possible role of Cytochrome-c in nematode system of apoptosis and further studies are needed to establish the precise role of Cytochrome-c in the process . The screening experiments with pharmacological agents and LPPs in this study were performed by treatment of developing embryos , released from adult female worms in vitro . To simulate a physiological scenario , intact live adult worms were treated with Plumbagin followed by analysis of their intra uterine embryonic stages for apoptosis ex vivo . Three independent sets of experiments yielded comparable results - the embryonic stages from the adult female parasites treated with Plumbagin revealed significantly higher degree of apoptosis in terms of expression of CED-3 [Figure 6A and D] , cleavage of caspase substrate PARP [Figure 6B and E] and fragmentation of chromosomal DNA [Figure 6C and F] . These observations suggests that the phenomenon of exogenous induction of apoptosis in developing embryos of nematode worms could be biologically relevant and that apoptosis inducing pharmacological agents could be used for blocking embryogenesis in adult female parasites in infected hosts . Similarly , host effector molecules such as lipid peroxidation products present in circulation of an infected host could potentially exert their anti-fecundity effects by induction of apoptosis in developing embryos in vivo . Anti-fecundity immunity in lymphatic filariasis is characterized by presence of adult worms in host in the absence of circulating microfilariae . The mechanism by which human as well as bovine hosts achieves this status in filariasis has not been studied so far [15] . The possibility that enhanced apoptosis of developing embryos during embryogenesis could lead to a state of amicrofilaraemia , in hosts harboring gravid adult female worms was addressed in this study for the first time , in a bovine model of filariasis [20] . Endogenous apoptosis was quantified in developing embryos of S . digitata , harvested from naturally infected bovine hosts [20] with two different parasitological status [e . g . i- cattle harboring both adult worms in peritoneum and microfilariae in circulation and ii- amicrofilaraemic cattle with only adult worms in peritoneum but no circulating microfilariae] by TUNEL staining of embryonic stages ex-vivo . Enhanced apoptosis was found in the developing embryos of adult female parasites , isolated from second category of bovine hosts [Figure 7] suggesting that increased apoptosis during embryogenesis [partly mediated by host factors] , could be responsible for significantly less production and release of viable microfilariae by gravid adult worms into the peripheral circulation of host leading to their amicrofilaraemic status . These findings suggest , induction of apoptosis in developing embryos could be a novel effector mechanism underlying anti-fecundity host immunity operational in helminthiasis .
Apoptosis is a genetically controlled conserved mechanism of cellular demise that plays an important role in a wide variety of physiological processes including embryogenesis , tissue homeostasis and disease progression [40] . During host parasite interaction , parasites tend to induce apoptosis in host cells as a strategy of survival and as a means of establishing infection in their hosts , by creating a site of immune privilege around them [41]–[43] . Similarly , host cells may induce apoptosis of parasites as a defense strategy which has been well documented in case of unicellular parasites only [43] . However , unlike protozoan parasites , metazoan parasites are difficult targets for the cells of host immune system . The disproportionately large size of metazoan helminth parasites vis-a-vis the host immune effector cells and prolonged survival of the former inside respective mammalian hosts has led to a perception that protective immunity may not be effectively operational against these pathogens [15] . Hence , induction of apoptotic death as a therapeutic approach seems to be more relevant in case of these extra cellular parasites . On the other hand , successful embryogenesis in adult females is a critical rate limiting step for survival and propagation of parasitic worms as well as pathology mediated by them [18] . Therefore , blockade of embryogenesis through induction of apoptosis in early embryonic stages offers a viable alternative for developing intervention strategies against them . But , the potential of therapeutic induction of apoptosis as an approach for drug development against metazoan helminth parasites has not been explored , so far . Hence , in the present study we have made an attempt to venture into this unexplored area of research in the biology of parasitic worms . Since , apoptosis in developing embryos of pathogenic nematodes has not been reported in literature earlier , we considered it essential to assess several conserved features of apoptosis including externalization of phosphatidyl serine , mitochondrial depolarization and activation of caspase family of cysteine proteases , fragmentation of chromosomal DNA and formation of hypo-diploid nuclei in embryonic stages of a filarial nematode S . digitata . We further validated the observed programmed cell death by using immuno fluorescence microscopy and scoring expression profile of nematode specific proteins related to apoptosis [e . g . CED-3 , CED-4 and CED-9] in the embryonic stages . Studies in the free living nematode C . elegans had identified three nematode specific proteins such as CED-3 , CED-4 and CED-9 associated with the process of apoptosis . Homolog of these three proteins has been shown to be involved in apoptosis in all most all systems studied so far [44] . The CED-3 protein is a member of the caspase family of cysteine proteases that executes the final cell death process . CED-4 protein is homologus to mammalian Apaf-1 that functions as a positive regulator of CED-3 whereas CED-9 protein is homologus to mammalian Bcl-2 and promotes cell survival in nematodes [45] . Further , redistribution of CED-4 could be demonstrated in apoptotic embryos of S . digitata , similar to earlier observations in apoptotic eggs of C . elegans [31] . Taken together , the above findings clearly suggest that the programmed cell death [PCD] observed in developing embryos of S . digitata in this study is a phenomenon of apoptosis . Even though the size of adult nematodes varies greatly , their eggs are usually of comparable size [46] . In general , size of nematode eggs , which might contain a fully formed first stage larvae ranges from 30 µm to 100 µm in greatest diameter [47] . Since standard protocols are currently available for isolating eggs of parasitic worms [48] and the sample injection port of conventional flow cytometers including BD- FACS calibur used in this study ranges up to 150 µm , the quantitative flow cytometry based assays for apoptosis evaluated in this study are expected to be amenable to embryonic stages of other parasitic worms as well . Our earlier demonstration of the compatibility of first stage larvae/ microfilariae of human filarial parasite Wuchereria bancrofti to flow cytometry analysis [28] offers credence to our hypothesis . Caspase activity is considered essential for general precipitation of typical nuclear features of apoptosis in mammalian cells . However , it is dispensable in several other instances for induction and execution of apoptosis [37] , [49] , [50] . In the present study we have convincingly demonstrated caspase activity in the apoptotic embryos of S . digitata , using multiple approaches [e . g . demonstrating cleavage of PARP - an evolutionarily conserved molecule targeted by active caspases , using assay for caspase activity with colorimetric caspase substrate Ac-DEVD-pNA , showing inhibition of phosphatidyl serine externalization by pan caspase inhibitor Z-VAD-FMK ( N-Benzyloxycarbonyl-Val-Ala-Asp[O-Me] fluromethyl ketone ) as well as studying intracellular expression profile of CED-3 , a homologue of caspase in nematodes] . In typical conserved apoptotic pathways in mammalian cells , the effector caspases invariably lead to activation of a nuclease [- caspase activated DNAse/CAD , by cleavage of its inhibitor , ICAD -inhibitor of caspase activated DNase [50]] that is primarily responsible for fragmentation of chromosomal DNA and formation of sub diploid nuclei during apoptosis [51] , [52] . Similar nuclear features of apoptosis were also observed in this study which corroborates activation of caspase family of cysteine proteases during induction of PCD in developing embryos in this study . Mitochondria are regarded as an integrative organelle in terms of apoptosis as it acts as a meeting point of both caspase dependent and independent path ways of cell death [36] . It is also known as the major source and target of intracellular ROS [53] . Diversion of electrons from mitochondrial respiratory chain is the primary source of intracellular ROS [53] which in its turn acts back upon mitochondria to bring about its depolarization and trigger the release of pro apoptotic factors including Cytochrome-c in to cytosol . The cytosolic Cytochrome-c then interacts with Apaf-1 to form an apoptosome which ultimately lead to activation of effector caspase and apoptosis [51] . Thus reactive oxygen species [ROS] are considered to be essential mediators of apoptosis in various eukaryotic systems [53] . It has been reported that , molecules stimulating formation of ROS trigger apoptosis , a process that is inhibited in the presence of antioxidants [32] . In this context , observations in the current study viz . , increased generation of ROS coupled with depolarization of mitochondria , redistribution of Cytochrome-c and cleavage of conserved intracellular caspase substrate PARP during induction of apoptotic death in developing embryos of filarial nematode S . digitata and reversal of all these features in presence of a known scavenger of ROS – NAC [21] , [22] suggest a role for ROS in mediating apoptosis in this study and indicate possible existence of mitochondrial pathway of apoptosis in pathogenic nematodes . These findings also provide possibilities to design new strategies to kill nematode parasites preferentially through induction of ROS mediated programmed cell death . Metazoan parasites elicit a unique mechanism of host immunity in their respective hosts commonly referred to as anti-fecundity immunity . This distinctive aspect of host immunity reduces output of viable embryos [e . g . eggs or microfilariae] by fecund female parasites into the peripheral circulation or tissues of infected hosts . However , the precise host mechanism underlying these anti-fecundity effects is yet to be established in literature . Taking into consideration the very high fecundity in parasitic worms , conventional methods to demonstrate anti-fecundity effects [e . g . measuring the circulating antigen; microscopic counting of the microfilariae in circulation or eggs in tissue homogenates , feces or urine of host and in uterine cavities of female worms by light microscopy etc . [54]-[56] do not offer reliable quantitative information on anti-fecundity effects of drugs or vaccines . In this context , the multiple quantitative flow cytometry based assays developed and evaluated in the present study for demonstration of apoptotic death in developing embryos of a filarial nematode , for the first time , represent a quantum improvement in our approach to understand the anti-fecundity effects mediated by drugs , vaccines and host immune cells or molecules in helminthic infections and are expected to find wide application in nematode biology . Further , induction of apoptotic death of developing embryos by LPPs observed in the present study; appears to be a possible effector mechanism of host against parasitic worms since , raised levels of plasma LPPs is an integral aspect of several parasitic diseases including helminth infections [19] . Usually cells far in excess are generated in metazoan organisms during early embryogenesis and many of them undergo apoptosis for embryogenic sculpturing of different tissues and organs in the late embryo [57] . In the present study the basal level of apoptosis was consistently found to be higher in pre larval embryonic stages as compared to microfilariae/larval stage-1 [L-1] [Table 1] . This quantitative difference in apoptosis among different developmental stages can be attributed to ongoing in-situ apoptosis during normal embryogenesis . Unlike pre larval embryonic stages which represent early stages of development in the filarial nematode S . digitata , microfilariae represent final stage of intra uterine development and hence are relatively free of ongoing in-situ apoptosis . Infected humans in endemic areas can be classified into three groups based on presence of Circulating Filarial Antigen/CFA [products of adult worms] and microfilariae in circulation , namely endemic controls [neither CFA nor microfilariae in circulation] , cryptic/amicrofilaraemic cases [with CFA but without microfilariae in circulation] and microfilariae carriers [with microfilariae in circulation] [15] . Several hypotheses have been proposed to explain the unusual cryptic/amicrofilaraemic status of human subjects with active filarial infection . Infestation with reproductively immature or unisexual worms , different anatomical location of male and female worms or anti microfilarial immunity [yet to be established in literature clearly] are some of the existing propositions in this regard [15] . The current study has used a bovine equivalent of this infection status observed in human filariasis [20] . Findings of the present study demonstrated enhanced in-situ apoptosis of developing embryos in adult female worms harvested from infected but amicrofilaraemic animals [cryptic equivalent of human filariasis] compared to that of microfilaraemic animals [Figure 7] in a bovine model of filariasis . This observation provides preliminary evidence to suggest that anti-fecundity host immunity involving in-situ induction of apoptotic death of developing embryos [which can potentially reduce the output of viable microfilariae] can be another plausible explanation for the peculiar parasitological status observed in human lymphatic filariasis . However , further studies are needed to precisely establish the role of anti-fecundity host immunity in determining the clinical status of hosts in lymphatic filariasis . In conclusion , findings of present study , constitute the first ever report on development and evaluation of flow cytometry based assays leading to clear demonstration of a common but hitherto unexplored phenomena i . e . apoptosis in developing embryos of a pathogenic nematode S . digitata . The observations in this study reveals that embryonic stages of metazoan filarial parasites are prone to caspase dependent apoptotic death , primarily mediated through induction of intra cellular ROS . Since apoptosis is a conserved biological process of cellular demise among metazoans , it's induction is expected to involve a closely similar mechanism , at least in a single group of animals i . e . among parasitic worms . Thus , compounds identified to have apoptogenicity towards one helminthic parasite/it's larval stages might prove effective against other helminthic pathogens , as well . In such an eventuality new drugs with broad spectrum anti helminthic activity with an established and common mode of action i . e . induction of apoptosis in embryonic stages will be a reality . The quantitative flow cytometry based assays for apoptosis evaluated in this study at a translational level , offers opportunities for developing automated high throughput screening assays for identification of anti-fecundity drugs and determining the efficacy of anti-fecundity vaccines to combat infections caused by helminth parasites . By providing a scope to understand the mechanism of an important phenomenon i . e . apoptosis in developing embryos of a pathogenic nematode , for the first time , the present study also offers leads to address other relevant issues of biological importance e . g . different forms of PCD , anti-fecundity host immunity , cell signaling and sex determination etc . during embryogenesis of parasitic worms . In addition , the present study can potentially further our understanding of genes that are critically important for embryo development and reproduction in parasitic worms , recently proposed to offer promise for developing alternative avenues of drug development against these metazoan parasites [8] . | Pathogenic nematodes currently infect billions of people around the world and pose serious challenges to the economic welfare and public health in most developing countries . At present , limitations of existing therapies warrant identification of new anti-parasitic drugs/drug targets to effectively treat and control neglected tropical diseases [NTD] caused by nematode pathogens . The current gold standard for measuring/screening drug effectiveness against most helminth parasites is in-vitro assessment of motility of parasites/larvae and larval development assays which fails to provide any conclusive idea about the precise mechanism of death of parasitic worms or their larval stages . Given the huge load of parasites or their larval stages in an infected host , a compound which shows promise in in-vitro/motility screening assays but induces necrotic death in parasites/larvae will be of limited use , as it may elicit severe inflammatory response in infected hosts . In this context , the present study , which demonstrates induction of apoptotic death in developing embryos of a pathogenic nematode as a potential drug target for the first time , and provides scope for high throughput screening of pharmacological agents for their apoptogenicity against nematode embryos , is a step forward to develop novel anti-parasitic measures to challenge NTD caused by nematode pathogens . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"biology",
"veterinary",
"science"
] | 2011 | Caspase Dependent Programmed Cell Death in Developing Embryos: A Potential Target for Therapeutic Intervention against Pathogenic Nematodes |
The HolC-HolD ( χψ ) complex is part of the DNA polymerase III holoenzyme ( Pol III HE ) clamp-loader . Several lines of evidence indicate that both leading- and lagging-strand synthesis are affected in the absence of this complex . The Escherichia coli ΔholD mutant grows poorly and suppressor mutations that restore growth appear spontaneously . Here we show that duplication of the ssb gene , encoding the single-stranded DNA binding protein ( SSB ) , restores ΔholD mutant growth at all temperatures on both minimal and rich medium . RecFOR-dependent SOS induction , previously shown to occur in the ΔholD mutant , is unaffected by ssb gene duplication , suggesting that lagging-strand synthesis remains perturbed . The C-terminal SSB disordered tail , which interacts with several E . coli repair , recombination and replication proteins , must be intact in both copies of the gene in order to restore normal growth . This suggests that SSB-mediated ΔholD suppression involves interaction with one or more partner proteins . ssb gene duplication also suppresses ΔholC single mutant and ΔholC ΔholD double mutant growth defects , indicating that it bypasses the need for the entire χψ complex . We propose that doubling the amount of SSB stabilizes HolCD-less Pol III HE DNA binding through interactions between SSB and a replisome component , possibly DnaE . Given that SSB binds DNA in vitro via different binding modes depending on experimental conditions , including SSB protein concentration and SSB interactions with partner proteins , our results support the idea that controlling the balance between SSB binding modes is critical for DNA Pol III HE stability in vivo , with important implications for DNA replication and genome stability .
Chromosome replication is performed by the replisome , a molecular machine present in all living organisms with strong structural and functional similarities [1]–[4] . Replisomes combine the action of a primosome and a polymerase , for which the enzymes from Escherichia coli have proved an invaluable model for understanding their function . The E . coli primosome is itself composed of two interacting enzymes , the hexameric DnaB helicase that opens double-stranded DNA and the DnaG primase that synthesizes leading- and lagging-strand primers . DNA is synthesized by the DNA polymerase III holoenzyme ( Pol III HE ) , composed of polypeptides encoded by 9 different genes . The holoenzyme is composed of three core polymerases [5] , each made up of a polymerase α subunit ( encoded by dnaE ) , a proofreading ε subunit ( dnaQ ) , and a θ stability factor ( holE ) . DNA binding by leading- and lagging- strand core polymerases is stabilized through interactions with the β-clamp ( dnaN ) . Lagging-strand synthesis is discontinuous and Okazaki fragments ( OF ) are joined by the ligase . The role of the third core polymerase is still under investigation; current models suggest that it replaces the lagging-strand polymerase when needed [6] . β-clamps are loaded onto DNA for replication initiation and for the synthesis of each OF by a complex called the clamp loader . The minimal clamp loader core is a pentameric complex containing a δ ( holA ) , a δ′ ( holB ) and three τ ( dnaX ) protein subunits . A χψ complex ( holC , holD ) connects this pentameric complex to DNA , as ψ ( HolD ) interacts with τ [7] , [8] and χ ( HolC ) interacts with the single-stranded DNA binding proteins ( SSB ) that cover the lagging-strand template [9]–[12] . The three-dimensional structure of the χψ complex has been determined , identifying the sites of interaction between ψ and τ , and between χ and SSB [12] , [13] . In vitro each clamp loader complex contains a single χψ complex , but four may be associated with the replisome in vivo [4] , [5] , [8] . How the three additional in vivo χψ complexes are organized is currently unknown . The clamp loader complex ensures replisome cohesion through interactions between τ and Pol III , τ and the DnaB helicase , and χ and SSB [4] . Clamps and clamp-loaders are universally conserved in structure and function , for example PCNA and RFC , respectively , in eukaryotes [2] . In contrast , ψ and χ have only been found in proteobacteria [1] , [14] , [15] and no homologous proteins have been reported in eukaryotes . Using E . coli mutants to analyze the role of the χψ complex provides an alternative approach toward understanding chromosome replication and the molecular mechanisms that underlie clamp-loader function . In vitro comparison of clamp loading and replication in the presence and absence of χψ have led to the identification of three main putative functions . Firstly , ψ-τ interactions stabilize the clamp loader complex , allowing it to form even at the low protein concentrations found in vivo [16] . The presence of ψ alone increases clamp-loader ATPase activity , and its affinity for DNA and the β-clamp [8] , [17] . Secondly , in the presence of SSB , owing to χ-SSB interactions , χψ increases the affinity of the clamp-loader complex to primer-template DNA , stimulates clamp loading activity and increases Pol III processivity [9] , [18] . Thirdly , χψ also promotes the in vitro displacement of primase from RNA primers , by switching from primase-SSB to χ-SSB interactions at the primer-template junction , and thus participating in primase recycling at replication forks [19] . The above properties indicate that χψ mainly acts on the lagging-strand template , which is SSB-coated and subject to clamp loading every 1–2 seconds . Accordingly , a holC deletion confers a hyper-recombination phenotype that can be explained by defective lagging-strand synthesis [20] . Moreover , a holD mutant was isolated in a screen for hyper-recombination mutants [21] . On the other hand , several lines of in vivo and in vitro evidence suggest that decreasing the cellular level of χ or ψ proteins affects both lagging- and leading-strand synthesis . A holD point mutation was shown to trigger replication fork reversal , which is caused by replication fork arrest and the subsequent annealing of leading- and lagging-strand ends to form a Holliday junction adjacent to a double-stranded DNA end [21] , [22] . In addition , a mutation affecting holC suppressed the growth defects caused by replication over-initiation , possibly because it slows down replication fork progression [23] . Finally , in vitro studies using χ variants with impaired SSB interaction capacity revealed defects in leading-strand synthesis and resulted in the production of shorter OFs , in agreement with the idea that holC and/or holD impairment affects the synthesis of both strands during replication [12] . Deletion of the holD gene strongly affects E . coli growth at 30°C and is lethal at higher temperatures . These defects could be partially suppressed by blocking the SOS response [24] . The E . coli SOS response is triggered by the accumulation of RecA-coated single-stranded DNA ( ssDNA ) followed by proteolysis of the LexA repressor , inducing the expression of more than 40 LexA-repressed DNA repair genes [25] , [26] . We have previously shown that the SOS-induced DinB and Pol II bypass polymerases are responsible for the deleterious effects of the SOS response on ΔholD mutant growth [24] . We proposed that a combination of replisome destabilization in the absence of ψ and displacement of the destabilized replisomes by these two SOS-induced polymerases was , at least in part , responsible for the poor growth of the ΔholD mutant [24] . In this work , we characterize a spontaneous suppressor mutation of the ΔholD growth defects and identify it as an ssb gene duplication . We find that this duplication also suppresses the growth defects of a ΔholC mutant but does not prevent ΔholD-induced expression of SOS response genes . We propose that ssb gene duplication directly compensates for the absence of the χψ complex by stabilizing the association of Pol III HE to DNA .
Because the ΔholD mutant is slow growing and accumulates suppressor mutations , we propagated it in the presence of a complementing plasmid that replicates from a conditional origin . The pAM-holD plasmid carries the wild-type holD gene and only replicates in the presence of the lac inducer isopropyl β-D thio-galactoside ( IPTG ) [24] . This plasmid was then cured prior to each experiment to restore the ΔholD mutant condition . Depending on the experiment , we either generated mixed cultures containing at least 99% cured cells by growing ΔholD [pAM-holD] cells in the absence of IPTG for about 15 generations , or isolated plasmid-less colonies by plating ΔholD [pAM-holD] cultures propagated in the absence of IPTG for 8 hours on IPTG-free medium . Cultures or colonies were confirmed to be free of both complementing plasmids and suppressor mutations . All strains used in this work carry a sfiA mutation ( sfiA::MudAplacZ or sfiA11 ) to prevent SfiA-mediated cell division blockage upon SOS induction . When grown at 30°C on minimal medium supplemented with casamino acids ( hereafter MM ) , ΔholD cells formed rare small colonies together with rapidly-growing colonies that we suspected to have acquired a suppressor mutation ( Figure 1 , Figure S1 , Figure S3 , [24] ) . Putative ΔholD suppressing clones were able to form normal sized colonies over two days when grown at 30°C on MM . We sequenced the genome of one such clone , designated JJC2394 , and compared its sequence to that of the parental ΔholD mutant . Two mutations were found in JJC2394: a leuO point mutation ( A233V ) and a 10 kb duplication ( Figure 2 ) . Reversing the leuO mutation to Leu+ by P1 transduction did not affect JJC2394 viability . In contrast , removing the duplication abolished most of the suppression effect , indicating that it is important for the suppressor phenotype ( Figure 1 ) . The duplication is flanked by 6 base pairs ( bp ) microhomology DNA sequences and lies between positions 4 266 351 and 4 276 297 on the E . coli MG1655 chromosome used as the reference strain ( Figure 2 ) . The duplicated region contains 10 genes: five of unknown function , aphA encoding a periplasmic phosphatase , uvrA encoding a nucleotide excision repair protein , soxS and soxR encoding the superoxide response activator , and ssb . Since ssb is the only one of the duplicated genes directly involved in DNA replication , we hypothesized that the presence of two copies of ssb might be responsible for the suppressor phenotype . To test this idea , we constructed a strain carrying an additional copy of ssb inserted into the argE locus , which is located approximately 120 kb from ssb . Strains harboring the argE::ssb insertion carry the same number of ssb genes as the ssb tandem duplication throughout the cell cycle , and the two ssb copies are stably maintained due to the distance separating them . Western analysis using anti-SSB antibodies showed that JJC2394 cells and cells containing the argE::ssb insertion expressed two to three times as much SSB protein as wild-type cells ( Figure S2 ) . Although over-production of SSB from a plasmid affects the growth of wild-type E . coli and induces the SOS response [27] , ssb gene duplication appears to have no deleterious effect since argE::ssb did not affect the growth of wild-type cells and did not induce the SOS response ( Figure 3 , Table 1 ) . Next , we analyzed the growth properties of ΔholD argE::ssb cells . The presence of two ssb gene copies restored wild-type levels of colony formation to ΔholD mutants when grown on MM and LB , at 30°C , 37°C and 42°C ( Figure 3 , Figure S3 ) . In fact , although the lexAind mutation allowed ΔholD cells to be propagated in liquid cultures [24] , ΔholD lexAind colonies were slower growing and reduced in number compared to ΔholD argE::ssb ( Figure 3 , Figure S3 ) . As preventing SOS induction improves ΔholD mutant growth [24] , we tested whether the ssb duplication acts by decreasing SOS-induction . SOS expression was measured using lacZ under the control of the SOS-induced promoter sfiA in ΔholD mutants and in suppressed strains . β-galactosidase expression was measured in cells propagated either at 30°C or after shifting to 42°C for 3 hours ( Table 1 ) . As reported previously , the ΔholD mutation induces the SOS response at both temperatures and this induction is prevented by recF inactivation [24] . In the JJC2394 spontaneous-suppression mutant , the SOS response was decreased compared to the ΔholD mutant at 30°C and 42°C , but was still significantly higher than in wild-type cells . The ΔholD argE::ssb mutant exhibited a similar SOS response to that of the ΔholD single mutant , showing that the suppression phenotype conferred by ssb duplication is not a consequence of SOS inactivation . SOS expression in ΔholD argE::ssb cells and in JJC2394 was largely RecF-dependent ( Table 1 ) , in the same way as in ΔholD cells . RecF , RecO and RecR proteins specifically promote RecA loading onto ssDNA gaps . Thus , RecF dependence suggests that the SOS response in ΔholD cells is induced by the accumulation of ssDNA gaps , possibly formed during lagging-strand synthesis . Inactivating the SOS response with a lexAind mutation did not further improve the capacity of ΔholD argE::ssb cells to form colonies at different temperatures ( Figure 3 , Figure S3 ) . This is consistent with the fact that SOS induction is not required for ΔholD argE::ssb cell viability and that the viability of this double mutant is equivalent to wild-type bacteria . For historical reasons , this work was realized in an AB1157 background . In the more commonly used MG1655 strain background , suppression of ΔholD growth defects by ssb gene duplication was observed at 30°C and 37°C but not at 42°C ( Figure 3 ) . In this case , introducing the lexAind mutation slightly improved ΔholD growth . Interestingly lexAind and argE::ssb suppressor mutations had additive effects on MG1655 ΔholD viability , with the MG1655 ΔholD lexAind argE::ssb mutant exhibiting similar plating efficiency to that of wild-type MG1655 at 30°C and 37°C . Nevertheless , in contrast to the AB1157 background , colonies were smaller than wild-type at 37°C and this strain was unable to propagate at 42°C ( Figure 3 ) . The higher SOS response levels in ΔholD argE::ssb cells compared to the ΔholD sup cells of JJC2394 suggested the presence of an additional mutation in the latter strain ( Table 1 ) . However , the mutation ( s ) responsible for this difference could not be identified from the chromosome sequence . lexA and recombination genes were intact and , accordingly , survival to UV irradiation was unaffected in JJC2394 ( Figure S4 ) . To test the effect of the 10 kb tandem duplication , the argE::ssb allele was introduced into JJC2394 resulting in the spontaneous loss of the duplication , presumably by homologous recombination . The resulting strain carried the argE::ssb allele instead of the spontaneous 10 kb duplication but was otherwise identical to JJC2394 . Significantly , the new strain showed the same SOS response levels as the ΔholD sup strain JJC2394 rather than the reconstructed ΔholD argE::ssb strain ( Table 1 , JJC6216 ) . Therefore , the lower expression of the SOS response in JJC2394 compared to ΔholD argE::ssb is not caused by another gene within the 10 kb duplication and remained unexplained . We previously showed that inactivating either of the dinB and polB SOS-induced polymerase genes improves ΔholD mutant viability at 37°C , while inactivating both restored growth at 42°C [24] . These results suggested that DinB and PolB polymerases participate in the destabilization of HolD-less Pol III HE upon SOS response induction . One possibility is that the restoration of ΔholD mutant growth by ssb duplication is linked to the destabilizing effect of SOS-induced DinB on Pol III HE . DinB levels increase about 8- to 10-fold upon SOS induction [25] , [28] , [29] , and increase to similar levels when expressed from a pSC101-derived vector [29] . Thus , we used the pSC101-derived vector pGB2 to compare the effects of increased expression of the wild-type dinB gene ( pGB-dinB ) . When tested in a lexAind background to prevent SOS induction , the transformation efficiency of [pGB-dinB] was similar for wild-type and ΔholD argE::ssb mutant bacteria in all conditions , confirming that 8-fold over-production of DinB is not deleterious to the ΔholD argE::ssb mutant ( JJC6133 Figure 4 , Table S2 and Figure S5 ) . Moreover , pGB-dinB was not deleterious for growth in HolD+ LexA+ or LexAdef backgrounds , confirming previous results showing that DinB expressed from a pSC101 replicon is not deleterious for growth , even in the absence of the LexA repressor [29] . In these conditions , DinB is expressed at 8- and 30-times the wild-type chromosomal level , respectively [29] , and replication in wild-type cells is only sensitive to the higher levels of DinB over-expression [30] , [31] . However , pGB-dinB could not be introduced into ΔholD argE::ssb or JJC2394 cells on MM ( Figure 4 , Table S2 ) ; on LB , ΔholD argE::ssb [pGB-dinB] clones were obtained at 37°C and 42°C but could not be propagated ( Figure S5 ) . These results suggest that ssb gene duplication does not stabilize ψ-less Pol III HE enough to compensate for the effect of pGB plasmid-mediated DinB overexpression combined with SOS response activation . To explore the mechanism behind this effect , we performed the same assay using a dinB gene lacking 5 C-terminal amino acids required for interaction with the β-clamp ( pGB-dinBΔC5 [32] ) . The absence of detrimental effects following expression of this deletion mutant ( pGB-dinBΔC5 , Figure 4 , Table S2 and Figure S5 ) demonstrates that a functional interaction between DinB and DnaN is required for the effects of 30-fold DinB overproduction . Moreover , it shows that this interaction induces the substitution of β-clamp DnaN-bound Pol III by DinB . Altogether , these results indicate that doubling SSB concentration in vivo protects the ΔholD mutant against the deleterious effects of a 8- to 10-fold increase in DinB , regardless of whether this increase is caused by SOS induction , or increased dinB gene expression from a ∼10 copy-number plasmid in the absence of SOS induction . These results suggest that doubling the amount of SSB stabilizes Pol III HE DNA binding in the absence of HolD and consequently improves resistance to physiological increases in DinB levels , such as those produced by SOS induction . However , the ΔholD argE::ssb mutant remains sensitive to ∼30-fold increases in DinB production , showing that even in the presence of twice the normal amount of SSB , the HolD-less Pol III HE complex is more sensitive than the wild-type holoenzyme to non-physiological DinB amounts . SSB interacts with a large number of DNA replication , recombination and repair proteins via its C-terminus in both E . coli and Bacillus subtilis ( reviewed in [33] , [34] ) . In order to test whether these interactions play a role in the suppression of ΔholD defects by ssb duplication , we constructed a strain in which the additional copy of ssb inserted into argE contains a five amino acid C-terminal deletion ( ssb-ΔC5; see Materials and Methods ) . The argE::ssb-ΔC5 allele did not affect wild-type growth and did not induce the SOS response ( Figure 5 , Table 1 ) , showing that expression of this SSB truncated protein does not affect the function of the wild-type protein . Growth of the ΔholD argE::ssb-ΔC5 mutant was tested on MM and on LB at different temperatures . Compared to the ΔholD single mutant , ΔholD argE::ssb-ΔC5 was only slightly more viable on MM at 30°C and rapidly acquired suppressor mutations ( Figure 5 , Figure S3 ) . We conclude from these experiments that ΔholD mutant growth defects can only be suppressed by an additional copy of ssb carrying an intact C-terminus . This result suggests that interaction ( s ) with SSB partner ( s ) are crucial for the rescue of ΔholD mutant by increased ssb gene dosage . The χψ complex ( HolC-HolD ) bridges the minimal clamp loader complex to SSB . We hypothesized that χ ( HolC ) might be the SSB interacting protein required for ΔholD growth defect suppression . If doubling the amount of SSB allows χ to act without ψ , introduction of ΔholC should abolish the suppression . Alternatively , if ssb gene duplication bypasses the need for the entire χψ complex , it should also suppress the growth defects of ΔholC and ΔholC ΔholD mutants . We tested these ideas using the ΔholC102::CmR deletion mutant [20] . Growth of the ΔholC mutant was strongly affected at 42°C and only slightly affected at 30°C and 37°C , both on MM and LB ( R . Maurer personal communication , Figure 6 and Figure S6 ) . It is worth noting that ΔholC is less deleterious for growth at 30°C and 37°C than ΔholD , suggesting a role for ψ-τ ( HolD-DnaX ) interaction in Pol III HE stability . We constructed pAM-holC and pAM-holCD plasmids that carry wild-type copies of holC or both holC and holD genes respectively , which were cured at the onset of each experiment ( see Materials and Methods ) . We analyzed ΔholC single , ΔholC argE::ssb double and ΔholC ΔholD argE::ssb triple mutants ( Figure 6 and Figure S6 ) and obtained similar results regardless of whether the strains were originally constructed in the presence of pAM-holC or pAM-holCD . The ssb gene duplication conferred viability to both ΔholC single and ΔholC ΔholD double mutants at all temperatures , although at 42°C ΔholC argE::ssb and ΔholC ΔholD argE::ssb colonies were slightly smaller than wild-type . Thus , doubling the amount of SSB suppresses growth defects caused by the absence of the entire χψ complex , regardless of whether χψ function is affected by the inactivation of holC , holD or both genes . We conclude that ssb duplication suppresses the growth defects caused by a HolCD-less Pol III holoenzyme via SSB interactions with a replisome protein other than χ .
In vivo , the ΔholD mutant accumulates gaps , as deduced from RecF-dependent constitutive SOS expression , and suffers from replication arrest and polymerase loss , as deduced from the occurrence of replication fork reversal and from its sensitivity to SOS-induced polymerases [21] , [24] . Accordingly , purified χ proteins containing a mutation that specifically affects SSB interaction were clearly impaired for both leading- and lagging-strand synthesis and for replisome stability [12] . The well-documented importance of χψ for lagging-strand synthesis in vitro [16] , [17] , [19] suggests that the gaps that induce the SOS-response in vivo are formed on the lagging strand . We observe that ssb gene duplication restores the viability of ΔholD mutant cells but does not prevent RecF-dependent SOS induction , thus does not suppress gap formation . Therefore , excessive gap formation is not directly responsible for the poor viability of the ΔholD mutant . Furthermore , ssb duplication restores normal ΔholD growth in the presence of an 8 to 10-fold excess of DinB , expressed either from the SOS-induced chromosomal copy or from a low copy plasmid in the absence of SOS induction . Consequently , we propose that an intrinsic lack of stability of HolD-less Pol III HE bound to DNA is responsible for the growth defects of the ΔholD mutant , and that ssb gene duplication acts by stabilizing the HolD-less Pol III holoenzyme . It is worth noting that competition by SOS-induced polymerases is not the only reason for HolD-less Pol III HE instability , as lexAind mutation does not suppress ΔholD growth defects as efficiently as ssb gene duplication . In the MG1655 background , suppression of ΔholD growth defects by either lexAind or ssb duplication is partial , showing that the effects of these two suppressor mutations are additive . We propose that a combination of decreased expression of SOS-induced polymerases ( lexAind ) and increased Pol III HE DNA stability ( ssb duplication ) is necessary to restore viability in this background . In AB1157 , where ssb duplication suppresses ΔholD growth defects quite efficiently , the additive effects of lexAind and ssb duplication are not directly detectable . The thermosensitivity of the ΔholD argE::ssb lexAind MG1655 mutant , also observed for the ΔholC AB1157 single mutant , is interesting since no protein is intrinsically sensitive to high temperature in these mutants . It cannot be accounted for solely by a higher number of replication forks per chromosome at 42°C compared to 30°C , since the number of replication forks per chromosome is also increased in rich medium and these mutants show no rich medium sensitivity . High temperature affects protein-protein and protein-DNA interactions and the sensitivity of these mutants to high temperature supports the idea that the primary defect of the HolCD-less Pol III holoenzyme is its intrinsic instability on DNA . In agreement with a direct role for HolD in clamp loader complex stability in vitro [8] , [17] , growth is clearly more affected at both 30°C and 37°C in ΔholD mutant cells than in cells lacking holC . In vitro SSB binds ssDNA in multiple binding modes , among which the two major forms are ( SSB ) 35 and ( SSB ) 65 , where 35 and 65 nucleotides , respectively , are wrapped around a SSB tetramer [35] . SSB proteins are also mobile on ssDNA , undergoing random diffusion along ssDNA mainly in the ( SSB ) 65 binding mode ( Zhou et al . 2011 ) . The ( SSB ) 35 binding mode is less mobile , more stable , and highly cooperative , forming protein clusters or filaments on DNA [35]–[37] . The binding mode is determined by salt concentration and by the SSB protein to ssDNA ratio . Increasing SSB concentration in vitro shifts the binding mode toward the ( SSB ) 35 form [36] . It has been proposed that the binding mode could also be influenced by protein interactors , and actually interaction between PriC and the C-terminal tail of SSB can also shift the ssDNA binding mode from ( SSB ) 65 to ( SSB ) 35 [38] . The primary effect of ssb gene duplication and the resulting increase in SSB protein concentration could involve a shift from the ( SSB ) 65 to the ( SSB ) 35 binding mode on the lagging-strand template at in vivo salt concentration . This phenomenon could compensate for the absence of the χψ complex if χ-SSB interaction is normally responsible for the shift , as has been hypothesized [36] . Nevertheless , it should be noted that to date the existence of different SSB binding modes , and their dependence on SSB concentration and on SSB-protein interactions have only been demonstrated in vitro and remain to be tested in vivo . A C-terminal SSB truncation promotes ssDNA binding and shifts the equilibrium toward the ( SSB ) 35 mode in vitro [10] , [36] , [39] , [40] . Therefore , it is unlikely that this deletion prevents a putative shift from ( SSB ) 65 to ( SSB ) 35 in cells expressing both wild-type and truncated SSB . Thus , the requirement for two intact ssb genes to suppress ΔholD mutant growth defects may instead reflect a need for SSB interaction ( s ) with one or more protein partner ( s ) [33] . Three replisome proteins have been reported to interact with SSB: χ , primase and the α polymerase ( DnaE ) [33] , [41] . χ is not required for suppression since ΔholC and ΔholC ΔholD mutants are also fully suppressed by the ssb gene duplication . The SSB and the primase appear to interact via SSB C-terminal amino acids and a specific region of the primase [42] . Primase could be a key SSB interacting protein for stabilization of the HolD-less Pol III holoenzyme , although its requirement for the OF synthesis and the high level of SOS induction in ΔholD argE::ssb cells suggests that gap formation during OF synthesis is not suppressed by ssb gene duplication . The SSB-DnaE interaction was detected in a Tap-Tag high-throughput analysis of E . coli proteins using DnaE as bait and SSB as prey [41] . Even though the protein regions involved in SSB and DnaE interaction have not yet been identified , this interaction could also be crucial for the growth of HolD-less Pol III containing cells . Interaction between the SSB C-terminus and a Pol III holoenzyme component other than χ has been shown to stimulate initiation complex formation [43] . In cells lacking χψ , DnaE-SSB interaction could be needed for OF initiation , and throughout lagging-strand synthesis for a stabilizing effect of the ( SSB ) 35 binding mode on the Pol III holoenzyme . In vitro experiments would be required to test these various hypotheses . It is remarkable that simply doubling the amount of SSB has such a large effect on viability , even though the ssb gene is not known to be regulated and strong SSB over-production is deleterious [27] . The striking effects of increased SSB expression on viability suggest that the in vivo SSB-DNA complex equilibrium is finely balanced between binding modes and can be switched by different factors , including SSB concentration and SSB interactors . It is noteworthy that ψ is present only in a few bacterial species [14] , and although χ is more widely distributed , it is not universal [1] , [14] , [15] . Since stabilization of Pol III can apparently be achieved by SSB interaction with a Pol III holoenzyme component other than HolC provided that the amount of SSB is doubled , bacteria which lack χψ may tune the stability of their Pol III holoenzyme via one of the minimal clamp-loader subunits , for example through SSB interactions that do not exist in E . coli , by a stronger interaction with DnaE , or by naturally expressing higher amounts of SSB than in E . coli , together with lower levels of competing SOS-induced polymerases .
Strains , plasmids and oligonucleotides used in this work are described in Table S1 . New mutations were constructed by recombineering as described in [44] and using DY330 [45] . All other strains were constructed by P1 transduction . pAM-holD plasmid was cured prior to each experiment by growing cells in the absence of IPTG and plasmid-less colonies were isolated on minimal medium glucose casaminoacids ( MM ) plates . We checked that less than 5% of cells in the culture contained pAM-holD and less than 1% had acquired a suppressor mutation . All mutations introduced by P1 transduction were checked by PCR and all new mutations constructed by recombineering were checked by PCR and sequencing . lexAind and recF mutations were tested by measuring UV sensitivity . For argE::ssb construction , DY330 was transformed by electroporation with a ssb-KanR PCR fragment flanked by 50 bp of homology with argE . For argE::ssbΔC5 , DY330- argE::ssb ( JJC5953 ) was transformed by electroporation with a PCR CmR fragment flanked by 50 bp homology with SSB C-ter sequence lacking the five last residues and 50 bp of homology with argE . In this construction the 5 last SSB residues are replaced by two stop codons and the KanR gene in argE::ssb is replaced by CmR . For pAM-holC construction , holC was PCR amplified from the chromosome and cloned into the pAM34 vector after digestion with PstI and HindIII; the resulting plasmid was verified by sequencing and complementation of the holC mutant . For pAM-holCD construction , holC was cloned from pAHM101 [12] in pAM-holD using Ssp1/BsaB1 and Xba1; the resulting plasmid was verified by PCR and complementation of the holC mutant . For spot assays , colonies formed in three days on MM at 30°C were suspended in MM salt medium . Serial 10-fold dilutions were then performed and 7 µl of dilutions 10−2 to 10−6 were spotted on three MM and three LB plates that were placed at 30°C , 37°C and 42°C . For pGB-dinB transformants , 10−2 to 10−5 dilutions of colonies obtained overnight on LB were used . In all cases , plates were scanned after 16–24 hours of incubation at 37°C and 42°C , and after 2 days of incubation at 30°C . Spot assays were performed at least twice for each strain . In addition , for viable strains colony forming units ( cfu ) were determined by plating appropriate dilutions of overnight MM cultures on MM and LB plates . The number of colony was counted after 16–24 hours of incubation for plates at 37°C and 42°C and after 48 hours of incubation for plates at 30°C . Each strain was tested at least three times and results confirmed the full viability observed in spot assays . Non-viable mutants could not be grown overnight and therefore the lack of viability was also checked by streaking several isolated plasmid-less colonies on LB and MM at 30°C , 37°C and 42°C . SSB and FtsZ proteins were detected in cell extracts using polyclonal chicken antibodies against SSB ( gift from MM Cox , University of Wisconsin-Madison ) and polyclonal rabbit antibodies against FtsZ ( gift from J Camberg , National Institutes of Health , Bethesda , Maryland ) . Cell extracts were prepared from a fixed amount of exponentially growing cells . The cells were resuspended in 100 µl of Laemmli Buffer ( Bio-Rad #161-0737 ) and incubated for 10 min at 100°C . Total cellular proteins were fractionated by SDS-PAGE on 12 . 5% gels and transferred to a Hybond Nitrocellulose membrane ( Amersham ) by electroblotting using a semidry transfer system . Immunodetection was carried out as described in the ECL+ kit ( Amersham ) . Western blots were revealed using LAS-3000 FujiFilm and quantified with ImageQuant . | Both replication polymerases and single-stranded DNA binding proteins ( SSB , which associate with single-stranded DNA exposed transiently during replication ) are ubiquitous and show high levels of functional and structural conservation across all species . Among the nine different polypeptides that compose the bacterial replicative polymerase , the HolC-HolD ( χψ ) complex interacts with SSB , and is crucial for normal growth in the model bacteria Escherichia coli . Interestingly , many bacterial species lack this complex , where its function is presumably carried out by other polymerase components . With the aim of better understanding HolC-HolD ( χψ ) complex function in E . coli , we isolated growth defect suppressor mutations of the holD mutant . We found that ssb gene duplication and the consequent doubling of SSB protein expression , renders the entire χψ complex dispensable for growth . We also show that growth-defect suppression requires the presence of the SSB C-terminal amino acids in both ssb gene copies . This C-terminal tail promotes interaction between SSB and its partner proteins . Thus , our results indicate that in vivo SSB concentration plays a key role in maintaining polymerase stability and replication efficiency , in a reaction that involves SSB interactions with protein partner ( s ) other than χψ . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"microbial",
"mutation",
"dna",
"replication",
"genetics",
"biology",
"and",
"life",
"sciences",
"dna",
"microbiology",
"dna",
"synthesis",
"gene",
"function"
] | 2014 | ssb Gene Duplication Restores the Viability of ΔholC and ΔholD Escherichia coli Mutants |
Visceral leishmaniasis ( VL ) is expanding in Brazil and in other South American countries , a process that has been associated with the urbanization of the disease . This study analyzes the spatial and temporal distribution of VL in the Brazilian state of Minas Gerais and identifies the areas with higher risks of transmission . An ecological study with spatial and time series analyzes of new confirmed cases of VL notified to the Brazilian Notifiable Disease Information System between 2002 and 2013 , considering the 12 mesoregions of Minas Gerais . Two complementary methodologies were used: thematic maps of incidence and Poisson ( log-linear ) generalized linear model . Thematic maps using crude and smoothed cumulative incidences were generated for four trienniums . Poisson Regression measured the variation of the average number of cases from one year to the following , for each mesoregion . The 5 , 778 cases analyzed revealed a heterogeneous spatial and temporal distribution of VL in Minas Gerais . Six mesoregions ( Central Mineira , Jequitinhonha , Metropolitan area of Belo Horizonte , Northwest of Minas , North of Minas , and Vale do Rio Doce ) were responsible for the expansion and maintenance of VL , with incidence rates as high as 26/100 , 000 inhabitants . The Vale do Rio Doce and Jequitinhonha mesoregions showed a considerable increase in the incidence rates in the last period studied . The other six mesoregions reported only sporadic cases and presented low and unsteady incidence rates , reaching a maximum of 1 . 2/100 , 000 inhabitants . The results contribute to further the current understanding about the expansion of VL in Minas Gerais and may help guide actions for disease control .
Until 1980 , visceral leishmaniasis ( VL ) was considered a strictly rural disease in Brazil , the main parasite ( Leishmania infantum ) reservoirs being foxes ( Dusicyon vetulus and Cerdocyon thous ) and marsupials ( Didelphis albiventris ) [1] . However , the epidemiological profile of VL shifted with the urbanization of the disease and domestic dogs became the main reservoir ( Canis familiaris ) [1 , 2] . Since the first VL epidemic in 1981 in Teresina , the capital of the state of Piauí located in northeastern Brazil , various epidemics occurred in other major urban centers in the northeast ( São Luís , Natal , and Aracajú ) , north ( Boa Vista and Santarém ) , mid-west ( Cuiabá and Campo Grande ) , and southeast ( Belo Horizonte and Montes Claros ) regions of the country [3] . The geographic expansion of VL is associated with the process of urbanization of the disease . Indeed , migration of people from rural endemic areas to urban centers , adaptation of the vector to the domestic environment , the presence of disease reservoirs such as domestic dogs , malnutrition , and the lack of basic sanitation are considered contributing factors to the urbanization and geographic expansion of VL [3–5] . In this context , the Visceral Leishmaniasis Control and Surveillance Program ( VLCSP ) was implemented in Brazil to reduce the risk of transmission , the lethality , and the morbidity rates of VL in urban and rural areas . The program has three main pillars: the treatment of human cases , control of canine reservoirs , and vector control [1] . Aiming to identify areas where VLCSP strategies should be prioritized , a temporal study ( 2001–2011 ) of the disease’s incidence was conducted and pointed out the southeastern Brazilian state of Minas Gerais as a priority [6] . The VL incidence rates in this state were 1 . 6 and 1 . 4 for 100 , 000 inhabitants in 2012 and 2013 , respectively . These rates were superior to those found for the entire southeast region ( 0 . 6 e 0 . 5/100 , 000 inhabitants ) and similar to the national rates ( 1 . 6/100 , 000 inhabitants ) [7] . The first cases of human VL in Minas Gerais were registered from 1940 in the North of Minas mesoregion [8] and , in the 1960’s , in the Vale do Rio Doce mesoregion [9] . In 1989 , the first autochthonous case in an urban area was registered in the municipality of Sabará [10] , which belongs to the metropolitan region of Belo Horizonte . Later , in 1994 , the first autochthonous case was registered in Belo Horizonte , the capital of the state [11] . Currently , some cities of the state of Minas Gerais are considered endemic for VL and have attracted studies regarding the disease . Among those , stands out Montes Claros [12] and Porteirinha [13] in the North of Minas mesoregion; Paracatu [14] in the Northwest of Minas mesoregion; and Belo Horizonte in the Metropolitan area of Belo Horizonte mesoregion [15 , 16] . Governador Valadares , located in the Vale do Rio Doce mesoregion , registered cases of human VL in the 1960’s [9] and , following disease control measures , the municipality became a silent area for VL [17] . Unfortunately , disease control actions were interrupted in the 1990’s [18 , 19] and reporting of VL cases resumed in 2008 [19 , 20] to the point that , currently , Governador Valadares is considered a re-emergent focus of intense VL transmission [19] . Several studies have attempted to understand VL urbanization and geographic expansion by means of spatial and temporal analyses . They studied the distribution and variation of the incidence rates of human [6 , 21–23] the distribution of canine infection cases [24 , 25] the abundance of phlebotomine sand flies [25–27] , and the temporal trends of the disease [23 , 28 , 29] . Others identified areas where VL control measures could be prioritized [30–32] . Studies conducted in other Brazilian states such as Pernambuco [21] , Mato Grosso do Sul [22] , São Paulo [33 , 34] and Maranhão [35] evaluated the dispersion of VL over time . To our knowledge , no study to date has evaluated the VL dispersion in the state of Minas Gerais and therefore more robust and updated research is needed to unveil the VL profile in this area . Indeed , a combination of different methodologies for spatial and temporal analyses of VL may be useful to understand the aggregation , maintenance , and dispersion patterns of the disease , not only in Minas Gerais , but also in other regions of Brazil . The present study analyzed the spatial and temporal distribution of VL in the state of Minas Gerais between 2002 and 2013 using two different methodologies to further understand , characterize , and quantify the expansion of the disease in the vast territory occupied by this Brazilian state . The results of this study identify within the state the areas that could be prioritized by the control and vigilance of VL , considering the specificities of each mesoregion of the state of Minas Gerais .
This study was approved by the Ethics Committee on Research of the Federal University of Minas Gerais ( UFMG ) under the number CAAE n . 45497015 . 3 . 0000 . 5149 . Only secondary information about the patients was collected from the Brazilian Notifiable Disease Information System ( SINAN/VL ) . Data were analyzed anonymously . This is an ecological study analyzing the spatial and temporal patterns of confirmed cases of VL notified to the SINAN/VL in the period between 2002 and 2013 , in which the units of analysis were the 12 mesoregions of Minas Gerais . Minas Gerais is one of the 27 federative units of Brazil , being located in the southeast region of the country . It has an area of 586 , 521 . 24 km² and ranks as the fourth largest state in territorial extension . With the highest number of municipalities ( 853 ) in the country , Minas Gerais is the second most populous Brazilian state with an estimated population of 20 , 997 , 560 in 2016 , and a population density of 33 . 41 inhabitants/km2 . The capital is Belo Horizonte [36] . The Brazilian Institute of Geography and Statistics ( IBGE ) divides Minas Gerais in 12 mesoregions ( Fig 1 ) : Campo das Vertentes; Central Mineira; Jequitinhonha; Metropolitan Area of Belo Horizonte; Northwest of Minas; North of Minas; West of Minas; South/Southwest of Minas; Triângulo Mineiro/Alto Paranaíba; Vale do Mucuri; Vale do Rio Doce; and Zona da Mata [36] . Each mesoregion is an aggregate of municipalities from the same geographic area presenting similar social and economic profiles and natural conditions . In Brazil , VL is a disease of compulsory notification , i . e . , in the case of clinically suspected VL , health professionals must fill in a specific SINAN/VL form to start up procedures to investigate the disease . Initially , data such as the home address of the patient , age , gender , schooling , occupation , date when the first symptoms erupted , date of notification , and clinical manifestations ( signs and symptoms ) are collected . Later , additional information such as laboratorial exam results , date of treatment commencement , medication used , and outcomes are added to the system . During the period analyzed in this study ( 2002–2013 ) , the SINAN/VL database changed platform from Windows-based version ( the ILeishVi 2002–2006 ) to Net version ( LEISHNET , 2007–2013 ) . Therefore , the databases had to be standardized and a new unified dataset had to be generated with the same variables for analyzes using a Microsoft Office Excel 2013 . In the period analyzed herein , 13 , 409 suspected cases of VL were registered in the SINAN/VL . The number of confirmed cases was 6 , 158 ( 46% ) , 6 , 904 ( 51 . 5% ) cases were discarded because the disease was not confirmed , and 347 ( 2 . 5% ) cases had notification forms lacking information regarding the course of the disease . Among the confirmed cases , only new cases were included in the present study , totaling 5 , 778 cases ( Fig 2 ) . The data were analyzed in two phases . First , thematic maps of the state of Minas Gerais were generated with the crude and the smoothed cumulative incidence rates of VL for each municipality within the mesoregions of the state . The software MapInfo 10 . 0 ( MapInfo Corporation , Troy , New York ) and TerraView 4 . 2 . 2 ( Instituto Nacional de Pesquisas Espaciais , INPE , SP , Brazil ) were used . The information regarding the estimated resident population of each municipality and the cartographic basis of the state were obtained from IBGE [37] . Because the analyzes covered 12 years ( 2002–2013 ) of notifications , the data involved was divided into four maps , each comprising three years of study: 2002–2004 , 2005–2007 , 2008–2010 and 2011–2013 . The intervals of incidence rates used were chosen using the software MapInfo 10 . 0 considering the quartile , mean , median , and minimum and maximum values . The cumulative crude incidence was calculated using Microsoft Office Excel 2013 , and the smoothed cumulative incidence was obtained with the software TerraView 4 . 2 . 2 . The local empiric Bayesian estimator , which allows for the estimation of the incidence of a municipality using the incidence rates of the neighboring municipalities converging to a local mean , was used . To evaluate the spatial variability of the data , a spatial proximity matrix using contiguity-based spatial weights was built . The elements of this matrix can take values 1 ( if geographical analytical units are adjacent ) or 0 ( otherwise ) [38] . Following the calculations , thematic maps for both the crude and the smoothed cumulative incidence rates were generated using MapInfo 10 . 0 . Next , Generalized Linear Models ( GLM ) , through Poisson Regression , using STATA version 12 . 0 software ( Stata Corp . , College Station , TX , USA ) , was used to quantify the variation of the average number of VL cases from one year to the next in each of the 12 mesoregions of Minas Gerais . A curve with the annual incidence rates was generated for each mesoregion . Visual inspection of the resulting graphs allowed the establishment of cut off points according to the trends of increase or decrease of the number of VL cases over time . Consequently , the time interval analyzed was divided differently for each mesoregion . The response ( or dependent ) variable was the “number of cases” , and the “year” was used as the independent variable in the temporal series . The logarithm of the population [log ( pop ) ] was used as the “Offset” term , i . e . , was a known component included for adjustment of the model . The equation used was as follows: Log ( μi ) =log ( popi ) +α+β*year Where: The equation ( eβ-1 ) *100% was used to obtain the variation of the mean number of VL cases from one year to the following .
The present study only included new cases of VL , which totaled 5 , 778 . Of these , 89% were confirmed by at least one diagnostic test ( ELISA and/or IFAT and/or parasitological ) . Also , the SINAN /VL’s platform ( LEISHNET , n = 3349 ) included one variable informed that 95% of the cases were considered positive by clinical-laboratory criteria , while only 5% were confirmed by the clinical-epidemiological criteria . Therefore , we consider that the cases studied were correctly classified as VL . Of the cases analyzed , 91% were from urban areas . The mesoregions of Central Mineira , Jequitinhonha , Metropolitan area of Belo Horizonte , Northwest of Minas , and North of Minas concentrated most VL cases ( 91% ) and the highest number of municipalities with registered cases ( 66% ) . Among those , the Metropolitan area of Belo Horizonte mesoregion was the one with the largest population , the highest number , and the widest variation of VL cases , considering the population of its municipalities . On the other hand , the mesoregions of Campo das Vertentes , South/Southwest of Minas , Vale do Mucuri , and Zona da Mata presented the lowest number of cases ( 1% ) and the lowest number of municipalities which reported VL cases ( 13% ) ( Table 1 ) . The variation of the incidence from 2002 to 2013 was obtained from the thematic maps of crude ( Fig 3 ) and smoothed ( Fig 4 ) cumulative incidence rates , for each municipality of the 12 mesoregions of Minas Gerais . The highest incidence rates concentrated in the mesoregions located in the north ( Northwest of Minas , North of Minas , and Jequitinhonha ) , east ( Vale do Rio Doce ) , and central ( Central Mineira and Metropolitan area of Belo Horizonte ) parts of the state ( Table 2 ) . During the first three trienniums , the mesoregions of Northwest of Minas and North of Minas presented areas with the highest crude cumulative incidence rates . In the last triennium , however , a discrete decrease was observed in the rates in both these mesoregions . The highest incidence rate throughout the study , and among all the mesoregions , was observed in the Northwest of Minas mesoregion in the second triennium ( 67 . 7/100 , 000 inhabitants ) . This mesoregion presented a considerable increase in the incidence rates from the first to the second triennium and a subsequent reduction in the last two trienniums . Nonetheless , it was the mesoregion with the highest incidence rate within the state of Minas Gerais in the last triennium ( 31 . 2/100 , 000 inhabitants ) . Paracatu , one of the most populous municipalities of the Northwest mesoregion , presented high VL incidence rates throughout the period of study ( approximately 60/100 , 000 inhabitants ) . There was an increase in the incidence rates during the full study period in the municipalities of Unaí and Brasilândia de Minas , where VL cases were already observed between 2002 and 2004 . Moreover , new cases arose in other municipalities ( such as João Pinheiro and Guarda Mor ) , indicating the expansion of the disease in this mesoregion ( Fig 3 ) . This information is visualized in the map of smoothed incidence ( Fig 4 ) . In the last triennium ( 2011–2013 ) , however , a slight reduction was observed in the number of municipalities that presented cases of VL ( Figs 3 and 4 ) . The North of Minas mesoregion presented a high number of municipalities with VL cases during the four trienniums evaluated ( 17 municipalities ) . Among them , Montes Claros , Porteirinha , Matias Cardoso , Janaúba , Capitão Enéas , Montalvânia , Francisco Sá , and Nova Porteirinha showed the highest incidence rates . Figs 3 and 4 show the reduction in the incidence rates of VL in the municipalities of this mesoregion over the period analyzed . The incidence reduced from 29 . 9/100 , 000 inhabitants in the first triennium to 10 . 9/100 , 000 inhabitants in the last . This mesoregion showed the largest reduction in the incidence between the first and the second periods , as shown by the smoothed incidence rate maps ( Fig 4 ) . The Vale do Rio Doce mesoregion had the largest increase in the incidence rates over the temporal series . Although incidence rates in this region remained stable during the first two trienniums ( 1 . 1 and 1 . 0/100 , 000 inhabitants , respectively ) , a considerable increase was observed in the last two ( 10 . 1 and 10 . 4/100 , 000 inhabitants , respectively ) . This increase reflected the expansion of VL to the east and the center of the mesoregion , particularly to the municipalities of Governador Valadares , Conselheiro Pena , Ipanema , Resplendor , and Tumiritinga ( Figs 3 and 4 ) . The Central Mineira and Metropolitan area of Belo Horizonte mesoregions had small oscillations in the incidence rates of VL throughout the study period ( Table 2 ) . Nevertheless , the incidence rates reduced in the final triennium ( 10 . 3 and 8 . 5/100 , 000 inhabitants , respectively ) in relation to the initial ( 11 . 8 and 11 . 1/ 100 , 000 inhabitants ) . The following municipalities showed incidence rates higher than 5/100 , 000 inhabitants in all four trienniums: Curvelo , Presidente Juscelino , and Inimutaba ( all within the Central Mineira mesoregion ) ; Belo Horizonte , Ribeirão das Neves , Sabará , Ibirité , Prudente de Morais , Vespasiano , Jaboticatubas , Sarzedo , Sete Lagoas , São Joaquim de Bicas , and Nova Lima ( all within the Metropolitan area of Belo Horizonte mesoregion ) ( Figs 3 and 4 ) . Slight oscillations in VL incidence rates were also observed in the Jequitinhonha mesoregion . This oscillation is difficult to detect in the maps because different municipalities presented VL cases at various times ( Figs 3 and 4 ) . During all the time intervals evaluated , only seven municipalities ( Berilo , Araçuaí , Jequitinhonha , Almenara , Itaobim , Diamantina , and Virgem da Lapa ) stood out with very intense colors in the maps . In the last triennium ( 2011–2013 ) the incidence rate increased ( 15 . 8/100 , 000 inhabitants ) in comparison with the three previous trienniums ( 10 . 7 for the first; 13 . 3 for the second , and 8 . 2/100 , 000 inhabitants for the third ) . The other mesoregions ( Campo das Vertentes , West of Minas , South/Soutwest of Minas , Triângulo Mineiro/Alto Paranaíba , Vale do Mucuri , and Zona da Mata ) reported few VL cases and , consequently , low incidence rates throughout the study ( Figs 3 and 4 ) . Thus , these mesoregions were less important , as well as their variations over time in relation to VL . Fig 5 shows the crude VL incidence rates obtained for each mesoregion between 2002 and 2013 . These rates fluctuated over time and from one mesoregion to another . The following mesoregions reached elevated incidence rates , ranging from 0 to 26/100 , 000 inhabitants: Central Mineira , Jequitinhonha , Metropolitan area of Belo Horizonte , Northwest of Minas , North , of Minas and Vale do Rio Doce . On the other hand , the following mesoregions presented low VL incidence rates , reaching a maximum of 1 . 2/100 , 000 inhabitants: Campos das Vertentes , West of Minas , South/Southwest of Minas , Triângulo Mineiro/Alto Paranaíba , Vale do Mucuri , and Zona da Mata ( Fig 5 ) . The VL incidence rates in the whole state of Minas Gerais varied from 1 . 6 to 3 . 5/100 , 000 inhabitants . The time periods used in the Poisson Regression model were selected by visual analysis of the graph presented in Fig 5 . For each mesoregion , different time intervals were chosen , considering the intervals of increase or reduction in the incidence rates per mesoregion . The results of the adjustment of the model are shown in Table 3 . A single period ( 2002–2013 ) was chosen to analyze the mesoregions of Campos das Vertentes , Central Mineira , South/Southwest of Minas , Vale do Mucuri , and Zona da Mata ) . This was because it was not possible to observe trends in growth or reduction in the average number of VL cases occurring in these mesoregions , given the instability in the number of cases observed in the temporal series analyzed ( Fig 5 ) . The results obtained with the Poisson regression corroborated the analysis of the graph , as the results of the model were not significant , except in the case of the mesoregion South/Southwest of Minas . Indeed , the variation in the average number of VL cases per year was relatively stable during the study period ( Table 3 ) . However , the slope coefficient ( β ) obtained for the South/Southwest of Minas mesoregion was -0 . 22 , indicating a reduction in the average number of cases between the beginning and the end of the period analyzed , despite the fluctuations observed over the years ( Fig 5 ) . The Metropolitan area of Belo Horizonte , Northwest of Minas and North of Minas mesoregions presented similar results . Table 3 showed that , in the initial periods , the average number of cases per year rose 5% , 152% , and 43% , respectively . In the last periods , the average number of cases reduced in 15% , 12% and 13% , respectively . Of note is the mesoregion Northwest of Minas , which exhibited a marked increase in the average number of VL cases between the years of 2002 and 2005 ( 152% ) . The time intervals analyzed in the West of Minas and Vale do Rio Doce mesoregions were divided into three periods . It was not possible to establish a significant increase or reduction in the average number of cases per year , in the first periods analyzed ( Table 3 ) . In the intermediate intervals , the coefficients were positive , demonstrating an increase in the average number of cases ( 108% in the West of Minas and 90% in Vale do Rio Doce ) . On the other hand , the last periods showed a reduction of 34% ( West of Minas ) and of 19% ( Vale do Rio Doce ) in the number of VL cases from one year to the next . Two periods were chosen to analyze the Triângulo Mineiro/Alto Paranaíba mesoregion: 2002–2009 and 2009–2013 . In the first period , the mesoregion presented a positive slope coefficient of the model ( 0 . 25 ) , with an increase of approximately 28% in the average number of VL cases from one year to the next . In the interval between 2009 to 2013 , the variations were not significant and it was not possible to establish a co-relation between the average number of cases from one year to the next . Thus , the average number of cases remained relatively constant during this period ( Table 3 ) . Three periods were chosen to analyze the Jequitinhonha mesoregion: 2002–2005 , 2005–2008 , and 2008–2013 . In the first interval , the model coefficient was 0 . 51 , indicating an increase of approximately 66% in the average number of cases from one year to the next . On the other hand , the coefficient was negative ( -0 . 39 ) in the period between 2005 and 2008 , demonstrating a reduction of 32% in the average number of cases . During the final period , this mesoregion was the only one that presented a positive coefficient ( 0 . 16 ) , with the average number of cases increasing by 17% from one year to the next .
The present study shows that VL had a heterogeneous spatial and temporal distribution in the state of Minas Gerais , in the period between 2002 and 2013 . Among the 12 existing mesoregions , six ( Central Mineira , Jequitinhonha , Metropolitan area of Belo Horizonte , Northwest of Minas , North of Minas , and Vale do Rio Doce ) were responsible for the expansion and maintenance of VL in the state . Among them , the Vale do Rio Doce and Jequitinhonha mesoregions presented a considerable increase in the incidence rates of the disease in the last triennium ( 2011–2013 ) . The North of Minas Gerais and Metropolitan area of Belo Horizonte mesoregions reduced the incidence rates in the last years of the study , despite the elevated number of VL cases . In the other six mesoregions ( Campo das Vertentes , West of Minas , South/Southwest of Minas , Triângulo Mineiro/Alto Paranaíba , Vale do Mucuri , and Zona da Mata ) , only sporadic cases of the disease were reported during the study period and , consequently , these regions showed low and unsteady VL incidence rates . VL is expanding in Brazil [2] and in other South American countries [39] . In Argentina [40 , 41] and Paraguay [42 , 43] a significant increase in the number of VL cases was observed in the last two decades , raising concerns about the spreading of the disease . Therefore , studies evaluating the process of expansion of VL and the spatial and temporal variation of its incidence are of great importance . The Metropolitan area of Belo Horizonte mesoregion comprises a large number of municipalities ( 105 ) and stands out from the other mesoregions of Minas Gerais for being the most urbanized and the most economically developed , and where the political , financial , commercial , educational , and cultural centers of the state are concentrated [36] . Dissemination and urbanization of VL became even more evident in the Metropolitan region of Belo Horizonte after an autochthonous case of VL was registered in the municipality of Sabará [10] . The results presented herein reveals that since 2002 this municipality has higher , albeit oscillating , VL incidence rates than other municipalities of the same mesoregion considered of high VL transmission risk such as Sarzedo , Jaboticatubas , Sete Lagoas , and Vespasiano . Indeed , the 1990’s saw increasing numbers of VL cases being registered in the Metropolitan area of Belo Horizonte mesoregion and that trend persisted until the early 2000’s [44] . Previous studies in this region pointed out that the VL cases occurred in non-rural areas , reinforcing the urbanization process of the disease [44] , which becomes more likely in localities of high population density ( Table 1 ) and with houses close to one another [45] . The present study suggests that this trend persisted at least up to 2008 since map analysis detected a decreasing trend only in the last years analyzed . Indeed , the Poisson regression adjustment indicated an increase in the average number of cases between the years of 2002 and 2008 ( 5% ) and a reduction between 2008 and 2013 ( -15% ) . Recent studies have shown that VL cases are heterogeneous in the capital Belo Horizonte [16 , 46] . This may be due to the city’s vast territorial extension , high population density , and different microenvironments [47] . The VLCSP implemented in Belo Horizonte in the 1990’s [48] is considered to take place in a systematic and orderly manner throughout the city [49] and this may explain the reduction in the VL incidence rates observed in the present study . It is possible that the situation of Belo Horizonte reflects in the adjacent municipalities as well as in the mesoregion of the Metropolitan area of Belo Horizonte . This may be due to the population densification in these areas [45 , 47] , which compromises the execution and maintenance of measures to control the disease . Therefore , both Belo Horizonte and its mesoregion presented similar incidence rates , with a gradual increase in the incidence of VL in the first trienniums and a decrease in the last . Different from the Metropolitan area of Belo Horizonte mesoregion , the North of Minas mesoregion presents more ancient cases of the disease [8] . Nevertheless , research performed in some municipalities of this mesoregion , such as Montes Claros [8 , 50] and Porteirinha [51] pointed out that the incidence of VL is decreasing in the last years . The authors attributed this decrease to the VL control actions that are being performed in these municipalities . This reduction in the number of cases in this mesoregion is in agreement with the results obtained herein . Accordingly , the results of the Poisson regression adjustment , revealed a reduction in the average number of VL cases in the North of Minas from 2004 onwards ( -13% ) and the thematic maps , presented similar data to those previously described in Montes Claros in the same period [8] . Indeed , while investigating the VL cases registered in Montes Claros between 2001 and 2007 , these authors observed that , from 2005 onwards , the incidence of VL reduced , but the municipality remained endemic for the disease and representing a serious public health problem [8] . Because VL is associated with poor socioeconomic conditions [2 , 12 , 31] and that Montes Claros is considered the municipality with the best socioeconomic indexes in this mesoregion , one can assume that the problem regarding VL extends to the whole of the North of Minas mesoregion . The thematic maps indicated that the Northwest of Minas and Vale do Rio Doce mesoregions had pronounced increase of VL cases during the study period . On the other hand , the Poisson regression showed that , despite the increase observed in some periods , there was a reduction in the incidence rates starting in 2005 in the Northwest of Minas and in 2010 in the Vale do Rio Doce . In 2005 , VL expanded to other municipalities in the Northwest of Minas mesoregion . Initially , Unaí and Paracatu stood out as typical examples of VL urbanization [14] . Later , neighboring municipalities , such as João Pinheiro and Brasilândia de Minas , also began to show a high incidence of the disease . Noteworthy , the four mentioned municipalities occupy a vast territorial extension in the mesoregion and kept a cumulative incidence superior to 20/100 , 000 inhabitants in the last years . The high incidence of VL in the first three periods analyzed in the thematic maps and its reduction during the last period in the mesoregions North of Minas and Northwest of Minas may be explained by their geographical proximity . The thematic maps generated for the Vale do Rio Doce mesoregion for the period between 2002 and 2004 reveal only five municipalities with cumulative incidence higher than 5/100 , 000 inhabitants . From 2008 onwards , expansion of VL is observed in this mesoregion with 10 municipalities presenting VL incidence rates higher than 5/100 , 000 inhabitants , in which stands out Governador Valadares . In this mesoregion , human VL cases started to be reported the 1960’s [9] . At that time , control measures such as the treatment of VL patients with Glucantime , elimination of dogs positive for VL , and use of the insecticidal dichlorodiphenyltrichloroethane–DDT in the domicile and peridomicile areas were established in the municipalities with an expressive number of cases . These measures resulted in a progressive reduction in the number of cases and the absence of VL in the years of 1978 and 1979 [52] . As an unfortunate consequence , disease control measures were interrupted in the 1990’s [18 , 19] . Since 2008 , studies have identified cases of the disease in Governador Valadares [19 , 20] , which may explain the increase in the incidence of VL in the mesoregion of Vale do Rio Doce in the same year and the increase in the average number of cases between 2006 and 2010 . The urbanization process that is taking place in the municipality of Governador Valadares and the interruption of the disease control measures previously implemented in this mesoregion [19] may be the reasons underlying the observed increase in the number of cases VL . These results indicate the relevant role of Governador Valadares on the increase in the number of cases of the Vale do Rio Doce mesoregion in the last years analyzed . The results presented herein showed the geographical expansion of VL in Minas Gerais between 2002 and 2013 . The data reveal a trend for VL persisting in municipalities that already presented cases , even when there was oscillation in the disease incidence rates . Despite this expansion , the incidence of the disease does not spread from the North region of the state to the South , that is , to mesoregions such as Campos das Vertentes and South/Southwest of Minas , which presented a consistently low number of cases . The persistence of the disease in the mesoregions located in the North of Minas Gerais may be related to socioeconomic [5 , 6 , 31] and environmental [6] factors . A study performed in Montes Claros showed that VL in this region is associated with poor domiciles and precarious sanitation conditions , which facilitates the accumulation of organic material and other factors that enable the life cycle of the main vector of VL , Lutzomya longipalpis [12] . The maintenance of the disease cycle explained by similar reasons was also observed in the Jequitinhonha ( Araçuaí ) [53] . The findings reported herein corroborate these previous observations , since the highest VL incidence rates were found in the North of Minas Gerais , a region that presents low Human Development Index [54] . Considering the environmental factors , several authors described what Sherlock ( 1996 ) [55] previously observed in the northeastern state of Bahia: that VL is occurring with higher frequencies in hot dry climates common to the northern areas of the Minas Gerais state [12 , 56 , 57] . These observations are in agreement with the results presented in the current study since it detected high incidence rates especially in the mesoregions located at the north of the state , which also have a hot and dry climate . We suggest that VL presence in the northern region of Minas Gerais is mainly due to two reasons: climate and socioeconomic factors . Indeed , we observed that VL is distributed in contiguous mesoregions , namely Northwest of Minas , North of Minas , and Jequitinhonha , all of which have low socioeconomic development and display a warm climate favorable to the development of the vector [1] . On the other hand , the southern mesoregions , namely West of Minas , Campo das Vertentes , Zona da Mata , and South / Southwest of Minas , present lower temperatures and better socioeconomic conditions . Interestingly , although the mesoregion Triangulo Mineiro / Alto Paranaíba has a hot climate , it also presents few VL cases . This mesoregion is the second largest economy in the state and has the highest Human Development Index . The Human Development Index of the Belo Horizonte Metropolitan mesoregion is one the highest in the state of Minas Gerais , but there is a great socioeconomic disparity between the municipalities composing this mesoregion . The high levels of social inequality , as represented by the existence of slums within the municipality of Belo Horizonte itself , may underlie the high number of VL cases observed in this city . This study has a few limitations that should be pointed out . Even though VL is of compulsory notification in Brazil , the real number of cases may be underestimated as the symptoms of the disease are unspecific , and some cases may go unreported . This might compromise calculations of the incidence and the estimates obtained through the Poisson regression . However , the extent of this underreporting is most probably minimum since SINAN covers all health systems ( public and private ) at its various levels of complexity . Furthermore , it is noteworthy that the medication used for treatment is solely dispensed by the Brazilian Health Public System and this measure has minimized underreporting . Some variables of this system , such as “Final Classification” ( cases that were confirmed , discarded or had incomplete forms ) and “Type of entry” ( New Case , Recurrence , Transference , and Ignored ) are not filled after closure of the case , thus , increasing the number of losses ( missing ) and compromise the analyzes . Lastly , areas bordering the state of Minas Gerais were not evaluated . Furthermore , an empiric local Bayesian approach was used in one of the sets of the thematic maps created . This method estimates the mean local incidence in each municipality taking into account the incidence values of the neighboring municipalities . Thus , the rates generated are corrected , smoothed , and less unstable . Maps built using this approach are , therefore , more informative and interpretative . The spatial and temporal epidemiology , together with the Poisson regression approach , allowed a more precise identification and quantification of areas of expansion and stabilization of VL , and the identification of regions that share similar spatial patterns . The findings reported herein may help to guide the implementation of actions for controlling the disease in the state of Minas Gerais . Given the worrisome expansion of VL in Brazil and in other South American countries , the results describing the spatial and temporal pattern of VL expansion in this vast geographic area may be relevant to researchers following the disease in other regions of Brazil and the world . | This article presents the spatial and temporal distribution of visceral leishmaniasis ( VL ) in Minas Gerais State and identifies the greater risk areas of transmission . This study is both timely and substantive because Minas Gerais is an important Brazilian state in the number of cases of visceral leishmaniasis . The results showed that during the 12-year time series the VL had a heterogeneous spatial and temporal distribution in the state of Minas Gerais . Among the 12 existing mesoregions , six ( Central Mineira , Jequitinhonha , Metropolitan area of Belo Horizonte , Northwest of Minas , North of Minas , and Vale do Rio Doce ) were responsible for the expansion and maintenance of VL in the state . Among them , the Vale do Rio Doce and Jequitinhonha mesoregions presented a considerable increase in the incidence rates of the disease in the last period . In the other six mesoregions only sporadic cases of the disease were reported during the study period . The results of in this study may contribute to a better understanding the dynamic of the disease in Minas Gerais . Also these findings can provide subsidies to assist the actions of the control program of VL . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"kala-azar",
"urban",
"geography",
"animal",
"types",
"medicine",
"and",
"health",
"sciences",
"domestic",
"animals",
"tropical",
"diseases",
"geographical",
"locations",
"social",
"sciences",
"parasitic",
"diseases",
"animals",
"neglected",
"tropical",
"diseases",
"inf... | 2017 | Spatial and temporal trends of visceral leishmaniasis by mesoregion in a southeastern state of Brazil, 2002-2013 |
Despite intensive study , most of the specific genetic factors that contribute to variation in human height remain undiscovered . We conducted a family-based linkage study of height in a unique cohort of very large nuclear families from a founder ( Jewish ) population . This design allowed for increased power to detect linkage , compared to previous family-based studies . Loci we identified in discovery families could explain an estimated lower bound of 6% of the variance in height in validation families . We showed that these loci are not tagging known common variants associated with height . Rather , we suggest that the observed signals arise from variants with large effects that are rare globally but elevated in frequency in the Jewish population .
Height is a classic genetically complex quantitative trait with high heritability ( ~80%-90% [1 , 2] ) . Despite intensive study , the genetic basis of variation in height remains mostly unexplained . Genome-wide association studies ( GWAS ) in hundreds of thousands of individuals have identified hundreds of common variants significantly associated with height [3 , 4] . However , the individual effect sizes of these variants are small , and all variants identified to date jointly explain only ~20% of the heritability of height . One proposed explanation for the gap between the overall heritability of height and that explained by common variants is the contribution of rare genetic variants with large phenotypic effects [5] . Recent studies lend support to this idea by identifying such variants and showing that their effect sizes are inversely related to their frequencies [6 , 7] . Several examples of large-effect variants that are rare globally but are more common in certain founder populations have been reported for height ( in Sardinians [8] and in Puerto Ricans [9] ) and diabetes ( in Greenlanders [10] ) , however , identifying associations between rare variants and traits of interest typically requires very large sample sizes [11] . For instance , 750 , 000 participants were required to identify 32 rare variants ( those with frequency <1% ) that affect height [7] . Rare variants can in principle be identified in family-based linkage studies with lower sample size requirements than association studies . 21 family-based studies of height have been conducted , mostly prior to the GWAS era [12–32] . Only a few of these studies detected Quantitative Trait Loci ( QTLs ) linked to height at the accepted level of genome-wide statistical significance , with the three most convincing findings all reported by a single study with the largest sample size of subjects from one ethnicity [23] . Moreover , few if any of the reported QTLs were replicated across multiple studies [14 , 23 , 32] . We sought to overcome some of the limitations of previous linkage studies by focusing on several factors that influence statistical power , including the minor allele frequency ( MAF ) of causal variants , the strength of linkage between causal and typed variants , and the quality of genotyping and phenotyping [33 , 34] . Specifically , we attempted to increase power by studying the genetic architecture of height in very large nuclear families ( 10 to 20 siblings per family ) from a founder ( Jewish ) population . We acquired highly accurate measurements of height , and we used dense genotyping arrays to fully reconstruct the inheritance patterns in these families . The use of large pedigrees drawn from a population with a small effective population size should increase the effective allele frequency of variants of interest , enabling their detection in a study with a modest sample size .
To increase the power to detect effects of rare variants , we sought to increase their effective frequency in our cohort by studying very large nuclear families . A rare variant carried by a parent of such a family automatically rises to a frequency of ~25% among the children . However , this effect is rapidly diluted when many unrelated small families are combined , as was done in all previous family studies of height . To minimize such dilution , we recruited 397 participants from 29 very large nuclear families containing 10 to 20 siblings per family ( mean = 12 siblings ) . In addition , siblings in eight of the nuclear families have first cousins in several other nuclear families in the cohort . Any variant segregating in our cohort has a minimum expected MAF of ~1% if it is present in only one family , and higher if it is present in multiple families , regardless of its frequency in the general population . See S1 Table for more details on the cohort . Another approach to increase MAF is to study a founder population , where some variants that are rare in a cosmopolitan population can rise to high frequencies due to a small number of founders and subsequent genetic drift . We therefore recruited our cohort from the Jewish population . Specifically , 80% of our cohort consists of Ashkenazi Jews ( AJ ) . Today’s 10 million AJ have an effective population size ( Ne ) of ~350 as a result of a founder event ~700 years ago [35] . This effective population size is small even compared to other founder populations such as Finland ( Ne≈3000 ) [36] and Iceland ( Ne≈5000 ) [37] . Two unrelated AJ on average share ~30 times more of their genome in long ( >3Mbps ) identity by descent ( IBD ) segments than two unrelated non-Jewish Caucasians [38] . As a result , variants that are rare in other populations can rise to high frequencies in AJ . Indeed , at least 40 Mendelian genetic diseases in AJ are caused predominantly by such founder mutations that are non-existent or rare in other populations [39 , 40] . The other 20% of our sample are Jews of other ethnicities . Previous studies showed that Jews who are not Ashkenazi are on average closer genetically to Ashkenazi Jews than to their non-Jewish neighbors [41] . To estimate the shift in allele frequencies between the European and AJ populations , we compared allele frequencies in whole genomes of 7509 non-Finnish Europeans and 151 Ashkenazi Jews , both from the gnomAD database [42] ( Fig 1 ) . The database contains 90 . 5 million bi-allelic high quality variants found in Europeans , and 82 . 5 million ( 91% ) of these are rare ( MAF<1% ) . 95% of the rare variants found in Europeans are not observed in the AJ sample , but 33 , 331 variants that are rare or not observed in Europeans are common in AJ ( allele frequency ≥5% ) , and 757 of them reach allele frequency ≥10% . To test whether the increase in the frequency of rare variants in AJ is a consequence of small sample size , we used the European allele frequencies as probabilities for randomly sampling 151 Europeans from gnomAD and repeated the analysis . No rare variant reached a frequency of ≥5% in this sub-sample ( Fig 1B ) . Increasing the MAF of a variant increases the power to detect that variant in several ways . First , power is dependent directly on MAF . Second , rare variants tend to have a lower quality of genotyping and imputation [33] , and increasing MAF allows for a higher quality of both , especially when using a population-matched reference panel for imputation . Third , the effect sizes of variants are typically inversely related to their frequencies in cosmopolitan populations [6 , 7] , and therefore causal variants that are rare elsewhere but are common in our study population may have larger effect sizes . The combination of these factors , together with our use of large families and accurate phenotyping ( see below ) , has the potential to increase power to detect Quantitative Trait Loci ( QTL ) that influence height . Most previous height studies used measurements acquired incidentally as part of studying other traits or diseases . Such measurements may suffer from low accuracy . To ensure phenotype accuracy and increase statistical power , we measured the height of each participant to the nearest 0 . 1 cm , with four technical repeats , and with one researcher ( D . Ze . ) conducting all measurements and using the same measurement system . We tested different methods to correct heights for age and sex and used a non-linear correction for age that maximized heritability in our sample . Nearly all previous family linkage studies of height [12–32] used sparse maps of microsatellite markers , which provide incomplete inheritance information content and a corresponding reduction in statistical power to detect QTL [43] . A dense map of single nucleotide polymorphisms ( SNPs ) , combined with multipoint linkage analysis , can generate near-perfect information content [43] , but existing computational tools are not well-suited to carry out multipoint analysis with hundreds of thousands of markers in very large families . To obtain fully informative inheritance patterns at every position in the genome , we genotyped ~630 , 000 SNPs and developed a method for identity-by-descent ( IBD ) reconstruction that leverages information from the large number of siblings in each family . Briefly , we compared genotypes of each sibling pair to identify IBD segments , and then used these segments identified across all pairs to reconstruct , at every genomic position , the fully informative inheritance pattern of all four parental haplotypes in the children ( S1 Fig ) . Every recombination event in a sibling was identified as a change in his or her IBD relationship pattern with the other siblings . Because the number of recombination events per chromosome per participant ( typically 1–2 ) is much smaller than the number of siblings in each of our families , we were able to identify these IBD switching events with high certainty , even in the absence of parental genotypes . To test the accuracy of the method , we calculated the average pair-wise IBD sharing for the genome . Two siblings shared zero , one or two alleles IBD for 24 . 9% , 50% and 25 . 1% of the genome , compared to the theoretical expectation of 25% , 50% and 25% . The IBD segments identified by our method provide near-perfect inheritance information for linkage analysis between genomic loci and height . To identify regions of the genome that co-segregate with height differences ( QTLs ) , we conducted linkage analysis by contrast tests ( ref . [44] ) . Briefly , at every position in the genome , we compared the heights of siblings that inherited one of the two possible haplotypes from a given parent to the heights of those who inherited the other haplotype . We used permutations of height among siblings in a family to calculate significance , which we express as an equivalent LOD score ( S2 Fig ) . It has been shown [45] that for sib pair analysis , genome-wide significance at an FDR of 5% corresponds to LOD≥3 . 6 ( equivalent to p-value ≤ 2X10-5 ) . We identified two significant QTLs , on chr14:71 , 145 , 000–71 , 855 , 000 bp ( hg19 genome coordinates ) with LOD = 4 . 13 , and on chr17:78 , 845 , 000–79 , 695 , 000 bp with LOD = 3 . 8 . It is important to note that our QTL sizes are large , spanning hundreds of thousands of base pairs ( as is typical in linkage analysis ) , compared to GWAS results that identify variants in LD blocks that have typical sizes of tens of thousands base pairs , and that sometimes can even indicate the causal SNPs if they are genotyped . Permutation analysis showed that 0 . 11 QTLs are expected by chance at this threshold , which corresponds to an empirical FDR of 5 . 5% , in close agreement with the theoretical value . The QTL on chromosome 14 includes two genes: PCNX , an oncogene [46] and homolog to a drosophila component of the Notch signaling pathway [47] , which functions in several developmental processes , and MAP3K9 , a Mitogen-activated protein kinase . The QTL on chromosome 17 contains 25 genes , including an interesting candidate , RPTOR—a binding Partner of Target of Rapamycin ( TOR ) which control cell growth [48] . Mutations in RPTOR have been shown to cause significant reduction in body size in flies [49] and mice [50] . We reasoned that because height is a complex trait that is governed by multiple loci with relatively small effects , an FDR analysis of larger sets of loci with lower locus-specific significance would be appropriate and could increase the sensitivity of our study . This approach has been previously employed in GWAS [51] , but was not attempted in previous linkage studies of height . To carry out such analysis , we compared the number of observed QTLs for a range of LOD score thresholds below 3 . 6 to the number expected in the absence of true signals , determined from permutations . As expected , lower thresholds resulted in larger numbers of detected QTLs , with an increasing FDR ( Fig 2A and 2B ) . Importantly , the total number of detected QTLs remained significant ( P<0 . 05 compared to permutations , Fig 2C ) from the strictest threshold ( LOD = 3 . 6; 2 QTLs; P = 0 . 009; FDR = 5 . 5% ) , down to LOD = 1 . 1 ( 24 QTLs; P = 0 . 019; FDR = 68%; S2 Table ) . We used the FDR to estimate the number of true QTLs as a function of decreasing detection thresholds ( Fig 2A ) . The number of true QTLs maximizes at LOD = 1 . 3 , where we estimate 8 . 8 true QTLs out of 20 that were detected ( P = 0 . 002; FDR = 56% ) . To estimate how much phenotypic variance can be explained by the detected QTLs , and to test whether the detected loci tag common SNPs identified by previous height GWAS [52] , we conducted variance partitioning in a cross validation framework designed to avoid potential overfitting that could result from detecting QTLs and estimating their effect sizes in the same dataset . We generated 100 training sets that each randomly sampled 2/3 of the families . In each set , we mapped QTLs as described above . The results were similar to those obtained with the full dataset ( S3 Fig ) , although on average , fewer QTLs were in the training sets as a consequence of the smaller sample size . We then used the 1/3 of the families held out of each training set as a test set , and simultaneously estimated the contributions of three different sources of genetic variance to height variance in a variance components model ( implemented in the software package GCTA [53] ) . Specifically , the model included variance component terms for the detected QTLs ( using SNPs from the detected QTLs in the training sets ) , the common SNPs associated with height by GWAS [52] , and the overall genomic relatedness among individuals; the latter controls for pedigree structure , and can also capture a polygenic signal of height . To further control for upward bias in the estimation of variance explained in the test set due to shared environment of family members and tagging of polygenic background , we used as a baseline the variance explained in GCTA in a similar model with the same number of SNPs but from random genomic segments similar in size to the real QTLs ( “random QTLs” ) . The reported variance explained by QTLs ( Fig 2D ) are after subtraction of variance explained by the “random QTLs” in the null model . The total variance explained by the QTLs increased as the detection threshold was lowered and a larger number of QTLs were included in the model , but the variance explained over and above the null model initially also increased ( Fig 2D ) . For example , at LOD≥3 , we detected an average of 0 . 74 estimated true QTLs per training set ( 1 QTL at FDR = 26% ) , and these explained an average of 1 . 3% of the phenotypic variance above the null model in the test sets . At a lower detection threshold of LOD = 1 . 9 , we estimated an average of 3 . 7 true QTLs ( 7 QTLs at FDR = 47 . 5% ) , and these explained on average 5 . 8% of the variance above the null model . For low detection thresholds ( LOD≤1 . 3 ) , the variance explained above the null model decreased , presumably because too many false discoveries were included . To test whether QTLs explained variance in height by tagging previously discovered height-associated common variants , we ran the variance components model with and without including the GWAS SNPs . The variance explained by the QTLs in both models was similar ( 0 . 1% difference , t-test P = 0 . 72 , S4 and S5 Figs ) . This result suggests that the QTLs we identified are novel and are not simply tagging common SNPs that were previously identified by GWAS . The GWAS SNPs explain in our variance components model 11 . 2±1 . 8% , uniformly across LOD thresholds regardless of the number of QTLs they compete against . This is similar to the amount of variance explained reported in the study which originally identified these GWAS variants 3 . To assess how much phenotypic variance can be explained by entire chromosomes , we estimated a genomic relatedness matrix ( GRM ) from all the SNPs on each autosome , and let all 22 GRMs compete together at the same time in a variance components model to explain phenotypic variance ( Fig 3 ) . We found no correlation between variance explained and chromosome length ( Pearson r = –0 . 08 , P = 0 . 71 ) , although we cannot rule out that such a correlation exists , as the standard errors in the estimation of single chromosome contributions are large . In contrast , the variance explained by chromosomes is highly correlated with the top LOD score on each chromosome ( Pearson r = 0 . 7 , P = 2 . 9x10-4; Spearman r = 0 . 6420 , P = 0 . 0016 ) . This correlation arises in part from the results for chromosome 14 , which explains the most variance in the variance components model ( 25%±7 . 8% ) and has the single most significant QTL . To test whether the correlation is driven solely by chromosome 14 , we omitted it from the analysis . The correlation fell but remained significant ( Pearson r = 0 . 53 , P = 0 . 015; Spearman r = 0 . 59 , P = 0 . 006 ) . These results suggest that at least in our sample , variance explained by some of the chromosomes captures contributions of small regions with large effects rather than solely infinitesimal contributions distributed throughout the entire chromosomes . To investigate this further , we simulated 100 sets of phenotypes from an infinitesimal model , in which normally distributed small effect sizes were randomly assigned to all SNPs in the genome while maintaining the overall heritability . For each simulated data set , we calculated the distribution of variance explained by entire chromosomes . The correlation between variance explained and chromosome length in the different simulated sets was r = 0 . 19±0 . 19 ( Mean±SD ) , stronger than the r = –0 . 08 observed for the real data , although the difference was not statistically significant ( P = 0 . 09 , S6 Fig ) . The observed variance explained by chromosome 14 was significantly higher than expected from the infinitesimal model . Only 5/2200 chromosomes in the simulated data sets explained as much as 25% of the variance ( Bonferroni corrected P = 0 . 05 , Fig 3C ) , and all of these five observations were for chromosomes that are longer than chromosome 14 .
Here , we studied height in a unique cohort of very large nuclear families from a founder population . This strategy was designed to increase the effective allele frequency of some variants that are otherwise rare , thereby also increasing our power to detect their effects on height . This approach enabled us to detect significant QTLs for height in a study with modest sample size . We also used FDR analysis to identify a larger number of QTLs that were highly significant as a set , despite few of the QTLs achieving significance individually . Using a variance components model , we showed that these QTLs explained 6% of the variance in height in a cross validation framework , and that they were not tagging common variants previously identified as associated with height by GWAS . The actual fraction of variance explained by the QTLs is likely higher because of the conservative nature of the estimation procedure . Further , we showed that the variance contributed by chromosome 14 , and possibly by some of the other chromosomes , arises at least in part from small regions with large effects ( which correspond to the detected QTLs ) , rather than solely from infinitesimal contributions distributed throughout the entire chromosomes . Taken together , these results suggest that variation in height in our sample arises from a combination of a small number of QTLs with large effects and a large number of common variants with small effects . Because the detected QTLs are not tagging previously identified common variants , they likely arise from variants that are elevated in frequency in the AJ population . Although we have not identified the specific variants underlying the QTLs , we speculate that candidate variants can be identified by sequencing the parents of the pedigrees and searching for variants that are rare in other populations , common in AJ , and follow segregation patterns consistent with the QTL signals . The approach described in this paper , coupled with recruitment of additional large families ( which are abundant in the Jewish population [54 , 55] ) , may provide further insights into the genetic basis of height and the role of population-specific vs . cosmopolitan variants , and may serve as a complement to GWAS for genetic investigations of other complex traits .
This study was approved by the IRB of Shaare Zedek Medical Center , Jerusalem , Israel ( IRB#131/12 ) , Princeton University , Princeton , NJ , USA ( IRB#0000006027 ) , and UCLA , Los Angeles , CA , USA ( IRB#14–000357 ) . In all cases written consent was obtained . Participants were recruited in Israel and in the US after receiving IRB approvals in both locations . All participants gave written informed consent , then filled a questionnaire about their growth process , medical history , lifestyle during growth years ( nutrition , sleep , physical activity , etc . ) , and ancestry origins and heights . Participants’ heights were measured with a Seca 213 mobile measurement system to the nearest 0 . 1 cm with 4 repeats ( participants stepped off and on the measurement system for repeated measurements ) . We also measured sitting height ( ±0 . 1 cm , 3 repeats ) and arm span ( ±1 cm , 3 repeats ) . The participants donated saliva samples into DNA Genotek OGR-500 tubes from which DNA was extracted by Ethanol precipitation . The samples were genotyped by RUDCR ( Rutgers ) using Affymetrix Axiom Biobank Array ( ~630 , 000 SNPs ) . The Heights distributions in our sample were normal with 210 males at 172 . 7cm ± 5 . 7cm ( Mean ± S . D . ) and 187 females at 161 . 6 ± 5 . 5 cm ( Mean ± S . D . ) and are representative of the Israeli population ( Mean = 173 . 7 cm for Israeli males and 160 . 3 cm for females [56] ) . See S1 Table for more details on the cohort . Longitudinal studies have shown that people “shrink” non-linearly with age , and that the rates of shrinkage also differ for men and for women [57 , 58] . We therefore used the data of a previous longitudinal study [57] to derive sex specific non-linear equations for height shrinkage as a function of age . Their data included measurements of the rate of height change for every decade for men and women separately from age 20 to 90 . They showed that men and women start shrinking around age 30 and that between age 30 and 80 the rate of height loss increases linearly with age ( i . e . constant acceleration of height loss ) . We plotted these rates and used a linear fit to derive the dependence of the rate of height loss on age . Height loss rate increases by 0 . 00416 cm per year for men , and by 0 . 00641 cm per year for women both starting at age 30 . We integrated these equations to receive quadratic formulas for height loss: Defining: H≡HeightLoss For males: From plotting longitudinal data [57] and fitting a linear regression: dHdAge=−0 . 00416×Age+0 . 124 Age when starting to shrink: dHdAge=0→Agestartshrinking=0 . 1240 . 00416=29 . 8 Defining t ≡ Age − 29 . 8 The rate of height loss after age 29 . 8: dHd ( t ) =−0 . 00416×t Integrating to get the height loss as a function of time ( after age 29 . 8 ) : H ( t ) =∫dHd ( t ) =∫−0 . 00416×t=−0 . 00208×t2+C H ( t=0 ) =0→C=0 H ( t ) =−0 . 00208×t2 For females: Finally , since participants reported ages in full years , we approximate: HeightLoss ( Age ) ={0 , Age≤30−0 . 00208× ( Age−30 ) 2 , Males , Age>30−0 . 003205× ( Age−30 ) 2 , Females , Age>30 To compare our quadratic correction for age to a linear one , we first estimated the linear correction from our data by using SOLAR [59] and applying age and sex as covariates in a linear model . The linear correction for height was 0 . 0995* ( Age-35 . 084 ) + Height corrected for sex . We then compared the linear and quadratic corrections by estimating ( using SOLAR ) the heritability of our cohort for heights corrected by the two models . The heritability of height after a quadratic correction to age was h2±S . E = 0 . 86±0 . 07 , higher than the h2±S . E = 0 . 81±0 . 08 achieved after a linear correction . This improvement of the quadratic model over the linear one might be an underestimation since the quadratic correction was estimated from a different cohort of a longitudinal study , while the linear correction was estimated with the same data that was used to estimate the heritability . For any further analysis we therefore used the quadratic correction . To correct height for sex we standardized ( z score ) the heights of females and males separately and then pooled the standardized heights together . To calculate power , we used ANNOVA F-test , as follows in Appendix A5 of [34]: The non-centrality parameter ( NCP ) : λ=f*n* ( MAF−MAF2 ) *β21−0 . 5*[h2−2* ( MAF−MAF2 ) *β2 ) where: f = number of families , n = number of siblings per family , MAF = Minor Allele Frequency , β = effect size ( assumed 1 SD ) , h2 = heritability ( assumed 80% ) and: Power=Pr[Fdf1 , df2 , λ>Fdf1 , df2 , [1−α]] Where df1 = 2*f , df2 = f*n-3 , α = significance level corresponding to the LOD detection threshold In R: chi_threshold = lod_threshold * 2 * log ( 10 ) a = pchisq ( chi_threshold , 1 , lower . tail = FALSE ) power = pf ( a , df1 , df2 , ncp , lower . tail = FALSE ) The Affymetrix Axiom Biobank Array was used for genotyping . It covers 628 , 679 SNPs distributed across the genome . Samples had on average 8043 ± 3291 ( mean ± S . D ) SNPs reported as “No Call” ( 1 . 3% ± 0 . 52% ) , and 48 , 787 ± 12881 SNP calls ( 7 . 8% ± 2% ) reported as “Low Quality” ( P>0 ) . To assess the experimental technical error , we genotyped one sample 4 times ( in different Axiom array 96 plates ) . For the 6 two-way comparisons , excluding all “No Calls” in each compared pair , we had an error rate of 9 . 1x10-3 ± 2 . 5x10-3 ( mean ± S . D ) i . e . 1 SNP in every ~110 SNP is called wrongly . However , most of these errors were in the “Low Quality” category , since while only ~8% of the calls had low quality scores , when they were dropped from the comparisons , the error rate dropped ~20 fold to 4 . 45 x10-4 ± 2 . 7x10-4 or 1 SNP called with an error in every ~2247 SNPs . Given these results , we did not use any SNP with a low quality call score for IBD calling . To reconstruct the inheritance patterns in each family , we used only the informative SNPs that are not homozygote identical in all siblings of the family . Typically , it reduced the number of useable SNPs in a family from ~630 , 000 to ~180 , 000 SNPs , reflecting the high level of genetic homogeneity in our population . To increase accuracy , we excluded any low quality SNP calls ( Axiom Biobank array calls with a “Confidence” value > 0 or a “No Call” , a total of approximately 4% of the SNPs of each participant ) . This exclusion left ~150 , 000 informative SNPs for comparison between each sib pair . For each sib pair , we compared the SNP calls along the genome to partition the genome into regions with opposite homozygous calls ( indicating a region with 0 shared alleles ) , regions with no opposite homozygous calls but heterozygote calls in one sib vs homozygote calls in the other ( indicating 1 shared allele ) and regions with only identical homozygote calls ( 2 shared alleles ) . To avoid false positives due to genotyping errors we required a stretch of at least 3 SNPs of the same type ( opposite homozygotes , one heterozygote vs one homozygote , or two identical homozygotes ) within 2Mbps , 1Mbps and 1Mbps accordingly , to declare a region as 0 , 1 or 2 alleles shared . ( See S1 Fig for an example ) . To correct for errors of IBD calls between each sib pair we used the multiple siblings’ information by comparing all the sib pairs IBD within a family . For example , if sib #1 and sib #2 ( 1–2 ) share 2 alleles in some region and so do sib pairs 1–3 , 1–4 , 2–3 and 2–4 , we would expect sibs 3–4 to also share 2 alleles in this region and therefore a 1 shared allele for this sib pair would be likely a false negative and we should correct it to 2 shared alleles . We corrected such contradictions by applying the minimal number of corrections to sib pairs IBD calls while prohibiting creating any new contradictions with other sib pairs by the correction . Examining the matrix of the shared allele numbers between all sibling pairs shows the siblings falling into 4 groups ( 4 possible values for lines in the matrix ) or less , as expected by the 4 possible combinations of grandparental alleles . Advancing along a chromosome , these matrices remain identical up to some chromosomal position where one of the siblings changes suddenly its allele shared numbers with all other siblings and move from one grandparental allele group into another , indicating a recombination event . Further , the new grandparental alleles combination that the sibling switch into allows us to infer in which of his two haplotypes the recombination event occurred . An example of a fully phased IBD reconstructed map of a nuclear family is in S1 Fig . We conducted QTL mapping by marker contrast tests ( [44] , chapter 16 ) . We divided the siblings of each family every 5000 bases along the genome into two groups according to which grandparental haplotype they inherited from a specific parent ( the IBD reconstruction ) . We performed simple linear regression and calculated the coefficient of determination ( R2 ) between the two groups and the siblings’ heights . We repeated the calculation for the two haplotypes inherited from the other parent at the same genomic position to get a total of two R2 scores ( one from each parent ) that correspond to the association of this position to height differences ( See example at S1 Fig ) . To adjust for the different number of haplotypes underlying the two correlations in different genomic positions , we conducted permutation analysis . We permuted randomly the heights of the siblings within each family 1000 times while keeping the genetics fixed and calculated the resulting R2 scores for each genomic position . We then calculated two local P values for each genomic position for each family as the number of times a similar or larger R2 was achieved in the permuted families compared to the real family . This procedure also normalizes the signal for family size . We then take the lower of the two single family local P values , which allows for capturing also dominance effects ( results however were robust to taking the average ) and combine the signals from all families by taking the average of these local P values over all the families at each position . We then repeat the same calculation for the 1000 height permuted sets of families , and calculate an empirical genome wide P value by counting how many times each local P value or lower from the real families appears anywhere in the genome in any of the height permuted families’ combinations . Lastly , to make it easier to compare to previous linkage studies , we transform the global genome wide P values into LOD scores by using the inverse of the Chi-Square cumulative distribution function , as described in [60] . The LOD scores for all chromosomes are plotted in S2 Fig . To estimate the genomic distance in which two nearby QTLs can be called as independent , we investigated how far along a chromosome a LOD score is still correlated to other LOD scores of nearby positions . We calculated the absolute difference of LOD scores between any two positions on any given chromosome , and averaged over all chromosomes to get the mean absolute difference of LOD scores as a function of genomic distance ( S7 Fig ) . As expected , close genomic regions show , on average , similar LOD scores ( due to linkage ) , and increasingly larger genomic distances show monotonically increasing absolute LOD differences . This holds true up to a genomic distance of ~33Mbps where the LOD score difference reaches its median level , and for larger distances the LOD score absolute difference fluctuates around its median value . We infer that association between genotype and phenotype at some genomic position has on average no influence through linkage over the association at a distance larger than 33Mbps . We therefore count regions with LOD scores above detection threshold as belonging to the same QTL if they are less than 33Mbps apart . A Quantitative Trait Locus ( QTL ) is called for every linkage signal peak that is above some LOD threshold . To avoid noise and linkage over large distances translating into multiple peaks we count peaks that are less than 33Mbps apart as one QTL . We define the QTL confidence interval as the region between the furthest positions from the QTL peak that are above 1 LOD drop from the peak’s LOD score and that are no further than ±33Mbps from the peak’s position . To calculate the empirical P value and false discovery rate ( FDR ) for the total number of QTLs , we use permutation analysis . We repeat the QTL detection procedure for the 1000 combinations of height permuted families and calculate the distribution of the number of QTLs that are expected by chance anywhere in the genome . We then compare the total number of QTLs detected for the real families to the distribution of the number of QTLs detected in the permutation analysis . We calculate the significance ( P value ) of the real families’ QTLs as the fraction of permutations that yield an equal or larger number of QTLs than the real families , and the FDR as the ratio between the average QTL number per permutation ( the expected number of QTLs by chance ) and the number of QTLs from the real families ( the observed number of QTLs ) . We define the estimated total number of true QTLs as the difference between the observed and expected QTL numbers . The required detection threshold to call linkage as genome-wide significant with a low false positive rate of 5% ( LOD score = 3 . 6 for sibling pairs , or an equivalent P value = 2X10-5 ) was established over two decades ago [45] and has been the gold standard for family studies . Using this strict threshold is appropriate for Mendelian traits where one expects only a single locus to influence the trait and given the high costs of perusing and fine mapping QTLs to find the causal genetic variant . Height however is a complex trait and is most probably governed by multiple loci with smaller effects . We therefore increased the power to detect QTLs by investigating the space of lower detection thresholds . We start from the traditional LOD = 3 . 6 threshold and gradually lower the detection threshold at constant increments of 0 . 1 LOD , thus increasing the sensitivity and detecting more QTLs at lower thresholds . Individually , QTLs that are identified below LOD = 3 . 6 might not be significant , but as a group , the total number of detected QTLs can be significant when compared to the total number of QTLs detected in the permutations . The increased sensitivity to identify more QTLs comes at the expense of reduced specificity ( higher FDR ) , i . e . more QTLs are detected at lower detection thresholds , but a larger fraction of these QTLs are false ( and it is not possible to know which specific QTLs are the false ones ) . However , as long as true positives accumulate faster than false positives , the absolute number of true QTLs that are identified increases . See Fig 2 for the P values and FDR for the total number of detected QTLs for different detection thresholds . See S2 Table for a list of all detected QTLs . All genomic coordinates refer to genome build hg19 . To correct genotyping and IBD inference errors , and to facilitate imputation , we phase the genotypes of all participants into haplotypes . We first use the phased IBD structure that we reconstructed in a previous step . For each SNP we compare the SNP calls of the children that belong to the four different possible combinations of grandparental haplotypes . If the siblings in one of these groups are homozygote for this SNP , both haplotypes are assigned with the called allele . If they are heterozygous , we examine the SNP calls for the two other groups that share with the group in question exactly one allele IBD . In the cases where heterozygote calls cannot be phased in the above manner ( e . g . when all siblings are heterozygotes ) we use information from the parents ( if existing in the data ) . First , for each chromosome we compare the haplotypes in the already solved positions to the SNP calls of the parents in order to identify which of the parents is carrying haplotypes 1 & 2 and which carries haplotypes 3 & 4 along the entire chromosome . We then use homozygous SNPs of the parents to phase heterozygous SNPs of the children . Using only the IBD we were able to phase ~95% of the SNPs . Unphased SNPs included SNPs where not enough siblings or parents had good quality genotype calls , SNPs where no informative homozygote participant existed or SNPs where phasing by different siblings or parents contradicted . To phase the remaining SNPs we used the phased haplotypes of the entire study population as reference . For each unphased SNP within a specific family , we use all the previously solved haplotypes in other families that include this SNP . We preferably use the longest haplotype that matches on both sides of the unknown call , and when there are contradicting cases we use the call that matches the majority of the solved haplotypes . When there is no usable information ( <0 . 1% of cases ) the call is determined to be the B allele ( due to the Affymetrix definition of the A and B allele on their chip ) . To test the accuracy of our IBD reconstruction and haplotype phasing we compared the raw genotype calls as read from the Affymetrix Axiom Biobank array , to the genotype calls that can be inferred back from the calculated haplotypes . Since the reconstructed IBD is used for phasing the haplotypes , mismatches between the experimentally read genotypes and the inferred genotypes can be caused either by experimental errors in genotyping , or by analysis errors in the IBD reconstruction or haplotype phasing . We observed two types of mismatch patterns . First , blocks of continuous mismatches between the observed and inferred genotypes , that are immediately adjacent to inferred recombination positions . Such blocks most likely indicate errors in the IBD calling due to errors in identification of the exact recombination position . The rate of this type of error was low , as blocks of mismatches covered <1% of all chromosomes of all participants . We reduced it further by correcting manually IBD calls that created mismatch blocks larger than 1Mbps and then iterated the haplotype phasing pipeline and the above quality control . The second type of errors is sparsely distributed mismatches . Since our nuclear families are large and typically at least several siblings share at each genomic position the same two grandparental haplotypes from their parents , if our IBD reconstruction is correct , the inferred genotypes can be in theory more accurate than the experimentally measured ones since experimental errors in one sibling could be corrected by the other siblings that share the same two grandparental haplotypes , or by other families who share a haplotype in this genomic region . When experimentally low quality SNP calls ( Confidence value > 0 in the Affymetrix array reading ) were included , the frequency of SNP calls differing between the experimentally measured and analytically inferred for the 397 participants was 9 . 1x10-3 ± 2 . 5x10-3 ( mean ± S . D ) , practically identical to the experimental error rate measured as mismatches between technical repeats . This suggests that most SNPs with experimental genotyping errors due to low quality are indeed corrected by the multiple siblings and parents IBD information . When low quality genotype calls are excluded from the comparison , the mismatch rate drops ~8 fold to 1 . 2x10-3 ± 4 . 9x10-4 ( mean ± S . D ) , i . e . from 1 error in every ~110 SNPs to 1 in every ~866 SNPs . The total number of mismatches that we observe between experimentally genotypes and analytically inferred genotypes is on average ~726 errors per sample for the 628 , 456 SNPs on the array ( excluding the Y chromosome and mitochondria ) , of which ~280 errors are expected to be errors coming from experimentally determining genotypes ( based on the expectation of an experimental error of 1 in every ~2247 SNPs ) , and ~446 errors are expected to come from errors in the IBD reconstruction and haplotype phasing pipeline . Overall these quality control calculations suggest a low error rate of 446/628 , 456 = 7 . 1x10-4 for our IBD reconstruction and haplotype phasing pipeline ( 1 in every 1409 SNPs ) , while correcting most of the individual sample genotyping experimental errors and SNPs that were not called . This is important as the inferred genotyped and not the experimentally measured genotypes are used for the downstream analysis of variance partitioning and predictions of height . To assess how much phenotypic variance can be explained by the QTLs , we conducted variance partitioning in a cross validation framework . We divided randomly our data into 100 training and corresponding test sets . Each training set contained 19 families and the corresponding test set the other 10 families . We constrained the random choice of families so that each training set contained two thirds of all participants and each test set one third of all participants ( ±2 participants ) . We conducted linkage analysis as detailed above on each training set and identified QTLs . We used only the autosomal QTLs for the next steps . We took the coordinates of the QTLs into a corresponding test set , and from the genotypes of the test set participants took 100 SNPs closest to each QTL peak ( ~1Mbps per QTL ) and concatenated these SNPs together ( e . g . for 7 QTLs we took the genotype calls of 700 SNPs ) . We chose to use 100 SNPs since using less SNPs explained less variance , probably since there are not enough SNPs to capture the exact genetic relatedness between participants in these regions ( e . g . two participants can be identical by state ( IBS ) for several SNPs , while not identical by descent ( IBD ) , for example if they are both homozygote to a common variant with high frequency ) . Increasing the GRMs to include more than 100 SNPs per QTL did not result in explaining more variance , suggesting that the genetic variance relevant to the trait is captured in the 100 SNPs ( ~1Mbps ) around a QTL . We used the SNPs in the QTLs as input to the software GCTA [53] to create a Genomic Relatedness Matrix ( GRM ) , representing the genetic similarity between all participants in the test set for the specific coordinates of the QTLs ( hence the QTL GRM ) . To control for pedigree structure , the overall genetic similarity between participants and shared environments , we built a second GRM based on all the genotyped SNPs of the genome ( ~350 , 000 informative SNPs , hence whole genome GRM ) . To control for QTLs explaining variance through tagging of known common variants for height , we constructed a third GRM from 650 GWAS [3] known common variants that we could genotype or impute ( hence GWAS GRM ) . We fit a joint variance component model , where variance in height is the sum of the random effects from GRMs ( covariance structures ) built from QTLs , the whole genome , and GWAS snps , as described above using GCTA [53] , with the EM algorithm . Since the QTLs GRM might explain some phenotypic variance simply because is correlated to the whole genome GRM and the GWAS GRM , we repeated the analysis for each test set 100 times , but with a random QTLs GRM made from random regions in the genome that have equal number ( 100 ) of SNPs as in the real QTLs GRM . We constrained these randomly chosen regions so that they maintain a minimal distance between QTLs similar to the one between real QTLs . We calculated the estimated variance explained by QTLs for each threshold , as the median variance explained by the 100 test set with the real QTLs , minus the median of the 10 , 000 sets with the random QTLs . It is important to note that random QTLs can be on top or in linkage with real QTLs , therefore our final estimation of variance explained by the real QTLs ( after deduction of the variance explained by the random QTLs ) might be an underestimation , especially for the low LOD detection thresholds where many regions of the genome are picked as QTLs and the chances that random QTLs are overlapping real QTLs is higher . Of the 697 SNPs that were previously identified to be associated with height [3] , only 145 were on our genotyping array . In order to impute the other 552 SNPs , we used the IMPUTE2 software [61 , 62] , with two reference panels—the 1000 genomes ( 5008 phased haplotypes ) [63] and a reference panel of Ashkenazi Jews ( 256 phased haplotypes ) [35] . Despite the much smaller sample size of the Ashkenazi Jews panel , it yielded higher imputation accuracy as tested by the IMPUTE2 concordance tables ( internal cross validation that IMPUTE2 conducts by imputing SNPs that are genotyped by the array ) . For SNPs that were confidently imputed by IMPUTE2 ( maximum posterior probability ≥ 0 . 9 ) the concordance between the imputed and genotyped SNPs was on average 98% when using the Ashkenazi Jews reference panel , compared to 97% when using the 1000 Genomes reference panel . Using the Ashkenazi Jews panel , we could impute 513 of the 552 missing SNPs , and the last 39 SNPs could be imputed by the 1000 Genomes reference panel . | Rare variants with large effects have been suggested to account for some of the missing heritability of height , but detection of such variants in genome-wide association studies ( GWAS ) requires very large sample sizes due to their low frequencies . Here , we designed a unique study of height , in which we sought to increase the effective frequency of rare variants in our sample . This was done by assembling a cohort of very large nuclear families , with an average of 12 and a maximum of 20 siblings per nuclear family . In this design , any variant segregating in our cohort has a minimum expected frequency of ~1% , regardless of its frequency in the general population . In addition , we recruited all participants from a founder ( Jewish ) population , in which some variants that are rare in a cosmopolitan population can rise to high frequencies . To further increase power , we developed methods to obtain highly accurate height measurements , genotype calls , and inheritance pattern reconstructions . These factors allowed us to detect many more QTLs than previous family-based studies of height , despite a modest sample size of only 397 participants . The approach described in this paper provides insights into the genetic basis of height and the roles of population-specific vs . cosmopolitan variants , and may serve as a complement to GWAS for genetic investigations of other complex traits . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genome-wide",
"association",
"studies",
"jewish",
"people",
"quantitative",
"trait",
"loci",
"variant",
"genotypes",
"alleles",
"genetic",
"mapping",
"ethnicities",
"mathematics",
"genome",
"analysis",
"molecular",
"genetics",
"discrete",
"mathematics",
"combinatorics",
... | 2019 | Analysis of the genetic basis of height in large Jewish nuclear families |
Expression of genes of the locus of enterocyte effacement ( LEE ) is essential for adherence of enterohemorrhagic Escherichia coli ( EHEC ) to intestinal epithelial cells . Gut factors that may modulate LEE gene expression may therefore influence the outcome of the infection . Because nitric oxide ( NO ) is a critical effector of the intestinal immune response that may induce transcriptional regulation in enterobacteria , we investigated its influence on LEE expression in EHEC O157:H7 . We demonstrate that NO inhibits the expression of genes belonging to LEE1 , LEE4 , and LEE5 operons , and that the NO sensor nitrite-sensitive repressor ( NsrR ) is a positive regulator of these operons by interacting directly with the RNA polymerase complex . In the presence of NO , NsrR detaches from the LEE1/4/5 promoter regions and does not activate transcription . In parallel , two regulators of the acid resistance pathway , GadE and GadX , are induced by NO through an indirect NsrR-dependent mechanism . In this context , we show that the NO-dependent LEE1 down-regulation is due to absence of NsrR-mediated activation and to the repressor effect of GadX . Moreover , the inhibition of expression of LEE4 and LEE5 by NO is due to loss of NsrR-mediated activation , to LEE1 down-regulation and to GadE up-regulation . Lastly , we establish that chemical or cellular sources of NO inhibit the adherence of EHEC to human intestinal epithelial cells . These results highlight the critical effect of NsrR in the regulation of the LEE pathogenicity island and the potential role of NO in the limitation of colonization by EHEC .
Enterohemorrhagic Escherichia coli ( EHEC ) , especially those belonging to the O157:H7 serotype , are foodborne pathogens and healthy rearing animals are the main reservoir . Human infection occurs through the ingestion of contaminated food . This primary infection yields to the development of intestinal disorders , including aqueous or bloody diarrhea . Moreover , EHEC express a cardinal and well-defined virulence factor , the Shiga-toxin ( Stx ) encoded by genes located in lysogenic lambdoid bacteriophages . Stx is produced in the gut lumen and crosses the epithelial barrier to reach the blood and the target organs including the kidneys . In this context , infected patients may develop life-threatening complications such as the hemolytic and uremic syndrome ( HUS ) , the main cause of renal failure in children in developed countries [1] . EHEC genes carried by the locus of enterocyte effacement ( LEE ) , a chromosomal pathogenicity island organized in 5 operons , encode bacterial factors implicated in the intimate adherence of these bacteria to intestinal epithelial cells [2] . These genes encode a type 3 secretion system ( T3SS; LEE1 , LEE2 , LEE3 ) , a translocon and a syringe ( LEE4 ) that allows bacteria to inject effectors in epithelial cells , such as the LEE5-encoded intimin receptor Tir; moreover , other proteins not carried by the LEE can be translocated by the T3SS into enterocytes [3] , [4] . The injected effectors and/or protein of the translocon itself interact with the host signal transduction , leading to actin polymerization and to microvilli effacement [2] , to regulation of the innate immune response [5] , [6] , and to increased electrolyte transport [7] . Regulation of gene expression within the LEE is known to be complex and governed by a large number of influences , including environmental cues or quorum sensing , and involves several specific or global regulators [8] , [9] . The first gene of the LEE1 operon , ler , encodes a transcriptional regulator that positively regulates the expression of all the other operons [9]–[11] . However a variety of extra-transcriptional mechanisms have also been involved in the regulation of LEE expression , though little detailed mechanistic information is available [12] . GadE ( YhiE ) and GadX ( YhiX ) are two main regulators of the acid fitness island involved in acid-resistance ( AR ) in E . coli K12 [13]–[15] . At acidic pH values , GadE and GadX positively regulate the gadA and gadBC genes , encoding the components of the glutamate-dependent AR . In E . coli O157:H7 , GadE has acquired additional functions and inversely coordinates expression of AR and LEE genes [16]: It has been proposed that , during passage through the human stomach , GadE protects EHEC by inducing the glutamate-dependent AR system and inhibits the unnecessary expression of the LEE genes , while environmental cues in the intestine lead to downregulation of gadE and upregulation of the LEE genes [16] . GadE has been shown to directly bind the ler ( LEE1 ) and sepZ ( LEE2 ) promoters in vitro [8] , but in vivo binding of GadE and the role of GadX have never been investigated . We have previously shown that nitric oxide ( NO ) decreases Stx2 synthesis by EHEC O157:H7 at the transcriptional level [17] . This occurs through the inhibition of the SOS response by the NO sensor nitrite-sensitive repressor ( NsrR ) [17] , the key regulator of the nitrosative stress in enterobacteria [18] . In this context , our aim was to investigate whether NO also modulates LEE gene transcription and therefore EHEC adhesion to epithelial cells . Here we show that NsrR is a direct positive regulator of the transcription of LEE1 , LEE4 and LEE5 genes and an indirect repressor of gadE and gadX genes . In the presence of NO , LEE1/4/5 activation is abrogated , GadE is induced and yields to gadX expression . Finally , we identify GadE and GadX as repressors of LEE4/5 and of LEE1 , respectively . Using a human intestinal epithelial cells/EHEC co-culture model we demonstrate that bacterial adhesion is inhibited in NO producing cells .
We first examined adhesion of the E . coli O157:H7 strain EDL933 to cultured Hct-8 intestinal epithelial cells in the presence of the NO donor NOR-4 . Exposure to NOR-4 at 200 µM or 500 µM did not cause any significant difference in the growth rate of EDL933 , as described [17] . However , EHEC adhesion to Hct-8 cells was dramatically inhibited when NOR-4 was added to the co-cultures ( Figs . 1A and 1B ) . The number of EHEC fixed to the cells was significantly decreased by 41±5% and 89±2% in the presence of 200 µM and 500 µM NOR-4 , respectively ( Fig . 1B ) . To further confirm this result , we analyzed the effect of endogenous NO released by enterocytes . Hct-8 cells were first treated for 24 h with a cytokine cocktail known to stimulate the inducible NO synthase ( iNOS ) expression [19] , washed , and then infected with the strain EDL933 in the presence or absence of the iNOS inhibitor N6- ( 1-iminoethyl ) -l-lysine ( l-NIL ) . There was less EHEC fixed to NO-producing epithelial cells than to control cells ( Figs . 1A and 1C ) . The inhibition of EHEC adherence to Hct-8 cells treated with cytokines was abolished by the use of l-NIL ( Figs . 1A and 1C ) . The expression of genes that represent the five operons of the LEE ( Fig . 2A ) was analyzed after treatment with NOR-4 for 6 h . NO was consistently generated in the bacteria culture medium and reached a plateau after 6 h ( Fig . S1 ) . The expression of ler ( LEE1 ) , espA ( LEE4 ) , tir and eae ( LEE5 ) was down-regulated by NO , while the transcription of sepZ ( LEE2 ) was induced by 2 . 4-fold ( Fig . 2B ) . The expression of the gene escV ( LEE3 ) was not modulated by NOR-4 ( Fig . 2B ) . Because GadE and GadX modulates LEE expression in EHEC and EPEC , respectively , [16] , [20] , we investigated the effect of NOR-4 on gadE and gadX transcription . As shown in Figure 2C , the expression of gadE and gadX was significantly induced by 2 . 4- and 2 . 7-fold in bacteria exposed to NOR-4 , respectively . Thereby , these data prompted us to wonder whether NO-dependent down-regulation of LEE1 , LEE4 and LEE5 requires GadE and/or GadX . Since the role of GadX and GadE on LEE expression is not well defined and is strongly dependent on the growth conditions [16] , [20] , [21] , we first analyzed the expression of ler , espA , and tir in EDL933 ΔgadE and ΔgadX mutants . When compared to the EDL933 strain , the mRNA levels of ler , espA and tir were increased by ∼1 . 4- , 2 . 3- , and 2-fold in the ΔgadE strain , respectively ( Fig . 3A ) ; these effects were reversed when the gadE mutant was trans-complemented with the gadE gene in a low copy number plasmid vector ( Fig . 3A ) . The gadX mutation was associated with a spontaneous increase of ler transcription and with a significant reduction of espA and tir gene expression ( Fig . 3A ) . The transcription of ler was repressed while the expression of espA was activated and that of tir was restored to the same level as the WT in the trans-complemented strain ( EDL933 ΔgadX-c; Fig . 3A ) . These data suggest that GadE represses the expression of LEE4 and LEE5 genes independently of Ler , and that GadX represses LEE1 but activates LEE4 and LEE5 gene expression . Interestingly , the NOR-4-dependent down-regulation of ler , espA , and tir was still observed in the ΔgadE , ΔgadX and ΔgadE/gadX mutants ( Fig . 3A ) , suggesting that another factor is implicated in the inhibition of LEE1/4/5 by NO . We next wonder whether GadE and GadX repressed the LEE independently from each other or whether GadX is epistatic to GadE as in E . coli K12 [22] . The expression of ler was similar in a ΔgadE/gadX double mutant and in the EDL933 ΔgadX strain ( Fig . 3A ) , indicating that GadX is epistatic to GadE in controlling LEE1 . Conversely , espA and tir mRNA levels were increased in EDL933 ΔgadE/gadX when compared to the WT strain , as in the ΔgadE strain , demonstrating that GadE is epistatic to GadX for the regulation of LEE4 and LEE5 . Therefore we investigated whether GadE controls gadX expression . Figure 3B shows a 33% decrease in gadX mRNA levels in the gadE mutant , indicating that GadE activates gadX expression . In addition , we observed 3 . 1-fold more gadE mRNA copies in the ΔgadX strain than in the WT strain ( Fig . 3B ) and gadE mRNA levels were dramatically reduced in the complemented strain ( Fig . 3E ) , suggesting that GadX is a repressor of gadE expression . Therefore , the moderate increase in ler expression observed in the ΔgadE strain ( Fig . 3A ) is likely due to the lower level of GadX in this strain and not to a direct effect of GadE on ler transcription . Lastly , the activation of gadX transcription by NOR-4 was suppressed in the gadE mutant , but not in the EDL933 ΔgadE-c strain ( Fig . 3B ) , while the NO-dependent induction of gadE mRNA expression was still observed in the ΔgadX strain ( Fig . 3B ) . These data indicate that NO activates gadX expression through GadE . NsrR is a transcriptional regulator that regulates gene expression in response to NO [18] . Therefore we investigated whether NsrR regulates gadE , gadX , and the LEE genes . In the absence of NO , the mRNA levels of ler , espA , and tir were 6 . 8 , 7 . 1 , and 14 . 3-fold lower in the ΔnsrR mutant than in the WT strain , respectively ( Fig . 4A ) . The expression of these genes was similar in the strains EDL933 and EDL933 ΔnsrR-c ( Fig . 4A ) . Moreover , the NO-dependent regulation of these LEE genes was abrogated in EDL933 ΔnsrR and was restored in the complemented strain ( Fig . 4A ) . Inversely , the transcription of gadE and gadX was significantly increased in the ΔnsrR mutant , but not in the complemented strain . The expression of these two genes was not affected by NOR-4 in the nsrR-deficient strain ( Fig . 4B ) . These data suggest that NsrR is a transcriptional activator of LEE1 , LEE4 , and LEE5 and a repressor of gadE , which in turn modulates gadX expression . NsrR loses its ability to regulate the expression of LEE and gad genes in the presence of NO . The investigation of GadE , GadX and/or NsrR direct binding to the gadE , gadX , and LEE promoter regions was performed by chromatin immunoprecipitation ( ChIP ) experiments using the EDL933 ΔgadE , ΔgadX , and ΔnsrR mutants expressing the 6-His-GadE , the 6-His-GadX , and the 6-His-NsrR fusion proteins , respectively . We first analyzed the gadX promoter described by Hommais et al . [15] , the three promoters described for gadE in E . coli K12 [22] ( Fig . S2 ) , and the gadA promoter as a positive control for GadE and GadX binding [23] . Surprisingly , we found that GadE and GadX did not bind to the gadX and gadE promoters , respectively ( Figs . 5A and 5B ) , indicating that activation of gadX by GadE and repression of gadE by GadX occur through indirect regulations . As expected , the binding of GadX and GadE to the gadA promoter region was observed ( Figs . 5A and 5B ) . Two ler promoters have been described in EHEC , the distal P1 promoter and a putative proximal P2 promoter ( Fig . S2 ) . The P1 promoter is common to EHEC and EPEC , while the P2 promoter is present only in EHEC [24]–[26] . Neither GadE ( Fig . 5A ) nor GadX ( Fig . 5B ) bound to either of these promoters ( Figs . 5A and 5B ) . These data indicate that GadE and GadX do not repress ler expression directly . The LEE4 promoter has been identified in EHEC upstream of sepL [27] , espA being the second gene of the operon . In EPEC , it has been shown that Ler-mediated activation of the LEE5 operon requires sequences between positions -198 and -75 relative to the transcriptional start site [28] . Two primer pairs overlapping this region have been designed for ChIP experiments , amplifying a LEE5 distal ( P1LEE5 ) and a LEE5 proximal ( P2LEE5 ) region ( Fig . S2 ) . ChIP experiments showed that neither GadE nor GadX bound to the LEE4 and LEE5 promoters ( Figs . 5A and 5B ) . Lastly , the binding of GadE and GadX to the LEE1/4/5 promoter regions was not modulated by NOR-4 . These data indicate that control of LEE4 and LEE5 expression by GadE and GadX is due to indirect effects . In contrast , NsrR bound to the distal LEE1 promoter ( P1LEE1 ) , to the LEE4 and LEE5 promoters , and to the promoter of hmpA , a well-know NsrR target gene ( Fig . 5C ) . Furthermore , NsrR binding to these promoter regions was inhibited when the bacteria were grown in the presence of NOR-4 ( Fig . 5C ) . We did not observed NsrR binding to the gadE and gadX promoters ( Fig . 5C ) . We thus performed bio-informatics analysis to identify putative NsrR-binding sites in the LEE1 , LEE4 LEE5 , gadE , and gadX promoters in the strain EDL933 . We used the homologous sequences of seven NsrR-binding sites described in E . coli K12 [29] to generate the sequence logo of the NsrR box in the strain EDL933 ( Fig . 5D ) . We then performed bioinformatics analysis on the LEE1 , LEE4 and LEE5 promoter sequences by the Gibbs Sampler Motif Software , using the matrix of the seven putative NsrR-binding sites of EDL933 . In agreement with the ChIP data , bioinformatics analysis identified sequences presenting high identity with the NsrR consensus binding site in the LEE1 ( P1 ) , LEE4 and LEE5 promoter regions ( Fig . 5D and Fig . S2 ) , but not in the promoters of gadE and gadX . The analysis indicated a 23 bp putative NsrR-binding site in the promoters of LEE1 ( 86 . 9% identity ) and LEE4 ( 78 . 2% identity ) , but only a second half-site NsrR-binding site in the LEE5 promoter ( 90 . 9% identity for the half site; Fig . 5D ) . In silico analyses performed using the BLAST program indicated that these putative binding sites are conserved in a number of EHEC and EPEC strains , but not in Citrobacter rodentium ( Fig . S3 ) , an attaching/effacing pathogen that infects rodents . Since NsrR has been exclusively described as a transcriptional repressor , we investigated the molecular mechanism underlying the direct activation of LEE gene expression by NsrR . For many transcriptional activators , increase of the transcription level results from the recruitment of RNA polymerase through direct interaction between the regulatory protein and one or several subunits of the polymerase [30] . We therefore examined if NsrR can interact with α and σ RNA polymerase subunits . To this end , His-tagged NsrR and hemagglutinin ( HA ) -tagged polymerase subunits α ( RpoA ) or σ38 ( RpoS ) were co-expressed in bacteria . His-NsrR was purified under native conditions using a nickel affinity resin and the different fractions were analyzed by western-blot . As positive controls , RpoA and RpoS were also co-expressed with His-Crp or His-Crl , respectively , two well-known interacting partners [31] , [32] . All tagged proteins were properly expressed as revealed by their immunodetection in the whole extract samples ( Fig . 6 ) . As expected , HA-RpoA and HA-RpoS co-eluted with His-Crp or His-Crl , respectively . No HA-tagged protein was detected in the His eluates of the negative controls , i . e . , bacteria expressing only HA-tagged proteins ( Fig . 6 ) . Importantly , HA-RpoA and HA-RpoS were also specifically recovered in the eluted fractions from the His-NsrR purifications . This finding demonstrates that NsrR can interact with the RNA polymerase complex and suggests that NsrR activates LEE gene expression through the recruitment of RNA polymerase . In order to confirm the role of NO , NsrR , GadE , and GadX in regulating LEE expression , we investigated the attachment of the regulatory mutants to HeLa cells after 6 h of infection in the presence or absence of NOR-4 . As expected , EDL933 adhered to HeLa cells and when NOR-4 was added to the co-culture the level of adhesion was dramatically reduced to that of the ΔescN mutant that lacks a functional T3SS ( Figs . 7A and 7B ) . The adhesion of the ΔgadE and ΔgadX strains was higher than that of the parent strain , correlating with the repressive effect of AR regulatory proteins on LEE gene expression ( Figs . 7A and 7B ) . Conversely , the nsrR mutant was less adherent than the WT strain ( Figs . 7A and 7B ) . The complementation of these three mutants restored the adhesion phenotype of the parental strain ( Figs . 7A and 7B ) . Under NO exposure , adherence properties were affected for the ΔgadE and ΔgadX mutants but not for the ΔnsrR mutant ( Figs . 7A and 7B ) , demonstrating that NsrR is the key regulator controlling the T3SS-dependent adhesion of EHEC in response to NO .
In the present report , we show that NO , a critical mediator of the host innate immune response , is a potent inhibitor of LEE gene expression in EHEC O157:H7 and consequently inhibits the adhesion of these pathogens to intestinal epithelial cells . We identified NsrR as an unrecognized regulator that controls the expression of LEE genes in response to NO , and we propose a regulatory model presenting the role of NsrR , GadE and GadX in LEE expression ( Fig . 8 ) . In the absence of NO ( Fig . 8A ) , NsrR directly activates LEE1 , LEE4 , and LEE5 gene expression , and indirectly represses gadE and therefore gadX expression . We also show that GadE indirectly activates gadX expression and represses LEE4 and LEE5 expression independently of Ler , while GadX inhibits gadE and LEE1 expression . When NsrR binds NO ( Fig . 8B ) , it is released from its target DNA , leading to loss of induction of LEE1/4/5 genes and to the up-regulation of gadE and , consequently , gadX . In this context , the NO-dependent LEE1 down-regulation is due to absence of NsrR-mediated activation and to the inhibitory effect of GadX . In parallel , the inhibition of LEE4 and LEE5 gene expression is due to absence of NsrR- and Ler-dependent activation and to increase of GadE level . This model assumes that repression of gadX expression by NsrR is mediated by GadE , which is consistent with the observation that the NO-dependent activation of gadX is abrogated in the ΔgadE and ΔnsrR mutants . NsrR is a key negative regulator of the nitrosative stress in enterobacteria [18] , [33] . NsrR has always been described as a transcriptional repressor . In addition , its DNA-binding activity is suppressed in the presence of NO , yielding to the expression of various genes mainly involved in NO detoxification [18] , [33] . In non-pathogenic E . coli , NsrR also regulates expression of genes involved in metabolism , motility , protein degradation , surface attachment , stress response and transmembrane transport [29] , [34] . Our data indicate that NsrR is also a repressor of the genes gadE and gadX . Nonetheless , the NsrR-dependent repression of gadX is probably mediated by GadE since the NO-dependent up-regulation of gadX is abrogated in the ΔgadE mutant . We did not find a sequence matching the NsrR consensus binding site in the gadE promoter , and ChIP experiments failed to demonstrate physical interaction between NsrR and the gadE promoter . Therefore , the effect of NsrR on gadE transcription is probably indirect and mediated by an unknown regulatory cascade controlled by NsrR . Here we provide compelling evidence that NsrR is a direct positive regulator of LEE1 , LEE4 , and LEE5 operons in EHEC by binding to their own promoters . Moreover , our data also suggest that NsrR acts as a transcriptional activator by recruiting RNA polymerase to promoter regions since NsrR is able to pull-down the α and σ38 subunits of the RNA polymerase . Supporting the concept that it may also be a transcriptional activator , it has been reported that NsrR activates virulence gene expression in Salmonella Typhimurium , in particular expression of genes important for eukaryotic cell adherence , invasion and intestinal translocation , and that an nsrR mutant is impaired in invasion of HeLa cells [35] . However , in silico analysis failed to identify an NsrR consensus binding site in the promoter regions of these genes , indicating that the positive regulatory effect of NsrR is probably indirect in this pathogen [35] . Moreover , using an E . coli K12 strain harboring a multicopy plasmid that titrates out NsrR , Filenko et al . have identified by a microarray analysis 22 transcripts that could be directly or indirectly activated by NsrR [34] . The NsrR binding site is a 23 bp palindrome sequence composed of two 11 bp half sites separated by a single nucleotide , and NsrR binds to DNA as a dimer [36] . However , a number of NsrR target promoters contain only a single half site [29] . Potential NsrR consensus sequence were identified in the LEE1 , LEE4 and LEE5 promoters , with a 23 pb putative NsrR-binding site in the LEE1 and LEE4 promoters , and a putative second half-site in the LEE5 promoter . It has been suggested that , when the NsrR binding site contains only a single half site , one NsrR monomer makes specific contact to the consensus half site and the other monomer forms nonspecific contact [37] . Alternatively , it has been suggested that NsrR binds as a tetramer to the complete binding motif and as a dimer when only one half site is conserved [29] . It is noteworthy that the putative NsrR binding sites identified in the LEE1 , LEE4 and LEE5 promoters are conserved in a number of other EHEC and EPEC strains , but not in C . rodentium , suggesting that NO also influences cell adhesion via NsrR in other E . coli attaching/effacing pathogenic human strains . Influence of GadE on LEE gene expression remains controversial . While Tatsuno et al . described an increased expression of LEE2 , LEE4 , and LEE5 in a ΔgadE mutant , which is not correlated with enhancement of ler expression [20] , KailasanVanaja et al . showed that GadE represses LEE expression by down-regulating ler transcription [16] . These discrepancies are proposed to be due to differences in growth medium and/or differences in the sensitivity of the assays used in each study . Interestingly , our data indicate that GadE may repress the expression of LEE4 and LEE5 via two regulatory cascades , mediated or not by Ler ( Figure 8 ) . On the one hand , we show that GadE inhibits LEE1 through GadX , because a decreased expression of gadX and an induction of LEE1 are observed in the gadE-deficient strain; this results in loss of Ler-dependent induction of LEE4/5 . On the other hand , the deletion of gadX is associated with an increased expression of ler and gadE , and with an inhibition of LEE4/5 , suggesting that GadE inhibits these operons independently of Ler . In accordance , the induction of espA and tir in the gadE mutant and in the ΔgadE/gadX strain demonstrates that GadX regulates LEE4/5 via the repression of gadE . However , although it has been shown in vitro that GadE can bind to the ler promoter in EHEC O157:H7 [8] , we did not observe such an interaction in vivo in our experiments; this difference is probably due to the presence of binding competitors in live bacteria . Regarding GadX , we show herein that it negatively regulates ler transcription in EHEC . However , the effect of GadX on LEE1 expression is indirect since no physical interaction between GadX and the LEE1 promoter has been demonstrated . Interestingly , it has been described in EPEC that LEE1 is down-regulated under conditions in which GadX is induced , namely at pH 5 . 5 or in contact to epithelial cells [21]; this occurs through the inhibition of the transcription of the per locus by GadX [21] . Because the perC homologue in EHEC , named pch , is involved in LEE1 induction [38] , it would be interesting to now determine the role of GadX on pch expression . The biological relevance of LEE1 , LEE4 , and LEE5 inhibition by NO is the decreased adhesion of E . coli O157:H7 to epithelial cells . When EHEC are ingested with the contaminated food , they first reach the stomach . It has been proposed that the acidic conditions of this ecological niche favor GadE induction and therefore limit EHEC adhesion to gastric tissues [16] . There is also abundant nonenzymatically formed NO in the gastric juice caused by acidification of nitrate and nitrite . In this context , we now propose that the NO-dependent LEE4/5 inhibition is a supplementary mechanism developed by EHEC to avoid their persistence in the stomach and to favor bacterial colonization in the colon . Moreover , we have shown in the present study that , not only a chemical source of NO , but also the reactive nitrogen species released by iNOS-expressing colonic epithelial cells inhibit the adherence of O157:H7 E . coli , and our previous work has identified NO as a potent inhibitor of Stx synthesis [17] . Together , these results suggest that NO might limit the infectious process and HUS development . Nonetheless , it has been described that EHEC inhibit the inducible transcription of iNOS in human enterocytes [19] , thus , by limiting NO production , EHEC might favor their own virulence by increasing the intimate adherence to the intestinal epithelium and Stx synthesis . We can therefore speculate that the issue of the crosstalk between EHEC and the host-derived NO might determine the outcome of the infection .
Strains and plasmids used in this study are listed in Table S1 . The EHEC O157:H7 strain EDL933 [39] was used throughout the study . The EDL933 ΔgadE and ΔgadX mutants and the ΔgadE/gadX double mutant were constructed using the one-step PCR-based method [40] , [41] . Mutants were verified by PCR to assess the loss of the gene and by RT-qPCR to confirm lack of expression of the gene of interest , using the primers listed in Table S2 . The ΔnsrR mutant strain has been previously described [17] . For complementation analysis and ChIP experiments , the gadE , gadX , and nsrR genes were amplified with the high fidelity polymerase Pfx50 ( Invitrogen ) and cloned under the control of the araC promoter into a low-copy plasmid containing a 6-histidine tag ( pBADHisA or pBADMycHisA; Invitrogen ) , or in pBAD33 . The cloned genes were checked by nucleotide sequencing , and their expression was analyzed by RT-qPCR . The 6-His-NsrR- , 6-His-GadE- , and 6-His-GadX-encoding genes were expressed at the same level than the WT genes . To verify the mutation of the gadE and gadX genes , we analyzed the acid resistance of the mutant strains [42]: Acid-resistance of the ΔgadE and ΔgadX mutants dropped to 0 and 1 . 41% of the parent strain , respectively; acid resistance was restored in the complemented mutant strains ( data not shown ) . A single colony of EDL933 or isogenic mutants was grown overnight in DMEM Low glucose containing 10 mM HEPES . These cultures were diluted in fresh medium to an OD600 = 0 . 03 and grown at 37°C . The medium was supplemented with ampicillin ( 50 µg/ml ) , kanamycin ( 50 µg/ml ) , chloramphenicol ( 25 µg/ml ) , L-arabinose ( 0 . 1 mM–0 . 5 mM ) , or the NO donor NOR-4 ( Enzo Life Science ) when required . The NsrR-binding sequence logo of the strain EDL933 was generated using homologous sequence of the seven NsrR-binding sites described previously by Partridge et al . in E . coli K-12 strain MG1655 [29] and the software Weblogo ( http://weblogo . berkeley . edu/logo . cgi ) . The probabilities of occurrence matrix from the seven homologous sequences in EHEC O157:H7 strain EDL933 served as a model for the identification of a consensus sequence in the promoter regions of the target genes using the online software Gibbs Motif Sampler ( http://ccmbweb . ccv . brown . edu/gibbs/gibbs . html ) . The sequence alignment of the LEE1 , LEE4 and LEE5 putative sites in other EHEC strains , in EPEC strains , and in C . rodentium was performed with the MEGA5 software . The pBADMycHisA::gadE , pBADHisA::gadX , and pBADMycHisA::nsrR plasmids , encoding 6His-GadE , 6His-GadX and 6His-NsrR , were electroporated into the respective mutants to avoid native protein interference . Overnight cultures of each strain in LB medium were diluted 1∶100 in 25 ml of fresh DMEM medium buffered with 10 mM HEPES , with or without NOR-4 . GadE and GadX expression was induced with 0 . 5 mM l-arabinose and NsrR with 0 . 1 mM l-arabinose . After 6 h of growth with shaking , ChIP was performed as described by Lannois et al . [43] with slight modifications . First , the protein-DNA complexes were cross-linked by treating bacteria with 1% formaldehyde at room temperature for 30 min . Bacteria were then washed twice with cold PBS and incubated for 30 min at 37°C in 0 . 7 ml of lysis buffer ( 10 mM Tris pH 8 , 50 mM NaCl , 10 mM EDTA , and 20% sucrose ) containing 10 mg/ml lysozyme ( Sigma ) . Then , 0 . 7 ml of 2X IP buffer ( 100 mM Tris pH 8 , 300 mM NaCl , 2% Igepal CA-630 , 0 . 5% Na deoxycholate ) containing 1 mM PMSF was added and samples were incubated 15 min at 37°C , cooled down on ice , sonicated , and incubated on ice for 1 min . Sonication was repeated 11 times to obtain a solution of fragmented chromatin . A 50 µl aliquot of each sample was treated with 100 µl TE containing 36 µg proteinase K for 2 hours at 37°C , incubated 8 hours at 67°C to reverse crosslinking , and the DNA was purified with the kit Qiaquick ( Qiagen ) ; this was termed as Input fraction . The rest of the fragmented chromatin was used to generate the IP fraction . After a 2 h-incubation with an anti-Histidine monoclonal antibody ( Sigma ) , protein G sepharose 50% ( 40 µl ) was added to each sample and incubated 1 hour at room temperature . The beads were washed twice with IP buffer , twice with 1 ml of ChIP wash buffer ( 10 mM Tris HCl pH 8 , 250 mM LiCl , 1 mM EDTA , 0 . 5% Igepal CA-630 , and 0 . 5% Na deoxycholate ) and twice with 1 ml of TE buffer . The beads were resuspended in 100 µl of elution buffer ( 50 mM Tris HCl pH 8 , 10 mM EDTA , 1% SDS ) , incubated 15 min at 65°C , and centrifuged at 9500× g for 1 min . The supernatants containing the immunoprecipitated DNA were collected and incubated with 100 µl TE containing 36 µg proteinase K 2 hours at 37°C and 8 hours at 65°C . DNA was purified with the Qiaquick kit ( Qiagen ) and amplified by qPCR using the primers listed in Table S2 and depicted in Fig . S2 . The enrichment of DNA targets was calculated as follows for each protein: the promoters of interest as well as a non-specific rpoA intragenic region were amplified with specific primers ( Table S2 ) . For each DNA target , we calculated the ratio between the copy number in the IP fraction and the Input fraction; each value was then divided with the ratio obtained for the non-specific rpoA intragenic region . Then the same ratio was calculated from the parent strain EDL933 containing the empty pBADMycHisA vector . Values higher than 20 , corresponding to twice the values obtained for the strain EDL933 containing the empty pBADmycHisA vector , indicate protein binding to the promoter of interest . For bacterial co-expression experiments , genes encoding NsrR , Crp or Crl were cloned into the first multiple cloning site of pCDFDuet-1 vector ( Novagen ) allowing expression of the proteins tagged with a N-terminal hexahistidine motif . Genes encoding RpoA or RpoS were cloned into the second multiple cloning site using PCR primers allowing the insertion of a N-terminal HA motif ( see Table S2 for primers ) . E . coli BL21 ( DE3 ) harboring the different constructs was grown at 37°C to OD600 nm of 0 . 7 , then induced with 1 mM IPTG and grown for an additional 2 h . After resuspension of bacteria with a 1/10e volume of lysis buffer ( 50 mM NaH2PO4 , 300 mM NaCl ) , samples were sonicated and centrifuged . Supernatants ( whole bacterial extracts ) were incubated with Ni-NTA beads at 4°C for 16 h . Beads were washed four times with lysis buffer containing 60 mM imidazole and bound proteins were eluted with lysis buffer containing 250 mM imidazole . Total RNA from bacteria was extracted using the TRI Reagent RNA Isolation Reagent ( Sigma ) . Each RNA sample ( 1 µg ) was reverse transcribed with Superscript II enzyme ( Invitrogen ) and random primers ( Invitrogen ) . The cDNAs and serial dilutions of EDL933 genomic DNA , which were used for the standard curves , were amplified with gene-specific primers ( Table S2 ) in the Eppendorf Mastercycler eprealplex ( Eppendorf ) apparatus . The results are presented as the ratios between the copy number of mRNA of the gene of interest and the copy number of rpoA mRNA . Samples were mixed with a 2X SDS-PAGE sample buffer , heated for 5 min at 100°C , resolved on 14% SDS-PAGE gels and blotted on PVDF membranes . Membranes were blocked in PBS-0 . 05% Tween 20 supplemented with 5% non-fat dry milk , then probed with murine monoclonal anti-HA or HRP-conjugated anti-HIS Abs ( Sigma; 1/4000 for each ) . An HRP-conjugated goat anti-murine IgG Ab ( Sigma ) was also used for the HA blots . Acquisitions were performed with a G:box system ( Syngene ) . The epithelial cell lines Hct-8 and HeLa were maintained in DMEM with 10% FCS , 10 mM Hepes , 100 U/ml penicillin , 100 µg/ml streptomycin at 37°C under 5% CO2 . Hct-8 cells were plated on LabTek ( Nunc ) , cultured for 7 days , and stimulated for 24 h with human IFN-γ ( 50 ng/ml ) , TNF-α ( 20 ng/ml ) , and IL-1β ( 5 ng/ml ) . HeLa cells were seeded into LabTek and grown for 24 h . These Hct-8 and HeLa cells were washed , and infected with bacteria with an MOI of 100 , in the presence or absence of NOR-4 or of the iNOS inhibitor l-NIL . After 4 washes with PBS , cells were fixed using 1 ml methanol for 15 min at −20°C and stained with Giemsa or May-Grünwald Giemsa for 30 min . The number of adherent bacteria per cell was counted using the AxioVision 4 software . The concentration of the stable oxidized products of NO , NO3− and NO2− , was measured using the Nitrite/Nitrate Assay Kit ( Cayman Chemical ) . All the data represent the mean ± SEM . Student's t test or ANOVA with the Newman-Keuls test were used to determine significant differences between two groups or to analyze significant differences among multiple test groups , respectively . | Enterohemorrhagic Escherichia coli ( EHEC ) O157:H7 are food-borne pathogens for humans causing bloody diarrhea and , especially in children under five years old , kidney damages leading to death in 5% of cases . Antibiotics are contra-indicated because they are suspected to increase the severity of the disease . Therefore , it is crucial to develop alternative preventive or therapeutic strategies to fight EHEC infection . To reach this goal , a deeper knowledge of host-pathogen interaction is required . A critical step in EHEC infection is the adhesion of bacterial cells to intestinal epithelial cells . In response to the bacterial infection , the host triggers an immune response directed against the pathogen . The current study shows that a main effector of this immune response , nitric oxide ( NO ) , dramatically reduces the capacity of EHEC to adhere to intestinal epithelial cells . We have investigated the molecular mechanisms involved and identified a NO-sensor regulator that controls the expression of the genes required for EHEC adhesion . This finding underlines that NO could be a potential protective factor limiting the development of EHEC-induced diseases and provides a new avenue of investigation for the development of therapeutic strategies against infections with O157:H7 bacteria . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"escherichia",
"coli",
"biochemistry",
"neurochemistry",
"immunity",
"neurochemicals",
"nitric",
"oxide",
"innate",
"immunity",
"microbial",
"pathogens",
"immunology",
"host-pathogen",
"interaction",
"biology",
"microbiology",
"pathogenesis",
"bacterial",
"pathogens"
] | 2014 | NsrR, GadE, and GadX Interplay in Repressing Expression of the Escherichia coli O157:H7 LEE Pathogenicity Island in Response to Nitric Oxide |
Methicillin-resistant Staphylococcus aureus ( MRSA ) is a leading cause of morbidity and death . Phenol-soluble modulins ( PSMs ) are recently-discovered toxins with a key impact on the development of Staphylococcus aureus infections . Allelic variants of PSMs and their potential impact on pathogen success during infection have not yet been described . Here we show that the clonal complex ( CC ) 30 lineage , a major cause of hospital-associated sepsis and hematogenous complications , expresses an allelic variant of the PSMα3 peptide . We found that this variant , PSMα3N22Y , is characteristic of CC30 strains and has significantly reduced cytolytic and pro-inflammatory potential . Notably , CC30 strains showed reduced cytolytic and chemotactic potential toward human neutrophils , and increased hematogenous seeding in a bacteremia model , compared to strains in which the genome was altered to express non-CC30 PSMα3 . Our findings describe a molecular mechanism contributing to attenuated pro-inflammatory potential in a main MRSA lineage . They suggest that reduced pathogen recognition via PSMs allows the bacteria to evade elimination by innate host defenses during bloodstream infections . Furthermore , they underscore the role of point mutations in key S . aureus toxin genes in that adaptation and the pivotal importance PSMs have in defining key S . aureus immune evasion and virulence mechanisms .
Staphylococcus aureus is a dangerous human pathogen that is responsible for thousands of deaths annually in the U . S . alone [1] . Virulence of S . aureus is due to a large repertoire of virulence factors , including immune evasion factors and aggressive cytolytic toxins [2] . S . aureus infections become particularly dangerous when they are caused by strains that are resistant to commonly used antibiotics . Methicillin-resistant S . aureus ( MRSA ) is of especially great concern , as identification of MRSA eliminates the therapeutic use of most beta-lactam antibiotics , which are antibiotics of first choice against pathogenic staphylococci . Many countries report high rates of methicillin resistance among hospital-associated infections caused by S . aureus [3] . In addition , community-associated strains of MRSA ( CA-MRSA ) have emerged over the last two decades that have the capacity to infect healthy individuals outside of hospital settings [4] . MRSA strains belonging to clonal complex ( CC ) 30 are a major cause of hospital-associated infections in the U . S . , Europe , and elsewhere [5]–[7] . Infections with CC30 MRSA present predominantly as bloodstream infections with complications such as hematogenous seeding [8] . Historical methicillin-susceptible CC30 strains ( phage type 80/81 ) caused serious , in part community-associated infections of the skin and lungs in addition to blood infections . In contrast to contemporary CC30 isolates , many historical phage type 80/81 clones had genes encoding the Panton-Valentine leukocidin ( PVL ) [7] . Furthermore , contemporary CC30 clones contain mutations in the global virulence regulator Agr ( agrC gene , non-synonymous mutation , G55R ) and the gene encoding α-toxin ( hla , STOP mutation ) [9] . The resulting overall lower expression of cytolytic toxins in contemporary compared to historic CC30 clones has been linked to the fact that contemporary CC30 clones predominantly cause hospital-associated infections [9] . PSMs are short , amphipathic , α-helical peptides with a major impact on S . aureus virulence [10] , [11] . The PSMα peptides of S . aureus in particular cause lysis of a variety of cell types , including neutrophils ( or polymorphonuclear leukocytes , PMNs ) , monocytes , erythrocytes , and osteoblasts [11]–[13] . The rather low target specificity of PSM-mediated cytolysis is due to the fact that lysis is believed to be receptor-independent [14] , [15] , which is reflected by the capacity of PSMs to lyse artificial vesicles [14] . In addition , PSMs have pro-inflammatory capacities that are receptor-dependent , leading for example to neutrophil chemotaxis and activation [11] , [14] . Similar to other S . aureus toxins , these immune-stimulatory activities are observed at sublytic concentrations [14] , [16] . PSMα3 is the by far most pro-inflammatory and cytolytic PSM of S . aureus [11] . Notably , the capacity of PSMα3 to elicit chemotaxis by neutrophils by far exceeds that of any other PSM of S . aureus [11] . Except for a variation in the PSM δ-toxin sequence ( serine substitution for glycine at position 10 in some strains ) , whose effect on peptide function has not yet been analyzed , naturally occurring variants of PSM peptides have not yet been reported; and in general , the consequences of non-synonymous variations in psm genes are not understood . Here , we report an allelic variation in the PSMα3-encoding gene that is characteristic of CC30 strains and leads to significantly lower cytolytic and chemotactic activity , and increased hematogenous seeding in a bacteremia model . For the first time , our study describes an allelic variant of a psm gene that has key biological consequences and whose appearance is strongly correlated with a specific MRSA lineage . Furthermore , our findings reveal a molecular mechanism supporting the notion that MRSA strains such as those of the CC30 lineage evolved to cause specific infections , in which reduction of the expression of immune-stimulatory toxins allows to evade recognition and elimination by host defenses .
When analyzing CC30 strains using an HPLC/MS-based PSM analysis method [17] , we realized that CC30 strains lacked the m/z values typical for PSMα3 , but instead produced a peptide with a changed mass ( Fig . 1A , B ) . The mass of that peptide was 49 Da higher than that of PSMα3 . Sequencing the psmα3 gene in selected CC30 strains and comparing to the published genome of the CC30 strain MRSA252 ( EMRSA16 ) [18] , we identified the cause for that difference as a non-synonymous mutation in the psmα3 gene in the codon coding for the C-terminal amino acid position ( TAC instead of AAC , leading to an exchange of asparagine with tyrosine ) . Notably , we found the mass corresponding to the PSMα3N22Y peptide in all CC30 strains that we analyzed , containing contemporary and historic strains ( Fig . 1C ) [9] , but never in any of the large number of S . aureus strains of different genetic backgrounds that we had analyzed over the last ∼5 years . Some CC30 strains did not produce any PSM peptide ( 9 among the 41 analyzed ) . Lack of PSM production is commonly found in about a quarter of strains in staphylococcal strain collections and due to non-functionality of the Agr system , which strictly regulates PSM production [19] , [20] . Furthermore , a BLASTP search only detected the PSMα3N22Y peptide in the CC30 strain MRSA252 , but not in any other sequenced S . aureus genome . Therefore , we conclude that the N22Y variant of PSMα3 is characteristic of the CC30 lineage . Contemporary CC30 MRSA strains contain two key mutations , one in the hla ( α-toxin ) and one in the agrC gene [9] . When introduced in a historic CC30 strain ( strain 22030 ) , both mutations caused a significant reduction in virulence in a mouse infection model; and the agrC mutation led to reduced production of RNAIII [9] , the intracellular effector molecule of the Agr system [21] . However , we did not detect a significant difference in the average PSMα3N22Y production comparing historic and contemporary CC30 strains from that previous study [9] ( Fig . 2 ) . Given that PSM production is a direct readout of Agr functionality [19] , our present results indicate that the agrC mutation in contemporary CC30 strains does not cause a significant reduction of Agr functionality in average in the contemporary CC30 strain population . We confirmed the results obtained with PSMα3 by analyzing production of δ-toxin , which is the translational product of RNAIII . In both PSMα3 and δ-toxin analyses , contemporary CC30 strains only showed a slightly reduced average production ( PSMα3 , 0 . 99×107 versus 1 . 16×108 intensity units; δ-toxin 3 . 18×108 versus 3 . 81×108 ) ; both differences were not significant ( Fig . 2 ) . In contrast , production of α-toxin as measured by densitometry of Western blot signals was strongly different between the historical CC30 strains and the contemporary strains bearing a stop codon mutation in the α-toxin gene ( Fig . 2 ) . As expected , no α-toxin could be detected in culture filtrates of the latter . These findings attribute a greatly more important impact of the hla than the agrC mutation to causing overall lower aggressive virulence of contemporary as compared to historic CC30 strains . Neutrophil chemotaxis is one of the most important pro-inflammatory activities exerted by PSMs via their interaction with formyl peptide receptors ( FPRs ) [10] , [14] . PSMs activate FPRs at nanomolar amounts , with activation of FPR2 ( also termed FPRL1 ) being ∼10 times stronger than that of FPR1 and FPR3 [14] . We analyzed the capacities of PSMα3 to cause neutrophil chemotaxis in comparison to PSMα3N22Y . PSMα3 showed a capacity to cause neutrophil chemotaxis that was about twice higher than that caused by PSMα3N22Y at all tested concentrations ( Fig . 3A ) . To test the impact of the psmα3 gene mutation in CC30 strains on chemotactic activity in the natural strain background , we changed the corresponding , last codon in the psmα3 coding sequence in the genome of strain MRSA252 by an allelic replacement-based strategy to the sequence found in other S . aureus strains ( from TAC to AAC , tyrosine to asparagine codon ) . Furthermore , we analyzed the impact of PSMα3 and PSMα3N22Y expressed constitutively from a plasmid in a PSM-negative strain background ( strain USA300 with all psm genes deleted ) [22] . PSMα3N22Y and PSMα3 expression was verified by HPLC/MS and found to be highly similar in the respective strain pairs ( Supplemental Fig . S1 ) . Chemotaxis of human neutrophils was significantly higher with the MRSA252 strain that was altered to express non-CC30 PSMα3 ( strain 252* ) compared to the wild-type strain MRSA252 expressing PSMα3N22Y ( Fig . 3B ) . Similarly , chemotaxis was significantly higher in the PSM-free strain expressing PSMα3 from a plasmid compared to the isogenic strain expressing PSMα3N22Y ( Fig . 3C ) . These results demonstrate that PSMα3N22Y causes decreased chemotaxis of human neutrophils compared to PSMα3 , also when expressed in its natural strain background . Next , we analyzed Ca2+ flux as a general measure to test for neutrophil activation , as used in previous studies [14] . HL60 cells stably transfected with different FPR receptor genes [14] were used to determine which FPR receptors are differentially activated by PSMα3 versus PSMα3N22Y ( Fig . 4A ) . Somewhat surprisingly , only FPR1 was activated differentially by the two peptides in this assay , but not FPR2 and FPR3 . As expected , both peptides activated FPR2 much stronger than FPR1 and FPR3 , but we detected no difference in the FPR2-activating potential between the two peptides . Using a one site-binding model , we obtained Kd values for FPR1 of 615 . 6±101 . 2 nM ( PSMα3 ) versus 3039±682 nM ( PSMα3N22Y ) , which represents a significant difference with a factor of ∼5 . In contrast , Kd values were not significantly different for FPR2 ( PSMα3 , 11 . 90±1 . 90 nM; PSMα3N22Y , 10 . 81±2 . 38 nM ) and FPR3 ( PSMα3 , 1377±285 nM; PSMα3N22Y , 953 . 0±253 . 9 ) . However , in human neutrophil chemotaxis experiments using specific inhibitors of FPR1 ( chemotaxis inhibitory protein of S . aureus , CHIPS ) and FPR2 ( FPRL1 inhibitory protein FLIPr ) [23] , [24] , we found that both FPR1 and FPR2 appear to be responsible for the observed differences between PSMα3 and PSMα3N22Y ( Fig . 4B ) . At a peptide concentration of 100 nM , we observed a difference of ∼factor 2 in chemotactic activity comparing the two peptides when FPR1 recognition was inhibited by CHIPS , likely attributable to FPR2-mediated recognition . Owing to the overall ∼10 times lower capacity to stimulate FPR receptors other than FPR2 , we observed no chemotaxis at that peptide concentration with addition of the FPR2 blocking agent FLIPr . However , at a 10 times higher peptide concentration of 1 µM , there was a very strong difference in chemotactic potential when FPR2-mediated recognition was blocked ( Fig . 4B ) , which is in agreement with the results of the transfected HL60 experiments and according to our results mainly attributable to differences in activation of FPR1 . Together , these results indicate that the differences in the pro-inflammatory potentials of PSMα3 and PSMα3N22Y are due to a considerable extent to differential activation of FPR1 , while the overall stronger activation of FPR2 does not differ that much between the two peptides . PSMα peptides are strongly cytolytic to a variety of cell types , which is believed to be a major determinant of their impact on disease progression observed in several manifestations of S . aureus disease , in particular infections of the skin [10] . PSM-mediated cytolysis is assumed to occur independently of receptor interaction by promoting disintegration of eukaryotic plasma membranes [14] , [15] , [25] . To compare the cytolytic capacities of PSMα3 and PSMα3N22Y , we first analyzed the degree to which synthetic peptides lysed human erythrocytes and neutrophils . PSMα3 showed significantly higher cytolytic capacities toward human neutrophils and erythrocytes compared to PSMα3N22Y ( Fig . 5A , B ) . Then , we analyzed the impact of PSMα3 and PSMα3N22Y on cytolysis when expressed from plasmids and the genome ( see above ) . In accordance with the results achieved with the synthetic peptides , culture filtrates of the PSMα3-expressing strain had significantly higher capacities to lyse human erythrocytes and neutrophils ( Fig . 5C–F ) . These findings show that the CC30 psmα3 gene mutation leads to significantly decreased cytolytic activity toward human neutrophils and erythrocytes , including in its original strain background . The CC30 MRSA lineage is a leading cause of sepsis and subsequent hematogenous complications in hospitals . Therefore , we used a mouse bacteremia ( renal abscess ) model with assessment of seeding into and abscess formation in the kidneys to evaluate the impact of the psmα3 mutation on CC30 pathogenic success . Both the contemporary CC30 MRSA strain 252 and the historic strain 22030 showed significantly increased seeding into kidneys and kidney abscess formation compared to the respective isogenic mutants in which the psmα3 gene was altered to the non-CC30 form ( Fig . 6A , B ) . Considerable bacterial numbers ( >1000/kidney ) were found in 70% of kidneys in mice infected with MRSA252 , while never in kidneys infected with the isogenic mutant expressing the non-CC30 form of PSMα3 ( Fig . 6A ) . Similar results were found for the 22030 strain pair ( 65% versus 20% ) ( Fig . 6B ) . Histological analyses of several kidneys from mice infected with MRSA252 showed clear signs of infection: they had developed tubulointerstitial nephritis that was associated with colonies of cocci ( Fig . 6 C , D ) . This was not seen in the kidneys of mice infected with the isogenic strain expressing the non-CC30 form of PSMα3 ( Fig . 6E ) . These results indicate that the specific features of the CC30 mutant of PSMα3 ( PSMα3N22Y ) result in an increased capacity to cause sepsis and subsequent complications . This effect appeared to be stronger in the contemporary clone MRSA252 than in the historic strain 22030 , possibly because the latter strongly expresses a series of other pro-inflammatory toxins that to a certain extent overshadow the effect of the PSMα3 mutation . PSMs of the α-type are known to have a strong impact on acute skin infections [11] , [26] . This is believed to be due primarily to their cytolytic activities , which they exert for example on neutrophils after phagocytosis [27] . We found that the CC30 form of PSMα3 ( PSMα3N22Y ) in strains MRSA252 and 22030 did not cause increased virulence in skin infections as measured by abscess sizes . Rather , abscess sizes were in general slightly larger in the MRSA252 strain expressing non-CC30 PSMα3 compared to the MRSA252 wild-type strain , but differences were not significant . No differences were detectable with the 22030 strain pair ( Supplemental Fig . S2 ) . This indicates that in contrast to bacteremia , the altered features of the CC30-type PSMα3 ( PSMα3N22Y ) do not significantly impact the outcome of skin infections .
In the present study , we report an allelic variant in a psm gene that is characteristic for a specific S . aureus lineage , namely CC30 , and shows reduced immune-stimulatory and cytolytic activities . Our results indicate that the increased capacity of CC30 strains to cause hematogenous complications is at least in part due to the production of this attenuated form of PSMα3 . Furthermore , the strongly different impact of that form of PSMα3 on bacteremia versus skin infection underscores that the two main features of PSMs , i . e . cytolytic and pro-inflammatory capacity , may have a strongly different influence on the development of distinct disease types . Given that a reduction in cytolytic activity can hardly explain increased virulence and hematogenous seeding , the differences that we detected in the bacteremia model are very likely due to the attenuated pro-inflammatory features of the attenuated form of PSMα3 in CC30 strains . These findings are in accordance with the notion that the innate immune system makes use of S . aureus toxins for pathogen recognition and evading that recognition is of benefit for bacterial survival in specific types of disease [28] . This notion is founded on several previous observations: first , S . aureus produces a series of molecules such as CHIPS or FLIPr that block toxin recognition by receptors on innate host defense cells , including the FPRs that recognize PSMs [14] , [23] , [24]; second , it was found that the pro-inflammatory properties of PVL [16] , [29] enhanced clearing of pneumonia and anti-PVL antibodies led to increased severity of infection [30]–[32] . Finally , persistent bacteremia has been reported to be associated with Agr dysfunction [33] . While the latter has been explained by an increased fitness cost associated with the production of RNAIII [34] , this correlation may also be explained by the lower expression of pro-inflammatory molecules in agr mutants ( such as PSMs and other toxins ) . All these observations indicate that toxin production is a double-edged sword for the bacteria: toxins not only serve to eliminate immune cells but also trigger the launch of host defenses , representing a sort of pathogen-associated molecular pattern . Importantly , our study is the first to provide an example in which a staphylococcal toxin variant appears to have evolved to circumvent pathogen recognition . The psmα3 allelic variant is present in all CC30 strains , including historic and contemporary strains . Our findings therefore do not answer the much-debated question which specific mutations are linked to the fact that the latter are predominantly restricted to the hospital setting [7] , [9] . However , our findings suggest that the hla and agrC mutations in contemporary CC30 clones may add to immune evasion by lowering pathogen recognition in a similar fashion as the PSMα3 attenuation , because they also result in an attenuation of the expression of pro-inflammatory toxins . These mutations , which are believed to make contemporary strains of CC30 better adapted to long-term colonization and persistence in the human host , may thus not decrease virulence in a general fashion as has been suggested [9] . They may rather shift the disease spectrum to types of infection in which the production of aggressive toxins that alert the immune system is counterproductive for the establishment and progression of infection . According to our results , the hla mutation plays a much greater role in that adaptation of contemporary CC30 clones to hospital-associated infections than the agrC mutation , Altogether , these findings attribute a key role to point mutations in major S . aureus toxins for virulence and correlation with specific disease types . Results that we previously obtained using an alanine exchange peptide library screen of PSMα3 indicated that the C-terminal amino acids are critical for interaction with FPR2 [35] . In general , the difference in receptor interaction between PSMα3 and PSMα3N22Y as determined in the present study supports this concept . However , exchange of the C-terminal asparagine with tyrosine ( CC30 PSMα3N22Y ) leads to a different change in receptor interaction as compared to an exchange with alanine ( PSMα3N22A ) , inasmuch as the former more strongly affects interaction with FPR1 . This indicates that the molecular details of PSM – FPR interaction differ between the different subtypes of FPRs . In conclusion , we here report an allelic variation of a key member of the PSM toxin family that significantly impacts in-vitro virulence phenotypes and in-vivo virulence of strains of the CC30 lineage . Our findings lend support to the general idea that mutations leading to lower pro-inflammatory potential are linked to the involvement in specific types of disease .
The used animal protocol ( LHBP1E ) was reviewed and approved by the Animal Use Committee at NIAID , NIH , according to the animal welfare act of the United States ( 7 U . S . C . 2131 et . seq . ) . Human neutrophils were isolated from venous blood of healthy volunteers in accordance with protocols approved by the Institutional Review Board for Human Subjects , NIAID , and the University of Tübingen , Germany . Informed written consent was obtained from all volunteers . The bacterial strains used in this study are contemporary and historic CC30 strains as published in a previous study [9] . MRSA252 is a widespread contemporary CC30 isolate and a frequent cause of hospital-associated infections [18] . LAC is a CA-MRSA strain of the pulsed-field type USA300 . Strains were grown in tryptic soy broth ( TSB ) with addition of tetracycline at 12 . 5 µg/ml when appropriate . Culture filtrates for all experiments were obtained from 50-ml cultures grown for 18 h in 100-ml flasks with shaking at 180 rpm . They were filtered with PES filters ( 0 . 2 µm pore size , Millipore ) and used fresh or stored at −20°C for further use . Only for the analysis of the CC30 strain collection ( Fig . 1C ) , strains were grown as 1-ml cultures in 5-ml tubes , with all other conditions being the same . Synthetic PSM peptides were obtained from commercial vendors at a purity of >95% . They all carry the N-terminal N-formyl methionine as common for PSMs , which are secreted into the medium without a signal peptide [36] . Peptide stock solutions were made in dimethyl sulfoxide at 10 mg/ml , which were diluted in RPMI 1640 ( Gibco ) for PMN or Dulbecco's phosphate-buffered saline ( DPBS ) ( Sigma ) for hemolysis assays . Allelic replacement was performed as previously described [11] using the plasmid pKOR1-based method [37] to change the psmα3 gene in MRSA252 and 22030 ( encoding PSMα3N22Y ) to non-CC30 psmα3 . Template DNA was from strain LAC ( USA300 ) . Oligonucleotides used are shown in Table 1 . Plasmids for PSM expression are based on plasmid pTXΔ , which is a derivative of the xylose-inducible plasmid pTX15 [38] , in which inducible was changed to constitutive expression by deletion of the xylR repressor gene [11] . Oligonucleotides for construction of the pTXΔpsmα3N22Y plasmid were used as reported [11] to amplify the psmα3 gene from genomic DNA of strain MRSA252 . Fidelity of the replacement was ascertained by DNA sequencing of the psmα operon and HPLC/MS of PSMs showing the exchanged mass of PSMα3 versus PSMα3N22Y . Additionally , production of PSMs was verified in the culture filtrates and cultures of strains used for all experiments prior to performing the experiment , to ensure that PSMα3 and PSMα3N22Y production was at the same level and no agr or other mutation impacting PSM levels had occurred ( see Supplemental Fig . S1 ) . PSM concentrations in culture filtrates grown for 16 h in TSB were measured using HPLC/MS as described [17] . Staphylococcal strains were grown in TSB at 37°C overnight . 10 µl of these cultures were inoculated into 1 ml of TSB and grown for 8 h . Culture supernatants were loaded onto 15% SDS-PAGE gels and run at 150 V for 1 h . Proteins in the gels were blotted on nitrocellulose membranes using an iBlot Western blotting system ( Life Technologies , Grand Island , NY ) . Blotted membranes were incubated with Odyssey blocking buffer ( LI-COR , Lincoln , NE ) for 1 h at room temperature . Anti-staphylococcal α-toxin rabbit serum ( 1∶2 , 000 , Sigma-Aldrich , St . Louis , MO ) was added to the blocking buffer and incubated for another hour at room temperature . Membranes were washed five times with washing buffer ( Tris-buffered saline containing 0 . 1% Tween-20 , pH 7 . 4 ) and incubated with 1∶10 , 000 diluted Cy5-labeled goat anti-rabbit IgG ( Life Technologies , Grand Island , NY ) in Odyssey blocking buffer in dark for 1 h at room temperature . Membranes were washed five times with the washing buffer and scanned with a Typhoon TRIO+ Variable Mode Imager ( GE Healthcare , Piscataway , NJ ) . The amount of α-toxin in the scanned image was quantified using ImageQuant TL software ( GE Healthcare , Piscataway , NJ ) . Human neutrophils were isolated from venous blood of healthy donors as described [24] , [39] . Neutrophil chemotaxis and Ca2+ flux were determined as previously described [11] , [24] . Briefly , neutrophils were subjected to a brief hypotonic shock with pyrogen-free water , washed , and suspended at 5×106 cells/ml . Chemotaxis of neutrophils was determined using a transwell system ( Costar ) with analysis of neutrophil migration using fluorescence labeling . Calcium fluxes were monitored as a surrogate marker for chemotaxis since it can be measured more robustly and accurately than chemotaxis . For measurement of Ca2+ fluxes , 5×106 neutrophils/ml were labeled with a fluorescent dye and analyzed with a FACScalibur ( Becton Dickinson ) as described [24] . For measuring the influence of CHIPS or FLIPr , 1×106 cells/ml were pre-incubated with CHIPS or FLIPr at final concentrations of 1 . 4 µg/ml or 0 . 5 µg/ml , respectively , for 20 min at room temperature under agitation . The CHIPS and FLIPr proteins ( a kind gift from K . van Kessel ) were prepared as described [23] , [24] . Measurements of 2 , 000 events were performed and Ca2+ flux was expressed as relative fluorescence corrected for buffer controls . For the measurement of neutrophil lysis , synthetic PSMs or ( diluted ) culture filtrates were added to wells of a 96-well tissue culture plate containing 106 PMNs and plates were incubated at 37°C . At the desired times , PMN lysis was determined by release of lactate dehydrogenase ( LDH ) ( Cytotoxicity Detection Kit , Roche Applied Sciences ) . Hemolysis assays were performed according to Wang et al . [11] Briefly , whole blood from humans was washed twice with ice-cold DPBS . The final percentage of blood was 2% ( v/v ) in ice-cold DPBS . Equal volumes of diluted synthetic PSMs or dilutions of culture filtrates were incubated with the 2% erythrocyte suspension in a total volume of 200 µl in 96 round-bottomed plates . After incubation at 37°C for 1 h , the plates were centrifuged at 233×g at 4°C for 5 min , supernatants were collected , and the optical density was measured at 540 nm . The mouse skin infection model was performed as described [11] . Briefly , six to eight week-old female Crl∶SKH1-hrBR hairless mice ( Charles River Laboratories ) were injected subcutaneously with ∼4×107 CFU of S . aureus strains 252 or 252* , or ∼5×106 CFU of S . aureus strains 22030 or 22030* in 50 µl of PBS in the left flank of the mouse . The length ( L ) and width ( W ) of the abscess or lesion caused by the bacterial infection was measured with an electronic caliper daily for 14 d post infection and calculated using the formula L×W . Typically , strain MRSA252 caused closed abscesses and strain 22030 open lesions . All animals were euthanized after completion of the entire procedure . All mouse experiments were performed blinded at the animal care facility of the NIAID , Building 33 , in compliance with the guidelines of the NIAID/NIH Institutional Animal Care and Use Committee . For the renal abscess model , S . aureus strains were inoculated from a pre-culture and grown to mid-exponential growth phase ( ∼2 h ) , harvested , washed , and diluted with sterile PBS . Six to eight week-old female CD-1 mice ( Charles River Laboratories ) were infected with 50 µl of bacteria in PBS via the tail vein . Mice received ∼1×107 CFU of S . aureus strains 252 or 252* , or ∼1×106 CFU of S . aureus strains 22030 or 22030* . Four days post infection terminal cardiac bleeds were performed . There were no bacteria found in the blood . All mice were euthanized by CO2 inhalation and kidneys were collected . One kidney of each mouse was placed into a 2-ml tube containing 1 ml of sterile PBS with 500 mg of 2 mm borosilicate glass beads ( Sigma ) . The kidney was homogenized in a Fast Prep bead beater ( Thermo Savant ) at 6 m/s for 20 s . The homogenates were diluted in PBS , plated onto TSB plates , and incubated overnight at 37°C for CFU counting . The other kidney was placed in 10% formalin ( Sigma ) for subsequent histopathological examination . Statistical analysis was performed using Graph Pad Prism version 6 . 02 . For the comparison of two groups , t-tests were used ( unpaired unless otherwise noted ) , for three or more , 1-way or 2-way ANOVA , as appropriate . All error bars depict the standard deviation . | Methicillin-resistant Staphylococcus aureus ( MRSA ) is a major cause of morbidity and mortality and a great concern for public health . The CC30 MRSA lineage is especially notorious for causing bloodstream infections with complications such as seeding into organs . In our study , we show that this lineage produces an attenuated form of a key S . aureus toxin with decreased pro-inflammatory features . Our results suggest that attenuation of this toxin allows the bacteria to evade recognition and subsequent elimination by host defenses , thereby increasing pathogen success during blood infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"methicillin-resistant",
"staphylococcus",
"aureus",
"infectious",
"diseases",
"bacterial",
"diseases",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"staphylococcus",
"medical",
"microbiology",
"host-pathogen",
"interactions",
"microb... | 2014 | Production of an Attenuated Phenol-Soluble Modulin Variant Unique to the MRSA Clonal Complex 30 Increases Severity of Bloodstream Infection |
Previous work has shown that asymmetry in viral phylogenies may be indicative of heterogeneity in transmission , for example due to acute HIV infection or the presence of ‘core groups’ with higher contact rates . Hence , evidence of asymmetry may provide clues to underlying population structure , even when direct information on , for example , stage of infection or contact rates , are missing . However , current tests of phylogenetic asymmetry ( a ) suffer from false positives when the tips of the phylogeny are sampled at different times and ( b ) only test for global asymmetry , and hence suffer from false negatives when asymmetry is localised to part of a phylogeny . We present a simple permutation-based approach for testing for asymmetry in a phylogeny , where we compare the observed phylogeny with random phylogenies with the same sampling and coalescence times , to reduce the false positive rate . We also demonstrate how profiles of measures of asymmetry calculated over a range of evolutionary times in the phylogeny can be used to identify local asymmetry . In combination with different metrics of asymmetry , this combined approach offers detailed insights of how phylogenies reconstructed from real viral datasets may deviate from the simplistic assumptions of commonly used coalescent and birth-death process models .
Genetic approaches to investigating infectious diseases are well-established , exploiting the naturally high genetic diversity in pathogen populations such as HIV and influenza to reconstruct both their evolutionary and epidemiological dynamics [1] . Phylogenies contain potentially large amounts of information on disease dynamics , and can help reveal the disease incidence and prevalence , changes in historical population size , and population substructure [2–4] . However , there can be confounding factors when trying to convert evolutionary dynamics into epidemiological quantities such as transmission rates , and ideally we want to be able to explicitly model viral transmission in an evolutionary framework , taking into account features such as the host population structure ( for example , differences in contact rates between groups of individuals ) and the natural course of infection ( for example , differences in infectiousness during the acute and chronic phases of HIV infection ) [5] . One way to investigate the extra biological complexity of such patterns is to consider the shape or branching structure of the phylogeny , a feature that is arguably underused despite being relatively straightforward to infer . Evidence of asymmetry in a tree reflects heterogeneity in the population that has arisen due to the processes by which a tree has grown [6]; previous work suggests that evidence of asymmetry in a phylogenetic tree can arise due to selection [2] , heterogeneity in contact rates [7] and population structure [5] . Since many tree models assume homogeneity in the population , it is important to be able to identify which parts of the tree might be driving asymmetry , and whether or not this is problematic under the modelling assumptions—preferably before running computationally expensive analyses . It is common to analyse viral datasets sampled over multiple timepoints . As viruses , including RNA and ssDNA viruses , evolve rapidly , phylogenetic reconstruction gives rise to trees with root-to-tip distances that reflect , in part , sampling times . However , such trees are more likely to be asymmetric , resulting in standard metrics developed for homochronous sampling being implicitly biased ( see Supplementary Information of Frost and Volz ( 2013 ) [5] ) . This is due to the fact that most metrics use the topological distance ( that is , the number of nodes traversed between two points in the tree ) , and isolates sampled earlier in the history of the phylogeny will tend to have fewer nodes between them and the root of the tree . In this paper , we propose a permutation-based approach that allows an observed phylogeny to be compared to random phylogenies with the same sampling and coalescence times . This approach can also be used to assess asymmetry throughout evolutionary history in a rooted tree , therefore also allowing areas of local asymmetry to be identified in addition to a single global value at the root of the tree . We demonstrate this approach on three datasets with different expected types of heterogeneity , illustrating the imprint of various transmission dynamics on viral phylogenies .
There are a number of ways to measure the balance of a phylogeny . Most approaches consider either the topological distance ( the number of nodes ) between two parts of the tree , for example Sackin’s index , or the balance of each internal node by comparing the number of leaves in the left and right subtrees below it , for example Colless’ index [8–11] . Here , we consider two measures of asymmetry: Sackin’s index [11] , and the number of cherries [12] , although the methodology can easily be extended to other metrics . Sackin’s index is the total topological distance between the leaves and root of the tree . If dj is the number of nodes to be traversed between each leaf j and the root , then Sackin’s index is the total over all leaves , I S = ∑ j d j . ( 1 ) In trees where the tips have been sampled at the same time , the expected Sackin’s index , 𝔼 ( IS ( n ) ) , for n isolates in the sample is given by: 𝔼 ( I S ( n ) ) = 2 n ∑ k = 2 n 1 k ( 2 ) under the Yule or coalescent models [13] . For large n , 𝔼 ( IS ( n ) ) ≈ 2n log ( n ) . Since the expected value of the Sackin’s index increases with the tree sample size , it is common to either divide the statistic by n ( i . e . the mean topological distance from root to tip ) , or use the following standardisation proposed by Leventhal et al . [7]: I ¯ S ( n ) = I S ( n ) - 𝔼 ( I S ( n ) ) 𝔼 ( I S ( n ) ) . ( 3 ) However , since the permutation method outlined in this paper compares an observed tree to those of the same size ( i . e . like with like ) , we simply use the non-standardised version here . We use the function sackin . test in the apTreeshape R package to test the hypothesis of asymmetry in the tree , comparing the observed value to 10 , 000 trees simulated under the Yule model [14 , 15] . A cherry is formed when two tips share a direct ancestor . In an asymmetric tree , tips generally coalesce with branches earlier in the ancestry of the tree , and therefore fewer cherries are expected than with a balanced tree . Under the Yule or coalescent model , the expected number of cherries , Cn , in a tree with n taxa is n/3 , and for a uniform tree is n/4 [12] . In addition , McKenzie and Steel showed that the number of cherries is asymptotically normal with C n - n / 3 2 n / 45 → 𝓝 ( 0 , 1 ) ( 4 ) under the Yule or coalescent model , and C n - n / 4 n / 16 → 𝓝 ( 0 , 1 ) ( 5 ) for a uniform tree [12] . These two metrics complement each other well , as the number of cherries reveals recent asymmetry in the tree , whereas Sackin’s index gives the asymmetry of the tree over the whole evolutionary history [5] . In addition , these metrics are only weakly correlated , unlike for example , the Sackin and Colless indices [16 , 17] . The ordering of nodes in a rooted tree means we can consider the asymmetry in the phylogenetic tree throughout the evolutionary period , and not just at the root . This asymmetry could be due to a small effect at each internal node accumulating throughout the tree , or due to one or more nodes with highly imbalanced subtrees below them . Calculating the asymmetry over the entire course of the tree allows us to identify local asymmetry , even when there may not be significant evidence for global asymmetry ( as obtained by considering the cumulative statistics at the root of the tree ) . There are two main types of event that can affect the shape of a phylogeny: a coalescence , and a new sampling event , which adds a tip . Sackin’s index and the number of cherries are both concerned with internal nodes rather than the tips , so we need only consider the former . At each coalescent event , we consider the contribution of that node to the overall metric . This results in a vector of n − 1 values , one for each ancestral node , giving a measure of how asymmetric the subtree below the node is ( Fig 1 ) . We can add these values cumulatively as we go backwards in time from the present towards the root , to investigate how asymmetry builds up over the course of the tree . For the number of cherries , calculating the effect of each individual node is straightforward—being 1 if the node is a cherry ( i . e . the direct ancestor of two tips ) and 0 if it is not . To calculate the Sackin’s index for each node , rather than count the topological distance to the root for each tip as the calculation of the Sackin’s index is usually presented , we instead consider the number of times each node is traversed going from the tip to the root . Namely , this is the number of tips found in the subtree below the node of interest . To obtain the distribution of possible values for each the statistics for an observed tree , we permute the tree whilst retaining the same tip sampling and internal node times ( Fig 2 ) . These simulated trees form a neutrally evolving null distribution of coalescent trees , conditioned on the same tip and internal node times as the observed tree . For n tips , there are n − 1 internal nodes . Starting at the time of the most recent internal node ( say , t1 ) and going backwards in time from the present at t = 0 , we consider all tips that were sampled more recently ( i . e . between t1 and t = 0 ) . Two of these tips are then chosen at random to coalesce , thus creating the internal node for t1 . This continues backwards in time for each node in turn , with the only difference that coalescences can be between sample tips and nodes that have already been produced via a coalescence between the time to node i , ti , and the present . The code to simulate permutations of an observed tree with the same sampling and coalescence times , and all the imbalance metrics considered above , were written in R [15] , and are available as part of the treeImbalance package on GitHub ( https://github . com/bdearlove/treeImbalance ) , and are in the Dryad Digital Repository: http://dx . doi . org/10 . 5061/dryad . v7817 . To obtain a distribution of possible values for each imbalance metric , 10 , 000 permutations of the observed tree with the same tip sampling and internal node times were generated . For each of these permuted trees , the number of cherries and Sackin’s index were calculated at each internal node and globally by computing the cumulative statistics at the root ( Fig 1 ) . The median trajectory of Sackin’s index and the number of cherries throughout the ancestry of the tree ( shown with a solid red line in plots ) was calculated by partitioning around the medoid with a single cluster using the function pam in the cluster R package [18] . The medoid represents the trajectory which has the least dissimilarity with all the other possible trajectories from the permutation test . This ensures that the median is obtained from within the set of permuted trajectories , thus ensuring it is a ‘viable’ trajectory , and overcomes issues associated with other methods ( such as calculating the mean or median statistic at each node ) , which do not necessarily force the trajectory to be monotonically increasing . At each internal node , the 95% confidence interval was calculated by inverting the hypothesis test around the medoid value at that timepoint [19] . The medoid was subtracted from the permuted trees , and then the critical points of this distribution are found where 2 . 5% of the values are as or more extreme ( with no interpolation ) . The confidence interval then is obtained by adding these back to the medoid . Calculating the 95% confidence in this way , as opposed to using quantiles or the variance , ensures that the value calculated is within the permuted dataset . Since considering the local imbalance at each node results in a multiple hypothesis test , several p-value adjustments were considered in order to control the family-wise error rate ( including the Bonferroni correction and methods proposed by Holm ( 1979 ) , Hochberg ( 1988 ) and Hommel ( 1988 ) [20–22] ) , and the false discovery rate ( including methods proposed by Benjamini and Hochberg ( 1995 ) and Benjamini and Yekutieli ( 2001 ) [23 , 24] ) . For the latter , we also investigated the q-value , which estimates the proportion of significant hypotheses that are false [25–27] . Results were generally consistent ( S1 Table ) , so here we report the most conservative adjustment , the Bonferroni correction , alongside the unadjusted p-values . The uncorrected values remain valuable since the purpose of the test is to identify potential deviations from the model for further investigation , rather than necessarily a strict hypothesis test . For the cumulative statistics , the Bonferroni correction is equal to the number of internal nodes , n − 1 . For the single node contribution to Sackin’s index , the correction is n − 2 , since at the root n tips will always be added . To illustrate the bias of standard metrics , we simulated two sets of trees—one set with homochronous sampling ( tips sampled at the same time ) and one set with heterochronous sampling ( tips sampled at different times ) . These were generated using Serial SimCoal [28] under a coalescent model with effective population size of 104 , with 100 tips sampled in the present for the homochronous sampling , and sampled over 10 time points each 1000 generations apart for the heterochronous sampling . A single tip-dated phylogeny is required as input for our permutation approach . These can be obtained via a number of methods , but for viral datasets , the use of BEAST [29] is most common . Before implementing the permutation test , the observed trees were checked for polytomies , which were subsequently resolved into randomly ordered dichotomies with zero branch lengths . Negative branches were set equal to zero . Tree files were available for the ebola virus [30] and influenza A virus [31] datasets in Newick format , and are available in the Dryad Digital Repository: http://dx . doi . org/10 . 5061/dryad . v7817 . For the within-host HIV dataset [32] , the sequences were aligned using MUSCLE v3 . 8 . 31 [33] and the maximum clade credibility tree ( MCC ) obtained using BEAST 1 . 8 with a GMRF Bayesian Skyride coalescent model [29] . The GTR model of nucleotide substitution [34] was used with an uncorrelated log-normal relaxed clock and a discretised gamma distribution with four categories was used to model rate heterogeneity across the sequence [35] . For the log-normal relaxed clock parameters , a uniform prior between 0 . 0 and 1 . 0 × 10100 was assumed for the mean , and an exponential with mean 1/3 for the standard deviation . A uniform ( Dirichlet ) prior was used for the nucleotide frequencies . The MCMC was run for 1 billion iterations , with a 10% burn-in period and samples saved every 10 , 000 iterations . The within-host HIV skyride plot was obtained from the observed tree in R using an approximate approach that employs an integrated nested Laplace approximation [36] .
We considered 98 influenza A virus H5N1 haemagglutinin sequences sampled from various bird species around seven locations in Asia ( as distributed with BEAST v1 . 8 . 0 , data originally collated by Wallace et al . ( 2007 ) [29 , 31] ) . Here , we would reasonably expect that there could be three main sources of asymmetry in the phylogeny: the temporal sampling , selection and population substructure in the form of host species and location . Using the standard Sackin’s index , the phylogeny is found to be extremely asymmetric ( p-value <0 . 0001 ) , though there is not enough evidence to reject the null hypothesis of asymmetry at the tips using the number of cherries ( p-value = 0 . 100 ) . However , when we condition on the heterochronous sampling and coalescence times using the permutation test , we find that there is no evidence for global asymmetry with either statistic ( Fig 4 ) . There may be evidence of individual nodes being more asymmetric than expected , with 12 nodes significant at an unadjusted significance level of 5% , though none remain significant after the Bonferroni correction ( Fig 4a ) . This suggests that the extreme result seen with the standard Sackin’s index was due to non-epidemiological effects , rather than heterogeneity in the population . However , the unadjusted p-values may still hint towards a deviation from the model so it could be worth investigating a model that allows for heterogeneity . The 2014 West Africa epidemic of ebola virus is the largest known outbreak of the virus , causing 25 , 791 cases and 10 , 689 deaths ( as of 15th April 2015 ) across Guinea , Liberia and Sierra Leone [37] . A recent study by Gire et al . [30] investigated sequences from 78 patients in Sierra Leone , suggesting a central African source to the outbreak in 2004 with continued human-to-human transmission , as opposed to punctuated re-transmission from a zoonotic source . A subsequent paper by Volz and Kosakovsky Pond [38] found strong evidence for superspreading , with much variance in the number of onward transmissions per individual , in contrast to the results of Stadler et al . who found that using two classes of transmission rates did not offer a significant improvement over an unstructured model [39] . Volz and Kosakovsky Pond note that this heterogeneity in transmission causes highly imbalanced phylogenies . Using a different method , Łuksza , Bedford and Lässig , identified a clade with a significantly higher growth rate than the ancestral clade it diverged from—again providing evidence for deviation from a simple randomly mixing model [40] . Similarly to the influenza data , the standard Sackin’s index showed evidence of global asymmetry ( p-value <0 . 0001 ) , whilst the null hypothesis could not be rejected for the number of cherries ( p-value = 0 . 141 ) . Fig 5 shows the trajectory plots for the same statistics using the permutation test , showing that when controlling for the tip sampling being heterochronous , there is no evidence for asymmetry . Again , this suggests that the extreme result was due to non-epidemiological effects rather than heterogeneity in the tree . This clearly does not fit with what previous work has revealed about the dynamics of the epidemic , and it may reflect the limited power of these statistics compared to models that take the full phylogeny into account . Within a single host infected with HIV , we might expect that selection driven by neutralising antibodies would be the primary driver of asymmetry in the phylogenetic tree of the viral envelope , as rates of diversifying selection are significantly higher in HIV-1 env in individuals with robust neutralising antibody responses [32] . However , this is not the only cause of asymmetry in a phylogeny . We re-examined the HIV env sequence data of a patient who was previously shown to have a slow rate of immune escape from neutralising antibodies [32] . There were 134 full-length env sequences available , collected from 13 time points sampled over 1 , 098 days of follow up ( Fig 6a ) . This phylogeny was found to be asymmetric with the standard Sackin’s index ( p <0 . 0001 ) , and was also significant using the number of cherries ( p = 0 . 028 ) . Correcting for the tip sampling with the permutation test , the number of cherries shows no evidence of global asymmetry in the phylogeny , though suggests there is some evidence of recent local asymmetry between 767 and 781 days from the present ( Fig 6b ) . Sackin’s index shows strong global asymmetry at the the root , which accumulates throughout the depth of the tree ( Fig 6c ) . Within this , there are six individual nodes identified as having more asymmetric than expected subtrees below them ( Fig 6d ) with the Bonferroni correction , and 20 at the unadjusted 5% level . If we consider the q-value instead , there are 14 nodes with a q-value of 2 . 5% in the upper tail ( for a 5% two-tailed test ) , suggesting that less than one of them ( 0 . 35 ) will be a false negative . Examining the skyline plot for these data ( S1 Fig ) does not indicate any deviations from the null model . There are two distinct clades circulating within the patient at the same time in the tree , and if these clades were non-overlapping in time , we would see a pronounced dip in the skyline plot . This is not the case , with the effective population size instead showing steady exponential growth . This pattern and treeshape is reminiscent of the inter-subtype competition identified by Ferguson , Galvani and Bush [41] .
In this paper , we have presented a framework to quantify asymmetry in phylogenetic trees where the tips have been sampled at different times . Previously , it has been highlighted that understanding the link between a tree topology and the evolutionary processes that gave rise to it is difficult [6 , 42] , which is further confounded by the fact that standard tests for asymmetry are implicitly biased in trees with heterochronous sampling [5] . The permutation test described here allows an observed phylogenetic tree to be compared to a distribution of coalescence trees , conditional on the same internal node and tip sampling structure . This is in contrast to the Temporal Clustering ( TC ) statistic proposed by Gray et al . [43] , which tests for a ‘temporal signal’ in a tip-dated phylogeny , whereby sequences sampled around the same time are found clustered together in the tree and among these is the ancestor of any clade with sampling dates closer to the present . Their statistic permutes the tips with a fixed tree , whereas the test presented here permutes the tree conditional on the observed temporal structure in the form of tip sampling dates and internal node times . Trees with high temporal clustering have a higher potential for false positives from the standard global tests . The three datasets presented in this paper all have a strong temporal signal according to their TC statistic . However , when we control for their temporal structures , they display different levels of asymmetry . Although only Sackin’s index and the number of cherries were illustrated here , the permutation test can be extended to other metrics of asymmetry including Colless’ index [10] , Shao and Sokal’s balance statistics B1 and B2 [44] , and the shape statistics of Agapow and Purvis [45 , 46] . These statistics use varying measures of topological distance to quantify asymmetry , meaning they tend to be biased when tips are sampled earlier in the tree and have fewer nodes connecting them to the root and other tips of the tree . Given that the number of cherries and Sackin’s index did not have the power to identify the asymmetry present in the ebola tree , it may well be worth considering a wider range of statistics alongside the permutation test if there is strong external suggestion of asymmetry in the tree . It is important to note that the branch lengths in a phylogeny can also convey important information about the dynamics of disease . The kernel function of Poon et al . [42] accounts for differences in branch lengths when comparing multiple trees , but cannot be used to statistically assess a single observed tree on its own . However , our permutation test could be used alongside this method to calculate the distance between the observed tree and simulated null trees . Additionally , the topology and the branch lengths of a viral phylogeny are not necessarily equivalent to the underlying transmission tree [47] , and therefore it is important to be aware of the possible discrepancy in equating asymmetry in the phylogeny with asymmetry in transmission . Generally , more complicated models will better fit the data . However , increased model complexity can be computationally intensive . As such , the model that is considered the best comes from a balance of the scientific relevance ( the biological plausibility ) , the goodness of fit , and complexity [48] . While this usually relies on some simplifying assumptions , these are often violated—such as the assumption of a randomly mixing population . As a result , it is important to bear in mind the overall fit of the model to data . In the Bayesian framework , posterior predictive simulation is widely used for model checking , but despite recommendations for its use in the literature [49–53] , it remains underutilised in the field of phylogenetics . In addition , these tests are often only possible alongside or once the analysis has been completed , after much computational effort . Since the base topology can often be recovered relatively quickly and accurately , our permutation test represents a quick method for checking whether the assumption of random mixing is supported , or whether there is evidence of asymmetry and therefore heterogeneity in the population . We simply test for evidence of asymmetry in an observed tree , which can arise in the tree due to many processes in the underlying population such as contact rates and population structure [5 , 7] . As evidenced with the within-host HIV data , it is not necessarily simple to interpret the underlying cause of local asymmetry being detected . It might be preferable to control for certain aspects of asymmetry occurring in the tree ( that is , allow for some specific asymmetry in the null model ) , and see if there is significant evidence for further imbalance beyond that expected under the null model . However , methods that have become standard for inferring structure in the phylogenetic tree , such as the phylogeographic approach of Lemey et al . [54] , make the assumption that the tree branching structure is not affected by the heterogeneity in the population ( i . e . the population is randomly mixing , and the discrete trait model is simply overlaid over the tree ) . Thus , our permutation test can be used to justify whether this is an appropriate assumption , or whether it might be more advisable to use a more complex model such as the structured coalescent [55 , 56] . Our approach is fast , has a free software implementation , and can offer important additional insights by highlighting potential lack of goodness-of-fit of commonly used coalescent and birth-death models . | Phylogenetic trees of viruses sampled from different individuals provide clues to the dynamics of transmission . The extent to which the tree is asymmetric may be influenced by biological factors such as differences in infectiousness or contact rates between individuals , but also by nuisance factors such as the pattern of sampling . We have devised a simple statistical test for asymmetry , which controls for sampling patterns and potentially complex temporal dynamics by conditioning on the sampling and coalescence times in a phylogeny , and can also detect whether specific clades in the phylogeny drive patterns of asymmetry . We apply our approach to data on HIV , influenza A virus H5N1 , and ebola virus . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Measuring Asymmetry in Time-Stamped Phylogenies |
To understand the molecular mechanisms underlying paramutation , we examined the role of Unstable factor for orange1 ( Ufo1 ) in maintaining paramutation at the maize pericarp color1 ( p1 ) and booster1 ( b1 ) loci . Genetic tests revealed that the Ufo1-1 mutation disrupted silencing associated with paramutation at both p1 and b1 . The level of up regulation achieved at b1 was lower than that at p1 , suggesting differences in the role Ufo1-1 plays at these loci . We characterized the interaction of Ufo1-1 with two silenced p1 epialleles , P1-rr′ and P1-prTP , that were derived from a common P1-rr ancestor . Both alleles are phenotypically indistinguishable , but differ in their paramutagenic activity; P1-rr′ is paramutagenic to P1-rr , while P1-prTP is non-paramutagenic . Analysis of cytosine methylation revealed striking differences within an enhancer fragment that is required for paramutation; P1-rr′ exhibited increased methylation at symmetric ( CG and CHG ) and asymmetric ( CHH ) sites , while P1-prTP was methylated only at symmetric sites . Both silenced alleles had higher levels of dimethylation of lysine 9 on histone 3 ( H3K9me2 ) , an epigenetic mark of silent chromatin , in the enhancer region . Both epialleles were reactivated in the Ufo1-1 background; however , reactivation of P1-rr′ was associated with dramatic loss of symmetric and asymmetric cytosine methylation in the enhancer , while methylation of up-regulated P1-prTP was not affected . Interestingly , Ufo1-1–mediated reactivation of both alleles was accompanied with loss of H3K9me2 mark from the enhancer region . Therefore , while earlier studies have shown correlation between H3K9me2 and DNA methylation , our study shows that these two epigenetic marks are uncoupled in the Ufo1-1–reactivated p1 alleles . Furthermore , while CHH methylation at the enhancer region appears to be the major distinguishing mark between paramutagenic and non-paramutagenic p1 alleles , H3K9me2 mark appears to be important for maintaining epigenetic silencing .
Paramutation , originally described at the r1 ( red1 ) locus in maize [1] , refers to the exchange of epigenetic information between two alleles in a heterozygote that leads to heritable change in expression of one of the alleles . In maize , regulatory genes involved in the synthesis of flavonoids -anthocyanins and phlobaphenes- have been extensively used to study paramutation . Thus far , paramutation has been described for four loci involved in flavonoid biosynthesis: r1 ( red1 ) and b1 ( booster1 ) encode basic helix-loop-helix ( bHLH ) transcription factors while pl1 ( purple plant1 ) and p1 ( pericarp color1 ) encode R2R3 Myb transcription factors [2] . Of these r1 , b1 , and pl1 regulate biosynthesis of anthocyanins while p1 regulates biosynthesis of phlobaphenes . Paramutation-like phenomena have also been reported for another gene in maize [3] , and other plants and animals [4] , [5] . Detailed characterization of paramutation at the b1 locus demonstrated that seven 853-bp tandem repeats located ∼100-kb upstream of the transcription start site are required for paramutation [6] , [7] . Cloning of mediator of paramutation1 ( mop1 ) revealed that an RNA-dependent RNA polymerase ( RDR ) , most similar to Arabidopsis RDR2 , is required for establishment and maintenance of paramutation [8] and indicated that RNA-mediated chromatin silencing regulates paramutation [9] . The role of RNA-mediated silencing mechanisms in paramutation has been further strengthened by the cloning of two additional maize genes . Of these , required to maintain repression6 ( rmr6 ) encodes largest subunit of RNA polymerase IV ( Pol IV ) similar to Arabidopsis NRPD1 [10] and Mop2/rmr7 encodes second-largest subunit similar to Arabidopsis NRPD2/E2 [11] , [12] which functions in both Pol IV and Pol V complexes [13] . Pol IV and Pol V are plant-specific polymerases which function in the biogenesis of small RNAs and in the RNA-mediated chromatin silencing pathway [13]–[15] . The p1 locus has been used as a model system to understand molecular and epigenetic mechanisms that regulate tissue-specific gene expression . p1-controlled red phlobaphene pigments accumulate in pericarp , cob glumes , husk , silk and tassel glumes . Alleles of p1 are identified based on a two-letter suffix , which denotes their expression in the pericarp and cob glume , respectively [16] , [17] . Variable pigmentation patterns of several p1 alleles have been attributed not only to DNA sequence differences , but also to differential epigenetic states [18]–[21] . For instance , P1-rr and P1-wr , two prototype p1 alleles have distinct phenotypes; P1-rr produces red pericarp and red cob glume while P1-wr specifies white pericarp and red cob glume phenotypes . These alleles have major structural differences: P1-rr is a single copy gene while P1-wr is composed of six or more gene copies tandemely arranged in a head-to-tail fashion [18] . Both alleles have similar coding and regulatory sequences and functional analyses have identified similar basal promoter and proximal enhancer regions [22] , [23] . However , P1-rr has fully pigmented pericarp whereas P1-wr accumulates no pigment in this tissue . The difference in pericarp expression pattern has been attributed to higher DNA methylation within the regulatory sequences of P1-wr in comparison to P1-rr which displays very low levels of DNA methylation [18] . Several p1 alleles that show a P1-rr-like pericarp and cob glume pigmentation phenotype and carry a P1-wr mutlicopy gene structure have been shown to be hypomethylated [21] . Likewise , loss of pericarp and cob glume pigmentation of P1-wr* , a silent epiallele of P1-wr , has been attributed to even denser cytosine methylation than that of P1-wr [20] . Presence of Unstable factor for orange1 ( Ufo1 ) , an un-cloned trans-acting dominant modifier , induces loss of cytosine methylation from P1-wr and P1-wr* , thereby relieving epigenetic suppression and leading to ectopic gain of phlobaphene pigmentation in various plant organs including dried silk , tassel glume , husk , and leaf sheath [20] , [24] , [25] . Ufo1-induced phenotypes show incomplete penetrance ( some progeny carrying the mutation completely lack the mutant phenotype ) and poor expressivity ( the extent of mutant phenotype is variable ) . Paramutation at the p1 locus was observed due to the interactions of the endogenous P1-rr allele with a transgene carrying fragments of the P1-rr regulatory region . The transgene was composed of a 1 . 2-kb P1-rr distal enhancer fragment ( P1 . 2 ) located 5-kb upstream of P1-rr transcription start site , a basal ( b ) p1 promoter fragment ( −236 to +326 untranslated leader ) , GUS coding region , and PinII terminator [26] . When plants with this transgene ( P1 . 2b::GUS ) were crossed with those carrying the P1-rr allele , a subset of the transgenic progeny showed a striking reduction in pericarp and cob glume pigmentation . The silenced state of P1-rr , designated as P1-rr′ , was inherited independent of the transgene , and it could silence the naïve P1-rr allele . Importantly , the silenced , paramutagenic state of P1-rr′ was associated with increased DNA methylation within the P1 . 2 fragment that is required for paramutation [26] . Additional transgenic experiments revealed that the P1 . 2 fragment is required and sufficient for paramutation [27] . The mop1-1 and Mop2-1 mutations disrupt paramutation of P1-rr to P1-rr′ demonstrating that RNA mediated mechanisms are involved in establishment of silencing associated with p1 paramutation [11] , [27] . Maintenance of silencing is less dependent on RNA mediated mechanisms , as up to three consecutive generations of exposure to the mop1-1 mutation were required for P1-rr′ up regulation while Mop2-1 had no effect on P1-rr′ silencing even after three generations of continuous exposure [11] . To further elucidate the mechanism ( s ) underlying epigenetic regulation of paramutation , we characterized involvement of Ufo1 in the regulation of silencing associated with b1 and p1 paramutation . We compared effects of the Ufo1-1 mutation on paramutagenicity and densities of DNA methylation within the P1 . 2 fragment in P1-rr′ and another spontaneous epimutation of P1-rr , P1-prTP ( patterned pericarp and red cob glume ) . This study highlights the role of histone modifications and DNA methylation and relationship between origin , epigenetic state and differential paramutagenic behavior of these epialleles derived from the common progenitor P1-rr allele . Possible mechanisms dictating different effects of Ufo1-1 on paramutation at b1 and p1 loci are discussed .
To test if Ufo1-1 reactivates single-copy silenced p1 alleles , P1-rr′ ufo1 plants were crossed with p1-ww Ufo1-1 ( Figure 1A ) . Of 239 F1 plants screened , 62 ( 26% ) showed gain of pericarp pigmentation while 177 ( 74% ) remained silent as indicated by a colorless pericarp except red pigmentation at the point where silk attaches to pericarp during seed development ( hereafter referred to as silk scarred phenotype ) . Similar to earlier published results [20] , [24] , the effect of Ufo1-1 on P1-rr′ silencing was not fully penetrant as only a subset of the F1 progeny showed gain of pigmentation with pericarp phenotypes varying from uniformly red or orange to red/orange variegation of pericarps . To test the heritability of reactivated P1-rr′ phenotypes , F1 plants showing gain of pericarp pigmentation were crossed with p1-ww[4Co63] ( Figure 1B ) . As expected , approximately half of the progeny ( 54 . 4% ) had colorless pericarp and cob glume specified by the homozygosity for the recessive p1-ww allele ears ( χ2 = 0 . 79; P = 0 . 38 ) . If Ufo1-1 mediated P1-rr′ up regulation were not heritable then 25% of total progeny is expected to carry Ufo1-1 and P1-rr′ and have up regulated pericarp and cob glume pigmentation phenotype , while 25% of progeny carrying P1-rr′ and a wild type ufo1 allele should have silenced silk scarred ears . Results of the analysis demonstrated that 37 . 8% of ears had silk scarred P1-rr′ phenotype and 7 . 8% had up regulated red or orange pericarp and cob glume pigmentation phenotypes . This segregation ratio indicates that the reactivated P1-rr′ state is not heritable and reverts back to silenced state after segregation of Ufo1-1 . The paramutagenic activity of individual P1-rr′ families is highly variable and the suppression of naive P1-rr by P1-rr′ can range from 0 to >90% [26] . We tested if P1-rr′ families with differential paramutagenic activity also differ for their extent of activation by Ufo1 . Representative P1-rr′ plants were crossed with p1-ww Ufo1-1 and each P1-rr′ plant was also crossed with the paramutable naïve P1-rr allele ( Figure 2 ) . Scoring the resultant progeny for frequency of paramutation revealed that P1-rr′ paramutagenicity varied between 40 and 81 . 6% , while reactivation by Ufo1-1 varied between 5 . 2% and 62 . 9% . Interestingly , frequency of reactivation by Ufo1-1 was lower in the progeny of the highly paramutagenic P1-rr′ ( families 3 and 4 ) as compared to the progeny of the low paramutagenic P1-rr′ ( families 1 and 2 ) . Statistical analysis revealed that high paramutagenic ability negatively correlated with frequency of reactivation by Ufo1-1 ( R2 = −0 . 86 ) . This result demonstrates that the mechanism ( s ) governing the paramutagenic activity also interfere with the effect of the Ufo1-1 mutation on P1-rr′ . The silenced state of P1-rr′ is characterized by increased methylation of the P1 . 2 enhancer element [26] . To test if the reactivation of P1-rr′ by Ufo1-1 involved hypomethylation within the P1 . 2 enhancer sequence , p1 fragment 15 was used as a probe to hybridize gel blots carrying leaf genomic DNA digested with methylation sensitive endonuclease HpaII ( Figure 3 ) and with a combination of HpaII and methylation insensitive endonuclease DraI ( data not shown ) . In P1-rr , two major fragments of approximately 1 . 2 and 1 . 1-kb are observed ( Figure 3A ) which originate from unmethylated HpaII sites flanking the fragment 15 in the upstream promoter and downstream of 3′ untranslated region ( UTR ) ( Figure 3B ) . Most of these sites are methylated in P1-rr′ resulting in the loss of 1 . 2 and 1 . 1-kb bands and appearance of higher molecular weight bands indicating increased DNA methylation at these HpaII sites ( Figure 3B ) . Examination of restriction patterns revealed loss of DNA methylation in all reactivated ( R ) and non-activated ( N ) P1-rr′/p1-ww; Ufo1-1/+ plants , as seen from the loss of high ( 8 . 1 and 6 . 5-kb ) molecular weight bands . However , additional loss of DNA methylation from HpaII sites that flank fragment 15 are observed in the R plants which results in the reappearance of 1 . 2 and 1 . 1-kb bands . This result demonstrates that partial loss of DNA methylation occurs in all reactivated P1-rr′/p1-ww; Ufo1-1/+ plants , but greater hypomethylation is observed in plants with up regulated pericarp and cob glume pigmentation phenotype . To assay methylation of individual cytosine residues , we performed genomic bisulfite sequencing on the upper strand of a 443-bp fragment of the P1 . 2 enhancer region , which is required for paramutation . Since P1-rr is expressed in pericarp and Ufo1-1 induced gain of pigmentation is more pronounced in this tissue , we used pericarp DNA for the methylation assay . In the functional P1-rr allele , almost all of the symmetric ( CG and CHG; H is A , T , or C ) , and asymmetric ( CHH ) cytosine sites were un-methylated ( Figure 4 ) . Silencing of P1-rr′ was associated with hypermethylation of most symmetric sites , and to a lesser extent , with that of asymmetric sites ( Figure 4 , Figure S1 ) . Additionally , symmetric cytosine methylation was higher in the 5′ end and it was reduced toward the 3′ end of the P1 . 2 fragment . Analysis of reactivated P1-rr′Ufo1-1 plants revealed that DNA methylation was dramatically reduced at all symmetric and asymmetric cytosines to a level comparable with that of the P1-rr ufo1 plants . Therefore , the DNA gel blot and bisulfite sequencing analyses demonstrated that Ufo1-1-mediated reactivation of P1-rr′ correlates with reduction of methylation within the P1 . 2 enhancer fragment . In an earlier study , Das and Messing [19] described a P1-pr ( patterned pericarp and red cob glume ) epiallele of P1-rr with reduced pericarp pigmentation . P1-pr silencing correlated with increased DNA methylation and failure of the P1 . 2 enhancer [28] to undergo tissue-specific chromatin remodeling in pericarp [19] , [29] . We characterized another independently isolated spontaneous epiallele of P1-rr , P1-prTP , which is phenotypically similar to P1-pr [19] and P1-rr′ [26] with colorless or silk scarred pericarp and light pink to light red cob glume ( Figure 5A ) . The silenced P1-prTP state is very stable as out of ∼1 , 000 plants screened , none showed any spontaneous gain of pericarp pigmentation ( data not shown ) . Extensive DNA gel blot analysis and partial sequencing revealed that the P1-prTP sequence is identical to P1-rr ( not shown ) . To test if P1-prTP silencing is associated with epigenetic modification , we compared DNA methylation of HpaII-digested P1-prTP and P1-rr . In P1-rr , the probe fragment 15 hybridized to five restriction fragments including the 1 . 2 kb and 1 . 1 kb bands ( Figure 5B ) . Loss of these bands and appearance of higher molecular weight bands in P1-prTP showed that the region around the P1 . 2 enhancer was methylated ( Figure 5C ) . Similarly , upon hybridization with the probe fragment 6 , P1-prTP produced higher molecular weight bands as compared to P1-rr indicating methylation within the promoter region . However , P1-prTP and P1-rr had similar methylation levels in the coding region as shown by hybridization with probe fragment 8B ( Figure S2 ) . Taken together , these results demonstrate that P1-prTP is an epiallele of P1-rr and that silencing of P1-prTP is associated with DNA methylation of upstream regulatory regions including the P1 . 2 enhancer . To test if P1-prTP participates in paramutation , it was crossed with P1-rr and resulting F1 progenies were scored for pericarp pigmentation . All 300 F1 plants screened showed red pericarp and red cob glume phenotype indicating that P1-prTP is non-paramutagenic to P1-rr and behaves as a recessive allele . Moreover , in the F2 generation , P1-rr and P1-prTP phenotypes segregated in the expected 3∶1 ratio further supporting that P1-prTP does not participate in paramutation ( data not shown ) . To summarize , genetic data indicated that P1-prTP is not paramutagenic to a naive P1-rr . To test if the Ufo1-1 mutation can disrupt the silenced state of P1-prTP , genetic crosses were performed ( Materials and Methods , Figure 6A ) . Analysis of 155 F1 plants revealed that P1-prTP was upregulated by the Ufo1-1 mutation and , similar to P1-rr′ , frequency of up regulation was low as only 21% of plants exhibited gain in pericarp and cob glume pigmentation ( Figure 6A ) . Further , the F2 progeny of Ufo1-1 expressing plants did not produce the expected phenotypic ratio of 9∶3∶3∶1 ( data not shown ) . To assay whether Ufo1-1-mediated reactivation of P1-prTP was heritable , F1 plants with an increased pericarp and cob glume pigmentation were crossed with colorless p1-ww[4Co63] plants carrying a wild type ufo1 allele ( Figure 6B ) . As expected , 50% of the progeny was homozygous for p1-ww and therefore produced colorless and pericarp and cob glume ( χ2 = 1 . 25; P = 0 . 263 ) . If the activated state of P1-prTP returned to a silenced state after segregation of Ufo1-1 , 25% of progeny was expected to have a silenced phenotype , while the remaining 25% , still carrying Ufo1-1 , would have increased pigmentation . There were 28 . 2% of silk scarred and 17% of red/variegated pericarp and cob glume individuals indicating that , similar to P1-rr′ , P1-prTP up regulation was not heritable in the absence of Ufo1-1 . Overall , these observations demonstrate that Ufo1-1 temporarily disrupts the silenced epigenetic state of P1-prTP leading to increased pericarp pigmentation . Non-heritable P1-prTP gain of pigmentation in the mutant Ufo1-1 background is reminiscent of Ufo1-1 interaction with other p1 alleles [20] , [24] . Gel blot analysis involving HpaII-digested leaf genomic DNA hybridized with p1 probe fragments did not detect any methylation differences between the silenced and reactivated P1-prTP plants ( Figure S2 ) . We also performed genomic bisulfite sequencing of the 443 bp of the P1 . 2 enhancer in pericarp of P1-prTP ufo1 and P1-prTP/p1-ww; Ufo1-1/ufo1 plants . In comparison to P1-rr′ , which is hypermethylated in symmetric and asymmetric contexts , P1-prTP is hypermethylated only in symmetric contexts ( Figure 4 ) . No decrease in DNA methylation was observed in pericarps of P1-prTP/p1-ww plants that were strongly up regulated by Ufo1-1 . These results are in striking contrast with P1-rr′ where Ufo1-1-up regulation correlated with dramatic reduction of cytosine methylation and demonstrates that similar phenotypes of P1-rr′ and P1-prTP are specified by distinct molecular mechanisms . Histone 3 lysine 9 dimethylation ( H3K9me2 ) is a histone modification associated with heterochromatin assembly and transcriptional silencing . To test if histone modifications are involved in the silencing of P1-rr′ and P1-prTP , chromatin immunoprecipitation ( ChIP ) followed by quantitative real-time PCR ( qPCR ) was performed to determine the H3K9me2 enrichments at the P1 . 2 region . The ChIP assays showed that the chromatin encompassing the 1 . 2 kb distal enhancer is significantly enriched for H3K9me2 in P1-rr′ and P1-prTP plants as compared to P1-rr ( Figure 7 ) . Thus , irrespective of their involvement in paramutation , silenced state of both alleles is associated with enrichment of the suppressive H3K9me2 mark . To further investigate if ufo1 is involved in maintenance of H3K9me2 , ChIP assays were performed at P1-rr′ and P1-prTP in the presence of Ufo1-1 . Interestingly , in the presence of Ufo1-1 , there is a dramatic reduction of H3K9me2 within the 1 . 2 kb distal enhancer region of P1-rr′ and P1-prTP ( Figure 7 ) . To summarize , these results demonstrate that ufo1 plays a role in maintaining repressive H3K9me2 histone marks at P1-rr′ and P1-prTP . Paramutation at the b1 locus occurs when the highly expressing , darkly pigmented B-I allele is exposed to a low expressing , lightly pigmented B′ allele in heterozygote [4] . While paramutable B-I is unstable and spontaneously reverts to B′ at variable frequencies , paramutagenic B′ is stable in wild type genetic backgrounds . To test if Ufo1-1-mediated disruption of epigenetic silencing associated with paramutation extends beyond the p1 locus , the silenced B′ allele was introgressed into the Ufo1-1 background and B′-specified plant pigmentation was evaluated ( Figure S3 ) . Examination of the segregating progeny revealed that B′-specified pigmentation increased in the mutant ( Ufo1-1/ufo1 ) , as evident from multiple , wide and darkly pigmented sectors on sheaths ( Figure 8A ) , husks , and tassels ( not shown ) . Extent of B′ pigmentation in the Ufo1-1 background was moderate and never reached solid dark purple plant pigment observed in B-I or dark pigmentation phenotypes observed for B′ in mop1-1 [8] , [9] and mop2-1 [11] mutant backgrounds . The Ufo1-1 mutation was penetrant in about half of F1 plants , and this number , but not intensity of B′ pigmentation , increased in later backcrosses ( BC2 and BC3 ) ( Figure 8B ) . Overall , these results show that the Ufo1-1 mutation disrupts the silenced state of B′ allele and therefore ufo1 is involved in the pathway that regulates silencing associated with paramutation at multiple loci .
Given the fact that P1-rr′ and P1-prTP are epialleles of a common progenitor allele ( P1-rr ) , their differential paramutagenic ability is intriguing; while P1-rr′ is able to communicate its chromatin state in trans to P1-rr , P1-prTP fails to do so . A notable difference between these epialleles is the nature of their origin: P1-rr′ silencing was induced by a transgene carrying P1 . 2 distal enhancer sequence , while P1-prTP originated spontaneously and causal factors for its origin are unknown . Establishment of transgene-mediated silencing at P1-rr′ requires RNA-dependent RNA polymerase activity of mop1 [27] and plant-specific RNA polymerase IV/V activity as evidenced by requirement for mop2 [11] implicating RNA mediated mechanism in the origin of this epiallele . The extent and distribution of cytosine methylation in the distal enhancer region of P1-rr′ and P1-prTP are strikingly different . In P1-rr′ , this region is methylated in all sequence contexts ( CG , CHG , and CHH ) , and CG and CHG methylation levels are higher in the 5′ and decrease toward the 3′ end , while CHH methylation is evenly distributed throughout the assayed region . These high and low methylation regions correspond , respectively , to the middle and the end of the sequence homology between the endogenous P1 . 2 enhancer and the transgene fragment that caused the initial silencing . This pattern of symmetric cytosine methylation and increased levels of asymmetric CHH methylation further points to the role of RNA-mediated mechanisms in P1-rr′ silencing . In P1-prTP , however , only CG and CHG sites are methylated and the methylation is uniformly high throughout the assayed region . Furthermore , while P1-rr′ can , on rare occasions , show spontaneous gain of function [26] , silenced state of P1-prTP is very stable . While it is possible that initial events leading to the silencing of P1-prTP involved RNA mediated mechanism , based on a lack of asymmetric DNA methylation , RNA signals do not appear to contribute to the maintenance of DNA methylation . In summary , difference in the levels of CHH methylation , a hallmark of RdDM , seems to be the only epigenetic mark that correlates with differential paramutagenic ability of the two alleles . While the paramutagenic ( P1-rr′ ) and non-paramutagenic ( P1-prTP ) alleles differ for their cytosine methylation patterns within the P1 . 2 enhancer , both display enrichment in H3K9me2 . Furthermore , while symmetric methylation persists at the enhancer in Ufo1-reactivated P1-prTP , levels of H3K9me2 are decreased in both reactivated alleles . Thus , H3K9me2 appears to be an indispensable repressive epigenetic mark for maintaining silencing at the paramutagenic and non-paramutagenic p1 epialleles . However , we cannot rule out that , both H3K9me2 and DNA methylation play important and mutually reinforcing roles in maintenance of silencing at both P1-prTP and P1-rr′ . In a similar study at the b1 locus , tandem repeats critical for paramutation exhibited tissue-independent DNA methylation , while enrichment in H3K9me2 and H3K27me2 was tissue-specific [30] . Based on these results , authors concluded that H3K9me2 does not play a role in the mitotic heritability of the silenced B′ state , but rather serves to reinforce silencing in a tissue-specific manner [30] . In our study , H3K9me2 enrichment was observed in pericarp tissue of P1-rr′ and P1-prTP pericarps , and loss of silencing in Ufo1-1 correlated with loss of H3K9me2 . However , it remains unclear whether this mark is involved in tissue-specific regulation and reinforcement of silencing , or it also plays a role in tissue-independent maintenance of silencing . Several studies have reported a positive correlation between cytosine , especially CHG , methylation and H3K9me2 marks [31]-[34] . In Arabidopsis , loss of KRYPTONITE , an H3K9-specific methyltransferase , results in a loss of both H3K9me2 and CHG methylation [35] and a genome-wide survey of H3K9me2 and CHG methylation has shown very high correlation between the two epigenetic marks [32] . In our study , CHG methylation and not H3K9me2 persists in the enhancer of reactivated P1-prTP Ufo1-1 plants indicating that H3K9me2 and CHG methylation may exist independent of each other . However , our data does not exclude the possibility that CHG methylation is crucial for establishment of silencing in P1-PrTP . The paramutagenic ability of P1-rr′ is highly variable; silencing of a naive P1-rr allele by P1-rr′ in independent families varied between 0–95% in the current and an earlier study [26] . We demonstrate that the paramutagenic ability of P1-rr′ was inversely correlated with Ufo1-1-induced reactivation . Thus , highly paramutagenic P1-rr′ stocks interfere with Ufo1-1-mediated reactivation in a manner not currently understood . Variable levels of p1 alleles reactivation have been attributed to incomplete penetrance of Ufo1-1 , and to the extent of epigenetic silencing of the p1 allele involved [20] , [24] . Ufo1-1 induces pericarp and cob glume pigmentation in moderately methylated P1-wr plants , but only cob glume pigmentation in a highly methylated P1-wr* . Interestingly , repeated back crosses of P1-wr* Ufo1-1 with p1-ww Ufo1-1 stock eventually leads to a gain of pericarp pigmentation ( R . Sekhon , P . Wang and S . Chopra , unpublished data ) . Thus , the highly silenced epigenetic states ( P1-wr* and highly paramutagenic P1-rr′ families ) may not show immediate gain of pericarp pigmentation in the presence of Ufo1-1 while moderately silenced states ( P1-wr ) can be readily perturbed . The Ufo1-1 mutant allele perturbs the organ-specific expression patterns of the multicopy p1 alleles P1-wr and P1-wr* [20] , [24] . These alleles do not participate in paramutation [20] , [26] . Our finding that presence of Ufo1-1 leads to reactivation of the paramutagenic P1-rr′ and B′ alleles indicates that the wild type factor is involved in maintenance of silencing imposed by paramutation . While the absence of RNA-dependent RNA polymerase MOP1 also leads to reactivation of silenced P1-rr′ and a gain of pericarp pigmentation in the P1-wr allele , several generations of absence of MOP1 are required for such activation [27] . In contrast , lack of ufo1 for one generation is sufficient to abolish silencing at the p1 locus . On the other hand , lack of ufo1 leads only to a partial reactivation of B′ while lack of MOP1 results in immediate disruption of silencing . It thus appears that epigenetic suppression at these two loci is mediated by distinct but overlapping pathways . Reactivation of the silenced loci in the absence of ufo1 suggests that it is directly involved in the epigenetic mechanism ( s ) responsible for the silencing . Ufo1-1-mediated reactivation of P1-rr′ and also in P1-wr [20] , [24] is associated with the loss of DNA methylation within regulatory regions whereas the de-repression of P1-prTP does not seem to involve any methylation modifications at the region tested . These results support our argument [20] that ufo1 may not be directly involved in establishing and/or maintaining DNA methylation . Given that reactivation of both P1-rr′ and P1-prTP in the Ufo1-1 background was associated with a loss of H3K9me2 marks , it appears that ufo1 is involved in maintaining these heterochromatic marks . Future studies to examine the role of Ufo1-1 in establishment and maintenance of silencing associated with paramutagenic and non-paramutagenic systems , and cloning of the gene will provide insights into ufo1-dependent mechanisms in epigenetic regulation of gene expression .
The maize P1-rr allele used in this study is derived from the P1-rr-4B2 genetic stock [36] . Origin of the P1-rr′ stock has been previously described [26] . The P1-rr′ allele used in this study was progeny of a homozygous ( P1-rr′/P1-rr′ ) plant that showed strong silencing and had colorless pericarp and light pink cob glume . The P1-prTP , a spontaneous epiallele of P1-rr , has been previously reported [26] and this epiallele is distinct from the previously characterized P1-pr allele [19] . A stock carrying Ufo1-1 has been described previously [24] . The Ufo1-1 was introgressed into the inbred line 4Co63 ( National Seed Storage Laboratory , Fort Collins , CO ) , which carries a null p1-ww allele . Since the p1-ww Ufo1-1 plants do not produce phlobaphene pigmentation , presence of Ufo1-1 was tested by crossing individual plants with P1-wr[W23] [17] . Ectopic gain of pigmentation in pericarp and other organs in the resulting F1 progeny confirmed the presence of Ufo1-1 in the p1-ww Ufo1-1 stock [24] . The stock carrying B′ allele of the b1 gene was obtained from E . H . Coe , Jr . ( University of Missouri , Columbia ) and this stock carries the Pl-sr allele of the pl gene that does not impart b1-specified pigmentation to the plant body . All plant stocks used in this study carry functional alleles for the structural genes required for anthocyanin and/or phlobaphene biosynthesis . Genomic DNA was isolated from the fifth or sixth leaf using CTAB method [37] . Genomic DNA was digested to completion using restriction enzymes , reagents , and protocols from Promega ( Madison , WI ) . DNA gel blot was performed as described previously [20] . DNA fragments of p1 used as probes in DNA gel blot analysis have been described previously [36] , [38] . The upper DNA strand of a 443-bp sub-fragment of the P1 . 2 fragment required for paramutation , was assayed by genomic bisulfite sequencing . Pericarp tissues were collected 18 days after pollination ( DAP ) from individual plants and genomic DNA was extracted using modified CTAB method [37] . For each genotype , DNA from two plants was subjected to genomic bisulfite sequencing and pooled results from the assay are presented . Eight micrograms of DNA was completely digested with suitable restriction enzymes that cut outside the amplified fragment of interest . The digested DNA was treated with sodium bisulfite following a previously published protocol [39] with modifications [20] . The promoter region was amplified using nested PCR primers [25] . The resulting PCR products were gel purified , cloned using a TOPO TA cloning kit ( Invitrogen , Carlsbad , CA ) , and 20–30 clones/ligation/genotype were sequenced . Due to the complex P1-rr locus structure , the region studied by bisulfite sequencing is present in three places ( see Figure 3B for details ) . The sequenced region is also present in p1-ww[4Co63] albeit with multiple indels and single nucleotide polymorphisms ( R . S . Sekhon and S . Chopra , unpublished ) . These sequence differences were used to omit the clones that originated from the null p1-ww allele . Percent methylation at each cytosine residue was calculated by dividing the number of clones methylated for the residue by the total number of clones for that residue for all of the amplified P1-rr′ and P1-prTP regions . ChIP assays were performed using pericarp tissues following a modified protocol as described previously [40] , [41] . Briefly , pericarp tissues were harvested at 18 DAP and cross-linked with 3% formaldehyde . The chromatin complex was then extracted and sheared to a size range of 0 . 5 to 1 kb fragments using a Bioruptor ( Diagenode , Denville , NJ ) . The anti-H3K9me2 antibody used for ChIP was kindly provided by Dr . Hiroshi Kimura [41] . This antibody was coupled with sheep anti-mouse IgG Dynabeads M-280 ( Invitrogen , Grand Island , NY ) and incubated with the sheared chromatin . A normal mouse IgG was used as no antibody control ( NoAb ) . The ChIPed DNA was further purified using QIAquick PCR Purification Kit ( QIAGEN , Valencia , CA ) and quantified with qPCR . The relative enrichment of H3K9me2 modification was normalized to the input DNA loaded in the ChIP reaction as described previously [40] . The primers specific to the P1 . 2 kb distal enhancer region used in this study are PW_RTF15-2F ( 5′-GACGTCTCACCGGCTCACA-3′ ) and PW_RTF15-2R ( 5′-ATGCAACGCAACGCTTTG-3′ ) . The relative differences between ChIP assay and input sample were determined using the percentage-of-input method ( see ChIP analysis; http://www . invitrogen . com/site/us/en/home/Products-and-Services/Applications/RNAi-Epigenetics-and-Gene-Regulation/Chromatin-Remodeling/Chromatin-Immunoprecipitation-ChIP/chip-analysis . html ) . Data shown in this study are representative result of three independent experiments . | Natural allelic variability is crucial for genetic improvement . While the genetic mechanisms leading to such variation have been studied in depth , relatively less is known about the role of epigenetic mechanisms in generation of allelic diversity . Paramutation is a phenomenon in which one allele can silence another allele in trans and , once established , such epigenetic silencing is heritable . To further understand the molecular components of paramutation , we characterized two epialleles of the pericarp color1 ( p1 ) gene of maize , which originated from a common progenitor; however , only one of these alleles is paramutagenic . Results show that , while both alleles have high levels of symmetric ( CG and CHG ) methylation in a distal enhancer element , only the paramutagenic allele has higher levels of asymmetric ( CHH ) methylation . Since CHH methylation is imposed and maintained through RNA–mediated mechanisms , these results indicate that paramutation at the p1 locus involves RNA–mediated silencing pathway . Further , both silent epialleles are reactivated in the presence of an unlinked dominant mutation Ufo1-1 , and reactivation is accompanied by the loss of suppressive histone mark H3K9me2 . Finally , we show that ufo1 is also required for epigenetic silencing at the booster1 locus and thus affects additional loci in maize that participate in paramutation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"plant",
"science",
"plant",
"biology",
"genetics",
"biology",
"genetics",
"and",
"genomics"
] | 2012 | Maize Unstable factor for orange1 Is Required for Maintaining Silencing Associated with Paramutation at the pericarp color1 and booster1 Loci |
Tuberculosis is a current major world-health problem , exacerbated by the causative pathogen , Mycobacterium tuberculosis ( Mtb ) , becoming increasingly resistant to conventional antibiotic treatment . Mtb is able to counteract the bactericidal mechanisms of leukocytes to survive intracellularly and develop a niche permissive for proliferation and dissemination . Understanding of the pathogenesis of mycobacterial infections such as tuberculosis ( TB ) remains limited , especially for early infection and for reactivation of latent infection . Signaling via hypoxia inducible factor α ( HIF-α ) transcription factors has previously been implicated in leukocyte activation and host defence . We have previously shown that hypoxic signaling via stabilization of Hif-1α prolongs the functionality of leukocytes in the innate immune response to injury . We sought to manipulate Hif-α signaling in a well-established Mycobacterium marinum ( Mm ) zebrafish model of TB to investigate effects on the host's ability to combat mycobacterial infection . Stabilization of host Hif-1α , both pharmacologically and genetically , at early stages of Mm infection was able to reduce the bacterial burden of infected larvae . Increasing Hif-1α signaling enhanced levels of reactive nitrogen species ( RNS ) in neutrophils prior to infection and was able to reduce larval mycobacterial burden . Conversely , decreasing Hif-2α signaling enhanced RNS levels and reduced bacterial burden , demonstrating that Hif-1α and Hif-2α have opposing effects on host susceptibility to mycobacterial infection . The antimicrobial effect of Hif-1α stabilization , and Hif-2α reduction , were demonstrated to be dependent on inducible nitric oxide synthase ( iNOS ) signaling at early stages of infection . Our findings indicate that induction of leukocyte iNOS by stabilizing Hif-1α , or reducing Hif-2α , aids the host during early stages of Mm infection . Stabilization of Hif-1α therefore represents a potential target for therapeutic intervention against tuberculosis .
Pulmonary tuberculosis ( TB ) , caused by the pathogen Mycobacterium tuberculosis ( Mtb ) , is a major world health problem and is a key priority for infectious disease research . The burden of TB has been exacerbated by the increasing occurrence of Mtb strains with resistance to multiple drug treatments , prioritising the need for understanding of the mechanistic basis of host-pathogen interactions during pathogenesis of disease in order to identify novel therapeutic strategies [1] . Upon infection Mtb are rapidly phagocytosed by host leukocytes , but are able to evade bacterial killing mechanisms and utilize the leukocytes as a niche in which to proliferate and disseminate [2] . Leukocyte infection initiates the recruitment of uninfected macrophages , neutrophils and T-cells , to form highly organised structures known as granulomas [3] , [4] . Mtb within granulomas can persist for many years and may eventually escape and disseminate during clinical reactivation , causing active disease [5] . The pathogenesis of both initial infection and reactivation of latent infection are not well understood , and further research into host signaling pathways at these stages may uncover novel , host-derived targets for therapeutic intervention against Mtb . Mycobacterial disease and hypoxia are intimately related . Human tuberculous granulomas are hypoxic environments , and it has been suggested that the relative hypoxia of granulomas contributes to the latent infection phenotype and the associated relative resistance of Mtb to host and pharmacological killing [6] , [7] . Hypoxia exerts its effects on host cell signaling predominantly through stabilization of Hypoxia Inducible Factor alpha ( HIF-α ) transcription factor . HIF-α is stability and activity is regulated by a group of oxygen sensitive enzymes: prolyl hydroxylases ( PHDs ) and Factor Inhibiting HIF ( FIH ) [8]–[10] . Oxygen dependent PHD activity leads to degradation of HIF-α , while hypoxia reduces PHD activity , stabilizing HIF-α , which joins a nuclear complex and transduces the hypoxic cellular response [11] . Three HIF-α isoforms have been identified in humans to date , of which HIF-1α is a key regulator of leukocyte function during both inflammation and a range of bacterial infections [12]–[15] . Normal host defense is dependent on HIF-1α expression , which activates and enhances leukocyte functionality [12] . We have previously shown in a zebrafish model of inflammation that stabilized Hif-1α delays inflammation resolution by reducing neutrophil apoptosis and reverse migration at the inflammation site [16] . Existing evidence suggests that the successful clearance of bacterial infections depends on normal HIF-α signaling , and furthermore , immune cell HIF-α is activated by bacterial challenge in normal oxygen levels , demonstrating the fundamental importance of this pathway to immune cell response to invading pathogens [13] , [17] . Despite the extensive work done on the effects of hypoxia on Mtb phenotype , the effects of HIF-α stabilization or downregulation in determining the outcome of host-mycobacterial interaction remains unknown , and presents a major research challenge requiring a combination of modern cell biology and genetic approaches in animal models . The zebrafish is a well-established model organism used to study a wide variety of human diseases [18] . Zebrafish embryos are easily manipulated genetically and their translucency allows for detailed microscopy studies . Mycobacterium marinum ( Mm ) , a natural fish pathogen and a close relative of Mtb , causes an infection in zebrafish that mimics key features of human TB , including the formation of caseating granulomas and development of latency [19] , [20] . Mm infection of zebrafish embryos has been successfully used to understand both host cell signaling and mycobacterial virulence determinants [21]–[24] . The Hif-α pathway can be manipulated in vivo in the zebrafish , both pharmacologically , using non-specific inhibitors of PHD enzymes such as dimethyloxaloylglycine ( DMOG ) , and genetically , by expression of dominant Hif-1α variants [16] , [25] . Using the zebrafish Mm model and Hif-α manipulation we aimed to understand the relationship between Hif-α signaling and the outcome of mycobacterial infection . We show that Hif isoforms , Hif-1α and Hif-2α have opposing effects on the host susceptibility to mycobacterial infection by demonstrating that stabilization of Hif-1α and downregulation of Hif-2α signaling decreases bacterial burden of Mm infection . Furthermore , using both genetic and pharmaceutical approaches , we show that this effect acts via a nitric oxide ( NO ) dependent mechanism . Our findings identify both Hif-α and NO signaling components as potential host-derived targets for therapeutic intervention against TB .
HIF-α signaling and its role during mycobacterial infection remain unclear . Upon Mm infection , zebrafish larvae develop early stage granulomas within several days , but levels of Hif-α signaling in early infection and larval granulomas are unknown . To detect levels of Hif-α signaling in Mm infection we utilized in situ hybridization for the known Hif-α target phd3 ( Figure S1A and S1B ) and the transgenic line Tg ( phd3:GFP ) i144 ( Figure 1A and 1B ) [16] , [25] . At 1 day post infection ( dpi ) phd3:GFP expression was observed in infected leukocytes ( Figure 1A ) . Larval granulomas at 6dpi showed only very low levels of phd3:GFP expression and are therefore unlikely to have stabilized Hif-α at this later stage ( Figure 1B and S1B ) . By crossing the Tg ( phd3:GFP ) i144 line with a line marking macrophages with a membrane targeted mCherry we showed that the phd3:GFP expression was found in infected macrophages at 1 dpi ( Figure 1C ) . This macrophage-specific upregulation of phd3:GFP expression by Mm infection was blocked by injection of RNA for dominant negative ( DN ) hif-1αb , indicating this is a Hif-1α dependent host response to Mm infection ( Figure 1D ) . Stabilization of Hif-α by DMOG treatment between 5 and 6 dpi did not reduce the bacterial burden at 6 dpi ( Figure 2A and 2B ) , suggesting that it is specifically early Hif-α stabilization which may represent a novel cellular defense mechanism to arm the leukocytes and cause Mm killing . Using a combination of pharmacological and genetic approaches to upregulate Hif-1α signaling we aimed to determine the effects on pathogenesis in the zebrafish Mm model . Treatment with DMOG from 4 hours pre-infection to 24 hours post infection significantly decreased bacterial burden compared to DMSO negative control embryos , assessed at 4 dpi by fluorescent imaging and pixel count analysis ( Figure 2C and 2D ) . This is consistent with stabilized Hif-1α aiding the host to combat Mm infection . To eliminate the possibility that DMOG affects bacterial growth we performed an in vitro assay for 24 hours of treatment , which showed that DMOG does not affect bacterial growth in culture , assessed by both OD600 reading and by plating to check viability ( Figure S2A and S2B ) . Hif-1α signaling was manipulated genetically by injecting RNA for dominant active ( DA ) and dominant negative ( DN ) hif-1αb variants [16] , [26] , [27] . Due to a genome duplication event in zebrafish there are two homologues of human HIF-1α , hif-1αa ( ZFIN: hif1αa , NCBI Reference Sequence: XM_001337574 . 2 ) and hif-1αb ( ZFIN: hif1αb , NCBI Reference Sequence: NM_200233 . 1 ) [28] . hif-1αb aligns more closely to the human sequence than hif-1αa , and contains all three highly conserved hydroxyl-regulated amino acids , while hif-1αa only has two . We have previously reported that dominant hif-1αb constructs are able to modulate Hif-1α signaling and that hif-1αb is the homologue responsible for the delay in resolution of inflammation observed when Hif-1α signaling is stabilized [16] , [25] . Therefore all dominant constructs used in this study are based on the hif-1αb homologue . Injection of DA hif-1αb RNA into one-cell stage embryos , followed by injection of Mm at 1 day post fertilization ( dpf ) , caused a decrease in bacterial burden at 4 dpi , when compared to both phenol red ( PR ) injected controls and DN hif-1αb injected embryos ( Figure 2E and 2F ) . The reduction in bacterial burden was comparable to that seen in DMOG treatment embryos ( Figure 2C and 2D ) . Furthermore , injection of DN hif-1αb blocked the anti-mycobacterial effect of DMOG without affecting the dynamics of infection in untreated wild-type fish ( Figure 2G ) . Leukocyte numbers were unaffected by dominant hif-1αb constructs at the timepoints of Mm injection ( 30 hpf ) and imaging ( 120 hpf , Figure 2H and 2I ) . These data indicate that Hif-1α stabilization aids the host in clearing Mm infection independently of a change in leukocyte number . Although bacterial burden was decreased by stabilized Hif-1α , general granuloma structure was morphologically normal in DA hif-1αb injected larvae compared to controls ( Figure 2I ) . The observation that granulomas are able to form in the presence of stabilized Hif-1 α suggests that in these fish , changes in pathogenesis occur before the appearance of larval granulomas . These data show that overexpression of stabilized Hif-1α pre-infection and during initial infection reduces the host's susceptibility to Mm , therefore we focused our further investigations on early stage Mm pathogenesis . Once phagocytosed by leukocytes , pathogens are subjected to a number of bactericidal processes , including exposure to reactive nitrogen species ( RNS ) [29] . Due to the highly reactive nature of RNS , direct measurement of production has previously proved problematic . To overcome this an anti-nitrotyrosine antibody has been used in other fish models [30] , and we have previously shown that this antibody can be used in wholemount zebrafish embryos after infection [31] . The advantage of measuring nitrosylation of tyrosine is that this is a stable protein formation , therefore detecting historical NO ( nitric oxide ) production . In the absence of infection , we found that anti-nitrotyrosine staining was specific to leukocytes at 2 dpf . As with other fish species , staining was found mainly in neutrophils , but also in macrophages ( Figure 3A ) . Nitrosylation levels are higher in neutrophils , partly due to a contribution of neutrophil-specific myeloperoxidase , which is able to nitrosylate tyrosine residues in the absence of peroxynitrite [30] . In the absence of infection , nitrosylation levels at 2 dpf were significantly higher after DA hif-1αb injection , compared to control and DN hif-1αb injected embryos ( Figure 3A and 3B ) , suggesting that DA hif-1α leukocytes are already in an activated state prior to infection . In control embryos , the presence of Mm infection at 1 dpi ( the equivalent timepoint of 2 dpf ) increased nitrosylation levels in leukocytes ( Figure 3A and 3C ) . Nitrosylation levels of DA hif-1αb leukocytes remain high in infected embryos compared to uninfected controls , but are lower than DA hif-1αb uninfected equivalents ( Figure 3A and 3C ) . Systemic stabilization of Hif-1α by DA hif-1αb RNA injection increased nitrosylation levels in neutrophils ( Figure 3A ) . In order to test the cell autonomy of the increase in nitrosylation observed after stabilization of Hif-1α , we transiently injected DNA constructs using neutrophil and macrophage specific promoters to express DA hif-1αb in mosaic . To express DA hif-1αb in specific leukocyte lineages a lysozyme C ( lyz ) promoter was used to drive expression in neutrophils ( Tg ( lyz:da-hif-1αb , IRES-nlsGFP ) construct , subsequently referred to as lyz:da-hif-1αb ) , and an mpeg1 promoter was used to drive expression in macrophages ( Tg ( mpeg1:da-hif-1αb , IRES-nlsGFP ) construct , subsequently referred to as mpeg:da-hif-1αb ) [32] , [33] . Both constructs contained a nuclear-localized GFP to visualize the mosaic expression in neutrophils or macrophages , expressed separately from the hif-1αb via an internal ribosome entry site ( IRES ) . When these constructs were injected into neutrophil or macrophage reporter lines expected leukocyte-specific , mosaic expression was observed ( Figure 4A , S3A , and S3B ) . When lyz:da-hif-1αb DNA was injected into Tg ( lyz:DsRED2 ) nz50 embryos or wildtype embryos ( with neutrophils labeled with a TSA post-mortem stain ) , GFP positive neutrophils had higher levels of nitrotyrosine compared to GFP negative neutrophils in the vicinity ( Figure 4A–C ) . These data support a link between stabilized Hif-1α and increased tyrosine nitrosylation in neutrophils . Although mosaic expressed mpeg:da-hif-1αb localized specifically to macrophages ( Figure S3B ) , GFP expression did not correlate to an increase in anti-nitrotyrosine . We previously observed that neutrophils represent the majority of cells with anti-nitrotyrosine labeling , although some mpx:GFP negative cells were labeled ( Figure 3A , white arrowhead ) . We confirmed that there are a proportion of macrophages labeled with anti-nitrotyrosine in the wildtype infected and uninfected scenario , although the frequency of labeling was much lower than observed for neutrophils ( Figure S3C ) . These data indicate that neutrophils are the main cell type that have tyrosine nitrosylation , and that this nitrosylation can be increased by stabilizing Hif-1α specifically in neutrophils . RNS are produced by the activity of the nitric oxide synthase ( NOS ) enzymes . There are three characterized forms of NOS , namely endothelial-NOS ( eNOS ) , neural-NOS ( nNOS ) , and the leukocyte specific inducible-NOS ( iNOS ) . iNOS expression has been shown to be increased in infected leukocytes , and is present in zebrafish leukocytes [34] . Increase of nitrosylation levels in neutrophils after infection is likely to be due to increased iNOS as the morpholino against nos2a , the zebrafish gene for iNOS [34] , was able to abrogate the increase in nitrotyrosine levels following infection ( Figure 3D ) . To confirm that the increase in nitrosylation observed after DA hif-1αb is due to iNOS we used a biochemical probe for NO and an antibody stain for iNOS . DAF-FM DA is a probe that measures NO levels directly [35] . In the absence of Mm , DAF-FM DA staining was increased after RNA injection of DA hif-1αb compared to controls ( Figure 5A ) . The signal of DAF-FM DA staining increased after Mm infection , indicating iNOS activity ( Figure 5B ) . Both the increasing effect on DAF-FM DA of DA hif-1αb in the absence of infection , and the increase after infection in controls , could be partially blocked using the nos2a morpholino , illustrating the iNOS specificity of these effects ( Figure 5A and 5B ) . Furthermore , using an iNOS antibody [36] , we were able to detect higher levels of iNOS protein in the DA hif-αb neutrophils in non-infected embryos , which were present only at low levels in controls ( Figure 5C ) . To assess the effect of blocking iNOS activity at the 1 dpi stage on bacterial burden at 4 dpi , we treated early infected embryos with NOS inhibitors . NOS activity was inhibited using the pan-NOS inhibitor L-NG-Nitroarginine methyl ester ( L-NAME ) and the iNOS specific inhibitor N6- ( 1-iminoethyl ) -L-lysine ( L-NIL ) at early stages of Mm infection [34] , [37] , [38] . Bacterial burden at 4 dpi was not significantly affected by either treatment in control embryos , although there was a trend towards increased infection levels after NOS inhibition ( Figure 5D and 5E ) . Both inhibitors were able to block the decreasing bacterial burden effect of DA hif-1αb at 4 dpi ( Figure 5D and 5E ) . Morpholino knockdown of nos2a was also able to block the decreased bacterial burden in DA hif-1αb injected embryos compared to PR injected controls ( Figure 5F ) . These data confirm that the positive effect of stabilized Hif-1α on the host to combat Mm infection is dependent on iNOS . In humans there are three different HIF-α transcription factors: HIF-1α , HIF-2α , and HIF-3α . It is becoming clear that HIF-2α is important in leukocyte biology [39] , [40] . In order to investigate the effects of hif-2α modulation , we synthesized dominant variants for hif-2α with the equivalent hydroxylation site mutations to the hif-1α variants [16] . As with Hif-1α , Hif-2α has two homologues in the zebrafish , hif-2αa ( ZFIN: epas1a , NCBI Reference Sequence: XM_690170 . 5 ) and hif-2αb ( ZFIN: epas1b , GenBank:DQ375242 ) . Unlike hif-1α homologues , hif-2α sequences are highly similar and both contain all three conserved hydroxylation sites . DA hif-2αa increased phd3 expression by in situ hybridization and phd3:GFP expression at 1 dpf , whilst DN hif-2αa decreased expression levels ( Figure S4A , B ) . DN hif-2αa also blocked the increase in phd3 expression associated with early Mm infection at the 1 dpi stage of pathogenesis ( Figure S4C ) . These data illustrate that zebrafish Hif-2α has similar effects on a well characterized target of the Hif-α transcription factor , phd3 , as the previously characterized zebrafish Hif-1α [16] , [25] . DA hif-2αa , while able to increase phd3 expression , had no effect on bacterial burden compared to controls ( Figure 6A and Figure 6B ) . DN hif-2αa reduced bacterial burden at 4 dpi to a similar level to that of DA hif-1αb ( Figure 6C and 6D ) . As was the case with dominant hif-1αb constructs , dominant hif-2αa variants had no effect on leukocyte numbers ( Figure 6E and 6F ) . Nitrosylation levels in DN hif-2αa were found to be high in non-infected leukocytes , as with DA hif-1αb , with no effect of DA hif-2αa ( Figure 7A and 7B ) . Mm infection caused a decrease in nitrosylation levels in DN hif-2αa leukocyte , as with DA hif-1αb ( Figure 7C ) . The reduction in bacterial burden seen with DN hif-2αa could be blocked by inhibition of iNOS ( Figure 7D–7F ) indicating that the reciprocal effects of Hif-1α and Hif-2α modulate susceptibility to mycobacterial infection via an iNOS dependent mechanism ( Figure 7G ) .
The rise in prevalence of multi-drug resistant TB creates an urgent need for novel , host-targeting therapies to complement existing antibiotics and to combat currently untreatable strains of Mtb [41] , [42] . Using a well-established zebrafish/Mm infection model of TB , we have identified a new anti-mycobacterial leukocyte phenotype being driven by Hif-1α stabilization and consequent iNOS activity . Our data identify that in vivo manipulation of Hif-α , a ubiquitous host-signaling pathway , can affect specific cellular mechanisms of pathogen handling , tipping the host-pathogen balance in favour of the host to decrease mycobacterial infection . Mtb infection and Hif-α signaling have only previously been linked in the context of the hypoxic , necrotic center of the fully formed caseating granuloma . The necrotic center has been previously identified in adult zebrafish granulomas , however larval granulomas are not necrotic , and to our knowledge the levels of Hif-α signaling in larval granulomas had not previously been investigated [19] . We did not detect high levels of Hif-α signaling , via elevated phd3 target gene expression , in larval granulomas . However , upregulated expression of phd3 was observed in infected macrophages at early stages of pathogenesis , before the formation of larval granulomas , and we were able to demonstrate that this increase in expression is Hif-α dependent . Upregulation and stabilization of HIF-1α is a known consequence of leukocyte activation during onset of infection , even in normoxia [13] , and remains high during pathogenesis of other types of bacterial infections leading to enhanced leukocyte function [13] , [15] . In contrast , we observed levels of Hif-α signaling in early Mm infection , but not at the granuloma stage , suggestive of a silencing of Hif-α signaling over the course of Mm pathogenesis . This observation is consistent with leukocyte transcriptional reprograming observed in human mycobacterial disease and with murine macrophages having upregulated HIF-1α in the presence of heat-killed Mtb but not in the presence of viable Mtb [43]–[45] . A dampening effect on Hif-α signaling could be a mechanism by which mycobacteria are able to form a protected niche in which to proliferate and disseminate , and this may be via manipulation of transcription factors such as HIF-1α . This possibility is supported by dominant negative hif-1αb having no effect on Mm bacterial burden , suggesting that Hif-1α signaling does not play a major role in leukocyte anti-mycobacterial activity during the normal pathogenesis of infection . Stabilization of HIF-1α during the pathogenesis of mycobacterial infection may represent a therapeutic opportunity to re-arm leukocytes and to inhibit further infection . The study of the zebrafish/Mm larval granuloma model has led to key discoveries for TB pathogenesis causing changes in treatment practises in the clinic [42] , [46] . However , the zebrafish embryo is an untapped resource for study of mycobacterial pathogenesis at the earliest stages of infection , before granuloma structures form . Genetic manipulation of Hif-1α during mycobacterial infection has not previously been explored , and further understanding of the roles of this critical host-signaling pathway may uncover Hif and its signaling components as future therapeutic targets for intervention against TB . A critical role of HIF-1α during bacterial infection was demonstrated in murine knockout studies showing that HIF-1α signaling is required for proper response to Streptococcal bacterial challenge [47] . Furthermore localized treatment with a drug that stabilized HIF-1α led to a decrease in proliferation of the skin pathogens , Pseudomonas aeruginosa and Acinetobacter baumanii , in a mouse abscess model [48] . Our data demonstrate , in vivo , that pharmacological or genetic stabilization of Hif-1α can aid the host in the fight against mycobacterial infection . Importantly , these data suggest that therapeutic upregulation of HIF-1α signaling could complement current antibiotic treatments in the fight against Mtb infection . In the zebrafish Mm model , the early treatment window of DMOG ( treatment between -4 hpi and 24 hpi , followed by wash-off then bacterial burden assessment at 4 dpi ) and the DA hif-1αb RNA injection at the one-cell stage , ( an effect that will be diluted as the embryo develops ) , indicate that Hif-1α stabilization at early stages of Mm infection , pre-granuloma formation , is causing the decrease bacterial burden at 4 dpi . This is further supported by the observation that embryo granulomas are able to form after DA hif-1αb injection despite a decrease in bacterial burden . Therefore , to understand the mechanism of action of early Hif-1α stabilization , we focused on the role of bacterial killing by the leukocytes during early stage infection . A major mechanism of bacterial killing is the use of RNS by leukocytes [29] , [49] . The leukocyte enzymatic producer of NO , iNOS , is a tightly regulated enzyme , and is a known target of a number of immune transcription factors , including HIF-1α [40] . We found that levels of NO , as assessed by protein nitrosylation , were consistently higher in neutrophils of infected embryos at early stages of infection . Morpholino reduction of iNOS confirmed that increased nitrosylation in neutrophils after infection was iNOS dependent . We showed that stabilization of Hif-1α is able to activate neutrophils to produce NO in the absence of bacterial challenge and that this could be achieved through neutrophil specific expression of DA hif-1αb , demonstrating a cell-autonomous effect . The increased level of NO in non-infected embryos with stabilized Hif-1α indicates a priming of neutrophils to bacterial challenge , leading to greater levels of RNS pre-infection . Early bacterial killing would lead to lower levels of bacterial survival , decreased dissemination and the observed decrease in bacterial burden at the granuloma stages of infection . This early priming of neutrophils could be blocked using iNOS inhibition , confirming that the reduced susceptibility to Mm infection due to Hif-1α stabilization is dependent on iNOS activity . The role of neutrophils in early mycobacterial infection is not fully understood . Macrophages have previously been thought the major leukocyte involved in the pathogenesis of mycobacterial infection [50] and are also the main leukocyte-type involved in the phagocytosis of intravenously injected mycobacteria and are present in abundance in the granuloma . However , more recently , potentially important roles of neutrophils during mycobacterial infection are becoming evident . Neutrophils are known to be able to undertake oxidative killing of Mm phagocytosed by macrophages in embryonic granulomas , a mechanism which may be important to control infection , however their role at early stages of infection has not been addressed [51] , [52] . We observed that enhanced nitrotyrosine levels early after infection were mostly detected in neutrophils , although in rare instances nitrosylation levels were detectable in both uninfected and infected macrophages . This was confirmed using the iNOS antibody , which also showed a mainly neutrophil localization . In the absence of infection there are detectable levels of nitrotyrosine in neutrophils , likely due to myeloperoxidase activity , an enzyme that is neutrophil-specific in zebrafish [53] . Myeloperoxidase can form NO-derived inflammatory oxidants and it has been shown that myeloperoxidase is responsible for the majority of tissue nitrosylation in fish in the absence of infection [30] , [54] . The anti-microbial effect of stabilized Hif-1α was abrogated after pharmacological and morpholino inhibition of iNOS . Neutrophils form the major leukocyte population that display protein nitrosylation after stabilization of Hif-1α , therefore we hypothesize that through an unknown mechanism , increased neutrophil iNOS levels leads to increased Mm killing at early stages of infection , ultimately decreasing bacterial burden . As discussed above , macrophages are the major leukocyte involved in the phagocytosis of Mm in this model [50] . However we found many instances where neutrophils contain internalized Mm in the first 24 hours post infection . Our observations indicate that neutrophils are able to internalize Mm before the presence of granuloma structures , and that they have elevated nitrosylation levels during infection . Therefore , a potential mechanism is that neutrophils are able to phagocytose Mm and increase iNOS after Hif-1α stabilization leading to enough early bactericidal activity to significantly reduce bacterial burden at later timepoints . However , it is clear that in the zebrafish embryo model , as in other models , macrophages are the major cell type with internalized Mm at early timepoints of infection . Therefore , a more likely hypothesis is that neutrophils with activated iNOS are able to interact with infected macrophages , either by transfer of live bacteria or transfer of reactive nitrogen species , leading to increased bactericidal activity . We therefore hypothesize that neutrophils play an important role alongside macrophages in early Mm bacterial killing , but the bactericidal mechanisms of this interaction are yet to be uncovered . The role of Hif-2α isoform stabilization in leukocytes during inflammation and infection has not been widely investigated in any in vivo model of infection . We set out to identify what effect Hif-2α modulation had on the outcome of Mm infection . Although stabilization of Hif-2α has the same increasing effect as Hif-1α stabilization on expression of the major Hif-α target gene , phd3 , we observed an opposing effect on Mm bacterial burden . We demonstrate that the decrease in bacterial burden was due to an increase in neutrophil nitrosylation after downregulation of Hif-2α . The decrease in bacterial burden after Hif-2α downregulation could be blocked by early iNOS inhibition and we hypothesize that this effect is mediated by differential roles of the Hif-α isoforms on the iNOS genetic pathway . HIF-1α and HIF-2α have been previously been shown to have opposing effects on iNOS in mammalian cultured macrophages , where HIF-1α stabilization transcriptionally upregulated iNOS , while HIF-2α stabilization decreased NO levels [40] . Our findings confirm that Hif-2α is able to upregulate the well-characterised Hif-α target phd3 while having opposing effects on nitrosylation , corroborating previous in vitro observations [40] . These data demonstrate the in vivo consequence on bacterial infection of the differential regulation of iNOS by Hif-α variants . The opposing effects of Hif-α isoforms on bacterial burden highlight the tight control of NO homeostasis in leukocytes . The potential of the HIF-α pathway for therapeutic intervention in other diseases , including cancer and ischemia , is widely recognized , however differential roles of HIF-α isoforms in these diseases are only recently coming to light [55] , [56] . These regulatory mechanisms , in part mediated by HIF-α , are complex and further studies are required before the regulation of HIF-α isoforms and iNOS during infection is fully understood . In conclusion , our data demonstrate that in vivo modulation of host Hif-α signaling during early Mm pathogenesis can lead to decreased burden of mycobacterial infection . Stabilization of Hif-1α , or reduction of Hif-2α , results in priming of neutrophil NO bactericidal activity leading to lower mycobacterial burden after challenge with infection . Our data highlight the delicate balance of HIF-α and iNOS signaling in leukocyte function during infection and highlight the important role of neutrophils during early stage Mm infection . Further understanding of the complex crosstalk between Hif-α and iNOS pathways during Mtb infection will help identify novel , host-targeted , therapeutic strategies against TB . NO priming of neutrophils by targeted Hif-α modulation , may decrease the level of initial Mtb infection and act to block the development of acute TB disease caused by re-activation and dissemination of latent Mtb infection . Host targeted strategies would be predicted to be beneficial against all types of TB , including multiple drug resistant strains , and may be less susceptible to therapy-resistance than antibiotic strategies , thereby reducing the global burden of TB .
Zebrafish lines were handled in compliance with the local animal welfare regulations and maintained according to standard protocols ( zfin . org ) . The breeding of adult fish was approved by the local animal welfare committee ( DEC ) of the University of Leiden . All protocols adhered to the international guidelines specified by the EU Animal Protection Directive 2010/63/EU . Zebrafish were maintained according to standard protocols [57] and local animal welfare regulations . Strains used were ABTL ( wildtype ) , Tg ( phd3:GFP ) i144 , Tg ( mpx:GFP ) i114 , Tg ( lyz:DsRED2 ) nz50 and Tg ( mpeg1:mCherryF ) ump2 [25] , [32] , [58] . Infection experiments were performed using M . marinum strain M ( ATCC #BAA-535 ) , containing the pSMT3-mCherry or pSMT3-Crimson vector [59] . Liquid cultures were prepared from bacterial plates [59] . Injection inoculum was prepared in 2% polyvinylpyrrolidone40 ( PVP40 ) solution ( CalBiochem ) as previously described [60] , [61] . 100 colony-forming units ( CFU ) of bacteria were injected into the caudal vein at 28 hpf as previously described [61] . Whole mount in situ hybridization of phd3 was carried out as previously described [16] , [25] . 6 dpi and 1 dpi Tg ( phd3:GFP ) i144 [25] larvae infected with Mm were embedded in 1% low melting point agarose ( Sigma Aldrich ) and transferred to a Leica DMIRBE inverted microscope with a Leica SP1 confocal scanhead for imaging with 40 or 63 times lenses . The pan hydroxylase inhibitor , DMOG ( dimethyloxaloylglycine , Enzo Life Sciences ) , was used at a 100 µM concentration as previously described [16] . DMSO solvent controls were used . Unless otherwise stated embryos were treated from 4 hours pre Mm infection to 24 hpi by addition to the embryo water . The inhibitors were then washed off with fresh embryo water and grown to 4 dpi for assessment of bacterial load as described below . Embryos were injected with dominant Hif-α RNA at the one cell stage as previously described [16] . hif-α variants used were dominant active ( DA ) and dominant negative ( DN ) hif-1αb ( ZFIN: hif1ab ) and hif-2αa ( ZFIN: epas1a ) ( primer sequences in Table S1 ) . Phenol red ( Sigma Aldrich ) was used as a vehicle control . Embryos were imaged at 4 dpi on a Leica MZ16FA Fluorescence Stereo Microscope . Brightfield and fluorescence images were generated with a Leica DC500 ( DFC420C ) camera . Bacterial loads were analysed using dedicated pixel counting software as previously described [23] . Larvae were fixed in 4% paraformaldehyde in PBS overnight at 4°C and leukocytes were immune-labeled using the l-plastin antibody as previously described [60] , [62] . Neutrophils were labeled with TSA ( TSAplus kit , Fluorescence Systems , Perkin Elmer Inc ) staining labeled neutrophils in fluorescein green fluorescence as previously described [63] . Two timepoints were chosen for this analysis , the timepoint of Mm injection ( 28–30 hpf ) and the timepoint of bacterial burden assessment ( 5 dpf ) . RNA groups were blinded prior to counting . Neutrophils and leukocytes in embryos and larvae were counted in the tail region using a Leica MZ16FA Fluorescence Stereo Microscope . The nos2a morpholino ( Genetools ) was used as previously reported [34] . A standard control morpholino ( Genetools ) was used as a negative control . Larvae were fixed in 4% paraformaldehyde in PBS overnight at 4°C and nitrotyrosine levels were immune-labeled with a rabbit polyclonal anti-nitrotyrosine antibody ( Merck Millipore 06-284 ) at a 1∶200 dilution of primary antibody , and were detected using an Alexa Fluor ( Invitrogen Life Technologies ) secondary antibody . Larvae were fixed in 4% paraformaldehyde in PBS overnight at 4°C and iNOS was immune-labeled with a rabbit polyclonal iNOS antibody ( BD Biosciences ) as previously described [36] . Detection was with goat anti-rabbit HRP-conjugated antibody ( Abcam , 1∶500 dilution ) and Cy3Plus TSA kit ( Perkin-Elmer ) . Embryos were imaged at 1 dpi , in the presence or absence of infection , embedded in 1% low melting point agarose ( Sigma Aldrich ) and transferred to a Leica DMIRBE inverted microscope with a Leica SP1 confocal scanhead for imaging with 40 or 63 times lenses . For quantification purposes acquisition settings and area of imaging ( in the caudal vein region ) were kept the same across groups . Corrected total cell fluorescence was calculated for each immune-stained cell using Image J as previously described [64] . The GFP of the Tg ( mpx:GFP ) i114 [58] was used to assess the leukocyte cell boundaries . The Tol2kit multisite gateway-based transposon system was used to make transgenic constructs to transiently and mosaic express DA hif-1α specifically in neutrophils and macrophages [65] . DA hif-1αb was recombined into the middle entry pDONR221 using BP Clonase ( Invitrogen ) . An LR Clonase ( Invitrogen ) Gateway reaction was performed with p5E-lyz ( neutrophil specific promoter ) or p5E-mpeg1 ( macrophage specific promoter , [33] ) , pDONR221-da-hif-1αb and p3E-IRES-nlsEGFPpA inserted into pDestTol2pA2 . The resulting plasmids , Tg ( lyz:da-hif-1αb , IRES-nlsGFP ) and Tg ( mpeg1:da-hif-1αb , IRES-nlsGFP ) , were microinjected with tol2 transposase RNA into the one cell stage embryo to express in transient , mosaically . Positive fish were screened for the transgene using the heart marker eGFP expression ( found in the Gateway vector ) , and positive cells were screened by confocal microscopy for the nuclear localized eGFP signal showing the expression of the transgene . Embryos were injected with the relevant hif-α construct and infected with Mm at 1 dpf . At the timepoint of infection DAF-FM DA was applied to the embryo water as previously described [35] . DAF-FM DA was washed off using embryo water at 1 dpi and imaged using confocal microscopy . The pan-NOS inhibitor L-NAME , ( NG-Nitro-L-arginine methyl ester , Tocris Bioscience ) , was used at 200 µM as previously described [34] , [37] . The iNOS inhibitor L-NIL ( N6- ( 1-iminoethyl ) -L-lysine , dihydrochloride , Tocris Bioscience ) was used at a 200 µM concentration [38] . DMSO solvent controls were used at corresponding concentrations for each treatment . Unless otherwise stated embryos were treated from 4 hours pre Mm infection to 24 hpi by addition to the embryo water . The inhibitors were then washed off with fresh embryo water and grown to 4 dpi for assessment of bacterial load as described above . Dominant active zebrafish hif-2αa was generated by successive rounds of site directed mutagenesis , each mutating a hydroxylation site into a non-hydroxylatable form as previously described for hif-1αb [16] . Dominant negative hif-2αa was generated by a truncation at the equivalent amino acid to the 330th amino acid in the human sequence , as previously described for hif-1αb [16] . All data were analysed ( Prism 5 . 0 , GraphPad Software ) using unpaired , two-tailed t-tests for comparisons between two groups and one-way ANOVA ( with Bonferonni post-test adjustment ) for other data . P values shown are: *P< . 05 , **P< . 01 , and ***P< . 001 . | Tuberculosis is a mycobacterial disease that was a major cause of death until the discovery of antibiotics in the mid-twentieth century . However , TB is once again on the rise , with the emergence of strains that are multi-drug resistant . Mycobacteria are specialists in evading immune cell killing and use host immune cells as a niche in which they can proliferate and survive latently , until subsequent re-activation and spreading causing life-threatening disease . Pharmaceutical reprogramming of the immune system to kill intracellular mycobacteria would represent a therapeutic strategy , effective against currently untreatable strains and less susceptible to drug resistance . Here we use an in vivo zebrafish model of TB to show that manipulation of the host genetic pathway responsible for detecting low oxygen levels ( hypoxia ) causes a decrease in mycobacterial infection . This antimicrobial effect was due to a priming of immune cells with increased levels of nitric oxide , a molecule that is used by immune cells to kill bacteria . Here we show in vivo manipulation of a host-signaling pathway aids the host in combatting mycobacteria infection , identifying hypoxic signaling as a potential target for future therapeutics against TB . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | Hypoxia Inducible Factor Signaling Modulates Susceptibility to Mycobacterial Infection via a Nitric Oxide Dependent Mechanism |
EBV causes human B-cell lymphomas and transforms B cells in vitro . EBNA3C , an EBV protein expressed in latently-infected cells , is required for EBV transformation of B cells in vitro . While EBNA3C undoubtedly plays a key role in allowing EBV to successfully infect B cells , many EBV+ lymphomas do not express this protein , suggesting that cellular mutations and/or signaling pathways may obviate the need for EBNA3C in vivo under certain conditions . EBNA3C collaborates with EBNA3A to repress expression of the CDKN2A-encoded tumor suppressors , p16 and p14 , and EBNA3C-deleted EBV transforms B cells containing a p16 germline mutation in vitro . Here we have examined the phenotype of an EBNAC-deleted virus ( Δ3C EBV ) in a cord blood-humanized mouse model ( CBH ) . We found that the Δ3C virus induced fewer lymphomas ( occurring with a delayed onset ) in comparison to the wild-type ( WT ) control virus , although a subset ( 10/26 ) of Δ3C-infected CBH mice eventually developed invasive diffuse large B cell lymphomas with type III latency . Both WT and Δ3C viruses induced B-cell lymphomas with restricted B-cell populations and heterogeneous T-cell infiltration . In comparison to WT-infected tumors , Δ3C-infected tumors had greatly increased p16 levels , and RNA-seq analysis revealed a decrease in E2F target gene expression . However , we found that Δ3C-infected tumors expressed c-Myc and cyclin E at similar levels compared to WT-infected tumors , allowing cells to at least partially bypass p16-mediated cell cycle inhibition . The anti-apoptotic proteins , BCL2 and IRF4 , were expressed in Δ3C-infected tumors , likely helping cells avoid c-Myc-induced apoptosis . Unexpectedly , Δ3C-infected tumors had increased T-cell infiltration , increased expression of T-cell chemokines ( CCL5 , CCL20 and CCL22 ) and enhanced type I interferon response in comparison to WT tumors . Together , these results reveal that EBNA3C contributes to , but is not essential for , EBV-induced lymphomagenesis in CBH mice , and suggest potentially important immunologic roles of EBNA3C in vivo .
Epstein-Barr virus ( EBV ) is a human gamma-herpesvirus that infects 90% of the world’s adult population [1] . EBV establishes a lifelong infection in the memory B cell compartment , and periodically reactivates to the lytic form of viral infection when B cells are stimulated by antigen and/or differentiate into plasma cells [1–3] . In addition to causing the clinical syndrome , infectious mononucleosis , EBV is associated with multiple human malignancies of B-cell and epithelial origin , including Burkitt lymphoma , Hodgkin lymphoma , diffuse large B cell lymphoma ( DLBCL ) , post-transplant lymphoproliferative disease ( PTLD ) , nasopharyngeal carcinoma and gastric carcinoma [4–6] . All EBV-associated malignancies are latently infected with EBV . EBV infection of B cells in vitro results in immortalized lymphoblastoid cell lines ( LCLs ) that produce lymphomas when injected into immunosuppressed mice , and much of our current understanding of the transforming functions of latent EBV proteins is derived from LCL models . However , while at least three different types of viral latency ( types I , II and III ) can occur in B cells during EBV infection in vivo , type III latency ( in which all 9 latency proteins are expressed ) is the only latency type that transforms B-cells in vitro into LCLs [6] . EBV-induced transformation of B cells in vitro is largely driven by the two major EBV-encoded oncoproteins , EBNA2 and LMP1 [7] . EBNA2 is a strong transcriptional activator that turns on the viral promoters driving latent viral gene transcription during type III latency , and also upregulates cellular genes such as c-Myc and Cyclin D2 that induce LCL proliferation [8–11] . EBNA2 directly interacts with the cellular RBPJK DNA binding protein ( that normally mediates intracellular Notch binding ) to produce a Notch-like signal [12 , 13] . The other major EBV oncoprotein , LMP1 , is a transmembrane protein that mimics the effect of constitutive CD40 signaling , which results in NF-kB activation [14–18] . The NF-kB pathway induces survival and proliferation of B cells and is activated in many human B-cell lymphomas [19] . In addition to EBNA2 , three other latent EBV nuclear proteins , EBNA3A , EBNA3B , and EBNA3C , also interact with RBPJK in LCLs , and both EBNA3A and EBNA3C are required for LCL generation in vitro [6] . The EBNA3 proteins regulate distinct but overlapping sets of cellular genes in vitro [20–25] . EBNA3A and EBNA3C collaborate to induce repression of the CDKN2A-encoded tumor suppressors , p16INK4a ( p16 ) and p14ARF ( p14 ) , an effect associated with increased H3K27me3 modification at the CDKN2A promoter [26–28] . Expression of the CDKN2A locus is activated by many cellular and viral oncoproteins , and the p14/p16 proteins encoded by this locus inhibit the ability of oncoproteins to transform cells by preventing cell cycle progression and inducing cellular senescence [29–34] . Turn-off of EBNA3C expression in established LCLs in vitro results in high level p16 expression and inhibits cellular proliferation [35 , 36] . Importantly , EBNA3C-deleted EBV can transform B cells containing a germline mutation inactivating p16 into LCLs in vitro [27] . Thus , the most essential role of EBNA3C in EBV-induced transformation of B cells in vitro is inhibiting oncogene-induced expression of p16 . Numerous other proposed functions of EBNA3C may also contribute to the viability of EBV-infected B cells in vivo . In addition to preventing p16-induced cellular senescence , EBNA3C attenuates an anti-proliferative DNA damage response that inhibits EBV transformation of primary B lymphocytes in vitro [37] , and upregulates expression of the cellular AICDA gene [38] . AICDA encodes the activation-induced cytosine deaminase protein ( AID ) , which is required for immunoglobulin class switching and somatic hyper-mutation of B cells [39] , and is thought to promote the c-Myc translocations that drive Burkitt lymphomas [40] . EBNA3C has also been implicated in preventing c-Myc induced apoptosis through inhibition of BIM activation in Burkitt lymphoma cell lines [34 , 41 , 42] . In addition , EBNA3C is reported to enhance stability of the B-cell survival protein , IRF4 [43] , and to inhibit p53 function in vitro [44] . Furthermore , although EBNA3 proteins interact with an RBPJK domain which is different from the domain that interacts with EBNA2 , they compete for RBPJK binding and limit EBNA2 activation of certain promoters [20 , 45–48] . Thus , EBNA3C could potentially affect EBV transformation in vivo by modulating EBNA2 function . EBNA3C also reduces expression of several different T-cell chemokines when expressed in the an EBV-negative Burkitt cell line in vitro [49] , and could potentially inhibit T-cell infiltration of tumors via this mechanism . Nevertheless , since many EBV+ human lymphomas do not express EBNA3C , it may be less important for growth of EBV+ lymphomas in vivo , suggesting that cellular mutations and/or signaling pathways may obviate the need for EBNA3C . In addition , the epigenetic modifications induced by EBNA3C during early infection may decrease the need for continued expression of EBNA3C at later time points [50] . Humanized mouse models provide an opportunity to study EBV infection in vivo in an environment containing human immune cells , and can lead to novel insights regarding the roles of EBV latency proteins . We previously used a humanized mouse model to show that an LMP1-deleted EBV mutant that is non-transforming in vitro can cause lymphomas in vivo due to the ability of human CD40 ligand-expressing CD4 T cells to provide an alternative source of CD40 signaling [51 , 52] . Here , we have used a cord blood-humanized mouse model to examine the phenotype of an EBNA3C-deleted EBV mutant in vivo . We find that this mutant is partially impaired for the ability to induce lymphomas in this model , although a subset of infected animals develop DLBCLs with delayed kinetics . By comparing the phenotypes of lymphomas containing wild-type ( WT ) versus EBNA3C-deleted viruses in this model , we have confirmed some previously reported in vitro functions of EBNA3C ( inhibition of p16 expression and activation of AICDA expression ) and have discovered potentially new functions ( including inhibition of T cell infiltration and type 1 interferon signaling ) . These results reveal that EBNA3C is important but not essential for the development of EBV-induced lymphomas in cord blood-humanized mice .
To examine the effect of EBNA3C loss on the ability of EBV to induce lymphomas in humanized mice in vivo , we inactivated the EBNA3C gene in the B95 . 8 EBV bacmid p2089 by inserting a stop codon near the start of the EBNA3C protein as described in the methods section . A revertant mutant was also constructed . 293 cell lines stably infected with either mutant and revertant bacmids were selected for as described in the methods . Infectious virion particles were produced by transfecting 293 cells with BZLF1 ( Z ) , BRLF1 ( R ) and BALF4 ( GP110 ) expression vectors . Virus was then concentrated and titered using the “green Raji cell assay” as previously described [53] . The cord blood-humanized mouse model , which we previously showed supports the development of lymphomas induced by a LMP1-deleted EBV mutant [51 , 52] , was used to compare lymphomagenesis outcomes from wild-type ( WT ) versus EBNA3C-deleted ( Δ3C ) EBV in vivo . CD34-depleted human umbilical cord blood was either mock infected or infected with 5 , 000 infectious particles ( green Raji units ) of WT or Δ3C EBV for 1 . 5 hours in vitro and then injected intraperitoneally ( i . p . ) into NSG ( NOD/LtSz-scid/IL2Rnull ) mice . Approximately 10 million nucleated cord blood cells ( containing a mixture of B cells , T cells , NK cells , monocytes and dendritic cells ) were injected into each mouse . In each experiment , WT and mutant virus infection was performed using the same cord blood sample , and experiments were repeated multiple times using different cord blood samples . We have previously shown that WT ( B95 . 8 bacmid ) EBV induces lymphomas in nearly all animals in this model , whereas mock-infected animals do not develop lymphomas [52] . Animals were sacrificed when they showed signs of illness , or otherwise euthanized at day 90 after cord blood cell injection . Animals were considered to have EBV-infected lymphomas if they had large numbers of EBNA2- ( or EBNA1- ) positive B cells invading one or more organs . As shown in Fig 1 , the Δ3C-infected animals required a significantly longer time to develop tumors ( 66–90 days ) compared to WT-infected animals ( 28–35 days ) . Since the revertant virus-infected lymphomas formed at a similar time-frame , and with a similar efficiency , as the WT virus-induced lymphomas ( Fig 1A ) , results from WT and revertant EBV infection were aggregated ( and referred to as “WT” EBV ) for the rest of the analyses shown in this paper . In addition to producing lymphomas at a delayed time point relative to WT EBV , Δ3C EBV-infected animals had significantly fewer lymphomas in the cord blood-humanized mouse model ( 10/26 infected animals ) in comparison to WT EBV ( 25/27 infected animals ) ( Fig 1B ) . We confirmed that lymphomas induced by the EBNA3C-mutant virus contained the expected mutation ( insertion of a stop codon at residue 2 in the EBNA3C protein ) by sequencing viral DNA isolated from lymphomas on formalin-fixed paraffin-embedded ( FFPE ) slides infected with WT or Δ3C viruses ( S1 Fig ) . These results confirm that EBNA3C expression contributes to efficient EBV-induced lymphoma formation in the cord blood-humanized mouse model . Nevertheless , since the Δ3C-infected animals developed significantly more tumors than mock-infected animals ( 0/32 ) , loss of EBNA3C expression does not completely prevent EBV from inducing lymphomas in cord blood-humanized mice . Interestingly , the great of majority of tumor-free Δ3C-infected animals had no evidence of persistent viral infection ( EBER+ cells by in situ hybridization ) at the time of euthanasia . The unexpected ability of the Δ3C virus to induce lymphomas in a subset of animals in this model allowed us to compare the characteristics of lymphomas infected with the two different types of viruses . To compare the phenotypes of lymphomas induced with the WT versus Δ3C viruses , tumors were stained with hematoxylin and eosin ( H&E ) . Animals infected with either the WT or Δ3C viruses developed aggressive DLBCLs that invaded various organs ( particularly the pancreas , as shown in Fig 2A , as well as the liver , gall bladder , abdominal lymph nodes and less commonly the spleen ) ( S1 Table ) . There was no consistent difference in the sizes of lymphomas induced by the WT versus Δ3C viruses . In each case , the tumors consisted of sheet-like expansions of morphologically atypical large cells with prominent nucleoli and irregular nuclei . The morphology ranged from immunoblastic to frankly anaplastic between different mice , but no consistent morphologic differences were noted between WT virus- versus Δ3C virus- induced lymphomas . To determine if DLBCLs have type III EBV latency , immunohistochemistry ( IHC ) staining and immunoblot analysis was performed using antibodies that detect either the EBNA2 or LMP1 proteins . Tumors infected with either the WT or Δ3C viruses expressed both EBNA2 and LMP1 ( Figs 2A , 2B and S2 ) , indicating that they each have type III latency . To determine if loss of EBNA3C affects the ability of EBNA2 to activate the LMP1 promoter ( which has previously been reported to be synergistically activated by the EBNA2/EBNA3C combination in vitro [54 , 55] ) , the relative number of LMP1-positive cells versus EBNA2-positive cells was quantitated . As shown in Fig 2C , WT- and Δ3C-infected lymphomas had a similar number of LMP1-positive cells relative to EBNA2-positive cells ( Fig 2C ) . To assess whether WT- and Δ3C-infected lymphomas contain similar numbers of EBV-infected B cells , we performed EBER , EBNA1 , and CD20 stains on adjacent slides and quantified the ratio of cells expressing each EBV marker to the number of CD20 ( a B-cell marker ) -positive cells . The number of cells expressing each EBV marker was similar to the number of CD20 positive cells in each tumor type ( S3 Fig and S2 Table ) . These results indicate that WT and EBNA3C-deleted viruses induce lymphomas containing a similar number of EBV-infected B cells . Given the ability of EBNA3C to inhibit p16 expression in LCLs , and the essential role that inhibition of p16 expression plays for LCL generation in vitro [36] , we next asked if p16 expression is altered in Δ3C virus-infected , versus WT virus-infected , lymphomas . As shown in Fig 2D and 2E , IHC staining of tumors using an antibody directed against p16 revealed a marked increase in p16 expression in the Δ3C virus-induced lymphomas . This result confirms that EBNA3C significantly inhibits p16 expression in EBV-infected lymphoma cells in cord blood-humanized mice . Although high-level p16 expression normally halts cellular proliferation by inhibiting the ability of cyclin D/CDK4/6 complexes to phosphorylate and inactivate pRb [33] , human melanomas and DLBCLs can escape the effect of p16 by inducing cyclin E expression [56 , 57] . High-level c-Myc expression ( which activates cyclin E expression ) can also bypass the p16 effect [58] . We therefore performed IHC using antibodies against CD20 ( a B-cell marker ) , cyclin E , and c-Myc to compare the levels of cyclin E and c-Myc in CD20+ B cells of Δ3C virus-infected versus WT virus-infected lymphomas ( Fig 2D ) . The Δ3C virus-infected lymphoma cells expressed both cyclin E and c-Myc at levels similar to that found in the WT virus-infected cells . This result suggests that Δ3C virus-infected lymphomas can bypass the cell cycle inhibitory effect of p16 by expressing cyclin E and/or c-Myc . Since the pro-apoptotic protein , BIM , is an important tumor suppressor for c-Myc-induced lymphomas [34 , 41] , and EBNA3C inhibits expression of BIM in Burkitt lymphoma cell lines in vitro [42] , we performed IHC analysis using antibodies against the EBV latency protein , EBNA1 , and BIM to compare BIM expression in WT virus- versus Δ3C virus-infected lymphoma cells . Consistent with previous results [42] , BIM expression was significantly higher in the Δ3C virus-infected lymphoma cells versus the WT virus-infected cells ( Fig 3A and 3B ) . Nevertheless , we found that the majority of EBNA1-positive cells did not co-express BIM in either tumor type ( Fig 3A and 3B ) . Many BIM+ cells that did not co-stain for EBNA1 ( presumably EBV-negative T cells ) were also observed ( Fig 3A ) . This result suggests that in addition to EBNA3C , other viral and/or cellular factors can suppress BIM expression in EBV-induced B-cell lymphomas with type III latency in the cord blood-humanized mouse model . We also compared the levels of the anti-apoptotic proteins , BCL2 and IRF4 , in the two tumor types . Both BCL2 and IRF4 expression are reported to be induced by NF-kB signaling [59–61] , which is activated by LMP1 . BCL2 inhibits apoptosis by inhibiting the pro-apoptotic proteins , BAX and BAK1 [62] . IRF4 is a surrogate immunohistochemical marker for the “activated B cell” subtype of DLBCLs and is also an essential survival factor for activated B cell lymphomas in humans [63 , 64] . EBNA3C has been shown to stabilize IRF4 expression in vitro , and regulates certain cellular genes in a complex with IRF4 [20 , 43] . IHC analysis using antibodies against BCL2 , IRF4 , and CD20 showed a similar level of both BCL2 and IRF4 in CD20 co-staining B cells in the WT virus- and Δ3C virus-induced tumors ( Fig 3A ) . These results suggest that EBNA3C is not required for IRF4 or BCL2 expression in EBV-infected lymphomas . Nevertheless , we cannot exclude the possibility that the essential role of IRF4 in activated DLBCLs selects for EBV-independent IRF4 expression in Δ3C virus-induced tumors . An EBV mutant deleted for the EBNA3B protein was previously shown to produce lymphomas with decreased T-cell infiltration ( in comparison to WT virus ) in humanized mice , possibly due to decreased expression of the T-cell chemokine , IP10 , in B cells infected with this virus [65] . In contrast , EBNA3C reduces expression of several different chemokines when expressed in the EBV-negative BJAB Burkitt cell line in vitro , including the CCL3 , CCL4 , CXCL10 and CXCL11 chemokines , which can each attract T cells to infected tissues [49] . Thus , EBNA3C could potentially act to inhibit T-cell infiltration of EBV-infected lymphomas in vivo . To determine if lymphomas infected with the Δ3C EBV have altered T-cell infiltration in comparison to wild-type virus infected lymphomas in cord-blood humanized mice , we compared the number of CD3-positive cells to the number of CD20-positive cells in each tumor type . This analysis revealed that Δ3C-virus induced lymphomas had increased T-cell infiltration compared to the WT-induced lymphomas ( Figs 4A , 4B and S4 ) . While the absolute number of B cells per 40X field was similar between WT- and Δ3C-induced lymphomas ( S4 Fig ) , the absolute number of T cells was increased . IHC analysis using antibodies against CD4 ( helper T cells ) and CD8 ( cytotoxic T cells ) showed a significant increase in both T-cell populations in the Δ3C-induced tumors compared to the WT-induced tumors ( Fig 4C and 4D ) . These results suggest that EBNA3C expression , in contrast to EBNA3B expression , may be associated with reduced T-cell infiltration into EBV-induced lymphomas . To further examine expression of viral and cellular genes in WT virus-induced , versus Δ3C virus-induced , lymphomas , RNA was isolated from three different tumors infected with each virus type , and RNA-seq was performed . In this analysis , RNA sequences were derived from a mixture of EBV-infected tumor cells and infiltrating T cells; contaminating mouse cell genes were removed from the bioinformatics analysis as described in the methods . To investigate the characteristics of the B cell populations of EBV-induced lymphomas infected with WT or Δ3C viruses , we compared the expression levels of the 40 human IGHV genes in each tumor as described in the methods . In most tumors ( both WT- and Δ3C-infected ) , a single dominant IGHV gene was utilized ( Fig 5 ) . One of the three Δ3C-induced lymphomas are examined was composed of two different major clones , and a mixture of other IGHV sequences ( Fig 5B ) . These results suggest that both WT EBV- and Δ3C-induced lymphomas are mainly derived from restricted expansions of a small initial number of EBV-infected B cells , rather than consisting of highly heterogeneous expansions of a broad selection of different EBV-infected B cells . Interestingly , 3 of the 6 tumors examined harbored the same dominant IGHV gene ( IGHV2-5 ) , and 2 of the tumors expressed the IGHV1-69 gene as a dominant clone ( Fig 5 ) . In comparison , IGH transcripts containing IGHV2-5 constitute only 0 . 05% of total IGH transcripts in normal cord blood , while transcripts containing IGHV1-69 constitute 9% of total IGH transcripts [66] . The CDR3 sequences of BCR clones in each tumor were analyzed as described in the methods . The major BCR clones in each tumor had distinct CDR3 sequences ( S5 Fig ) . To determine if WT virus-infected , or Δ3C virus-infected lymphomas had undergone somatic hyper-mutation , the CDR3 sequences of dominant IGH clones were also analyzed using publicly available software ( IMGT ) as described in the methods . This analysis revealed that the CDR3 sequences had undergone little if any somatic hyper-mutation in either WT EBV- or Δ3C virus-infected lymphomas ( S5 Fig ) . Together , these results indicate that the usage of certain variable heavy chain genes may be selected for during the development of EBV-derived B-cell tumors in the cord blood-humanized mouse model , but somatic hyper-mutation is unlikely to contribute to the formation of these tumors . To investigate the T cell infiltrates of EBV-induced lymphomas infected with the WT or Δ3C viruses , we compared the expression levels of the 42 different human TCR beta variable ( TRBV ) genes in each tumor ( Fig 6 ) . All tumors contained a variety of TCRs . CDR3 sequences of some TCR beta chains in each tumor were also determined by analysis of RNAseq results as described in the methods . A number of different CDR3 sequences were present , with no single CDR3 identified in multiple tumors ( S4 Table ) . Together , the analysis of the IGH and TCR beta gene expression patterns indicate that EBV-induced tumors contain restricted B-cell populations , and heterogeneous T-cell populations . To examine the pattern of EBV gene expression in two different Δ3C virus-infected , versus two different WT virus-infected , lymphomas , RNA-seq reads were mapped to the EBV genome ( Fig 7 ) ; transcripts were also mapped to each strand of the viral genome ( S6 Fig ) . Similar levels of lytic gene expression ( primarily derived from leftward transcripts ) were observed in the presence and absence of EBNA3C ( S6 Fig ) . Interestingly , the two Δ3C-infected tumors appeared to have somewhat increased levels of rightward transcripts mapping to the EBNA2 and BHRF1 genes in comparison to the two WT virus-infected tumors ( Figs 7 and S6 ) . In addition , the levels of the leftward LMP1 transcript appeared to be potentially higher in the two Δ3C-infected tumors . Since the RNAseq analysis shown does not normalize the number of viral transcripts to the number of B cells in each specimen , we also compared the number of EBV LMP1 , BHRF1 and EBNA2 transcripts to the number of the B-cell specific PAX5 transcript in each tumor sample ( Fig 8A ) , and performed qPCR analysis on cDNA derived from the same tumors ( Fig 8B ) . Although these results suggest that the BHRF1 and EBNA2 transcripts may be higher in the Δ3C-infected tumors , since only two tumors were examined for each tumor type further studies are required to confirm these findings , especially given that the EBNA2 protein levels were similar on immunoblot analysis of tumors ( Fig 2B ) . We did not detect BHRF1 expression by IHC or immunoblot analysis in either tumor type ( S7 Fig ) . It is possible that the increased BHRF1 transcripts in the Δ3C-infected tumors are due to increased levels of the BHRF1 non-coding miRNAs , which are known to promote tumor growth through multiple mechanism ( s ) [67 , 68] . RNA-seq results were also used to compare human cellular gene expression in Δ3C virus-infected versus WT virus-infected tumors . Two tumors of each type had sufficient RNA-seq quality to perform further analysis . RNA-seq results were derived using total tumor tissue , and thus represent a mixture of genes expressed in EBV-infected B cells , infiltrating T cells , and mouse cells ( removed by bioinformatic analysis ) . Since differences in the number of B cells present in each tumor sample , and the presence of infiltrating T cells , could potentially obscure differences in cellular gene expression specific to the B-cell population , we initially compared the expression of a B-cell specific gene , PAX5 , and the expression of several different well-known EBNA3C target genes in the Δ3C virus-infected versus WT virus-infected tumors . As shown in Fig 9A , Δ3C virus-infected and WT virus-infected tumors expressed very similar levels of PAX5 , suggesting a similar number of B cells in each tumor sample . In contrast , expression of the AICDA gene ( activated by EBNA3C in vitro ) was decreased over 500-fold in Δ3C virus-infected ( versus WT virus-infected ) lymphomas , and expression of four different cellular genes down-regulated by EBNA3C in vitro ( CDKN2A , COBLL1 , ADAMDEC1 , and ADAM28 [26 , 69] ) was significantly increased . These results confirm that RNA-seq analysis of the total tumor cell population detects altered regulation of known EBNA3C target genes in the Δ3C-infected tumors . To examine the consequence of EBNA3C loss on cellular gene expression more globally , we performed gene set enrichment analysis ( GSEA ) to identify specific pathways altered in the Δ3C virus-infected ( versus WT virus-infected ) lymphomas . One of the most significantly down-regulated pathways in the Δ3C-induced tumors was “Hallmark E2F Targets” ( Fig 9B and 9C ) . This result is consistent with increased expression of p16 in Δ3C virus-infected lymphomas ( Fig 2C ) , given that p16 inhibits E2F1 expression by preventing phosphorylation/inactivation of pRb by the cyclin D/CDK4/6 complex [29 , 33] . E2F1 activates cell cycle progression , and thus decreased expression of E2F target genes is also consistent with our finding that Δ3C virus-infected lymphomas occur less frequently , and at later time points , in comparison to the WT virus-infected lymphomas ( Fig 1 ) . Since Δ3C virus-infected lymphomas continue to proliferate ( albeit more slowly ) in the presence of high-level p16 , we also compared expression of specific cellular genes involved in activating cell cycle progression in the Δ3C-induced versus WT-induced tumors , including cyclin E , CDK4 , CDK6 , cyclins D1 , D2 , and D3 , c-Myc , E2F1 , and cyclin A ( Fig 9D ) . The expression of these genes was not significantly increased or decreased in the Δ3C-induced tumors compared to the WT-induced tumors , although in most cases ( except for the cyclin D1 and CDK6 genes ) expression was less in the EBNA3C-deleted tumors . The continued transcription of genes required for cell cycle progression , along with c-Myc and cyclin E protein expression ( Fig 2C ) , in the Δ3C virus-infected lymphomas likely allows these tumors to partially escape the effect of high p16 levels . GSEA analysis also revealed an increase in the “Hallmark Interferon Alpha response pathway” in Δ3C virus-infected tumors ( Fig 10A ) . Examples of alpha interferon-stimulated genes differentially regulated in the Δ3C virus-infected versus WT virus-infected tumors ( including IFI44 , OASL , ISG15 , IFIT2 , IFIT3 ) are shown in Fig 10B . Increased ISG15 protein expression in Δ3C-induced tumors was confirmed using IHC staining ( Fig 10C ) . qPCR analysis of IFNα and IFNβ transcripts in cDNA derived from Δ3C virus-infected versus WT virus-infected tumors revealed increased expression of IFNα ( but not IFNβ ) in the Δ3C virus-infected tumors ( Fig 10D ) . Since EBV infection of B cells in vitro induces a type 1 interferon response [70 , 71] , and type I interferon inhibits EBV-induced transformation of B cells in vitro [72] , these data suggest that EBNA3C may help to repress this response in EBV-infected lymphomas in vivo . Additional studies are required to identify the mechanism ( s ) underlying this effect and to determine if the source of the interferon signal is primarily B cell- and/or T cell-derived in our model . Molecular signature analysis of the RNA-seq data also showed a marked increase in pathways associated with T-cell activity in the Δ3C-induced tumors ( Fig 11A ) , consistent with our histologic findings . Differentially regulated genes in one of the molecular signatures , the “GO_Immune_Response gene list” included genes expressed in cytotoxic T cells , including CD8A , perforin1 , and granzyme B , as well as T-cell chemokines including CCL5 ( RANTES ) , CCL20 , and CCL22 ( Fig 11B ) . The increase in CD8A expression is consistent with our finding that Δ3C-induced tumors have a greater number of infiltrating T cells ( Fig 4 ) . We confirmed increased expression of the CD8A , perforin1 , granzyme B , CCL5 , and CCL20 genes in the Δ3C-induced tumors by performing qPCR analysis on cDNA isolated from tumor tissues ( Fig 11C–11G ) . In addition , we performed IHC co-staining for CCL5 and EBNA2 , and detected cells expressing both CCL5 and EBNA2 in Δ3C-induced tumors ( Fig 11H ) . These results suggest that EBNA3C inhibits T-cell responses to EBV-infected B cells by blocking expression of chemokines that attract T cells .
EBNA3C is expressed in human B-cell lymphomas that have type III latency , and in a subset of human Burkitt lymphomas [6 , 73] . The EBV EBNA3C protein plays an essential role in promoting EBV transformation of B cells into LCLs in vitro , and this effect is largely due to the ability of EBNA3C ( in collaboration with EBNA3A ) to inhibit expression of the CDKN2A tumor suppressor locus . Nevertheless , most EBV-infected lymphomas in humans have restrictive forms of viral latency that do not express EBNA3C [74] , indicating that EBV-infected lymphoma cells in vivo can sometimes proliferate even without EBNA3C . Furthermore , although EBNA3C regulates numerous different cellular genes in addition to CDKN2A , whether these other EBNA3C-regulated cellular genes have important in vivo functions remains unclear . Here we have used a cord blood-humanized mouse model to examine the phenotype of an EBNA3C-deleted EBV mutant in vivo in the presence of human T cells . We find that EBNA3C is not essential for the ability of EBV to cause lymphomas in this model , although mice infected with this mutant have fewer tumors compared to mice infected with wild-type virus , and tumors occur at later time points . Furthermore , we demonstrate that the monoclonal B-cell lymphomas induced by the EBNA3C-deleted mutant occur despite high-level p16 expression and reduced activation of E2F-responsive cellular genes . In addition , we find that loss of EBNA3C expression in EBV-induced lymphomas in vivo is associated with enhanced expression of T-cell chemokines , increased T-cell infiltration , and increased type I interferon signaling . These results suggest that EBNA3C is not only critical for reducing p16 expression , but may also affect the T-cell response and type I interferon expression in EBV-infected lymphomas . Our results confirm that EBNA3C inhibits p16 expression in EBV-infected B cells and show that loss of EBNA3C significantly attenuates the ability of EBV to form lymphomas in cord blood-humanized mice . Nevertheless , the finding that a subset of mice infected with the EBNA3C-deleted virus eventually develop aggressive lymphomas suggests that these tumors have developed mechanism ( s ) to at least partially thwart the growth inhibitory effect of p16 . Although inhibition of p16 transcription via promoter DNA methylation is a relatively common mechanism used by human tumors to prevent p16 accumulation [75] , we found that Δ3C virus-infected lymphomas continued to express high-level p16 , and thus must have developed other mechanisms to circumvent its growth inhibitory effect . However , this bypass of p16 function was not complete , since a major GSEA signature in Δ3C virus-infected tumors was reduced expression of “Hallmark E2F genes” . This is an expected downstream signature of enhanced p16 expression , since p16 inhibits the ability of cyclin D-CDK4/6 complexes to phosphorylate and inactivate pRb , and pRb inhibits E2F1 transcription . Although the precise mechanism ( s ) that allow Δ3C virus-infected lymphomas to proliferate despite high-level p16 expression are not yet fully understood , our results here suggest several ( not mutually exclusive ) possibilities . Cyclin E/CDK2 complexes ( which are not inhibited by p16 ) , like cyclin D-CDK4/6 complexes , can phosphorylate and inactivate pRb , and both human melanomas and DLBCLs expressing high-level p16 have been reported to escape p16 repression by activating cyclin E expression [56–58] . High level c-Myc can likewise allow tumors to escape p16 effects [58] . Our RNA-seq results suggest that cyclin E and c-Myc expression is not significantly lower in Δ3C virus-infected lymphomas versus WT virus-infected lymphomas , and our IHC results showed that both cyclin E and c-Myc are expressed at the protein level in Δ3C virus-infected lymphomas ( Fig 2D ) . Thus , continued c-Myc and/or cyclin E expression likely plays a key role in allowing the Δ3C virus-infected lymphomas to proliferate . EBNA2 activates c-Myc in EBV-infected B cells [9] , and we found that Δ3C virus-infected lymphoma cells express at least as much EBNA2 as the wild-type virus-infected lymphomas at both the RNA ( Figs 7 and 8 ) and protein ( Fig 2A and 2B ) levels . Thus , EBNA2 might promote growth of Δ3C virus-infected lymphomas , even in the presence of high-level p16 , by inducing c-Myc ( and downstream cyclin E ) expression . Since high-level c-Myc expression ( as observed in both the wild-type virus-infected , and Δ3C virus-infected , lymphomas ) promotes apoptosis , our findings here also indicate that Δ3C virus-infected lymphomas have developed mechanism ( s ) to escape c-Myc-induced apoptosis in the absence of EBNA3C . EBNA3C inhibits expression of BIM , a major tumor suppressor for c-Myc-induced B-cell lymphomas , in Burkitt lymphoma cell lines in vitro [42] . BIM promotes apoptosis by sequestering anti-apoptosis proteins in the BCL2 family . Although we found that BIM level was significantly decreased in the WT virus-infected versus Δ3C virus-infected lymphomas , BIM was not expressed in most EBV-infected B cells in the Δ3C virus-infected lymphomas , despite the high-level c-Myc expression . Cellular BIM level is controlled through multiple different pathways , including phosphorylation by ERK kinases that promote Bim degradation [76 , 77] , and both LMP1 and LMP2A can activate ERKs [78 , 79] . LMP1 also activates the NF-kB pathway , which protects B cells from apoptosis by inducing expression of the anti-apoptotic proteins , BCL2 and IRF4 . We found that BCL2 and IRF4 are highly expressed in both Δ3C virus-infected , and WT virus-infected lymphomas , even though EBNA3C stabilizes IRF4 protein in vitro [43] . These results suggest that the anti-apoptotic functions of other EBV proteins expressed in B cells with type III latency ( including LMP1 , LMP2A and EBNA3A [80] ) at least partially substitute for the anti-apoptotic effect of EBNA3C . In addition , increased transcription of the EBV BCL2 homologue , BHRF1 , in Δ3C virus-infected lymphomas ( as suggested in Figs 7 and 8 ) could potentially enhance expression of this latent viral protein may also compensate for loss of EBNA3C expression by inhibiting apoptosis [81] . However , we have been unable to confirm that Δ3C virus-infected lymphomas have increased BHRF1 protein expression by either IHC analysis ( perhaps due to inefficient detection with currently available antibodies using IHC staining ) or immunoblot analysis ( S7 Fig ) . In this study , we also demonstrated that EBV-induced tumors in the cord blood-humanized mouse model ( both wild-type EBV- and Δ3C virus-induced ) are dominated by expansions of a limited number of individual B cells , rather than consisting of highly heterogeneous expansions of a broad selection of different EBV-infected B cells . This finding strongly suggests that the Δ3C virus is driving development of these malignant lymphomas and is not simply persisting as a “passenger virus” in a non-malignant B-cell inflammatory infiltrate . The absence of B-cell lymphomas in NSG mice injected with mock-infected cord blood also confirms that EBV infection is essential for these lymphomas . Interestingly , although the frequency of IGHV2-5 usage in IGH genes expressed in normal cord blood is less than 1% [66] , IGHV2-5 usage was observed in the dominant IGH clones in 3 of the 6 tumors examined in this study ( Fig 5 ) . IGHV1-69-containing IGH transcripts ( obtained from two separate donors ) were also over-represented in the EBV-induced tumors in this study relative to the frequency of such transcripts in normal cord blood . These results suggest that expression of specific IGH variable regions may cooperate with EBV infection to induce lymphomas in the cord blood-humanized mouse model , particularly since specific variable IGH gene usage ( and specific CDR3 sequences ) support the growth of human DLBCLs and CLL tumors by enhancing BCR signaling [82–85] . However , since the antigens recognized by the tumor derived-antibodies in our model are not yet identified , we cannot exclude the possibility that B cells expressing these antibodies are selected for in EBV-infected cord blood-humanized mice because they recognize viral and/or cellular ( including mouse ) antigens that stimulate their proliferation . EBNA3C has been reported to increase AID expression in EBV-infected B cells in vitro [38] , and we found that loss of EBNA3C expression strongly decreased AID expression in EBV-induced lymphomas in the cord blood-humanized model . Although AID is required for IG class-switching and BCR somatic hyper-mutation , we did not find any differences in the amount of class switching or somatic hyper-mutation of antibodies in wild-type virus-infected versus Δ3C virus-infected lymphomas . This lack of difference likely reflects our finding that even the wild-type EBV-infected B cells do not undergo efficient IG class switching or somatic hyper-mutation in the cord blood-humanized mouse model . RNA-seq analysis showed that almost all IG transcripts in both wild-type EBV- and Δ3C virus-infected lymphoma cells were of the IGM rather than IGG types , and our sequencing of the CDR3 motifs of antibodies expressed in lymphomas revealed few if any mutations . The lack of class switching and somatic hyper-mutation in EBV-infected cells may be due to inefficient germinal center-like B cell/T cell interactions in the cord blood-humanized mouse model . Nevertheless , since EBNA3C has been shown to induce somatic hyper-mutation of LCLs in vitro even in the absence of any T cell help [38] , the previously described ability of EBNA2 to inhibit AID expression and class switching [86 , 87] may also explain this result . EBNA2-expressing tonsillar cells in patients with infectious mononucleosis do not undergo somatic hyper-mutation [87] , consistent with our results here . Thus , the relative level of EBNA2 versus EBNA3C protein expression in EBV-infected B cells may determine whether class switching and somatic hyper-mutation occurs . In contrast to the restricted B-cell population in tumors , we found that all tumors contained a variety of TCRs , with no single CDR3 identified in multiple tumors . Nevertheless , we did find that certain TRBV genes were used more frequently in some tumors relative to their use in the normal adult peripheral blood cell population . The antigens recognized by the T cell CDR3 motifs in our model are not yet identified and the TCR sequences we observed do not match previously identified “public” TRBV CDR3 motifs reported to recognize EBV antigens . Further studies will be required to determine if the TRBV CDR3 motifs reported here are unique to specific cord blood donors , and to identify the viral and/or cellular antigens recognized by these T cells . We have also discovered some other , previously unsuspected , potential functions of EBNA3C using the cord blood-humanized mouse model that are difficult to study in vitro . The RNA-seq data revealed that Δ3C-induced tumors have a greatly increased type I interferon response compared to the wild-type virus-infected lymphomas , and this effect was confirmed by both qPCR analysis of tumor cDNA , and by ISG15 IHC staining . We also found that IFNα is expressed at higher levels in the Δ3C-induced tumors . Since we used human-specific primers to quantify interferon alpha expression in tumors ( Fig 10D ) , the source of increased interferon alpha must be a human cell . However , technical issues have limited our ability to specifically isolate enough viable B cells and/or T cells from lymphomas to determine if the enhanced interferon response in Δ3C-induced lymphomas is primarily due to interferon produced in B cells , T cells or other cell types . Single-cell RNAseq analysis of lymphomas in this humanized mouse model ( in which both human B cells are T cells are intermixed ) may help to define the site ( s ) of interferon production in the future . The mechanism ( s ) by which loss of EBNA3C expression leads to an increased interferon response in EBV-infected lymphomas in the cord blood-humanized mouse model is not yet determined . EBER expression in EBV- infected B cells induces strong type I interferon signaling [70 , 71] . In addition , the ability of EBV infection in B cells to globally induce LTR expression of endogenous retroviruses [88] ( a potent stimulator of type I interferon signaling [89 , 90] ) , is another potential mechanism by which EBV may induce this pathway . Although we favor a model whereby EBNA3C primarily attenuates the type I interferon response in EBV-infected B cells , the known ability of EBERs to enter uninfected cells via exosomes [91] may also play a role . Although alpha interferon can also be derived from plasmacytoid dendritic cells [92] , we have found that EBV-induced lymphomas in cord blood-humanized mice contain relatively few human plasmacytoid dendritic cells . Since interferon alpha inhibits the ability of EBV to transform B cells in vitro [72] , and blocks its ability to lytically reactivate [93] , the ability of EBNA3C to decrease interferon alpha production may enhance both latent and lytic infection . In addition , decreased interferon production in EBV-infected B cells likely benefits the virus in vivo by decreasing inflammation and T-cell responses . Loss of EBNA3C expression also resulted in an increased number of tumor-infiltrating T cells as determined by IHC staining , and RNA-seq data showed increased expression of genes associated with cytotoxic T cells ( including CD8 , perforin , and granzyme B ) . RNA-seq analysis also revealed that several different T-cell chemokines ( including CCL5 , CCL20 , and CCL22 ) are over-expressed in the Δ3C virus-infected , versus WT virus-infected , lymphomas , a finding confirmed by qPCR analysis of tumor cell-derived cDNA . In addition , CCL5 ( RANTES ) was shown to be expressed at the protein level in Δ3C virus-infected B cells by IHC . These results suggest that EBNA3C inhibits expression of CCL5 in EBV-infected B cells in the cord blood-humanized mouse model . In contrast , we did not find that several different T-cell chemokines previously shown to be inhibited by EBNA3C in an EBV-negative Burkitt cell line in vitro ( CCL3 , CCL4 , CXCL10 and CXCL11 ) [49] were affected by EBNA3C loss in the CBH model , suggesting that the precise effects of EBNA3C may be context-specific . Together , the previous in vitro studies , and our results here , suggest that the increased expression of chemokines that attract T cells in Δ3C virus-infected B cells may be a major mechanism leading to increased T-cell infiltration of Δ3C virus-infected lymphomas . However , since the Δ3C virus-infected lymphomas occur at later time points ( 60–90 days ) compared to WT virus-induced lymphomas ( 30–35 days ) , we cannot exclude the possibility that the increased time required for lymphomagenesis allows for improved anti-EBV T cell responses , and thus increased T-cell infiltration into lymphomas . Finally , we suspect that the in vivo tumor microenvironment in cord blood-humanized mice may also play a critical role in promoting the development of Δ3C virus-induced tumors . We previously demonstrated that CD40L-expressing CD4 T cells can substitute for LMP1 expression in the cord blood-humanized mouse model [51] . Future studies may identify additional growth factors provided by the in vivo tumor microenvironment that allow EBV-infected B cells with stricter forms of viral latency ( where the EBV-encoded oncoproteins such as EBNA2 , LMP1 and EBNA3C are not expressed ) to proliferate in less immunogenic forms of EBV infection .
All animal work experiments were approved by the University of Wisconsin-Madison Institutional Animal Care and Use Committee ( IACUC ) and conducted in accordance with the NIH Guide for the care and use of laboratory animals ( protocol numbers M005197 and M005214 ) . We anesthetized mice using isoflurane and euthanized animals by performing cervical dislocations on anesthetized mice [53] . All experiments performed in this study were in the context of the p2089 ( B95 . 8 EBV ) BACmid . p2089 , which expresses green fluorescent protein ( GFP ) and a hygromycin resistance gene has been described previously [94] . The ΔEBNA3C BACmid was constructed via the GS1783 E . coli-based En Passant method [95 , 96] by inserting a single nucleotide substitution changing the second amino acid residue to a stop codon . Details on mutagenesis methodology and subsequent derivation of all EBV-positive HEK293 cell lines are described in [97] . Note that because the BM2710 invasive E . coli ( used to infect HEK293 cells ) are resistant to Chloramphenicol , the BACmids were made Kanamycin-resistant pre-transfer to the BM2710 E . coli . To construct ΔEBNA3C-Revert , all steps were reversed starting with the Kanamycin resistant ΔEBNA3C BACmid . The Chloramphenicol insert was created by amplifying the chloramphenicol gene with the corresponding homology arms from the parental p2089 BACmid . All primers used for construction and confirmation of ΔEBNA3C and the corresponding revertant ( ΔEBNA3C-Revert ) are provided in Table 1 . HEK293 cells latently infected with WT EBV or EBNA3C-mutant virus were maintained in Dulbecco modified Eagle medium ( DMEM ) supplemented with 10% fetal bovine serum ( FBS ) , and 1% penicillin-streptomycin ( pen-strep ) . and 100 ug of Hygromycin B . Infectious viral particles were produced from 293 cell lines stably infected with the WT or mutant viruses following transfection with EBV BZLF1 , BRLF1 and gp110 expression vectors as previously described [53] . The titer of EBV was determined on Raji cells by using the green Raji cell assay as previously described [53] . Immunodeficient NSG ( NOD/LtSz-scid/IL2Rγnull ) mice were purchased from Jackson labs ( catalogue number: 005557 ) . Commercially available CD34-depleted human cord blood mononuclear cells ( AllCells ) were infected with 5000 Green Raji Units of wild-type or mutant EBV strains in vitro for 1 ½ hours at 37°C , and a minimum of 10 million cells were injected intraperitoneally into 3 to 5-week-old NSG mice which were age-matched [51 , 53] . Mice were kept for a maximum of 90 days after injection of cord blood , or were euthanized due to tumor symptoms prior to day 90 . Studies were performed using three different donors . Following euthanasia , multiple different organs ( including the lungs , spleen , pancreas , liver , gallbladder , mesenteric fat , and abdominal lymph nodes ) were formalin fixed and then examined by using a variety of techniques to determine if animals had persistent EBV infection and/or EBV-positive lymphomas and to assess the viral protein expression pattern . Samples from all EBV-infected animals were examined by H&E staining to determine if tumors were present and to assess the types of tumors in each animal . Tumors from at least 10 different animals infected with the Δ3C or wild-type viruses also underwent IHC staining by using the antibodies listed in Table 2 , as previously described [53 , 98] . Coauthor Erik A . Ranheim , a board-certified hematopathologist , performed the pathological analysis of the tumors . In some animals , EBER in situ hybridization studies were performed by using the PNA ISH detection kit ( DakoCytomation ) as previously described [53] . For quantification of IHC results , at least 2 random fields of view were selected per animal , photographed , and then counted by 2 independent observers . The counts of positively staining cells were averaged across observers . For each ratio ( i . e . , CD3 to CD20 ) , stains were done on adjacent slides , and counts were completed on the same field of view . These analyses were performed on 5 WT tumors , 5 revertant virus tumors , and 10 Δ3C virus tumors ( S3 Table ) , using which tumors from all three donors . Frozen tumor samples were homogenized using Cellcrusher tissue pulverizer system ( Cellcrusher ) as per the manufacturer’s instructions . Immunoblotting was performed as described previously [99] . Briefly , tumor tissue was lysed in SUMO buffer plus protease inhibitors and then sonicated and centrifuged . Equivalent amounts of protein were separated in sodium dodecyl sulfate-10% or 15% polyacrylamide gel electrophoresis ( SDS-PAGE ) gels and transferred to nitrocellulose membranes . Membranes were blocked in 5% milk and then incubated with the appropriate primary antibodies diluted in 5% milk in 1X PBS and 0 . 1% Tween 20 ( PBS-T ) . The primary antibodies used are described in Table 2 . After being washed , the membranes were incubated with the appropriate horseradish peroxidase-conjugated secondary antibodies ( Pierce , Waltham , MA ) in 5% milk–1X PBS-T for 1 h at room temperature and then washed again . Bound antibodies were visualized by use of enhanced chemiluminescent reagent ( Pierce ) according to the manufacturer’s instructions . RNA samples were harvested from frozen primary tumor tissue by homogenizing tissue with the Covaris Cryoprep Pulverizer . The pulverized tissue was then lysed in Trizol reagent ( Invitrogen ) . RNA was isolated using phenol/chloroform extraction and total RNA was quantitated using a NanoDrop 2000 Spectrophotometer machine ( ThermoFisher ) . RNA-seq libraries were prepared using the TruSeq Stranded mRNA Prep kit ( illumina ) . Libraries were sequenced using an illumina HiSeq2500 platform at the University of Wisconsin Biotechnology Center DNA Sequencing Facility . RNA Seq data analysis was conducted by BioInfoRx ( Madison , WI ) . Briefly , the fastQC program was used to verify raw data quality of the Illumina reads . The sequence data were mapped to human , mouse and virus genomes separately . The hg19 human genome and Ensembl gene annotations ( v75 ) , GRCm38 ( mm10 ) mouse genome and Ensembl gene annotations ( v82 ) were used for mapping . The raw sequence reads were mapped to the genome using Subjunc aligner from Subread [100]; then , we took the human and mouse alignment files ( bam files ) and assigned a read to one of the following categories ( in the order listed below ) : 1 . Mapped to both ( mapping scores for both human and mouse >20 , the difference between human and mouse mapping scores < = 10 ) ; 2 . Mapped mostly to human ( human mapping score>20 , mapping to human genome better than to mouse genome ( higher mapping score and fewer mismatches ) ) ; 3 . Mapped mostly to mouse ( mouse mapping score>20 , mapping to mouse genome better than to human genome ( higher score and fewer mismatches ) ) ; 4 . Mapped to human ( human mapping score minus mouse mapping score >10 ) ; 5 . Mapped to mouse ( mouse mapping score minus human mapping score >10 ) ; 6 . No mapping ( both mapping scores are 0 ) ; 7 . Others . Reads from categories 2 and 4 were combined as human only reads ( 17–42% of all reads ) , and reads from categories 3 and 5 were combined as mouse only reads ( 18–42% of all reads ) . These mouse only and human only alignment bam files were compared against corresponding gene annotation GFF files , and raw counts for each gene were generated using the featureCounts tool from Subread , with around 30–47% of reads overall assigned to human genes , and around 31–51% of reads overall assigned to mouse genes . The raw counts data were normalized using the TMM normalization method [101] in the program edgeR , and the normalized gene counts were transformed to log2 scale using the voom method from the R Limma package [102] , then used for differential expression analysis . Functional interpretation of the differentially expressed genes was conducted based on GO terms , KEGG pathway and GSEA [103] methods . EBV transcripts were analyzed by aligning the fastq files to an indexed B95-8 EBV genome using Burrows-Wheeler Aligner ( BWA ) [104] . SAMtools was used to generate sorted BAM files . A pileup of aligned reads was constructed as a Wig file using a python script . RPKM measurements for EBV genes were derived by normalizing the number of aligned reads to the total number of reads aligned to human or EBV transcripts . cDNA was made from RNA collected from frozen tumor tissues using methods described above . The extracted RNA was then treated with DNase , followed by reverse transcription using random primers and GoScript reverse transcriptase ( RT ) ( Promega ) . Real-time PCR was performed on the reverse-transcribed cDNA by using iTaq Universal SYBR green mix ( Bio-Rad ) in a Bio-Rad CFX96 machine . cDNA ( 1 uL ) was used for 40 cycles of 15 s at 95°C and 30 s at 60°C [105] . The CFX Maestro software ( Biorad ) was used to collect Cq values . Cq values were either normalized to the CD20 ( MS4A1 ) value ( EBV genes and B-specific genes ) or total GAPDH . Delta Cq values were determined by calculating 2^ ( Normalized Cq ) which were then plotted . All primers used in qPCR analysis are listed in Table 3 [105 , 106] . FASTQ files were first processed with Trim Galore ! v0 . 4 . 4 ( Cutadapt v1 . 14 ) with default parameters to remove adapter sequences and low quality bases and reads . Following trimming , BCR and TCR clonotypes were predicted using miXCR ( v2 . 1 . 11 ) as described in the documentation [107] . MiXCR extendAlignments was used to extend incomplete TCR CDR3 sequences using germline annotations . The amount of somatic hyper-mutation in IGHV sequences was analyzed using IMGT/Vquest software ( http://www . imgt . org/IMGT_vquest/vquest ) . All bar graphs were constructed in Microsoft Excel and standard error was used for error bars in graphs . Kaplan-Meier analysis and dot plots were constructed in MSTAT statistical software . Fisher Exact , Log-Rank analysis and Wilcoxon rank-sum test was performed using MSTAT statistical software ( https://mcardle . wisc . edu/mstat/ ) . A p-value of < . 05 was considered significant in all tests used . The RNA-seq data reported in this paper have been deposited in the GEO database and are under the GEO accession number GSE113070 .
While this manuscript was being revised , another group reported that an EBNA3C-deleted virus is able to establish latency in another humanized mouse model [108] . | Epstein-Barr virus is associated with multiple human B-cell malignancies and transforms B cells in vitro into immortalized cell lines . The EBV protein , EBNA3C , inhibits expression of the tumor suppressor protein , p16 , and is required for EBV transformation of B cells in vitro . Here , we show that an EBNA3C-deleted virus ( Δ3C ) has a reduced ability ( in comparison to the wild-type control virus ) to cause B-cell lymphomas in a cord blood-humanized mouse model , although a subset of infected animals develop lymphomas at later time points . These tumors had high-level expression of the cell cycle inhibitor , p16 , yet also expressed the cell cycle activators , c-Myc and cyclin E . Tumors also expressed pro-survival proteins , including BCL2 and IRF4 . Unexpectedly , Δ3C tumors had increased T-cell infiltration , enhanced expression of several different T-cell chemokines , and a RNA-seq signature suggestive of an increased type 1 Interferon response . These results reveal that EBNA3C promotes EBV-induced lymphomas in the cord blood-humanized mouse model ( but is not absolutely required ) , and suggest that EBNA3C may have immune modulatory effects in vivo . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods",
"Note",
"in",
"Proof"
] | [
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"cell",
"cycle",
"and",
"cell",
"division",
"immunology",
"cell",
"processes",
"cancers",
"and",
"neoplasms",
"oncology",
"hematologic",
"cancers",
"and",
"related",
"disorders",
"animal",... | 2018 | An EBNA3C-deleted Epstein-Barr virus (EBV) mutant causes B-cell lymphomas with delayed onset in a cord blood-humanized mouse model |
Sleep spindles are brief oscillatory events during non-rapid eye movement ( NREM ) sleep . Spindle density and synchronization properties are different in MEG versus EEG recordings in humans and also vary with learning performance , suggesting spindle involvement in memory consolidation . Here , using computational models , we identified network mechanisms that may explain differences in spindle properties across cortical structures . First , we report that differences in spindle occurrence between MEG and EEG data may arise from the contrasting properties of the core and matrix thalamocortical systems . The matrix system , projecting superficially , has wider thalamocortical fanout compared to the core system , which projects to middle layers , and requires the recruitment of a larger population of neurons to initiate a spindle . This property was sufficient to explain lower spindle density and higher spatial synchrony of spindles in the superficial cortical layers , as observed in the EEG signal . In contrast , spindles in the core system occurred more frequently but less synchronously , as observed in the MEG recordings . Furthermore , consistent with human recordings , in the model , spindles occurred independently in the core system but the matrix system spindles commonly co-occurred with core spindles . We also found that the intracortical excitatory connections from layer III/IV to layer V promote spindle propagation from the core to the matrix system , leading to widespread spindle activity . Our study predicts that plasticity of intra- and inter-cortical connectivity can potentially be a mechanism for increased spindle density as has been observed during learning .
Sleep marks a profound change of brain state as manifested by the spontaneous emergence of characteristic oscillatory activities . In humans , sleep spindles consist of waxing-and-waning bursts of field potentials oscillating at 11–15 Hz lasting for 0 . 5–3 s and recurring every 5–15 s . Experimental and computational studies have identified that both the thalamus and the cortex are involved in the generation and propagation of spindles . Spindles are known to occur in isolated thalamus after decortication in vivo and in thalamic slice recordings in vitro [1 , 2] , demonstrating that the thalamus is sufficient for spindle generation . In in-vivo conditions , the cortex has been shown to be actively involved in the initiation and termination of spindles [3] as well as the long-range synchronization of spindles [4] [5] . Multiple lines of evidence indicate that spindle oscillations are linked to memory consolidation during sleep . Spindle density is known to increase following training in hippocampal-dependent [6] as well as procedural memory [7] memory tasks . Spindle density also correlates with better memory retention following sleep in verbal tasks [8 , 9] . More recently , it was shown that pharmacologically increasing spindle density leads to better post-sleep performance in hippocampal-dependent learning tasks [10] . Furthermore , spindle activity metrics , including amplitude and duration , were predictive of learning performance [11–13] , suggesting that spindle event occurrence , amplitude , and duration influence memory consolidation . In human recordings , spindle occurrence and synchronization vary based on the recording modality . Spindles recorded with magnetoencephalography ( MEG ) are more frequent and less synchronized , as compared to those recorded with electroencephalography ( EEG ) [14] . It has been proposed that the contrast between MEG and EEG spindles reflects the differential involvement of the core and matrix thalamocortical systems , respectively [15] . Core projections are focal to layer IV , whereas matrix projections are widespread in upper layers [16] . This hypothesis is supported by human laminar microelectrode data which demonstrated two spindle generators , one associated with middle cortical layers and the other superficial [17] . Taken together , these studies suggest that there could be two systems of spindle generation within the cortex and that these correspond to the core and matrix anatomical networks . However , the network and cellular mechanisms whereby the core and matrix systems interact to generate both independent and co-occurring spindles across cortical layers are not understood . In this study , we developed a computational model of thalamus and cortex that replicates known features of spindle occurrence in MEG and EEG recordings . While our previous efforts have been focused on the neural mechanisms involved in the generation of isolated spindles[5] , in this study we identified the critical mechanisms underlying the spontaneous generation of spindles across different cortical layers and their interactions .
Histograms of EEG and MEG gradiometer inter-spindle intervals are shown in Fig 1C . For neither channel type are ISIs distributed normally as determined by Lilliefors tests ( D2571 = 0 . 1062 , p = 1 . 0e-3 , D4802 = 0 . 1022 , p = 1 . 0e-3 ) , suggesting that traditional descriptive statistics are of limited utility . However , the ISI at peak of the respective distributions is longer for EEG than it is the MEG . In addition , a two-sample Kolmogorov-Smirnov test confirms that EEG and MEG ISIs are not drawn from the same distribution ( D2571 , 4802 = 0 . 079 , p = 1 . 5e-9 ) . While the data where not found to be drawn from any parametric distribution with 95% confidence , an exponential fit ( MEG ) and lognormal fit ( EEG ) are shown in red overlay for illustrative purposes . These data are consistent with previous empirical recordings [18] and suggest that sleep spindles have different properties across superficial vs . deep cortical layers . To investigate the mechanisms behind distinct spindle properties across cortical locations as observed in EEG and MEG signals , we constructed a model of thalamus and cortex that incorporated the two characteristic thalamocortical systems: core and matrix . These systems contained distinct thalamic populations that projected to the superficial ( matrix ) and middle ( core ) cortical layers . Four cell types were used to model distinct cell populations: thalamocortical relay ( TC ) and reticular ( RE ) neurons in the thalamus , and excitatory pyramidal ( PY ) and inhibitory ( IN ) neurons in each of three layers of the cortical network . A schematic representation of the synaptic connections and cortical geometry of the network model is shown in Fig 2 . In the matrix system , both thalamocortical ( from matrix TCs to the apical dendrites of layer 5 pyramidal neurons ( PYs ) located in the layer 1 ) and corticothalamic synapses ( from layer 5 PYs back to the thalamus ) formed diffuse connections . The core system had a focal connection pattern in both thalamocortical ( from core TCs to PYs in the layer III/IV ) and corticothalamic ( from layer VI PYs to the thalamus ) projections . Because spindles recorded in EEG signal reflect the activity of superficial layers while MEG records spindles originating from deeper layers ( Fig 1 and [19] ) , we compared the activity of the model’s matrix system , which has projections to the superficial layers , to empirical EEG recordings and compared the activity in model layer 3/4 to empirical MEG recordings . In agreement with our previous studies [3 , 5 , 20 , 21] , simulated stage 2 sleep consisted of multiple spindle events involving thalamic and cortical neuronal populations ( Fig 3 ) . During one such typical spindle event ( highlighted by the box in Fig 3A and 3B ) , cortical and thalamic neurons in both the core and matrix system had elevated and synchronized firing ( Fig 3A bottom ) , consistent with previous in-vivo experimental recordings [22] . In the model , spindles within each system were initiated from spontaneous activity within cortical layers and then spread to thalamic neurons , similar to our previous study[5] . The spontaneous activity due to miniature EPSPs in glutamergic cortical synapses led to fluctuations in membrane voltage and sparse firing . At random times , the miniature EPSPs summed such that a small number of locally connected PY neurons spiked within a short window ( <100ms ) , which then induced spiking in thalamic cells through corticothalamic connections . This initiated spindle oscillations in the thalamic population mediated by TC-RE interactions as described before [20 , 23 , 24] . Thalamic spindles in turn propagated to the neocortex leading to joint thalamocortical spindle events whose features were shaped by the properties of thalamocortical and corticothalamic connections . In this study , we examined how the process of spindle generation occurs in a thalamocortical network with mutually interacting core and matrix systems , wherein the thalamic network of each system is capable of generating spindles independently . Based on the anatomical data [16] , the main difference between the modeled core and matrix systems was the radii or fanout of connections in thalamocortical and corticothalamic projections ( in the baseline model , the fanout was 10 times wider for the matrix compared to the core system ) . Furthermore , the strength of each synaptic connection was scaled by the number of input connections to each neuron [25 , 26] , leading to weaker individual thalamocortical projections in the matrix as compared to the core . These differences in the strength and fanout of thalamocortical connectivity resulted in distinctive core and matrix spindle properties ( see Fig 3A , right vs left ) . First , both cortical and thalamic spindles were more spatially focal in the core system as only a small subset of neurons was involved in a typical spindle event at any given time . In contrast , within the matrix system spindles were global ( involving the entire cell population ) and highly synchronous across all cell types . These results are consistent with our previous studies [5] and suggest that the connectivity properties of thalamocortical projections determine the degree of synchronization in the cortical network . Second , spindle density was higher in the core system compared to the matrix system . At every spatial location in the cortical network of the core system , the characteristic time between spindles was shorter compared to that between spindles in the matrix system ( Fig 3A left vs right ) . In order to quantify the spatial and temporal properties of spindles , we computed an estimated LFP as an average of the dendritic synaptic currents for every group of contiguous 100 cortical neurons . LFPs of the core system were estimated from the currents generated in the dendrites of layer 3/4 neurons while the LFP of the matrix system was computed from the dendritic currents of layer 5 neurons , located in the superficial cortical layers ( Fig 2 ) . After applying a bandpass filter ( 6–15 Hz ) , the spatial properties of estimated core and matrix LFP ( Fig 3C ) closely matched the MEG and EEG recordings , respectively ( Fig 1 ) . In subsequent analyses , we used this estimated LFP to further examine the properties of the spindle oscillations in the core and matrix systems . We identified spindles in the estimated LFP using an automated spindle detection algorithm similar to that used in experimental studies ( details are provided in the method section ) . The spindle density , defined as the number of spindles occurring per minute of simulation time , was greater in the core compared to the matrix ( Fig 4A ) as confirmed by an independent-sample t-test ( t ( 18 ) = 7 . 06 , p<0 . 001 for across estimated LFP channels and t ( 2060 ) = 19 . 2 , p<0 . 001 across all spindles ) . The results of this analysis agree with the experimental observation that MEG spindles occur more frequently than EEG spindles . While the average spindle density was significantly different between the core and matrix , in both systems the distribution of inter-spindle intervals peaks below 4 seconds and has a long tail ( Fig 4B ) . A two sample KS test comparing the distributions of inter-spindle intervals confirmed that the intervals were derived from different distributions ( D1128 , 932 = 0 . 427 , p<0 . 001 ) . The peak ISI of the core was shorter than that of the matrix system , suggesting that the core network experiences shorter and more frequent quiescence periods than the matrix population . Furthermore , maximum-likelihood fits of the probability distributions ( red line in Fig 4B ) confirmed that the intervals of spindle occurrence cannot be described by a normal distribution . The long tails of the distributions suggest that a Poisson like process , as oppose to a periodic process , is responsible for spindle generation . This observation is consistent with previous experimental results [18 , 27] and suggests that our computational model replicates essential statistical properties of spindles observed in in vivo experiments . Several other features of simulated core and matrix spindles were similar to those found in experimental recordings . The average spindle duration was significantly higher in the core compared to the matrix system ( Fig 4C ) as confirmed by independent-sample t-test ( t ( 2060 ) = 16 . 3 , p<0 . 001 ) . To quantify the difference in the spatial synchrony of spindles between the core and matrix systems , we computed the spatial correlation [28] between LFP groups at different distances ( measured by the location of a neuron group in the network ) . The correlation strength decreased with distance for both systems ( Fig 4D ) . However , the spindles in the core system were less spatially correlated overall when compared to spindles in the matrix system . Simultaneous EEG and MEG measurements have found that about 50% of MEG spindles co-occur with EEG spindles , while about 85% of EEG spindles co-occur with MEG spindles [29] . Further , a spindle detected in the EEG signal is found to co-occur with about 66% more MEG channels than a spindle detected in MEG . Our model generates spindling patterns consistent with these features . The co-occurrence probability revealed that during periods of spindles in the matrix system , there was about 80% probability that core was also generating spindles ( Fig 4E ) . In contrast , there was only a 40% probability of observing a matrix spindle during a core system spindle . An independent-sample t-test confirmed this difference between the systems across estimated LFP channels ( t ( 14 ) = 31 . 4 , p<0 . 001 ) . Furthermore , we observed that the number of LFP channels that were simultaneously activated during a spindle event in the core system was higher when a spindle co-occurred in the matrix versus times when the spindles only occurred in the core ( Fig 4F , t ( 14 ) = 67 . 2 , p<0 . 001 ) . This suggests that the co-occurrences of spindles in both systems are rare events but lead to the wide spread activation in both the core and matrix when they take place . Finally , we examined the delay between spindles in the core and matrix systems ( Fig 4G ) . We observed that on average ( red line in Fig 4G ) , the spindle originated from the core system then spread to the matrix system with a mean delay of about 300 ms ( delay was measured as the difference in onset times between co-occurring spindles within a window of 2 , 500 ms; negative delay values indicate spindles in which the core preceded matrix ) . The peak at -750 ms corresponds to spindles originating from the core system , while the peak at +750 ms suggests that at some network sites , spindles originated in the matrix system and then spread to the core system . While there were almost no events in which the matrix preceded the core by more than 1 sec ( right of Fig 4G ) , many events occurred in which the core preceded the matrix by more than 1 sec ( left of Fig 4G ) . In sum , these results suggest that spindles were frequently initiated locally in the core system , then propagate to and spread throughout the matrix system . This can trigger spindles at the other locations of the core , so eventually , even regions in the core system that were not previously involved become recruited . These findings explain the experimental result that spindles are observed in more MEG channels when they also co-occur in the EEG [29] . We leveraged our model to examine factors that may influence spindle occurrence across cortical layers . The main difference between the core and matrix systems in the model was the breadth or fanout of the thalamic projections to the cortical network . Neuroanatomical studies suggest that the core system has focused projections while matrix system projects widely [16] . Here , we assessed the impacts of this characteristic by systematically varying the connection footprint of the thalamic matrix to superficial cortical regions , while holding the fanout of the thalamic core to layer 3/4 projections constant . We also modulated the corticothalamic projections in proportion to the thalamocortical projections . Using the estimated LFP from the cortical layers corresponding to core and matrix system , respectively , we quantified various spindle properties as the fanout was modulated . Spindle density ( the number of spindles per minute ) in both layers was sensitive to the matrix system’s fanout . ANOVA confirmed significant effects of fanout and layer location , as well as an interaction between layer and fanout ( fanout: F ( 6 , 112 ) = 66 . 4; p<0 . 01 , Layer: F ( 1 , 112 ) = 65 . 18; p<0 . 01 and interaction F ( 6 , 112 ) = 22 . 8; p<0 . 01 ) . When the matrix and core thalamus had similar fanouts ( ratio 1 and 2 . 5 in Fig 5B ) , we observed a slightly higher density of spindles in the matrix than in the core system . This observation is consistent with the properties of these circuits ( see Fig 2 ) , wherein the matrix system contains direct reciprocal projections connecting cortical and thalamic subpopulations and the core system routes indirect projections from cortical ( layer III/IV ) neurons through layer VI to the thalamic nucleus . When the thalamocortical fanout of the matrix system was increased to above ~5 times the size of the core system , the density of spindles in the matrix system was reduced to around 4 spindles per minute . Interestingly , the density of spindles in the core system was also reduced when the thalamocortical fanout of the matrix system was further increased to above ~10 times of that in the core system ( ratio above 10 in Fig 5B ) . This suggests that spindle density in both systems is determined not only by the radius of thalamocortical vs . corticothalamic projections , but also by interactions between the systems among the cortical layers . We further expound on the role of these cortical connections in the next section . We also examined the effect of thalamocortical fanout on the distribution of inter-spindle intervals ( Fig 5C ) . Although the mean value was largely independent of the projection radius , a long tailed distribution was observed for all values of fanout in the core . Contrastingly , in the matrix system the mean and peak of the inter-spindle interval shifted to the right ( longer intervals ) with increased fanout . With large fanouts , the majority of matrix system spindles had very long periods of silence ( 10-15s ) between them . This suggests that thalamocortical fanout determines the peak of the inter-spindle interval distribution , but does not alter the stochastic nature of spindle occurrence . The degree of thalamocortical fanout also influenced the co-occurrence of spindles in the core and matrix systems ( Fig 5D ) . Increasing the fanout of the matrix system reduced spindle co-occurrence between two systems . This reduction resulted mainly from lower spindle density in both layers . However , the co-occurrence of core spindles during matrix spindles was higher for all values of fanout when matrix thalamocortical projections were at least 5 times broader than core projections . This suggests that the difference in spindle co-occurrence between EEG and MEG as observed in experiments [14] depends mainly on the difference in the radius of thalamocortical projections between the core and matrix systems , while overall level of co-occurrence is determined by the interaction between cortical layers . We examined how spatial correlations during periods of spindles vary depending on the fanout of thalamocortical projections . The spatial correlation quantifies the degree of synchronization in the estimated LFP signals of network locations as a function of the distance between them . As expected , increasing the distance reduced the spatial correlation ( Fig 4D ) . We next measured the mean value of the spatial correlation for each fanout condition . The mean correlation increased as a function of the fanout in the matrix system ( Fig 5A ) . However , the spatial correlation within the core , and between the core and matrix systems , did not change with increases in the fanout , suggesting that the spatial synchronization of core spindles is largely influenced by thalamocortical fanout but not by interactions between the core and matrix systems as was observed for spindle density . Does intra-cortical excitatory connectivity between layer 3/4 of the core system and layer 5 of the matrix system affect spindle occurrence ? To answer this question , we first varied the strength of excitatory connections ( AMPA and NMDA ) from the core to matrix pyramidal neurons ( Fig 6A and 6B ) . Here the reference point ( or 100% ) corresponds to the strength used in previous simulations , i . e . half the strength of a within-layer connection . The spindle density varied with the strength of the interlaminar connections ( Fig 6A ) . For low connectivity strengths ( below 100% ) , the spindle density of the matrix system was reduced significantly , while at high strengths ( above 140% ) the matrix system spindle density exceeded that of the control ( 100% ) . There were significant effects of connection strength and layer on the spindle density , as well as an interaction between the two factors ( connection strength: F ( 5 , 96 ) = 24 . 7; p<0 . 01 , layer: F ( 5 , 96 ) = 386 . 6; p<0 . 01 and interaction F ( 5 , 96 ) = 36 . 9; p<0 . 01 ) that suggests a layer-specific effect of modulating excitatory interlaminar connection strength . Similar to the spindle density , spindle co-occurrence between the core and matrix systems also increased as a function of interlaminar connection strength , reaching 80% for the both core and matrix at 150% connectivity . In contrast , changing the strength of excitatory connections from layer 5 to layer 3/4 had little effect on the spindle density , ( Fig 6C ) . Taken together , these results suggest that the strength of the cortical core-to-matrix excitatory connections is one of the critical factors in determining spindle density and co-occurrence among spindles across both cortical lamina and the core/matrix systems .
Using computational modeling and data from EEG/MEG recordings in humans we found that the properties of sleep spindles vary across cortical layers and are influenced by thalamocortical , corticothalamic and cortico-laminar connections . This study was motivated by empirical findings demonstrating that spindles measured in EEG have different synchronization properties from those measured in MEG [14 , 29] . EEG spindles occur less frequently and more synchronously in comparison to MEG spindles . Our new study confirms the speculation that anatomical differences between the matrix thalamocortical system , which has broader projections that target the cortex superficially , and the core system , which consists of focal projections which target the middle layers , can account for the differences between EEG and MEG signals . Furthermore , we discovered that the strength of corticocortical feedforward excitatory connections from the core to matrix neurons determines the spindle density in the matrix system , which predicts a specific neural mechanism for the interactions observed between MEG and EEG spindles . There were several novel findings in this study . First , we developed a novel computational model of sleep spindling in which spindles manifested as a rare but global synchronous occurrence in the matrix pathway and a frequent but local occurrence in the core pathway . In other words , many spontaneous spindles occurred locally in the core system but only occasionally did this lead to globally organized spindles appearing in the matrix system . As a result , only a fraction of spindles co-occurred between the pathways ( about 80% in matrix and 40% in core pathway ) . This is consistent with data reported for EEG vs MEG in vivo ( Fig 1 ) . In contrast , in our previous models [3 , 5] , spindles were induced by external stimulation and always occurred simultaneously in the core and matrix systems , but with different degrees of internal synchrony . In addition , these studies did not examine how the core and matrix pathways interact during spontaneously occurring spindles . Second , in this study we found that the distribution of the inter-spindle intervals between spontaneously occurring spindles in both the core and matrix pathways had long tails similar to a log-normal distribution . This result is consistent with analyses of MEG and EEG data reported in this study and in our prior study [18] . In our previous models [3 , 5] , spindles were induced by external stimulation and the statistics of spontaneously occurring spindles could not be explored . Third , we demonstrated that the strength of thalamocortical and corticothalamic connections determined the density and occurrence of spontaneously generated spindles . The spindle density was higher in the core pathway as compared to the matrix pathway with high co-occurrence of core spindles with matrix spindles . These findings were corroborated with experimental evidence from EEG/MEG recordings . Finally , we reported that laminar connections between the core and matrix could be a significant factor in determining spindle density , suggesting a possible mechanism of learning . When the strength of these connections was increased in the model , there was a significant increase in spindle occurrence , similar to the experimentally observed increase in spindle density following recent learning [10] . The origin of sleep spindle activity has been linked to thalamic oscillators based on a broad range of experimental studies [2 , 30 , 31] . The excitatory and inhibitory connections between thalamic relay and reticular neurons are critical in generating spindles [20 , 23 , 32 , 33] . However , in intact brain , the properties of sleep spindles are also shaped by cortical networks . Indeed , the onset of a spindle oscillation and its termination are both dependent on cortical input to the thalamus [3 , 34 , 35] . In model studies , spindle oscillations in the thalamus are initiated when sufficiently strong activity in the cortex activates the thalamic network , and spindle termination is partially mediated by the desynchronization of corticothalamic input towards the end of spindles [3 , 32] . However , in simultaneous cortical and thalamic studies in humans , thalamic spindles were found to be tightly coupled to a preceding downstate , which in turn was triggered by converging cortical downstates [36] . Further modeling is required to reconcile these experimental results . In addition , thalamocortical interactions are known to be integral to the synchronization of spindles [5 , 33] . In our new study , the core thalamocortical system revealed relatively high spindle density produced by focal and strong thalamocortical and corticothalamic projections . Such a pattern of connectivity between core thalamus and middle cortical layers allowed input from a small region of the cortex to initiate and maintain focal spindles in the core system . In contrast , the matrix system had relatively weak and broad thalamocortical connections requiring synchronized activity in broader cortical regions in order to initiate spindles in the thalamus . We previously reported [5] that ( 1 ) within a single spindle event the synchrony of the neuronal firing is higher in the matrix than in the core system; ( 2 ) spindle are initiated in the core and with some delay in the matrix system . The overal density of core and matrix spindle events was , however , the same in these earlier models . In the new study we extended these previous results by explaining differences in the global spatio-temporal structure of spindle activity between the core and matrix systems . Our new model predicts that the focal nature of the core thalamocortical connectivity can explain the more frequent occurrence of spindles in the core system as observed in vivo . The strength of core-to-matrix intracortical connections determined the probability of core spindles to “propagate” to the matrix system . In our new model core spindles remained localized and have never involved the entire network , again in agreement with in vivo data . We observed that the distribution of inter-spindle intervals reflects a non-periodic stochastic process such as a Poisson process , which is consistent with previous data [18 , 27] . The state of the thalamocortical network , determined by the level of the intrinsic and synaptic conductances , contributed to the stochastic nature of spindle occurrence . Building off our previous work [21] , we chose the intrinsic and synaptic properties in the model that match those in stage 2 sleep , a brain state when concentrations of acetylcholine and monoamines are reduced [37–39] . As a consequence , the K-leak currents and excitatory intracortical connections were set higher than in an awake-like state due to the reduction of acetylcholine and norepinephrine [40] . The high K-leak currents resulted in sparse spontaneous cortical firing during periods between spindles with occasional surges of local synchrony sustained by recurrent excitation within the cortex that could trigger spindle oscillations in the thalamus . Note that this mechanism may be different from spindle initiation during slow oscillation , when spindle activity appears to be initiated during Down state in thalamus [35] . Furthermore , the release of miniature EPSPs and IPSPs in the cortex was implemented as a Poission process that contributed to the stochastic nature of the baseline activity . All these factors led to a variable inter-spindle interval with long periods of silence when activity in the cortex was not sufficient to induce spindles . While it is known that an excitable medium with noise has a Poisson event distribution in reduced systems [41] , here we show that a detailed biophysical model of spindle generation may lead to a Poission process due to specific intrinsic and network properties . Layer IV excitatory neurons have a smaller dendritic structure compared to Layer V excitatory neurons [42] . Direct recordings and detailed dendritic reconstructions have shown large post-synaptic potentials in layer IV due to core thalamic input [42 , 43] . We examined the role of thalamocortical and corticothalamic connections in a thalamocortical network with only one cortical layer ( S1 Fig ) . We found that increasing the synaptic strength of thalamocortical and corticothalamic connections both increased the density and duration of spindles , however it did not influence their synchronization ( S1A Fig ) . In contrast , changing fanout led to an increase in spindle density , duration , and synchronization . Furthermore , we examined the impact of thalamocortical and corticothalamic connections individually without applying a synaptic normalization rule ( see Methods ) . We observed that the thalamocortical connections had a higher impact on spindle properties than corticothalamic connections ( S1B Fig ) . In our full model with multiple layers , which included a weight normalization rule and wider fanout of the matrix pathway ( based on experimental findings[16] ) , the synaptic strength of each thalamocortical synapse in the core pathway was higher than that in the matrix pathway . The exact value of the synaptic strength was chosen from the reduced model to match experimentally observed spindle durations , as observed in EEG/MEG and laminar recordings [17] . The simultaneous EEG and MEG recordings reported here and in our previous publications [14 , 29] revealed that ( a ) MEG spindles occur earlier compared to the EEG spindles and ( b ) EEG spindles are seen in a higher number of the MEG sensors compared to the spindles occurring only in the MEG recordings . This resembles our current findings , in which the number of regions that were spindling in the core system during a matrix spindle was higher than when there was no spindle in the matrix system . Further , the distribution of spindle onset delays between the systems indicates that during matrix spindles some neurons of the core system fired early , and presumably contributed to the initiation of the matrix spindle , while others fired late and were recruited . Taken together , all the evidence suggests a characteristic and complex spatiotemporal evolution of spindle activity during co-occurring spindles , where spindles in the core spread to the matrix and in turn activate wider regions in the core leading to synchronized activation across cortical layers that is reflected by strong activity in both EEG and MEG . Thus , the model predicts that co-occurring spindles could lead to the recruitment of the large cortical areas , which indeed has been reported in previous studies [28 , 44] . At the same time , local spindles occurring in the model within deep cortical layers may correspond to the local spindles observed in some studies [45] , or may be even hidden from empirical recordings because of their localized and low amplitude properties . Finally , regional differences in thalamocorical and corticocortical connections could explain the characteristic regional and spatial patterns of spindles observed in human recordings [15] . The correspondence of the matrix vs core thalamocortical system to EEG vs MEG recordings was proposed originally to explain the differences between properties of the EEG and MEG spindles [14 , 46] . Several lines of evidence support this hypothesis and was reviewed by Piantoni et al [15] . MEG and EEG share neural generators , but differences arise due to the biophysics of their cancellation patterns as they project from cortex to sensor . We have recently applied a biophysical forward model of MEG and EEG generation from a large-scale thalamocortical model that is similar to the model used in this study [47] . As hypothesized , in this combined neural/biophysical model , core-dominant spindles were more MEG-weighted than matrix-dominant spindles . Human EEG studies have previously reported the existence of two types of spindles based on frequency—slow and fast spindles ( 9–12 and 12-15Hz ) . In our study , there was small difference ( core-13 . 8 ( subharmonic at 6 . 7Hz ) and matrix-14 . 4 Hz ( subharmonic-7 . 2Hz ) ) in the frequency between core and matrix spindles . However , this difference was much smaller than between fast and slow spindles ( which are 9–12 and 12-15Hz ) reported in vivo . These results are consistent with laminar recordings of spindles in humans [17] , where the average frequency of middle vs upper layer spindles does not differ significantly . Furthermore , in intracranial recordings , both slow and fast spindles occur after down states [35 , 36] , as opposed to the reported occurrence of slow spindles before down states in the EEG [48] . Taken together , these findings suggest that the properties of the thalamocortical and corticothalamic connections explored here are not sufficient to explain the origin of fast vs . slow spindles as observed in rodent studies . The relationship between spindle phase and spike timing of cortical neurons has been examined by previous studies , though their findings have been contradictory . While some studies have shown a preference in phase [45] , others have shown no such preference [49] . In a recently published analysis of spindles recorded in the thalamus and cortex of humans [35] , we found that thalamic spindling appears to drive cortical spindling , including both LFP and high gamma activity . This suggests that the nature of thalamic connections could influence the phase preference of spiking during spindle oscillations . In this new study , we measured the phase of cortical neurons’ spiking during spindles ( S2 Fig ) and we observed that , in our model , neurons both in the core and the matrix systems had a higher preference to spike at the peak of the spindle oscillation ( corresponding to the oscillation phase values close to pi or–pi ) . We also compared the variability of the spiking phase in the matrix pathway versus the core one ( S2C Fig ) . The statistical test comparing the phase of spiking in the core versus matrix systems aggregated for all cortical neurons was not significant ( two sample KS test of phase distribution , KS statistic = 0 . 12 , p = 0 . 44 ) . However , normalized probability of spiking for the different spindle oscillation phases ( obtained from the normalized histogram binned at 100 intervals ) was significantly different between the core and the matrix for many values of phase . This suggests that the phase of spiking in the matrix pathway has a trend for being more variability than in the core pathway . Increase in spindle density following learning is a robust experimental finding that suggests a role for sleep in memory consolidation [2 , 6 , 7 , 10 , 50 , 51] . However , the neural mechanisms that increase spindle density after learning are not known . The hippocampal CA1 region projects to both superficial and deep layers of the rodent prefrontal cortex [52 , 53] . In addition , experiments with simultaneous recordings from cortex and hippocampus report that during NREM sleep , the two structures show coordinated activity [54 , 55] , underlying spike sequence replay and the reactivation of memories that were recently learned while awake [56] . In our study , we found that activation of layer 3/4 of the neocortex triggers spindles that propagate between the core and matrix systems and eventually lead to spindle recruitment in wide regions of the neocortex . Based on these findings , we predict that hippocampal input to superficial cortical layers ( layer 2/3/4 ) during NREM sleep can induce local activation and spindles in the core system , which then propagate to the matrix system thereby activating large cortical regions in both layers , potentially contributign to memory consolidation . Elevated spindle density may arise due to the changes in the cortical microcircuit , of which excitatory interlaminar connections form the main component [57] . This circuit is implicated as a site of sensory coding and learning [58] . In this study , we identified that connection strength from the core to the matrix , but not vice verse , was critical in determining spindle density . This predicts that the increase in spindle density following a hippocampal-dependent task may arise from the strengthening of feedforward projections from middle to superficial cortical layers . In sum , our study identified a rich set of the local and global network mechanisms involved in the propagation and interactions of spindles across different cortical structures . While spindle activity in the model arises from thalamic circuits , our study supports the idea that thalamocortical and intracortical projections significantly shape the properties of spindling activity and that this may explain the characteristic changes of spindle density associated with sleep-related memory replay .
The human research reported in this study was approved by the institutional review board at Partners Healthcare Network . Written informed consent was directly obtained from all subjects prior to their participation . Extracranial electromagnetic fields were recorded in 4 healthy adults ( 3 female ) . Subjects did not report any neurological problems including sleep disorders , epilepsy , or substance dependence . In addition , subjects did not consume caffeine or alcohol on the day of recording . A whole-head MEG system with integrated EEG cap ( Elekta Neuromag ) was used to collect 204 planar gradiometers and 60 EEG channels . EEG data were referenced to an averaged mastoid . Additional details concerning data collection can be found in [46] . Sleep staging was performed by three neurologists according to standard criteria ( Rechtschaffen and Kales , 1968 ) . Data analyzed came from a 17 . 5 ± 3 . 4 ( mean ± SD ) minute period of stage 2 sleep . Data were acquired at 603 . 107 Hz . Gross artifacts and bad channels were excluded manually . The continuous data were band-pass filtered to between 0 . 1 and 30 Hz and ICA ( Delorme and Makeig , 2004 ) was used to remove the ECG component . Spindles were automatically detected in each MEG and EEG channel using a method modified from [45] . The 10–16 Hz analytic signal was extracted from the data using the Hilbert transform and the envelope obtained by computing its elementwise modulus . The spindle-band envelope was smoothed with a Gaussian kernel ( 300 ms width , 40 ms σ ) . Putative spindles were initially marked as contiguous regions of the smoothed spindle-band envelope where the envelope amplitude was more than 2 standard deviations above the mean . Marked regions were then expanded until amplitude dropped below 1 standard deviation above the mean . Putative spindles shorter than 500 ms and longer than 2 s were excluded from further analysis . Inter-spindle intervals ( ISIs ) were computed from spindle center to center . Outlying ISIs longer than 20 seconds were excluded , as these are likely caused by false negatives in spindle detection . ISIs from all subjects and all channels were pooled together to form a single distribution for EEG and gradiometer data , respectively . | The density of sleep spindles has been shown to correlate with memory consolidation . Sleep spindles occur more often in human MEG than EEG recordings . We developed a thalamocortical network model that is capable of spontaneous generation of spindles across cortical layers and that captures the essential statistical features of spindles observed empirically . Our study predicts that differences in thalamocortical connectivity , known from anatomical studies , are sufficient to explain the differences in the spindle properties between EEG and MEG which are observed in human recordings . Furthermore , our model predicts that intracortical connectivity between cortical layers , a property influenced by sleep preceding learning , increases spindle density . Results from our study highlight the role of intracortical and thalamocortical projections on the occurrence and properties of spindles . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"learning",
"medicine",
"and",
"health",
"sciences",
"sleep",
"brain",
"electrophysiology",
"brain",
"electrophysiology",
"social",
"sciences",
"neuroscience",
"learning",
"and",
"memory",
"physiological",
"processes",
"clinical",
"medicine",
"cognitive",
"psychology",
"b... | 2018 | Thalamocortical and intracortical laminar connectivity determines sleep spindle properties |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.