A Varroa destructor protein atlas reveals molecular underpinnings of developmental transitions and sexual differentiation
2017; Elsevier BV; Volume: 16; Issue: 12 Linguagem: Inglês
10.1074/mcp.ra117.000104
ISSN1535-9484
AutoresAlison McAfee, Queenie W. T. Chan, Jay D. Evans, Leonard J. Foster,
Tópico(s)Plant and animal studies
ResumoVarroa destructor is the most economically damaging honey bee pest, weakening colonies by simultaneously parasitizing bees and transmitting harmful viruses. Despite these impacts on honey bee health, surprisingly little is known about its fundamental molecular biology. Here, we present a Varroa protein atlas crossing all major developmental stages (egg, protonymph, deutonymph, and adult) for both male and female mites as a web-based interactive tool (http://foster.nce.ubc.ca/varroa/index.html). We used intensity-based label-free quantitation to find 1,433 differentially expressed proteins across developmental stages. Enzymes for processing carbohydrates and amino acids were among many of these differences as well as proteins involved in cuticle formation. Lipid transport involving vitellogenin was the most significantly enriched biological process in the foundress (reproductive female) and young mites. In addition, we found that 101 proteins were sexually regulated and functional enrichment analysis suggests that chromatin remodeling may be a key feature of sex determination. In a proteogenomic effort, we identified 519 protein-coding regions, 301 of which were supported by two or more peptides and 169 of which were differentially expressed. Overall, this work provides a first-of-its-kind interrogation of the patterns of protein expression that govern the Varroa life cycle and the tools we have developed will support further research on this threatening honey bee pest. Varroa destructor is the most economically damaging honey bee pest, weakening colonies by simultaneously parasitizing bees and transmitting harmful viruses. Despite these impacts on honey bee health, surprisingly little is known about its fundamental molecular biology. Here, we present a Varroa protein atlas crossing all major developmental stages (egg, protonymph, deutonymph, and adult) for both male and female mites as a web-based interactive tool (http://foster.nce.ubc.ca/varroa/index.html). We used intensity-based label-free quantitation to find 1,433 differentially expressed proteins across developmental stages. Enzymes for processing carbohydrates and amino acids were among many of these differences as well as proteins involved in cuticle formation. Lipid transport involving vitellogenin was the most significantly enriched biological process in the foundress (reproductive female) and young mites. In addition, we found that 101 proteins were sexually regulated and functional enrichment analysis suggests that chromatin remodeling may be a key feature of sex determination. In a proteogenomic effort, we identified 519 protein-coding regions, 301 of which were supported by two or more peptides and 169 of which were differentially expressed. Overall, this work provides a first-of-its-kind interrogation of the patterns of protein expression that govern the Varroa life cycle and the tools we have developed will support further research on this threatening honey bee pest. The Varroa destructor mite is the most devastating pest for Western honey bees (Apis mellifera) (1.Cornman S.R. Schatz M.C. Johnston S.J. Chen Y.P. Pettis J. Hunt G. Bourgeois L. Elsik C. Anderson D. Grozinger C.M. Evans J.D. Genomic survey of the ectoparasitic mite Varroa destructor, a major pest of the honey bee Apis mellifera.BMC Genomics. 2010; 11: 602Crossref PubMed Scopus (99) Google Scholar, 2.Le Conte Y. Ellis M. Ritter W. Varroa mites and honey bee health: Can Varroa explain part of the colony losses?.Apidologie. 2010; 41: 353-363Crossref Scopus (401) Google Scholar, 3.Dietemann V. Pflugfelder J. Anderson D. Charrière J.-D. Chejanovsky N. Dainat B. de Miranda J. Delaplane K. Dillier F.-X. Fuch S. Gallmann P. Gauthier L. Imdorf A. Koeniger N. Kralj J. Meikle W. Pettis J. Rosenkranz P. Sammataro D. Smith D. Yañez O. Neumann P. Varroa destructor: Research avenues towards sustainable control.J. Apicultural Res. 2012; 51: 125-132Crossref Scopus (112) Google Scholar). This obligate parasite feeds on honey bee hemolymph (blood), simultaneously weakening its host, suppressing the innate immune system, and transmitting debilitating viruses (see Rosenkranz et al. (4.Rosenkranz P. Aumeier P. Ziegelmann B. Biology and control of Varroa destructor.J. Invertebrate Pathol. 2010; 103: S96-S119Crossref PubMed Scopus (1000) Google Scholar) for a comprehensive review on Varroa biology). Varroa's natural host is the Eastern honey bee (A. cerana), and millions of years of coevolution have led A. cerana to develop various tolerance mechanisms, thereby minimizing the mite's negative impact on these colonies (5.Rath W. Co-adaptation of Apis cerana Fabr., and Varroa jacobsoni Oud.Apidologie. 1999; 30: 97-110Crossref Scopus (118) Google Scholar, 6.Lin Z. Page P. Li L. Qin Y. Zhang Y. Hu F. Neumann P. Zheng H. Dietemann V. Go east for better honey bee health: Apis cerana is faster at hygienic behavior than A. mellifera.PLoS ONE. 2016; 11: e0162647PubMed Google Scholar, 7.Boot W.J. Calis J.N. Beetsma J. Hai D.M. Lan N.K. Toan T.V. Trung L.Q. Minh N.H. Natural selection of Varroa jacobsoni explains the different reproductive strategies in colonies of Apis ceranaApis mellifera.Experiment. Appl. Acarol. 1999; 23: 133-144Crossref Scopus (30) Google Scholar). However, in the mid-1900s, the mite jumped hosts to A. mellifera—the bee species that is most commonly used for active crop pollination today—which is less effective at defending itself (4.Rosenkranz P. Aumeier P. Ziegelmann B. Biology and control of Varroa destructor.J. Invertebrate Pathol. 2010; 103: S96-S119Crossref PubMed Scopus (1000) Google Scholar, 6.Lin Z. Page P. Li L. Qin Y. Zhang Y. Hu F. Neumann P. Zheng H. Dietemann V. Go east for better honey bee health: Apis cerana is faster at hygienic behavior than A. mellifera.PLoS ONE. 2016; 11: e0162647PubMed Google Scholar). Managed A. mellifera colonies infested with Varroa have shorter lifespans than uninfested colonies unless they are actively treated with miticides (8.Fries I. Imdorf A. Rosenkranz P. Survival of mite infested (Varroa destructor) honey bee (Apis mellifera) colonies in a Nordic climate.Apidologie. 2006; 37: 564-570Crossref Scopus (141) Google Scholar, 9.Korpela S. Aarhus A. Fries I. Hansen H. Varroa jacobsoni Oud. in cold climates: Population growth, winter mortality and influence on the survival of honey bee colonies.J. Apicultural Res. 1992; 31: 157-164Crossref Scopus (70) Google Scholar), causing serious negative economic impacts (10.Currie R. Gatien P. Timing acaricide treatments to prevent Varroa destructor (Acari: Varroidae) from causing economic damage to honey bee colonies.The Canadian Entomologist. 2006; 138: 238-252Crossref Google Scholar, 11.Vanengelsdorp D. Meixner M.D. A historical review of managed honey bee populations in Europe and the United States and the factors that may affect them.J. Invertebrate Pathol. 2010; 103: S80-S95Crossref PubMed Scopus (789) Google Scholar, 12.Guzman-Novoa E. Eccles L. Calvete Y. McGowan J. Kelly P.G. Correa-Benitez A. Varroa destructor is the main culprit for the death and reduced populations of overwintered honey bee (Apis mellifera) colonies in Ontario, Canada.Apidologie. 2010; 41: 443-450Crossref Scopus (277) Google Scholar). Despite being responsible for significant colony losses, very little is known about the molecular biology of the Varroa mite. Since the egg, protonymph, and deutonymph life stages (Fig. 1) only exist when the foundress mite (reproductive female) is actively reproducing within capped honey bee brood comb (4.Rosenkranz P. Aumeier P. Ziegelmann B. Biology and control of Varroa destructor.J. Invertebrate Pathol. 2010; 103: S96-S119Crossref PubMed Scopus (1000) Google Scholar), they are seldom observed and are tedious to sample. Furthermore, male mites (even as adults) die soon after the adult honey bee emerges, so even though they are obviously important factors in mite reproduction, our knowledge of their basic molecular biology is extremely limited. Research on Varroa has focused on its role as a vector for viruses (13.Francis R.M. Nielsen S.L. Kryger P. Varroa-virus interaction in collapsing honey bee colonies.PLoS ONE. 2013; 8: e57540Crossref PubMed Scopus (209) Google Scholar, 14.Levin S. Sela N. Chejanovsky N. Two novel viruses associated with the Apis mellifera pathogenic mite Varroa destructor.Sci. Rep. 2016; 6: 37710Crossref PubMed Scopus (37) Google Scholar, 15.Erban T. Harant K. Hubalek M. Vitamvas P. Kamler M. Poltronieri P. Tyl J. Markovic M. Titera D. In-depth proteomic analysis of Varroa destructor: Detection of DWV-complex, ABPV, VdMLV and honeybee proteins in the mite.Sci. Rep. 2015; 5: 13907Crossref PubMed Scopus (36) Google Scholar, 16.Di Prisco G. Pennacchio F. Caprio E. Boncristiani Jr., H.F. Evans J.D. Chen Y. Varroa destructor is an effective vector of Israeli acute paralysis virus in the honeybee, Apis mellifera.J. Gen. Virol. 2011; 92: 151-155Crossref PubMed Scopus (174) Google Scholar, 17.Gisder S. Aumeier P. Genersch E. Deformed wing virus: Replication and viral load in mites (Varroa destructor).J. Gen. Virol. 2009; 90: 463-467Crossref PubMed Scopus (186) Google Scholar, 18.Ryabov E.V. Wood G.R. Fannon J.M. Moore J.D. Bull J.C. Chandler D. Mead A. Burroughs N. Evans D.J. A virulent strain of deformed wing virus (DWV) of honeybees (Apis mellifera) prevails after Varroa destructor-mediated, or in vitro, transmission.PLoS Pathog. 2014; 10: e1004230Crossref PubMed Scopus (230) Google Scholar), their response to pheromone cues (19.Ziegelmann B. Lindenmayer A. Steidle J. Rosenkranz P. The mating behavior of Varroa destructor is triggered by a female sex pheromone.Apidologie. 2013; 44: 314-323Crossref Scopus (19) Google Scholar, 20.Del Piccolo F. Nazzi F. Della Vedova G. Milani N. Selection of Apis mellifera workers by the parasitic mite Varroa destructor using host cuticular hydrocarbons.Parasitology. 2010; 137: 967-973Crossref PubMed Scopus (49) Google Scholar, 21.Nazzi F. Le Conte Y. Ecology of Varroa destructor, the major ectoparasite of the western honey bee, Apis mellifera.Annu. Rev. Entomol. 2016; 61: 417-432Crossref PubMed Scopus (194) Google Scholar), attempts to control it via RNAi (22.Garbian Y. Maori E. Kalev H. Shafir S. Sela I. Bidirectional transfer of RNAi between honey bee and Varroa destructorVarroa gene silencing reduces Varroa population.PLoS Pathog. 2012; 8: e1003035Crossref PubMed Scopus (92) Google Scholar, 23.Campbell E.M. Budge G.E. Bowman A.S. Gene-knockdown in the honey bee mite Varroa destructor by a non-invasive approach: Studies on a glutathione S-transferase.Parasit. Vectors. 2010; 3: 73Crossref PubMed Scopus (65) Google Scholar, 24.Campbell E.M. Budge G.E. Watkins M. Bowman A.S. Transcriptome analysis of the synganglion from the honey bee mite, Varroa destructor and RNAi knockdown of neural peptide targets.Insect Biochem. Mol. Biol. 2016; 70: 116-126Crossref PubMed Scopus (34) Google Scholar), and host shifts (25.Andino G.K. Gribskov M. Anderson D.L. Evans J.D. Hunt G.J. Differential gene expression in Varroa jacobsoni mites following a host shift to European honey bees (Apis mellifera).BMC Genomics. 2016; 17: 926Crossref PubMed Scopus (12) Google Scholar). At the time of writing, there have only been two previous Varroa proteomic investigations, one of which focused on viral proteins (15.Erban T. Harant K. Hubalek M. Vitamvas P. Kamler M. Poltronieri P. Tyl J. Markovic M. Titera D. In-depth proteomic analysis of Varroa destructor: Detection of DWV-complex, ABPV, VdMLV and honeybee proteins in the mite.Sci. Rep. 2015; 5: 13907Crossref PubMed Scopus (36) Google Scholar) and the other identifying fewer than 700 proteins within one developmental stage (26.Surlis C. Carolan J.C. Coffey M.F. Kavanagh K. Proteomic analysis of Bayvarol® resistance mechanisms in the honey bee parasite Varroa destructor.J. Apicultural Res. 2016; 55: 49-64Crossref Scopus (7) Google Scholar). Global protein expression changes associated with developmental transitions and sexual differentiation are yet unknown. The Varroa genome was first sequenced in 2010 (1.Cornman S.R. Schatz M.C. Johnston S.J. Chen Y.P. Pettis J. Hunt G. Bourgeois L. Elsik C. Anderson D. Grozinger C.M. Evans J.D. Genomic survey of the ectoparasitic mite Varroa destructor, a major pest of the honey bee Apis mellifera.BMC Genomics. 2010; 11: 602Crossref PubMed Scopus (99) Google Scholar) and was accompanied by a provisional gene annotation that will be updated shortly. Gene annotations are living databases and, particularly with newly sequenced species, they undergo continuous refinement as more omic data become available. Unfortunately, the more evolutionarily distant a species is from well-annotated species typically used for orthology delineation and gene prediction training sets, the less accurate the predictions become. Such is the case for Varroa. Proteogenomics (27.McAfee A. Foster L.J. Proteogenomics: Recycling public data to Improve genome annotations.Methods Enzymol. 2017; 585: 217-243Crossref PubMed Scopus (5) Google Scholar, 28.Nesvizhskii A.I. Proteogenomics: Concepts, applications and computational strategies.Nat. Methods. 2014; 11: 1114-1125Crossref PubMed Scopus (0) Google Scholar) can help overcome this problem by sequencing the expressed protein regions in a relatively unbiased survey of the genomic landscape. Since protein expression is dynamic throughout an organism's life cycle, high-resolution omics data that cross developmental stages and sexes are very well-suited for this purpose. Investigating global protein expression profiles throughout development of both sexes simultaneously provides a foundational understanding of Varroa biology and creates an opportunity to improve upon existing gene annotations. We present here the first Varroa proteome crossing all major developmental stages (egg, protonymph, deutonymph, adult) of both males and females, where distinguishable (Fig. 1). Through a proteogenomics effort, we identified 519 new protein-coding regions—301 of which are supported by two or more peptides. We also analyze the chemical properties of these sequences and their sequence similarity to other organisms to investigate reasons why underannotation continues to be a problem. We identified 3,102 proteins overall, nearly half (1,433) of which were significantly differentially expressed through development and 101 of which were differentially expressed between sexes. Functional enrichment suggested that carbohydrate and amino acid metabolism underpin developmental transitions, so we investigated proteins involved in glycolysis and the Krebs cycle in detail. Cuticle formation is clearly a process associated with mite aging, and closer analysis suggests the mites utilize different chitin structural proteins as they mature. In addition, chromatin remodeling and positive regulation of transcription may be key factors in sexual differentiation. Building on our previous honey bee protein atlas (29.Chan Q.W. Chan M.Y. Logan M. Fang Y. Higo H. Foster L.J. Honey bee protein atlas at organ-level resolution.Genome Res. 2013; 23: 1951-1960Crossref PubMed Scopus (42) Google Scholar), we provide a web-based interactive platform (http://foster.nce.ubc.ca/varroa/index.html) where researchers can query proteins for visual displays of expression patterns, enabling further hypothesis generation and maximizing the utility of this information for the scientific community. Varroa mite families were collected from a single A. mellifera colony in the fall of 2016 in Vancouver, Canada. In a large-scale population genomics study, the authors found that the genetic variation of Varroa within colonies accounted for by far the largest fraction of genetic variation compared with between colonies and between apiaries (30.Dynes T.L. De Roode J.C. Lyons J.I. Berry J.A. Delaplane K.S. Brosi B.J. Fine scale population genetic structure of Varroa destructor, an ectoparasitic mite of the honey bee (Apis mellifera).Apidologie. 2016; 2016: 1-9PubMed Google Scholar); therefore, sampling mites from a single colony was sufficient. Eggs, foundresses, adult daughters, and adult sons were transferred directly to microfuge tubes using a soft paintbrush, whereas protonymphs and deutonymphs were transferred to a Petri dish and sorted under a dissecting microscope according to the identification guides available at http://idtools.org/id/mites/beemites and http://extension.msstate.edu/publications (publication number: P2826) via the University of Michigan and the Mississippi State University, respectively. Approximately 50 individuals were pooled for each replicate (seven developmental stages, n = 3 for each stage). All samples were immediately frozen at −72 °C until protein extraction. Protein was extracted by homogenizing each mite stage with ceramic beads as previously described (31.McAfee A. Collins T.F. Madilao L.L. Foster L.J. Odorant cues linked to social immunity induce lateralized antenna stimulation in honey bees (Apis mellifera L.).bioRxiv. 2016; Google Scholar). Clarified lysate was precipitated overnight with four volumes of 100% ice cold acetone, and the pellet was washed twice with ice cold 80% acetone. After allowing residual acetone to evaporate (∼15 min), the protein pellet was solubilized in urea buffer (6 m urea, 2 m thiourea in 10 mm HEPES, pH 8) and ∼30 μg (determined via the Bradford Assay) was reduced, alkylated, and digested with Lys-C then trypsin as previously described (32.Foster L.J. De Hoog C.L. Mann M. Unbiased quantitative proteomics of lipid rafts reveals high specificity for signaling factors.Proc. Natl. Acad. Sci. U.S.A. 2003; 100: 5813-5818Crossref PubMed Scopus (729) Google Scholar). Peptides were acidified (one volume 1% TFA), desalted on a high capacity C18 STAGE tip (33.Rappsilber J. Ishihama Y. Mann M. Stop and go extraction tips for matrix-assisted laser desorption/ionization, nanoelectrospray, and LC/MS sample pretreatment in proteomics.Anal. Chem. 2003; 75: 663-670Crossref PubMed Scopus (1795) Google Scholar), solubilized in Buffer A (0.1% formic acid), and quantified in technical triplicate using a peptide fluorometric assay (Pierce; cat: 23290). 2 μg of peptides per sample were analyzed on an EasynLC-1000 chromatography system (Thermo) coupled to a Bruker Impact II Q-TOF mass spectrometer. The LC C18 columns included a fritted trap column and pulled-tip, 50-cm analytical column produced and packed in-house (34.McAfee A. Harpur B.A. Michaud S. Beavis R.C. Kent C.F. Zayed A. Foster L.J. Toward an upgraded honey bee (Apis mellifera L.) genome annotation using proteogenomics.J. Proteome Res. 2016; 15: 411-421Crossref PubMed Scopus (16) Google Scholar, 35.Beck S. Michalski A. Raether O. Lubeck M. Kaspar S. Goedecke N. Baessmann C. Hornburg D. Meier F. Paron I. Kulak N.A. Cox J. Mann M. The Impact II, a very high-resolution quadrupole time-of-flight instrument (QTOF) for deep shotgun proteomics.Mol. Cell. Proteomics. 2015; 14: 2014-2029Abstract Full Text Full Text PDF PubMed Scopus (121) Google Scholar). Peptides were separated using a 165-min linear gradient of increasing Buffer B as specified in the LCParms.txt file embedded within the Bruker data folders (available at www.proteomexchange.org, accession: PXD006072). Buffers A and B were 0.1% formic acid and 0.1% formic acid, 80% acetonitrile, respectively. The instrument was set to the same parameters as described in our previous publication under "Analysis of PTMs" (34.McAfee A. Harpur B.A. Michaud S. Beavis R.C. Kent C.F. Zayed A. Foster L.J. Toward an upgraded honey bee (Apis mellifera L.) genome annotation using proteogenomics.J. Proteome Res. 2016; 15: 411-421Crossref PubMed Scopus (16) Google Scholar), except the scanned mass range was 200–2,000 m/z, the top 20 precursors were fragmented at a 5-Hz spectral rate, and the lower precursor intensity threshold was 300 counts. For the proteogenomics analysis, the Varroa spectra were searched against a six-frame translation of the publicly available Varroa genome sequence (PRJNA33465) using MaxQuant (v.1.5.3.30) to identify new protein-coding regions (minimum ORF length was set to 100 amino acids). All viruses known to infect A. mellifera and Varroa were also included in the database. Honey bee proteins were not included after a follow-up sequence similarity analysis indicated that only five of the proteins identified in this search matched to bees. MaxQuant search settings included: trypsin cleavage specificity, two allowed missed cleavages, fixed carbamidomethyl modification, variable oxidated methionine and N-terminal acetylation, 0.07 Da precursor mass tolerance, 35 ppm fragment mass tolerance, and 1% protein and peptide FDR calculated based on reverse hits. The peptide (scores, modifications, precursor mass, and m/z) and protein (protein groups, accessions, number of assigned peptides, unique peptides, and % coverage) identification information contained within the main MaxQuant output files (summary.txt, peptides.txt, proteinGroups.txt, parameters.txt), and the protein database (165,951 entries) are available at PXD006072. We also include protein accessions, numbers of distinct peptides, and percentage protein coverage for each protein group in supplemental Table 5. Annotated spectra are available through MS-viewer (search key: wuh30b9smr). Peptides identified in the six-frame translation search but which were not present in the canonical protein database were used as anchors to retrieve the corresponding ORFs from the genome using a simple Perl script. This yielded 524 new protein-coding sequences. Of these, 301 were flanked by two or more peptides spanning at least 50 amino acid residues. We used a two-way ANOVA (factors: amino acid and new/known sequence origin) to compare amino acid composition between this set of 301 new protein-coding sequences and 902 sequences bounded by known peptides that were identified in the same six-frame translation search. We used these 902 sequences, which were also generated by the MaxQuant six-frame translation algorithm, because protein-coding sequences generated by more sophisticated algorithms (as with the canonical Varroa annotation) could generate different sequence properties simply due to the algorithm being different. These 902 sequences, however, were both a product of the six-frame translation and part of the canonical protein database. Next, we used the same approach to compare nucleotide positions within codons (factors: nucleotide position and sequence origin). We also compared the adenine and thymine frequency of the new coding regions, known coding regions, and genome sequences that were broken into 1-kb segments in silico (n = 384,129) using a one-way ANOVA (three levels) with a Tukey HSD (honestly significant difference) post-hoc test. We have included the Perl script modules used in these analyses as Supplemental File 1. To survey these proteins for orthology with other species and to retrieve GO terms, we performed Blast2GO (v4.0) using default parameters. We reasoned that these sequences might have been missed in the Varroa annotation effort if they only share sequence similarity to evolutionarily distant species; therefore, we queried them against the nonredundant protein collection with no taxonomic restrictions. Five sequences showed significant homology to honey bee sequences and were removed from the list of new protein sequences, leaving 519 in total. We searched the mass spectrometry data using the same parameters as above, except label-free quantitation (LFQ) was enabled (with min ratio count = 1), and a composite protein database was used that included all proteins in the most recent Varroa gene annotation1 (the final protein database is included at the ProteomeXchange accession below), the 519 protein sequences identified above, all viral sequences known to infect A. mellifera or Varroa, and all proteins contained within the A. mellifera OGSv3.2 annotation. Since A. mellifera biological material is Varroa's sole food source, we expected to find a substantial number of honey bee proteins within our samples. The final database totaled 32,110 entries and is available at PXD006072, along with the MaxQuant peptide and protein identification information as described under "Proteogenomics." We also include a table with protein accessions, numbers of distinct peptides, percentage protein coverage, and LFQ measurements (supplemental Table 6). Honey bee proteins include an Amel tag in the accession, new protein-coding regions from the six-frame translation include a True or False tag in the accession (indicating the DNA template strand relative to the indicated contig), virus sequences are represented by a single gi number or Uniprot identifier, and all other sequences (excluding contaminants and reverse hits) belong to Varroa. Annotated spectra are available at MS-viewer (search key: msmx6z444s). All seven developmental stages were collected in biological triplicate with ∼50 individuals pooled to create each replicate. Since each stage is a pool of many individuals, even a relatively low replication of n = 3 represents a large sample of the population. Only proteins with six or more observations (out of 21) were included in differential expression analysis across developmental stages. For the differential expression across sexes, this was relaxed to three or more observations to avoid excluding proteins that are stage and sex specific. Visual inspection of log-transformed LFQ intensity histograms confirmed the data for each replicate were distributed normally prior to analyzing with an ANOVA. For the developmental stage analysis, missing values were not imputed because they were too numerous, forcing a bimodal distribution upon imputation for some samples, which violates one assumption of the t test. For the sexual differentiation analysis, which excluded egg and protonymph stages, this was not the case, so missing values were imputed (width = 0.3, downshift = 1.5) to capture sex-specific proteins as previously described (36.Tyanova S. Temu T. Sinitcyn P. Carlson A. Hein M.Y. Geiger T. Mann M. Cox J. The Perseus computational platform for comprehensive analysis of (prote)omics data.Nat. Methods. 2016; 13: 731-740Crossref PubMed Scopus (3529) Google Scholar). Differential expression analysis across developmental stages was performed as previously described (31.McAfee A. Collins T.F. Madilao L.L. Foster L.J. Odorant cues linked to social immunity induce lateralized antenna stimulation in honey bees (Apis mellifera L.).bioRxiv. 2016; Google Scholar) except Perseus v1.5.6.0 was used, missing values were not imputed, and the ANOVA (one factor, seven levels) p values were Benjamini Hochberg-corrected at 5% FDR. For the analysis of sexually regulated genes, the female (deutonymph, adult daughter, and foundress) and male (deutonymph, adult son) samples were pooled as n = 9 and n = 6, respectively. A t test was then performed and subjected to the Benjamini Hochberg correction at 5% FDR. All hierarchical clustering analyses were performed in Perseus using average Euclidian distance (300 clusters, maximum 10 iterations). We performed functional enrichment analysis on two sets of proteins: 1) Varroa proteins that were differentially expressed through development and 2) Varroa proteins that were differentially expressed between sexes. For all protein sets, we retrieved GO terms using Blast2GO (v4.0) with default parameters, first searching against all arthropods, then sequences with missing GO terms were searched again against the entire nonredundant protein collection. GO terms were exported after running the GO-Slim function. We then performed a gene score resampling analysis with ErmineJ v3.0.2 (37.Lee H.K. Braynen W. Keshav K. Pavlidis P. ErmineJ: Tool for functional analysis of gene expression data sets.BMC Bioinformatics. 2005; 6: 269Crossref PubMed Scopus (207) Google Scholar), using log-transformed q values (from the previous differential expression analysis) for protein score. We considered a GO term significantly enriched if the Benjamini Hochberg-corrected gene score resampling p value was less than 0.10. The web-based interactive Varroa protein atlas was built using the framework previously described for the honey bee protein atlas (29.Chan Q.W. Chan M.Y. Logan M. Fang Y. Higo H. Foster L.J. Honey bee protein atlas at organ-level resolution.Genome Res. 2013; 23: 1951-1960Crossref PubMed Scopus (42) Google Scholar). Procuring an accurate protein database is critically important for proteomics applications. The first Varroa draft gene set was published in 2010 (1.Cornman S.R. Schatz M.C. Johnston S.J. Chen Y.P. Pettis J. Hunt G. Bourgeois L. Elsik C. Anderson D. Grozinger C.M. Evans J.D. Genomic survey of the ectoparasitic mite Varroa destructor, a major pest of the honey bee Apis mellifera.BMC Genomics. 2010; 11: 602Crossref PubMed Scopus (99) Google Scholar) along with the initial genome sequence (ADDG00000000.1); however, a new genome build was just released (ADDG00000000.2) with annotation refinement efforts underway.1 A new gene set will soon to be released, and we have made the new protein database provisionally available through ProteomeXchange (PXD006072). To test the accuracy of the new gene set compared with the first draft, we searched our complete Varroa proteomics data against both versions and found that greater than twofold more unique peptides were identified using the refined annotation (Fig. 2A). Overall, we identified nearly 20,000 unique peptides corresponding to 3,102 protein groups at 1% peptide and protein FDR (Fig. 2B) representing the first global survey of Varroa protein expression. To maximize the utility of this information for researchers, we incorporated the quantified proteins into an interactive Varroa protein atlas (http://foster.nce.ubc.ca/varroa/index.html). The atlas features a searchable database of the quantifi
Referência(s)